University of Groningen
Electron transfer properties in the Prussian Blue analogues RbxMn[Fe(CN)6]y·zH2O
Vertelman, Esther
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to
cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2009
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Vertelman, E. J. M. (2009). Electron transfer properties in the Prussian Blue analogues
RbxMn[Fe(CN)6]y·zH2O s.n.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 18-06-2017
Electron Transfer Properties in
the Prussian Blue Analogues
RbxMn[Fe(CN)6]y·zH2O
Esther Jacoba Martha Vertelman
Cover Pictures: Scanning Electron Microscopy images of several samples
(NH4)xMn[Fe(CN)6]y·zH2O,
Rb0.92Mn[Fe(CN)6]0.95·1.03H2O
and
NaxMn[Fe(CN)6]y·zH2O) measured by G.H. ten Brink at the University of
Groningen.
Cover design:
P. Dijkstra.
The work described in this thesis was performed in the group “Molecular Inorganic
Chemistry” of the University of Groningen, The Netherlands, with financial support
provided by the Zernike Institute for Advanced Materials.
Printed by: Wöhrmann Print Service
Zernike Institute for Advanced Materials Ph. D. thesis series 2009-06
ISSN 1570-1530
ISBN 978-90-367-3785-2
RIJKSUNIVERSITEIT GRONINGEN
Electron Transfer Properties in the
Prussian Blue Analogues
RbxMn[Fe(CN)6]y·zH2O
Proefschrift
ter verkrijging van het doctoraat in de
Wiskunde en Natuurwetenschappen
aan de Rijksuniversiteit Groningen
op gezag van de
Rector Magnificus, dr. F. Zwarts,
in het openbaar te verdedigen op
vrijdag 19 juni 2009
om 16.15 uur
door
Esther Jacoba Martha Vertelman
geboren op 24 april 1980
te Purmerend
Promotor:
Prof. dr. B. Hessen
Copromotor:
Dr. P. J. van Koningsbruggen
Beoordelingscommissie:
Prof. dr. A. Bleuzen
Prof. dr. ir. P.H.M. van Loosdrecht
Prof. dr. T.T.M. Palstra
Contents
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science............... 1
1.1 Introduction .................................................................................................... 1
1.2 The Structure of Prussian Blue Analogues .................................................... 2
1.3 Magnetic Properties in Prussian Blue Analogues .......................................... 3
1.4 Other Properties of Prussian Blue Analogues .............................................. 10
1.5 Overview of Thesis ...................................................................................... 12
1.6 References .................................................................................................... 14
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic
Properties of RbxMn[Fe(CN)6]y·zH2O.................................................................... 21
2.1 Introduction .................................................................................................. 21
2.2 Experimental Section ................................................................................... 21
2.3 Results and Discussion................................................................................. 24
2.4 Conclusions .................................................................................................. 43
2.5 Acknowledgements ...................................................................................... 46
2.6 References .................................................................................................... 46
Chapter 3. The Influence of Synthetic Conditions on the Physical Properties of
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues ................................................ 49
3.1 Introduction .................................................................................................. 49
3.2 Experimental Section ................................................................................... 49
3.3 Results and Discussion................................................................................. 51
3.4 Conclusions .................................................................................................. 64
3.5 References .................................................................................................... 65
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in
Derivatives of RbxMn[Fe(CN)6]y·zH2O.................................................................. 67
4.1 Introduction .................................................................................................. 67
4.2 Experimental Section ................................................................................... 67
4.3 Results and Discussion................................................................................. 69
4.4 Conclusions .................................................................................................. 81
4.5 Acknowledgements ...................................................................................... 82
4.6 References .................................................................................................... 82
Chapter 5. Light- and Temperature-Induced Electron Transfer in Single Crystals of
RbMn[Fe(CN)6]·H2O.............................................................................................. 85
5.1 Introduction .................................................................................................. 85
5.2 Experimental Section ................................................................................... 85
5.3 Results and Discussion................................................................................. 86
5.4. Tentative Model Explaining the Presence of Different
Iron/Manganese Sites ......................................................................................... 94
5.5. Conclusions ................................................................................................. 96
5.6 Acknowledgements ...................................................................................... 97
5.7 References .................................................................................................... 97
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O .. 99
6.1 Introduction .................................................................................................. 99
6.2 Experimental Section ................................................................................... 99
6.3 Results and Discussion............................................................................... 100
6.4 Conclusions ................................................................................................ 116
6.5 Acknowledgements .................................................................................... 117
6.6 References .................................................................................................. 117
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of
AxMn[Fe(CN)6]y.zH2O (A = Na+, K+, Cs+, NH4+, N(CH3)4+, Sr2+) ....................... 119
7.1 Introduction ................................................................................................ 119
7.2 Experimental Section ................................................................................. 120
7.3 Results and Discussion............................................................................... 121
7.4 Discussion and Conclusions....................................................................... 129
7.5 References .................................................................................................. 133
Appendix I - X-Ray Powder Diffraction Profiles for Samples in
Chapters 2, 3 and 7 ............................................................................................... 135
Appendix II - Crystallographic Details of the Crystal Structure Described in
Chapter 5 .............................................................................................................. 141
Nederlandse Samenvatting ................................................................................... 147
Dankwoord ........................................................................................................... 153
List of Publications............................................................................................... 155
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 1. The Use of Prussian Blue Analogues in
Modern Day Science
1.1 Introduction
Around 1700 Prussian Blue (PB, FeIII4[FeII(CN)6]3·14H2O) was accidentally
discovered by Diesbach, a painter from Berlin who actually tried to create a red
coloured paint.1 It consists of [FeIII(CN)6] units linked to FeII via bridging cyano
ligands. Both iron ions are octahedrally surrounded. The C-bounded Fe ion has
always 6 cyano ligands, while for the other Fe ion its shell can be completed with
water. The latter occurs when [Fe(CN)6] defects are present. This gives a cubic
structure2 in which several interstitial sites are present. In these interstitial sites
alkali ions or water can be present (see Figure 1.1). The intense blue colour comes
from an interaction of the different charges on the two iron ions. Since its
discovery Prussian Blue has been used by painters as a relatively cheap blue dye
and also the blue colour of blueprints originate from this colour.
b)
a)
Fe
C
N
C
C
N
C
Fe
N
N
C
N
C
N
Fe
C
C
Fe
N
C
N
C
Fe
N
N
N
C
C
Fe
C
C
M
C
N
N
Fe
H
H
H
Fe
M'
M'
O
H
O
C
C
C
H
N
M
M
M
M' H
Fe
Fe
N
N
M'
N
C
N
N
N
Fe
N
C
C
C
C
M'
N
Fe
C
N
N
C
N
N
N
C
M
Fe
Fe
C
N
N
C
Fe
C
C
C
C
N
Fe
N
N
N
Fe
Fe
M
O H
H
O M
H
O
M
M'
H
M'
M
M'
Figure 1.1 a) Part of the structure of Prussian Blue (alkali ions and water molecules in the
interstitial sites are omitted for clarity). b) an [M’(CN)6] defect in Prussian Blue analogues
(solid lines between M and M’ represent a bridging cyanide, alkali ions and water
molecules in the interstitial sites are omitted for clarity)
Although PB has now been known for several centuries, it is still widely used in
chemical and physical research. Especially the Prussian Blue analogues (PBA,
AxM[M’(CN)6]y·zH2O, A = (alkali) cation) in which one of the iron ions or both
have been replaced by a different metal ion receive considerable interest.
This chapter focuses on the use of PBAs in modern day science. Explicitly only
properties of materials with the generic formula AxM[M’(CN)6]y.zH2O in which A+
= (alkali) cation and M, M’ = (d-block) metal ion are described. For reviews on
1
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
polymeric materials in which one or two of the CN ligands have been replaced by a
different ligand or materials with the formula AxM[M’(CN)8]y.zH2O the reader is
referred to literature.3 This chapter starts with the general structure of PBAs
(section 1.2), next the magnetic properties of PBAs are discussed (section 1.3) and
other physical properties (section 1.4). The chapter ends with an overview of the
research described in this thesis.
1.2 The Structure of Prussian Blue Analogues
The stoichiometry of PBAs is represented by the general formula
AxM[M’(CN)6]y·zH2O, with A = (alkali) cation. In the case that M is a divalent
metal ion and M’ a trivalent ion two extreme stoichiometries exist:
AM[M’(CN)6]·H2O (stoichiometric) and M3[M’(CN)6]2·zH2O (i.e. no alkali
cation). Most of the times, though, the stoichiometry lies between these two
extremes.
Generally, the structure of PBAs shows a cubic lattice similar to the lattice of PB
itself (Figure 1.1) where either the C bound Fe ion or the N bound Fe ion or both
have been replaced by a different metal ion.4 Usually the space group is Fm-3m or
F-43m. The difference in the two space groups lies in the precise distribution of the
interstitial ions: in Fm-3m the ions are equally and randomly distributed over the
interstitials sites, whereas in F-43m two different interstitial sites exist (see Figure
1.2). In the latter case, the cation has a preference for one of these sites.
Occasionally, a tetragonal space group is found for PBAs, in which one of the axes
of the cubic space group is elongated with respect to the other two axes.4s,5 This
Figure 1.2 The two different interstitial sites present (represented by grey spheres) in a unit
cell with space group F-43m. Solid lines indicate the (200), (020) and (002) planes and do
not represent a chemical bond.
2
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
occurs mostly when a larger cation such as N(CH3)4+ 4s,5a is incorporated or when a
Jahn-Teller active metal ion such as CuII 5b is present. In special cases, when the
C-M’-C or N-M-N angles are not right angles, a hexagonal space group is found.6
1.3 Magnetic Properties in Prussian Blue Analogues
As is already clear, in PBAs two different transition metal ions can be present.
Each of these metals in turn can have unpaired electrons and thus give rise to a
magnetic moment when placed in a magnetic field. In this way, it is possible to
tune the magnetic properties of PBAs by selectively choosing the specific metal
ions involved without altering the structure too much. Indeed, several reviews on
the magnetic properties of PBAs have already appeared.7
1.3.1 Long-Range (Anti) Ferromagnetic Interactions in Prussian Blue Analogues
The group of Verdaguer et al.8 has explained how magnetic interactions in PBAs
take place: In a AxM[M’(CN)6]y·zH2O system the M’ site is always surrounded by
six C bonded cyanides and is thus always in a low spin state. Only the t2g orbitals
are occupied in this M’ ion (no dn M’ ion is known with n > 6 in [M’(CN)6] ions).
The M ion is surrounded by either 6 N bonded cyanides (when x = 1) or by 4 N
bonded cyanides and 2 water molecules (when x = 0) (or something in between
these two extremes). In either case the M ion is expected to be high spin, although
in borderline cases8,9 low spin behaviour is found. When overlap10 is present
between the electrons in t2g orbitals of the M’ ion and the electrons in eg orbitals of
the M ion, ferromagnetic interactions take place. Conversely, overlap between
electrons in the t2g orbitals of the M’ ion and the electrons in t2g orbitals of the
M ion lead to antiferromagnetic interactions. The larger the energy difference
between the two different t2g orbitals, the larger the antiferromagnetic
interactions.11 Indeed, Verdaguer et al.8 have proved this concept for the
MII3[CrIII(CN)6]2·zH2O series.
In this way, several PBAs are known in which long range ferro- or ferrimagnetism
below a critical temperature, the Curie temperature (Tc), takes place. Some of them
show this ordering only below low temperatures (Tc < 50 K),12 but more elevated
ordering temperatures (50 K < Tc < 300 K) have been found too13 and even several
compounds exist with a Curie temperature above room temperature.14 The highest
Tc found was 376 K for KVII[CrIII(CN)6].2H2O.0.1KCF3SO3.14b As it turns out, the
specific magnetic properties of PBAs are not only dependent on the nature of the
metal(s) and the stoichiometry, but some evidence exists that they might also be
dependent on the specific grain size and shape of the involved material.15
1.3.2 Changes in Magnetisation upon External Stimuli
Several photo-induced magnetic phenomena were found. Elaborate research was
performed on the photo-induced charge transfer in AxCo[Fe(CN)6]y.zH2O (A =
alkali cation) and RbxMn[Fe(CN)6]y.zH2O and these will be explained in more
detail in section 1.3.5 and 1.3.6 respectively.
3
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
Fe[Cr(CN)6]2/3.5H2O shows a decrease in magnetisation below Tc (21 K) when it is
irradiated with 360 – 450 nm light.16 This was explained by a mechanism in which
the ferromagnetic coupling between FeII and CrIII has changed due to the fact that
the spins of both metal ions were not perfectly aligned after the irradiation.
Micelles, which contain CdS and PB, show a decrease in magnetisation after
irradiation with UV light. The magnetisation can be restored after annealing at
room temperature. The entire process can be repeated several times.17
It is also possible to have changes in the magnetisation when applying pressure.
Generally, low spin metal ions occupy a smaller volume than the corresponding
high spin metal ions due to the fact that in the latter case the nonbonding eg orbitals
are occupied. Compounds are known to adapt a different structure which occupy a
smaller volume when pressure is applied. One of the possibilities can be to change
the spin of one of the metal ions from high spin to low spin. Consequently the
magnetisation of the compound will become smaller. For instance, when
K0.4Fe4[Cr(CN)6]2.8.16H2O is subjected to pressure the total magnetisation becomes
less and the Curie temperature shifts to lower values.18 The explanation is possibly
the spin crossover of the FeII ion from high spin to low spin. In the case of
Mn3[Cr(CN)6]2.12H2O applying pressure results in a reduction of the magnetisation
but an increase in Tc.19 By applying pressure to Ni3[Cr(CN)6]2.12H2O only a
reduction in the magnetisation is observed.19
As will be discussed in section 1.4.3 several species can be intercalated in the
interstitial sites, most notably solvent molecules and (alkali) cations. By
incorporation of these molecules the local crystal field around the metal ions
changes to a different crystal field. When the crystal field around the metal ions
change the magnetic properties of these metal ions can change. In the case of
M3[Fe(CN)6]2.xH2O with M = Mn, Co, Ni, Cu removal of the water gives a higher
value of Tc and a reduction of the magnetisation.20 Martinez-Garcia et al. explain
this by proposing a weakening of the π* back donation of the iron atom, leading to
a reduction in the overlap between the metal centres and consequently a reduction
of the ferromagnetic coupling. When the materials CoIIxMnII1-x[CrIII(CN)6]2/3.zH2O
with x = 1 or 0.41 are placed in a humid environment the magnetisation increases
and the Tc rises with increased humidity.21 The explanation is that the CoII ion near
Cr(CN)6 defects and on the surface changes from 4-fold to 6-fold coordination with
increased humidity (by taking up water molecules from the atmosphere that expand
the coordination sphere of the ions) and thus the magnetic interaction between CoII
and CrIII changes from ferromagnetic to antiferromagnetic coupling. The same
principle applies when in this material water molecules are replaced with ethanol
molecules.22 Desolvation of CrII0.5CrIII[VII(CN)6].zMeOH leads to significant
reduction of the Curie temperature from 110 K to 25 K.23 Dehydration and
rehydration of K0.2Mn1.4[Cr(CN)6].6H2O leads to a reversible change of Tc of 66 K
for the hydrated species and 99 K for the dehydrated species.24 Again the change in
4
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
coordination number for the MnII ions near defects is a possible explanation for this
phenomenon.
1.3.3 FeII Spin Crossover in CsFe[Cr(CN)6]
Interestingly, in CsFeII[CrIII(CN)6] the FeII ion undergoes spin crossover from low
spin to high spin at 238 K when heated, while the reverse process happens at 211 K
when cooled, giving it a hysteresis width of 27 K.25 The spin crossover process can
also be induced by X-ray illumination26 and an applied pressure shifts the entire
hysteresis to higher temperatures with unchanged hysteresis width.27
1.3.4 Magnetic Pole Inversion in Prussian Blue Analogues of the Type
AxMaM’1-a[M’’(CN)6]y.zH2O
When one of the metal ions (notably the N-bound metal) is a mixture of two
different types of metal ions several new interesting magnetic properties can arise.
Due to different types of interaction between M-M’’ and M’-M’’ (i.e.
ferromagnetic and antiferromagnetic coupling) magnetic pole inversion can be
observed.28 Magnetic pole inversion occurs when the magnetisation changes from
positive to negative values or vice versa. The extent of the magnetic pole inversion
is dependent on the precise ratio between M and M’.28c Magnetic pole inversion
can be photo-induced28b,d,e and/or pressure induced.28f
1.3.5 Charge Transfer in CoFe PBAs
In 1996 Sato et al. noticed that when iridiating K0.2Co1.4-[Fe(CN)6].6.9H2O with red
light (660 nm) the magnetisation below the Curie temperature was increased and
the Curie temperature increased as well (Figure 1.3). Conversely, when irradiating
with blue light (450 nm) the reverse process took place.29 The process could be
repeated several times.
Figure 1.3. Difference in Field Cooled
magnetisation of K0.2Co1.4[Fe(CN)6].6.9H2O
before and after irradiation with 660 nm
light. Figure taken from reference 29.
It was found that the process involved a reversible charge transfer combined with
low spin to high spin conversion from low spin FeIII and high spin CoII (the high
temperature, HT, phase) to low spin FeII and low spin CoIII (the low temperature,
LT, phase; Figure 1.4).30
5
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
Heating,
660 nm
FeII
CN
e-
CoIII
FeIII
Cooling,
450 nm
CoII
CN
e-
eg
t2g
Figure 1.4. Charge transfer process in AxCo[Fe(CN)6]y.zH2O compounds. The bottom part
of the figure depicts the occupation of the d orbitals by electrons in the metal ions.
Soon it was discovered that also the CoFe PBAs with Na+, Rb+ and Cs+ as an alkali
cation showed this phenomenon.31 However, when no alkali cation was present no
electron transfer takes place,32 nor when too much alkali cation was present.31d,33
The group of Bleuzen and Verdaguer31d,33,34 proposed a plausible mechanism for
these observations. In order to have maximum charge transfer Co and Fe should be
available in stoichiometric amounts, hence the incapability towards charge transfer
for the alkali cation free compound. However, because the metal ions change their
radii upon the charge transfer and the Fe – C and Co – N distances change, a
flexible network is required. This flexible network is provided by the presence of
Fe(CN)6 defects. Indeed, in the vicinity of these defects the bond lengthening
around Co is easier.35 This explains the observation that the compounds with too
much alkali cation do not show electron transfer. The precise maximum amount of
alkali cation for the destruction of the charge transfer depends on the specific
cation. Furthermore, it turns out that the nature of the cation is of importance too:
the most efficient charge transfer compound seems to be with Na+ present.36
Later it was found that a similar charge transfer process takes place under the
influence of temperature:37 when heating the magnetic susceptibility increases
abruptly around a temperature ranging from 225 to 300 K (again, the precise
temperature depends on the specific stoichiometry and the specific alkali cation).
Conversely, when cooling down, the reverse process takes place at a temperature
some 50 K lower, giving the entire process a broad hysteresis width. The precise
transition temperatures depend strongly on the specific stoichiometry of the
compound. At no temperature, however, 100% of FeIII or FeII ions is present.
In 2002 Shimamoto et al.38 found out that within the hysteresis loop the
AxCo[Fe(CN)6]y·zH2O system could be transformed from the LT to the HT phase
6
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
with a one-shot-laser-pulse (6 ns) of 532 nm with at least 45 mJ pulse-1 cm-2 power
density. The possibility of local heating of the sample was discarded, because in a
reference experiment with a laser of 1001 nm, the process did not occur, even when
a power density of 688 mJ pulse-1 cm-2 was applied. Later on the reverse process
was established with a one-shot-laser-pulse (8 ns) of 2.38 eV with a power density
of 9.2 . 1020 photons cm-3.39
Also pressure-induced magnetisation changes are possible: when pressure is
applied (1 – 8 bar) on K0.4Co4[Fe(CN)6]2.7·18H2O, K0.28Co4[Fe(CN)6]2.76·18H2O or
Cs0.7Co4[Fe(CN)6]y·16H2O the compounds gradually change from the HT phase to
the LT phase with increasing pressure.40 With higher pressures the transition
temperatures increased to temperatures around room temperature.
In order to use the Co – Fe PBA system for possible applications in molecular
electronics, it is useful to have thin films of it. Both KxCo[Fe(CN)6]y·zH2O and
RbxCo[Fe(CN)6]y·zH2O films have been prepared, (unfortunately, the precise
stoichiometry is not given by the authors) and the films showed the same electron
transfer properties as the bulk AxCo[Fe(CN)6]y·zH2O compounds.41 Care should be
taken however, in the precise method of fabrication of the film. When the substrate
is not rinsed with water between subsequently dipping in solutions of the starting
materials, the resulting film shows light induced magnetisation. When the substrate
is rinsed with water after each dipping step, the resulting film shows light induced
demagnetisation, in contrast to the behaviour of the bulk material.42 As it turns out,
also the orientation of the film with oriented material towards the magnetic field is
of importance: when oriented parallel the magnetisation increases upon irradiation,
while it decreases when oriented perpendicular to the field.43 However, when the
films are thicker than ~200 nm this is no longer the case.44 Another striking effect
is found for nano particles of RbxCo[Fe(CN)6]y·zH2O: the ordering temperatures
and coercive fields evolve with particle size.45 Almost no charge transfer happens
in particles of sizes smaller than 7 nm. The larger the particle size, the more the
material behaves like the microcrystalline bulk material, in that the Tc and the
amount of light induced magnetisation approaches that in the bulk.
Discrete molecular entities of only 4 Fe and 4 Co ions are also shown to possess
electron transfer capability. Recently, it was found that a ‘cage’ of FeIII4CoII4
converts to FeII4CoIII4 and vice versa under the influence of temperature and light.46
1.3.6 Charge Transfer in MnFe PBAs
A few years after the discovery of the charge transfer in Co – Fe PBA Ohkoshi et
al. found another material displaying a temperature induced change in
magnetisation: RbMn[Fe(CN)6].47 On heating, the magnetic susceptibility changes
abruptly from 3.16 cm3 K mol-1 at 285 K to 4.65 cm3 K mol-1 at 320 K. On cooling,
the reverse process takes place between 245 K and 200 K. The entire process had
an unusual hysteresis width of 73 K (Figure 1.5).
7
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
Figure 1.5. The variation of the
magnetic susceptibility with temperature
for RbMn[Fe(CN)6]. Figure taken from
reference 47
At first it was believed that the process was due to a spin crossover of the MnII ion
from intermediate spin at low temperatures to high spin at high temperatures.47
Later, however, Fe-C and Mn-N distances as found by refining X-ray powder
diffraction profiles and X-ray absorption fine structure experiments indicated that it
was due to a charge transfer between Mn and Fe ions.48 At low temperatures the
system consists of low spin FeII and high spin MnIII (LT phase), whereas at high
temperatures low spin FeIII and high spin MnII are present (HT phase,
Figure 1.6).48,49 Furthermore, the transition is accompanied by a structural phase
change from cubic in the HT phase (F-43m) to tetragonal (I-4m2) in the LT phase,
due to a Jahn-Teller distortion of the MnIII ion. The LT phase orders
ferromagnetically below 12 K.49i,50 Similar to the AxCo[Fe(CN)6]y.zH2O materials
(section 1.3.5) the broadness of the hysteresis and the specific temperatures at
which the temperature induced phase transitions take place is dependent on the
specific stoichiometry.51 In this case though, the most complete charge transfer
takes place when the ratio is close to Rb:Mn:Fe = 1:1:1.
Soon after the observation of temperature induced changes also photo-induced
phase transfer from the LT to the HT phase were realised. This transfer could take
place at 3 K using a one-shot-laser-pulse of 532 nm with at least 9.3 mJ cm-2
pulse-1 power density.50,52 The discovery of the reverse phase transfer from this
photo-induced HT phase back to the LT phase took quite some time: only very
recently it was found that irradiation with 410 nm light gives the reverse phase
transition.53
When pressure is applied to the RbxMn[Fe(CN)6]y.zH2O system in the LT phase a
third phase is observed.54 It has the same electronic state as the LT phase, however,
its structure is now in the P-4n2 space group. Rapidly cooling the HT phase leads
to ‘trapping’ of the HT phase at very low temperatures.55 The trapped phase is
slightly different from the photo-induced HT phase at low temperatures in that it
has a somewhat larger unit cell, but the space group is still F-43m.
8
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Heating,
532 nm
FeII
CN
-
e
MnIII
FeIII
Cooling,
410 nm
MnII
CN
eeg
t2g
Figure 1.6. Charge transfer process in RbxMn[Fe(CN)6]y.zH2O. The bottom half indicates
the occupation of the d orbitals by electrons in the metal ions.
When RbMn[Fe(CN)6] is irradiated with laser light (λ = 1064 nm) the evolving of
532 nm light is observed, indicating second harmonic generation by the material.56
The intensity of the 532 nm light increased with the square of the intensity of the
incident light. Interestingly, the second harmonic generation is five times stronger
in the HT phase than in the LT phase.
In Rb0.82Mn[Fe(CN)6]0.94.H2O Ohkoshi et al.57 found a hysteresis loop in
polarization with electric field, which according to the authors indicate that
ferroelectricity is present. Because Rb0.94Mn[Fe(CN)6]0.98.0.4H2O does not show
this hysteresis loop, the authors related the observed ferroelectricity in the previous
compound to mixing of FeII, FeIII, Fe vacancies, MnII and Jahn-Teller distorted
MnIII.
In contrast to the Co – Fe PBA system, it seems to be difficult to incorporate alkali
cations other than Rb+ in the Mn – Fe PBA system. Some examples exist where a
cesium ion is incorporated in the system,58 (Cs1.78MnII[FeII(CN)6]0.78[FeIII(CN)6]0.22,
Cs1.57MnII[FeII(CN)6]0.57-[FeIII(CN)6]0.43, Cs1.51MnII[FeII(CN)6]0.51[FeIII(CN)6]0.49 and
Cs0.94MnII[FeII(CN)6]0.21[FeIII(CN)6]0.70·0.8H2O) however, in these samples an
excess of inactive FeII(CN)6 needs to be present in order to have an electron
transfer from MnII to FeIII(CN)6 and vice versa. When FeIII(CN)6 exceeds FeII(CN)6
the system is inactive. Also one sample exists which shows electron transfer in
which very little K+ is present,59 but also oxygen and acetate is present and it is
unclear what their role is in the electron transfer process. At room temperature its
molecular formula is believed to be K0.2MnII0.66MnIII1.44-[FeII0.2FeIII0.8(CN)6]
O0.66(CH3COO)1.32].7.6H2O based on elemental analysis and IR spectroscopy. The
authors believe that below 10 K the ratio of FeIII/FeII increases, in contrast to the
process in RbxMn[Fe(CN)6]y.zH2O where this ratio decreases with decreasing
temperature.
9
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
Thin layers or single crystals are easier to handle than powdered samples. In view
of possible applications of this system, it would thus be of importance to have thin
layers or single crystals of RbxMn[Fe(CN)6]y.zH2O. To our knowledge, however,
no thin layers have been prepared successfully. Furthermore, the known single
crystals to date,60 do not exhibit the electron transfer capability.
1.4 Other Properties of Prussian Blue Analogues
Although most research done on PBAs nowadays concerns magnetic properties,
some other properties are investigated too. In this section the literature concerning
zero and negative thermal expansion (section 1.4.1), electrochemical (section 1.4.2)
and intercalation properties (section 1.4.3) is discussed.
1.4.1 Zero and Negative Thermal Expansion in Prussian Blue Analogues
Normally, when a compound is heated, the vibrations of the atoms become more
active and as a result the distance between the atoms increases and the total volume
of the compound becomes larger. This phenomenon is called thermal expansion.
Sometimes though, the volume of the compound remains the same (zero thermal
expansion) or decreases (negative thermal expansion) with increasing temperature.
The reason for the occurrence of these counter-intuitive phenomena may be that a
different, more closed, structure becomes apparent with higher temperatures.
Zero and negative thermal expansion have been found for several PBAs.61 For the
series MII[PtIV(CN)6] with M = Mn, Fe, Co, Ni, Cu, Zn, Cd the smallest negative
expansion was found for M = NiII and the largest for M = CdII.61a,d The differences
in thermal expansion factors lie in the differences in the metal cyanide binding
interactions, and consequently the different energies of the transverse vibrations of
the cyanide bridge, with the most flexible lattices showing the largest negative
thermal expansion.
1.4.2 Prussian Blue Analogues in Electrochemistry
Quite some literature concerning PBAs in electrochemistry is about an electrode
modified with a layer of PBA. The PBA layer is capable of catalysing an oxidation
or reduction of a certain compound at a specific potential. The resulting current at
this voltage is then a direct measure of the concentration of the compound and as
such the modified electrode can be used as a sensor for the compound.
Roughly three ways are distinguished to modify an electrode with a thin film.62
Often the electrode is scanned through a range of voltages (cyclic voltammetry)
while being inserted in a solution of the starting materials.63 The change in
potential can change the solubility of the materials and the PBA will precipitate on
the electrode. It is also possible to start with a reduced or oxidized form of one of
the starting materials in solution and then oxidize or reduce this by applying a
specific voltage with the electrode. By oxidizing or reducing the material the
solubility changes and thus a layer is formed by precipitation.64 Finally, it is
10
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
possible to successively ‘dip’ the electrode in a solution of one of the starting
materials and then in a solution of the other. Repeating this cycle several times will
generate a layer of the PBA.65
The resulting modified electrode then can be characterised by several methods.
Often the voltammogram of the electrode is used as a characterisation,66 X-ray
diffraction67 or the stoichiometry is determined with XPS.68 The voltammogram of
the electrode is very sensitive to the used electrolyte, so care should be taken when
only using the first method as characterisation.69
In this way PBA based sensors have been prepared for the detection of hydrazine,70
hydroxylamine,70b,d,71 hydrogen peroxide,72 ascorbic acid70d,73 and several other
compounds.74 The best results for hydrazine were obtained for a carbon fiber
microelectrode modified with a mixed cobalt and copper hexacyanoferrate layer
which had a detection limit as low as 5.10-7 M.70f In the case of hydroxylamine
good results were found for a glassy carbon electrode modified with hybrid coppercobalt hexacyanoferrate films71 (2.1.10-7 M as a lower detection limit), while for
hydrogen peroxide a glassy carbon electrode modified with PB nano arrays was
best72g (detection limit 1.10-6 M). For ascorbic acid a carbon ceramic electrode
modified with a terbium hexacyanoferrate layer gave very good results with a
detection limit of 2.10-7 M.73b
1.4.3 Intercalation of (Small) Molecules and Ions
When looking at the structure of PBAs one notices that in the interstitial sites space
is available in which several ions or small molecules can be intercalated. The size
of the interstitial sites can be tuned by using different metal ions and this can be
used to selectively remove particles of the right size. Unfortunately, various
definitions of the ease of uptake are in use: either mg/g, milliequivalents or
percentages of the original concentration of the compound to be intercalated.
Combined with the often undetermined stoichiometry of the used compound it is
not always possible to convert these definitions into one another and therefore quite
hard to compare the different types of research with respect to the ‘best’ material to
be used in the intercalation of a specific compound.
It is already known for quite some time that cesium ions perfectly fit in the
interstitial sites and thus PBAs can be used for the removal of radioactive cesium
ions from a solution.75 Most probably in the uptake of Cs+ the original alkali cation
is released (ion exchange), although it is also possible that the divalent metal ions
are replaced in the process.
Recently, the production of energy from the oxidation of hydrogen has received
considerable interest due to its relatively ‘green’ process: only water is released.
The problem at the moment is that the needed hydrogen is so volatile in ambient
conditions that it is not safe to use it as such. Therefore compounds are needed in
which the hydrogen can be stored and later released at the right time. In this respect
11
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
PBAs can also be of interest: several PBAs have been examined with respect to
their capacity in the storage of hydrogen.76
Other particles have been intercalated in PBAs too: smaller molecules like N2, CO2
and H2O 77 or larger like aminopyridines and tryptophan.78 Also several charged
particles have been inserted in PBAs.79 In the case of singly charged cations, like
Cs+, the original alkali cation is replaced or (part of) the divalent metal ions. In the
case of doubly charged particles part of the doubly charged metal ions from the
original PBA structure is replaced by the incoming divalent metal ion.79b,c,g
1.5 Overview of Thesis
This thesis focuses on the charge transfer properties in the Prussian Blue Analogue
RbxMn[Fe(CN)6]y.zH2O. As explained in section 1.3.6 these compounds have two
stable states: at room temperature MnII is high spin (HS) and FeIII is low spin (LS),
comprising the high temperature (HT) phase. On cooling the two metal ions show
an electron transfer resulting in HS MnIII and LS FeII, thus forming the low
temperature (LT) phase. The hysteresis width of this process is broad: as large as
138 K was found for Rb0.64Mn[Fe(CN)6]0.88.1.7H2O.55a At low temperatures the
charge transfer process can be induced by light irradiation by 532 nm (LT Æ HT)
and the reverse by irradiating with 410 nm.
In this thesis two main questions are addressed:
What are the precise requirements for the RbxMn[Fe(CN)6]y·zH2O system to show
the electron transfer process?
To what extent is it possible to change and tune the physical properties of the
RbxMn[Fe(CN)6]y·zH2O system?
1.5.1 Chapter 2 – The Influence of Defects on the Electron Transfer and Magnetic
Properties of RbxMn[Fe(CN)6]y.zH2O
This chapter focuses on acquiring a more detailed insight in the relationship
between structural features and electron transfer capabilities of
RbxMn[Fe(CN)6]y·zH2O analogues. It is shown in which way small variations in
the synthetic procedures may give rise to materials that differ considerably in
composition and their consequential physical behaviour. The obtained materials are
characterised with the use of X-ray powder diffraction, magnetic measurements
and differential scanning calorimetry. Furthermore, IR, Raman, 57Fe Mössbauer
and X-ray photoelectron spectroscopy are exploited as well for characterisation.
1.5.2 Chapter 3 – The Influence of Synthetic Conditions on the Physical Properties
of RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
The precise stoichiometry of the RbxMn[Fe(CN)6]y·zH2O materials is of
tremendous importance in determining its physical properties. The more the ratio
approaches the ideal Rb:Mn:Fe = 1:1:1 formula, the more efficient the electron
12
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
transfer from the high temperature phase (LS FeIII and HS MnII) to the low
temperature phase (LS FeII and HS MnIII) and vice versa. Therefore there is a need
for well-defined reaction conditions that lead to a specific stoichiometry.
Unfortunately, the literature shows that the stoichiometry of Prussian Blue
analogues are not so easy to reproduce. In this chapter the synthetic conditions for
a reproducible stoichiometry are sought. Special attention is given to the influence
of temperature during the synthesis and the addition speed of the MnII containing
solution to the [FeIII(CN)6] containing solution.
1.5.3 Chapter 4 – Identifying the Fraction of Charge Transfer Active Material in
Derivatives of RbxMn[Fe(CN)6]y·zH2O
The physical properties of RbxMn[Fe(CN)6]y·zH2O depend strongly on the actual
sample stoichiometry. In compounds with Fe(CN)6 defects, oxygen atoms of water
molecules coordinate to the Mn ions instead of the nitrogen atoms of the cyanide
ions and the average environment of the manganese ions changes to
Mn(OH2)x(NC)6-x. As a result, the ligand field strength of some of the Mn ions
becomes weaker and consequently the redox potential of the MnIII/MnII couple
changes leading finally to its inability to reduce FeIII(CN)6 to FeII(CN)6. If the water
content is sufficiently high, the redox potential changes to such an extent that the
high temperature phase becomes stable at very low temperatures. This important
role of the incorporated water molecules, which may be coordinated as well as noncoordinated, suggests that the physical properties of these compounds may change
with time, but this occurrence has not been discussed in literature. This chapter
focuses on the way in which the physical properties of RbxMn[Fe(CN)6]y·zH2O
compounds might change in time due to the possibility of water uptake.
Furthermore, it is shown how room temperature 57Fe Mössbauer spectroscopy can
be used in order to predict the fraction of charge transfer active material.
1.5.4 Chapter 5 – Light- and Temperature-Induced Electron Transfer in Single
Crystals of RbMn[Fe(CN)6]·H2O
A further development of the field of Prussian Blue analogues as a component for
molecular electronics would highly benefit from an extended knowledge on the
relationship between structural features and the material’s physical properties,
which ideally requires single crystals. Nevertheless, the number of crystal
structures of Prussian Blue analogues with AM[M’(CN)6]·zH2O stoichiometry
known to date is small and limited to compounds that do not exhibit physical
properties relevant in terms of applications. In this chapter single crystals of
RbMn[Fe(CN)6]·H2O have been obtained which show light- and temperatureinduced switching, in which only 50% of the metal ions participate. A model is
introduced to explain the partial switching behaviour.
13
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
1.5.5 Chapter 6 – Further Characterisation of Single Crystals of
RbMn[Fe(CN)6]·H2O
In chapter 5 single crystals of RbMn[Fe(CN)6]·H2O were introduced. As was clear
from the magnetic measurements, Raman spectroscopy and 57Fe Mössbauer
spectroscopy these single crystals only showed a switching behaviour of 50%. This
is in contrast to the behaviour that was found for powdered samples with a similar
stoichiometry, which show nearly complete switching behaviour. In this chapter
the single crystals are investigated in further detail. Specifically temperature
dependent X-ray powder diffraction, more detailed 57Fe Mössbauer spectroscopy,
X-ray photoemission spectroscopy and electron spin resonance spectroscopy is
used. Furthermore, for the first time for RbxMn[Fe(CN)6]y·zH2O compounds the
thermal conductivity as a function of temperature is investigated.
1.5.6 Chapter 7 – Attempted Synthesis, Characterisation and Physical Properties
of AxMn[Fe(CN)6]y.zH2O Compounds (A = Na+, K+, Cs+, NH4+, N(CH3)4+ and Sr2+)
In general, compounds of the series AxMn[Fe(CN)6]y.zH2O with A = cation, have
mostly been prepared with A = Rb+ and very few samples are known with A =
Cs+.80 For AxCo[Fe(CN)6]y·zH2O all different alkali cations (except Fr+) have been
incorporated and in all these cases it is possible that the electron transfer process
takes place.81 It is therefore surprising that in the case of AxMn[Fe(CN)6]y.zH2O the
only known examples (with electron transfer) are with either Rb+ or with Cs+. In
this chapter it was tried to incorporate cations other than Rb+ into the
AxMn[Fe(CN)6]y.zH2O system and determine their influence on the physical
properties of the system.
1.6 References
1
2
3
4
14
Ludi, A., J.Chem.Edu., 1981, 58, 1013
Herren, F.; Fischer, P.; Ludi, A.; Halg, W., Inorg.Chem., 1980, 19, 956
For reviews on these materials see: a) Cernak, J.; Orendac, M.; Potocnak, I.; Chomic, J.;
Orendacova, A.; Skorsepa, J.; Feher, A., Coord.Chem.Rev., 2002, 224, 51; b) Ohba, M.; Okawa,
H., Coord.Chem.Rev., 2000, 198, 313; c) Stasicka, Z.; Wasielewska, E., Coord.Chem.Rev., 1997,
159, 271; d) Dei, A., Angew.Chem.Int.Ed., 2005, 44, 1160; e) Sato, O., J.Photochem.Photobio.C,
2004, 5, 203; f) Bernhardt, P. V.; Bozoglian, F.; Macpherson, B. P.; Martinez, M.,
Coord.Chem.Rev., 2005, 249, 1902; g) Tanase, S.; Reedijk, J., Coord.Chem.Rev., 2006, 250,
2501; h) Einaga, Y., J.Photochem.Photobi.C, 2006, 7, 69; i) Sato, O.; Tao, J.; Zhang, Y.-Z.,
Angew.Chem.Int.Ed., 2007, 46, 2152
a) Zhou, P.; Xue, D.; Luo, H.; Chen, X., Nano Lett., 2002, 2, 845 ; b) Beall, G. W.; Milligan, W.
O.; Korp, J.; Bernal, I., Inorg.Chem., 1977, 16, 2715; c) Mullica, D. F.; Gage, J. L.; Sappenfield,
E. L., J.Solid State Chem., 1994, 112, 312; d) Ludi, A.; Gudel, H.-U.; Ruegg, M., Inorg.Chem.,
1970, 9, 2224; e) Gravereau, P.; Garnier, E., Acta Cryst., 1984, C40, 1306; f) Cao, M.; Wu, X.;
He, X; Hu, C., Chem.Comm., 2005, 2241; g) Basset Brown, D.; Shriver, D. F., Inorg.Chem.,
1969, 8, 37; h) Shriver, D. F.; Basset Brown, D., Inorg.Chem., 1969, 8, 42; i) Galet, A.; Munoz,
M. C.; Real, J. A., Inorg.Chem., 2006, 45, 4583; j) Loos-Neskovic, C.; Fedoroff, M.; Garnier, E.,
Talanta, 1989, 36, 749; k) Ayers, J. B.; Waggoner, W. H., J.Inorg.Nucl.Chem., 1971, 33, 721; l)
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
5
6
7
8
9
10
11
12
13
14
Farmer, J. M.; Mullica, D. F.; Kautz, J. A., Inorg.Chem.Comm, 1999, 2, 456; m) Mullica, D. F.;
Zielke, J. T.; Sappenfield, E. L., J.Solid State Chem., 1994, 112, 92; n) Beall, G. W.; Milligan,
W. O.; Petrich, J. A.; Swanson, B. I., Inorg.Chem., 1978, 17, 2978; o) Swanson, B. I.,
Inorg.Chem., 1976, 15, 253; p) Mullica, D. F.; Oliver, J. D.; Milligan, W. O.; Hills, F. W.,
Inorg.Nucl.Chem.Lett., 1979, 15, 361; q) Cola, M.; Ganzerli Valentini, M. T.; Borroni, P. A.,
J.Inorg.Nucl.Chem., 1978, 40, 1041; r) Dix, M. F.; Rae, A. D., J.Oil.Col.Chem.Assoc., 1978, 61,
69; s) Ziegler, B.; Witzel, M.; Schwarten, M.; Babel, D., Z.Naturforschung B, 1999, 870; t) Ludi,
A.; Gudel, H.-U., Struct.Bonding, 1973, 14, 1; u) Wojdel, J. C.; Bromley, S. T., J.Phys.Chem.B,
2006, 110, 24294; v) Holmes, S. M.; Whelpley, A. S.; Girolami, G. S., Polyhedron, 2007, 26,
2291; w) Arai, M.; Miyake, M.; Yamada, M., J.Phys.Chem.C, 2008, 112, 1953
a) Witzel, M.; Ziegler, B.; Babel, D., Z.Anorg.Allg.Chem., 2000, 626, 471; b) Kosaka, W.;
Ishihara, T.; Yashiro, H.; Taniguchi, Y.; Hashimoto, K.; Ohkoshi, S.-I., Chem.Lett., 2005, 34,
1278
Zhang, L.-P.; Zhou, X.-J.; Mak, T. C. W.; Tanner, P. A., Polyhedron, 2007, 26, 4019
a) Verdaguer, M.; Galvez, N.; Garde, R.; Desplanches, C., Interface, 2002, 28 ; b) Ohkoshi, S.I.; Hashimoto, K., J.Photochem.Photobio.C, 2001, 2, 71; c) Verdaguer, M.; Girolami, G. S. In:
Magnetism: Molecules to Materials V; Wiley-VCH: Weinheim, Germany, 2005, Ed: Miller, J. S.
and Drillon, M., 283; d) Ruiz, E.; Rodriguez-Fortea, A.; Alvarez, S.; Verdaguer, M.,
Chem.Eur.J., 2005, 11, 2135
Verdaguer, M.; Bleuzen, A.; Marvaud, V.; Vaissermann, J.; Seuleiman, M.; Desplanches, C.;
Scuiller, A.; Train, C.; Garde, R. ; Gelly, G. ; Lomenech, C. ; Rosenman, I. ; Veillet, P.; Cartier,
C.; Villain, F.; Coord. Chem. Rev. 1999, 190-192, 1023
CrII, MnIII and CoII
this is no direct overlap, but rather a communication via the orbitals of the bridging ligands, also
known as superexchange
Goodenough, J.B.; Magnetism and the Chemical Bond; John Wiley & Sons, New York, 1963
a) Zentko, A.; Bokor, M.; Lukacova, M.; Marysko, M.; Mihalik, M.; Mitroova, Z.; Zentkova,
M., Phys.Stat.Sol., 2003, 196, 340; b) Broschova, M.; Mihalik, M.; Kavecansky, V.; Seberini,
M.; Suski, W., Czechoslovak J.Phys., 2002, 52, 325; c) Entley, W. R.; Girolami, G. S.,
Inorg.Chem., 1994, 33, 5165; d) Kirk, M. L.; Rubie, N. D.; Hybl, L., Polyhedron, 2001, 20,
1741; e) Franz, P.; Ambrus, C.; Hauser, A.; Chernyshov, D.; Hostettler, M.; Hauser, J.; Kramer,
K.; Stoeckli-Evans, H.; Pattison, P.; Burgi, H.-B.; Decurtins, S., J.Am.Chem.Soc., 2004, 126,
16472; f) Kumar, A.; Yusuf, S. M.; Keller, L., Phys.Rev.B, 2005, 71, 054414; g) Klenze, R.;
Kanellakopulos, B.; Trageser, G.; Eysel, H. H., J.Chem.Phys., 1980, 72, 5819; h) Ohkoshi, S.-I.;
Hashimoto, K., Chem.Phys.Lett., 1999, 314, 210; i) Kurtz, W.; Babel, D., Solid State Comm,
1983, 48, 277; j) Buschmann, W. E.; Miller, J. S., Inorg.Chem., 2000, 39, 2411; k) Kumar, A.;
Yusuf, S. M., Phys.B, 2005, 362, 278; l) Ng, C. W.; Ding, J.; Chow, P. Y.; Gan, L. M.; Quek, C.
H., J.Appl.Phys., 2000, 87, 6049
a) Dong, W.; Ou-Yang, Y.; Song, H.-B.; Liao, D.-Z.; Jiang, Z.-H.; Yan, S.-P.; Cheng, P.,
Inorg.Chem., 2006, 45, 1168; b) Manuel, E.; Evangelisti, M.; Affronte, M.; Okubo, M.; Train,
C.; Verdaguer, M., Cond.Mat., 2006, 0602589; c) Manuel, E.; Evangelisti, M.; Affronte, M.;
Okubo, M.; Train, C.; Verdaguer, M., Phys.Rev.B, 2006, 73, 172406; d) Sato, O.; Iyoda, T.;
Fujishima, A.; Hashimoto, K., Science, 1996, 271, 49; e) Lukacova, M.; Kiss, L. F.; Marysko,
M.; Mihalik, M.; Mitroova, Z.; Stopka, J.; Zentko, A.; Zentkova, M., Phys.Stat.Sol A, 2003, 196,
240
a) Dujardin, E.; Ferlay, S.; Phan, X.; Desplanches, C.; Cartier di Moulin, C.; Sainctavit, P.;
Baudelet, F.; Dartyge, E.; Veillet, P.; Verdaguer, M., J.Am.Chem.Soc., 1998, 120, 11347 ; b)
Holmes, S. M.; Girolami, G. S., J.Am.Chem.Soc., 1999, 121, 5593; c) Hatlevik, O.; Buschmann,
E.; Zhang, J.; Manson, J. L.; Miller, J. S., Advanced Mater., 1999, 11, 914; d) Ferlay, S.; Mallah,
15
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
16
T.; Ouahes, R.; Veillet, P.; Verdaguer, M., Nature, 1995, 378, 701; e) Ohkoshi, S.-I.; Mizuno,
M.; Hung, G.-J.; Hashimoto, K., J.Phys.Chem.B, 2000, 104, 9365
a) Culp, J. T.; Park, J.-H.; Benitez, I. O.; Meisel, M. W.; Talham, D. R., Polyhedron, 2003, 22,
2125; b) Tozawa, M.; Ohkoshi, S.-I.; Kojima, N.; Hashimoto, K., Chem.Comm., 2003, 1204; c)
Sun, H.-L.; Shi, H.; Zhao, F.; Qi, L.; Gao, S., Chem.Comm., 2005, 4339; d) Catala, l.; Gloter, A.;
Stephan, O.; Rogez, G.; Mallah, T., Chem.Comm., 2006, 1018; e) Guari, Y.; Larionova, J.;
Molvinger, K.; Folch, B.; Guerin, C., Chem.Comm., 2006, 2613; f) Catala, l.; Gacoin, T.; Boilot,
J.-P.; Riviere, E.; Paulsen, C.; Lhotel, E.; Mallah, T., Advanced Mater., 2003, 15, 826
Nuida, T.; Hozumi, T.; Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I., J.Solid State Electrochem.,
2007, 11, 763
Taguchi, M.; Yagi, I.; Iyoda, T.; Einaga, Y., J.Am.Chem.Soc., 2006, 128, 10978
Coronado, E.; Gimenez-Lopez, M. C.; Levchenko, G; Romero, F. M.; Garcia-Baonza, V.;
Milner, A.; Paz-Pasternak, M., J.Am.Chem.Soc., 2005, 127, 4580
Zentkova, M.; Arnold, Z.; Kamarad, J.; Kavecansky, V.; Lukacova, M.; Mat'as, S.; Mihalik, M.;
Mitroova, Z.; Zentko, A., J.Phys.:Condens.Matter, 2007, 19, 266217
a) Martinez-Garcia, R.; Knobel, M.; Reguera, E., J.Phys.:Condens.Matter, 2006, 18, 11243; b)
Martinez-Garcia, R.; Knobel, M.; Reguera, E., J.Phys.Chem.B, 2006, 110, 7296
Ohkoshi, S.-I.; Arai, K.-I.; Sato, Y.; Hashimoto, K., Nature Mater., 2004, 3, 857
Sato, Y.; Ohkoshi, S.-I.; Arai, K.-I.; Tozawa, M.; Hashimoto, K., J.Am.Chem.Soc., 2003, 125,
14590
Nelson, K. J.; Giles, I. D.; Troff, S. A.; Arif, A. M.; Miller, J. S., Inorg.Chem., 2006, 45, 8922
Lu, Z.; Wang, X.; Liu, Z.; Liao, F.; Gao, S.; Xiong, R.; Ma, H.; Zhang, D.; Zhu, D.,
Inorg.Chem., 2006, 45, 999
a) Kosaka, W.; Nomura, K.; Hashimoto, K.; Ohkoshi, S.-I., J.Am.Chem.Soc., 2005, 127, 8590;
b) Le Guennic, B.; Borshch, S. A.; Robert, V., Inorg.Chem., 2007, 46, 11106
Papanikolau, D.; Margadonna, S.; Kosaka, W.; Ohkoshi, S.-I.; Brunelli, M.; Prassides, K.,
J.Am.Chem.Soc., 2006, 128, 8385
Papanikolau, D.; Kosaka, W.; Margadonna, S.; Kagi, H.; Ohkoshi, S.-I.; Prassides, K.,
J.Phys.Chem.C, 2007, 111, 8086
a) Ohkoshi, S.-I.; Hozumi, T.; Utsunomiya, M.; Abe, M.; Hashimoto, K., Phys.B, 2003, 329333, 691; b) Ohkoshi, S.-I.; Hashimoto, K., J.Am.Chem.Soc., 1999, 121, 10591; c) Ohkoshi, S.I.; Iyoda, T.; Fujishima, A.; Hashimoto, K., Phys.Rev.B, 1997, 56, 11642; d) Ohkoshi, S.-I.;
Hashimoto, K., Interface, 2002, 34; e) Ohkoshi, S.-I.; Yorozu, S.; Sato, O.; Iyoda, T.; Fujishima,
A.; Hashimoto, K., Appl.Phys.Lett., 1997, 70, 1040; f) Egan, L.; Kamenev, K.; Papanikolau, D.;
Takabayashi, Y.; Margadonna, S., J.Am.Chem.Soc., 2006, 128, 6034; g) Ohkoshi, S.-I.;
Fujishima, A.; Hashimoto, K., J.Am.Chem.Soc., 1998, 120, 5349; h) Nuida, T.; Hozumi, T.;
Kosaka, W.; Sakurai, S.; Ikeda, S.; Matsuda, T.; Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I.,
Polyhedron, 2005, 24, 2901
Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K., Science, 1996, 272, 704
Verdaguer, M., Science, 1996, 272, 698
a) Yokoyama, T.; Kigushi, M.; Ohta, T.; Sato, O.; Einaga, Y.; Hashimoto, K., Phys.Rev.B, 1999,
60, 9340; b) Yokoyama, T.; Ohta, T.; Sato, O.; Hashimoto, K., Phys.Rev.B, 1998, 58, 8257; c)
Varret, F.; Constant-Machado, H.; Dormann, J. L.; Goujon, A.; Jeftic, J.; Nogues, M.;
Bousseksou, A.; Klokishner, S. I.; Dolbecq, A.; Verdaguer, M., Hyperfine Inter., 1998, 113, 37;
d) Escax, V.; Bleuzen, A.; Cartier dit Moulin, C.; Villain, F.; Goujon, A.; Varret, F.; Verdaguer,
M., J.Am.Chem.Soc., 2001, 123, 12536
Sato, O.; Einaga, Y.; Fujishima, A.; Hashimoto, K., Inorg.Chem., 1999, 38, 4405
Bleuzen, A.; Lomenech, C.; Escax, V.; Villain, F.; Varret, F.; Cartier dit Moulin, C.; Verdaguer,
M., J.Am.Chem.Soc., 2000, 122, 6648
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
a) Cartier dit Moulin, C.; Villain, F.; Bleuzen, A.; Arrio, M.-A.; Sainctavit, P.; Lomenech, C.;
Escax, V.; Baudelet, F.; Dartyge, E.; Gallet, J.-J.; Verdaguer, M., J.Am.Chem.Soc., 2000, 122,
6653 ; b) Champion, G.; Escax, V.; Cartier dit Moulin, C.; Bleuzen, A.; Villain, F.; Baudelet, F.;
Dartyge, E.; Verdaguer, M., J.Am.Chem.Soc., 2001, 123, 12544
Escax, V.; Bleuzen, A.; Itie, J. P.; Munsch, P.; Varret, F.; Verdaguer, M., J.Phys.Chem.B, 2003,
107, 4763
Bleuzen, A.; Escax, V.; Itie, J. P.; Munsch, P.; Verdaguer, M., C.R.Chimie, 2003, 6, 343
a) Goujon, A.; Varret, F.; Escax, V.; Bleuzen, A.; Verdaguer, M., Polyhedron, 2001, 20, 1339 ;
b) Shimamoto, N.; Ohkoshi, S.-I.; Sato, O.; Hashimoto, K., Inorg.Chem., 2002, 41, 678
Shimamoto, N.; Ohkoshi, S.-I.; Sato, O.; Hashimoto, K., Chem.Lett., 2002, 486
Liu, H. W.; Matsuda, T.; Gu, Z. Z.; Takahashi, K.; Cui, A. L.; Nakajima, R.; Fujishima, A.; Sato,
O., Phys.Rev.Lett., 2003, 90, 167403
Ksenofontov, V.; Levchenko, G; Reiman, S.; Gütlich, P.; Bleuzen, A.; Escax, V.; Verdaguer, M.,
Phys.Rev.B, 2003, 68, 024415
a) Yamamoto, T.; Umemura, Y.; Sato, O.; Einaga, Y., Chem.Lett., 2004, 33, 500; b) Moore, J.
G.; Lochner, E. J.; Ramsey, C.; Dalal, N. S.; Stiegman, A. E., Angew.Chem.Int.Ed., 2003, 42,
2741; c) Yamamoto, T.; Umemura, Y.; Sato, O.; Einaga, Y., J.Am.Chem.Soc., 2005, 127, 16065
Park, J.-H.; Frye, F.; Lane, S.; Cizmar, E.; Huh, Y. D.; Talham, D. R.; Meisel, M. W.,
Polyhedron, 2005, 24, 2355
Park, J.-H.; Cizmar, E.; Meisel, M. W.; Huh, Y. D.; Frye, F.; Lane, S.; Talham, D. R.,
Appl.Phys.Lett., 2004, 85, 3797
Frye, F. A.; Pajerowski, D. M.; Lane, S. M.; Anderson, N. E.; Park, J.-H.; Meisel, M. W.;
Talham, D. R., Polyhedron, 2007, 26, 2281
Pajerowski, D. M.; Frye, F. A.; Talham, D. R.; Meisel, M. W., New J.Phys., 2007, 9, 222
Li, D.; Clerac, R.; Roubeau, O.; Harte, E.; Mathoniere, C.; Le Bris, R.; Holmes, S. M.,
J.Am.Chem.Soc., 2008, 130, 252
Ohkoshi, S.-I.; Tokoro, H.; Utsunomiya, M.; Mizuno, M.; Abe, M.; Hashimoto, K.,
J.Phys.Chem.B, 2002, 106, 2423
a) Yokoyama, T.; Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.; Okamoto, K.; Ohta, T.,
Phys.Rev.B, 2002, 66, 184111 ; b) Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.;
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., J.Phys.Soc.Jpn., 2002, 71, 2078
a) Osawa, H.; Iwazumi, T.; Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.; Shoji, H.; Hirai, E.;
Nakamura, T.; Nanao, S.; Isozumi, Y., Solid State Comm, 2003, 125, 237 ; b) Kato, K.;
Morimoto, Y.; Takata, M.; Sakata, M.; Umekawa, M.; Hamada, N.; Ohkoshi, S.-I.; Tokoro, H.;
Hashimoto, K., Phys.Rev.Lett., 2003, 91, 25502 ; c) Ishiyama, H.; Maruta, G.; Kobayashi, T.;
Takeda, S., Polyhedron, 2003, 22, 1981 ; d) Umekawa, M.; Hamada, N.; Kodama, A.;
Moritomo, Y., J.Phys.Soc.Jpn., 2004, 73, 430 ; e) Moritomo, Y.; Kato, K.; Kuriki, A.; Takata,
M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., J.Phys.Soc.Jpn., 2003, 72, 2698 ; f)
Tokoro, H.; Ohkoshi, S.-I.; Matsuda, T.; Hashimoto, K., Inorg.Chem., 2004, 43, 5231 ; g) Kato,
K.; Moritomo, Y.; Tanaka, H.; Tokoro, H.; Ohkoshi, S.-I.; Takata, M., J.Phys.Soc.Jpn., 2007,
76, 123602 ; h) Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I., J.Magnetism and Magnetic Mater.,
2007, 310, 1422 ; i ) Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K., Coord.Chem.Rev., 2005, 249,
1830
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., Appl.Phys.Lett., 2003, 82, 1245
a) Cobo, S.; Fernandez, R.; Salmon, L.; Molnar, G.; Bousseksou, A., Eur.J.Inorg.Chem., 2007,
1549; b Ohkoshi, S.-I.; Matsuda, T.; Tokoro, H.; Hashimoto, K., Chem.Mater., 2005, 17, 81
Tokoro, H.; Matsuda, T.; Hashimoto, K.; Ohkoshi, S.-I., J.Appl.Phys., 2005, 97, 10M508
Tokoro, H.; Matsuda, T.; Nuida, T.; Moritomo, Y.; Ohoyama, K.; Loutete Dangui, E. D.;
Boukheddaden, K.; Ohkoshi, S.-I., Chem.Mater., 2008, 20, 423
17
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
54
55
56
57
58
59
60
61
62
63
64
65
66
67
18
Moritomo, Y.; Hanawa, M.; Ohishi, Y.; Kato, K.; Takata, M.; Kuriki, A.; Nishibori, E.; Sakata,
M.; Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K., Phys.Rev.B, 2003, 68, 144106
a) Tokoro, H.; Miyashita, S.; Hashimoto, K.; Ohkoshi, S.-I., Phys.Rev.B, 2006, 73, 172415; b)
Tokoro, H.; Matsuda, T.; Miyashita, S.; Hashimoto, K.; Ohkoshi, S.-I., J.Phys.Soc.Jpn., 2006,
75, 085004
Nuida, T.; Matsuda, T.; Tokoro, H.; Sakurai, S.; Hashimoto, K.; Ohkoshi, S.-I., J.Am.Chem.Soc.,
2005, 127, 11604
Ohkoshi, S.-I.; Tokoro, H.; Matsuda, T.; Takahashi, H.; Irie, H.; Hashimoto, K.,
Angew.Chem.Int.Ed., 2007, 46, 3238
a) Matsuda, T.; Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I., Dalton Transactions, 2006, 5046; b)
Matsuda, T.; Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I., J.Appl.Phys., 2007, 101, 09E101; c)
Ishiji, K.; Matsuda, T.; Tokoro, H.; Iwazumi, T.; Hashimoto, K.; Ohkoshi, S.-I.,
J.Phys.Chem.Solids, 2007, 68, 2158
a) Bhattacharjee, A.; Saha, S.; Koner, S.; Ksenofontov, V.; Reiman, S.; Gütlich, P., J.Magnetism
and Magnetic Mater., 2006, 302, 173; b) Bhattacharjee, A.; Saha, S.; Koner, S.; Miyazaki, Y.,
J.Magnetism and Magnetic Mater., 2007, 312, 435
a) Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I., Z.Anorg.Allg.Chem., 2007, 633, 1134;
b) Matsuda, T.; Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I., Acta Cryst., 2008, E64,
i11
a) Goodwin, A. L.; Chapman, K. W.; Kepert, C. J., J.Am.Chem.Soc., 2005, 127, 17980; b)
Pretsch, T.; Chapman, K. W.; Halder, G. J.; Kepert, C. J., Chem.Comm., 2006, 1857; c)
Margadonna, S.; Prassides, K.; Fitch, A. N., J.Am.Chem.Soc., 2004, 126, 15390; d) Chapman, K.
W.; Chupas, P. J.; Kepert, C. J., J.Am.Chem.Soc., 2006, 128, 7009
Pournaghi-Azar, M. H.; Sabzi, R., J.Solid State Electrochem., 2002, 6, 553
a) Rutkowska, I. A.; Stroka, J.; Wrona, P. K.; Galus, Z., J.Electroanalytical Chem., 2004, 564,
199; b) Chen, C., J.Electroanalytical Chem., 2002, 521, 29; c) Lezna, R. O.; Romagnoli, R.; de
Tacconi, N. R.; Rajeshwar, K., J.Electroanalytical Chem., 2003, 544, 101; d) Yang, C.; Wang,
C. H.; Wu, J.-S.; Xia, X., Electrochim.Acta, 2006, 51, 4019; e) Liu, Y.-Q.; Yan, Y.; Shen, H.-X.,
Chin.J.Chem., 2005, 23, 1165; f) Tao, W.; Liu, Y.; Pan, D.; Nie, L.; Yao, S., J.Colloid Interface
Sc., 2004, 275, 257; g) Roka, A.; Varga, I.; Inzelt, G., Electrochim.Acta, 2006, 51, 6243; h)
Vittal, R.; Gomathi, H., J.Phys.Chem.B, 2002, 106, 10135; i) Gao, Z.; Wang, G.; Li, P.; Zhao,
Z., Electrochim.Acta, 1991, 36, 147; j) Yu, H.; Sheng, Q. L.; Li, L.; Zheng, J.-B.,
J.Electroanalytical Chem., 2007, 606, 55
a) Itaya, K.; Ataka, T.; Toshima, S., J.Am.Chem.Soc., 1982, 104, 3751; b) Nakanishi, S.; Lu, G.;
Kothari, H. M.; Bohannan, E. W.; Switzer, J. A., J.Am.Chem.Soc., 2003, 125, 14998; c) Wu, Y.;
Pfennig, B. W.; Bocarsly, A. B.; Vicenzi, E. P., Inorg.Chem., 1995, 34, 4262
a) Itaya, K.; Ataka, T.; Toshima, S., J.Am.Chem.Soc., 1982, 104, 4767; b) Umemura, Y.,
J.Phys.Chem.B, 2002, 106, 11168; c) Umemura, Y.; Einaga, Y.; Yamagishi, A., Mater.Lett.,
2004, 58, 2472; d) Zamponi, S.; Kijak, A. M.; Sommer, A. J.; Marassi, R.; Kulesza, P. J.; Cox, J.
A., J.Solid State Electrochem., 2002, 6, 528; e) Kulesza, P. J.; Malik, M. A.; Berrettoni, M.;
Giorgetti, M.; Zamponi, S.; Schmidt, R.; Marassi, R., J.Phys.Chem.B, 1998, 102, 1870; f) Wang,
F.; Wang, J.; Chen, H.; Dong, S., J.Electroanalytical Chem., 2007, 600, 265; g) Han, S.; Chen,
Y.; Pang, R.; Wan, P.; Fan, M., Ind.Eng.Chem.Res., 2007, 46, 6847
a) Tacconi, N. R.; Rajeshwar, K.; Lezna, R. O., Chem.Mater., 2003, 15, 3046 ; b) Itaya, K.;
Uchida, I.; Neff, V. D., Acc.Chem.Res., 1986, 19, 162; c) Jin, W.; Toutianoush, A.; Pyrasch, M.;
Schnepf, J.; Gottschalk, H.; Rammensee, W.; Tieke, B., J.Phys.Chem.B, 2003, 107, 12062; d)
Scholz, F.; Dostal, A., Angew.Chem.Int.Ed., 1995, 34; e) Kulesza, P. J.; Zamponi, S.; Malik, M.
A.; Berrettoni, M.; Wolkiewicz, A.; Marassi, R., Electrochim.Acta, 1998, 43, 919
Kaplun, M. M.; Smirnov, Y. E.; Mikli, V.; Malev, V. V., Russ.J.Electrochem., 2001, 37, 914
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
68
69
70
71
72
73
74
75
a) Benedetto, G. E. De; Guascito, M. R.; Ciriello, R.; Cataldi, T. R. I., Anal.Chim.Acta, 2000,
410, 143; b) Bo, A. L.; Lin, X. Q., Talanta, 1999, 49, 717; c) Sauter, S.; Wittstock, G.; Szargan,
R., Phys.Chem.Chem.Phys., 2001, 3, 562
a) Kulesza, P. J.; Malik, M. A.; Zamponi, S.; Berrettoni, M.; Marassi, R., J.Electroanalytical
Chem., 1995, 397, 287 ; b) Toyoda, Y.; Katoh, N.; Kuwabara, K., Mater.Science Engineering B,
2004, 108, 271; c) Chen, S.-M.; Peng, K.-T.; Lin, K.-C., Electroanalysis, 2005, 17, 319; d)
Vittal, R.; Kim, K.-J.; Gomathi, H.; Yegnaraman, V., J.Phys.Chem.B, 2008, 112, 1149
a) Yu, H.; Zheng, J.-B., Chin.J.Chem., 2007, 25, 503; b) Chen, S.-M.; Wu, M.-H.;
Thangamuthu, R., Electroanalysis, 2008, 20, 178; c) Zheng, J.; Sheng, Q.; Shen, Y.,
J.Electroanalytical Chem., 2007, 611, 155; d) Chen, S.-M.; Liou, C.-Y.; Thangamuthu, R.,
Electroanalysis, 2007, 19, 2457; e) Narayanan, S. S.; Scholz, F., Electroanalysis, 1999, 11, 465;
f) Mo, J.-W.; Ogorevc, B.; Zhang, X.; Pihlar, B., Electroanalysis, 2000, 12, 48
Cui, X.; Hong, L.; Lin, X., Anal.Sc., 2002, 18, 543
a) Sheng, Q. L.; Yu, H.; Zheng, J. B., Electrochim.Acta, 2007, 52, 4506; b) Garjonyte, R.;
Malinauskas, A., Sensors Actuators B, 1998, 46, 236; c) Wang, Q.; Zhang, L.; Qiu, L.; Sun, J.;
Shen, J., Langmuir, 2007, 23, 6084; d) Florescu, M.; Brett, C. M. A., Anal.Lett., 2004, 37, 871;
e) Pauliukaite, R.; Florescu, M.; Brett, C. M. A., J.Solid State Electrochem., 2005, 9, 354; f)
Fiorito, P. A.; Goncales, V. R.; Ponzio, E. A.; Cordoba de Torresi, S. I., Chem.Comm.,
2005, 366; g) Xian, Y.; Hu, Y.; Liu, F.; Xian, Y.; Feng, L.; Jin, L., Biosens.Bioelec., 2007, 22,
2827; h) Tao, W.; Pan, D.; Liu, Y.; Nie, L.; Yao, S., Anal.Biochem., 2005, 338, 332
a) Shi, Y.; Zhou, B.; Wu, P.; Wang, K.; Cai, C., J.Electroanalytical Chem., 2007, 611, 1; b)
Sheng, Q.; Yu, H.; Zheng, J., J.Electroanalytical Chem., 2007, 606, 39; c) Senthil Kumar, S.;
Sriman Narayanan, S., Chem.Pharm.Bull., 2006, 54, 963; d) Wang, G.; Meng, J.; Liu, H.; Jiao,
S.; Zhang, W.; Chen, D.; Fang, B., Electrochim.Acta, 2008, 53, 2837
a) Abbaspour, A.; Ayatollahi Mehrgardi, M., Anal.Chem., 2004, 76, 5690 ; b) Xun, Z.; Cai, C.;
Lu, T., Electroanalysis, 2004, 16, 674; c) Abe, T.; Taguchi, F.; Tokita, S.; Kaneko, M.,
J.Mol.Catalysis A, 1997, 126, L89; d) Wang, P.; Yuan, Y.; Zhu, G., Fresenius J.Anal.Chem.,
2001, 369, 153; e) Chen, S.-M.; Liao, C.-J.; Vasantha, V. S., J.Electroanalytical Chem., 2006,
589, 15; f) Wang, P.; Jing, X.; Zhang, W.; Zhu, G., J.Solid State Electrochem., 2001, 5, 369; g)
Shi, L.-H.; Wu, T.; Wang, M.-J.; Li, D.; Zhang, Y.-J.; Li, J.-H., Chin.J.Chem., 2005, 23, 149; h)
Souza Castilho, M. de; Stradiotto, N. R., Talanta, 2008, 74, 1630; i) Shankaran, D. R.;
Narayanan, S. S., Sensors Actuators B, 2002, 86, 180; j) Liu, M.-C.; Li, P; Zhang, L.; Xian, Y.Z.; Ding, H.-C.; Zhang, C.-L.; Zhang, F.-F.; Jin, L.-T., Chin.J.Chem., 2005, 23, 983; k) Liu, M.;
Li, P.; Cheng, Y.; Xian, Y.; Zhang, C.; Jin, L., Anal.Bioanal.Chem., 2004, 380, 742; l)
Shankaran, D. R.; Narayanan, S. S., Bull.Chem.Soc.Jpn., 2002, 75, 501; m) Shaidarova, L. G.;
Davletshina, L. N.; Budnikov, G. K., J.Anal.Chem., 2006, 61, 502; n) Lowinsohn, D.; Bertotti,
M., Anal.Biochem., 2007, 365, 260; o) Gao, Z.; Zhou, X.; Wang, G.; Li, P.; Zhao, Z.,
Anal.Chim.Acta, 1991, 244, 39; p) Tseng, K.-S.; Chen, L.-C.; Ho, K.-C., Electroanalysis, 2006,
18, 1306; q) Chang, C.-Y.; Chau, L.-K.; Hu, W.-P.; Wang, C.-Y.; Liao, J.-H., Microporous
Mesoporous Mater., 2008, 109, 505
a) Martinez-Garcia, R.; Reguera, E.; Rodriguez, J.; Balmaseda, J.; Roque, J., Powder
Diffraction, 2004, 19, 255 ; b) Milyutin, V. V.; Gelis, V. M.; Klindukhov, V. G.; Obruchikov, A.
V., Radiochemistry, 2004, 46, 479; c) Nilchi, A.; Malek, B.; Ghanadi Maragheh, M.; Khanchi,
A., J.Radioanal.Nucl.Chem., 2003, 258, 457; d) Abdel raouf, M. W.,
J.Chem.Technol.Biotechnol., 2003, 79, 22; e) Mardan, A.; Ajaz, R.; Mehmood, A.; Raza, S. M.;
Ghaffar, A., Separation Purification Tech., 1999, 16, 147; f) Jain, A. K.; Singh, R.; Bala, C.,
Bull.Chem.Soc.Jpn., 1983, 56, 1269; g) Lehto, J.; Harjula, R.; Wallace, J.,
J.Radioanal.Nucl.Chem., 1987, 111, 297; h) Lehto, J.; Szirtes, L., Radiat.Phys.Chem., 1994, 43,
261; i) Lehto, J.; Paajanen, A.; Harjula, R., J.Radioanal.Nucl.Chem., 1992, 164, 39; j) Aziz, M.;
Beheir, Sh. G., J.Radioanal.Nucl.Chem., 1995, 191, 53; k) Samanta, S. K.; Thomas, M. K.;
19
Chapter 1. The Use of Prussian Blue Analogues in Modern Day Science
76
77
78
79
80
81
20
Theyyunni, T. K., J.Radioanal.Nucl.Chem., 1995, 199, 183; l) Yu, Q.; Steen, W. A.; Jeerage, K.
M.; Jiang, S.; Schwartz, D. T., J.Electrochem.Soc., 2002, 149, E195
a) Kaye, S. S.; Long, J. R., J.Am.Chem.Soc., 2005, 127, 6505; b) Chapman, K. W.; Southon, P.
D.; Weeks, C. L.; Kepert, C. J., Chem.Comm., 2005, 3322; c) Hartman, M. R.; Peterson, V. K.;
Liu, Y.; Kaye, S. S.; Long, J. R., Chem.Mater., 2006, 18, 3221; d) Chapman, K. W.; Chupas, P.
J.; Maxey, E. R.; Richardson, J. W., Chem.Comm., 2006, 4013; e) Kaye, S. S.; Long, J. R.,
Chem.Comm., 2007, 4486; f) Reguera, L.; Balmaseda, J.; Castillo, L. F. del; Reguera, E.,
J.Phys.Chem.C, 2008, 112, 5589
a) Balmaseda, J.; Reguera, E.; Fernandez, J.; Gordillo, A.; Yee-Madeira, H.,
J.Phys.Chem.Solids, 2003, 64, 685; b) Chapman, K. W.; Chupas, P. J.; Kepert, C. J.,
J.Am.Chem.Soc., 2005, 127, 11232; c) Reguera, E.; Fernandez-Betran, J.; Duque, J., Polyhedron,
1994, 13, 479; d) Iton, L. E.; Beal, R. B.; Hodul, D. T., J.Mol.Catalysis, 1983, 21, 151; e) Ali, S.
R.; Ahmad, J.; Kamaluddin, Colloid Surf.A, 2004, 236, 165; f) Kubica, B.; Tuteja-Krysa, M.;
Misiak, R.; Tronh thi Thu My; Kubica, M.; Stobinski, M.; Godunowa, H.,
J.Radioanal.Nucl.Chem., 2003, 258, 167; g) Roque, J.; Reguera, E.; Balmaseda, J.; RodriguezHernandez, J.; Reguera, L.; Castillo, L. F. del, Microporous Mesoporous Mater., 2007, 103, 57
a) Alam, T.; Tarannum, H.; Ravi Kumar, M. N. V.; Kamaluddin, Talanta, 2000, 51, 1097; b)
Alam, T.; Tarannum, H.; Kumar, N.; Kamaluddin, J.Colloid Interface Sc., 2000, 224, 133; c)
Ali, S. R.; Alam, T.; Kamaluddin, Astrobiol., 2004, 4, 420; d) Alam, T.; Kamaluddin, Colloid
Surf.A, 2000, 162, 89; e) Viladkar, S.; Alam, T.; Kamaluddin, J.Inorg.Biochem., 1994, 53, 69
a) Hao, X.; Schwartz, D. T., Chem.Mater., 2005, 17, 5831; b) Mikhailov, O. V.; Tatarintseva, T.
B.; Naumkina, N. I.; Lygina, T. Z., Russ.J.Coord.Chem., 2004, 30, 639; c) Mikhailov, O. V.;
Tatarintseva, T. B.; Naumkina, N. I.; Kolgina, V. A., Russ.J.Coord.Chem., 2003, 29, 630; d)
Rosseinsky, D. R.; Lim, H.; Jiang, H.; Chai, J. W., Inorg.Chem., 2003, 42, 6015; e) Imanishi, N.;
Morikowa, T.; Kondo, J.; Yamane, R.; Takeda, Y.; Yamamoto, O.; Sakaebe, H.; Tabuchi, M.,
J.Power Sources, 1999, 81-82, 530; f) Kubica, B.; Godunowa, H.; Tuteja-Krysa, M.; Stobinski,
M.; Misiak, R., J.Radioanal.Nucl.Chem., 2004, 262, 721; g) Mikhailov, O. V.; Tatarintseva, T.
B.; Naumkina, N. I.; Lygina, T. Z., Russ.J.Gen.Chem., 2004, 74, 7; h) Lehto, J.; Haukka, S.;
Blomberg, M., J.Solid State Chem., 1991, 90, 79; i) Gimenez-Romero, D.; Bueno, P. R.; GarciaJareno, J. J.; Gabrielli, C.; Perrot, H.; Vicente, F., J.Phys.Chem.B, 2006, 110, 19364
Section 1.3.6
Section 1.3.5
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 2. The Influence of Defects on the Electron
Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O1
2.1 Introduction
This research was triggered by the intriguing results reported on the properties of
RbMn[Fe(CN)6].2,3 At room temperature, the material contains high spin MnII (S =
5/2) and low spin FeIII (S = 1/2), whereas cooling of the compound induces a
structural phase transition associated with a charge transfer4 yielding high spin
MnIII (S = 2) and low spin FeII (S = 0) and resulting in a ferromagnetic ordering
below Tc = 12 K. A similar process has also been found to be induced by light at
3 K,5 and there is an indication that it can also be triggered by X-ray illumination of
14.575 keV.6 RbMn[Fe(CN)6] 7 represents an appealing object for further study. A
vital advantageous criterion supporting this is the possibility of reaching a maximal
photo-efficiency, which is a consequence of its fairly 1:1 stoichiometric Mn:Fe
ratio. This is in fact a drawback of the family of related non-stoichiometric
CoIII-FeII PB analogues,8 where the volume changes accompanying the charge
transfer require a flexibility of the 3D-network generated by Fe defects, thus
preventing the optimization of the quantum yield by restricting the number of
photo-active Co-Fe pairs.
This chapter focuses on acquiring a more detailed insight in the relationship
between structural features and electron transfer capabilities of
RbxMn[Fe(CN)6]y·zH2O analogues. It is shown in which way small variations in
the synthetic procedures may give rise to materials that differ considerably in
composition and their consequential physical behaviour. The obtained materials are
characterised with the use of X-ray powder diffraction, magnetic measurements
and differential scanning calorimetry. Furthermore, IR, Raman, 57Fe Mössbauer
and X-ray photoelectron spectroscopy are exploited as well for characterisation.
2.2 Experimental Section
2.2.1. Synthesis
All chemicals were purchased from Sigma-Aldrich and used without further
purification.
Sample 2.1 was prepared by slowly adding (addition time: 20 minutes) an aqueous
solution (50 mL) containing K3Fe(CN)6 (0.1 M) and RbCl (1 M) to an aqueous
solution (50 mL) containing MnCl2·4H2O (0.1 M).
Sample 2.2 was prepared by slowly adding (addition time: 20 minutes) an aqueous
solution (50 mL) containing MnCl2·4H2O (0.1 M) to an aqueous solution (50 mL)
containing K3Fe(CN)6 (0.1 M) and RbCl (1 M).
21
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
Sample 2.3 was prepared by simultaneously adding equal amounts of an aqueous
solution (50 mL) containing MnCl2·4H2O (0.1 M) and an aqueous solution (50 mL)
containing both K3Fe(CN)6 (0.1 M) and RbCl (1 M) together over a period of 5
hours.
Sample 2.4 was prepared in a similar fashion as sample 2.2, but now the complete
synthesis took place in the dark. Characterization of this sample took place in
daylight (except for the XPS measurements, where the sample was shielded from
all light except the X-rays).
All solutions were stirred mechanically and kept at a temperature of 50°C during
the addition procedure and for an additional hour after completion of the addition.
A brown powder precipitated and this was centrifuged and washed twice with
distilled water of room temperature. The powder was allowed to dry in air for
about 12 hours at room temperature.
2.2.2 Elemental Analysis
Elemental analysis of Rb, Mn, Fe, (Inductive Coupling Plasma Atomic Emission
Spectroscopy after demineralization in H2SO4/HNO3) C, H (combustion above
850°C, then IR-detection) and N (combustion above 850oC, then
quantogravimetry) was performed at the analysis facility of CNRS in Vernaison,
France. The O atom was presumed to be the only other element present and
obtained by difference to 100%.
2.2.3 X-Ray Powder Diffraction
Measurements were carried out in Bragg-Brentano geometry using a Bruker D8
Advance diffractometer operating with Cu Kα radiation. The finely ground powder
was attached to the sample holder with vaseline. Data were collected between 2θ =
10o and 2θ = 70o with a step size of 0.02o, measuring for 4 seconds per step. The
sample was rotated at 60 rpm. The resulting diffraction profiles were analyzed
using the GSAS9 software suite using both the high temperature (F-43m) and the
low temperature phase (I-4m2) as found by Kato et al.10 and Moritomo et al..11
2.2.4 FTIR Spectroscopy
Spectra were recorded at room temperature on a Interspec 200-X FTIR
spectrofotometer by Interspectrum with the use of KBr mulled tablets. The
resolution was higher than 1 cm-1.
2.2.5 Raman Spectroscopy
Room temperature polarized Raman scattering measurements in the spectral region
2000 - 2300 cm-1 were performed in a backscattering configuration, using a microRaman spectrometer (T64000- Jobin Yvon) which consists of a double grating
monochromator acting as a spectral filter and a polychromator which disperses the
scattered light on a liquid nitrogen cooled CCD. The spectral resolution of the
experiments was higher than 1 cm-1. The second harmonic of a Nd:YVO4 laser
22
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
(532 nm) (Verdi – Coherent) was used as a light source. A portion of the laser
output (0.2 mW) was focused on the sample using a 50x microscope objective. The
power density on the sample was of the order of 10 W m-2. A study of
photostability of the samples revealed that at excitation densities above 100 W m-2
a darkening of the irradiated spot takes place which indicates photo-induced
changes. In the case of ten times lower excitation density, i.e. 10 W m-2, Raman
spectra measured on the same spot are stable at RT and there are no noticeable
differences between spectra recorded at different spots.
2.2.6 57Fe Mössbauer Spectroscopy
Mössbauer spectra were collected in an exchange gas cryostat. A conventional
constant acceleration spectrometer was used, equipped with a 57Co/Rh source kept
at RT and calibrated with Fe metal. Isomer shift values are reported relative to iron
metal (α-Fe). The spectra were fitted on a PC with a least squares minimization
procedure assuming Lorentzian line shapes.
2.2.7 Magnetic Measurements
Magnetic measurements were performed on a Quantum Design magnetometer with
a superconducting quantum interference device. The sample was prepared by
putting 20-30 mg of the compound (accurately weighed) between two pieces of
cotton wool in a gelcap. In the magnetic susceptibility measurements, first the
sample was slowly cooled from room temperature to 5 K. Then the field was kept
constant at 0.1 T while the temperature varied from 5 to 350 K and back to 150 K.
The measurements were corrected for the diamagnetic contribution with Pascal
constants, but not for the diamagnetic contribution of the sample holder.
The inverse magnetic susceptibility data were fitted with a straight line. From the
fit the Curie constant (C) and the Curie-Weiss temperature (θ) were calculated via
the formula χ M =
C
, where χM = molar magnetic susceptibility and T =
T −θ
temperature.
In the magnetisation measurements first the temperature was varied from 5 K to
20 K in the absence of a magnetic field (zero field cooling, ZFC), then the field
was kept constant at 0.01 T and the temperature was varied from 20 K to 5 K (field
cooling, FC). Finally the field was switched off again and the temperature varied
from 5 K to 20 K (remanence).
2.2.8 Differential Scanning Calorimetry (DSC)
The DSC measurement was performed in the 100–330 K temperature range at a
scan rate of 10 K min-1, using a differential scanning calorimeter Q1000 from TA
Instruments. The measurements were carried out using 4.81 mg of powdered
sample 2.3 sealed in an aluminium pan with a mechanical crimp. Temperature and
enthalpy calibrations were made with standard samples of Indium, using its melting
23
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
transition (429.76 K, 3.296 kJ mol-1) and with a sample of Mn3GaC, using its ferroto antiferromagnetic transition (171.83 K, 713 J mol-1). An overall accuracy of
±0.5 K in temperature and ±5% in the enthalpy contents is estimated. The
uncertainty for the determination of the anomalous enthalpy and entropy increases
at lower temperatures due to the difficulty in tracing the baselines (especially for
the cooling mode).
2.2.9 X-Ray Photoelectron Spectroscopy (XPS)
For the XPS measurements evaporated gold films supported on glass were used as
substrates. They were cleaned in an ozone discharge (UV-Ozone Photoreactor TM
PR100, Ultra Violet Products) for 15 min and sonicated in ethanol for 15 min
immediately before being employed. Prussian Blue analogue powder samples were
dispersed in distilled-deionised water and after stirring for 15 min, a small drop of
the suspension was left to dry in air on the substrate.
The samples were introduced through a load look system into an SSX-100 (Surface
Science Instruments) photoelectron spectrometer with a monochromatic Al Kα
X-ray source (hν = 1486.6 eV). The base pressure in the spectrometer during the
measurements was 1·10-9 Torr. The photoelectron take off angle was 37o. The
energy resolution was set to 1.4 eV. Sample charging was compensated for by
directing an electron flood gun supplying 0.1 eV kinetic energy electrons onto the
sample and covering the sample holder with a Au grid. Since Prussian Blue
analogues are insulating compounds, XPS binding energies have to be referenced
to an internal reference of the sample. The nitrogen 1s signal at 398 eV (cyanide
groups) was chosen for this purpose and it was supposed that the variation of the
N1s binding energy when the environment changes during phase transitions or
when comparing different compounds are negligible compared to the changes in
the 2p core level binding energies of the metal ion which is bound to it.12 The XP
spectra showed no X-ray induced changes. Spectral analysis included Shirley or
Tougaard background subtraction, and peak deconvolution employing Gaussian
functions, in a least squares curve-fitting program (WinSpec) developed at the
LISE, University of Namur, Belgium.
2.3 Results and Discussion
2.3.1 General
Analytical data of the RbxMn[Fe(CN)6]y·zH2O compounds are compiled in table
2.1. Based on these findings, the yields of the samples have been determined as
(based on Mn): 96% (sample 2.1), 86% (sample 2.2), 82% (sample 2.3) and 84%
(sample 2.4). Contrary to the materials reported by the group of Ohkoshi and
Hashimoto,13 for which no specific details of the synthetic procedure were given,
the samples under discussion here unambiguously appear to be hydrated and the
composition is non-stoichiometric. The deviation from the perfect Rb:Mn:Fe
stoichiometry of 1:1:1 for RbxMn[Fe(CN)6]y·zH2O directly correlates with the
24
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
lattice water content of the samples. Apart from the exclusion of daylight, the
synthetic procedures for sample 2.2 and 2.4 were kept as similar as possible,
however, a difference in the stoichiometry of the two samples is noted. In the
course of our research numerous samples have been synthesized and characterized.
Apparently, stable compounds are formed which contain several water molecules
per Mn ion. This contribution will only focus on four representative samples
originating from four different preparation methods.
Table 2.1. Experimental (exp) and calculated (cal) weight percentages of the various
samples.
Sample
2.1
2.2
2.3
2.4
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
%
Rb
14.96
14.96
22.24
22.24
22.75
22.75
19.97
19.97
%
Mn
16.42
16.42
15.51
15.51
15.09
15.08
15.83
15.83
%
Fe
14.31
14.35
15.20
15.02
15.04
15.04
14.96
15.22
%
C
18.91
18.52
20.00
19.38
20.19
19.41
19.64
19.64
%
N
21.28
21.59
22.61
22.61
22.89
22.63
22.40
22.90
%
H
1.44
1.58
0.25
0.59
0.18
0.57
0.67
0.72
Calculated Composition
Rb0.59Mn[Fe(CN)6]0.86·2.63H2O
Rb0.92Mn[Fe(CN)6]0.95·1.03H2O
Rb0.97Mn[Fe(CN)6]0.98·1.03H2O
Rb0.81Mn[Fe(CN)6]0.95·1.24H2O
Exp = experimental values, Cal = calculated values.
The standard deviations in the observed percentages is ± 0.3%
2.3.2 X-Ray Powder Diffraction
An assessment of the phases present within the samples has been made by
analyzing the X-ray powder diffraction patterns recorded at room temperature. For
all samples these show Bragg reflections characteristic for both the low
temperature phase (LT; tetragonal space group I-4m2, Z = 2) and the high
temperature phase (HT; cubic space group F-43m, Z = 4). In the following, we will
explicitly use the word “phase” while referring to a pure form of the material either
in its MnII-FeIII or MnIII-FeII form, whereas the word “configuration” will be
applied when dealing with a material in which fractions of both the MnII-FeIII and
MnIII-FeIII form are present.
Interpretation of the diffraction profiles of these mixed-phases was made based on
the results reported for synchrotron-radiation X-ray powder diffraction studies
performed at temperatures where the compounds Rb0.7Mn1.15[Fe(CN)6]·2H2O 14 and
RbMn[Fe(CN)6] 11 are present as single-phase materials. The X-ray diffraction data
of RbMn[Fe(CN)6] could be interpreted for the low temperature phase at 120 K (I4m2) and for the high temperature phase (F-43m) at 240 K.11 The evolution from
HT to LT phase involves the elongation of one of the cubic axes resulting in a
tetragonal space group with an accompanying reduction of the number of volume
units present in the unit-cell from Z = 4 to Z = 2. For Rb0.7Mn1.15[Fe(CN)6]·2H2O
the X-ray powder diffraction study was carried out at 295 K and the data could be
analyzed using the cubic Fm-3m space group with Z = 4 for modelling the HT
phase.14 In fact, the HT space groups used in both studies only differ in that the
fourfold symmetry axis in F-43m is replaced by a mirror plane in Fm-3m implying
25
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
that in the Fm-3m HT phase only one atomic position for Rb+ is taken into account,
whereas in the F-43m HT phase two individual Rb+ sites exist. Both the Fm-3m
and F-43m spacegroups obey the same reflection conditions15 (hkl, h+k=2n,
h+l=2n) and can therefore not be distinguished by reflection conditions.
In the Rietveld structure refinement procedure [GSAS9] used to analyze the X-ray
diffraction patterns of the present RbxMn[Fe(CN)6]y·zH2O compounds the
implementation of the tetragonal I-4m2 LT phase and the cubic F-43m HT phase
yielded the most satisfying results.16 The model consists of using two different Rb
positions in each phase, which is in agreement with the analysis reported by
Moritomo et al.11 and is summarized in table 2.2. The starting point was the same
relative Rb+ occupancy for both phases, namely 92.1% for Rb position 1 and 7.9%
for the alternative Rb position 2. Since the current samples contain varying
amounts of lattice water molecules, these have been included in the fitting
procedure according to Margadonna et al..14 Following their method used for the
non-stoichiometric Rb0.7Mn1.15[Fe(CN)6]·2H2O material, the O atoms of the H2O
molecules were placed at positions where Fe(CN)6 deficiencies occur, i.e. the O
atom takes the place of the missing N cyanide atom. This means that on the N
position (x) N atoms and (1-x) O atoms were placed (see also figure 2.1 and table
2.2). The defects as described in figure 2.1 do not lead to new reflections, but will
lead to a difference in intensities of the reflections. The remaining water O atoms
were placed in a similar fashion on tetrahedral interstitial positions within the 3D
framework where Rb+ vacancies occur. The fractions of the Rb ions and O1 and O2
were allowed to vary in such a way, that the total Rb fraction never exceeded the
observed value by elemental analysis and the total fraction on a certain position
never exceeded 1. Furthermore, the fractions of the various atoms in both phases
were kept identical. In order to increase the stability of the refinement, atoms on
similar positions were fixed with the same isotropic atomic displacement
parameters, Uiso (i.e. UisoMn = UisoFe; UisoRb1 = UisoRb2 = UisoO1 = UisoO2; UisoN
= UisoC = UisoO3). The line shapes of the reflections were fitted with a pseudoN C
C
N
N C
N
C
C N
C
C
N
N
N
C
N
C N
H
N C
N
N
H
N
C
N
C
N
H
H
N
H
O
C N
Fe
C
O
C
N
H
H
C
C
O
O
N
C
C
N C
H
Mn
N
C
C N
C
N C
C
N
N
N
C
N C
26
C N
N
C
C N
C
N C
C
N
Defect of
Fe(CN)6
Figure 2.1. Visualisation of an
Fe(CN)6 defect in the threedimensional Prussian Blue-type lattice
of RbxMn[Fe(CN)6]y·zH2O. Rb+ ions
and non-coordinated lattice water
molecules are omitted for clarity; both
occupy the tetrahedral interstitial sites.
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Table 2.2. Starting positions and specific restrictions for the refinement of the X-ray
powder diffraction profiles for samples with general stoichiometry RbxMn[Fe(CN)6]y·zH2O.
HT phase F-43m, Z = 4
x
y
z
Occupancya
Uisob
Fe
Mn
Rb1
Rb2
C
N
O1
O2
O3
0
½
¼
¾
0.1822
0.2918
¼
¾
0.3095
0
½
¼
¾
0
0
¼
¾
0
0
½
¼
¾
0
0
¼
¾
0
y
1
0.921x
0.079x
y
y
1-0.921x
1-0.079x
1-y
A
A
B
B
C
C
B
B
C
LT phase I-4m2, Z = 2
Fe
Mn
Rb1
Rb2
C1
C2
N1
N2
O1
O2
O3
O4
x
0
0
0
0
0
0.200
0
0.311
0
0
0
0.311
y
0
0
½
½
0
0.200
0
0.311
½
½
0
0.311
z
0
½
¼
¾
0.180
0
0.285
0
¼
¾
0.285
0
Occupancy
y
1
0.921x
0.079x
y
y
y
y
1-0.921x
1-0.079x
1-y
1-y
D
D
D
D
D
D
D
D
D
D
D
D
Positions from Moritomo et al.11
a
x, y and z come from the stoichiometry as obtained by elemental analysis in RbxMn[Fe(CN)6]y·zH2O
b
A, B, C and D indicate the same values for Uiso
See text for further details
Voigt function in which both the Gaussian and Lorentzian broadening components
were allowed to vary.
From the analysis of the X-ray powder data it can be seen that the LT phase is
present in all samples to varying degrees, since small peaks are present at 2θ =
24.5o and 35o. The LT phase at room temperature should be considered as inactive
material rather than material that has not been converted to the HT phase. Sample
2.3 was the only sample in which an estimate of the amount of LT phase could be
determined: 4.7(4)%. In all other samples the peaks indicative of the LT phase,
were too small, which obstructed the obtaining of meaningful results in the
refinement procedure.
It is to be noted that the agreement with the recorded diffraction patterns may be
considered as being rather poor. Of course, certain obtained parameters (most
importantly the lattice parameters) are more reliable than other parameters (bond
distances, occupancies and phase fractions) since the first depends on the reflection
position only and the latter on the intensity as well. In addition, the present
investigation is even more challenging than those reported previously in that the
27
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
Table 2.3. Selected details for the refined diffraction profiles with F-43m of the various
samples.
Fe-C
(Å)
Mn-N
(Å)
Occupation
4c with Rb
%HT
wRp
Rp
2.1
Lattice
constant a
HTa (Å)
10.5493(7)
1.92(6)
2.22(5)
0.480(7)
100*
18.89
14.64
Reduced χ2
(number of
variables)
2.736 (19)
2.2
10.5664(6)
1.91(6)
2.24(5)
0.92
100*
22.16
16.39
5.477 (18)
2.3
10.5597(3)
1.88(6)
2.24(5)
0.97
95.3(4)**
20.92
16.38
9.691 (23)
2.4
10.5586(9)
1.92(9)
2.22(7)
0.816(7)
100*
25.39
20.61
6.065 (19)
Sample
Sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample
2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
Values in parentheses are standard deviations. Values without parentheses were kept constant during
the fitting, in order to ensure a stable refinement.
a
HT stands for the high temperature phase F-43m, LT for the low temperature phase I-4m2.
* Peaks show the presence of LT phase I-4m2, but refinement procedure does not yield meaningful
values for the parameters in the refinement.
** 4.7(4)% LT phase I-4m2 present: a = 7.29(2) Å, c = 10.31(6) Å
present diffraction profiles include the co-existence of both the LT and HT phase.
In that respect, the results of our attempt to model these diffraction data may be
considered quite reasonable. To a certain extent it may be unanticipated that the
fitting of X-ray powder data on such mixed-phase materials would yield such
relatively reliable results, as the occurrence of two phases might involve
boundaries or cross sections within the crystallites where the structure could be
rather amorphous. It may well be that the change from HT to LT phase within these
materials might occur in a relatively smooth fashion as it only involves the
elongation of one of the cubic axes to yield the tetragonal space group. This
elongation may be quite subtle and could proceed gradually, i.e. no precise
domains with one phase or the other need to be present.
Table 2.2 also shows the calculated Mn-N and Fe-C distances in the structures. All
Fe-C distances are comparable to the literature values from 1.929(4) Å to
1.918(1) Å in single crystals of RbxMn[Fe(CN)6]y·zH2O determined at 293 K and
90 K, which is indicative of LS FeIII.17 Similarly the Mn-N distances are indicative
for HS MnII and comparable to literature values (2.105(13) Å at 90 K and
2.205(5) Å at 293 K).17
2.3.3 Vibrational Spectroscopy
Spectroscopic investigation of the compounds 1-4 was carried out in the spectral
window 2000 – 2300 cm-1, i.e. the range of the CN stretching frequencies, being a
fingerprint of structural and electronic changes occurring in Prussian Blue
analogues. The CN stretching frequency of the free CN ion in aqueous solution is
2080 cm-1, while upon coordination to a metal ion it shifts to higher frequencies.18
28
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Figure 2.2. Room temperature FTIR
spectra for sample 2.1
(Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample
2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O),
sample 2.3
(Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and
sample 2.4
(Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
Sample 2.1
Intensity (a.u.)
Sample 2.2
Sample 2.3
Sample 2.4
2300
2200
2100
2000
190
-1
ν (cm )
The FTIR spectra, shown in figure 2.2, all show characteristic absorption bands in
the CN stretching region. The positions and line widths of these absorptions are
listed in table 2.4.
Assignments are based on literature data.19,20 From a comparative analysis of CN
stretching vibrations for various M-(CN)-M’ derivatives it was concluded that the
position of the υCN vibration appears to be considerably more sensitive to the
C-bound metal ion and its oxidation state than it is to the N-bound metal ion and its
29
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
Table 2.4. FTIR spectroscopic data (CN stretching bands) for RbxMn[Fe(CN)6]y·zH2O
Sample
2.1
2.2
2.3
2.4
Position (FWHM)a band 1
(cm-1)
2151 (40)
2152 (7)
2154 c
2152 (26)
Position (FWHM)a band 2
(cm-1)
Position (FWHM)a band 3
(cm-1)
2072 (82)
2101 (91)
2101 (88)
2090 (108)
b
b
b
Sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample
2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
a
Full Width at Half Maximum
b
Band 2 is composed of two bands: one around ~2100 cm-1 and one around ~2075 cm-1, the first is
the most pronounced
c
Very weak
valency.19 This is also the case for cyanide groups establishing a linkage between
two different metal ions in varying valence states.20
In fact, the absorption around 2150 cm-1 was assigned to the stretching of the CN
ligand bridged between FeIII and MnII (FeIII-CN-MnII), while the absorptions
around 2100 cm-1 and 2075 cm-1 were assigned as the two vibrations with different
symmetry of the CN ligand between FeII and MnIII (FeII-CN-MnIII).13b,21 Other CN
vibrations are not directly distinguishable. The different CN vibrations are not very
pronounced, this is probably due to overlap with OH vibrations of water molecules,
although the vibrations of water are not situated near the CN stretching vibration
frequencies. Depending on the extent of electron transfer that has occurred, both
the high temperature (high spin MnII and low spin FeIII) and low temperature
electronic configuration (high spin MnIII and low spin FeII) may be present at room
temperature.13b It has to be noted that the absorption intensity of the band around
2075 cm-1 (indicative for the LT configuration) is much larger than that of the
bands for the HT configuration. This means that a small amount of LT phase will
already give a relatively large absorption at 2075 cm-1. Sample 2.3 shows the
largest amount of LT phase in the refinement of the X-ray powder diffraction, and
this is confirmed in the IR spectrum of this sample which shows the largest peak
around 2075 cm-1. Care has to be taken when directly comparing the data obtained
from IR spectra and from X-ray powder diffraction: the latter shows only the
results of the crystalline fraction, while the first also shows the amorphous fraction.
It is interesting to notice that the position of cyano stretching vibration 1 remains
virtually unaltered throughout the series and its position is in agreement with the
assignment mentioned above, whereas the combined vibrations 2 and 3 seem to
reveal a dependence on the composition of the sample. In fact, the IR data of
samples 2.2, 2.3 and 2.4 having a lattice water content close to 1 show a close
similarity with the IR spectroscopic results reported for RbMn[Fe(CN)6] 13b,21 in
that they exhibit a rather sharp absorption at around 2150 cm-1 together with a
relatively broad band composed of two absorptions around 2100 cm-1 and
2075 cm-1.
30
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
The close to 2.6 fold hydrated sample 2.1 also displays a sharp peak at about
2150 cm-1 (vibration 1), whereas vibration 2 (2100 cm-1) is now almost not present.
This shift to lower energy would be consistent with a weakening of the CN bond,
implying that increased hydration of the material might involve a slight elongation
of the CN linkage within the FeII-CN-MnIII framework.
Unfortunately, an assignment of the observed vibrational bands in terms of specific
vibrational modes which are purely based on group theory considerations is
severely hampered by the non-stoichiometric nature of the materials. The presence
of water molecules, different sites for the Rb+ ions, defects such as missing
Fe(CN)6 units, etc. eventually lead to a lowering of the local symmetry. This is true
for both the IR and the Raman spectroscopic data.
In contrast to IR absorption spectra, Raman spectra are not obscured by water lines,
which gives a cleaner access to specific molecular vibrations. Raman spectra of the
studied compounds in the spectral region 2000-2300 cm-1 are dominated by two
strong Raman bands which are centred in the vicinity of 2165 cm-1 and 2156 cm-1
(see figure 2.3). In addition, a rather weak Raman band (~5% of the total
amplitude) is found around 2080 cm-1. In between 2080 cm-1 and 2156 cm-1 an
irregular background is observed for some samples. The measured Raman spectra
can be interpreted in analogy with the IR absorption data13b where the CN
stretching mode corresponding to the IR band centred around 2150 cm-1 is
attributed to CN ligands linking FeIII and MnII ions, while the broad IR band
Figure 2.3. Room temperature Raman
spectra of sample 2.1
(Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample
2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O),
sample 2.3
(Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and
sample 2.4
(Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
Intensity (a.u.)
Sample 2.1
Sample 2.2
Sample 2.3
Sample 2.4
2050
2100
2150
ν (cm-1)
2200
2250
31
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
centred around 2095 cm-1 is assigned to CN ligands located between FeII and MnIII
ions. The band around 2080 cm-1 should probably be assigned as the CN ligand
vibrating between FeII-CN-MnII.22
In table 2.5 the parameters obtained by multi-Lorentzian fitting of the Raman
spectra depicted in figure 2.3 are presented. Sample 2.1 exhibits noticeably broader
Raman bands and the largest amount of “irregular background”, while the fit
parameters obtained for samples 2.2-4 are similar. Presumably the variations in
relative intensities and width of bands 1 and 2 for sample 2.1 are due to its
increased deviation from a perfect RbMn[Fe(CN)6] stoichiometry and associated
larger water content. The activation of vibrational modes by the lowering of the
local symmetry may give rise to various bands resulting in the observed broadening
of the Raman bands and the “irregular background”. This finding is also in
agreement with the correlation found by Cobo et al.: 23 the larger the deviation
from the perfect 1:1:1 stoichiometry, the broader the bands around 2170 and
2159 cm-1. It is unclear where the intensity differences between the two absorptions
come from and this has also not been investigated in literature.
Table 2.5. Raman spectroscopic data (CN stretching bands) for RbxMn[Fe(CN)6]y·zH2O
Sample
2.1
2.2
2.3
2.4
Position (FWHM)a band 1
(cm-1)
2165.7 (6.3)
2165.0 (3.6)
2165.0 (4.3)
2165.0 (4.5)
Position (FWHM)a band 2
(cm-1)
2156.3 (8.5)
2156.3 (4.2)
2156.1 (4.8)
2156.1 (4.7)
Position (FWHM)a band 3 (cm-1)
2080.6 (5.9)
2079.6 (4.6)
2080.1 (10.7)
2079.7 (4.4)
Sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample
2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
a
Full Width at Half Maximum
2.3.4 57Fe Mössbauer Spectroscopy
The 57Fe Mössbauer spectra recorded at 80 K of the four samples are displayed in
figure 2.4a. The fit parameters are shown in table 2.6. The experimental data were
analyzed with two sub components, one singlet (Site 1) and one doublet (Site 2).
The singlet corresponds to low spin FeII (S = 0) ions situated in a symmetrical
octahedral coordination sphere of CN ligands and the doublet to low spin FeIII
(S = ½) ions. The values of the hyperfine parameters are in agreement with those
reported in the literature.24 The negative isomer shift found for all singlets for low
spin FeII is comparable for the same species in NaxCo[Fe(CN)6]y·zH2O.25
The strongly bonded CN ligands shield effectively the iron atom from the outer
environment. The iron atom adopts a low spin configuration in both valence states.
The spectrum recorded for sample 2.1 at 80 K consists of a superposition of the
low spin FeII singlet and the low spin FeIII doublet both having the same isomer
shift. The area fractions of the resonance lines are 19(3)% for FeII and 81(3)% for
FeIII, which is in agreement with the magnetic data (see below) showing that the
material mainly consists of the HT configuration and does not exhibit any
32
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
temperature dependent electron transfer process upon cooling. The broadness of
the spectral features deserves particular attention, as the relatively large line widths
could be assigned to the presence of attached Mn ions which vary in valency and
also in geometry: MnII being about Oh and MnIII D4h (so Jahn-Teller distorted) and
this will give a range of slightly different Fe geometries.
b) Room temperature
a) 80 K
80 K
100
100
Sample 2.1
Sample 2.1
98
98
Relative Transmission (%)
Relative Transmission (%)
Room temperature
96
100
Sample 2.2
99
100
98
Sample 2.3
96
94
100
Sample 2.4
98
96
-4
96
100
Sample 2.2
99
100
98
Sample 2.3
96
94
100
Sample 2.4
98
96
-2
0
v (mm/s)
2
4
-4
-2
0
2
4
v (mm/s)
Figure 2.4. 57Fe Mössbauer spectra recorded at 80 K (a) and room temperature (b) for
sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O),
sample 2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4
(Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
The resonance lines for low spin FeII and low spin FeIII are considerably broadened
at 80 K. This feature together with the superposition of the low spin FeII and low
spin FeIII resonance lines, will severely hamper the exact determination of the
spectral contributions of both ions, particularly in those instances where the
spectral contribution of the low spin FeIII ions is relatively low. In fact, in the
spectra recorded at 80 K, the signature for the low spin FeIII doublet can still be
clearly distinguished for sample 2.2 and 2.4 in the form of shoulders at the outer
parts of the overlapping low spin FeII and low spin FeIII absorption lines. The
spectral contribution of low spin FeIII can therefore with certainty be estimated to
be of the order of 7(3)% for sample 2.2 and 29(3)% for sample 2.4. For sample 2.3,
33
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
this characteristic feature on which the low spin FeIII contribution can be based is
not clearly visible, since in this instance it is entirely obscured by the broad single
line originating from low spin FeII. Hence, it cannot be excluded that low spin FeIII
may contribute to a minor extent to the spectral density, however, its exact area
fraction cannot be accurately determined, so there remains an uncertainty in the
FeIII percentage present at this temperature.
Figure 2.4b shows the 57Fe Mössbauer spectra of the four samples at room
temperature. The spectrum of sample 2.1 consists of a doublet and a singlet. The
doublet is assigned to low spin FeIII, whereas the singlet is assigned to low spin
FeII. The relative areas of both species are 78(3)% and 22(3)% respectively, which
is more or less the same as the area fractions of these species at 80 K for this
sample. It thus confirms that the sample does not show any electron transfer.
The spectrum of sample 2.2 shows only a singlet at room temperature. On the basis
of the hyperfine parameters and the strong temperature dependence of the
quadrupole splitting, this singlet can be assigned either to LS FeII or LS FeIII in a
local perfect cubic environment.26 These two species cannot be distinguished solely
on the isomer shift. However, based on the Raman data which show mostly two
bands at 295 K (ascribed to CN stretching between HS MnII and LS FeIII),
combined with the high value found for χMT at room temperature we assign this
singlet to LS FeIII. Although quite rare, a singlet for LS FeIII in 57Fe Mössbauer
spectra has been reported before.27 In analogy to sample 2.2 the singlets in the 57Fe
Mössbauer spectra of samples 2.3 and 2.4 are assigned to low spin FeIII in perfect
cubic environment and the doublet in sample 2.4 to low spin FeIII with a slight
deviation of the cubic environment.
Table 2.6. 57Fe Μössbauer parameters resulting from least-squares fits of the spectra of
RbxMn[Fe(CN)6]y·zH2O
Site 1
Sample
δ (mm/s)
2.1
2.2
2.3
2.4
-0.04
-0.04
-0.03
-0.04
2.1
2.2
2.3
2.4
-0.17
-0.16
-0.16
-0.13
∆ (mm/s)
80 K
0.00
0.00
0.00
0.00
Room temperature
0.00
0.00
0.00
0.00
Site 2
Γ (mm/s)
A/Atot
(%)
0.48
0.37
0.32
0.36
0.32
0.36
0.30
0.28
δ (mm/s)
∆ (mm/s)
Γ (mm/s)
A/Atot
(%)
19
93
100
71
-0.08
-0.04
0.71
0.84
0.42
0.21
81
7
-0.05
0.76
0.33
29
22
100
100
70
-0.18
0.41
0.30
78
-0.12
0.40
0.29
30
Sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample
2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
The most probable errors for the hyperfine parameters are ±0.01 mm s-1 and for the area ±3%.
34
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
2.3.5 Magnetic Measurements
The magnetic behaviour of the four samples is shown in figure 2.5 in the form of
χMT versus T plots, with χM the molar magnetic susceptibility per formula unit and
T the temperature.
Assuming that the materials consist exclusively of either the LT or HT phase and
taking a g-value equal to 2, the theoretical χMT values would be 3.00 cm3 K mol-1
for an {S1;S2} = {2;0} system (high spin MnIII and low spin FeII) and 4.75 cm3 K
mol-1 for a paramagnetic {S1;S2} = {5/2;1/2} system (high spin MnII and low spin
FeIII). These values have not been found experimentally as all materials contain
both LT and HT configuration to various extents at all temperatures, resulting in an
underestimation of the χMT value at temperatures where the HT configuration is
expected, and an overestimation of these values where the LT configuration is
expected. Secondly, all RbxMn[Fe(CN)6]y·zH2O materials show significant,
although different deviations from the perfect Rb:Mn:Fe stoichiometry of 1:1:1. At
the extremes of the magnetic curves, a major fraction of the sample (denoted by y)
5.5
3
-1
χMT (cm K mol )
Sample 2.2
Sample 2.1
5.0
4.5
4.0
3.5
3.0
5.5
Sample 2.3
Sample 2.4
3
-1
χΜT (cm K mol )
5.0
4.5
4.0
3.5
3.0
0
50
100
150
200
T (K)
250
300
350
0
50
100
150
200
250
300
350
T (K)
Figure 2.5. Temperature dependence of χMT for sample 2.1
(Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample 2.3
(Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O)
35
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
is then either formed by an {S1;S2} = {2;0} system (high spin MnIII and low spin
FeII) or an {S1:S2} = {5/2;1/2} system (high spin MnII and low spin FeIII), whereas a
minor fraction (given by 1-y) is represented by non-magnetically coupled Mn(NC)6
entities, which most probably contain high spin MnII ions (S = 5/2). This has the
effect of underestimating the χMT value at temperatures where the HT configuration
is expected, while an overestimation of this value occurs at temperatures where the
LT configuration is expected.
The χMT versus T curve for sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O) indicates
that it does not exhibit a conversion from the high to the low temperature
configuration. To our knowledge, this is the first sample belonging to the family of
RbxMn[Fe(CN)6]y·zH2O compounds, that does not show the temperature-induced
interconversion between the high and the low temperature configuration.
The χMT value at high temperature, 4.07 cm3 K mol-1, is significantly lower than
the χMT value of 4.75 cm3 K mol-1 (calculated for g = 2) that would be expected for
a complete population of the HT configuration. This feature can be explained by
the observed significant lack (14%) of Fe(CN)6 entities and the presence of a
relatively small fraction of LT configuration. The presence of such a minor portion
consisting of high spin MnIII (S = 2) and low spin FeII (S = 0) is in agreement with
the results of various types of spectroscopic measurements reported in the present
chapter.
The existence of a small amount of LT phase can also be inferred from the χMT
behaviour at low temperature. In this temperature range the occurrence of a peak in
the χMT curve is due to the superposition of a signal due to the ferromagnetic
ordering of the residual LT phase and the predominant χMT signal of the HT phase
exhibiting a paramagnetic behaviour.
The data for sample 2.2 recorded during heating from 5 - 350 K first show a
decrease in χMT until it reaches the value of 3.07 cm3 K mol-1 at 255 K, then it
increases sharply to a value of 4.50 cm3 K mol-1 at 300 K. Upon cooling, the χMT
value abruptly decreases at 242 K to reach a value of 3.14 cm3 K mol-1 at 195 K. A
thermal hysteresis width of 52 K characterized by T1/2↓ = 234 K and T1/2↑ = 286 K
has been detected. The χMT versus T plot for sample 2.3 has a fairly similar shape,
however, the hysteresis width has increased to 57 K (T1/2↓ = 240 K, T1/2↑ = 297 K).
The limiting values for χMT encompassing this hysteresis loop are 3.17 cm3 K mol-1
at 277 K and 4.76 cm3 K mol-1 at 333 K.
Sample 2.4 exhibits a much broader hysteresis (width 86 K, T1/2↓ = 197 K, T1/2↑ =
283 K). In addition, the value of χMT for the LT phase is significantly higher (3.58
cm3 K mol-1), whereas the one for the HT phase is of a similar order (4.76 cm3 K
mol-1).
The magnetic behaviour of the electron transfer active compounds is comparable to
that reported for RbMn[Fe(CN)6] which displays a thermal hysteresis width of
36
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
73 K (T1/2↓ = 231 K, T1/2↑ = 304 K).13c The limiting values for this hysteresis loop
in χMT are 3.2 cm3 K mol-1 at 200 K and 4.7 cm3 K mol-1 at 320 K.
Examining these χMT values while taking into account that these samples are
somewhat short in spin-carrying Fe(CN)6 entities, leads to the conclusion that the
temperature-induced switching process is incomplete in all instances, with the
extent of electron transfer being the largest for sample 2.3. All compounds show
the general feature that the residual HT fraction present at low temperatures
appears to be larger than the residual LT fraction at higher temperatures.
Figure 2.6 shows the magnetisation versus temperature curves of the four samples.
These magnetisation curves were done in a field of 0.01 T which is not a saturated
field for this type of material (saturation occurs arround ~1 T). Sample 2.1 does not
show any significant magnetisation, but sample 2.2, 2.3 and 2.4 do exhibit
spontaneous magnetisation below Tc = 12 K. The order of the magnetisation
decreases from sample 2.3 to sample 2.2 to sample 2.4. The ordering temperature
of 12 K exactly corresponds to that what has been reported for RbMn[Fe(CN)6].21a
3000
60
Sample 2.1
Sample 2.2
2500
50
M (EMU/mol)
2000
40
1500
30
1000
20
500
10
0
0
5000
-500
400
Sample 2.3
4000
Sample 2.4
350
M (EMU/mol)
300
3000
250
200
2000
150
1000
100
50
0
-1000
0
2
4
6
8
10
12
T (K)
14
16
18
20
-50
2
4
6
8
10
12
14
16
18
20
T (K)
Figure 2.6. Temperature dependence of the magnetisation for sample 2.1
(Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample 2.3
(Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O). Notice
the differences in the scale of the y-axis. ■ = Zero Field Cooled, ○ = Field Cooled,
x = Remanence
37
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
Because only the LT configuration shows magnetisation below 12 K, the
magnitude of the magnetisation is determined by the extent to which the LT phase the {S1;S2} = {2;0} system consisting of high spin MnIII and low spin FeII - is
present at low temperatures. The 57Fe Mössbauer data recorded at 80 K give
valuable information in that the spectral contribution of low spin FeII varies with
the increasing magnetisation values, yielding 100% for sample 2.3, 93% for sample
2.2 and only 71% of low spin FeII ions for sample 2.4. An assessment of the
contribution of this LT electronic configuration based on the χMT value determined
at temperatures below the thermal hysteretic feature is hampered by it also
consisting of a fraction of residual HT electronic configuration and a small
percentage of high spin MnII ions. Both of these factors lead to the observation of a
χMT value that will be higher than expected for a pure LT configuration, however
an analysis may be given focusing on determining these various contributions to
χMT. The amount of high spin MnII ions is proportional to the number of Fe(CN)6
defects present in the various RbxMn[Fe(CN)6]y·zH2O compounds, hence its
contribution may be expected to vary as 1-y. This leads to the conclusion that the
correction for the MnII contribution will be the largest for sample 2.2 and 2.4,
whereas it should be somewhat smaller for sample 2.3. The contribution arising
from the residual HT configuration present at low temperatures may be estimated
from the 57Fe Mössbauer spectroscopic data recorded at 80 K, which show that the
low spin FeIII content - indicative for the fraction of HT phase - varies from sample
2.3 (0%), sample 2.2 (7%) to sample 2.4 (29%). Therefore, the purely LT
configuration-based χMT value lies closest to the value found for sample 2.3, and
the values found for sample 2.2 and 2.4 show a larger deviation from this. In view
of this analysis, the observed χMT values for the various samples do indeed reflect
the trend set out by the magnetisation values.
The inverse magnetic susceptibility data from 15-265 K (not shown here) were
fitted with a straight line yielding θ values of +9.0 ± 0.2 K, +9.6 ± 0.2 K and +3.2 ±
1.1 K for samples 2.2, 2.3, and 2.4, respectively. From the χMT values at the lower
limit of the hysteresis loops, as well as from the 57Fe Mössbauer spectroscopic data
recorded at 80 K, it is evident that sample 2.2 and 2.3 contain relatively small
amounts of HT configuration, whereas sample 2.4 contains a far larger percentage
of this configuration. In the determination of the θ values, an increasing amount of
HT configuration will result in an underestimation of the ferromagnetic coupling.
Despite this inaccuracy in the θ values, it can still be safely concluded that a weak
ferromagnetic coupling occurs between neighbouring metal ions in the LT phase.
This ferromagnetic exchange coupling in the LT phase has been proposed to arise
from a mechanism of mixed-valence electron delocalization of the Mn ions similar
as reported for the FeIII ions in Prussian Blue FeIII4[FeII(CN)6]3·14H2O.13b,28
2.3.6 Differential Scanning Calorimetry (DSC)
DSC measurements were carried out on 4.81 mg of sample 2.3
Rb0.97Mn[Fe(CN)6]0.98·1.03H2O. These measurements were performed in the
38
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
temperature range of 100-330 K for both heating and cooling processes. The heat
capacity anomalies corresponding to the HT ↔ LT transitions after subtracting the
baselines are shown in figure 2.7. The critical temperatures, defined as the
maximum of the peaks, are 244 K and 298 K for the cooling and heating mode,
respectively. These values are in good agreement with the characteristic transition
temperatures, T1/2↓ = 240 K and T1/2↑ = 297 K, obtained from the magnetic
measurements.23 The experimentally determined enthalpy and entropy variations
associated with the LT ↔ HT phase transition are ∆H = 18(1) kJ mol-1 and ∆S =
60(5) J K-1 mol-1 for the heating mode and ∆H = 19(1) kJ mol-1 and ∆S =
80(5) J K-1 mol-1 for the cooling mode. The higher values for the cooling mode are
related to the difficulty tracing the baseline, which must contribute the tail
observed in the heat capacity anomaly at low temperature.
Cooling
Heating
1500
-1
-1
∆Cp (J mol K )
2000
1000
Figure 2.7. Heat capacity anomalies
(on heating and cooling) corresponding
to the thermal induced transition for
sample 2.3
(Rb0.97Mn[Fe(CN)6]0.98·1.03H2O), after
subtraction of the baseline.
500
0
150
200
250
300
350
T(K)
Although the experimental values from Cobo et al.29 are of the same order
(∆H = 11 – 13.6 kJ mol-1, ∆S = 48 – 59 J K-1 mol-1) as our results, they are still
lower than our measured values. However the studied samples in reference 29 were
Rb0.58Mn[Fe(CN)6]0.82·4.04H2O and Rb0.85Mn[Fe(CN)6]0.96·1.40H2O, that is
compounds with vacancies at the Rb and Fe sites. The existence of vacancies in
RbxMn[Fe(CN)6]y·zH2O produces a incomplete conversion between the LT and HT
electronic configuration in the thermal phase transition, hence the determined
variation in thermodynamic parameters will necessarily be smaller than the actual
molar enthalpy and entropy changes for a complete phase transition. The spin-only
contribution to the change in entropy was calculated by Cobo et al.29 to be 7.3 J K-1
mol-1. Hence, the observed entropy change is largely due to the lattice and/or
vibrational changes.
2.3.7 X-Ray Photoelectron Spectroscopy (XPS)
This technique is a direct method to identify the oxidation states in compounds and
to give quantitative information about the elemental composition of the Prussian
Blue analogues.13e,30 Figure 2.8 shows the survey spectrum of sample 2.2 at room
39
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
N 1s
Figure 2.8. X-ray photoemission
survey spectrum of sample 2.2
(Rb0.92Mn[Fe(CN)6]0.95·1.03H2O).
Photoemission and Auger lines are
labelled.
C 1s
Mn KLL
Fe 2p
Fe LMM
Intensity (a.u.)
Mn 2p
1000
O 1s
Rb 3p
Rb 3d
Mn 3p
Fe 3p
800
600
400
200
0
Binding Energy (eV)
temperature. As labelled in the spectrum, characteristic photoelectron and Auger
peaks of Fe, Mn, C, N, Rb and O are clearly distinguishable. We now discuss the
different spectral regions corresponding to the various elements separately. Figure
2.9 shows the XP spectrum of the whole Fe 2p region (a) and a close-up on the Fe
2p3/2 region (b) for sample 2.2. The fit was obtained using a Tougaard background
(figure 2.9a) and Gaussian functions for each peak. The close-up of the Fe 2p3/2
peak (figure 2.9b) demonstrates that it consists of three contributions. One at
708.5 eV binding energy the FeII line, one at 710.1 eV the FeIII main line and the
last at 711.7 eV the satellite of FeIII at 1.6 eV higher binding energy than the main
peak, with an intensity of 0.22 times that of the main peak.31 We can therefore
compute the ratio of FeII/FeIII in sample 2.2 from the relative intensities of the
photoemission lines and obtain 69(1)% of FeIII and 31(1)% of FeII. Consequently,
the Mn 2p spectra should present spectral features representative of 69% of MnII
and 31% of MnIII. Figure 2.10 represents the Mn 2p region of the XP spectrum.
The two spin-orbit split contributions centred at 641.8 eV and 653.8 eV binding
energy are both very broad due to multiplet splitting.31,32 If we concentrate, as we
did for Fe, on the 2p3/2 contribution, we notice that according to the literature the
MnII and MnIII peaks are expected at 641.8 eV and 642.5 eV, respectively,13e
however, with our XP resolution we cannot separate these two lines. To verify that
we indeed obtain the expected ratio between MnII and MnIII, we therefore tried to
simulate the spectrum starting from the MnO and Mn2O3 spectra reported in the
literature31 and multiplying them by the respective percentages of FeIII and FeII
quoted above. This procedure reproduces the experimental data very well (dashed
line in figure 2.10) except for an increase of 0.8 eV in the binding energy
(compared to the value found in the literature for the manganese oxides) probably
due to the different environments present in the rubidium manganese
hexacyanoferrate compound. Thus, our spectral analysis is consistent with the
coexistence of two electronic configurations in the sample, one containing FeIII and
40
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Figure 2.9. X-ray photoemission
spectrum of sample 2.2
(Rb0.92Mn[Fe(CN)6]0.95·1.03H2O)
(a) and close-up of the Fe 2p3/2
region after background
subtraction (b). The fit was
performed using a Tougaard
background and Gaussian
functions for each peak.
Binding Energy (eV)
740
730
720
710
Intensity (a.u.)
a)Fe 2p
b)Fe 2p3/2
714
712
710
708
706
Binding Energy (eV)
II
Mn and the other containing FeII and MnIII. Importantly, the XP spectral analysis
confirms the presence of identical percentages of MnII and FeIII or MnIII and FeII.
This would be in agreement with the presence of distinct LT and HT
configurations. All samples were analyzed in the same way and the respective
percentages of FeIII are reported in table 2.7. The percentage of FeIII (indicative for
the HT-phase) increases from sample 2.1 (60%), via sample 2.2 (69%), sample 2.4
(75%), to sample 2.3 (76%). This order is the same as the order in the amount of
LT configuration present in the 57Fe Mössbauer spectra at 80 K, the relative
contribution of the band around 2150 cm-1 in IR spectroscopy at room temperature,
and the inverse of the extent of magnetisation at 5 K. The amount of LT phase
found by the X-ray powder diffraction does not confirm this trend.
Figure 2.10. X-ray photoemission
spectrum of the Mn 2p region for
sample 2.2
(Rb0.92Mn[Fe(CN)6]0.95·1.03H2O).
The simulated spectrum was
generated from the MnO and
Mn2O3 signals31 multiplied by the
respective percentages of FeIII and
FeII determined for the same
sample (details see text).
Mn 2p
Intensity (a.u.)
2
simulation
665
660
655
650
645
640
635
Binding Energy (eV)
41
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
Figure 2.11. X-ray photoemission
spectrum of the Fe 2p3/2 region for
sample 2.4
(Rb0.81Mn[Fe(CN)6]0.95·1.24H2O)
prepared in the dark.
Intensity (a.u.)
Fe 2p3/2
714
712
710
708
706
Binding Energy (eV)
Additionally we checked the influence of exposure to daylight during the synthesis.
The XP spectrum of Fe (figure 2.11) of sample 2.4 prepared (and measured) in the
“dark” still shows the presence of two components corresponding to FeII and FeIII
at the same binding energies as for sample 2.2 exposed to white light during
synthesis and XPS analysis.
Now a brief comment on rubidium: for all four samples the Rb 3d5/2 XPS spectra
(not shown) present a peak at 110.5 eV. In the Prussian Blue analogues the
rubidium atoms are non-coordinated in the interstitial sites (see figure 1.2) and
hence similar to those in alkali intercalation compounds. This explains why we find
a binding energy which coincides with that of Rb+ intercalated in TiSe2 and TiTe2
layered materials.33
The O 1s region of the photoemission spectra are shown in figure 2.12 for all four
samples. Four components are needed to reconstruct the spectral shape: the first
two components centred at 529.4 eV and 530.6 eV representing the coordinated
water to MnIII and MnII, respectively,34 i.e. in the absence of Fe(CN)6 entities the
octahedral coordination about the manganese ions is completed by aqua ligands;
the third component centred at 532.1 eV binding energy represents the unavoidable
contamination of the samples due to exposure to air,35 while the highest binding
energy component at 533.7 eV is due to the non-coordinated water molecules
inside the Prussian Blue analogue samples.36 The intensity of the peak
corresponding to non-coordinated water in samples 2.1 and 2.2 is higher than that
due to coordinated water, while the two intensities are comparable in sample 2.3. In
sample 2.4 the intensity of the peak corresponding to coordinated water is instead
higher than that of non-coordinated water. This means that samples 2.1 and 2.2
contain a higher amount of non-coordinated water compared to water at defects,
while in sample 2.4 the situation is reversed. In sample 2.3 the amount of noncoordinated water molecules is equal to the number of the defects. The relative
atomic percentages of oxygen from coordinated and non-coordinated water
determined from the O 1s peak areas are reported in table 2.7.
42
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Figure 2.12. X-ray photoemission
spectrum of the O 1s region for sample
2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O),
sample 2.2
(Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample
2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and
sample 2.4
(Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
Intensity (a.u.)
Sample 2.1
Sample 2.2
Sample 2.3
Sample 2.4
538
536
534 532 530 528
526
524
Binding Energy (eV)
If we compare now the elemental composition of the samples deduced from the Xray photoemission peak areas reported in table 2.7 (C has been excluded from this
analysis because a small amount of contamination is always present at the surface
of these air exposed samples37), we note a disagreement between the stoichiometric
formula obtained by elemental analysis and the atomic percentages deduced from
XPS measurements. These differences can be attributed to the fact that elemental
analysis is a bulk sensitive technique while XPS is a surface probe. In addition,
because XPS has to be performed in ultra-high vacuum, the samples loose water
from the surface of the sample.
Table 2.7. Atomic composition of the samples as deduced from the XPS analysis. The data
are referenced to the number of nitrogen atoms, assumed to be 6.
Sample
Fe
Mn
Rb
N
O
non-coordinated
water molecules
2.1
2.2
2.3
2.4
1.0(1)
1.0(1)
1.1(1)
1.1(1)
1.0(1)
1.0(1)
1.0(1)
1.1(1)
0.94(5)
1.02(5)
1.02(5)
0.69(3)
6
6
6
6
0.47(5)
0.50(5)
0.25(3)
0.16(2)
O
coordinated
water
molecules
0.23(3)
0.27(3)
0.34(4)
0.69(5)
FeIII (%)
60(1)
69(1)
76(1)
75(1)
Sample 2.1 (Rb0.59Mn[Fe(CN)6]0.86·2.63H2O), sample 2.2 (Rb0.92Mn[Fe(CN)6]0.95·1.03H2O), sample
2.3 (Rb0.97Mn[Fe(CN)6]0.98·1.03H2O) and sample 2.4 (Rb0.81Mn[Fe(CN)6]0.95·1.24H2O).
2.4 Conclusions
The compounds of general formula RbxMn[Fe(CN)6]y·zH2O clearly represent an
intriguing class of compounds. The obtained compositions of the
RbxMn[Fe(CN)6]y·zH2O compounds in terms of their differing in stoichiometry and
water content may be explained by the variation in the synthetic procedures
involved. Complexation of MnII and FeIII(CN)6 is extremely fast and irreversible as
43
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
can be seen from the instantaneous precipitation of the material when two solutions
containing each ingredient are added together. Therefore, concentration effects are
extremely important in determining the nature of the product formed. So, when
starting with a solution of MnCl2·4H2O and adding K3Fe(CN)6 (as in sample 2.1),
the MnII concentration is high at the initial stages of the precipitation and FeIII(CN)6
defects will be formed within the material. Instead of N-donating FeIII(CN)6
entities, water molecules will complete the presumably still octahedral coordination
about the MnII ion. On the contrary, when the solutions are added in a reversed
order (as in sample 2.2), less FeIII(CN)6 defects will be formed. When the solutions
are added simultaneously (sample 2.3), the product formation process shows little
difference compared to the one in which MnCl2·4H2O is added to K3Fe(CN)6.
From the results reported previously13,21,24 one could have concluded that singlephase materials had exclusively been obtained and that the solid state material
would undergo an electron transfer process enabling the interconversion between
the so-called high temperature (high spin MnII and low spin FeIII) and low
temperature phase (high spin MnIII and low spin FeII). This transformation is
accompanied by a change from cubic to tetragonal symmetry, which is associated
with the Jahn-Teller distortion of the MnIII centre. The present results point into the
direction that these redox processes do occur in a much more subtle way and can to
a certain extent be controlled during the synthesis, where concentration effects may
have a paramount influence on the nature of the resulting solid state material. In
fact, it may be concluded that redox processes have occurred to varying extents
during the synthesis and this is confirmed by all solid compounds containing both
LT and HT configuration to varying degrees at room temperature. The HT phase
clearly predominates at this temperature as indicated by the relatively high χMT
values obtained for all compounds at room temperature. From the evolution of
these χMT values, the fraction of LT configuration present may be estimated to
decrease in the range from sample 2.1, sample 2.2, sample 2.4 to sample 2.3.
Because the relative contributions of both phases obtained from the X-ray powder
diffraction study may not be considered as extremely accurate, this sequence is not
found in this experimental procedure. The IR spectroscopic data agree with the
trend mentioned above in that the relative contribution of the higher frequency
absorption around 2150 cm-1 attributed to the FeII-CN-MnIII cyano linkage present
in the LT phase is most pronounced in sample 2.1, samples 2.2 and 2.4 and least in
sample 2.3. The Raman data also agree with this series, in that the relative
contribution of the band located at 2165 cm-1 increases in going from sample 2.2,
sample 2.3, sample 2.4 to sample 2.1. The same trend is observed for the band at
2080 cm-1.
The XPS experiments have been very valuable in having yielded direct evidence
for the coexistence of identical percentages of Fe and Mn ions in oppositely
matching oxidation states, in agreement with the presence of the FeII MnIII LT and
the FeIII MnII HT electronic configuration. Importantly, this information suggests
44
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
that the redox processes operative within these materials only involve electron
transfer from Mn to Fe and vice versa, in agreement with the proposed electron
transfer mechanism.
The magnetic data for the electron transfer active compounds show a unique trend
correlating with the LT configuration content at room temperature increasing from
sample 2.4, sample 2.3 to sample 2.2. Sample 2.4 containing the LT electronic
configuration to the lesser extent appears to exhibit the broadest hysteresis (86 K),
whereas the thermal hysteresis width diminishes considerably with increasing
fraction of the LT configuration yielding 57 K for sample 2.3 and only 52 K for
sample 2.2. Considering that the electron transfer transitions during heating and
cooling are first order in origin, the temperatures at half maximum may be
determined by the extent to which domains of non-active electron transfer entities
are present in the sample. This may lead to a reduced cooperative behaviour, which
might be comparable to for instance metal dilution effects observed in FeII spin
crossover compounds.38
Sample 2.1 represents the first example within the family of
RbxMn[Fe(CN)6]y·zH2O compounds of a material that does not exhibit the
temperature-induced interconversion between the high and the low temperature
configuration. This difference in magnetic behaviour compared to that of the other
samples may be related to its composition showing the largest deviation from a
perfect RbMn[Fe(CN)6] stoichiometry and an associated considerably larger
number of lattice water molecules.
These factors seem to be the key towards explaining the observed difference in
physical behaviour of the samples. The closer the sample has a ratio of Rb:Mn:Fe
to 1:1:1, the less water is present. One water molecule is positioned on the Rb+
vacancies. The remaining water molecules fill the positions of the Fe(CN)6 defects
as depicted in figure 2.1. The latter has the consequence that the Mn ions get more
water molecules in their coordination shell resulting in the formation of presumably still octahedral - Mn(NC)6-a(H2O)a entities. The replacement of the
relatively strong N-donating cyano ligands by the weaker aqua donors will lead to
a significant lowering of the ligand field strength. Another important consequence
of this is that the redox potential of the Mn entity will change dramatically, such
that the close to stoichiometric compounds (sample 2.2, 2.3 and 2.4) containing the
MnII(NC)6/MnIII(NC)6 redox couple display electron transfer behaviour, whereas
sample 2.1 containing MnII(NC)6-a(H2O)a/MnIII(NC)6-a(H2O)a does not show any
sign of the occurrence of electron transfer processes.
The magnetisation versus temperature curves for the electron transfer active
compounds show spontaneous magnetisation below Tc = 12 K; the extent of this
magnetisation inversely correlates with the residual fraction of HT electronic
configuration present at lower temperatures. The 57Fe Mössbauer spectroscopic
data recorded at 80 K show that the low spin FeIII content - indicative for the
45
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
fraction of HT phase - varies from sample 2.3 (0%), sample 2.2 (7%) to sample 2.4
(29%). This increasing amount of residual HT configuration within the series
correlates with the increased hydration of the samples. In fact, the electron transfer
active materials contain two forms of HT electronic configurations at higher
temperatures, i.e. redox-active MnII(NC)6 entities, that are converted to the LT
configuration on cooling, and redox-inactive MnII(NC)6-a(H2O)a chromophores
responsible for the residual HT fraction at low temperatures. The relative
contribution of the latter is enhanced upon increased hydration of the samples.
This study has shown in which way small, but significant, variations in the
composition of the materials may be obtained by modifications in the synthetic
procedure. These adjustments of the molecular composition allow the controlled
tuning and optimization of the physical properties - the extent of magnetisation and
the thermal hysteresis width - associated with the metal-to-metal charge transfer
process within this family of RbxMn[Fe(CN)6]y·zH2O Prussian Blue-type
analogues.
2.5 Acknowledgements
The Raman spectra in this chapter have been measured by Audrius Pugzlyz, the
57
Fe Mössbauer spectra by Thomas Bakas, the DSC measurements have been done
by Miguel Castro and the XPS spectra have been measured by Enrico Maccallini.
2.6 References
1
2
3
4
5
6
7
8
46
This chapter has been based on the following publication: Vertelman, E.J.M.; Maccallini, E.;
Gournis, D.; Rudolf, P.; Bakas, T.; Luzon, J.; Broer, R.; Pugzlyz, A.; Lummen, T.T.A.;
Loosdrecht, P.H.M. van; Koningsbruggen, P.J. van; Chem. Mater. 2006, 18, 1951
Moritomo, Y.; Hanawa, M.; Ohishi, Y.; Kato, K.; Takata, M.; Kuriki, A.; Nishibori, E.; Sakata,
M.; Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K., Phys.Rev.B, 2003, 68, 144106
a) Ohkoshi, S.-I.; Tokoro, H.; Utsunomiya, M.; Mizuno, M.; Abe, M.; Hashimoto, K.,
J.Phys.Chem.B, 2002, 106, 2423; b) Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.;
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., J.Phys.Soc.Jpn., 2002, 71, 2078; c) Kato, K.;
Morimoto, Y.; Takata, M.; Sakata, M.; Umekawa, M.; Hamada, N.; Ohkoshi, S.-I.; Tokoro, H.;
Hashimoto, K., Phys.Rev.Lett., 2003, 91, 25502; d) Moritomo, Y.; Kuriki, A.; Ohoyama, K.;
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.; Hamada, N., J.Phys.Soc.Jpn., 2003, 72, 456; e)
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., Appl.Phys.Lett., 2003, 82, 1245; f) Tokoro, H.;
Ohkoshi, S.-I.; Matsuda, T.; Hashimoto, K., Inorg.Chem., 2004, 43, 5231; g) Ohkoshi, S.-I.;
Matsuda, T.; Tokoro, H.; Hashimoto, K., Chem.Mater., 2005, 17, 81; h) Moritomo, Y.; Kato, K.;
Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., J.Phys.Soc.Jpn.,
2003, 72, 2698
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K., J.Phys.Soc.Jpn., 2003, 72, 2698
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., Appl.Phys.Lett., 2003, 82, 1245
Margadonna, S.; Prassides, K.; Fitch, A. N., Angew.Chem.Int.Ed., 2004, 43, 6316
See section 1.3.6
Section 1.3.5; a) Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K., Science, 1996, 271, 49; b)
Bleuzen, A.; Lomenech, C.; Escax, V.; Villain, F.; Varret, F.; Cartier dit Moulin, C.; Verdaguer,
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
M., J.Am.Chem.Soc., 2000, 122, 6648; c) Cartier dit Moulin, C.; Villain, F.; Bleuzen, A.; Arrio,
M.-A.; Sainctavit, P.; Lomenech, C.; Escax, V.; Baudelet, F.; Dartyge, E.; Gallet, J.-J.;
Verdaguer, M., J.Am.Chem.Soc., 2000, 122, 6653; d) Escax, V.; Bleuzen, A.; Cartier dit Moulin,
C.; Villain, F.; Goujon, A.; Varret, F.; Verdaguer, M., J.Am.Chem.Soc., 2001, 123, 12536; e)
Champion, G.; Escax, V.; Cartier dit Moulin, C.; Bleuzen, A.; Villain, F.; Baudelet, F.; Dartyge,
E.; Verdaguer, M., J.Am.Chem.Soc., 2001, 123, 12544; f) Verdaguer, M., Science, 1996, 272,
698
Larson, A. C.; Von Dreele, R. B. General Structure Analysis System (GSAS) 2004, Los Alamos
National Laboratory Report LAUR 86-748.
Kato, K.; Morimoto, Y.; Takata, M.; Sakata, M.; Umekawa, M.; Hamada, N.; Ohkoshi, S.-I.;
Tokoro, H.; Hashimoto, K., Phys.Rev.Lett., 2003, 91, 25502
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K., J.Phys.Soc.Jpn., 2002, 71, 2078
a) Cataldi, T. R. I.; De Benedetto, G. E.; Bianchini, A., J.Electroanalytical Chem., 1998, 448,
111; b) Yatsimirskii, K. B.; Nemoshkalenko, V. V.; Nazarenko, Y.,
J.Electron.Spectrosc.Relat.Phenom., 1977, 10, 239
a) Moritomo, Y.; Hanawa, M.; Ohishi, Y.; Kato, K.; Takata, M.; Kuriki, A.; Nishibori, E.;
Sakata, M.; Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K., Phys.Rev.B, 2003, 68, 144106; b)
Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K., Coord.Chem.Rev., 2005, 249, 1830, (And references
therein) c) Ohkoshi, S.-I.; Tokoro, H.; Utsunomiya, M.; Mizuno, M.; Abe, M.; Hashimoto, K.,
J.Phys.Chem.B, 2002, 106, 2423; d) Moritomo, Y.; Kuriki, A.; Ohoyama, K.; Tokoro, H.;
Ohkoshi, S.-I.; Hashimoto, K.; Hamada, N., J.Phys.Soc.Jpn., 2003, 72, 456; e) Tokoro, H.;
Ohkoshi, S.-I.; Matsuda, T.; Hashimoto, K., Inorg.Chem., 2004, 43, 5231; f) Yokoyama, T.;
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.; Okamoto, K.; Ohta, T., Phys.Rev.B, 2002, 66,
184111; g) Osawa, H.; Iwazumi, T.; Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.; Shoji, H.; Hirai,
E.; Nakamura, T.; Nanao, S.; Isozumi, Y., Solid State Comm, 2003, 125, 237; h) Tokoro, H.;
Ohkoshi, S.-I.; Matsuda, T.; Hozumi, T.; Hashimoto, K., Chem.Phys.Lett., 2004, 388, 379; i)
Umekawa, M.; Hamada, N.; Kodama, A.; Moritomo, Y., J.Phys.Soc.Jpn., 2004, 73, 430
Margadonna, S.; Prassides, K.; Fitch, A. N., Angew.Chem.Int.Ed., 2004, 43, 6316
International tables for Crystallography, Edited by Authier, A., Kluwer Academic Publishers,
2003
The measured, calculated and difference diffraction profiles can be found in appendix I.
a) Chapter 5; b) Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I.; Z. Anorg. Allg. Chem.
2007, 633, 1134; c) Matsuda, T.; Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I.; Acta
Cryst. E 2008, i11
Nakamoto, K. In: Spectra of inorganic and coordination compounds; Wiley: New York, 1986,
ed: 4
Buschmann, W. E.; Ensling, J.; Gütlich, P.; Miller, J. S., Chem.Eur.J., 1999, 5, 3019
Sato, O.; Einaga, Y.; Fujishima, A.; Hashimoto, K., Inorg.Chem., 1999, 38, 4405
a) Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., Appl.Phys.Lett., 2003, 82, 1245; b) Ohkoshi, S.-I.;
Matsuda, T.; Tokoro, H.; Hashimoto, K., Chem.Mater., 2005, 17, 81
a) Bertrán, J. F., Pascual, J.B., Hernández, M., Rodríguez, R., Reactivity of Solids, 1988, 5, 95100; b) Nakamoto, K., Infrared and Raman Spectra of Inorganic and Coordination Compounds,
Wiley-Interscience, 1986, ISBN 0-471-01066-9; c) Reguera. E. , Bertrán J.F., Díaz, C., Blanco,
J., Rondón, S., Hyperfine Interact., 1990, 53, 391; d) Shinamoto, N., Ohkoshi, S.-I., Sato, O.,
Hashimoto, K., Chem. Lett. 2002, 31, 486; e) Sato, O., Einaga, Y., Fujishima, A., Hashimoto, K.,
Inorg. Chem. 1999, 38, 4405; f) Hester, R. E., Nour, E. M., J. Chem. Soc., Dalton Trans. 1981,
939
Cobo S.; Fernandez R.; Salmon L.; Molnar G.; Bousseksou A. Eur. J. Inorg. Chem. 2007, 1549
47
Chapter 2. The Influence of Defects on the Electron Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
48
a) Sato, O.; Einaga, Y.; Fujishima, A.; Hashimoto, K., Inorg.Chem., 1999, 38, 4405; b)
Greenwood, N. N.; Gibb, T. C. In: Mössbauer spectroscopy; Chapman and Hall: London, 1971
a) Einaga, Y.; Sato, O.; Iyoda, T.; Kobayashi, Y.; Ambe, F.; Hashimoto, K.; Fujishima, A.;
RIKEN Rev. 1997, 16, 41; b) Gawali-Salunke, S.; Varret, F.; Maurin, I.; Enachescu, C.;
Malarova, M.; Boukheddaden, K.; Codjovi, E.; Tokoro, H.; Ohkoshi, S.; Hashimoto, K.;
J.Phys.Chem.B 2005, 109, 8251
N. N. Greenwood, T. C. Gibb, Mössbauer spectroscopy, Chapman and Hall Ltd., London 1971.
a) He, Y., Tang, G., Liang, F., Huang, Y., Chen, Z., Phys. B, 2007, 393, 143; b) Reguera, E.,
Fernández-Bertrán, J., Hyperfine Interactions, 1994, 88, 49-58
Mayoh, B.; Day, P., J.Chem.Soc.Dalton Trans., 1976, 1483
Cobo S.; Fernandez R.; Salmon L.; Molnar G.; Bousseksou A. Eur. J. Inorg. Chem. 2007, 1549
Sauter, S.; Wittstock, G.; Szargan, R., Phys.Chem.Chem.Phys., 2001, 3, 562
Oku, M.; Wagatsuma, K.; Konishi, T., J.Electron.Spectrosc.Relat.Phenom., 1999, 99, 277
a) Oku, M.; Hirokawa, K.; Ikeda, S., J.Electron.Spectrosc.Relat.Phenom., 1975, 6, 451; b) Oku,
M.; Matsuta, H.; Wagatsuma, K., J.Chem.Soc.Faraday Trans., 1996, 92, 2759
Stoltz, S. E.; Starnberg, H. I.; Holleboom, L. J., Phys.Rev.B, 2005, 71, 125403
a) Feliu, S.; Perez-Revenga, M. L., Appl.Surface Science, 2004, 229, 112 ; b) Katsoyiannis, I. A.;
Zouboulis, A. I., Water research, 2004, 38, 1992
Barr, T. L.; Yin, M., J.Vac.Sci.Technol.A, 1992, 10, 2788
Andersson, K.; Nikitin, A.; Pettersson, L. G. M.; Nilsson, A.; Ogasawara, H., Phys.Rev.Lett.,
2004, 93, 196101
Barr, T. L. In: Practical Surface Analysis; Ed: Briggs, D. and Seah, M. P.; John Wiley:
Chinchester, 1990.
a) Jung, J.; Schmitt, G.; Wiehl, L.; Knorr, K.; Spiering, H.; Gütlich, P., Z.Phys.B, 1996, 100,
523; b) Martin, J.-P.; Zarembowitch, J.; Dworkin, A.; Haasnoot, J. G.; Codjovi, E., Inorg.Chem.,
1994, 33, 2617; c) Martin, J.-P.; Zarembowitch, J.; Bousseksou, A.; Dworkin, A.; Haasnoot, J.
G.; Varret, F., Inorg.Chem., 1994, 33, 6325
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 3. The Influence of Synthetic Conditions on the
Physical Properties of RbxMn[Fe(CN)6]y·zH2O Prussian
Blue Analogues
3.1 Introduction
As was demonstrated in the previous chapter, the precise stoichiometry of the
RbxMn[Fe(CN)6]y·zH2O materials is of tremendous importance in determining the
physical properties. The more the stoichiometry approaches the ideal Rb:Mn:Fe =
1:1:1 formula, the more complete the reversible electron transfer from the high
temperature (HT) phase (LS FeIII and HS MnII) to the low temperature (LT) phase
(LS FeII and HS MnIII) takes place. Therefore the need for well-defined reaction
conditions that lead to a specific and reproducible stoichiometry is high.
Unfortunately, the literature shows that the compositions of Prussian Blue
analogues are not so easy to reproduce. For instance, Cobo et al.1 claim to have
performed at least 3 syntheses for samples in the exact same way, but nonetheless
do not find the same stoichiometry and consequently the compounds do not have
the same physical properties. Since their synthetic details do not elaborate on the
addition speed or the stirring speed, a difference in these parameters during their
synthesis might be the reason for the different stoichiometries. The variation in the
specific addition speed and the stirring speed might also be the reason that sample
2.2 and 2.4 in the previous chapter do not have the same stoichiometry, although
this might have been expected. The influence of the concentration of the starting
materials on the stoichiometry of RbxMn[Fe(CN)6]y·zH2O has been demonstrated
previously:2 the lower the concentration of RbCl during the synthesis (with
constant MnCl2 and K3[Fe(CN)6] concentration), the more Rb and [Fe(CN)6]
defects occur in the resulting material.
In this chapter attempts have been made to find the synthetic conditions for a
reproducible stoichiometry. Special attention is given to the influence of
temperature during the synthesis (samples 3.1-6) and the addition speed (samples
3.7-11). Two more samples (sample 3.12-13) have been made to check for
reproducibility. All other conditions (for instance, concentration of the starting
materials, stirring speed, glassware and used heaters) have been kept exactly equal.
It will be shown that the temperature during the synthesis has little influence
(although it should not be too high) and that the most influence on the
stoichiometry and consequently the physical properties has the addition speed.
3.2 Experimental Section
3.2 1. Synthesis
All chemicals were purchased from Sigma-Aldrich, of analytical grade and used
without further purification. A solution of 0.495 g MnCl2·4H2O in 25 mL of H2O
49
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
on
the
Physical
Properties
of
(2.5 mmol, 0.1 M) was added to a mixed solution of 0.823 g K3[Fe(CN)6]
(2.5 mmol, 0.1 M) and 3.023 g RbCl (25 mmol, 1 M) in 25 mL of H2O. The
addition speed was kept constant at various defined speeds with a syringe pump
model 352 of Sage Instruments (see table 3.1). The solution was stirred at a
constant speed of 5.5 rps. The temperature of the combined RbCl and K3[Fe(CN)6]
solution was kept constant at selected temperatures for the various samples with
either an ice-, water-, or an oil bath (see table 3.1). The temperature of samples 3.7
– 13 was kept at 45ºC, because it was noted that in samples 3.1 – 6 the sample with
this temperature had the best stoichiometry (i.e. Rb:Mn:Fe closest to 1:1:1) and the
electron transfer properties were the most complete. In order to prevent evaporation
of the water at higher temperatures and to make sure that all other reaction
conditions were equal for all samples a reflux condenser was placed on top of the
reaction vessel. A brown powder precipitated and this was centrifuged and washed
twice with distilled water of room temperature. The samples were dried over night
in vacuum.
Table 3.1. Details of the synthesis of the various samples.
Sample
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
Addition speed (mL h-1)
6
6
6
6
6
6
Instantaneous (~5s for 25 mL)
30
15
6
3
6
6
Temperature (oC)
4
24
45
64
74
91
45
45
45
45
45
45
45
3.2.2 Elemental Analysis
For details on the elemental analysis, see section 2.2.2.
3.2.3 X-Ray Powder Diffraction
For details on the X-ray powder diffraction measurements, see section 2.2.3.
3.2.4 FTIR Spectroscopy
For details on the FTIR spectroscopy measurements, see section 2.2.4
3.2.5 Magnetic Measurements
For details on the magnetic susceptibility measurements, see section 2.2.7.
After the magnetic susceptibility measurements the magnetisation measurements
were carried out in which the field was kept constant at 0.01 T while the
temperature was varied from 20 K to 5 K.
50
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
3.3 Results and Discussion
3.3.1 Elemental Analysis
Table 3.2 shows the analytical data for the various samples. In all cases the
calculated Rb, Mn, Fe and N weight percentages correspond well with the
experimental weight percentages. The experimental C content is often (but
somehow not always) higher than the calculated values (~0.5%). It is not known
how this is possible, but since the fraction of C is never varied on its own, but
always coupled with the fractions of Fe and N in [Fe(CN)6] it is still believed that
the proposed formulas are correct. The experimental weight percentage of H is
often lower than the calculated weight percentage. The absolute value of H is small
and therefore hard to determine with great accuracy.
In order to check for reproducibility in the composition, in total 4 samples were
prepared in exactly the same manner: samples 3.3, 3.10, 3.12 and 3.13. The found
Rb:Mn:Fe ratios for these four samples are all very close: the fraction of Rb ions
per Mn ion varies from 0.91 to 0.96 whereas the Fe fraction varies from 0.97 to
Table 3.2. Experimental (exp) and calculated (cal) weight percentages of the various samples.
Sample
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Rb
Mn
Fe
C
N
H
20.13
20.19
20.76
20.76
21.26
21.26
20.63
20.63
21.34
21.34
21.19
21.19
15.49
15.49
20.51
20.13
20.72
20.72
22.07
22.07
23.20
23.20
22.12
22.12
22.79
22.79
15.00
15.00
15.00
15.00
15.03
15.03
14.18
14.46
15.14
15.14
14.27
14.27
16.20
16.20
15.33
15.69
15.71
15.71
15.52
15.52
15.06
14.77
15.32
15.32
15.33
15.33
14.69
14.56
14.87
14.69
14.80
14.80
14.32
14.29
14.88
14.81
14.49
14.34
14.26
14.21
15.02
15.02
15.03
15.03
15.24
15.34
14.95
15.07
15.45
15.45
15.44
15.32
20.14
18.79
20.29
18.95
20.23
19.10
20.33
18.44
20.16
19.11
18.50
18.50
19.02
18.34
20.03
19.38
20.07
19.39
19.67
19.91
20.30
19.44
20.36
19.94
20.34
19.76
22.30
21.91
22.64
22.10
22.63
22.27
22.70
21.51
22.28
22.28
21.04
21.57
21.22
21.39
22.55
22.60
22.52
22.62
22.94
22.48
22.61
22.67
22.97
23.25
22.80
23.05
0.55
1.07
0.46
0.95
0.32
0.84
0.30
1.19
0.61
0.82
0.47
1.13
1.30
1.61
<0.3
0.80
0.62
0.73
0.51
0.30
<0.3
0.54
<0.2
0.44
<0.2
0.42
Proposed formula
Yield
(%.
based
on Mn)
Rb0.86Mn[Fe(CN)6]0.95·1.94H2O
90
Rb0.89Mn[Fe(CN)6]0.96·1.73H2O
>99
Rb0.91Mn[Fe(CN)6]0.97·1.53H2O
84
Rb0.92Mn[Fe(CN)6]0.97·2.25H2O
86
Rb0.91Mn[Fe(CN)6]0.96·1.48H2O
86
Rb0.95Mn[Fe(CN)6]0.99·2.16H2O
79
Rb0.61Mn[Fe(CN)6]0.86·2.71H2O
86
Rb0.82Mn[Fe(CN)6]0.94·1.39H2O
77
Rb0.85Mn[Fe(CN)6]0.94·1.27H2O
79
Rb0.91Mn[Fe(CN)6]0.97·0.90H2O
74
Rb1.01Mn[Fe(CN)6]1.00·1.00H2O
72
Rb0.93Mn[Fe(CN)6]0.99·0.78H2O
76
Rb0.96Mn[Fe(CN)6]0.98·0.75H2O
83
51
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
the
Physical
Properties
of
b)
1.02
1.02
1.00
1.00
0.98
0.98
0.96
0.96
Fe fraction
Fe fraction
a)
on
0.94
0.92
0.90
0.88
0.94
0.92
0.90
0.88
0.86
1
10
100
-1
Addition speed (mL h )
0.86
0
20
40
60
80
100
0
Temperature of Synthesis ( C)
Figure 3.1. Variation of the Fe fraction with a) addition speed at T = 45oC, the dashed line
represents a linear fit on the data points b) temperature of synthesis, at an addition speed of
6 mL h-1.
0.99. This relatively small variation in Rb and Fe fractions indicates that it is
possible to reproduce the stoichiometry of Prussian Blue analogues.
When looking at the influence of the temperature (see also figure 3.1b) it seems as
if there is a slight tendency for the stoichiometry to become more ‘perfect’ (i.e.
Rb:Mn:Fe becomes closer to 1:1:1) with higher temperature. This trend is not
completely clear, though. Especially when taking into account the variation in the
stoichiometry of samples 3.3, 3.10, 3.12 and 3.13 (the samples that were
synthesized in the exact same way).
The trend for the addition speed is much more pronounced: the slower the addition
speed, the more the stoichiometry becomes perfect (figure 3.1a). It was possible to
find an empirical relation for this trend for the given specific reaction conditions in
terms of the Fe:Mn ratio in the resulting material: Fe = 1.027(9) log[v] − 0.067(6) ,
Mn
with v the addition speed of the MnCl2·4H2O solution in mL h-1. When the addition
speed is very low it can be understood that the system has more time to replace all
H2O molecules with [Fe(CN)6] units in all 6 available positions around the Mn ion.
When this time is not sufficient the H2O molecule will not be replaced and thus the
Mn ion has less [Fe(CN)6] units in its coordination sphere.
On the other hand, the determined water content of the samples (figure 3.2a) shows
a much larger variation with changes in the temperature and addition speed during
the synthesis. The maximum water content that can be present in a sample with
RbxMn[Fe(CN)6]y·zH2O stoichiometry is (2-x) + 6(1-y), with the first term the
water molecules that are present at the interstitial sites and the second the water
molecules that are attached to Mn on a position of a [Fe(CN)6] defect, instead of
52
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
a)
b)
5.0
1.0
4.5
0.9
3.5
Rb fraction
H2O fraction
4.0
3.0
2.5
2.0
1.5
0.8
0.7
0.6
1.0
0.5
0.0
0.5
0.80
0.85
0.90
Fe fraction
0.95
1.00
0.85
0.90
0.95
1.00
Fe fraction
Figure 3.2. a) Variation of the water fraction with Fe fraction, dashed line represents the
correlation that was found by Cobo et al.1 for the H2O:Fe ratio. b) Variation of the Rb
fraction with Fe fraction, the dashed line represents charge neutrality. Squares are samples
3.1-6, stars samples 3.7-11 and circles samples 3.12-13
the N of the CN bridging ligand. Often (samples 3.1 – 7), the water content in the
samples is larger than expected on the basis of the proposed formula (for example
sample 3.7 has 2.71 water molecules, whereas 2.23 water molecules is expected).
Therefore, it might also be conceivable that water is positioned on the Fe site of a
[Fe(CN)6] defect. IR spectroscopy does show the presence of vibrations indicative
of H-bonded water molecules,3 thereby confirming this proposal. We thus have two
different types of water molecules: one that is attached to Mn and the other that is
not. Most likely, the bound water molecules will be much harder to release in
vacuum than the non coordinated water molecules. It is expected that these
attached water molecules will always be present in the samples, in order to ensure
the anticipated hexa coordination around the Mn ion. On the other hand, the non
coordinated water molecules might only be incorporated if they are present during
the bridging of the Mn and the Fe by the CN ligand, and since this bridging is
extremely fast, it can be understood that these water molecules make the difference
between the various water contents of the similar samples (samples 3.3, 3.10, 3.12
and 3.13). Furthermore, it can be possible that water molecules are incorporated in
the interstitial sites when subjected to a moist environment, although the IR spectra
of the samples did not change much in time.4,5 Nonetheless, the water content in
the present samples seems to be similar to the water-Fe correlation that was found
previously by Cobo et al.1 (dotted line in figure 3.2a) although most samples
studied in this chapter have more water molecules. Cobo et al. have compared the
stoichiometry of 16 samples of RbxMn[Fe(CN)6]y·zH2O, which were both from
53
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
on
the
Physical
Properties
of
literature and prepared by themselves (in the same way as the preparation of
sample 2.2), in order to arrive at this correlation.
In figure 3.2b the variation of the fraction of Rb with the Fe fraction is plotted. The
dashed line shows the expected composition for a neutral compound in which FeIII,
MnII, RbI and CN- are present. It can be seen that all points are close to this line,
indicating that little redox processes have taken place.
Since the variation of the Rb fraction and H2O fraction with Fe fraction is
consistent with expectations and literature,1 in the remaining of this chapter only
the variation of the various physical parameters with the Fe content will be
considered.
3.3.2 X-Ray Powder Diffraction
All diffraction profiles obtained at room temperature were fitted with the cubic
space group F-43m, the structure for the HT phase as found for the single crystals
(see chapter 5) and powdered samples.6 The starting positions and specific
restrictions for the refinement are shown in table 2.2. A maximum of O atoms
(from H2O) was placed on the N position where [Fe(CN)6] defects are present in
order to ensure an octahedral environment around the Mn ions. The Rb and the
remaining O atoms were both placed on the interstitial sites 4c and 4d. The
refinement of the fractions of these atoms was restricted in such a way that the total
fraction of Rb and the total fraction of O never exceeded the experimental value
deduced from the elemental analysis and that the total occupancy of the two sites
never exceeded 1. From the X-ray powder patterns simulated for most samples
small negative values were found for the Rb fraction on site 4d. Since negative
fractions are chemically meaningless, in this case the total fraction of Rb was
placed on position 4c and the fractions were kept constant. Once (sample 3.7) the
best agreement was obtained when the fractions of Rb1 = Rb2 and O1 = O2. In
order to stabilize the refinement the isotropic atomic displacement parameters
(Uiso) were restricted to vary in such a way that atoms placed on similar positions
had the same Uiso (i.e. UisoMn = UisoFe, UisoRb1 = UisoRb2 = UisoO1 = UisoO2 and
UisoN = UisoC = UisoO3). Again, the lineshapes of the reflections were fitted with a
pseudo-Voigt function in which the Lorentzian and Gaussian broadening
components were allowed to vary. In general all atoms are placed on special
positions, except for the N, C and the coordinated O. The C and N positions were
allowed to vary and from these positions the Fe-C, Mn-N and C-N distances were
calculated. The fraction of the coordinated O was too small to be able to find
anything meaningful from the refinement of its position.
For most samples taking into account this HT structure only was sufficient to fit the
diffraction profiles. Sometimes though, small peaks were visible around 2θ = 24.5o
and 35o. These peaks are indicative of the tetragonal structure for the LT phase,
I-4m2, as found by Moritomo et al..6b For the present samples this LT phase
present at room temperature should be considered as inactive LT phase rather than
54
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
LT phase that has not yet been transformed into HT phase. Most of the times these
peaks were too small to refine and it was thus not possible to determine an estimate
of the amount of LT structure present, except for samples 3.5 and 3.6. In these two
samples the same atomic positions as in the literature were used. The fractions of
the atoms were kept the same as the fractions that were already found by
refinement with the HT structure. Furthermore, the Uisos of all atoms were kept
equal (see table 2.2).
Selected details of the refined diffraction profiles can be found in table 3.3. The
measured, calculated and difference profile curves can be found in Appendix I. The
refinements should not be considered excellent, but, taking into account the
uncertainties such as the presence of the LT structure, the isotropic atomic
displacement parameters and the precise distribution of the Rb ions and O atoms, it
is sufficient for indicating the structure of the samples. Furthermore, the
determined lattice parameters (see below) and the Fe-C, Mn-N and C-N distances
are all comparable to values reported in literature for similar compounds.7 Again of
course with the addition that the lattice parameters are more reliable than the bond
distances due to the fact that for the determination of the lattice parameters only the
position of the reflections is important, whereas for the bond distances the
intensities also have to be taken into account.
Table 3.3. Selected details for the refined diffraction profiles with F-43m of the various
samples.
Mn-N
C-N (Å)
fRb in wRp
Rp
Reduced χ2
(Å)
4c
(%)
(%)
(#variables)
10.5565(4) 1.83(4)
2.304(32) 1.144(15) 0.86
18.79
14.09 7.497 (18)
3.1
10.5595(3) 1.88(5)
2.26(4)
1.142(16) 0.89
19.46
15.21 8.200 (18)
3.2*
10.5575(4) 1.88(6)
2.25(5)
1.141(17) 0.91
19.38
14.71 7.915 (18)
3.3*
10.5593(3) 1.88(6)
2.26(5)
1.142(17) 0.92
21.75
16.62 9.620 (18)
3.4*
10.5596(3) 1.90(5)
2.24(4)
1.137(15) 0.91
19.97
15.56 7.324 (26)
3.5**
2.25(6)
1.143(24) 0.94
25.73
18.57 13.98 (24)
3.6*** 10.5557(4) 1.89(8)
10.5462(4) 1.982(18) 2.164(15) 1.127(14) 0.305 15.53
11.63 3.994 (16)
3.7*
10.5564(6) 1.89(8)
2.25(7)
1.139(21) 0.82
20.45
14.95 7.493 (18)
3.8
10.5589(6) 1.89(8)
2.25(7)
1.138(23) 0.85
21.71
16.44 7.583 (18)
3.9*
10.5569(7) 1.92(6)
2.23(5)
1.131(20) 0.91
22.71
17.68 7.303 (17)
3.10
10.5612(4) 1.87(6)
2.27(5)
1.143(19) 1.00
22.78
17.22 9.324 (18)
3.11
10.5606(3) 1.90(5)
2.25(4)
1.139(16) 0.93
20.11
15.06 7.387 (18)
3.12
10.5614(3) 1.86(4)
2.275(33) 1.144(15) 0.96
17.72
13.73 6.491 (18)
3.13*
Standard deviations are shown in parentheses. See text for further details.
* Peaks show presence of I-4m2 space group, too small to fit and make an estimate of the amount.
** 5(2)% of I-4m2 space group present
*** 11(4)% of I-4m2 space group present
Sample
a (Å)
Fe-C (Å)
Since sample 3.5 (prepared at 74oC) has a contribution of the tetragonal structure
I-4m2 of 5(2)% and in sample 3.6 (prepared at 91oC) it is as high as 11(4)%, it
seems as if the higher the temperature during the synthesis, the more this LT
structure becomes present. Because in both these samples the contribution of these
55
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
on
the
Physical
Properties
of
peaks is still too low to obtain any accurate information about the Fe-C and Mn-N
distances, it is impossible to draw conclusions about the charges on the metal ions
and so it is hard to determine if this structure is indeed the LT phase. The presence
of this assumed LT structure at higher temperatures seems counterintuitive and the
reason behind these observations is unclear.
The value of the lattice parameter of the cubic space group F-43m versus Fe
content is shown in figure 3.3b. It seems as if the lower the Fe content, the smaller
the volume of the unit cell. This can be understood as a lower Fe content means a
lower Rb content and a higher H2O content, which means more flexibility in
general and thus apparently a smaller volume.
In figure 3.3a the Fe-C and Mn-N distances are shown. The in general short Fe-C
distances of around 1.89 Å, are indicative for LS FeIII, as is expected for the HT
phase. Taking into account the generally large standard deviations of ~0.06, they
are in nice agreement with the found Fe-C distances ranging from 1.929(4) Å
(determined at 293 K) to 1.918(1) Å (determined at 90 K) for single crystals of
RbxMn[Fe(CN)6]y·zH2O.7 Similarly, the Mn-N distances of around 2.25 Å indicate
HS MnII and are in agreement with the found Mn-N distances ranging from
2.205(5) Å (determined at 293 K) to 2.105(13) Å (determined at 90 K) in single
crystals.7,8
b)
a)
10.562
2.4
10.560
10.558
2.2
10.556
a (A)
Distances (A)
2.3
10.554
2.1
10.552
2.0
10.550
1.9
10.548
1.8
10.546
0.85
0.90
0.95
Fe fraction
1.00
0.85
0.90
0.95
1.00
Fe fraction
Figure 3.3. a) Variation of the Mn-N (black) and Fe-C (grey) distances with Fe fraction as
deduced by refinement of the X-ray powder diffraction profiles. Dashed lines represent the
maximum and minimum values found in literature for Mn-N and Fe-C distances.7 b)
Variation of the cell axis a of the F-43m space group with Fe fraction as determined by
refinement of the X-ray powder diffraction profiles. Squares are samples 3.1-6, stars
samples 3.7-11 and circles samples 3.12-13
56
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
The Fe-C and Mn-N distances in samples 3.3, 3.10, 3.12 and 3.13 are all identical
within the standard deviations, which is expected for samples that were prepared
under the same conditions. The lattice parameters of samples 3.3 and 3.10 are
identical within the standard deviations, as are the lattice parameters of samples
3.12 and 3.13. The two sets of parameters are not identical within the standard
deviations, but both sets are close to one another. Also the X-ray powder
diffraction data thus indicate that it is possible to reproduce the physical properties
of Prussian Blue analogues.
3.3.3 FTIR Spectroscopy
In contrast to X-ray powder diffraction measurements, which measure extended
information for the crystalline fraction only, FTIR spectroscopy measures local
information for both the crystalline and non crystalline fractions. For each sample
an FTIR spectrum was taken at room temperature. The CN stretching region (1900
– 2300 cm-1) is shown in figure 3.4 for all samples. Specific details of these spectra
are given in table 3.4.
Three different lineshapes have been found in the CN stretching region which are
centred around 2070 cm-1, 2110 cm-1 and 2150 cm-1. Possible overlap with
vibrations of water molecules can obscure the assignment. In agreement with the
previous chapter and Raman data from the literature,1 the peak centred around
2150 cm-1 is assigned as the CN stretching between FeIII and MnII (FeIII-CN-MnII)
and the peaks centred around 2070 cm-1 and 2110 cm-1 as the two CN stretching
frequencies of different symmetry in FeII-CN-MnIII.
Table 3.4. FTIR spectroscopic data (CN stretching bands) for RbxMn[Fe(CN)6]y·zH2O
Sample
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
xc1
(FWHM)
2070 (32)
2071 (36)
2070 (50)
2072 (39)
2070 (54)
2071 (30)
2073 (29)
-
xc2
(FWHM)
2104 (32)
2104 (36)
2104 (46)
2104 (37)
2104
2103
2103 (131)
2103 (99)
2103 (78)
2102 (108)
2103 (108)
2102 (114)
xca = position of peak a in cm-1, FWHM = Full Width at Half Maximum in cm-1
xc3
(FWHM)
2150 (21)
2150
2151
2153
2152 (26)
2152 (23)
2152 (25)
2154 (7)
2154 (13)
2154 (12)
2155
2154 (7)
2153
57
Intensity (a.u.)
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
Sample 3.1
Sample 3.2
Sample 3.3
Sample 3.4
Sample 3.5
Sample 3.6
Sample 3.7
Sample 3.8
Sample 3.9
Sample 3.10
Sample 3.11
Sample 3.12
Sample 3.13
2300
2200
2100
on
2000
the
Physical
Properties
of
1900
-1
ν (cm )
2300
2200
2100
2000
1900
-1
ν (cm )
Figure 3.4. Room temperature FTIR spectra for sample 3.1
(Rb0.86Mn[Fe(CN)6]0.95·1.94H2O), sample 3.2 (Rb0.89Mn[Fe(CN)6]0.96·1.73H2O), sample 3.3
(Rb0.91Mn[Fe(CN)6]0.97·1.53H2O), sample 3.4 (Rb0.92Mn[Fe(CN)6]0.97·2.25H2O), sample 3.5
(Rb0.91Mn[Fe(CN)6]0.96·1.48H2O), sample 3.6 (Rb0.95Mn[Fe(CN)6]0.99·2.16H2O), sample 3.7
(Rb0.61Mn[Fe(CN)6]0.86·2.71H2O), sample 3.8 (Rb0.82Mn[Fe(CN)6]0.94·1.39H2O), sample 3.9
(Rb0.85Mn[Fe(CN)6]0.94·1.27H2O), sample 3.10 (Rb0.91Mn[Fe(CN)6]0.97·0.90H2O), sample
3.11 (Rb1.01Mn[Fe(CN)6]1.00·1.00H2O), sample 3.12 (Rb0.93Mn[Fe(CN)6]0.99·0.78H2O) and
sample 3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O).
58
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
When comparing the present IR spectra with those for the samples described in
chapter 2 two different patterns of IR absorptions can be distinguished. Sample 2.1
of chapter 2 and sample 3.7 of this chapter both show the pronounced presence of
the band around 2150 cm-1 and the band around 2070 cm-1, whereas the band
around 2110 cm-1 is almost not distinguishable. These samples have in common
that both do not show any electron transfer (see also section 3.3.4). Although the
presence of the bands around 2110 cm-1 and 2070 cm-1 indicate the presence of the
LT phase, the X-ray powder diffraction data for both these samples did not show a
pronounced presence of LT phase. This is probably due to the fact that the intensity
with which the FeII-CN-MnIII absorption is present in the IR spectra is higher than
the intensity with which the FeIII-CN-MnII vibrates. In other words, a (very) small
fraction of LT phase is strongly visible in the IR spectra of both these materials.
(b)
(a)
140
2140
120
2130
100
FWHM (cm )
2120
-1
-1
Position (cm )
2150
2110
2100
2090
2080
80
60
40
20
2070
0.85
0.90
0.95
Fe fraction
1.00
0
0.85
0.90
0.95
1.00
Fe fraction
Figure 3.5. Variation of position (a) and width (b) of the 3 absorptions present in the IR
spectra with Fe fraction. ■ = Peak 1, ○ = peak 2, ∆ = peak 3
All samples that do show electron transfer have both the absorption around
2110 cm-1 and around 2070 cm-1, which are sometimes merged into one broad
band. The absorption around 2150 cm-1 is much less present in these electron
transfer active materials and sometimes even hardly visible (as is the case for for
instance sample 3.11 of this chapter).
The IR spectroscopic data are reproduced upon using fairly identical synthetic
conditions (sample 3.10, 3.12 and 3.13). On the other hand, the spectrum of sample
3.3 (which was prepared in the exact same manner), looks somewhat different from
these spectra in that the bands around 2110 cm-1 and 2070 cm-1 are resolved rather
than being merged into the broad band that is present in the other 3 samples.
One spectrum is not in line with either of the described spectra and that is the
spectrum of sample 3.6. It mostly shows the presence of the absorption around
2070 cm-1 and the other two absorptions are hardly visible. This means that it does
59
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
on
the
Physical
Properties
of
not have cyano ligands which are stretching between FeIII and MnII (indicative for
the HT phase). In the previous section sample 3.6 showed the presence of a
considerable amount of material in the tetragonal I-4m2 space group, which is also
believed to be indicative for the LT phase. However, care must be taken when
comparing the X-ray powder diffraction data with the IR spectra. The powder
diffraction data only indicates that a fraction of the structure is present, due to its
relatively low contribution no data could be obtained about the distances in this
structure. Furthermore, in X-ray powder diffraction data only the crystalline part is
visible, whereas in FTIR spectra both crystalline and amorphous fractions are
present. It is of course still possible that FeIII and MnII (the charges of the ions in
the HT phase) are present in this structure rather than FeII and MnIII (the charges of
the ions in the LT phase). Magnetic susceptibility measurements (section 3.3.4)
indicate that sample 3.6 is mostly in the HT phase.
3.3.4 Magnetic Measurements
Figure 3.6 shows the variation of χMT with temperature of sample 3.6, 3.7 and 3.13,
where χM is the molar magnetic susceptibility and T the temperature. In general the
temperature variation of χMT of the other samples is similar as that of sample 3.13.
Specific details of the magnetic data of all samples are shown in table 3.5.
In agreement with the χMT values samples 2.2 and 2.3 the χMT value for sample
3.13 slowly decreases from 3.25 cm3 K mol-1 to 3.00 cm3 K mol-1 as the
temperature is raised from 125 K to 280 K. It then abruptly increases to 4.76 cm3 K
mol-1 at 320 K. When the temperature is decreased χMT slowly decreases to
4.60 cm3 K mol-1 at 250 K. When lowering the temperature even further it
decreases abruptly to reach the value of 3.20 cm3 K mol-1 at 205 K. The
temperatures encompassing this hysteresis loop are T1/2↑ = 303 K and T1/2↓ =
242 K.
5.50
Figure 3.6. Temperature
dependence of χMT for sample 3.6
(Rb0.95Mn[Fe(CN)6]0.99·2.16H2O,
■), 3.7
(Rb0.61Mn[Fe(CN)6]0.86·2.71H2O,
○) and 3.13
(Rb0.96Mn[Fe(CN)6]0.98·0.75H2O,
∆).
5.25
3
-1
χMT (cm K mol )
5.00
4.75
4.50
4.25
4.00
3.75
3.50
3.25
3.00
2.75
125 150 175 200 225 250 275 300 325 350
T (K)
60
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Table 3.5. Selected details of the magnetic measurements.
Sample
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
χMT at
330 K
(cm3 K
mol-1)
5.14(7)
4.86(6)
5.05(4)
5.15(4)
5.44(7)
4.91(9)
4.77(1)
4.87(1)
4.90(3)
4.94(4)
5.11(3)
4.68(9)
4.64(3)
χMT at
250 K
(cm3 K
mol-1)
3.66(7)
3.35(7)
3.49(3)
3.42(8)
3.74(7)
4.1(2)
3.27(1)
3.54(2)
3.35(5)
3.40(2)
3.2(1)
3.16(5)
CHT
(cm3 K
mol-1)
CLT
(cm3 K
mol-1)
θHT
(K)
θLT
(K)
M at 5 K
(cm3
mol-1)
T1/2↑
(K) a
T1/2↓
(K) b
∆T
(K)
5.47(3)
5.19(4)
5.38(2)
5.44(3)
5.46(4)
4.04(1)
4.88(1)
5.18(1)
5.19(2)
5.23(2)
5.35(2)
4.84(3)
4.76(1)
3.69(1)
3.45(1)
3.67(1)
3.33(1)
3.48(2)
4.12(1)
3.343(4)
3.57(1)
3.29(1)
3.27(1)
3.06(1)
3.00(1)
-20(8)
-16(10)
-22(5)
-20(8)
-11(12)
59(3)
-13.3(1)
-21(3)
-20(4)
-20(7)
-17(7)
-9(8)
-9.7(2)
1(2)
-0(2)
-9(1)
7(1)
11(3)
4.462(6)
-3.5(8)
0(1)
4(1)
9(1)
9.52(5)
12.82(1)
4095(41)
4080(35)
4088(24)
4695(30)
4860(61)
3541(16)
65.34(5)
4272(14)
3938(21)
4138(30)
4915(27)
4255(31)
4017.3(5)
295
297
297
301
288
282
296
294
297
304
298
303
226
236
231
243
236
227
237
222
232
242
244
242
69
61
66
58
52
55
59
72
65
62
54
61
Standard deviations are shown in parentheses.
a
The standard deviation in T1/2↑ is ~1K
b
The standard deviation in T1/2↓ is ~5K
For a paramagnetic {S1;S2} = {5/2;1/2} system (the HT phase), a χMT value of
4.75 cm3 K mol-1 is expected. This value is not found in all samples and often the
experimental value is somewhat higher. For a paramagnetic {S1;S2} = {2;0} system
(the LT phase) a χMT value of 3.00 cm3 K mol-1 is expected. Again, the measured
χMT values are, in general, higher than this value. Since both the χMT values for the
HT configuration and the LT configuration are higher than expected it is possible
that it is a general deviation consistent throughout the entire measurement, possibly
coming from a small impurity or problems we experienced with the calibration of
the instrument (see footnote 9). It might also be possible that the orbital momentum
was not quenched, although it is unclear how it is possible that these samples
would have this feature whereas none of the samples reported in literature have it.
At the temperatures where the LT phase is expected, the higher value of χMT is also
partially due to the deviation from the 1:1:1 stoichiometry, since it is assumed that
the excess Mn ions do not show an electron transfer and thus remain in the HS MnII
state, thus giving a relatively higher χMT value. However, this deviation from the
perfect stoichiometry is not enough to explain the higher χMT values on its own.
Furthermore, part of the sample does not show any electron transfer and remains in
the HT phase, thus giving even higher values than expected for a pure LT phase.
The temperature variation of χMT of sample 3.7 shows a slow increase from
4.44 cm3 K mol-1 at 135 K to 4.76 cm3 K mol-1 at 330 K. In cooling, it shows the
same decrease, indicating that no charge transfer takes place in this temperature
region for this sample and the sample is completely in the HT configuration. The
shape of the magnetic susceptibility measurements for this sample is the same as
that of sample 2.1, which also did not show electron transfer.
61
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
on
the
Physical
Properties
of
Sample 3.6 shows a completely different shape of the hysteresis loop than the other
samples. In the previous sections it was already noted that this sample was always
different from the rest of the samples. From 125 K to 255 K the χMT value
decreases from 4.33 cm3 K mol-1 to 4.09 cm3 K mol-1. When heating to higher
temperatures it increases to 5.03 cm3 K mol-1 at 300 K before decreasing again to
4.90 cm3 K mol-1 at 330 K. In cooling it increases to 5.32 cm3 K mol-1 at 240 K and
then decreases to 4.41 cm3 K mol-1 at 205 K. Cooling the sample even further gives
a magnetic susceptibility identical to that of sample 3.7, as if the sample again has
the same magnetic properties as at room temperature. The magnetic behaviour of
sample 3.6 is not similar to the behaviour of, for instance, sample 2.4 which shows
a partial charge transfer. This different behaviour of the magnetic susceptibility in
sample 3.6 from the rest of the present samples (especially the non-horizontal
shape during the charge transfer, but horizontal lines outside the charge transfer)
has been reproduced for a sample which has been made at a temperature of 96oC. It
is therefore believed that its behaviour is due to the relatively high synthetic
temperature, which leads to an unknown occurrence that manifests itself in the
magnetic susceptibility and other measurements.
The inverse magnetic susceptibility was fitted with a straight line and from this the
Curie constants (C) and the Curie-Weiss constants (θ) were calculated. In figure
3.7 the variation of these constants with the Fe content is given. In general, the C of
the HT configuration is higher than the C of the LT configuration, except for
sample 3.6, where these values are almost similar. Since the value of C is always
close to the value of χMT at room temperature the reasoning behind this was
b)
5.6
5.4
5.2
5.0
4.8
4.6
4.4
4.2
4.0
3.8
3.6
3.4
3.2
3.0
60
40
θ (K)
3
-1
C (cm K mol )
a)
20
0
-20
0.85
0.90
0.95
Fe fraction
1.00
0.85
0.90
0.95
1.00
Fe fraction
Figure 3.7. Variation of the Curie constant (C, a) and the Curie-Weiss constant (θ, b) with
Fe fraction as found by fitting the inverse magnetic susceptibility. Grey data points are of
the LT configuration and black data points of the HT configuration. Squares are samples
3.1-6, stars are samples 3.7-11 and circles are samples 3.12-13.
62
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
explained earlier. The θ’s of the HT configuration are lower than those of the LT
configuration, indicating that the HT configuration exhibits a stronger
antiferromagnetic coupling between nearest neighbour Mn and Fe ions than the LT
configuration. The only sample that is completely off in this respect is sample 3.6,
which shows a very high positive θ for the HT configuration and a much smaller
positive value for the LT configuration. The deviating behaviour of this sample
from the other samples has already been found in the powder diffraction refinement
and in the IR spectrum (see previous sections), but it is also shown in the shape of
the magnetic susceptibility measurements.
The variation of the T1/2↑, T1/2↓ and ∆T with Fe fraction is shown in figure 3.8. In
analogy with the literature,1 the temperature at which the HT to LT phase transition
takes place is less sensitive to the Fe fraction than the transition temperature for the
reverse process. According to the literature,1 the more Fe defects are present, the
broader the hysteresis width. This trend is not directly reflected in figure 3.8b. It is
unclear why this is the case, but possibly the use of different temperatures during
the synthesis or different precipitation times leads to other defects in the system
that are not reflected in the stoichiometry but do have an effect on the hysteresis
width.
a)
b)
300
290
270
∆T (K)
T1/2 (K)
280
260
250
240
230
220
0.94 0.95 0.96 0.97 0.98 0.99 1.00 1.01
Fe fraction
74
72
70
68
66
64
62
60
58
56
54
52
50
0.94 0.95 0.96 0.97 0.98 0.99 1.00 1.01
Fe fraction
Figure 3.8. a) Variation of T1/2↑ (black) and T1/2↓ (grey) with Fe content, b) Variation of the
hysteresis widths with Fe content. Squares are samples 3.1-6, stars are samples 3.8-11 and
circles are samples 3.12-13.
Most samples (except sample 3.7) show spontaneous magnetisation below 12 K
under a field of 0.01 T.10 The magnetisation data of a representative sample is
shown in figure 3.9a. The magnitude of the magnetisation at 5 K is a direct
measure of the amount of LT phase that is present, because only the LT phase
shows ferromagnetic ordering below 12 K, and hence a contribution to the
magnetisation (see chapter 2). The variation of the magnetisation at 5 K under a
63
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
on
the
Physical
Properties
of
field of 0.01 T with Fe content is shown in figure 3.9b. According to the findings
reported in chapter 2 more Fe-defects mean a less complete transition and thus less
LT phase is present at 5 K which is responsible for the ferromagnetic ordering.
Again, the present data do not directly confirm this finding, and the reasoning
behind this is unclear. Again, it might be the case that different temperatures during
synthesis or different precipitation times leads to defects in the system that are not
reflected in the stoichiometry but do have an effect on the magnetic properties.
a)
b)
5000
1000
0
4
6
8
10
12
T (K)
14
16
18
20
-1
2000
4000
3
-1
3000
3
M (cm mol )
4000
M at 5K (cm mol )
5000
3000
2000
1000
0
0.85
0.90
0.95
1.00
Fe fraction
Figure 3.9. a) Magnetisation versus temperature for sample 3.12
(Rb0.93Mn[Fe(CN)6]0.99·0.78H2O b) Variation of the magnetisation at 5 K with Fe fraction.
Squares are samples 3.1-6, stars are samples 3.7-11 and circles are samples 3.12-13.
The samples that were prepared under fairly identical conditions are samples 3.3,
3.10, 3.12 and 3.13. The magnetic measurements of samples 3.3 and 3.10 are quite
similar to each other and also those of samples 3.12 and 3.13 are identical within
the standard deviations. The magnetic susceptibility of the two sets of samples are
not directly similar. However, it seems as if both samples 3.3 and 3.10 have a
general deviation of the χMT values throughout the entire measurement of
~0.25 cm3 K mol-1. This deviation might come from a calibration error of the
machine.9 If this general deviation is subtracted for both entire measurements, these
samples have much more consistent values with samples 3.12 and 3.13 and thus the
physical properties of Prussian Blue analogues can be reproduced with fairly
identical synthetic conditions.
3.4 Conclusions
In this chapter the influence of temperature and addition speed during the synthesis
on the composition of RbxMn[Fe(CN)6]y·zH2O samples was investigated and its
consequential influence on their physical properties. In general, the temperature
during the synthesis has little influence on the composition of the samples.
64
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Although a too high temperature during the synthesis (at least higher than 74ºC, but
lower than 91ºC) leads to strange properties of the compound. In contrast to the
other samples the sample prepared at 91oC has a relatively high contribution of the
tetragonal structure I-4m2, only a vibration around 2070 cm-1 is visible in the CN
stretching region and the magnetic susceptibility measurements have a completely
different shape. It is unclear at this stage where these properties originate from,
although it might be due to a difference in the redox potentials or the Gibbs free
energy, which are both dependent on temperature.
On the other hand, the addition speed has a major influence on the composition: the
slower the addition speed, the more the metal stoichiometry becomes perfect (i.e.
Rb:Mn:Fe = 1:1:1). It was possible to reproduce the Rb:Mn:Fe ratio within a
standard deviation of 0.05. The water content per sample is less reproducible. This
might be due to the occurrence that water will only be incorporated in the
interstitial sites if it is present at the moment that the CN ligand bridges Mn and Fe.
It might also be conceivable that water is incorporated when the sample is exposed
to a moist environment, although IR spectra did not change significantly in time.
The water attached to Mn next to an [Fe(CN)6] defect will always be incorporated
in order to ensure the anticipated six coordinated Mn ion.
The measurements performed on the 4 samples which were prepared under similar
conditions show that the physical properties of these samples are comparable. For
most techniques one of the samples has a deviation in the results from the other 3
samples. However, the deviating sample is a different sample in almost all
techniques.
The trends in the physical properties of the series of compounds do not always
conform to the trends reported in the literature.1,11 This might be due to the
different temperatures during the synthesis or to the differences in precipitation
time which can lead to defects in the system that are not reflected in the
stoichiometry but do have an effect on the magnetic properties of the compounds.
3.5 References
1
2
3
4
5
Cobo, S.; Fernandez, R.; Salmon, L.; Molnar, G. ; Bousseksou, A. ; Eur. J.Inorg. Chem. 2007,
1549.
Ohkoshi, S.-I.; Matsuda, T.; Tokoro, H.; Hashimoto, K. ; Chem. Mater. 2005, 17, 81.
H-bonded water molecules give a broad band in vibrational spectra in the OH-stretching region
from 3200 – 3600 cm-1. Raman spectroscopy in the OH strectching region on mixtures of H2O
and HDO/D2O in acetonitrile, E.J.M. Vertelman, Master thesis, University of Groningen, 2002
FTIR spectra always show a significant amount of water in each sample of
RbxMn[Fe(CN)6]y·zH2O samples, coming from both the sample itself and water molecules
present in air. It is therefore difficult to estimate a change in water content on IR spectra alone.
Reversible water intercalation has been studied for Ni[Fe(CN)6]y·zH2O and M3[Co(CN)6]y·zH2O
(M = Mn, Co, Ni, Cu, Zn, Cd) compounds. It was shown that the intercalation of water vapor is
65
Chapter 3. The Influence of Synthetic Conditions
RbxMn[Fe(CN)6]y·zH2O Prussian Blue Analogues
6
7
8
9
10
11
66
on
the
Physical
Properties
of
possible but below a temperature of 30oC was very low. a) Yu, Q.; Steen, W.A.; Jeerage, K.M.;
Jiang, S.; Schwartz, D.T., J. Electrochem. Soc. 2002, 149(6), E195; b) Roque, J.; Reguera, E.;
Balmaseda, J.; Rodriguez-Hernandes, J.; Reguera, L.; Castillo, L.F.del, Microporous
Mesoporous Mater. 2007, 57.
a) Kato, K.; Morimoto, Y.; Takata, M.; Sakata, M.; Umekawa, M.; Hamada, N.; Ohkoshi, S.-I.;
Tokoro, H.; Hashimoto, K., Phys.Rev.Lett., 2003, 91, 25502; b) Moritomo, Y.; Kato, K.; Kuriki,
A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K., J.Phys.Soc.Jpn., 2002,
71, 2078
a) Chapter 5; b) Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I.; Z. Anorg. Allg. Chem.
2007, 633, 1134; c) Matsuda, T.; Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I.; Acta
Cryst. E 2008, i11
a) Dong, W.; Zhu, L.-N.; Song, H.-B.; Liao, D.-Z.; Jiang, Z.-H.; Yan, S.-P.; Cheng, P.; Gao, S..
Inorg. Chem. 2004, 43, 2465; b) Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.;
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.. J. Phys. Soc. Jpn. 2003, 72, 2698
None of the RbxMn[Fe(CN)6]y·zH2O compounds reported in literature show a χMT value that is
higher than 4.75 cm3 K mol-1 (the maximum value expected for a paramagnetic {S1:S2} =
{5/2;1/2} system). Most measurements shown in this thesis are also not above this value (taking
into account the error values). However, some samples (especially 3.1, 3.3-3.5 and 3.11) show a
χMT value that is noticeably higher. It is unclear how this is possible and several measurements
have been taken to confirm the values. It is to be noted, though, that all the measurements with
strange values have taken place after the MPMS machine was broken down and repaired. The
other users did not have this problem.
This field is smaller than the field at which saturation magnetisation takes place, which has an
order of magnitude of ~1 T.
Chapter 2
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 4. Identifying the Fraction of Charge Transfer
Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O1
4.1 Introduction
In chapter 2 and 3 it was shown that the physical properties of
RbxMn[Fe(CN)6]y·zH2O depend strongly on the actual sample stoichiometry. In
compounds with Fe(CN)6 defects, oxygen atoms of water molecules coordinate to
the Mn ions instead of the nitrogen atoms of the cyanide ions and the average
environment of the manganese ions changes to Mn(OH2)x(NC)6-x. As a result, the
ligand field strength of the Mn ions becomes weaker and consequently the redox
potential of the MnIII/MnII couple changes leading finally to its inability to reduce
[FeIII(CN)6] to [FeII(CN)6]. If the water content is sufficiently high, the redox
potential changes to such an extent that the HT phase (LS FeIII and HS MnII) form
becomes stable at very low temperatures. This important role of the incorporated
water molecules, which may be coordinated as well as non-coordinated, suggests
that the physical properties of these compounds may change with time due to
possible incorporation of water molecules from the environment, but this
occurrence has not been discussed in literature.
In the present chapter, three derivatives of RbxMn[Fe(CN)6]y·zH2O with different
stoichiometries are characterised in detail: Rb0.96Mn[Fe(CN)6]0.98·0.75H2O,
Rb0.94Mn[Fe(CN)6]0.88·2.17H2O and Rb0.61Mn[Fe(CN)6]0.86·2.71H2O.
Also the effect of the stoichiometry change on their crystallographic and electronic
structure as well as on their spectroscopic and magnetic properties is discussed.
4.2 Experimental Section
4.2.1 Synthesis
All chemicals of analytical grade were purchased from Sigma-Aldrich and used
without further purification.
The first sample is the same as sample 3.13 in chapter 3. Its synthesis and
characterisation can be found there.
In contrast with the “usual” synthesis route of the other two samples, which have
been obtained from a pure aqueous solution, sample 4.1 has been obtained from a
water/ethanol mixture by instantaneous mixing of a 50 mL ethanolic solution of
MnCl2·4H2O (0.1 M) and a 50 mL aqueous solution containing both K3Fe(CN)6
(0.1 M) and RbCl (0.8 M). All solutions were heated at a temperature of 50°C
before the addition procedure. The brown precipitate was filtered, washed twice
with water and dried in air at room temperature. For sample 4.1, it was not possible
67
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
to determine an accurate chemical formula because of its instability in time and
also because of the presence of different valencies of iron in this compound (see
section 4.3.2). From elemental analysis, for which the acquiring of the results took
more than 1 month (Measured: Rb = 22.42%, Mn = 15.23% and Fe = 13.69%;
calculated: Rb = 22.26%, Mn = 15.22%, Fe = 13.62%) the relative proportions of
the metal ions were derived as Rb/Mn/Fe = 0.94/1/0.88 and an approximate
formula as Rb0.94Mn[Fe(CN)6]0.88·2.17H2O.2 The magnetic, Raman spectroscopic,
Mössbauer spectroscopic and DSC measurements, reported below, were performed
on freshly prepared samples of sample 4.1, but the acquisition of Mössbauer data
took longer than one week. It should be noted that the thermodynamical parameters
(∆H, ∆S) as well as the molar magnetic susceptibility could not be accurately
evaluated for this sample due to the approximated molar mass and the relatively
high amount of residual fractions remaining unaltered during the phase transition.
The third sample is sample 3.7 in chapter 3. Its synthesis and characterisation can
be found there.
4.2.2 Elemental Analysis
Analysis of C, H and N were performed after combustion at 850°C, using IR
detection and gravimetry by means of a Perkin Elmer 2400 series II device. Fe,
Mn, Rb and K content were determined by the Service Central d’Analyse du CNRS
(Vernaison) using ICP–AES (Inductively Coupled Plasma Atomic Emission
Spectroscopy) after acid digestion in H2SO4/HNO3. Standard deviations in the
measured metal contents were typically around 2 rel.%. The O atoms were
assumed to be the only other elements present in the samples and the H2O content
was obtained by difference to 100%. (The K fraction was found to be negligible.)
4.2.3 Magnetic Measurements
The magnetic properties were measured at various cooling and heating rates in a
magnetic field of 0.1 T using a Quantum Design MPMS superconducting quantum
interference device magnetometer. The experimental data were corrected for the
diamagnetic contribution using Pascal’s constants.
4.2.4 Differential Scanning Calorimetry (DSC)
DSC analysis was carried out on a Netsch DSC 204 instrument under helium
purging gas (20 cm3 min-1) at a heating/cooling rate of 10 K min-1. Temperature
and enthalpy were calibrated using the melting transition of standard materials (Hg,
In, Sn). The uncertainty in the transition enthalpy (∆HHL) and entropy (∆SHL) is
(over)estimated to be ~10% because the determination of the accuracy of the
measurements involves a rather complicated modelling of the baseline, which then
needs to be subtracted. The repeatability of the measurements is within 2-3%.
68
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
4.2.5 Raman Spectroscopy
Raman spectra were collected between 300 K and 80 K using a LabRAM-HR
(Jobin-Yvon) Raman micro-spectrometer and a Linkam THMS-600 cryostage. The
632.8 nm line of a HeNe laser was used as the excitation source, the laser power on
the sample was typically 0.1 mW. A spectral resolution of ~1 cm-1 was obtained.
Least square fittings of the Raman peaks have been carried out with the assumption
of Lorentzian line shapes.
4.2.6 X-Ray Powder Diffraction
For the detailed refinement of the powder diffraction profiles of samples 3.13 and
3.7, see chapter 3.
4.2.7 57Fe Mössbauer Spectroscopy
Fe Mössbauer spectra have been recorded using a conventional constantacceleration type spectrometer equipped with a 50 mCi 57Co source and a flowtype, liquid helium cryostat. Spectra of the powder samples (~30 mg) were
recorded between 5 and 300 K. Least square fittings of the Mössbauer spectra have
been carried out with the assumption of Lorentzian line shape using the Recoil
software package.3
57
4.3 Results and Discussion
4.3.1 Sample 3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O)
The χMT value, where χM stands for the molar magnetic susceptibility and T for
temperature, of the freshly prepared4 sample 3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O)
decreases abruptly from 4.6 cm3 K mol-1 at 250 K to 3.2 cm3 K mol-1 at 200 K upon
cooling and, conversely, as the sample is warmed up from 130 K, the χMT value
increases around 285 K and reaches the initial χMT value at 325 K (Figure 4.1). The
values of χMT in the high and low temperature regions are in agreement with the
theoretical values expected for the sum of the spin only values of {LS FeIII; HS
Figure 4.1. Temperature
dependence of χMT of sample 3.13
(Rb0.96Mn[Fe(CN)6]0.98·0.75H2O)
in the cooling and heating modes:
(○) freshly prepared sample (less
than one week between measuring
and preparation), („) after 6
months storage in ambient
conditions.
4.5
4.0
3
-1
χMT (cm K mol )
5.0
3.5
3.0
100
150
200
250
300
350
T (K)
69
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
(b)
310
240
300
220
T1/2 (K)
T1/2 (K)
(a)
290
280
270
200
180
160
260
140
0.7
0.8
0.9
1.0
0.7
0.8
0.9
1.0
Rb fraction
Rb fraction
Figure 4.2 Phase transition temperatures in the heating (a) and cooling modes (b) as a
function of the Rb fraction in samples 3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O) (large grey
squares) and 3.7 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O) (large grey circles). Data in black have
been reported in reference 5 for a series of different RbxMn[Fe(CN)6]y·zH2O
stoichiometries. (Lines are inserted to guide the eye.) The half filled grey circles correspond
to the stoichiometry of the HT configuration in sample 4.1 estimated from the Raman line
widths.
-0.20
-0.21
-0.1
-0.22
-0.23
-0.24
265
270
275
280
285
-1
Heat flow (mW mg )
T (K)
sample 4.1
-0.2
150
200
250
300
350
250
300
350
0.0
-0.5
sample 3.13
150
200
T (K)
70
Figure 4.3. Heat capacity anomalies
recorded in the heating mode for samples
3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O)
and 4.1 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O).
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
MnII} and {LS FeII; HS MnIII}, respectively, revealing thus a complete transition,
which would indeed be in agreement with the approximately 1:1 Mn:Fe ratio. The
width of the thermal hysteresis loop (61 K) is defined by T1/2↓ = 242 K and T1/2↑ =
303 K. These temperatures correlate with the general trends reported by Cobo et
al.,5 i.e. when the Rb content is close to 1 and the water content tends to go towards
0, the phase transition shifts to higher temperatures and the hysteresis width
decreases (see Figure 4.2). The magnetic measurements on this sample were
reproduced (on a different machine with the use of the same molecular weight and
Pascal’s constants) after the compound had aged for six months in ambient storage
conditions. These reveal a smaller hysteresis (T1/2↓ = 250 K and T1/2↑ = 303 K) and
an upward displacement of the absolute values of χMT (Figure 4.1). On the whole,
however, if one takes into account the measurement uncertainties the overall
difference appears to be relatively small.
The first-order phase transition in an aged sample 3.13 was also evidenced by DSC
measurements (Figure 4.3) and the associated enthalpy and entropy changes were
determined as ∆H = 15 kJ mol-1 and ∆S = 53 J K-1 mol-1.6 These values are in
agreement with literature data.5,7 As discussed in reference 5 this large entropy
change has, for the most part, a vibrational origin. The completeness of the
293 K
Intensity (a.u.)
180 K
sample 3.13
80 K
293 K
Figure 4.4. Raman spectra of samples
3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O),
4.1 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O) and
3.7 (Rb0.61Mn[Fe(CN)6]0.86·2.71H2O)
acquired at 293 K (cooling mode), 180 K
(heating mode) and 80 K (cooling mode).
180 K
sample 4.1
80 K
sample 3.7
2050
293 K
2100
2150
2200
-1
ν (cm )
Table 4.1 Raman spectroscopic data (CN stretching bands) for RbxMn[Fe(CN)6]y·zH2O
Sample
3.13
4.1
3.7
Temperature
(K)
293
180
80
293
180
80
293
Mode 1
(FWHM)
(cm-1)
2168 (3.5)
2167
2170
2170 (4.9)
2169
2170 (4)
2168 (6.0)
Mode 2
(FWHM)
(cm-1)
2159 (3.5)
2159 (6.0)
2161
2160 (3)
2159 (9.2)
Mode 3
(FWHM)
(cm-1)
2113 (3)
2115 (2)
2120
2111 (6)
2122
Mode 4
(FWHM)
(cm-1)
2094 (5)
2096 (3)
2082
2091 (10)
2082
71
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
transition was confirmed by Raman spectroscopy (Figure 4.4), which revealed two
CN stretching modes around 2159 and 2168 cm-1 at 293 K (indicative for CN
stretching in FeIII-CN-MnII, the HT phase), while the LT configuration is
characterized by two modes around 2094 and 2113 cm-1 (CN stretching in FeII-CNMnIII). The modes indicative for the LT phase are not present in the HT phase,
while the opposite is only present to a very small extent. The narrow peaks
observed in the Raman spectra (FWHM = ~3.5 cm-1) confirm nicely the proposed
correlation5 between the stoichiometry of the samples and the Raman line widths,
i.e. the more deviation from a ‘perfect’ stoichiometry, the broader the modes in the
HT configuration (see Figure 4.5).
12
2159 cm-1
2170 cm-1
-1
fwhm (cm )
10
8
6
4
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Rb fraction
Figure 4.5. Evolution of the Raman line
widths of the CN stretching modes
around 2159 and 2170 cm-1 as a function
of the Rb fraction for samples 3.13
(Rb0.96Mn[Fe(CN)6]0.98·0.75H2O) (▼),
4.1 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O) (●)
and 3.7 (Rb0.61Mn[Fe(CN)6]0.86·2.71H2O)
(▲). Data in black have been reported
previously in reference 5 for a series of
different RbxMn[Fe(CN)6]y·zH2O
stoichiometries. (Straight lines represent
linear fits on the data.) The grey arrows
indicate the stoichiometry (x) of the HT
configuration in sample 4.1 as estimated
from the Raman line widths.
Selected 57Fe Mössbauer spectra of the aged sample 3.13 acquired in the cooling
and heating modes (HT and LT configurations, respectively) within the hysteresis
region at 265 K are shown in Figure 4.6. Values of the hyperfine parameters
obtained from the least squares fitting procedure of the spectra recorded in the 29380 K region are listed in Table 4.2. In the cooling mode at 265 K the spectrum
corresponding to LS FeIII (HT phase) can be fitted with a doublet with an isomer
Figure 4.6. 57Fe Mössbauer spectra of
sample 3.13
(Rb0.96Mn[Fe(CN)6]0.98·0.75H2O)
acquired at 265 K in the cooling (○) and
heating („) modes
Rel. transmission (%)
100
95
90
85
-4
-2
0
v (mm/s)
72
2
4
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
265 K
ln[A]
2.2
2.0
1.8
0
50
100
150
200
250
300
Figure 4.7. Temperature dependence of
the natural logarithm of the total area of
57
Fe Mössbauer absorptions of sample
3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O).
(The lines represent linear fits on data
obtained in the pure LT and HT
configurations.) Most points have been
acquired in the cooling mode, except for
the upper point at 265 K and the point at
295 K, which have been acquired in the
heating mode.
T (K)
shift δ = -0.133(2) mm s-1 and a quadrupole spitting QS = 0.097(2) mm s-1. In the
heating process, the spectrum corresponding to LS FeII (LT phase) can be fitted
with a doublet with an isomer shift δ = -0.119(1) mm s-1 and a quadrupole spitting
QS = 0.124(8) mm s-1. In literature8 the observed values for LS FeII are δ = -0.01 –
-0.085 mm s-1 and QS = 0 – 0.24 mm s-1, and for LS FeIII δ = -0.06 – -0.17 mm s -1
and QS = 0.285 – 1.02 mm s-1 in compounds of AxCo[Fe(CN)6]y·zH2O (A = Na+,
K+, Rb+). In other words, taking into account the Mössbauer hyperfine parameters,
distinguishing the two forms is not straightforward. However, the Mössbauer peak
areas appear significantly different (~20%) indicating that the recoilless fractions
differ in the two configurations. Indeed, the plot of the total area (A) as a function
of the temperature (Figure 4.7.) reveals clearly a discontinuity at the phase
transition. This sudden change of A reflects a marked variation in the LambMössbauer factor, i.e. in the lattice dynamics. In a more quantitative manner, using
the Debye model in the high temperature limit one can determine the so-called
Mössbauer Debye temperature, θM from the temperature dependence of A using the
relationship:9
d(ln[A])/dT = -3Eγ2/Mc2kBθM2
where Eγ is the Mössbauer transition energy, kB the Boltzmann constant, c the
velocity of light and M has been taken as the mass of the ‘bare’ Mössbauer atom
(57Fe in the present case). From the linear least squares fit of ln[A] versus T it was
possible to determine the values of θM as 265 (± 24) K and 333 (± 14) K in the HT
and LT configurations, respectively. The higher value of θM in the LT
configuration signifies that the crystal lattice is more rigid in this configuration.
This finding is in agreement with the X-ray diffraction data10 which reveals a
strong lattice contraction when going from the HT to the LT configuration and also
with the calorimetric results which reveal a significantly higher entropy in the HT
configuration.5,7
The X-ray powder diffraction profile of sample 3.13 obtained at room temperature
(see chapter 3) was satisfactorily fitted using the F-43m model reported for single
73
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
crystals.11 The atom fractions on each crystallographic site were fixed to those
deduced from the elemental analysis, with the exception of the fractions of the Rb
and O atoms. For Rb, there are two crystallographic distinct positions in the F-43m
space group: 4c and 4d (hereafter denoted as Rb1 and Rb2, respectively). It was
assumed that the octahedral environment of the Mn ion is always maintained by
coordination of water molecules, therefore the O3 fraction was kept constant on
0.02. During structural refinement, a series of constraints involving the Rb and O
fractions were used: [Rb1] + [Rb2] = 0.96, [O1] + [O2] = 0.63 (=0.75 – 6 x 0.02),
[Rb1] + [O1] = 0.96 and [Rb2] + [O2] = 0.63. This resulted in small (~-0.02))
negative occupation fractions for Rb2 and O1. Since this does not make sense
chemically, all Rb ions were placed on Rb1 and the remaining O atoms were
placed on O2. Constraints were also placed on the isotropic atomic displacement
parameters, Uiso: UisoMn = UisoFe, UisoRb1 = UisoRb2 = UisoO1 = UisoO2, UisoC =
UisoN = UisoO3. The determined Fe-C and Mn-N distances of 1.86(4) Å and
2.28(3) Å are indicative of LS FeIII and HS MnII, respectively, and match those in
the crystal structure of the HT configuration determined at 293 K (Mn-N =
2.205(5) Å, Fe-C = 1.929(4) Å), taking into account the uncertainties in the present
model.11 After refinement of the profile using this cubic space group, weak
Table 4.2. Least-squares fitted 57Fe Mössbauer data for sample 3.13
(Rb0.96Mn[Fe(CN)6]0.98·0.75H2O) obtained in the cooling (293 – 80 K) and heating (265 K,
293 K) modes. (Values in parentheses are standard deviations. Values without standard
deviations were fixed)
T
(K)
293
285
275
265
255
245
235
225
215
205
195
180
160
140
120
100
80
265*
293*
Doublet
IS
(mm s-1)
-0.151(2)
-0.138(3)
-0.137(2)
-0.133(2)
-0.131(1)
-0.126(2)
-0.123(3)
-0.121
-0.117
-0.114
QS
(mm s-1)
0.07(2)
0.177(9)
0.159(7)
0.097(2)
0.049(3)
0.026(8)
0
0.12
0.12
0.12
Γ/2
(mm s-1)
0.152(5)
0.162(7)
0.155(5)
0.162(7)
0.155(5)
0.157(8)
0.17(1)
0.16
0.16
0.16
Area
(%)
100
100
100
100
100
100
100
78(6)
53(9)
20(10)
0
0
0
0
0
0
0
0
0
Doublet
IS
(mm s-1)
QS
(mm s-1)
Γ/2
(mm s-1)
-0.094
-0.091
-0.087
-0.086(2)
-0.080(2)
-0.074(4)
-0.069(2)
-0.060(2)
-0.054(1)
-0.050(2)
-0.119(1)
-0.147(1)
0.12
0.12
0.12
0.117(9)
0.12(1)
0.145(2)
0.12(1)
0.13(1)
0.117(7)
0.12(1)
0.124(8)
0.11(1)
0.16
0.16
0.16
0.16(4)
0.168(6)
0.159(9)
0.164(7)
0.159(7)
0.156(3)
0.151(6)
0.165(4)
0.164(4)
Area
(%)
0
0
0
0
0
0
0
22(6)
47(1)
80(10)
100
100
100
100
100
100
100
100
100
IS: isomer shift (with reference to metallic iron at 293 K), QS: quadrupole splitting, Γ/2: half-height
width
* Obtained in heating mode, in contrast to the rest which was obtained in the cooling mode.
74
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
unindexed peaks were still present at 2θ values of 24.5o and 35o. These peaks are
indicative of the LT-phase (space group I-4m2) reported by Moritomo et al.12
4.3.2 Sample 4.1 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O
Figure 4.8a shows the product of the molar magnetic susceptibility and the
temperature for the freshly prepared sample 4.1 as a function of temperature
measured at a rate of 2 and 0.3 K min-1, respectively. At room temperature the χMT
value is ~4.6 cm3 K mol-1. The first cooling was carried out at a rate of 2 K min-1
and a partial transition was observed between 160 and 130 K leading to a partially
quenched HT configuration. At 80 K the χMT value reaches ~4.0 cm3 K mol-1,
which corresponds to a mixture of ~40% LT and ~60% HT phase. In the heating
mode, around 130 K, the quenched, metastable HT configuration relaxes and
consequently the χMT value decreases to reach ~3.5 cm3 K mol-1 at 175 K. This
value indicates the presence of a residual HT fraction even after the relaxation. As
shown in Figure 4.8a, if the sample is slowly cooled down at a rate of 0.3 K min-1
one can avoid quenching the system. The width of the thermal hysteresis loop
(135 K) is defined by T1/2↓ = 147 K and T1/2↑ = 282 K. Such a large hysteresis
width is an appealing property for the application of these materials in various
devices (e.g. memories). This hysteresis confirms the trend that
RbxMn[Fe(CN)6]y·zH2O materials with a high concentration of Fe(CN)6 defects
leads to a broad hysteresis. It is interesting to note that in the warming mode for
both experiments, the phase transition occurs in two steps with a very small plateau
around 280 K. This unprecedented result may have a crystallographic origin. The
two steps, although very minimal, are also reflected in the DSC curve of the sample
(Figure 4.3), which shows two peaks in the heating mode at 270 and 275 K.
The incomplete nature of the transition in sample 4.1 was confirmed by Raman
spectroscopy (Figure 4.4.) which revealed the presence of a small fraction of the
reduced LT phase (υCN = 2082 and 2120 cm-1) in the HT configuration and a
Figure 4.8. (a) Temperature dependence
of χMT of the freshly prepared sample 4.1
(Rb0.94Mn[Fe(CN)6]0.88·2.17H2O)
recorded at different cooling rates: 2 K
min-1 („) and 0.3 K min-1 ({). (b)
Temperature dependence of χMT
(recorded at a cooling rate of 0.3 K min-1)
of sample 4.1 after 1 month ageing (U),
followed by a heat treatment at 403 K
(▼).
a
4.0
3.5
3
-1
χMT (cm K mol )
4.5
4.5
b
4.0
3.5
100
150
200
250
300
350
T (K)
75
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
significant quantity of residual HT phase (2161 and 2169 cm-1) in the LT
configuration at 180 K (heating mode). It should be noted that at 80 K the Raman
spectrum of this compound corresponds to the HT configuration even if the sample
is cooled down at a very slow cooling rate. This observation reflects that the
exciting laser light instantaneously transforms the LT configuration into the HT
configuration, which then remains trapped due to the slow relaxation kinetics at
this temperature. The power of the laser has an influence in this, but this question
has not been studied in detail. Care was taken in minimising the laser power to
such an extent that no significant laser heating occurs in the sample.
The stoichiometry of sample 4.1 deviates significantly from the general trend
observed previously5 (Figure 4.9.), which means that the sample does not consist of
a homogeneous phase. This fact is also reflected in the Raman and Mössbauer
spectra. It is therefore difficult to discuss in this case the correlation between the
sample stoichiometry and the phase transition temperatures. On the other hand, it is
possible to measure the broadening of the Raman spectral lines, which correspond
to the HT configuration. It appears that the two CN stretching modes in the HT
configuration display a significant line broadening: the FWHM values measured at
room temperature are 4.9 and 6.0 cm-1 for the bands centred around 2170 and
2159 cm-1, respectively. This line broadening is in fairly good agreement with the
(a)
(b)
5
1.0
H2O fraction
4
Rb fraction
0.8
0.6
3
2
1
0
0.4
0.80
0.85
0.90
0.95
1.00
Fe fraction
0.80
0.85
0.90
0.95
1.00
Fe fraction
Figure 4.9. Rb fraction (a) and numbers of H2O molecules (b) as a function of Fe fraction in
samples 3.13 (Rb0.96Mn[Fe(CN)6]0.98·0.75H2O) (▼), 4.1 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O)
(●) and 3.7 (Rb0.61Mn[Fe(CN)6]0.86·2.71H2O) (▲). Data in black have been reported
previously in reference 5 for a series of different RbxMn[Fe(CN)6]y·zH2O stoichiometries.
The half-filled grey circle corresponds to the stoichiometry (x) of the HT configuration in
sample 4.1 estimated from the Raman line widths. (Straight lines represent linear fits on the
data.)
76
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
relatively large amount of Fe(CN)63- vacancies and also with the observed phase
transition temperatures (see Figure 4.5).5
Sample 4.1 appeared to be unstable in ambient storage conditions, in that its
physical properties change to the HT configuration with time most probably due to
hydration phenomena. As shown in Figure 4.8b, after one month at ambient
conditions, only a small fraction (~5%) of the compound still exhibits charge
transfer. This temporal evolution of sample 4.1 towards a stable HT configuration
suggests that an additional amount of water molecules has entered the crystal
lattice to such an extent that the {LS FeIII; HS MnII} form becomes stable
regardless of temperature. The most plausible explanation of this phenomenon is
water uptake from the air.13 Indeed, one should notice that this sample contains a
large amount of Fe(CN)6 vacancies and the presumably octahedral coordination
sphere of manganese ions must be completed by, for instance, water molecules. If
this hypothesis is correct a supplementary heat treatment may allow removing a
certain amount of water molecules14 and the charge transfer should thus be
restored. A thermal treatment of sample 4.1 (HT form) was carried out in an oven
at 130°C during 5 days in air. As shown by the magnetic measurements (Figure
4.8b), ~80% of the original amount of charge transfer is restored. The sample
exhibits a large hysteresis loop following this treatment even if the final phase
transition temperatures are somewhat different from those observed in the starting
material.
Transmission
Mössbauer spectra of compound 4.1 (freshly prepared) at selected temperatures are
shown in Figure 4.10 and values of the hyperfine parameters obtained from the
least squares fitting procedure are listed in Table 4.3. At room temperature the
Mössbauer spectrum can be satisfactorily fitted by using three components: two
quadrupolar doublets and one singlet with approximately 40/45/15% area ratios,
respectively. On the basis of the hyperfine parameters and the strong temperature
dependence of the quadrupole splitting the doublets can be clearly assigned to low
Figure 4.10. 57Fe Mössbauer spectra of
sample 4.1
(Rb0.94Mn[Fe(CN)6]0.88·2.17H2O)
acquired in the cooling and heating
modes at 210 K
cooling
-2
-1
0
1
2
heating
-2
-1
0
V (mm s -1)
1
2
77
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
spin FeIII species in two different distorted cubic environments.15 On the sole basis
of the isomer shifts, the singlet component can be assigned either to LS FeII or LS
FeIII in a local cubic symmetry. However, in agreement with the Raman results, the
singlet component can be straightforwardly attributed to a LS FeII species
corresponding to the reduced form of the complex (probably FeII-MnII because of
the values of the magnetic susceptibility measurements at this temperature16). The
temperature dependence of the Mössbauer parameters indicates that the species
with the largest quadrupole splitting (about 45% area ratio) is not involved in the
phase transition. When the sample is cooled down to 170 K, the area of the LS FeIII
doublet with the small quadrupole splitting decreases and, conversely, the area of
the LS FeII singlet increases. At 80 K this FeIII species is converted almost
completely into the corresponding FeII species. The main finding here is that the
Mössbauer data gives unambiguous evidence for the existence of two structurally
different LS FeIII environments in the sample at room temperature, with only one of
them being involved in the thermal phase transition. The difference between the
two iron sites must probably be traced back to the different occupation of the
interstitial sites either by rubidium ions or by water molecules.17
Table
4.3.
Least-squares
fitted
Mössbauer
data
for
sample
4.1
(Rb0.94Mn[Fe(CN)6]0.88·2.17H2O) obtained in the cooling (293 – 80 K) and heating (170 K,
210 K) modes. (Values in parentheses are standard deviations. Values without standard
deviations were fixed)
T
(K)
293
210
170
140
80
170
210
IS
(mm
s-1)
-0.15
(1)
-0.116
(8)
-0.10
(1)
-0.08
(1)
-0.076
(8)
-0.103
(7)
-0.113
(1)
Doublet
QS
Γ/2
(mm
(mm
s-1)
s-1)
0.46
0.153
(1)
(8)
0.55
0.175
(1)
(5)
0.64
0.173
(1)
(6)
0.70
0.166
(1)
(5)
0.77
0.181
(2)
(2)
0.62
0.172
(3)
(3)
0.54
0.169
(1)
(5)
Area
(%)
IS
(mm
s-1)
45(8)
-0.15
44(2)
42(6)
36(4)
36(1)
34(1)
37(2)
-0.12
(1)
-0.090
(2)
-0.075
(3)
-0.065
Doublet
QS
Γ/2
(mm
(mm
s-1)
s-1)
0.14
0.13
(2)
0.170
0.20
(5)
(1)
0.15
0.23
(2)
0.15
0.28
(2)
0.12
0.34
(6)
Singlet
Area
(%)
39(10)
38(3)
42(4)
29(6)
3(2)
IS
(mm
s-1)
-0.099
(6)
-0.08
(1)
-0.067
(3)
-0.055
(2)
-0.049
(4)
-0.079
(3)
-0.097
(5)
Γ/2
(mm
s-1)
0.118
(6)
0.150
(5)
0.160
(1)
0.160
(9)
0.21
0.20
(2)
0.19
(5)
Area
(%)
16(3)
18(2)
16(3)
35(4)
61(2)
66(2)
63(3)
IS: isomer shift (with reference to metallic iron at 293 K), QS: quadrupole splitting, Γ/2: half-height
width
To confirm the electronic state of the quenched configuration of compound 4.1
(freshly prepared) was cooled down to 5 K at a rate of 20 K min-1 and Mössbauer
spectra were measured between 5 and 20 K (Figure 4.11, Table 4.4). At 20 K the
spectrum can be satisfactorily fitted by using three components; two quadrupolar
doublets and one singlet with approximately 42/38/20% area ratios, respectively.
The approximately equal proportions for similar assigned signals determined at
78
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
room temperature confirm the total quenching of the system, i.e. the HT phase is
still present at 20 K. At 15 K, the intensities of both quadrupole doublets start to
decrease and an enlarged magnetic component appears spontaneously. Below 15 K,
the magnetic sextet was fitted by a least-squares fitting procedure and the obtained
local magnetic hyperfine field has a value of about 200 kOe, This value is much
higher than predicted by the -220<Sz> rule for a single unpaired electron, but
presumably there will also be orbital and dipolar terms.9 These findings are in
reasonable agreement with the report of Ohkoshi et al.10 who revealed an
antiferromagnetic ordering around 11.5 K in the HT phase.
Figure 4.11. 57Fe Mössbauer spectrum of
the quenched HT configuration of sample
4.1 (Rb0.94Mn[Fe(CN)6]0.88·2.17H2O)
recorded at 5 K
Table
4.4.
Least-squares
fitted
Mössbauer
data
for
sample
4.1
(Rb0.94Mn[Fe(CN)6]0.88·2.17H2O) following a rapid cool-down to 20 K. (Values in
parentheses are standard deviations. Values without standard deviations were fixed)
Doublet
T
(K)
20
15
10
5
Doublet
Singlet
IS
(mm
s-1)
QS
(mm
s-1)
Γ/2
(mm
s-1)
IS
(mm
s-1)
QS
(mm
s-1)
Γ/2
(mm
s-1)
IS
(mm
s-1)
Γ/2
(mm
s-1)
-0.076
(7)
-0.069
(7)
0.847
(6)
0.86
(2)
0.190
(4)
-0.072
(8)
0.346
(5)
0.184
(9)
-0.040
(2)
0.21
0.21
-0.070
0.36
0.21
-0.021
-0.014
(10)
0.000
(11)
0.21
0.21
0.21
Sextet
IS
(mm
s-1)
H
(KOe)
-0.14
(1)
-0.08
(1)
-0.07
(1)
130.5
(7)
184.8
(9)
191.4
(8)
Area
(%)
42/38/20/0
34/11/14/41
0/0/12/88
0/0/11/89
IS: isomer shift (with reference to metallic iron at 293 K), QS: quadrupole splitting, Γ/2: half-height
width
4.3.3 Sample 3.7 (Rb0.61Mn[Fe(CN)6]0.86·2.71H2O)
Figure 4.12 displays the evolution of χMT of sample 3.7 as a function of
temperature. The magnetic behaviour clearly shows that the sample does not
exhibit an electron transfer, which is in agreement with the literature for a
RbxMn[Fe(CN)6]y·zH2O compound with this stoichiometry.5,18 The χMT value at
room temperature is 4.75 cm3 K mol-1. This value is exactly what would be
expected for a paramagnetic {S1:S2} = {5/2:1/2} spin system. The room
79
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
χMT (cm K mol )
5.0
Figure 4.12. Temperature dependence of
χMT of sample 3.7
(Rb0.61Mn[Fe(CN)6]0.86·2.71H2O) in the
cooling and heating modes
-1
4.5
3
4.0
3.5
3.0
100
150
200
250
300
350
T (K)
temperature Raman spectrum of sample 3.7 is characterized by two large peaks
around 2159 and 2168 cm-1 with FWHM = 9.2 and 6.0 cm-1, respectively (Figure
4.4). These large line widths reflect the significant deviation from a perfect
RbMn[Fe(CN)6] stoichiometry of the sample and are in good agreement with the
lack of charge transfer in this sample (see Figure 4.5). The Raman spectrum reveals
also the presence of a small amount of the reduced form in this sample (2122 and
2082 cm-1) beside the HT form. This reduced form of iron appears also from the fit
of the room temperature Mössbauer spectrum, which displays a singlet and a
quadrupolar doublet (Figure 4.13, Table 4.5). Indeed, the former species can be
assigned to a small amount of FeII species, while the latter is attributed to FeIII ions.
It is interesting to notice the relatively large quadrupole splitting (0.44 mm s-1) in
the FeIII sub-spectrum. This value is comparable with values observed in sample
4.1 as well as in single crystals of RbMn[Fe(CN)6]·H2O for the ferric ions which do
not show electron transfer.11 This is in contrast with the case of the electron transfer
active FeIII species (observed in samples 3.13 and 4.1 as well as in single crystals,11
which display either a singlet or a doublet with very small quadrupole splitting
(~0.15 mm s-1) in their 57Fe Mössbauer spectrum at room temperature. One may
therefore suggest that the more the environment of the FeIII ion deviates from the
perfect cubic symmetry of the Prussian Blue structure, the more the quadrupole
Figure 4.13. 57Fe Mössbauer spectrum of
sample 3.7
(Rb0.61Mn[Fe(CN)6]0.86·2.71H2O)
acquired at 295 K
Rel. transmission (%)
100
98
96
94
-4
-2
0
2
-1
v (mm s )
80
4
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
splitting increases. In this way, the room temperature Mössbauer spectrum is a
direct measure of the electron transfer capability to be detected at lower
temperatures.
The X-ray diffraction profile of sample 3.7 was fitted in the same way as for
sample 3.13 in the space group F-43m (see chapter 3). For Rb and O two scenarios
were tested: one in which the Rb ions were equally distributed over the two
interstitial sites (4c and 4d) and the vacancies on these sites were filled by O atoms,
and another in which all the Rb ions were placed on the 4c position and the
remaining vacancies on both sites were filled by O atoms. Of these two scenarios
the first clearly gave the best agreement. Unfortunately, refinement of the fractions
of Rb and O led to an unstable refinement. The determined Fe-C and Mn-N
distances were 1.98(2) Å and 2.16(2) Å, respectively, which are close to the values
found for single crystals.11 The profile showed the same weak unindexed peaks as
in the profile of sample 3.13. From the refinement of these X-ray powder
diffraction profiles combined with findings reported in the literature11,12,19 it is clear
that in the structure of compounds which are electron transfer active, the majority
of the Rb ions are located on the 4c position in the F-43m space group and the noncoordinated water molecules are located on the 4d position. Therefore, it appears
that beside a favourable stoichiometry, the distribution of the Rb ions and water
molecules over the interstitial sites is also of crucial importance for determining the
electron transfer activity. Indeed, in the case of single crystals11 it was shown that
even a difference of 75:25 in the distribution of the Rb ions over the two interstitial
sites leads to a reduction of 50% in the electron transfer efficiency.
Table
4.5.
Least-squares
fitted
Mössbauer
data
for
sample
(Rb0.61Mn[Fe(CN)6]0.86·2.71H2O) (Values in parentheses are standard deviations.)
T (K)
Doublet
Singlet
IS (mm s )
QS (mm s )
Γ/2 (mm s )
Area (%)
IS (mm s )
Γ/2 (mm s-1)
-0.156(4)
0.44(2)
0.155(9)
84(13)
-0.13(1)
0.15(7)
-1
293
3.7
-1
-1
-1
IS: isomer shift (with reference to metallic iron at 293 K), QS: quadrupole splitting, Γ/2: half-height
width
4.4 Conclusions
To observe in which way the spectral (Raman and Mössbauer) and structural
characteristics as well as the charge transfer (CT) phase transition vary with the
sample stoichiometry physical and spectroscopic characterization of
RbxMn[Fe(CN)6]y·zH2O samples with different stoichiometries were carried out. In
samples with a stoichiometry which is close to Rb:Mn:Fe = 1:1:1 the transition is
more complete and the hysteresis loop is relatively narrow. The line widths of the
modes present in the CN stretching region in Raman spectroscopy become
narrower. This is in line with previous reports in literature.5 An unprecedented
finding in this chapter is that the temporal stability of the samples depends strongly
81
Chapter 4. Identifying the Fraction of Charge Transfer Active Material in Derivatives of
RbxMn[Fe(CN)6]y·zH2O
on the synthesis conditions. In certain cases (samples 3.13 and 3.7) the magnetic
properties and notably the CT phase transition appear quite stable in time for
several months, but in the case of sample 4.1 the phase transition is completely
suppressed within a few weeks. Interestingly, a heat treatment allows the recovery
of the phase transition suggesting that exchange of water molecules between the
lattice and the atmosphere has a huge influence on the magnetic properties. Further
investigation of this phenomenon might be interesting in view of the application of
this material in humidity sensors.
Another new feature in this chapter is that the room temperature 57Fe Mössbauer
spectra give a direct measure of the amount of electron transfer active material
present in the sample. The Fe ions in the part of the material that exhibit electron
transfer show a doublet with a small quadrupole splitting (~0.10 mm s-1) for LS
FeIII, indicating a very small deviation from cubic symmetry. On the other hand, Fe
ions in the inactive material show a doublet with significantly larger quadrupole
splitting (~0.45 mm s-1). The relative areas of these doublets are directly
proportional to the fraction of electron transfer active material present in the
sample. It is interesting to note on the other hand that the hyperfine Mössbauer
parameters do not allow for a straightforward distinction between the LS FeII and
LS FeIII species in the LT and HT configurations and the transition is better
observed through the variation of the Lamb-Mössbauer factors.
From the results of the refinement of the X-ray powder diffraction profiles may be
concluded that in electron transfer active materials the majority of the Rb ions is
located on one of the two possible interstitial sites in the space group F-43m,
thereby presumably ensuring the cubic environment of the Fe and Mn ions. In
contrast, in non electron transfer active materials the Rb ions appear to be equally
distributed over the two interstitial sites, thus apparently contributing to creating a
less cubic environment for the metal ions.
4.5 Acknowledgements
Sample 4.1 has been synthesised by the group of Azzedine Bousseksou. All
measurements done on this sample have been performed by this group as well as
the DSC and Raman measurements on samples 3.13 and 3.7.
4.6 References
1
2
82
This chapter has been based on the following publication: Salmon, L.; Vertelman, E.J.M.;
Murgui, C.B.; Cobo, S.; Molnár, G.; Koningsbruggen, P.J. van; Bousseksou, A.; Eur. J. Inorg.
Chem., 2009, 760
Because the acquiring of the elemental analysis took more than one month (more than the time
between sample synthesis and first measurements), the sample has had time to incorporate water
molecules from the environment. Therefore the CHN data were assumed to be misleading.
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Because the incorporation of water was unanticipated before any measurements were done, the
sample has not been placed under a protective environment.
http://www.isapps.ca/recoil/
less than 1 week between preparation and measurement
Cobo, S.; Fernandez, R.; Salmon, L. ; Molnár, G. ; Bousseksou, A.; Eur. J. Inorg. Chem., 2007,
1548
A small bump is present around 325 K. Possibly this is due to a small portion of the sample that
transits at a little higher temperature, which is not resolved in the magnetic measurements.
Assuming that the calorimetric effect of this transition is comparable to the calorimetric effect of
the transition around 300 K, this amounts to ~10% of the total transition.
Luzon, J. ; Castro, M. ; Vertelman, E.J.M. ; Gengler, R.Y.N.; Koningsbruggen, P.J. van ;
Molodtsova, O. ; Knupfer, M. ; Rudolf, P. ; Loosdrecht, P.H.M. van ; Broer, R.; J.Phys.Chem.C
2008, 112, 5742
a) Gawali-Salunke, S.; Varret, F.; Maurin, I.; Enachescu, C.; Malarova, M.; Boukheddaden, K.;
Codjovi, E.; Tokoro, H.; Ohkoshi, S.; Hashimoto, K.; J. Phys. Chem. B 2005, 109, 8251; b) Sato,
O.; Einaga, Y.; Fujishima, A.; Hashimoto, K.; Inorg. Chem. B 1999, 38, 4405; c) Einaga, Y.;
Sato, O.; Iyoda, T.; Kobayashi, Y.; Ambe, F.; Hashimoto, K.; Fujishima, A.; RIKEN Rev. 1997,
16, 41
N. N. Greenwood, T. C. Gibb, Mössbauer spectroscopy, Chapman and Hall Ltd., London 1971
Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K. ; Coord. Chem. Rev. 2005, 249, 1830
Chapter 5
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K.. J. Phys. Soc. Jpn. 2002, 71, 2078
No FTIR spectra have been taken for this sample to confirm this proposal. FTIR spectra always
show a significant amount of water in each sample of RbxMn[Fe(CN)6]y·zH2O samples, coming
from both the sample itself and water molecules present in air. It is therefore difficult to estimate
a change in water content on IR spectra alone.
It is certain that water is present in the freshly prepared sample, but it is not completely certain to
what extent. Heating the sample will therefore always result in the loss of water molecules, and
cannot give an estimate of the change in water molecules after storing in ambient conditions.
This is the reason why thermo gravimetric analysis can not be used as an indication for the
amount of water molecules that is lost.
N. N. Greenwood, T. C. Gibb, Mössbauer spectroscopy, Chapman and Hall Ltd., London 1971
Indeed, if we would assume only FeIII, MnII, RbI and CN-, the ratio of Rb:Mn:Fe should have
been 0.94:1:0.98. However, the concentration of Fe(CN)6 is 0.88, which is lower, thereby
confirming the possibility that FeII rather than FeIII is present.
See also chapter 5 and 6 where this possibility is looked at in more detail.
Chapter 2
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K. J. Phys. Soc. Jpn. 2003, 72, 2698
83
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 5. Light- and Temperature-Induced Electron
Transfer in Single Crystals of RbMn[Fe(CN)6]·H2O1
5.1 Introduction
A further development of the research of the PBA (magnetic) properties would
highly benefit from an extended knowledge on the relationship between structural
features and the material’s physical properties, which ideally requires single
crystals. Nevertheless, the number of crystal structures of PBAs with
AM[M’(CN)6]·zH2O stoichiometry known to date is small and limited to
compounds that do not exhibit physical properties relevant in terms of
applications.2,3,4 This also applied to single crystals of a rubidium manganese
hexacyanoferrate which have been obtained recently.5 Its inactivity in displaying
electron transfer is probably due to the unbalanced redox potentials of the
individual metal sites. This is a consequence of the unfavourable stoichiometry
(Rb0.61Mn[Fe(CN)6]0.87·1.7H2O) involving Fe(CN)6 vacancies and an associated
number of water molecules as has been suggested previously.6 This chapter reports
on single crystals of RbMn[Fe(CN)6]·H2O that have been obtained which show
light- and temperature-induced switching, in which only 50% of the metal ions
participate.
5.2 Experimental Section
5.2.1 Synthesis
All chemicals were purchased at Sigma Aldrich and used without further
purification. Single crystals of RbMn[Fe(CN)6]·H2O have been obtained by a
modified version of the crystallization method developed by Kepert et al.7 (Figure
5.1). A 10 mL vial was placed in a 100 mL jar with screw cap. The jar was filled
with an aqueous solution of 0.5 M RbCl at room temperature to approximately
Aqueous solutions of:
RbCl
MnCl2
Figure 5.1 Schematic
representation of the slow
diffusion method used to
obtain single crystals of
RbMn[Fe(CN)6]·H2O
K3Fe(CN)6
85
Chapter 5. Light- and
RbMn[Fe(CN)6]·H2O
Temperature-Induces
Electron Transfer
in
Single
Crystals
of
0.5 cm above the top of the vial. 2 mL of an aqueous solution of 1 M MnCl2·4H2O
(cooled on ice) was injected at the bottom of the vial and 2 mL of an aqueous
solution of 1 M K3Fe(CN)6 (cooled on ice) was injected at the bottom of the large
jar. The jar was closed with the screw cap and placed in a water bath of 45°C for 2
days. After 2 days dark brown cubic crystals had been formed. These were
filtrated, washed with water and 96% ethanol of room temperature, and allowed to
dry in air for 1 day, affording RbMn[Fe(CN)6]·H2O (sample 5.1) in a 47% yield.
Anal. calcd. for RbMn[Fe(CN)6]·H2O: C 19.46, N 22.69, H 0.54. Found: C 19.41,
N 22.75, H 0.59.
5.2.2 Magnetic Measurements
For details on the X-ray powder diffraction measurements, see section 2.2.3.
5.2.3 FTIR Spectroscopy
For details on the FTIR spectroscopy measurements, see section 2.2.4
5.2.4 Raman Spectroscopy
Inelastic light scattering experiments in the CN stretching region were performed
in a 180o backscattering configuration, using a triple grating micro-Raman
spectrometer (T64000-Jobin Yvon), equipped with a liquid nitrogen cooled CCD
detector. The frequency resolution was better than 1 cm-1. The samples were placed
in a liquid helium cooled optical flow-cryostat (Oxford Instruments). The
temperature was stabilised with an accuracy better than 0.1 K throughout the whole
temperature range. A fraction of the second harmonic output of a Nd:YVO4 laser
(532.6 nm, Verdi-Coherent) was focused on the sample using a 50x microscope
objective. All reported spectra were recorded in parallel polarization mode, with a
total integration time of 15 minutes.
5.2.5 57Fe Mössbauer Spectroscopy
For details on the 57Fe Mössbauer spectroscopy, see section 4.2.7.
5.2.6 X-Ray Crystallography
All details on the crystallographic data collection and refinement can be found in
Appendix II.
5.3 Results and Discussion
5.3.1 Description of the Crystal Structure of RbMn[Fe(CN)6]·H2O
The best structural model at 293(1) K is based on the cubic space group F-43m (a =
10.521(2) Å, figure 5.2). Fe is in an octahedral environment formed by 6 C atoms
of the cyano ligands bridging the Fe and Mn ions. Rb and O (H2O) are
stoichiometrically present and located in the fully occupied interstitial sites. Rb is
distributed over two positions (4c: s.o.f. 0.7512(4) and 4d: s.o.f. 0.2488(4)). The
interatomic distances of 1.929(4) Å for Fe – C and 2.205(5) Å for Mn – N are
86
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
indicative for low spin (LS) FeIII and high spin (HS) MnII, respectively,4,8,9 forming
the high temperature (HT) configuration of the material.10 Unfortunately, the
structure at 100 K could not be elucidated, despite numerous efforts. The data
recorded at 100 K yielded an inconclusive structural model, caused by the
diffraction measurements revealing many diffuse spots which did not result in a
realistic unit cell. In addition about 20% of the reflections did not fit in this
orientation matrix. See for further details on the elucidation of the structure at
100 K Appendix II.
a)
b)
Figure 5.2. a. Perspective ORTEP drawing of part of RbMn[Fe(CN)6]·H2O
(50% probability level). b. Representation of the unit cell of
RbMn[Fe(CN)6]·H2O (F-43m).
5.3.2 Magnetic Measurements
The temperature dependence of χMT (with χM being the molar magnetic
susceptibility per formula unit and T the temperature) of RbMn[Fe(CN)6]·H2O is
shown in figure 5.3a. At room temperature the χMT value is 4.66(5) cm3 K mol-1.
This corresponds well to the χMT value of 4.75 cm3 K mol-1 expected for a weakly
coupled paramagnetic {S1;S2} = {5/2;1/2} material (the HT configuration,
assuming a g-value of 2). Upon cooling χMT decreases around 250 K until it
reaches a value of 4.04(4) cm3 K mol-1 at 150 K. The T1/2↓ value for this process is
237 K. For a paramagnetic {S1;S2} = {2;0} material (HS MnIII and LS FeII, the low
temperature (LT) configuration) a χMT value of 3.00 cm3 K mol-1 is expected,
which is lower than the value observed in the experiment. If exactly 50% of the
material is switched from the HT to the LT configuration, the expected χMT value is
3.89 cm3 K mol-1; this is close to the value found in our experiment. It therefore
seems feasible that approximately 50% of the MnFe pairs have transformed to the
LT configuration on cooling, which is intriguing considering that microcrystalline
samples of comparable stoichiometry exhibit a virtually complete conversion.6 The
87
Chapter 5. Light- and
RbMn[Fe(CN)6]·H2O
Temperature-Induces
a)
Electron Transfer
in
Single
Crystals
of
b)
2000
4.75
1800
1400
M (emu/mol)
3
-1
χMT (cm K mol )
1600
4.50
4.25
4.00
1200
1000
800
600
400
200
3.75
0
50
100
150
200
T (K)
250
300
350
0
-200
0
5
10
15
20
T (K)
Figure 5.3a Temperature dependence of χMT of RbMn[Fe(CN)6]·H2O b Temperature
dependence of the magnetisation of RbMn[Fe(CN)6]·H2O
reverse process takes place at a T1/2↑ of 292 K. The broad hysteresis of 55 K
reflects the cooperativity of the electron transfer process.
Figure 5.3b shows the magnetisation versus temperature for a sample of single
crystals under a field of 0.01 T.11 In analogy with powdered samples of
RbxMn[Fe(CN)6]y·zH2O the sample shows ferromagnetic ordering below 12 K. The
magnetic ordering in the LT phase, and hence the observed increase in χMT, has
been proposed to arise from a mechanism of mixed-valence electron delocalization
of the Mn ions similar as reported for the FeIII ions in Prussian Blue
FeIII4[FeII(CN)6]3·14H2O.10,12 The value of the magnetisation at 4 K is 1819 emu
mol-1, which is roughly half the value of the magnetisation that was found for the
electron transfer active samples in chapter 3. The applied field is below the
saturation field, but it is the same field as was applied to the samples in chapter 2
and 3. This magnetisation value again confirms the 50% switching ratio, because
only the LT phase shows a ferromagnetic ordering.
5.3.3 FTIR spectroscopy
Figure 5.4 shows the FTIR spectrum of RbMn[Fe(CN)6]·H2O at room temperature
in the CN stretching region. It consists of a pronounced peak at 2150 cm-1, a small
shoulder at 2100 cm-1 and a pronounced peak at 2072 cm-1. In agreement with the
previous chapters and the Raman data (see below) the peak centred around
2150 cm-1 is assigned as the CN stretching between FeIII and MnII (FeIII-CN-MnII)
and the peaks centred around 2072 cm-1 and 2100 cm-1 as the two CN stretching
frequencies of different symmetry in FeII-CN-MnIII. The spectrum shows neither
the spectrum at room temperature as shown by electron transfer inactive samples
(sample 2.1 and 3.7; pronounced presence of bands around 2070 cm-1 and
2150 cm-1, intensity of band at 2150 cm-1 is higher than that of 2070 cm-1, no bands
88
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Figure 5.4 Room temperature FTIR
spectrum of RbMn[Fe(CN)6]·H2O
Intensity (a.u.)
Sample 5.1
2300
2200
2100
2000
1900
-1
ν (cm )
distinguishable around 2100 cm-1) nor that as shown at room temperature by
electron transfer active samples (samples 2.2-2.4; samples 3.1-3.5; samples 3.83.13; band at 2150 cm-1 hardly present and the bands at 2070 cm-1 and 2100 cm-1
merged together as one broad band). In this respect the spectrum of the present
sample is in between these two types of spectra: all three bands are visible and the
band at 2070 cm-1 is more pronounced than the band at 2150 cm-1. This seems to
indicate again that only partial electron transfer takes place.
5.3.4. 57Fe Mössbauer Spectroscopy
57
Fe Mössbauer spectra of single crystals of RbMn[Fe(CN)6]·H2O at 293 K and
50 K are shown in figure 5.5. Table 5.1 shows the values of the hyperfine
parameters obtained from the least square fitting procedure. At 50 K and 293 K the
Mössbauer spectra can be satisfactorily fitted by two components, a single line and
Figure 5.5. 57Fe Mössbauer spectra of
RbMn[Fe(CN)6]·H2O at 50 and 293 K
101
100
Relative Transmission (%)
99
98
97
96
95
94
50K
100
98
96
94
293K
92
-4
-2
0
2
4
v (mm/s)
89
Chapter 5. Light- and
RbMn[Fe(CN)6]·H2O
Temperature-Induces
Electron Transfer
in
Single
Crystals
of
a quadrupolar doublet, with area ratios 51(6):49(6) at 50 K and 43(7):57(7) at
293 K, respectively. Prior to measuring the room temperature spectrum the sample
was warmed to 350 K to ensure complete switching to the HT phase. On the basis
of the hyperfine parameters and the strong temperature dependence of the
quadrupole splitting, the doublet observed at 293 K can be assigned to a low spin
(LS) FeIII species in a slightly distorted cubic environment.13 The singlet in the high
temperature phase can be assigned either to LS FeII or LS FeIII in a local cubic
environment. These two species cannot be distinguished on the value of the isomer
shifts alone.13 However, based on the Raman spectroscopic data (section 5.3.4)
which show only two bands at 295 K (ascribed to CN stretching between HS MnII
and LS FeIII) we assign this singlet to LS FeIII. A singlet for LS FeIII in 57Fe
Mössbauer spectra has been reported before.14 At 50 K the Raman data clearly
show multiple bands in the CN stretching region. These bands show that (part of)
the compound has switched to LS FeII and HS MnIII, therefore the singlet at 50 K
has been assigned to LS FeII. These data suggest that in the HT phase two LS FeIII
sites with different local environment are present, only one of which switches upon
lowering the temperature. The present assignment of the 57Fe Mössbauer spectra
reported in this chapter is in agreement with the analysis of the spectra reported in
chapter 2 and 4.
The partial switching observed in the magnetic susceptibility measurements can be
explained by proposing two or more distinct iron sites differing in electron transfer
capability. This is supported by the 57Fe Mössbauer spectrum recorded at 293 K
revealing approximately equal area fractions for two different LS FeIII sites (Figure
5.5). Inequivalent Fe and Mn environments may be generated by a particular
distribution of the different unit cells that originate from the observed disorder of
the Rb ions (and water molecules) over the two interstitial sites, 4c and 4d in a 3:1
ratio, in the space group F-43m (see section 5.4).
Table 5.1. 57Fe Mössbauer parameters for crystals of RbMn[Fe(CN)6]·H2O
T (K)
50
293
δ (mm/s)
-0.033(4)
-0.067(5)
-0.138(2)
-0.150(2)
∆ (mm/s)
0
0.78(2)
0
0.43(1)
Γ/2 (mm/s)
0.23(2)
0.20(1)
0.16(2)
0.145(7)
A/Atot
51(6)%
49(6)%
43(7)%
57(7)%
δ: isomer shift relative to natural iron at room temperature
∆: quadrupole splitting
Γ/2: half width at half maximum
A/Atot: relative area of the sub spectrum
Standard deviations are given in parentheses. Values without parentheses were fixed for the fit.
5.3.5. Raman Spectroscopy
Figure 5.6 shows the evolution of the 532 nm excited Raman spectrum of the CN
stretch vibrations in a single crystal of RbMn[Fe(CN)6]·H2O upon cooling. At
temperatures above the phase transition (T > 250 K) two strong bands are observed
at 2155 and 2164 cm-1. At lower temperatures, new spectral features appear at the
90
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
-2
P = 1700 W cm
T = 295 K
-2
250 K
-2
200 K
-2
150 K
-2
100 K
-2
50 K
-2
4K
-2
4K
Intensity (a.u.)
1700 W cm
1700 W cm
1700 W cm
1700 W cm
1700 W cm
1275 W cm
3000 W cm
2000
2050
2100
2150
2200
-1
ν (cm )
Figure 5.6. Raman scattering spectra (black squares) of a single crystal of
RbMn[Fe(CN)6]·H2O at various temperatures. Multiple lorentzian line shapes were
combined to obtain a fit. The grey and white dashed areas represent vibrations
characteristic for the HT and LT configuration, respectively. The excitation density (P) is
indicated. The spectra are normalized with respect to the integrated intensity of the HT lines
and are shown with an offset for clarity.
cost of these lines, mostly at lower energies. These evolve into several new peaks
at 2086, 2091, 2110, 2130 and 2202 cm-1. The vibrational stretching frequencies of
the CN moiety in Prussian Blue analogues are highly sensitive to the charges of the
metal ions surrounding the cyano group. Typical ranges of the vibrational
stretching frequencies of cyano ligands bridging between two 3d metal ions are
listed in table 5.2. A general rule of thumb hereby is that as the oxidation state of
either one of the metal centres increases, so does the vibrational frequency and vice
versa.
In the HT phase of RbMn[Fe(CN)6]·H2O, (LS FeIII and HS MnII), only
FeIII-CN-MnII cyano bridges exist. From table 5.2 can be inferred that these lines in
the Raman spectrum of the HT phase (at 2155 and 2164 cm-1) occur in the expected
frequency range. Group theory analysis shows that the vibrational stretching mode
of the free CN- ion (A1 symmetry) splits up into an A1, an E and a T2 normal mode,
91
Chapter 5. Light- and
RbMn[Fe(CN)6]·H2O
Temperature-Induces
Electron Transfer
in
Single
Crystals
of
Table 5.2. Typical vibration frequencies of CN moieties in various types of cyano
bridges.16 M = 3d transition metal
Cyano bridge
FeII – CN – MII
FeII – CN – MIII
FeIII – CN – MII
FeIII – CN – MIII
Typical frequency range
2065-2110 cm-1
2095-2140 cm-1
2146-2185 cm-1
2180-2210 cm-1
when the CN moiety is placed on the C2v site of the F-43m (with Td2 symmetry)
space group. Of these normal modes the A1 and E modes are expected to be
observed in the parallel polarization spectra of figure 5.6. Therefore, the two lines
observed in the HT phase spectra are assigned to the A1 and E normal modes of the
CN moiety in the FeIII-CN-MnII cyano bridge, in agreement with the assignment of
Cobo et al.15 The assignment for the CN stretching region of the Raman spectra at
high temperatures in this chapter is in agreement with the assignment of the Raman
spectra at room temperature reported in chapter 2.
Magnetic measurements and 57Fe Mössbauer spectroscopic measurements indicate
that in the crystals only 50% of the 3d metal centres undergo the charge transfer
transition. As a result, the situation at lower temperatures is such, that there are
four types of 3d metal centres present: LS FeIII, HS MnII, LS FeII and HS MnIII.
Consequently, at temperatures below the phase transition of the material, there are
four types of cyano bridges (Fe-CN-Mn) in the system. The entities FeIII-CN-MnII
and FeII-CN-MnIII are the respective HT and LT configurations of the material.
However, the transition being only partial results in the presence of FeIII-CN-MnIII
and FeII-CN-MnII units as well. Consequently, in the LT phase of the crystalline
material, we expect to see peaks arising from all types of cyano bridges. In
addition, each of these vibrational modes are expected to split into multiple normal
modes due to the crystallographic symmetry, which further complicates the
vibrational spectrum. In fact, assuming that the LT phase has a lower symmetry
compared to the HT phase, the vibrational mode for each type of cyano ligand
present is expected to split even more than is the case for the one cyano bridge in
the HT phase. Indeed, the experimental vibrational spectrum at low temperatures
shows a large number of lines in the region 2050-2210 cm-1. The lines of the HT
phase (grey lines in figure 5.6) are still present in these spectra and are assigned to
the normal modes of the FeIII-CN-MnII cyano unit. In addition, the new line
observed at 2202 cm-1, reveals the presence of an FeIII-CN-MnIII bridge (the only
bridge that would explain a higher frequency vibration), consistent only with a
partial transition. The corresponding frequencies of the FeII-CN-MnII and FeII-CNMnIII entities are expected and observed at lower frequencies, i.e. in the
overlapping ranges 2065-2110 cm-1 and 2095-2140 cm-1, respectively. However,
considering the large number of overlapping lines expected and observed, it is non
trivial to give an accurate full assignment of the observed LT phase modes (white
dashed area in figure 5.6). Still, the observed lower frequency lines are considered
as ‘markers’ of the LT phase, since they are only expected to be observed below
92
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
the (partial) charge transfer transition temperature. Also, similar modes have been
observed and assigned to the LT configuration of RbMn[Fe(CN)6]·H2O
previously.15
It is clear from figure 5.6 that spectra below the transition (with temperature 100,
50 and 4 K (P = 1275 W cm-2)) show lines characteristic for both the LT and HT
configuration, indicating a partial switching, consistent with the magnetic
susceptibility and 57Fe Mössbauer spectroscopic data. In contrast, the 4 K spectrum
recorded with P = 3000 W cm-2 shows lines originating from the HT configuration
only. This is quite understandable in view of the photo-activity of the material.
This allows for optical switching from the LT configuration to a metastable HT
configuration,17 above a certain excitation power density. Here, the 4 K spectra
demonstrate complete optical switching using 532 nm excitation light. The c.w.
excitation power threshold for complete switching at 4 K is found to be ~2000 W
cm-2 for 532 nm light.
Comparing the temperature evolution of the Raman spectrum with the magnetic
measurements, it appears that the phase transition is lagging in the Raman data,
meaning that the transition seems to start above 200 K, being completed at 50 K.
This can be explained by considering the photo-activity of the material. As was
observed before,18 the material is known to be photo-excited to the HT
configuration by 532 nm light when in the LT configuration. Consequently, the
spectrum at temperatures where this photo-excitation occurs is the combined result
of the temperature-induced HT to LT conversion and the light-induced LT to HT
phase. The efficiency of this photo-conversion is temperature dependent and seems
to be the highest in temperature regions of bistability of the HT and LT
configurations, i.e. near the hysteresis loop in the magnetic data. When combined,
these effects can explain the apparent downshift in temperature of the hysteresis
loop in the Raman data. The photo-excitation of the sample is also reflected in the
optical properties of the material as a colour change from dark brown to light
brown of the sample surface on light irradiation. While this colour change has been
observed on the sample surface at temperatures of 200, 150 and 100 K, it is not
observed below 60 K. In the magnetically ordered phase (below 12 K), the colour
change is again observed upon photo-conversion to the metastable HT
configuration. Since there seems to be no photo-excitation between 60 and 12 K,
the spectrum at 50 K is considered to be the only ‘pure’ LT phase Raman spectrum,
showing both HT and LT configurations present at low temperatures.
The temperature dependence of the spectrum in the 2135-2180 cm-1 range shows
an evolution from two distinct lines above the charge transfer transition (the grey
HT lines in figure 5.6) to a more broadened structure below the transition. This can
be explained by the above considerations as well: in this range, many different
overlapping lines are expected to give rise to spectral weight, making it difficult to
resolve their separate contributions. The 2135-2180 cm-1 range is fitted with only
93
Chapter 5. Light- and
RbMn[Fe(CN)6]·H2O
Temperature-Induces
Electron Transfer
in
Single
Crystals
of
two lorentzians throughout the whole temperature range, which leads to the
observed, rather complex, artificial temperature dependence of the peak heights of
these lines.
5.4. Tentative Model Explaining the Presence of Different
Iron/Manganese Sites
There are several reasons to believe that the consistently observed partial switching
is an intrinsic property of the crystals. Firstly, magnetic susceptibility
measurements recorded for independently prepared samples confirm the
approximately 50% switching ratio. Secondly, it is also detected by 57Fe Mössbauer
spectroscopy and Raman spectroscopy. The latter technique is therefore crucial and
is in this instance very conclusive, as the spectra evidencing reproducible extents of
partial switching have repeatedly been recorded on different individual single
crystals. In addition, a sample made of powdered crystals also shows the same
switching ratio.
The presence of different second coordination spheres around the iron ions (and
most probably also the manganese ions) can be explained by the disorder in the Rb
and O occupation observed in the crystal structure. Here one of the possible models
that could explain the presence of two different Fe (and concomitant Mn) sites is
given. This is not the only possible model.
Single crystal X-ray analysis reveals that the Rb ions (and water molecules) are
disordered over the two interstitial sites, 4c and 4d in the space group F-43m, in a
3:1 ratio. As a result, there is a distribution of two different types of cells, A and B
(figure 5.7), which are related by a centre of inversion.
A
B
Figure 5.7. The two types of cells, A and B. Spheres indicate the Rb
positions; interstitial sites without a sphere contain water molecules.
Solid lines represent bridging CN ligands, metal ions are positioned at
the vertices of these solid lines. See also Figure 5.2b.
94
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
The presence of two distinct cells can account for the presence of different iron and
manganese environments. For example, an iron atom that is sandwiched in between
neighbouring cells of type A and type B, has a different geometry (II, figure 5.8)
than one that is located between two cells of the same type (I, figure 5.8).
In case of a non-random distribution of cells A and B, for example the one
represented in figure 5.9a, one can think of a “super unit cell” in this case
constructed from 8 building blocks that are constituted by the two possible cells, A
and B. Careful evaluation of the iron and manganese environments in this “super
unit cell” reveals three distinct iron (and manganese) sites in a 2:1:1 ratio (figure
5.9b). It is assumed that only particular types of iron environments (although
unknown at this stage) will allow for valence tautomerism, which would thus
account for the 50% switching observed in the magnetic measurements.
Unfortunately, with the crystallographic data obtained in this study, it is not
possible to resolve such a presumed higher positional ordering.
A random ordering of the cells results in a large variety of iron and manganese
sites. It is unlikely that a random distribution of cells A and B can account for the
reproducible 50% switching observed for compounds originating from various
independent preparations.
I
II
Figure 5.8. Two different iron (and concomitant manganese) sites.
Dark grey spheres are Rb ions originating from cell A, light grey
spheres from cell B; interstitial sites without a sphere contain water
molecules. Solid lines represent bridging CN ligands, metal ions are
positioned at the vertices of these solid lines.
95
Chapter 5. Light- and
RbMn[Fe(CN)6]·H2O
A
Temperature-Induces
Electron Transfer
in
Single
Crystals
of
B
Figure 5.9. a) Distribution of cells A (clear box) and B (shaded box), creating a “super unit
cell”. b) Three possible iron and manganese geometries present in the “super unit cell”
(from top to bottom: 2:1:1 ratio). Dark grey spheres represent rubidium ions that originate
from cell type A (¾ of the total amount of cells); Light grey spheres originate from cell
type B (¼ of the total amount of cells). Interstitial sites without spheres contain water
molecules. Solid lines represent bridging CN ligands, metal ions are positioned at the
vertices of these solid lines.
5.5. Conclusions
The crystal structure of the Prussian Blue analogue RbMn[Fe(CN)6]·H2O was
determined, and it was shown that the crystals undergo a partial, reversible charge
transfer between the HS MnII, LS FeIII and the HS MnIII, LS FeII configuration
under the influence of temperature. Raman spectroscopy also shows a transition
from the LT to the HT configuration upon irradiation with light in single crystals.
The partial switching in the crystals might be assigned to distinctly different Fe
sites brought about by unequal Rb/H2O distributions.
The availability of single crystals gives the opportunity to perform certain
measurements that cannot be done on powdered samples. These additional
measurements will help in further understanding the specific processes involved in
the temperature- and light-induced electron transfer in RbMn[Fe(CN)6]·H2O
Prussian Blue Analogues.
96
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
5.6 Acknowledgements
The Raman spectra in this chapter have been measured by Tom Lummen, the 57Fe
Mössbauer spectra by Gabor Molnár and the crystallographic measurements have
been done by Auke Meetsma.
5.7 References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
This chapter has been based on the following publication: Vertelman, E.J.M.; Lummen, T.T.A.;
Meetsma, A.; Bouwkamp, M.W.; Molnar, G.; Loosdrecht, P.H.M. van; Koningsbruggen, P.J.
van; Chem. Mater. 2008, 20, 1236
Witzel, M.; Ziegler, B.; Babel, D.; Z. Anorg. Allg. Chem. 2000, 626, 471
Ziegler, B.; Witzel, M.; Schwarten, M.; Babel, D.; Z. Naturforsch.B 1999, 870
Dong, W.; Zhu, L.-N.; Song, H.-B.; Liao, D.-Z.; Jiang, Z.-H.; Yan, S.-P.; Cheng, P.; Gao, S.;
Inorg. Chem. 2004, 43, 2465
Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S.-I.; Z. Anorg. Allg. Chem. 2007, 633, 1134
See chapter 2
Personal communication: Chapman, K. W.; Southon, P. D.; Kepert, C. J.; 2006
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K.; J. Phys. Soc. Jpn. 2002, 71, 2078
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K.; J. Phys. Soc. Jpn. 2003, 72, 2698
Ohkoshi, S.-I.; Tokoro, H.; Hashimoto, K.; Coord. Chem. Rev. 2005, 249, 1830
This field is lower than the field where saturation occurs. The order of magnitude for this
saturation field is ~0.1 T
Mayoh, B.; Day, P.; J. Chem. Soc., Dalton Trans., 1976, 1483
N. N. Greenwood, T. C. Gibb, Mössbauer spectroscopy, Chapman and Hall Ltd., London 1971.
a) He, Y., Tang, G., Liang, F., Huang, Y., Chen, Z., Phys. B, 2007, 393, 143-146; b) Reguera, E.,
Fernández-Bertrán, J., Hyperfine Interact., 1994, 88, 49-58
Cobo, S., Fernández, R., Salmon, L., Molnár, G., Bousseksou, A., Eur. J. Inorg. Chem. 2007,
1549-1555
a) Bertrán, J. F., Pascual, J.B., Hernández, M., Rodríguez, R., Reactivity of Solids, 1988, 5, 95100; b) Nakamoto, K., Infrared and Raman Spectra of Inorganic and Coordination Compounds,
Wiley-Interscience, 1986, ISBN 0-471-01066-9; c) Reguera. E. , Bertrán J.F., Díaz, C., Blanco,
J., Rondón, S., Hyperfine Interact., 1990, 53, 391; d) Shinamoto, N., Ohkoshi, S.-I., Sato, O.,
Hashimoto, K., Chem. Lett. 2002, 31, 486; e) Sato, O., Einaga, Y., Fujishima, A., Hashimoto, K.,
Inorg. Chem. 1999, 38, 4405; f) Hester, R. E., Nour, E. M., J. Chem. Soc., Dalton Trans. 1981,
939
Tokoro, H.; Ohkoshi, S.-I.; Hashimoto, K.; Appl. Phys. Lett. 2003, 82, 1245
Moritomo, Y., Hanawa, M.,Ohishi, Y.,Kato, K.,Takata, M., Kuriki, A., Nishibori, E, Sakata, M.,
Ohkoshi, S., Tokoro, H.,Hashimoto, K., Phys. Rev B 2003, 68, 144106
97
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 6. Further Characterisation of Single Crystals of
RbMn[Fe(CN)6]·H2O1
6.1 Introduction
In the previous chapter single crystals of RbMn[Fe(CN)6]·H2O were introduced. As
was clear from the magnetic measurements, Raman spectroscopy and 57Fe
Mössbauer spectroscopy, these single crystals only showed a switching behaviour
of 50%. This is in contrast to the behaviour that was found for powdered samples
with a similar stoichiometry, which show nearly complete switching behaviour.2 In
this chapter the single crystals are investigated in further detail. Specifically,
temperature dependent X-ray powder diffraction, more detailed 57Fe Mössbauer
spectroscopy, X-ray photoemission spectroscopy and electron spin resonance
spectroscopy is used. Furthermore, the thermal conductivity as a function of
temperature is investigated.
6.2 Experimental Section
6.2.1.Synthesis and Characterisation
The single crystals (RbMn[Fe(CN)6]·H2O) are the same crystals as sample 5.1 in
chapter 5. Its detailed synthesis and characterisation can be found there.
For the X-ray powder diffraction measurements a few single crystals of sample 5.1
have been carefully grinded with the use of a mortar and pestle.
In order to compare the found results for the single crystals with the properties of a
powdered sample, sample 3.3 of chapter 3 (Rb0.91Mn[Fe(CN)6]0.97·1.53H2O) has
been used in the electron spin resonance measurements. Its detailed synthesis and
characterisation can be found there.
6.2.2 X-ray Powder Diffraction
X-ray diffraction measurements were carried out in transmission geometry using a
Huber diffractometer operating with Mo Kα radiation and equipped with a G670
Guinier camera with integrated imaging plate system. Data were collected in the
temperature range 300 K to 125 K to 300 K with a step size of 25 K. The total
integration time was 3 hours per diffraction profile using a closed-cycle
refrigerator. The resulting diffraction profiles were analyzed using the GSAS
software suite.3
6.2.3.57Fe Mössbauer Spectroscopy
All details on the 57Fe Mössbauer spectroscopy have been described in paragraph
4.2.7.
99
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
6.2.4. X-Ray Photoemission Spectroscopy (XPS)
X-ray photoemission spectroscopy (XPS) data were collected using a Scientat R4000 spectrometer equipped with a monochromatic Al Kα X-Ray source
(hν = 1486.6 eV); the photoelectron take off angle was 90° at an X-probe Surface
Science Instrument SSX-100 with a 37° photoelectron take off angle equipped with
a monochromatic Al Kα X-Ray source (hν = 1486.6 eV). The spot diameter was
600 µm, and a microscope was used for sample alignment. An electron flood gun
was used to compensate for sample charging. The spectrometer operated at a base
pressure of 1·10-9 Torr. Evaporated gold films supported on mica served as
substrates. Samples were produced following two different procedures: (1) the
crystalline sample was grinded and dispersed in distilled-deionised (18 mΩ) water,
stirred for 5 minutes, and a few drops of the suspension were left to dry in air on a
substrate, (2) a few crystals were spread on the bare gold/mica substrate (crystals
were held in position by a grid screwed on top of the sample holder). For procedure
(1) the samples were introduced into ultra high vacuum as soon as dry. For both
procedures the samples were placed on a nitrogen cooled probe equipped with a
Lakeshore 331 cryogenic temperature controller. Temperatures ranged from 350 to
140 K. All binding energies were referenced to the nitrogen signal emitted by the
cyanide groups at 398 eV.4 No X-ray induced sample degradation was detected.
Spectral analysis included a Tougaard background subtraction5 and peak
deconvolution employing Gaussian line shapes using the WinSpec program
developed at the LISE laboratory, University of Namur, Belgium.
6.2.5. Electron Spin Resonance (ESR) Spectroscopy
Electron spin resonance spectra were recorded on powdered and single crystal
samples (resp. Rb0.91Mn[Fe(CN)6]0.97·1.53H2O and RbMn[Fe(CN)6]·H2O) at 9.41
GHz using a Bruker Elexsys and a 222.4 GHz spectrometer built at the Budapest
University of Technology and Economics. A comparison of spectra of the same
sample of dilute MnII in MgO under similar conditions shows similar sensitivities
for the two spectrometers.
6.2.6.Thermal Conductivity Measurements
The thermal conductivity was measured on cube-shaped crystals with typical
dimensions of (0.6 x 0.6 x 1) mm3 using a two sensor-two heater technique.6 The
temperature gradient was determined using a differential Au/Fe-Chromel
thermocouple.
6.3 Results and Discussion
6.3.1. X-Ray Powder Diffraction
The variation of the powder diffraction profiles of single crystals with temperature
can be found in figure 6.1. At room temperature the diffraction profile shows peaks
indicative for the space group F-43m (the HT phase) only. This is the same space
100
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
group as was found in chapter 5. When cooling down, these peaks decrease in
intensity while at the same time other peaks grow. These peaks are clearly at a
different position than the peaks of the HT configuration and hence indicate the
growth of a structure with a different symmetry than the HT phase. The new peaks
are indicative for the space group I-4m2 (the LT phase). This space group was
found previously by Moritomo et al. for the LT phase of RbxMn[Fe(CN)6]y·zH2O
powders.7 The peaks of the HT phase never completely vanish. When heating the
sample back to room temperature the peaks for the LT phase decrease again in
intensity, but still some residual intensity is visible at room temperature. The
diffraction patterns at 300 K before and after cooling and warming up again are not
the same. This indicates that the sample switches back to the HT phase upon
heating, but the switching is not yet complete at room temperature.
I-4m2
Intensity (a.u.)
300
275
250
225
200
175
150
125 heating
150
175
200
225
250
275
300 cooling
F-43m
5.0
7.5
10.0
12.5
15.0
17.5
20.0
22.5
25.0
27.5
30.0
o
2θ( )
Figure 6.1. Temperature variation of X-ray powder diffraction profiles of single crystals
(RbMn[Fe(CN)6]·H2O). Arrows pointing upwards indicate peaks indicative for space group
I-4m2 (the LT phase) and arrows pointing down indicate peaks indicative for space group
F-43m (the HT phase).
The profiles were refined with the space group F-43m for the HT phase, as found
for the single crystals previously.8 For the LT phase the space group I-4m2 as
found by Moritomo et al.7 was used. The space group P4/mmm reported by Tokoro
et al.9 was also tried for the LT phase, but gave less good fits. Since the positions
and fractions have already been found by single crystal X-ray diffraction, and in
order to reduce the amount of refined parameters all positions and fractions were
101
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
kept constant. Only the isotropic atomic displacement parameters (Uisos), lattice
parameters and phase fractions were refined. Of course, the positions and fractions
are best to be found by single crystal X-ray diffraction, whereas lattice parameters
are easier to find with powder X-ray diffraction. The lineshapes of the reflections
were fitted with a pseudo-Voigt function in which the Gaussian and Lorentzian
broadening components were allowed to vary. The used set-up gives rise to
assymetric peak shapes at low diffraction angles and therefore the detector height
over diffractometer radius and sample height over diffractometer radius were fixed
with optimised values that were found by refining the lineshapes of the reflections
of a sample of LaB6.
The measured, calculated and difference diffraction profiles for 300 K and 125 K
in cooling mode (the two profiles at the extremes of the temperature range) can be
found in Figure 6.2. Table 6.1 gives selected details of the fit of the different
diffraction profiles. Although the values for wRp, Rp and reduced χ2 can be
considered quite good, the difference profiles indicate a relatively poor fit.
Therefore, the phase fractions should only be considered as an indication, the
lattice parameters can be considered quite good, since these only depend on the
reflection positions and not on the intensities.
The variation of the volume per formula unit (in cooling mode) can be found in
figure 6.3. The volume per formula unit decreases with decreasing temperature for
the HT phase and for the LT phase.
b)
Intensity (a.u.)
Intenisty (a.u.)
a)
5
10
15
20
25
2θ
30
35
40
45
5
10
15
20
25
30
35
40
45
2θ
Figure 6.2. Observed (+), calculated and difference (grey) diffraction profiles of powdered
single crystals (RbMn[Fe(CN)6]·H2O) at 300 K in cooling mode (a) and at 125 K (b). At
300 K only the cubic space group F-43m was used, at 125 K both the cubic space group
F-43m and the tetragonal space group I-4m2 were used. For details see text.
102
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
290
Figure 6.3. Variation of the volume per
formula unit of powdered single crystals
(RbMn[Fe(CN)6]·H2O) with temperature in
cooling mode for the HT phase (■) and the
LT phase (●) as deduced from the powder
diffraction profiles.
HT phase
LT phase
288
3
V (A )
286
262
260
258
100
150
200
250
300
T (K)
The variation of the phase percentages (in cooling and heating mode) with
temperature can be found in figure 6.4. At the beginning of the measurement the
percentage of HT phase in the diffraction profiles is 100% at room temperature.
When cooling down it decreases around 250 K and stabilises at ~50% around
200 K. When heating the sample, the percentage of HT phase increases around
275 K. Figure 6.4 shows the temperature dependent magnetic susceptibility as well.
On cooling, the percentage of HT phase decreases at the same temperature as
where the magnetic susceptibility decreases. In heating mode, the percentage of HT
phase increases at a somewhat higher temperature than the temperature where the
magnetic susceptibility increases. Unfortunately, the used machine had a maximum
temperature of 300 K, which is just below the temperature were the sample returns
completely to the HT phase (325 K) according to the magnetic susceptibility
measurements and therefore it was not possible to see if the HT structure is
completely recovered and at what temperature. The overall temperature variation
of the HT phase is in good agreement with the temperature variation of the
magnetic susceptibility.
5.00
fHT in cooling mode
4.75
fHT in heating mode
100
χ MT
80
4.25
60
4.00
40
3.75
20
3.50
100
150
200
T (K)
250
300
0
fHT (%)
4.50
120
3
-1
χMT (cm K mol )
In order to test the model proposed in chapter 5 another refinement was tried on the
profile at 125 K, a temperature in the LT phase well below the phase transition. In
the model it was assumed that only Fe and Mn sites with a completely cubic
environment brought about by the distribution of the Rb ions and water molecules
over the interstitial sites exhibit the electron transfer. If this is the case, in the
Figure 6.4. Variation of the percentage
of HT-phase with temperature of single
crystals (RbMn[Fe(CN)6]·H2O) in
cooling (▼) and heating (▲) mode, as
deduced from the refinement of the Xray powder diffraction profiles. Lines are
for guiding the eye. The variation of the
magnetic susceptibility with temperature
(grey squares) is included.10
103
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
compound that has partially switched to the LT phase two forms are present: (1)
the HT phase with symmetry F-43m where the Rb ions and water molecules are
equally distributed over the two interstitial sites and (2) the LT phase I-42m where
the Rb ions occupy one of the interstitial sites and the water molecules the other.
The results of this refinement are also shown in table 6.1 and the measured,
calculated and difference profiles can be found in figure 6.5. The differences
between figure 6.2b and figure 6.5 displaying the results obtained for either of the
models are minimal. This can also be concluded from the wRp, Rp and reduced χ2
values of the two different refinements which are also very comparable. This
indicates that it is not possible to distinguish between these two possible scenarios.
Intensity (a.u.)
Figure 6.5. Observed (+), calculated and
difference (grey) X-ray powder
diffraction profiles for
RbMn[Fe(CN)6]·H2O measured at 125 K
with Rb ions and water molecules in the
HT phase (space group F-43m) equally
distributed over the interstitial sites and
Rb ions distributed over one interstitial
site in the LT phase (space group I-4m2).
For details see text.
5
10
15
20
25
30
35
40
45
2θ
Table 6.1. Selected details of the refined X-ray powder diffraction profiles of powdered
single crystals (RbMn[Fe(CN)6]·H2O).
T (K)
aLT (Å)
cLT (Å)
300
275
250
225
200
175
150
125
7.099(2)
7.112(3)
7.111(2)
7.123(2)
7.117(2)
10.441(4)
10.412(8)
10.368(5)
10.273(6)
10.233(5)
150
175
200
225
250
275
300
7.117(2)
7.120(2)
7.116(3)
7.150(5)
7.114(4)
7.129(6)
7.129(3)
10.265(5)
10.204(3)
10.179(8)
10.04(1)
10.15(1)
10.13(2)
10.384(8)
125
7.119(2)
10.200(7)
104
aHT (Å)
HT phase (%)
wRp
Cooling mode
100
4.88%
100
5.54%
100
5.43%
56(1)
5.23%
49(1)
4.67%
47.5(7)
5.04%
45.7(7)
6.67%
42.8(6)
6.52%
Heating mode
10.488(2)
42.6(7)
7.91%
10.479(3)
42.8(8)
6.48%
10.481(2)
44.1(8)
8.35%
10.476(2)
44.0(8)
8.51%
10.474(2)
44.4(7)
7.32%
10.473(2)
46.7(6)
4.64%
10.474(2)
50.6(9)
4.56%
Application of model proposed in chapter 5
10.476(2)
45.5(7)
6.61%
10.501(1)
10.491(1)
10.492(1)
10.478(2)
10.480(2)
10.485(2)
10.488(2)
10.481(2)
Rp
Reduced χ2
(number of
variables)
3.16%
3.69%
3.56%
3.34%
3.05%
3.41%
4.09%
4.07%
2.733 (15)
3.487 (11)
3.372 (11)
3.088 (16)
2.479 (16)
2.860 (16)
5.038 (16)
4.811 (16)
4.59%
4.09%
4.47%
4.44%
4.09%
3.23%
3.07%
7.064 (16)
4.717 (16)
7.764 (16)
8.151 (16)
6.000 (16)
2.369 (16)
2.264 (16)
4.03%
4.947 (16)
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
6.3.2. 57Fe Mössbauer Spectroscopy
57
Fe Mössbauer spectra have been recorded in the temperature range 255 – 125 K
in cooling mode. Figure 6.6 shows the 57Fe Mössbauer spectra at 255 K and 125 K.
Each spectrum consists of one doublet and at least one singlet. The doublet can be
assigned to low spin FeIII. The singlets can be either assigned to low spin FeII or to
low spin FeIII in a perfect cubic environment. In analogy with the assignments
made in chapter 2, 4 and 5 for these sites, at temperatures above the phase
transition (255 K) the singlet is assigned as low spin FeIII in a fairly cubic
environment. At temperatures below the phase transition (150 K, 125 K) the singlet
is assigned as low spin FeII. At 200 K the system is still in its phase transition, and
here two singlets are needed for the fitting of the spectrum: one for low spin FeIII
and one for low spin FeII.
The least-squares fitted Mössbauer data can be found in Table 6.2. In all cases the
chemical isomer shift, widths, quadrupole splitting and area fractions were allowed
to vary, except for the spectrum where three sites were present. Here, the chemical
isomer shift of the singlet sites were kept constant11 in order to increase the
stability of the refinement.
Figure 6.7 shows the variation of the area fractions of the three sites with
temperature. The least-squares fitted Mössbauer data can be found in Table 6.2.
The area fraction of the doublet stays more or less constant around 50% throughout
the temperature region, indicating that ~50% of the Fe sites does not show an
electron transfer upon cooling. More specifically, the low spin FeIII with a slight
Figure 6.6. 57Fe Mössbauer spectra of
single crystals of RbMn[Fe(CN)6]·H2O
at 255 K (a) and 125 K (b). Both spectra
were recorded in the cooling mode.
100
98
Absorption (%)
96
94
92
125 K
100
98
96
255 K
-4
-2
0
2
4
v (mm/s)
105
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
deviation from a perfectly cubic environment does not show electron transfer,
thereby confirming the proposed model from chapter 5. The area fraction of the
singlet for low spin FeII increases with decreasing temperature, while at the same
time the area fraction of the singlet for low spin FeIII decreases. This illustrates the
partial electron transfer from high spin MnII to low spin FeIII when cooling down.
Figure 6.7. Variation of the area
fractions of the three Fe sites with
temperature in cooling mode of single
crystals (RbMn[Fe(CN)6]·H2O).
Explanation in the text. Lines are for
guiding the eye.
III
singlet LS Fe
II
singlet Fe
III
doublet LS Fe
100
90
80
Area (%)
70
60
50
40
30
20
10
0
0
50
100
150
200
250
300
T (K)
Table 6.2. Least-squares fitted Mössbauer data for single crystals (RbMn[Fe(CN)6]·H2O)
obtained in the cooling mode.
T
(K)
293*
255
200
150
125
50*
IS
(mm
s-1)
-0.150
(2)
-0.1444
(6)
-0.117
(8)
-0.087
(9)
-0.084
(10)
-0.067
(5)
Doublet
QS
Γ/2
(mm
(mm
s-1)
s-1)
0.43
0.145
(1)
(7)
0.46
0.16
(5)
(3)
0.55
0.18
(5)
(3)
0.64
0.18
(4)
(3)
0.71
0.17
(3)
(3)
0.78
0.20
(2)
(1)
Area
(%)
57(7)
62(30)
59(22)
45(15)
34(12)
49(6)
Singlet 1
Γ/2
(mm
s-1)
0.16
(2)
0.17
(8)
0.18
-0.1126
(22)
IS
(mm
s-1)
-0.138
(2)
-0.134
(7)
Area
(%)
IS
(mm
s-1)
Singlet 2
Γ/2
(mm
s-1)
Area
(%)
43(7)
38(32)
15(29)
-0.0858
-0.062
(6)
-0.056
(5)
-0.033
(4)
0.190
(99)
0.21
(4)
0.24
(4)
0.23
(2)
27(22)
55(16)
66(13)
51(6)
Standard deviations are indicated in parentheses. Values without parentheses were kept constant
during the fitting procedure.
* Taken from chapter 5.
6.3.3. X-Ray Photoemission Spectroscopy (XPS)
XPS is a direct method to identify the surface composition of a compound as well
as the oxidation state of the various elements it contains. Like previous studies
have proved,12 this technique is well suited to trace a phase transition as a function
of temperature. The charge transfer between the metallic ions leads to a change in
the oxidations states, which can be quantified by XPS.
106
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
XPS spectra have been recorded between 350 K and 130 K both in cooling and
heating mode. The spectra are shown in Figure 6.8. To calculate the conversion
yield of the compound during the phase transition, the attention has been focused
on the Fe 2p photoemission line, because of its high photoemission cross section
and the spin configuration of the Fe shell. The recorded spectra of the Fe
photoemission line are straightforward to analyse and relatively clear compared to
the Mn photoemission line (which, due to its high spin state, presents a
photoemission peak containing many components13). Figure 6.8 shows the Fe 2p3/2
core level photoemission spectrum (fits and raw data) of a single spot on a sample
of single crystals prepared with method (1) (see experimental details) for various
temperatures. The cooling and warming rate between the measurements at different
temperatures was in the order of 2 K min-1. Similarly to the observation made in
previous work on a powdered sample14 the Fe 2p3/2 signal consists of three distinct
features: the FeII line at 708.5eV binding energy, the FeIII line at 710.0 eV and the
FeIII satellite at 711.4eV. Investigation of the surface FeII/FeIII ratio on a different
spot of the same sample (not shown) shows that it is not uniform throughout the
sample (see table 6.3).
By comparing the area of the peaks after background subtraction one can deduce
that the sample initially (at 350 K) consists of 79% FeIII and 21% FeII.15 Upon
cooling, a decrease in the FeIII peak intensity is observed, whereas the FeII peak
intensity increases. Those changes can be explained by the charge transfer between
FeIII and MnII ions which is described as LS FeIII; HS MnII Æ LS FeII; HS MnIII.
Fe2p3/2
350 K (a)
Intensity (a.u.)
130 K
Figure 6.8. Fe 2p3/2 core level
photoemission spectra of powdered
crystals of RbMn[Fe(CN)6]·H2O
collected at 350 K, 140 K, 200 K and
350 K. The fits to the raw data are also
plotted. (a) refers to the 350 K spectrum
taken before starting the cooling cycle,
whereas (b) refers to the measurement
done after warming up to 350 K.
200 K
350 K (b)
714
712
710
708
706
Binding energy (eV)
107
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
Table 6.3. Variation of the FeII/FeIII ratio at room temperature as measured by XPS
spectroscopy on single crystals prepared with procedure (2) as described in the
experimental details.
Entry
% FeII
% FeIII
1
46.8
53.2
2
47.0
53.0
3
56.4
43.6
4
54.9
45.1
5
60.1
39.9
6
41.1
58.9
7
46.7
53.3
Average
50.5
49.5
Indeed, as known from the magnetic susceptibility measurements at a temperature
of 350 K the sample is in the HT configuration whereas at 140 K the sample is in
the LT configuration. The low temperature measurements reveal a configuration
composed of 47% FeIII and 53% FeII at 140 K in the cooling mode and 45% FeIII
and 55% FeII at 200 K in the heating mode.16 When the temperature returns back to
350 K the XPS spectrum shows 71% FeIII and 29% FeII. These values correspond
nicely to the ratios found for the spectrum at 350 K in the cooling mode, indicating
that the sample has completely returned to its original configuration at this
temperature.
The transition percentage at the surface of the material is about 33% which is lower
than the 50% found by the magnetic susceptibility measurements, powder
diffraction data and 57Fe Mössbauer spectroscopy. One first possible reason for this
mismatch is the fact that XPS spectroscopy is a surface technique and the other
techniques mentioned are all bulk techniques. When looking at the literature, this
mismatch in surface and bulk techniques is seen more often.12,14,17 A further
explanation can be given when focusing on the spectra recorded at room
temperature for a sample prepared following procedure (2) as described in the
experimental section. In that particular case the 600 µm X-ray spot allowed us to
measure single crystals individually. The data collected on various crystals, show
at room temperature a surface composition ranging from as low as 41.1% FeII and
58.9% FeIII up to as much as 60.1% and FeII 39.9% FeIII. The average ratio is
50.5% FeII 49.5% FeIII (see table 6.3). It turns out, from measurements performed
on one single crystal at different angles between the surface and the electron beam,
that the ratio of FeIII/FeII depends on the precise angle between the surface of the
crystal and the electron beam.18 This observation could in part explain the large
differences between the found ratios.
6.3.4. Electron Spin Resonance (ESR) Spectroscopy
ESR studies have been done on sample 3.3 (Rb0.91Mn[Fe(CN)6]0.97·1.53H2O) and
on single crystals of sample 5.1 (RbMn[Fe(CN)6]·H2O) at excitation frequencies of
108
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Figure 6.9. Electron spin
resonance at 222.4 GHz at room
temperature of sample 3.3
(Rb0.91Mn[Fe(CN)6]0.97·1.53H2O).
Ref.
.
RbMn[Fe(CN)6] H2O defects
7.85
7.90
0
7.95
2
8.00
8.05
4
6
8
Magnetic Field(T)
9.4 and 222.4 GHz (g = 2 resonance at 0.33 T and 7.95 T.) Figure 6.9 shows an
ESR spectrum of sample 3.3 at 222.4 GHz and room temperature. Figures 6.10 and
6.11 show the ESR spectra of sample 5.1 at 9.4 GHz and 222.4 GHz at various
temperatures. It can be seen that the ESR spectra of both the powder and the single
crystals show no strong signal arising from the bulk of the material. The ESR of
both samples arises from MnII and possibly FeIII ions at structural defects weakly
interacting with the magnetic ions of the bulk of the structure. This is inferred from
the small ESR intensity and the weak sensitivity of the ESR spectrum to the HT to
LT phase transition.
The experimental results on sample 3.3 are in good agreement with a recent multi
frequency ESR study of a deuterated powder (Rb0.81Mn1.1[Fe(CN)6]·D2O) by
Pregelj et al.19 These authors showed that the ESR arises from defect sites which
are weakly affected by the HT – LT phase transition between 150 and 300 K and
Figure 6.10. Orientation dependence of the
ESR spectra of a single crystal
(RbMn[Fe(CN)6]·H2O) at 9.4 GHz and 5K.
The peak at highest fields shifts about 10
mT between 0 and 45°. Other components
of the spectrum are, however, little
dependent on orientation.
90°
75°
60°
45°
30°
15°
0°
0.0
0.1
0.2
2.079
2.022
0.3
0.4
0.5
0.6
Magnetic Field (T)
109
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
(a)
(b)
single crystal
9 GHz
powder
g=2.022
65K
ESR absorption derivative
45K
296 K
15K
9K
151 K
5K
2.079
110 K
0.2
0.3
0.4
(c)
150 K
300 K
7.8
7.9
8.0
Magnetic field (T)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Magnetic Field (T)
Magnetic field (T)
8.1
8.2
Figure 6.11.(a) ESR spectra of sample 5.1
(RbMn[Fe(CN)6]·H2O at 9.4 GHz measured
in cooling. The magnetic field is along the
same undetermined direction at the different
temperatures. The vertical line corresponds
to the position of the 9.4 GHz ESR in a
powdered sample (g = 2.022). (Spectra
corrected for background impurity lines not
originating from the sample). (b) ESR
spectra of sample 5.1
(RbMn[Fe(CN)6]·H2O) at 9.4 GHz broaden
and grow in intensity faster than 1/T below
the ferromagnetic transition at 12 K. The
complex temperature dependent spectra of
sample 5.1 differ significantly from the
spectra sample 3.3. (Uncorrected spectra).
(c) ESR spectra at 222.4 GHz at 300 K and
150 K of sample 5.1 (RbMn[Fe(CN)6]·H2O.
Note the large anomalous frequency
dependence between 9.4 and 222.4 GHz.
The spectra are similar to the spectra of the
powdered sample at 222.4 GHz.
the ferromagnetic ordering at 12 K of the bulk. Surprisingly for a compound
consisting of HS MnII (S = 5/2) and LS FeIII (S = 1/2) ions, in the HT configuration
no ESR of the bulk is observed. A lower limit of 1 T was found for the HT bulk
line width which corresponds to an unusually fast spin relaxation rate.
110
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
6.3.4.1 The Origin of the Electron Spin Resonance Lines
The possible reasons why the ESR of the bulk material was not observed is first
discussed. The lack of a spin resonance in the LT phase is not surprising: FeII is not
magnetic and MnIII has an S = 2 spin. In MnIII the shell is not half filled and spin
relaxation related to orbital effects broaden the ESR beyond observability. On the other
hand, the lack of any ESR above the phase transition is not easily explained. Crystals
with MnII (S = 5/2) and FeIII (S = 1/2) spin state ions have narrow ESR lines with
gyromagnetic factors near g = 2.20 Yet, if there is a bulk ESR resonance of the material,
it must be extremely broad, at least two orders of magnitude broader than the ESR of
the MnII defects in the same system. Since the line width must be more than 1 T, at
9 GHz the life time broadening is more than the Larmor frequency and there is no ESR
of the bulk at all. It is suggested that in the HT configuration there is a small mixture
with the S = 2 and S = 0 states into the pure S = 5/2 and 1/2 states of the MnII and FeIII
ions. Above the phase transition, charge fluctuations give rise to a fast spin relaxation
and a very broad or no ESR. A weak exchange interaction between Mn and Fe with
well defined magnetic states does not broaden the ESR. Although the g factors are
slightly different, the two magnetic species in pure HT states would have a common
exchange narrowed ESR line. Fine splitting from crystal fields is relatively small for
the half filled 3d5 shell of MnII which is one of the most common ESR probes. There is
no zero field splitting for S = 1/2 FeIII ions either, the ESR of this ion is not strongly
anisotropic and has been frequently observed. Crystal field anisotropies (fine structure
splitting) and the dipolar interactions are ineffective in magnetically dense systems like
Prussian Blue analogues. If the exchange interactions between Mn and Fe ions are
larger than the dipolar and single ion crystal field energies then the ESR is „exchange
narrowed” i.e. there is a common narrow ESR resonance of the bulk. Thus we suggest
that the fact that a bulk ESR resonance for HT RbxMn[Fe(CN)6]y·zH2O is not observed
is due to unusually fast spin relaxation.
Fe(CN)6 defects of the type described in chapter 2 are the likely source of the ESR signal
observed in powders and in the single crystals at high frequency. These defects are
Fe(CN)6 vacancies replaced by 6 H2O molecules. The 6 Mn ions (probably together with
8 first neighbour FeIII ions) surrounding the defect form a well isolated magnetic cluster
observed by ESR. The ions are in a MnII state at all temperatures and have a relatively
narrow ESR, which is little influenced by the phase transition of the bulk. Coupling of
Mn first neighbours within the cluster is via the intermediate Fe ions. It is assumed that
there is a ferromagnetic exchange coupling of the order of 10 K between first neighbour
Mn ions within the cluster.21 This cluster of 6 weakly interacting MnII on a cube
surrounding a defect explains qualitatively the characteristics of the ESR spectra:
i.) the nonlinear variation of the line position with frequency, i.e. that the g factor
varies from g = 2.022 at ωL/2π = 9.4 GHz to 2.000 at 222.4 GHz
ii.) the line width, ∆H(ωL), in the powder is nearly independent of frequency, ωL at
temperatures well above the ferromagnetic transition. A minimum in ∆H(ωL) was
observed by Pregelj et al..19
111
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
The 6 Mn ions have one of the Fe first neighbours replaced by H2O, thus the
individual Mn ions are in anisotropic environments. Such ions, if isolated from
each other, would have a strongly anisotropic ESR due to crystal field “fine
structure” splitting and g factor anisotropy. In the cubic arrangement of the 6
interacting MnII ions, the g factor anisotropy of the single ions, ∆g, and the fine
structure are narrowed due to motions in the lattice. A small exchange energy, J, is
sufficient to narrow the line. J > ∆g/g ħωL and J > ħωD are the conditions for
motional narrowing where ωD and ∆g are characteristic values of the fine structure
splitting and g factor anisotropy. On the other hand, at low ESR frequencies the
crystal field shifts the average line position. The nonlinear shift is proportional to
ωD2/ωL, in second order. The non linear shift is nearly isotropic for cubic clusters
and will appear as anomalously large g shifts at low frequencies. The small
frequency dependence of the line width with a shallow minimum at 52 GHz
observed by Pregelj et al19 arises from the isotropic shift of the motionally
narrowed fine structure at low frequencies and a small g factor anisotropy due to
incomplete motional narrowing of the g factor anisotropy of individual Mn ions of
the cluster.
6.3.4.2 Electron Spin Resonance in Sample 3.3
In the ESR spectra of sample 3.3 a single strong ESR line is observed at all
temperatures between 4 and 320 K. The intensity of this ESR line is proportional to
the spin susceptibility of the ESR active species and was measured in a powdered
sample at 9 GHz using a CuSO4·5H2O reference. At ambient temperatures the
intensity is much less than expected for a resonance arising from one HS MnII
(a)
(b)
0.0225
0.0220
0.032
Heating
Cooling
0.0210
0.0205
0.028
0.0200
Line width (T)
Line width (T)
0.0215
0.0195
0.0190
0.0185
0.0180
0.024
0.020
0.0175
0.0170
0.0165
0.016
150
200
250
T (K)
300
0
50
100
150
200
250
300
350
T (K)
Figure 6.12. (a) Temperature dependence of the ESR line width in spectra of a single
crystal of sample 5.1 (RbMn[Fe(CN)6]·H2O) at 222.4 GHz. (b). Thermal hysteresis of ESR
line width in sample 3.3 (Rb0.91Mn[Fe(CN)6]0.97·1.53H2O) measured at 9.4 GHz.
112
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
paramagnetic ion per formula unit. The ESR is little affected by the HT – LT
transition. A small hysteretic increase is observed between 150 and 300 K but the
intensity follows a Curie law within experimental accuracy (Figure 6.12b) between
50 and 300 K. The defect line is slightly broadened by the HT to LT phase
transition in cooling between 200 and 150 K. The line width remains broad in the
heating cycle showing that the compound remains in the LT configuration up to
high temperatures. The strong line width increase at low temperatures indicates the
transition of the paramagnetic to the ferromagnetic phase. The intensity of the
spectrum of the powdered sample corresponds to a concentration of MnII ions
surrounding a Fe(CN)6 defect of about 2.5% (comparable to the value of [Mn] –
[Fe] in the compound). No broad large intensity line of the bulk magnetic ions was
found. As shown in Figure 6.9, at 222.4 GHz no resonance is observed within a
range of 0 to 9 T apart from the narrow (about 20 mT line width) resonance arising
from the 2.5% MnII ions surrounding Fe(CN)6 defects. Thus the ESR of the bulk is
presumably very broad: a 1 T broad resonance with the intensity of the bulk MnII
ions would have been easily detected.
Above 50 K the spin susceptibility of defects is proportional to the inverse
temperature. Below 50 K this ESR increases faster than the Curie susceptibility and
at 5 K it is several times larger than expected for a constant concentration of nearly
free magnetic defects. The line width at 9 GHz ESR frequencies first decreases
slightly as the temperature decreases from 50 K and then increases rapidly as the
sample is cooled below 12 K, the ferromagnetic ordering temperature of the bulk.
At 222.4 GHz the line width increase is observed already at 25 K, i.e. at much
higher temperatures than for the low field ESR measurement.
6.3.4.3 Electron Spin Resonance in Sample 5.1
In the 9 GHz ESR spectrum of sample 5.1 at least three ESR active defect sites
with different line widths and resonance fields are resolved at ambient
temperatures. The line positions and widths are strongly temperature dependent
(Figure 6.11a and b). At low temperatures (below 110 K) the most intense line is at
about g = 2.079. On the other hand, at 222.4 GHz only a single line at about
g = 2.00, with a line width similar to the line in the spectra of sample 3.3, was
observed (Figure 6.11c).
The temperature dependence of the ESR line width (Figure 6.12a) has only a small
anomaly in the range of the phase transition between the HT (FeIII (S=1/2), MnII
(S = 5/2)) and the LT (FeII (S = 0), MnIII (S = 2)) phases.22 A small thermal
hysteresis of the line width appears in the range of 150 to 300 K. The sample was
annealed at 320 K before the thermal cycling experiment. The line width is slightly
narrower in the HT configuration then in the LT configuration. Since magnetisation
measurements on similar powder samples show the phase transition clearly it was
concluded that the magnetic ions are observed (most probably MnII) at defects for
113
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
which the phase transition does not take place but is weakly affected by the
transition at neighbouring sites.
In the single crystals a complicated spectrum composed of several lines is observed
at low frequencies. This may arise from defects with lower symmetries together
with the site attributed to clusters of 6 MnII ions surrounding a Fe(CN)6 defect. At
high frequencies the ESR corresponds to the latter cubic site only. The other defect
sites observed at lower frequency only may have lower symmetry.
6.3.4.4 Ferromagnetic Ordering Below 12 K
The ESR intensity increases faster than 1/T as the ferromagnetic ordering
temperature is approached from above (figure 6.11c). This is due to a
superparamagnetic behaviour at low temperatures of clusters essentially isolated
from the bulk. In this model exchange interactions within a cluster are of the order
of J = 10 K and exchange interactions with the bulk are much weaker than this.
Above 50 K the spin susceptibility of defects is to a good approximation equal to
that of non-interacting Mn and Fe ions in the HT phase. At temperatures below J/k
the defect clusters are superparamagnetic. In a magnetic field the magnetisation of
the cluster saturates at much higher temperatures than for non-interacting ions. The
disordered demagnetising fields at the ferromagnetic transition of the bulk broaden
the defect ESR. Exchange interactions between defect clusters and the bulk are
very weak and the common ferromagnetic resonance of the bulk and the defects is
not observed down to 5 K.
The temperature dependence of the ESR intensity is in agreement with a low
temperature superparamagnetic behaviour. Deviations from the Curie law in the
ESR intensity increase gradually: they become apparent below 50 K, i.e. well
above Tc and defect site magnetism becomes fully ordered only at temperatures
below 5 K. On the other hand the ferromagnetic ordering temperature of the bulk is
well defined, in low fields the ESR line width increase sets in abruptly at
Tc = 12 K.
As stressed already in the previous discussion, the ESR active sites are well
isolated, a significant coupling to the bulk would result in a fast spin relaxation and
a large line broadening. Spin relaxation in the bulk is extremely fast in the HT
configuration and is without doubt even faster in the LT configuration, since HS
MnIII is well known to show rapid relaxation and LS FeII does not show any ESR.
On the other hand, the HT to LT transition is accompanied by a very small line
width change of 2 mT only.
The conditions for the observation of the ferromagnetic resonance (FMR) of the
bulk are much less stringent than for the paramagnetic resonance. Nevertheless the
FMR was not observed. The ESR intensity variation is in agreement with the
susceptibility of clusters at defects becoming superparamagnetic at low
temperatures. Below Tc the modest intensity increase is easily understood for a
114
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
cluster of interacting Mn ions while it is much less than the intensity increase
expected for the FMR of the bulk. The line width of defects is increased by only
10 mT below the ferromagnetic ordering temperature. This corresponds to a very
small coupling, and even this small broadening can be from long range dipolar
interactions, i.e. inhomogeneous demagnetizing fields.
6.3.5. Thermal Conductivity Measurements
Figure 6.13 shows the thermal conductivity of RbMn[Fe(CN)6]·H2O as a function
of temperature. Starting from low temperatures up to 150 K the thermal
conductivity κ monotonically increases with rising temperature and has extremely
small absolute values. Such a behaviour is typically found for the “lattice” thermal
conductivity in amorphous substances like glasses, even though the current
material is clearly crystalline. The notion of κ being hampered by the amorphous
character of the material (from a conductivity point of view) is corroborated by the
comparison of our data with the theoretical minimum of the thermal conductivity
κmin. The theoretical minimum has been calculated with the use of a model of
Einstein by Cahill et al.23 and it was confirmed by measurements on a crystal with
glasslike conductivity. This conductivity can be regarded a lower limit of κ of
insulating amorphous materials. Apparently, the measured κ is only slightly larger
than κmin up to 150 K. Above 150 K a hysteretic steep increase of κ by a factor of
~2 is observed which is in correspondence with measurements of the magnetic
susceptibility. This anomaly is related to the temperature driven switching of MnFe
pairs between a HT and a LT configuration. The difference in thermal
conductivities between both configurations indicates an increased scattering of
phonons for the LT configuration and thus a higher degree of disorder in this
configuration as compared to the HT configuration. Note that it was not possible to
acquire data while heating between 260 K and 300 K because of an increasing
strain on the crystal that lead to some cracks. A field of up to 14 T did not show
any influence on the thermal conductivity and hence the switching behaviour.
1.2
Figure 6.13. Thermal conductivity of a
single crystal of RbMn[Fe(CN)6]·H2O
and theoretical minimum thermal
conductivity κmin
0.8
heating
-1
-1
κ (WK m )
cooling
0.4
κmin
0.0
0
100
200
300
T (K)
115
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
6.4 Conclusions
In this chapter the single crystals that have been obtained in chapter 5 have been
investigated in more detail with the use of temperature dependent X-ray powder
diffraction, 57Fe Mössbauer, XPS and ESR spectroscopy and thermal conductivity
measurements. All techniques show differences between the high temperature and
low temperature configurations. The temperatures at which the two configurations
switch are in agreement with the magnetic measurements from chapter 5 in all
techniques (i.e. T1/2↓ = 237 K and T1/2↑ = 292 K).
From the previous chapter it was already known that the crystals show a switching
behaviour of about 50% upon cooling. From the X-ray powder diffraction
measurements in this chapter it becomes clear that indeed roughly 50% of the room
temperature structure switches to a new structure with I-4m2 symmetry.
Unfortunately, it was not possible to determine the distribution of the Rb ions and
water molecules over the interstitial sites in either the HT phase structure or the LT
phase structure with X-ray powder diffraction. From the 57Fe Mössbauer
spectroscopy measurements however, it becomes clear that two different low spin
FeIII sites are present in a 1:1 ratio at room temperature. One of them is in a
perfectly cubic environment and the other has a deviation from this cubic
environment. Only the first site changes into low spin FeII upon cooling, thereby
indicating that the iron sites (and most probably the Mn sites too) need a perfect
cubic environment to be able to switch. This is in agreement with the findings from
chapter 4. The differences between the two environments for the Fe and Mn ions
could indeed be caused by the specific distribution of the Rb ions and water
molecules over de the interstitial sites.
Due to the availability of single crystals it was now possible for the first time to perform
thermal conductivity measurements on compounds of RbxMn[Fe(CN)6]y.zH2O. These
measurements on the low temperature phase show that the conductivity is only
slightly higher than that of amorphous glasses. This indicates that the LT phase has
a large disorder, whereas the conductivity of the HT phase is much larger and
therefore the disorder in the HT phase is considerably less. These findings are in
agreement with the fact that it was possible to determine the HT structure, but not
the LT structure with single crystal X-ray diffraction. The disorder in the LT phase
could be caused by the various Fe sites and the superposition of the two structures.
Furthermore, ESR measurements indicate that, besides the defects of the type
described in chapter 2 which are present both in powders and single crystals, the
single crystals show defects in the LT phase, at low frequencies which are not
present in the powders.
The XPS spectroscopy measurements also indicate a switching between the two
configurations, but since this technique is a surface technique rather than a bulk
technique the absolute values of the found FeIII/FeII ratios are not in agreement with
the ratios found by bulk techniques. This indicates that at the surface of the crystals
116
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
a different switching behaviour takes place than in the bulk of the crystals. The
XPS spectra have been vital though, in showing the complete returning to the
original state after a complete cycle of cooling and heating has taken place.
The precise structure of the low temperature configuration in the single crystals is
still unclear. Although highly likely, the proposed structure in chapter 5 remains a
proposal. Possibly with the aim of different measurement techniques the question
about the low temperature structure could be answered.
6.5 Acknowledgements
The XPS spectra in this chapter have been measured by Régis Gengler, the ESR
spectra by Agnes Antal and the thermal conductivity measurements have been
done by Nikolai Hlubek.
6.6 References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
This chapter was based on the following publication: Vertelman, E.J.M.; Gengler, R.; Antal, A.;
Hlubek, N.; Molnar, G.; Hess, C.; Bousseksou, A.; Rudolf, P.; Jánossy, A.; Büchner, B.;
Koningsbruggen, P.J. van; 2009, In preparation
Chapter 2 and 3
Larson, A. C.; Von Dreele, R. B. General Structure Analysis System (GSAS) 2004, Los Alamos
National Laboratory Report LAUR 86-748.
a) Cataldi, T. R. I.; De Benedetto, G. E.; Bianchini, A., J.Electroanalytical Chem., 1998, 448,
111; b) Yatsimirskii, K. B.; Nemoshkalenko, V. V.; Nazarenko, Y.,
J.Electron.Spectrosc.Relat.Phenom., 1977, 10, 239
Tougaard, S.; J. Vac. Sci. Tech. A 2005, 23(4), 741
Hess, C.; Baumann, C.; Ammerahl, U.; Büchner, B.; Heidrich-Meisner, F.; Brenig, W.;
Revcolevschi, A.; Phys. Rev. B 2001, 64, 184305
Moritomo, Y.; Kato, K.; Kuriki, A.; Takata, M.; Sakata, M.; Tokoro, H.; Ohkoshi, S.-I.;
Hashimoto, K., J.Phys.Soc.Jpn., 2002, 71, 2078
Chapter 5
Tokoro, H.; Matsuda, T.; Nuida, T.; Moritomo, Y. ; Ohoyama, K. ; Loutete Dangui, E.D. ;
Boukheddaden, K.; Ohkoshi, S.-I.; Chem. Mater. 2008, 20, 423
The variation of the magnetic susceptibility with temperature that is shown here is for nonpowdered single crystals. The magnetic susceptibility is also measured for powdered single
crystals, but only at the extreme temperatures (330 K and 125 K). The measured values were
similar as for the crystals at these temperatures.
Obtained by extrapolating the chemical isomer shifts for these sites obtained in chapter 4 for
sample 3.13.
Lummen, T.T.A.; Gengler, R.Y.N.; Rudolf, P.; Lusitani, F.; Vertelman, E.J.M.;
Koningsbruggen, P.J. van; Knupfer, M.; Molodtsova, O.; Pireaux, J.-J.; Loosdrecht, P.H.M. van;
J. Phys. Chem. C 2008 112, 14158
Gupta, R.P.; Sen, S.K.; Phys. Rev. B 1975, 12(1), 15
Luzon, J.; Castro, M.; Vertelman,E.J.M.; Gengler, R.Y.N.; Koningsbruggen, P.J. van;
Molodtsova, O.; Knupfer, M.; Rudolf, R.; Loosdrecht, P.H.M. van; Broer, R.; J. Phys. Chem. A
2008, 112, 5742
The standard deviation in the determined FeIII/FeII ratio is ~1% at temperatures close to room
temperature and increases to ~3% when the temperature has decreased to 140 K
117
Chapter 6. Further Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O
16
17
18
19
20
21
22
23
118
The absolute resolution of the instrument was better at 200 K compare to 140 K due to less
differential charging effect (which is responsible of the peak broadening), therefore the peaks are
better resolved at 200 K.
Chapter 2
When one single crystal was measured in one position with respect to the electron beam its
spectrum showed 41% Fe(II). When subsequently measuring at a different (undetermined) angle
its spectrum showed 48% Fe(II).
Pregelj, M.; Zorko, A.; Arcon, D.; Margadonna, S.; Prassides, K.; Tol, H. Van; Brunel, L.C.;
Ozarowski, A.; J. Magnetism and Magnetic Materials 2007, 316, e680
Abragam, A.; Bleaney, B.; Electron Paramagnetic Resonance of Transition Ions Clarendon
Press Oxford 1970 p.440
If Fe is in the S = 1/2 FeIII state then the coupling between Mn is ferromagnetic independently of
the sign of the Mn Fe coupling.
The line width is defined as the full width at half height of the best fitting Lorentzian curve.
Cahill, D.G.; Watson, S.K.; Pohl, R.O.; Phys. Rev. B 1992, 46, 6131
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Chapter 7. Attempted Synthesis, Characterisation and
Physical Properties of AxMn[Fe(CN)6]y.zH2O (A = Na+,
K+, Cs+, NH4+, N(CH3)4+, Sr2+)
7.1 Introduction
In general, compounds of the series AxMn[Fe(CN)6]y.zH2O with A = cation, have
mostly been prepared with A = Rb+ 1 and in addition few samples are known with
A = Cs+.2 The RbxMn[Fe(CN)6]y·zH2O compounds show a structural phase transformation from cubic to tetragonal upon cooling associated with charge transfer
from LS FeIII and HS MnII at room temperature to LS FeII and HS MnIII at lower
temperatures. This process has a hysteresis width which depends on the amount of
Fe(CN)6 defects present: a relatively small hysteresis of 52 K was found for
Rb0.97Mn[Fe(CN)6]0.98.1.03H2O,3 whereas the broadest hysteresis of 116 K was
found for Rb0.73Mn[Fe(CN)6]0.91.1.4H2O.4
In the case of Cs+ only four compounds have previously been reported in
literature.2 These compounds are: Cs1.78MnII[FeII(CN)6]0.78[FeIII(CN)6]0.22,
Cs1.57MnII[FeII(CN)6]0.57[FeIII(CN)6]0.43, Cs1.51MnII[FeII(CN)6]0.51[FeIII(CN)6]0.49and
Cs0.94MnII[FeII(CN)6]0.21[FeIII(CN)6]0.70·0.8H2O. All samples have at room
temperature a mixture of both FeII(CN)6 and FeIII(CN)6 present, of which the first
always remains in that valency on cooling. The first three samples show charge
transfer from MnII and FeIII(CN)6 to MnIII and FeII(CN)6 at transition temperatures
of (T1/2↑, T1/2↓) (207 K, 225 K), (190 K, 231 K) and (175 K, 233 K), respectively. It
seems as if when too little FeII(CN)6 is present no electron transfer takes place,
thereby indicating that its presence at room temperature is a prerequisite for the
system to undergo electron transfer. The authors have not investigated this
observation in detail.
In a comparable system, AxCo[Fe(CN)6]y.zH2O, the LT phase consists of LS CoIII
and LS FeII whereas the HT phase is comprised of HS CoII and LS FeIII.5,6 For this
system all different alkali cations (except Fr+) have been incorporated and in all
these cases it is possible to have electron transfer. It is therefore surprising that in
the case of AxMn[Fe(CN)6]y.zH2O the only reported examples (with electron
transfer) are with either Rb+ or with Cs+.
Table 7.1. (Pauling) radii of the cations used in this study
Cation
Na+
Sr2+
K+
NH4+
Rb+
Cs+
N(CH3)4+
(Pauling) radius (pm)
98 7
110 7
133 7
142 7
148 7
167 7
307 8
119
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
In this chapter it was tried to incorporate cations other than Rb+ into the
AxMn[Fe(CN)6]y.zH2O system and to determine their influence on the physical
properties of the material. The selected cations are Na+, K+, Cs+, NH4+, N(CH3)4+
and Sr2+. The most striking difference between the cations is their size (see table
7.1). Most used cations are monovalent, thereby differing with Rb+ mostly in size
only. One cation (Sr2+) is divalent, which might give rise to other properties, such
as a different crystal packing or the distribution of the cation in the crystal lattice.
NH4+ has additional features in that it has also the possibility to form hydrogen
bonds with water or can act as a Brønsted acid.
7.2 Experimental Section
7.2.1 Synthesis
All chemicals were purchased at Sigma Aldrich and used without further
purification.
In the synthesis of all samples an aqueous solution of 25 mL with 0.495 g of
MnCl2.4H2O (0.1 M, room temperature, solution A) was added to an aqueous
solution of 25 mL with 0.823 g of K3Fe(CN)6 (0.1 M, 43oC, solution B). The
(alkali) cation was incorporated either only in solution B or in both solutions (see
table 7.2).
The addition speed was kept constant at 6 mL h-1 with a syringe pump model 352
of Sage Instruments. The resulting solution and solution B was stirred with a
magnetical stirrer at 5.5 rps and kept at 43oC. In all cases a brown powder
precipitated instantaneously, which was centrifuged and washed twice with H2O of
room temperature. The powder was then allowed to dry in vacuum for 1 night.
Table 7.2. Details of the synthesis of the various samples
Sample
Cation source
7.1
NaCl
7.2
KCl
7.3
CsCl
-
7.4
NH4Cl
-
7.5
N(CH3)4Cl
-
7.6
SrCl2.6H2O
-
a
Amount cation in A
8.7 g, ~6 M,
Saturateda
7.4 g, ~4 M,
Saturateda
Amount
cation in B
8.7 g, ~6 M,
Saturateda
7.4 g, ~4 M,
Saturateda
2.104 g,
0.50 M
6.7 g, ~5 M,
Saturateda
2.740 g,
1.0 M
3.333 g,
0.50 M
Cation/Fe ratiob
60
40
5
50
10
5
The given amount is only an estimate, since not all material actually dissolved
b
The ratio is calculated for solution B only
7.2.2 Elemental Analysis
For the details of the elemental analysis of K/Na/Cs/Sr, Mn, Fe, C, H and N see
section 2.2.2
120
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
7.2.3 FTIR Spectroscopy
For details of the FTIR spectroscopy see section 2.2.4.
7.2.4 X-Ray Powder Diffraction
For details of the X-Ray diffraction measurements see section 2.2.3.
7.2.5 Magnetic Measurements
For details of the magnetic measurements see section 2.2.7.
7.3 Results and Discussion
7.3.1 Elemental Analysis
The results of the elemental analysis can be found in table 7.3. The correlation
between the Fe fraction and the cation fraction is depicted in figure 7.1a. The
dashed line in figure 7.1a is based on charge neutrality of the material assuming
A+, CN- and {MnII, FeIII} or {MnIII, FeII}. When comparing the found
stoichiometries with the prediction given by this line, in the case of the samples
with monovalent cations the trend for charge neutrality (Figure 7.1a) is similar as
for the Rb ions, i.e. the points are close to the dashed line. For sample 7.5 this
agreement with the charge neutrality line might be a coincidence, since X-ray
powder diffraction studies (see section 7.3.3) shows that this material likely
consists of about 68 weight% of a cubic phase Mn3[Fe(CN)6]2·4H2O and 32
weight% of a tetragonal phase [N(CH3)4]Mn[Fe(CN)6]·3H2O.
b)
a)
1.0
7.6
7.1
H2O fraction
A fraction
0.8
0.6
0.4
7.2
7.4
7.5
0.2
0.0
0.6
7.5
5.0
7.3
4.0
7.2
3.5
7.4
7.3
3.0
7.6
0.7
4.5
0.8
Fe fraction
0.9
1.0
0.75
0.80
0.85
0.90
7.1
0.95
Fe fraction
Figure 7.1. (a) Variation of the cation fraction (A) with Fe fraction. The dashed line indicates
charge neutrality (based on the assumption of A+, CN- and {MnII, FeIII} or {MnII, FeIII}).
(b)Variation of the water fraction with Fe fraction. The dashed line was reported in literature for
RbxMn[Fe(CN)6]y·zH2O compounds.9
121
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
Elemental analysis showed that Sr had not been incorporated in sample 7.6.10 In
addition the Mn:Fe ratio is close to 0.76, yielding a general composition of
Mn[Fe(CN)6]0.76·4.99H2O.
The variation of the water fraction with Fe fraction is shown in figure 7.1b. The
dashed line was found for the water/iron ratio in several samples of
RbxMn[Fe(CN)6]y·zH2O by Cobo et al.11 Possibly this trend in water/iron ratio
could also apply to AxMn[Fe(CN)6]y·zH2O samples with A not equal to Rb. Both
samples with cesium and sodium are quite far from this line and contain
considerably more water than expected. The other samples have a water content
that either is expected for samples with the found Mn:Fe ratio, or lower which can
be explained by the assumption that water is only incorporated when it is present
during the precipitation of the sample.12
The variation of the Fe fraction with the size of the cation is displayed in figure 7.2.
The assumption has been made that the amount of incorporation of the various
cations has largely to do with size, since this general observation was found both
for the AxCo[Fe(CN)6]y·zH2O compounds by Bleuzen et al.13 and for
AxMn[Fe(CN)6]y·zH2O compounds that were made in the present study.14 First of
all, the largest cation present (tetramethylammonium) is incorporated to a very
small extent.15 This is probably due to the fact that it is too large for the interstitial
sites.16 Small cations (K+ and NH4+ 17) are incorporated to a comparable extent as
N(CH3)4+. Due to the fact that already saturated solutions of KCl of NH4Cl had
been used to prepare these samples, the only possibility to increase the amount of
these cations in the final compound would be to reduce the concentrations of
MnCl2 and of K3[Fe(CN)6] in order to ensure a higher A:Mn/Fe ratio in the
solutions, possibly leading to a larger amount of incorporated A cations. Cesium is
incorporated very readily: even half the concentration necessary to prepare similar
Rb compounds already gives a compound which, based on the stoichiometry and
compared to a similar stoichiometry with RbxMn[Fe(CN)6]y·zH2O, might be able to
show electron transfer. Unfortunately, the magnetic measurements (section 7.3.4)
0.96
0.94
7.1
Figure 7.2. Variation of the Fe fraction
with cation size.
7.3
0.92
Fe fraction
0.90
0.88
0.86
0.84
0.82
0.78
0.76
0.74
7.5
7.2
0.80
7.6
7.4
75 100 125 150 175 200 225 250 275 300 325
Ionic radius (pm)
122
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
indicate that this sample does not show charge transfer. Sample 7.1 on the other
hand shows a very surprising result: solely based on the size of sodium we would
not expect this cation to be incorporated in such a large amount. This might be due
to the fact that NaCl has a higher solubility in water and thus the A:Mn/Fe ratio is
in this case large enough to ensure the higher inclusion.
Table 7.3. Experimental (exp) and calculated (cal) weight percentages of the various
samples.
A
7.1
Na+
7.2
K+
7.3
Cs+
7.4
NH4+
7.5a
N(CH3)4+
7.6
Sr2+
a
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
Exp
Cal
%A
%Mn
%Fe
%C
%N
%H
Proposed formula
5.72
5.73
4.35
4.35
26.51
26.51
97
ppm
0
16.75
16.77
18.19
17.97
13.19
13.19
18.73
18.73
16.41
16.41
16.06
16.04
13.96
14.22
12.08
12.45
15.07
15.07
13.14
13.14
21.66
20.70
18.76
18.35
16.41
16.06
19.82
19.45
21.70
21.70
24.76
24.14
21.40
21.40
18.73
18.73
24.28
24.28
20.94
21.16
1.42
1.86
2.42
2.65
0.58
1.46
2.71
2.92
3.62
4.15
Na0.82Mn[Fe(CN)6]0.94.
3.02H2O
K0.34Mn[Fe(CN)6]0.78.
4.02H2O
Cs0.83Mn[Fe(CN)6]0.93.
3.02H2O
[NH4]0.34Mn[Fe(CN)6]0.79.
3.58H2O
[N(CH3)4]0.33Mn[Fe(CN)6]0.79.
4.91H2O
17.99
13.73
18.41
20.83
2.64
18.00
13.84
17.86
20.83
3.30
Mn[Fe(CN)6]0.76.
4.99H2O
Yield
(%,Mn)
67
46
90
66
51
64
The actual composition seems to consist of 32% [N(CH3)4Mn[Fe(CN)6]·4H2O and 68% of
Mn3[Fe(CN)6]2·3H2O.
7.3.2 FTIR Spectroscopy
The FTIR spectra of all samples in the CN stretching region can be found in figure
7.3. The results for this infrared spectroscopic study are summarized in table 7.4.
As for the spectra reported in chapter 2 and 3, two peaks are present. Peak 1 is in
all cases positioned around 2070 cm-1, which is attributed to the CN stretching in
FeII-CN-MnIII, and peak 2 around 2150 cm-1 is attributed to FeIII-CN-MnII. The
spectra for the present samples are similar to the spectrum of sample 2.1 in chapter
2 and sample 3.7 in chapter 3 (the two RbxMn[Fe(CN)6]y·zH2O samples that do not
show electron transfer), i.e. the peak around 2100 cm-1 is never clearly present.
Unfortunately, it is not possible to determine the relative contributions of the two
configurations, since it is unknown with what relative scattering intensity the two
contribute. Based on this assignment as well as on the analogy explained in chapter
3 it is expected that none of these samples shows electron transfer and this turns
out to be indeed the case (see section 7.3.4).
Table 7.4. FTIR spectroscopic data (CN stretching bands) for AxMn[Fe(CN)6]y·zH2O
Sample
7.1
7.2
7.3
7.4
7.5
7.6
Peak 1 (FWHM*)
2071 (23)
2069 (24)
2078 (33)
2065 (29)
2072 (25)
2067 (24)
Peak 2 (FWHM*)
2145 (35)
2147 (31)
2152 (18)
2148 (30)
2153 (32)
2149 (37)
* FWHM = Full Width at Half Maximum
123
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
Sample 7.2
Sample 7.3
Sample 7.4
Sample 7.5
Sample 7.6
Intensity (a.u.)
Sample 7.1
2300
2200
2100
-1
ν (cm )
2000
1900
2300
2200
2100
2000
1900
-1
ν (cm )
Figure 7.3. Room temperature FTIR spectra for sample 7.1
(Na0.82Mn[Fe(CN)6]0.94.3.02H2O), sample 7.2 (K0.34Mn[Fe(CN)6]0.78.4.02H2O), sample 7.3
(Cs0.83Mn[Fe(CN)6]0.93.3.02H2O), sample 7.4 ((NH4)0.34Mn[Fe(CN)6]0.79.3.58H2O), sample
7.5 ([N(CH3)4]0.33Mn[Fe(CN)6]0.79.4.91H2O) and sample 7.6 (Mn[Fe(CN)6]0.76.4.99H2O).
7.3.3 X-Ray Powder Diffraction
Most Prussian Blue analogues adopt the F-43m space group at room temperature.18
Therefore, all diffraction profiles have been fitted with a model based on this space
group, in the same way as was done for the compounds described in chapter 2 and
3. Selected details of the refined structural models can be found in table 7.5. The
measured, calculated and difference diffraction profiles can be found in Appendix I.
124
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
In contrast to the powder diffraction profiles of the RbxMn[Fe(CN)6]y·zH2O
compounds, none of the present diffraction profiles contain peaks that are
indicative of the I-4m2 space group. This structure was found for the LT phase in
the samples with Rb as a cation.
In the case of samples 7.1 and 7.3 this refinement with the cubic F-43m space
group is very satisfactorily. In all other cases small peaks between 2θ = 15o and 25o
cannot be accounted for using this model. These remaining peaks are all on more
or less the same positions. Presumably, these peaks indicate the same side product
in all samples. Because the elemental analysis gave no indication of the presence of
other ions or atoms than the ones already found, the side product has to consist of
Fe, Mn, CN- and H2O entities.
The unindexed peaks in the diffraction pattern of Sample 7.5 have the highest
intensity. Hence, this material is very suitable for further investigation of these
features. For this sample the unindexed peaks were fitted using the XFit19 software
and the resulting peak positions were input for the Treor9020 and Taup21
autoindexing programs. All of the peaks could be indexed using a unit cell of 16.3
x 16.3 x 17.3 Å. The only known structure found with these dimensions containing
either Fe or Mn was [N(CH3)4]MII[M’III(CN)6].3H2O (M = Mn, Cd, M’ = Mn, Cr,
Fe, Co) with space group I4, Z = 10.22 This structure was added to the refinement
as a second phase next to the cubic F-43m structure. The result of this refinement
was that roughly 33 weight% of the compound consisted of the tetragonal structure
and the rest had the cubic structure. Since this weight percentage is very
comparable to the amount of tetramethylammonium incorporated in the compound
a new final refinement was tried: a combined approach consisting of the cubic
structure F-43m for Mn3[Fe(CN)6]2·4H2O and the tetragonal structure I4 for
[N(CH3)4]Mn[Fe(CN)6]·3H2O.22 Unfortunately, in the second structure a total of 35
atoms is present on non special positions and the intensity of the peaks is relatively
Table 7.5. Selected details for the refined diffraction profiles with F-43m of the various
samples.
Sample
a (Å)
Fe-C (Å)
7.1
7.2
7.3
7.4
7.5a
10.6049(4)
10.5135(4)
10.5725(5)
10.5167(3)
10.4996(7)
ac:
10.4984(4)
at:
16.2457(9)
ct: 17.254(2)
10.5014(4)
1.925(15)
1.928(17)
1.838(17)
1.918(11)
1.960(34)
7.5b
a
7.6
1.942(16)
Mn-N
(Å)
2.245(13)
2.220(15)
2.310(14)
2.221(9)
2.204(29)
2.18(9)
C-N (Å)
1.133(12)
1.109(22)
1.139(15)
1.119(14)
1.09(4)
1.12(8)
fA in
4c
0.41
0.17
0.415
0.17
0.165
0
wRp
(%)
18.96
21.48
11.59
18.59
30.98
Rp
(%)
13.86
15.90
8.84
13.26
21.28
Reduced χ2
(#variables)
6.924 (17)
8.796 (18)
1.900 (18)
6.362 (18)
18.98 (17)
17.05
12.72
5.761 (24)
21.55
15.54
9.370 (18)
Fitted with only the cubic space group F-43m similar to the rest of the samples
b
Fitting with both cubic space group F-43m (weight fraction 68.1(4)%) and tetragonal space group I4
(weight fraction 31.9(4)%), for details see text
125
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
low. It was therefore not possible to refine the atom positions or the atom fractions
and only the lattice parameters and weight fraction of the total structure could be
obtained. For a more detailed analysis of the tetragonal structure single crystals are
needed, which to this date proved to be impossible to obtain. Presumably the side
product present in samples 7.2, 7.4 and 7.6 has the same tetragonal structure. The
available space of the large N(CH3)4+ cation in the interstitial sites could be
replaced by the used cation combined with H2O to fill the large interstitial sites in
this structure (i.e. A(H2O)n+ instead of N(CH3)4+).
The distances for Fe – C of ~1.9Å and Mn – N of ~2.2Å found in the cubic
structure for all samples are in the range for LS FeIII and HS MnII, respectively.23
The variation of the lattice parameters of the cubic structure with cation size can be
found in figure 7.4. Generally with larger cation size, the lattice parameter a
increases, which is as expected, since the larger cation has to fit in the interstitial
sites, apparently necessitating an expansion of the crystal lattice. However, the
sample with sodium as a cation is completely off in this trend: it has the smallest
cation size and the largest lattice parameter. Probably this is due to the relatively
high incorporation of sodium in the lattice in contrast to the other cations, although
it is still expected that the compound with cesium should have larger lattice
parameters than the sample with sodium. The small lattice parameter found for the
sample with tetramethylammonium is due to the fact that it is not incorporated in
the cubic F-43m structure and hence the lattice parameter is more sensitive to the
incorporated water than to the size of tetramethylammonium.
10.62
Figure 7.4. Variation of the lattice
parameter a of the cubic space group
F-43m with cation size, as found by
refinement of X-ray powder diffraction
profiles.
7.1
10.60
7.3
a (A)
10.58
10.56
10.54
7.4
10.52
10.50
7.6
7.2
7.5
75 100 125 150 175 200 225 250 275 300 325
Cation radius (pm)
7.3.4 Magnetic Measurements
The variation of χMT with temperature for all samples is shown in figure 7.5 and
the details are summarized in table 7.6. Samples 7.1 to 7.4 show a magnetic
behaviour which is comparable to a sample that does not show any electron
transfer.24 At room temperature these samples all have a χMT value which is close
to or somewhat lower than 4.75 cm3 K mol-1 indicating that most of the compound
has the LS FeIII and HS MnII configuration, the value being lowered by some
126
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
residual LT phase. In samples 7.1 to 7.4 the presence of some residual LT phase is
visible since an apparent increase of χMT is present at low temperatures. The
pronounced presence of the peak around 2070 cm-1 in the IR spectra indicates that
LT phase is indeed present at room temperature. The X-Ray powder diffraction
profiles gave no indication of the I-42m structure (indicative for the LT phase in
the analogous RbxMn[Fe(CN)6]y·zH2O compounds), but it is of course possible to
have LS FeII and HS MnIII present in the cubic F-43m structure.
7
0
100
150
200
250
300
00
35
50
100
150
200
250
300
350
7
S a m p le 7 .2
S a m p le 7 .1
6
6
5
5
4
4
3
3
2
2
1
1
M
3
-1
χ T (cm K mol )
50
0
7
0
7
S a m p le 7 .3
6
5
5
4
4
3
3
2
2
1
1
M
3
-1
χ T (cm K mol )
6
S a m p le 7 .4
0
7
0
7
6
6
5
5
4
4
3
3
2
2
1
1
M
3
-1
χ T (cm K mol )
S a m p le 7 .6
S a m p le 7 .5
0
0
50
100
150
200
T (K )
250
300
35
00
50
100
150
200
250
300
0
350
T (K )
Figure 7.5. Magnetic susceptibility measurements for sample 7.1
(Na0.82Mn[Fe(CN)6]0.94.3.02H2O), sample 7.2 (K0.34Mn[Fe(CN)6]0.78.4.02H2O), sample 7.3
(Cs0.83Mn[Fe(CN)6]0.93.3.02H2O), sample 7.4 ([NH4]0.34Mn[Fe(CN)6]0.79.3.58H2O), sample
7.5 ([N(CH3)4]0.33Mn[Fe(CN)6]0.79.4.91H2O) and sample 7.6 (Mn[Fe(CN)6]0.76.4.99H2O).
127
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
Table 7.6. Selected details of the magnetic measurements
Sample
7.1
7.2
7.3
7.4
7.5
7.6
χMT at 300 K (cm3 K
mol-1)
4.58 (5)
4.62 (3)
4.56 (6)
4.26 (3)
0.47 (4)
6.34 (6)
C (cm3 K mol-1)
θ (K)
4.665(3)
4.579(3)
4.660(1)
4.110(1)
0.464(1)
6.467(8)
-6.1(3)
-1.9(2)
-4.881(4)
-3.066(2)
-9.83(1)
-11(2)
If it is assumed that the present samples need the same stoichiometry for electron
transfer as the samples with Rb as a cation, then based on their stoichiometry it is
not expected that samples 7.2 and 7.4 show any electron transfer. The reason for
this is explained in chapter 2: Fe(CN)6 defects are replaced by water molecules and
the Mn ions are now surrounded by Mn(CN)6-x(OH2)x instead of 6 cyano ligands.
In contrast to MnII(CN)6/MnIII(CN)6 the redox couple MnII(CN)6-x(OH2)x/
MnIII(CN)6-x(OH2)x is not able to reduce FeIII(CN)6 to FeII(CN)6, thereby explaining
the lack of electron transfer in compounds with a stoichiometry which is far from
Rb:Mn:Fe = 1:1:1. However, based on the stoichiometry of sample 7.1 and 7.3
electron transfer for these two samples is expected, but is not found in the
measured temperature region. From the powder diffraction refinement (section
7.3.3) for these two samples it is found that the cations are equally distributed over
the two interstitial sites, which would therefore suggest (as in chapter 4 and 5), that
only Fe sites (and most probably Mn sites) with a perfectly cubic surrounding are
capable of inducing electron transfer.
Sample 7.5 shows a rather low value of χMT of 0.47 cm3 K mol-1 at room
temperature. Such a low value indicates a total spin of S = 0.6. From the IR spectra
we know that either FeIII (S = ½) and MnII (S = 5/2) are linked to the cyano bridge
or FeII (S = 0) and MnIII (S = 2). It is not possible to combine the spins of the metal
ions in such a way to create this total spin of 0.6. Possibly its different crystal
structure gives rise to this different behaviour, although it is not clear how.
Sample 7.6 has a rather high value of χMT of 6.34 cm3 K mol-1 at room temperature.
This value indicates a total spin of S = 3.0. Similar to sample 7.5 the IR spectrum
indicates the presence of either FeIII and MnII or FeII and MnIII. One possibility is
that LS FeIII and HS MnII order ferromagnetically to arise at this high spin value,
but this is only speculation. Above all, the observed value for θ (see below) is in
contrast with this speculation. More research is needed to determine the precise
spin states of the metal ions in this sample.
Figure 7.6 shows the variation of the Curie (C) and the Curie-Weiss constant (θ)
with cation size. The values for C are more or less the same for all samples, with
the exception of samples 7.5 and 7.6. Since C is always close to the value of χMT at
128
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
a)
b)
7
7.2
7.6
-2
7.4
7.1
4
7.2
-4
7.3
7.3
7.1
7.4
θ (K)
5
-1
C (cm K mol )
6
3
3
-6
-8
2
7.5
1
7.5
0
-10
7.6
-12
100
150
200
250
300
Ion radius (pm)
100
150
200
250
300
Ion radius (pm)
Figure 7.6. Variation of C (a) and θ (b) with ion radius, as found by fitting the inverse
magnetic susceptibility with a straight line.
300 K, the reasoning for these values is already explained in the part describing the
χMT values. All values of θ are negative, indicating antiferromagnetic coupling
between Mn and Fe neighbours for all samples.
7.4 Discussion and Conclusions
7.4.1 Synthetic considerations
In chapter 2 it was shown that for powdered samples the most efficient electron
transfer compounds have a ratio of Rb:Mn:Fe close to 1:1:1. The reasoning behind
this observation lies in the coordination around the Mn ion: when more water
molecules are surrounding Mn it is not able to reduce FeIII(CN)6 to FeII(CN)6. It
was assumed that this reasoning also applies to compounds with a different cation
than Rb. The use of a different cation (with a different size) might give changes in
the flexibility of the lattice which would therefore give changes in the electron
transfer properties of the compounds, similar to the charge transfer properties of
the AxCo[Fe(CN)6]y·zH2O compounds.5 As was seen in chapter 3, the best way to
get a 1:1:1 ratio is by slowly adding the solution containing MnCl2·4H2O to the
solution with K3[Fe(CN)6]. Because the outcome in the RbxMn[Fe(CN)6]y·zH2O
compounds originating from an addition speed of 3 or 6 mL h-1 was very
comparable (see chapter 3) an addition speed of 6 mL h-1 was used in the synthesis
of all present samples.
In the case of the alkali cations it was noticed that the smaller the cation, the less of
it was incorporated when the same concentration (1 M) of the cation salt was used
in the synthesis procedure as in the synthesis of the Rb samples reported in the
129
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
previous chapters.13 In literature it was shown that a smaller concentration of RbCl
used in the synthesis results in more Rb defects in the manganese hexacyanoferrate
material and consequently less Fe(CN)6 is incorporated as well.4 So probably more
concentrated solutions of NaCl and KCl would give a higher amount of the alkali
cation in the compound. Therefore, saturated solutions of NaCl and KCl were used
during the precipitation of sample 7.1 and 7.2.
Because small cations (Na+ and K+) were incorporated less and cesium cations
were incorporated more easily it was assumed that the capability of the cations to
be included in the final compounds had to do with size: the smaller the cation the
less it would be integrated. Since the ammonium ion has a comparable size as the
Rb ion, the first attempt was to prepare a compound with the same concentrations
used for the cation salt as was used in the synthesis for Rb. This gave a compound
with a very low Mn:Fe ratio,25 therefore a saturated solution of ammonium chloride
was used in the preparation of sample 7.4.
For the same reasoning, in the synthesis of sample 7.3 and 7.5 half the
concentration (0.5 M) that was used for RbCl was employed for the cation salt. In
the case of sample 7.3 this concentration (0.5 M) seems to be sufficient to give a
stoichiometry which, when compared to the stoichiometry of the
RbxMn[Fe(CN)6]y.zH2O compounds, would give a compound which might be able
to show electron transfer. For tetramethylammonium however, this concentration
of the cation salt (0.5 M) gave a high concentration of Fe(CN)6 defects, possibly
because the ion is too large to fit in the interstitial sites (see 7.3.1). This is as
expected, since the space that is present in the interstitial sites is large enough for a
sphere with a radius of 265 pm16 and the radius of tetramethylammonium is larger.
Therefore, no further attempts have been made to improve the stoichiometry of
sample 7.5.
For sample 7.6, we used a completely different type of cation since it is doubly
charged instead of singly charged. We would therefore expect a maximum of 0.5
Sr2+ ions per Mn ion. That is why only 0.5 M of SrCl2 was used in the preparation
of sample 7.6. It turned out that the Sr2+ ions were not incorporated at all (section
7.3.1) and the magnetic properties of the compound were so different (section
7.3.4) from the rest of the AxMn[Fe(CN)6]y.zH2O compounds, that also in this case
no further attempts have been made to improve the stoichiometry.10
7.4.2 Sample 7.1
The elemental analysis indicates that sample 7.1 has a stoichiometry of
Na0.82Mn[Fe(CN)6]0.94.3.02H2O. When comparing this stoichiometry with the
trends known for powdered samples of RbxMn[Fe(CN)6]y.zH2O (i.e. the more the
Rb:Mn:Fe ratio approaches 1:1:1, the more complete the charge transfer),24 it is
expected that this sample shows electron transfer. However, the magnetic
susceptibility measurements show no electron transfer in the measured temperature
region. The shape of the IR spectrum in the CN stretching region is in agreement
130
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
with this observation: it is the same as the IR spectrum for RbxMn[Fe(CN)6]y·zH2O
samples that do not show electron transfer (i.e. two pronounced bands around 2070
cm-1 and 2150 cm-1 and no band around 2100 cm-1). A possible explanation for this
can be inferred from the refinement of the X-ray powder diffraction profile: the
sodium ions are equally distributed over the two different interstitial sites. In
chapter 4 and 5 it was shown that RbxMn[Fe(CN)6]y·zH2O samples which did not
show electron transfer have their Rb ions and water molecules equally distributed
over the two interstitial sites, which leads to the generation of different Fe and Mn
sites. Each of these different Fe and Mn sites differ in the charge transfer
capability. Possibly this also applies to samples with Na+ as a cation. Another
possibility is that the smaller sodium allows for a more open and flexible crystal
structure, which might be the reason that the electron transfer process does not take
place in this compound.
7.4.3 Sample 7.2
The found stoichiometry for sample 7.2 is K0.34Mn[Fe(CN)6]0.78.4.02H2O. The
small K+ content is probably due to its relatively small size. From chapter 2 and 3 it
is known that RbxMn[Fe(CN)6]y.zH2O samples with a comparable stoichiometry do
not show any electron transfer. Sample 7.2 does not show any electron transfer, as
is clear from the magnetic measurements. The IR spectrum is in agreement with
this observation: the IR spectrum shows no band around 2100 cm-1 similar as the
RbxMn[Fe(CN)6]y·zH2O samples that do not show a phase transition.
7.4.4 Sample 7.3
The Cs+ cation is larger than Rb+, but not too large for the interstitial sites. It is
therefore expected that it is readily incorporated, as indeed is the case: a
stoichiometry of Cs0.83Mn[Fe(CN)6]0.93.3.02H2O was found. As with sample 7.1, in
case it would be a Rb analogue for this stoichiometry a sample is expected that is
capable of showing electron transfer. Unfortunately, again no switching is
observed. Similar to sample 7.1, the Cs is equally distributed over the two
interstitial sites, which could be the reason why the compound does not switch.
Again, the IR spectrum shows results that are in agreement with this behaviour.
When comparing this sample with the CsxMn[Fe(CN)6]y.zH2O samples from the
literature which do show (partial) electron transfer,2 the most striking difference is
the fact that in our case almost no FeII(CN)6 is present at room temperature. The
difference in the synthesis of this sample and the samples in literature2 is that the
ratio between MnII:Cs+ in the starting solutions ranges from 1:50 to as high as
1:200, rather than the ratio of 1:5 used in this study. It is unclear why in the
samples in literature part of the FeIII(CN)6 has been reduced to FeII(CN)6, since the
starting materials are the same as in the present synthetic method.
7.4.5 Sample 7.4
The size of ammonium is very similar to the size of rubidium. It is therefore
expected that it will just as easily be incorporated in the system as the Rb ion.
131
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
However, this turns out not to be the case. Even an increase of the concentration of
the ammonium salt in the starting solution to a saturated solution still gives a
sample whose stoichiometry is only [NH4]0.34Mn[Fe(CN)6]0.79.3.58H2O. It is
unclear why this cation is not readily incorporated, possibly the hydrogen bonding
ability of NH4+ could be a reason or the fact that it can act as a Brønsted acid. But,
as expected for a sample with this stoichiometry it does not show any switching
between the HT and the LT phase. The results of the IR spectroscopic study are in
agreement with a sample that does not show electron transfer.
7.4.6 Sample 7.5
Of the used cations in this study, tetramethylammonium has the largest size. It is
even larger than the available space in the interstitial sites and this is probably the
reason that it is difficult to be incorporated in the system: an overall stoichiometry
of only [N(CH3)4]0.33Mn[Fe(CN)6]0.79.4.91H2O was found. The compound that was
synthesised probably consists of two different materials with different structure:
[N(CH3)4]Mn[Fe(CN)6]·3H2O
with
a
tetragonal
I4
structure
and
Mn3[Fe(CN)6]2·4H2O with a cubic F-43m structure. This sample
{[N(CH3)4]Mn[Fe(CN)6]·4H2O}0.32{Mn3[Fe(CN)6]2·3H2O}0.68 is very different to
the other samples: its value of χMT is much lower than the value observed for the
other samples. Possibly its different crystal structure gives rise to this different
behaviour. Further attempts are needed to purify the two separate compounds and
investigate the physical properties of each individually.
7.4.7 Sample 7.6
In contrast to the other used cations strontium has a divalent charge rather than
monovalent. Strangely it is not incorporated at all in the system. The found
stoichiometry is Mn[Fe(CN)6]0.76.4.99H2O. However, when expecting a divalent
manganese and a trivalent iron ion together with six cyano ligands, the ratio
between Mn and Fe should be more in the region of 3:2 rather than the found 4:3
(compare for instance the Mn3[Fe(CN)6]2·4H2O phase in sample 7.5). We thus have
more Fe(CN)6 present than expected, which indicates that some redox process has
taken place. The found high value of χMT of 6.34 cm3 K mol-1 is also different to
the values of the other samples.26 More research is needed to determine the exact
valencies and spin states of the metal ions in sample 7.6.
7.4.8 Overall considerations
From this set of samples, the physical measurements on samples 7.1-7.4 can be
explained by analogy with the RbxMn[Fe(CN)6]y·zH2O samples. The reasoning that
none of these samples shows electron transfer capabilities lies either in the fact that
too many [Fe(CN)6] defects are present or that the A cations are randomly
distributed over the two interstitial sites. The lack of electron transfer possibilities
in the first case is explained in chapter 2 and in the second case in chapter 4 and 5.
All these samples show characteristics that are comparable to inactive electron
132
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
transfer compounds. The reason why [Fe(CN)6] defects would be present or why
the A cations would be randomly distributed is not clear as yet.
Samples 7.5 and 7.6 show behaviour that has not been seen before for
AxMn[Fe(CN)6]y·zH2O compounds and therefore it is also not clear what exactly
causes these physical properties.
7.5 References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
See for a review: Ohkoshi, S.-I.; Tokoro, I.; Hashimoto, K.; Coord.Chem. Rev. 2005, 249, 1830,
and references therein.
a) Matsuda, T.; Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I.; J. Chem. Soc.,Dalton transactions
2006, 5046; b) Matsuda, T.; Tokoro, H.; Hashimoto, K.; Ohkoshi, S.-I.; J. Appl. Phys. 2007, 101,
09E101; c) Ishiji, K.; Matsuda, T.; Tokoro, H.; Iwazumi, T.; Hashimoto, K.; Ohkoshi, S.-I.; J.
Phys. Chem. Solids 2007, 68, 2158
Chapter 2
Ohkoshi, S.-I.; Matsuda, T.; Tokoro, H.; Hashimoto, K.; Chem. Mater. 2005, 17, 81
a) Bleuzen, A.; Lomenech, C.; Escax, V. Villain, F.; Varret, F.; Cartier dit Moulin, C.;
Verdaguer, M.; J. Am. Chem. Soc. 2000, 122, 6648; b) Cartier dit Moulin, C.; Villain, F.;
Bleuzen, A.; Arrio, M.-A.; Sainctavit, P.; Lomenech, C.; Escax, V.; Baudelet, F.; Dartyge, E.;
Gallet, J.-J.; Verdaguer, M.; J. Am. Chem. Soc. 2000, 122, 6653; c) Escax, V.; Bleuzen, A.;
Cartier dit Moulin, C.; Villain, F.; Goujon, A.; Varret, F.; Verdaguer, M. ; J. Am. Chem. Soc.
2001, 123, 12536 ; d) Champion, G.; Escax, V.; Cartier dit Moulin, C.; Bleuzen, A.; Villain, F.;
Baudelet, F.; Dartyge, E.; Verdaguer, M. ; J. Am. Chem. Soc. 2001, 123, 12544.
See also chapter 1
Table 39A in BINAS, Informatieboek HAVO-VWO voor het onderwijs in de
natuurwetenschappen, 1992, 3rd edition, Wolters-Noordhoff, Groningen, The Netherlands, p. 78
Calculated with the use of N – C bond (147 pm), C – H bond (108 pm), vanderWaals radius of H
(120 pm) and N – C – H angle (109.45o). The used parameters come from table 53 in BINAS,
Informatieboek HAVO-VWO voor het onderwijs in de natuurwetenschappen, 1992, 3rd edition,
Wolters-Noordhoff, Groningen, The Netherlands, p. 100
Cobo, S.; Fernández, R.; Salmon, L.; Molnár, G.; Bousseksou, A.; Eur. J. Inorg. Chem. 2007,
1549
A duplo elemental analysis of the same sample gave a Sr content of 123 ppm.
Cobo, S.; Fernández, R.; Salmon, L.; Molnár, G.; Bousseksou, A.; Eur. J. Inorg. Chem. 2007,
1549
See chapter 3
Bleuzen, A.; Escax, V.; Itie, J.-P.; Munsch, P.; Verdaguer, M.; C.R. Chimie 2003, 6, 343
Compounds of Na0.11Mn[Fe(CN)6]0.74·4.43H2O (A), K0.23Mn[Fe(CN)6]0.74·3.75H2O (B) and
Cs0.87Mn[Fe(CN)6]0.84·1.95H2O (C) had been obtained, in the synthesis with identical
concentrations of the cation salt as used in the synthesis for the Rb compounds. Calculated for A:
Na 0.84%, Mn 18.64%, Fe 14.08%, C 18.17%, N 21.19%, H 3.03%; Observed: Na 0.82%, Mn
18.64%, Fe 14.08%, C 18.55%, N 20.85%, H 2.75%. Calculated for B: K 3.12%, Mn 19.08%, Fe
14.31%, C 18.46%, N 21.53%, H 2.63%; Observed: K 3.12%, Mn 19.08%, Fe 14.31%, C
18.34%, N 21.11%, H 2.54%. Calculated for C: Cs 30.02%, Mn 14.30%, Fe 12.27%, C 15.83%,
N 18.46%, H 1.02%; Observed: Cs 30.02%, Mn 14.30%, Fe 12.72%, C 16.31%, N 18.46%, H
0.57%.
CH3 bending frequencies (~1400 cm-1) are present in the IR spectrum, thereby indicating that
N(CH3)4+ is indeed present
133
Chapter 7. Attempted Synthesis, Characterisation and Physical Properties of AxMn[Fe(CN)6] y·zH2O
16
17
18
19
20
21
22
23
24
25
26
134
The diameter of the cavity is 5.29 Å in the HT structure (see chapter 5 and appendix II). The
radius is then 2.65 Å, which is smaller than the radius of N(CH3)4+.
NH bending (3000 – 3600 cm-1) and NH stretching frequencies (1600 cm-1) overlap with OH
bending and OH stretching frequencies in the IR spectra. However, the IR spectrum of this
sample does look different in the mentioned frequency regions from IR spectra of the other
samples.
See chapter 1
Cheary, R.W.; Coelho, A.A.; XFIT 1996, http://www.ccp14.ac.uk/tutorial/xfit-95/xfit.htm
Werner, P.-E.; Eriksson, L.; Westdahl, M.; J. Appl. Cryst. 1985, 18, 367
Taupin, D.; J. Appl. Cryst. 1973, 6, 380
Witzel, M.; Ziegler, B.; Babel, D.; Z.Anorg.Allg.Chem. 2000, 626, 471
See also chapter 5
See chapter 2 and 3
According to elemental analysis performed by electron microscopy a Mn:Fe ratio of 1:0.75 was
obtained.
In a similar synthesis LiCl was used as a cation source. Electron microscopy measurements on
the resulting sample indicated the inclusion of K+ (from K3[Fe(CN)6]) rather than Li+, the found
stoichiometry was K0.15Mn[Fe(CN)6]0.73·4.05H2O (calculated: Mn 19.10%, Fe 14.08%, C
18.17%, N 21.19%, H 2.84%; observed: Mn 19.10%, Fe 14.08%, C 18.90%, N 20.57%, H
2.84%). The elemental analysis of sample 7.6 does not indicate that K+ was incorporated. It is
therefore expected that the presence of Sr2+ in the synthesis of sample 7.6 gives rise to the
strange physical properties.
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Appendix I - X-Ray Powder Diffraction Profiles for
Samples in Chapters 2, 3 and 7
Intensity (a.u.)
Intensity (a.u.)
Observed (+), fitted and difference (grey) powder diffraction profiles for samples
reported in chapter 2 (figures I2.1-I2.4), chapter 3 (figures I3.1-I3.13) and chapter
7 (figures I7.1-I7.7).
10
20
30
40
50
60
70
10
2θ
20
30
40
50
60
70
2θ
Rb0.92Mn[Fe(CN)6]0.95.1.03H2O fitted
with cubic space group F-43m
Intensity (a.u.)
Figure I2.2 Sample 2.2
Rb0.59Mn[Fe(CN)6]0.86.2.63H2O fitted with
cubic space group F-43m
Intensity (a.u.)
Figure I2.1 Sample 2.1,
10
20
30
40
50
60
70
2θ
10
20
30
40
50
60
70
2θ
Figure I2.3 Sample 2.3
Figure I2.4 Sample 2.4
Rb0.97Mn[Fe(CN)6]0.98.1.03H2O fitted with
cubic space group F-43m and tetragonal space
group I-4m2
Rb0.81Mn[Fe(CN)6]0.95.1.24H2O fitted
with cubic space group F-43m
135
Intensity (a.u.)
Intensity (a.u.)
Appendix I – X-Ray Powder Diffraction Profiles for Samples in Chapters 2, 3 and 7
10
20
30
40
50
60
70
10
20
30
2θ
50
60
70
Figure I3.2 Sample 3.2
Rb0.89Mn[Fe(CN)6]0.96.1.73H2O fitted with
cubic space group F-43m
Intensity (a.u.)
Intensity (a.u.)
Figure I3.1 Sample 3.1
Rb0.86Mn[Fe(CN)6]0.95.1.94H2O fitted with
cubic space group F-43m
10
20
30
40
50
60
70
2θ
Figure I3.3 Sample 3.3
Rb0.91Mn[Fe(CN)6]0.97.1.53H2O fitted with
cubic space group F-43m
136
40
2θ
10
20
30
40
50
60
70
2θ
Figure I3.4 Sample 3.4
Rb0.92Mn[Fe(CN)6]0.97.2.25H2O fitted with
cubic space group F-43m
Intensity (a.u.)
Intensity (a.u.)
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
10
20
30
40
50
60
70
10
20
30
2θ
40
50
60
70
2θ
Figure I3.6 Sample 3.6
Rb0.95Mn[Fe(CN)6]0.99.2.16H2O fitted with
cubic space group F-43m and tetragonal
space group I-4m2
Intensity (a.u.)
Intensity (a.u.)
Figure I3.5 Sample 3.5
Rb0.91Mn[Fe(CN)6]0.96.1.48H2O fitted with
cubic space group F-43m and tetragonal space
group I-4m2
10
20
30
40
50
60
70
2θ
Figure I3.7 Sample 3.7
Rb0.61Mn[Fe(CN)6]0.86.2.71H2O fitted with
cubic space group F-43m
10
20
30
40
50
60
70
2θ
Figure I3.8 Sample 3.8
Rb0.82Mn[Fe(CN)6]0.94.1.39H2O fitted with
cubic space group F-43m
137
Intensity (a.u.)
Intensity (a.u.)
Appendix I – X-Ray Powder Diffraction Profiles for Samples in Chapters 2, 3 and 7
10
20
30
40
50
60
70
10
20
30
2θ
50
60
70
Figure I3.10 Sample 3.10
Rb0.91Mn[Fe(CN)6]0.97.0.90H2O fitted with
cubic space group F-43m
Intensity (a.u.)
Intensity (a.u.)
Figure I3.9 Sample 3.9
Rb0.85Mn[Fe(CN)6]0.94.1.27H2O fitted with
cubic space group F-43m
10
20
30
40
50
60
70
2θ
Figure I3.11 Sample 3.11
Rb1.01Mn[Fe(CN)6]1.00.1.00H2O fitted with
cubic space group F-43m
138
40
2θ
10
20
30
40
50
60
70
2θ
Figure I3.12 Sample 3.12
Rb0.93Mn[Fe(CN)6]0.99.0.78H2O fitted with
cubic space group F-43m
Intensity (a.u.)
Intensity (a.u.)
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
10
20
30
40
50
60
70
10
20
30
2θ
40
50
60
70
2θ
Figure I.7.1 Sample 7.1
Na0.82Mn[Fe(CN)6]0.78.3.02H2O fitted with
cubic space group F-43m
Intensity (a.u.)
Intensity (a.u.)
Figure I3.13 Sample 3.13
Rb0.96Mn[Fe(CN)6]0.98.0.75H2O fitted with
cubic space group F-43m
10
20
30
40
50
60
70
2θ
Figure I7.2 Sample 7.2,
K0.34Mn[Fe(CN)6]0.78.4.02H2O fitted with
cubic space group F-43m
10
20
30
40
50
60
70
2θ
Figure I7.3 Sample 7.3,
Cs0.83Mn[Fe(CN)6]0.93.3.02H2O fitted with
cubic space group F-43m
139
Intensity (a.u.)
Intensity (a.u.)
Appendix I – X-Ray Powder Diffraction Profiles for Samples in Chapters 2, 3 and 7
10
20
30
40
50
60
70
10
20
30
2θ
50
60
70
Figure I7.5 Sample 7.5,
[N(CH3)4]0.33Mn[Fe(CN)6]0.79.4.91H2O,
fitted with cubic space group F-43m
Intensity (a.u.)
Intensity (a.u.)
Figure I7.4 Sample 7.4,
(NH4)0.34Mn[Fe(CN)6]0.79.3.58H2O fitted with
cubic space group F-43m
10
20
30
40
50
60
70
2θ
Figure I7.6 Sample 7.5,
[N(CH3)4]0.33Mn[Fe(CN)6]0.79.4.91H2O, fitted
with both the cubic space group F-43m and
the tetragonal space group I4
140
40
2θ
10
20
30
40
50
60
2θ
Figure I7.7 Sample 7.6,
Mn[Fe(CN)6]0.76.4.99H2O fitted with
cubic space group F-43m
70
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Appendix II - Crystallographic Details of the Crystal
Structure Described in Chapter 5
II.1 Experimental
A crystal with the dimensions of 0.31 x 0.28 x 0.21 mm was mounted on top of a
glass fiber and aligned on a Bruker1 SMART APEX CCD diffractometer (Platform
with full three-circle goniometer). The diffractometer was equipped with a
4K CCD detector set 60.0 mm from the crystal. The measurement took place at
293(1) K. Intensity measurements were performed using graphite monochromated
Mo-Kα radiation from a sealed ceramic diffraction tube (SIEMENS). Generator
settings were 50 KV/ 40 mA. SMART was used for preliminary determination of
the unit cell constants and data collection control. The intensities of reflections of a
hemisphere were collected by a combination of 3 sets of exposures (frames). Each
set had a different φ angle for the crystal and each exposure covered a range of 0.3o
in ω. A total of 1800 frames were collected with an exposure time of 10.0 seconds
per frame. Data integration and global cell refinement was performed with the
program SAINT.1 The final unit cell was obtained from the xyz centroids of 1705
reflections after integration. Intensity data were corrected for Lorentz and
polarization effects, scale variation, for decay and absorption: a multi scan
absorption correction was applied, based on the intensities of symmetry related
reflections measured at different angular settings (SADABS)2, and reduced to Fo2.
The program suite SAINTPLUS was used for space group determination (XPREP).1
The unit cell3 was identified as cubic. Space group F-43m was derived from the
systematic extinctions and discriminated from other candidate space groups, which
comply with the same extinction conditions, during the structure determination
process. Examination of the final atomic coordinates of the structure did yield an
extra crystallographic symmetry element: an inversion centre, but refinement with
space group Fm-3m did not lead to chemically reasonable Fe-C and Mn-N
distances.
The initial coordinates4,5 were taken from a refined, powder structure. The
positional and anisotropic displacement parameters were refined. The Rb atom is
disordered over two equivalent positions (4c & 4d). The s.o.f. of the major fraction
of the component of the disorder model refined to a value of 0.7512(4). From the
elemental analysis it was clear that the ratio (CN) / H is 3:1. The oxygen (from
water) is located on the same positions as Rb and also is disordered. The s.o.f. of
Rb1 is set equal to the s.o.f. of O2 and v.v. The positional-, anisotropic
displacement parameters were set to be equal and the relative s.o.f. was refined.
Final refinement on F2 carried out by full matrix least squares techniques
converged at wR(F2) = 0.1004 for 181 reflections and R(F) = 0.0348 for 170
reflections with Fo ≥ 4.0 σ(Fo) and 16 parameters. The final difference Fourier map
141
Appendix II – Crystallographic Details of the Crystal Structure Described in Chapter 5
was essentially featureless: no significant peaks (0.8(2) e/Å3) having chemical
meaning above the general background were observed.
The crystal appeared to be an inversion twin:6,7,8,9 x parameter refined to 0.69(11).
The crystal exhibited some secondary extinction for which the Fc2 values were
corrected by refinement of an empirical isotropic extinction parameter10
(0.0025(6)).
The positional and anisotropic displacement parameters were refined on F2 with
full matrix least squares procedures minimizing the function Q = ∑h[w(│(Fo2) k(Fc2)│)2], where w = 1/[σ2(Fo2) + (aP)2 + bP], P = [max(Fo2,0) + 2Fc2] / 3, F0
and Fc are the observed and calculated structure factor amplitudes, respectively;
ultimately the suggested a (=0.0596) and b (= 1.7609) were used in the final
refinement.
Crystal data and numerical details on data collection and refinement are given in
Table II.1. Final fractional atomic coordinates, equivalent displacement parameters
and anisotropic displacement parameters for the atoms are given in Table II.2 and
Table II.3, respectively. Neutral atom scattering factors and anomalous dispersion
corrections were taken from International Tables for Crystallography.10
All refinement calculations and graphics were performed on a HP XW6200 (Intel
XEON 3.2 GHz) / Debian-Linux computer at the University of Groningen with the
program packages SHELXL11 (least square refinements), a locally modified version
of the program PLUTO12 (preparation of illustrations) and PLATON13 package
(checking the final results for missed symmetry with the MISSYM option, solvent
accessible voids with the SOLV option, calculation of geometric data and the
ORTEP11 illustrations).
II.2 Results
All atoms of the asymmetric unit are located on a special position, but with
different site symmetry: Rb1 (4c): -43m (4/96), Rb2 (4d): -43m (Rb1 and Rb2 are
partly occupied with an s.o.f. of 0.7512(4) and 0.2488(4), respectively), O1 (4c): 43m (4/96), O2 (4d): -43m (O1 and O2 are partly occupied with an s.o.f. of
0.2488(4) and 0.7512(4), respectively), Mn (4b): -43m (4/96), Fe (4a): -43m
(4/96), N (24f) : 2.mm (24/96), C (24f) : 2.mm (24/96). The cubic unit cell contains
five units: free Rb cations, free oxygen atoms from water and a three dimensional
network of (-Fe-C-N-Mn-); Fe and Mn are each surrounded by six CN ligands.
II.3 Attempts to determine the structure at 100K
Slow cooling of the crystal to 100 K and subsequent diffraction measurements at
100 K resulted in many diffuse spots and a unit cell by dirax:14 a, b, c = 7.4348,
7.4564, 7.4572 and α, β, γ = 60.164, 60.217, 60.162. Apart from these unit cell
parameters being rather unrealistic, also about 20% of the reflections did not fit in
this orientation matrix. Noting that the crystallographic data recorded at room
temperature revealed twinning, this may also be the case at 100 K. In addition,
142
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
from X-ray diffraction on powdered samples it is known that the low temperature
form of the material has a tetragonal symmetry,4,5 and the crystal may have a lack
in preference for the elongation of one axis when going from cubic to tetragonal.
Furthermore, the partial switching of the material will give another complication in
the elucidation of the crystal structure at 100 K. Similar results were obtained on
various single crystals.
Table II.1a Crystal data and details of the structure determination
Moiety_Formula
Formula_Weight, g.mol-1
Crystal system
Space group, no.
A, Å
B, Å
C, Å
V, Å3
Θ range unit cell: min.-max., deg; reflections
Formula_Z
Space group_Z
Z’ (= Formula_Z / Space group_Z)
ρcalc, g.cm-3
F(000), electrons
µ(Mo K α ), cm-1
Color, habit
Approx. crystal dimension, mm
Rb(H2O)FeMn(CN)6
370.37
cubic
F-43m, 216
10.521(2)
10.521(2)
10.521(2)
1164.6(4)
3.35 - 29.76 ; 1705
4
96
1/24
2.112
704
64.68
Dark brown, cube
0.31 x 0.28 x 0.21
Table II.1b Data collection
Radiation type; λ, Å
Monochromator
Measurement device type
Detector Area resolution (pixels / mm)
Temperature, K
Measurement method
θ range; min. max., deg
Index ranges
Min.- Max. absorption transmission factor
X-ray exposure time, h
Total data
Unique data
Data with criterion: (Fo ≥ 4.0 σ (Fo))
∑ [|Fo2 - Fo2 (mean)|] / ∑ [Fo2]
Rsig = ∑ σ(Fo2) / ∑ [Fo2]
Rint =
Mo K α , 0.71073
Graphite
Bruker SMART APEX
CCD area-detector diffractometer
4096 x 4096 / 62 x 62 (binned 512)
293(1)
ϕ- and ω-scans
3.35, 28.22
h: -12→14; k: -13→14; l: -13→14
0.1821 – 0.2733
7.8
2657
181
170
0.0231
0.0116
143
Appendix II – Crystallographic Details of the Crystal Structure Described in Chapter 5
Table II.1c Refinement
Number of reflections
Number of refined parameters
Isotropic secondary-extinction coefficient, g
Final agreement factors:
wR(F2) = [ ∑ [w(Fo2 - Fc2)2] /
Weighting scheme: a, b
∑ [w(Fo2)2]]1/2
181
16
0.0024(11)
0.1004
0.0596, 1.7609
W = 1/[σ (Fo ) + (aP) + bP]
And P = [max(Fo2,0) + 2Fc2] / 3
2
R(F) =
2
2
∑ (||Fo| - |Fc||) / ∑ |Fo |
For Fo > 4.0 σ (Fo)
0.0348
Absolute-Structure parameter Flack's x
GooF = S = [ ∑ [w(Fo - Fc ) ] / (n-p)]
n = number of reflections
p = number of parameters refined
Residual electron density in final
Difference Fourier map, e/Å3
Max. (shift/σ) final cycle
Mean. (shift/σ) final cycle
2 2
2
1/2
0.69(11)
1.287
-0.9, 0.8(2)
<0.001
<0.001
Table II.2 Final fractional atomic coordinates and equivalent isotropic displacement
parameters with s.u.'s in parentheses for RbMn[Fe(CN)6]•H2O.
Atom
Fe
Mn
N
C
Rb1a
Rb2b
O1b
O2a
X
0
0
0
0
¼
¾
¼
¾
*) Ueq = 1/3 ∑ i ∑ jUijai*aj*ai.aj
[a] Indicates a s.o.f. of 0.7512(4)
[b] Indicates a s.o.f. of 0.2488(4)
y
0
0
0
0
¼
¾
¼
¾
Z
0
½
0.2904(5)
0.1828(4)
¼
¾
¼
¾
Ueq (Å2)*
0.0170(3)
0.0149(3)
0.0493(11)
0.0281(9)
0.1340(12)
0.210(5)
0.1340(12)
0.210(5)
Table II.3 Anisotropic (displacement) parameters (Å2) for RbMn[Fe(CN)6]•H2O
Fe
Mn
N
C
Rb1
Rb2
O1
O2
U11
0.0170(6)
0.0149(6)
0.0585(18)
0.0326(13)
0.134(2)
0.210(9)
0.134(2)
0.210(9)
U22
0.0170(6)
0.0149(6)
0.0585(18)
0.0326(13)
0.134(2)
0.210(9)
0.134(2)
0.210(9)
U33
0.0170(6)
0.0149(6)
0.022(2)
0.0192(18)
0.134(2)
0.210(9)
0.134(2)
0.210(9)
Thermal vibration amplitudes (Å2)
3
F(h) = Fo(h) exp (-2π2
3
Σ Σ hihIa aj *Uij)
i
*
i=1 j=1
or
F(h) = Fo(h) exp (-8π2Uiso(sin(θ)/λ)2)
144
U23
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
U13
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
U12
0.0(-)
0.0(-)
0.001(4)
0.004(3)
0.0(-)
0.0(-)
0.0(-)
0.0(-)
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
II.4 References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Brucker, SMART, SAINTPLUS and XPREP Area Detector Control and Integration Software
Smart Apex Software Reference Manuals, Bruker Analytical X-ray Instruments Inc.: Madison,
Wisconsin, USA, 2000
Sheldrick, G.M. SADABS Multi-Scan Absorption Correction Program, version 2.03, University
of Göttingen: Germany, 2001
Spek, A. L., J. Appl. Crystallogr. 1988, 21, 578-579
Moritomo, Y., Kato, K., Kuriki, A., Takata, M., Sakata, M., Tokoro, H., Ohkoshi, S.-I.,
Hashimoto, K., J. Phys. Soc. Jpn. 2002, 71, 2078-2081
Moritomo, Y., Kato, K., Kuriki, A., Takata, M., Sakata, M., Tokoro, H., Ohkoshi, S.-I.,
Hashimoto, K., J. Phys. Soc. Jpn. 2003, 72, 2698-2698
Flack, H. D., Acta Cryst. 1983, A39, 876-881
Flack, H. D., Bernardinelli, G., Acta Crystallogr. 1999, A55, 908-915
Flack, H. D., Bernardinelli, G., J. Appl. Crystallogr. 2000, 33, 1143-1148
Herbst-Irmer, R., Sheldrick, G. M., Acta Crystallogr. 1998, B54, 443-449
Wilson, A. J. C. International tables for crystallography, Kluwer Academic Publishers:
Dordrecht, The Netherlands, 1992, Vol. C.
Sheldrick G.M., SHELXL-97. Program for the Refinement of Crystal Structures, University of
Göttingen: Germany, 1997
Meetsma, A., PLUTO. Molecular Graphics Program, University of Groningen: The
Netherlands, 2006
Spek, A.L., PLATON. Program for the Automated Analysis of Molecular Geometry (A
Multipurpose Crystallographic Tool), version August 2006, University of Utrecht: The
Netherlands Spek, A.L., J.Appl. Crystallogr. 2003, 36, 7.
Duisenberg, A. J. M., J. Appl. Cryst. 1992, 25, 92-96
145
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Nederlandse Samenvatting
Ieder element dat bestaat heeft een specifiek aantal protonen (positief geladen
deeltjes in de kern van het atoom) en electronen (negatief geladen deeltjes die in
een wolk om de kern heen zweven).
Voor ieder electron dat om de kern heen zweeft, is er een specifiek gebied waar het
electron zich mag bevinden. Deze specifieke gebieden worden orbitalen genoemd.
Ieder orbitaal dat om een atoomkern aanwezig is kan maximaal twee electronen
bergen: een electron dat rechtsom draait en een electron dat linksom draait. De
draaiing van de electronen wordt spin genoemd en de waarde van de spin is -½ (↑)
voor een electron dat de ene kant opdraait en +½ (↓) voor een electron dat de
andere kant opdraait. Indien een orbitaal twee electronen bergt is de totale spin dus
0. Als er echter maar één electron in het orbitaal aanwezig is, is de spin ½. Indien
een atoom/ion orbitalen heeft die half gevuld zijn, heeft dat atoom/ion een
electronen spin. Spin leidt tot een magnetisch moment op de volgende manier:
χT =
g e2 S ( S + 1)
8
Waarbij S de totale spin van het atoom/ion is, ge is een constante die g-factor wordt
genoemd, χ is de magnetische susceptibiliteit en T is de temperatuur. Hoe groter
dus de spin van een deeltje, hoe groter de magnetische susceptibiliteit.
Sommige orbitalen liggen laag in energie en andere orbitalen liggen veel hoger in
energie (vergelijk dit met een ladder die steeds hoger gaat). Afhankelijk van het
aantal electronen en de energie dat het atoom heeft zijn de orbitalen gevuld.
Het vullen van de orbitalen volgt twee algemene regels:
1. De orbitalen die het laagste in energie liggen worden als eerste gevuld.
2. Indien meerdere orbitalen dezelfde energie hebben, gaan electronen eerst
orbitalen half vullen voordat een orbitaal volledig wordt gevuld. Op deze
manier wordt de repulsie tussen de electronen geminimaliseerd.
Orbitaal 2, 3, 4
Figuur 1. Opvulling van orbitalen met
electronen in een atoom met 5 electronen
en 4 orbitalen.
Orbitaal 1
Stel je voor dat we een atoom hebben met 5 electronen en 4 orbitalen (zie figuur 1).
Het eerste orbitaal ligt veel lager in energie dan de andere drie orbitalen die alle
drie dezelfde energie hebben. Het eerste electron zal in orbitaal 1 gaan zitten,
immers deze is veel lager in energie. Het tweede electron zal dit ook doen, maar
met een tegenovergestelde spin. Het derde electron past niet meer in orbitaal 1 en
147
Nederlandse Samenvatting
zal dus in orbitaal 2 moeten gaan (of orbitaal 3 of 4, want ze hebben immers
allemaal dezelfde energie). Het vierde electron zal in orbitaal 3 gaan (of 4, maar
niet in 2, vanwege de repulsieve krachten). Het laatse electron gaat vervolgens in
orbitaal 4. Op deze manier hebben we 3 halfgevulde orbitalen verkregen en de
totale spin van dit atoom is dus 3/2 en dit atoom zal een magnetisch moment
hebben.
a)
b)
Figuur 2. Opvullen van orbitalen met electronen in een atoom met 5 electronen en 5
orbitalen waarbij het energieverschil klein is (a) en waarbij het energieverschil groot is (b).
Orbitaal 1, 2 en 3 liggen laag in energie en orbitaal 4 en 5 liggen hoger in energie.
Stel je nu eens voor dat we een atoom hebben met 5 orbitalen en 5 electronen (zie
figuur 2). Van de 5 orbitalen zijn er 3 orbitalen die laag in energie liggen en 2 die
hoger in energie liggen, echter de afstand tussen de energieniveaus is niet heel erg
groot. De eerste 3 electronen zullen de onderste 3 orbitalen ieder half vullen. Het
vierde (en het vijfde) electron kunnen kiezen:
1. In het vierde (of vijfde) orbitaal, maar dan moet het het energieverschil
tussen de orbitalen overbruggen.
2. In het eerste (of tweede) orbitaal, maar dan heeft het repulsieve krachten
met electron 1 (of 2).
Afhankelijk van het precieze energieverschil tussen de orbitalen krijg je in dit geval
of optie 1 (klein energieverschil, figuur 2a) of optie 2 (groot energieverschil, figuur
2b). Optie 1 heeft een totale spin van 5/2 en wordt een high-spin complex genoemd,
terwijl optie 2 een totale spin heeft van ½ en daarom low-spin complex heet. De
beide complexen hebben ook een verschillende waarde voor de magnetische
susceptibiliteit. De precieze afstand tussen de orbitalen hangt af van de omgeving
van het atoom.
De meeste moleculen die een centraal metaalion hebben zijn of high-spin of lowspin complexen.1 Het is echter mogelijk dat het energieverschil tussen de orbitalen
van dezelfde grootte is als de repulsieve krachten tussen de electronen. In dit geval
kun je een spin-crossover complex krijgen: bij lage temperatuur (en dus weinig
energie voor de electronen) is het een low-spin complex en bij hoge temperatuur
(en dus meer energie voor de electronen) is het een high-spin complex. Deze
complexen krijgen dan, afhankelijk van de temperatuur, verschillende waardes
voor de magnetische susceptibiliteit en hebben op die manier twee toestanden een
‘1’ en een ‘0’. Als de overgang van low-spin naar high-spin bij een andere
148
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
temperatuur gebeurd dan van high-spin naar low-spin is er een temperatuur gebied
ontstaan waarbij de twee toestanden beiden kunnen bestaan en is er een soort
geheugen element gevormd.
In dit proefschrift is er gekeken naar de magnetische eigenschappen van
RbxMn[Fe(CN)6]y·zH2O verbindingen (figuur 3). In deze verbindingen zitten twee
metaalionen (mangaan, Mn en ijzer, Fe). De twee metaalionen zijn verbonden via
een brug van cyanide (CN): de C zit vast aan het ijzer ion en de N zit vast aan het
mangaan ion. In de “vakjes” tussen de metaalionen en de cyanides zitten de
rubidium ionen (Rb) en watermoleculen (H2O).
Figuur 3. De eenheidscel van
RbxMn[Fe(CN)6]y·zH2O verbindingen. Deze
verbinding strekt in alle dimensies uit.
In dit materiaal zijn zowel ijzer als mangaan aanwezig die allebei de mogelijkheid
hebben om zowel low-spin als high-spin te zijn. De spin overgang in het materiaal
gebeurt op een aparte manier: ze wisselen een electron uit, ze hebben een interne
redoxreactie (zie figuur 4). Bij lage temperatuur (de lage temperatuur fase) heeft
FeII 6 electronen en is het low-spin en heeft MnIII 4 electronen en is high-spin. Bij
warmte,
groen
FeII
CN
-
e
MnIII
FeIII
kou,
violet
MnII
CN
eeg
t2g
Figuur 4. De interne redoxreactie die plaats vindt in verbindingen van
RbxMn[Fe(CN)6]y·zH2O. De onderste helft van de figuur geeft de opvulling van de
orbitalen weer.
149
Nederlandse Samenvatting
hoge temperatuur (de hoge temperatuur fase) gaat er een electron van ijzer naar
mangaan en krijgen we FeIII met 5 electronen en low-spin en MnII met 5 electronen
en high-spin. Dit proces is reversibel en kan meerdere malen worden herhaald.
Deze interne redoxreactie kan ook worden veroorzaakt door het materiaal te
beschijnen met licht van een bepaalde kleur: met groen gekleurd licht (532 nm)
gaat de verbinding van de lage naar de hoge temperatuur fase en met violet
gekleurd licht (410 nm) gaat de verbinding van de hoge naar de lage temperatuur
fase.
De variatie van de magnetische susceptibiliteit met temperatuur staat in figuur 5.
Als de verbinding verwarmd wordt blijft de susceptibiliteit rond 3.0 cm3 K mol-1
tot aan ongeveer 300 K (27oC), waarna hij omhoog schiet naar 4.75 cm3 K mol-1.
Bij afkoelen gebeurd het omgekeerde bij ongeveer 225 K (-48oC). Er is dus een
zeer groot temperatuurgebied waarbij de verbinding zowel hoge temperatuur als
lage temperatuur fase kan zijn. Dit heet een hysterese.
Figuur 5. De variatie van de magnetische
susceptibiliteit met temperatuur voor
RbMn[Fe(CN)6]. De figuur komt uit
referentie 2.
Het onderzoek in dit proefschrift is erop gericht om uit te vinden hoe het
mechanisme van de interne redoxreactie in RbxMn[Fe(CN)6]y·zH2O verbindingen
precies werkt. Verder is er gekeken of de interne redoxreactie kan worden verfijnd
zodat het temperatuurbereik van de hysterese kan worden veranderd naar iedere
gewenste temperatuur.
In hoofdstuk 2 is er gekeken naar de invloed van [Fe(CN)6] defecten op de interne
redoxreactie. Het blijkt dat hoe minder defecten er zijn hoe completer de interne
redoxreactie en hoe smaller de hysterese. Dit kan als volgt worden begrepen: als er
een [Fe(CN)6] defect is (zie figuur 6), wordt mangaan niet meer omringd door 6
cyanides maar 5 cyanides en één water molecuul. Het redoxpotentiaal van
mangaan verandert daardoor op zo’n manier dat het niet meer in staat is om de
interne redoxreactie uit te voeren. Als er te veel defecten zijn, komen er teveel
mangaan ionen die dit niet meer kunnen en daardoor wordt de volledige verbinding
inactief. De meest complete interne redoxreactie vindt dus plaats als
Rb:Mn:Fe = 1:1:1.
150
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
In Hoofdstuk 3 wordt omschreven hoe de specifieke synthesemethode gevolgen
heeft voor de verhouding van Rb:Mn:Fe in het gevormde materiaal (de
stoichiometrie). In principe wordt het materiaal gemaakt door een oplossing van
MnCl2 toe te voegen aan een gemengde oplossing van K3[Fe(CN)6] en RbCl. Bij
dit toevoegen wordt instantaan een bruin poeder gevormd. Het blijkt dat de
temperatuur tijdens de synthese weinig invloed heeft op de stoichiometrie,
alhoewel de temperatuur niet boven de 85oC moet uitkomen omdat er dan dingen
gebeuren in de verbindingen waarvan niet duidelijk is wat het is. De meeste
invloed op de stoichiometrie heeft de toevoegsnelheid van de oplossing waar
mangaanionen inzitten aan de oplossing van ijzerhexacyanide- en rubidiumionen.
Hoe langzamer dit gaat, hoe beter de stoichiometrie wordt: het systeem heeft
voldoende tijd om alle watermoleculen die oorspronkelijk om mangaan zaten te
vervangen door de cyanides die aan ijzer vastzitten.
N C
C
N
N C
N
C
C N
C
C N
N
C
C N
C
N
N
N
C
N
C N
H
N C
N
N
H
N
C
Figuur 6. Een [Fe(CN)6] defect in
RbxMn[Fe(CN)6]y·zH2O. De Rb ionen en
water moleculen tussen de metaal ionen
en cyanides in is weggelaten voor de
duidelijkheid.
C
N
H
H
N
H
O
C N
N
C
O
C
N
H
H
Fe
C
O
O
N
C
C
N C
H
C
N
C
C N
C
Mn
N
N
C
N C
N C
C
N
C
N C
C
N
Defect of
Fe(CN)6
Omdat de watermoleculen van grote invloed zijn op de eigenschappen van de
RbxMn[Fe(CN)6]y·zH2O verbindingen is er in Hoofdstuk 4 bekeken of de
uitwisseling met water moleculen uit de lucht invloed heeft op het veranderen van
de interne redoxreactie met de tijd. Afhankelijk van de precieze synthese heeft dit
veel invloed (indien de synthese is uitgevoerd met een mengsel van water en
methanol) of relatief weinig invloed (indien de synthese is uitgevoerd met alleen
maar water als oplosmiddel). Verder blijkt ook dat het 57Fe Mössbauer spectrum bij
kamertemperatuur een directe aanwijzing geeft of en in welke mate een interne
redoxreactie kan plaatsvinden.
Voor veel karakterisatie- en meettechnieken is het handiger om kristallen te hebben
dan poedervormige materialen. In Hoofdstuk 5 wordt beschreven hoe voor de
eerste keer kristallen van RbMn[Fe(CN)6]·H2O zijn verkregen. Deze kristallen
vertonen de interne redoxreactie onder invloed van temperatuur en van licht. In
tegenstelling tot wat kan worden verwacht op basis van de stoichiometrie vertonen
de kristallen slechts een overgang van 50%. De precieze verdeling van de rubidium
ionen en watermoleculen over de ruimtes tussen de metaalionen en cyanides in, is
151
Nederlandse Samenvatting
een mogelijke verklaring hiervoor. Hoofdstuk 6 gaat dieper in op de karakterisatie
van deze kristallen.
Hoofdstuk 7 laat zien in hoeverre andere ionen dan rubidium kunnen worden
opgenomen in de ruimtes tussen de metaalionen en de cyanides. Kalium,
ammonium, tetramethylammonium en strontium worden slecht opgenomen.
Natrium en cesium worden wel goed opgenomen, maar de gevonden verbindingen
vertonen geen interne redoxreactie.
1
2
152
In dit geval wordt alleen maar naar de bovenste 5 orbitalen gekeken (de d-orbitalen),
alle lager gelegen orbitalen zijn volledig gevuld en worden genegeerd
Ohkoshi, S.-I.; Tokoro, H.; Utsunomiya, M.; Mizuno, M.; Abe, M.; Hashimoto, K.,
J.Phys.Chem.B, 2002, 106, 2423
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
Dankwoord
Aan het einde van dit proefschrift en het einde van mijn promotietraject wil ik een
aantal mensen bedanken zonder wie het niet mogelijk was dit geheel te volbrengen.
Allereerst wil ik Petra van Koningsbruggen bedanken voor het initiëren van dit project,
het aandragen van mogelijke invalshoeken en het mij volledig vrij laten in mijn
onderzoek. Petra, veel succes met de volgende stappen in je carrière. Ook Bart Hessen
wil ik bedanken; we hebben elkaar weliswaar niet al te veel gezien, maar de weinige
keren dat we hebben gesproken en vooral gemaild kwamen er altijd nuttige suggesties
naar voren. Bart, ook jij succes met je volledig nieuwe carrière.
Vervolgens wil ik de leden van de leescommissie bedanken voor hun grondige analyse
van dit proefschrift en de snelle afhandeling: thank you Anne Bleuzen, Paul van
Loosdrecht (ook voor het optreden als een soort van promotor op de achtergrond) and
Thom Palstra.
The project I have worked on was not a single research project on my side, but was part
of a larger project between the groups of Ria Broer, Paul van Loosdrecht, Petra Rudolf
and Petra van Koningsbruggen. I would like to thank Audrius, Enrico, Fillipo, Javier,
Luminita, Paul, Petra, Régis, Ria and Tom for the nice cooperation and the many
useful discussions on all the results.
Due to this and other cooperations some of the results presented in this thesis have
been measured by people other than me: All syntheses have been done by myself,
except for the synthesis of sample 4.1 which was done by the group of Azzedine
Bousseksou. All measurements on this sample have been performed by them as well.
All X-ray powder diffraction profiles have been measured and analysed by myself, but
I would like to thank Jacob Baas for helping me with the handling of the hardware and
Graeme Blake for helping me interpreting some of the profiles. Even though my
undergraduate research consisted solely on Raman spectroscopy, none of the Raman
spectra in the present thesis have been measured by myself: the Raman spectra from
chapter 2 have been measured by Audrius Pugzlyz, those in chapter 4 by the group of
Azzedine Bousseksou and those in chapter 5 by Tom Lummen. The 57Fe Mössbauer
spectra in chapter 2 have been measured by Thomas Bakas, those in chapter 5 by
Gabor Molnár and the rest of them by myself. I would especially like to thank Gabor in
this respect with help on the interpretation of all 57Fe Mössbauer spectra. All magnetic
measurements in this thesis have been measured by myself, but I would like to thank
Mylène Sage and Jacob Baas for teaching me how to use the MPMS machine and
helping me in case of troubles with the hardware. The DSC measurement in chapter 2
has been measured by Miguel Castro and those in chapter 4 by the group of Azzedine
Bousseksou. The XPS measurements in chapter 2 have been measured by Enrico
Maccallini and those in chapter 6 by Régis Gengler. Even though none of the XPS
measurements done by myself have made it to the thesis I would still like to thank
Régis, Fillipo and Fabrizio for teaching me how to perform an XPS measurement. The
crystal structure presented in chapter 5 has been elucidated by Auke Meetsma. Auke,
dank je dat je altijd klaar stond om antwoord te geven op vragen en als we weer eens
153
Dankwoord
iets hadden bedacht dat mogelijk zichtbaar was in de structuur. Ik hoop dat we niet al te
vervelend waren in dit opzicht. Marco Bouwkamp has helped me with developing the
model proposed in chapter 5. The ESR spectra in chapter 6 have been measured by
Agnes Antal and Andras Jánossy. I would like to thank Wesley Browne for answering
the questions I had on ESR spectroscopy. The thermal conductivity measurements in
chapter 6 have been performed by Nikolai Hlubek, Christian Hess and Bernd Büchner.
Wesley Browne has thought me how to perform CV measurements. Unfortunately,
none of these measurements are to be found in this thesis. The scanning electron
microscopy pictures on the cover of this thesis (who said the BORG of Star Trek
doesn’t exist?) have been measured by Gert ten Brink.
During the four/five years of my PhD many people have been present at the labrooms
of the first floor of building 18. I would like to thank all of these people for many
coffee/teabreaks, lunches and “fruithapjes”. We have discussed many things during
these breaks, both useless and useful. Too many people have been present to thank you
all personally in this respect, but I would like to thank the rest of the “van
Koningsbruggen group” Eddy, Heloïse and Andy for sharing the same
“responsibilities”. Kees, jij hebt je bachelor onderzoek gedaan bij mij. Jouw resultaten
hebben geleid tot het vervolgonderzoek beschreven in hoofdstuk 3. Succes met jouw
verdere chemische carriere. Martijn ook jij bedankt dat je hebt “meegespeeld” met het
bouwen van de Legorobot, ondanks dat je eigenlijk naar een groen of rood prutje had
moeten kijken. Helaas zijn de resultaten behaald met deze robot niet terug te vinden in
dit proefschrift. Ook een speciaal woord van dank aan Marco Bouwkamp die altijd
klaar stond als ik weer een luisterend oor nodig had bij (niet zulke) chemische
problemen.
Buiten het lab heb ik gelukkig ook nog tijd gehad voor een sociaal leven. In verband
hiermee wil ik meerdere groepen van mensen bedanken voor de broodnodige
afwisseling. Ten eerste het orkest Harmonie’67 en met name de (steeds maar weer
wisselende) saxsectie en het bestuur. Het was en is altijd heel fijn om op de
maandagavond samen muziek te maken en even al het andere te vergeten. Ook
Marjolijn en Sarah wil ik bedanken voor de wekelijkse etentjes en het vele
bioscoopbezoek. Marion, Marije, Liselotte, Petra en Trieneke, jullie ook bedankt voor
alle fijne mailconversaties en de vele bijkletsweekendjes, wat mij betreft moeten we
weer eens een andere stad bezoeken! Verder de vijfdejaars voor de jaarlijkse
weekendjes weg, vele weekenden schilderen (zijn jullie nu eens klaar met verhuizen?)
en het oud en nieuw vieren, maar dan vooral niet met oud en nieuw.
Tenslotte wil ik mijn ouders, Mirjam en Arno en Henk bedanken voor het altijd in mij
blijven geloven, trots te blijven en mij te steunen “ook al snap ik er geen ene reet van”.
Het was fijn om te weten dat er altijd mensen achter je blijven staan.
Als aller, allerlaatste (maar zeker niet het allerminste) wil ik Peter heel erg bedanken
voor de dagelijkse steun en begrip als ik er weer eens helemaal doorheen zat. Peter ik
hou van je en ik hoop dat we nog heel veel jaren samen zullen blijven.
154
Electron Transfer Properties in the Prussian Blue Analogues RbxMn[Fe(CN)6] y·zH2O
List of Publications
Characterisation of Single Crystals of RbMn[Fe(CN)6]·H2O Vertelman, E.J.M.;
Gengler, R.; Antal, A.; Hlubek, N.; Molnar, G.; Hess, C.; Bousseksou, A.; Rudolf,
P.; Jánossy, A.; Büchner, B.; Koningsbruggen, P.J. van; 2009, In preparation
Interplay Between the Charge Transport Phenomena and the Charge-Transfer
Phase Transition in RbxMn[Fe(CN)6]y·zH2O Molnar, G.; Cobo, S.; Mahfoud, T.;
Vertelman, E.J.M.; Koningsbruggen, P.J. van; Demont, P.; Bousseksou, A.;
J.Phys.Chem. C 2009, 113, 2586
Valence-Tautomeric RbMnFe Prussian Blue Analogues: Composition and Time
Stability Investigation Salmon, L.; Vertelman, E.J.M.; Murgui, C.B.; Cobo, S.;
Molnár, G.; Koningsbruggen, P.J. van; Bousseksou, A.; Eur. J. Inorg. Chem. 2009,
760
Bulk and Surface Switching in Mn-Fe-Based Prussian Blue Analogues Lummen,
T.T.A.; Gengler, R.Y.N.; Rudolf, P.; Lusitani, F.; Vertelman, E.J.M.;
Koningsbruggen, P.J. van; Knupfer, M.; Molodtsova, O.; Pireaux, J.-J.;
Loosdrecht, P.H.M. van; J. Phys. Chem. C 2008 112, 14158
Prediction of the Equilibrium Structures and Photomagnetic Properties of the
Prussian Blue Analogue RbMn[Fe(CN)6] by Density Functional Theory Luzon, J.;
Castro, M.; Vertelman,E.J.M.; Gengler, R.Y.N.; Koningsbruggen, P.J. van;
Molodtsova, O.; Knupfer, M.; Rudolf, R.; Loosdrecht, P.H.M. van; Broer, R.; J.
Phys. Chem. A 2008, 112, 5742
Light- and Temperature-Induced Electron Transfer in Single Crystals of
RbMn[Fe(CN)6]·H2O Vertelman, E.J.M.; Lummen, T.T.A.; Meetsma, A.;
Bouwkamp, M.W.; Molnar, G.; Loosdrecht, P.H.M. van; Koningsbruggen, P.J.
van; Chem. Mater. 2008, 20, 1236
Crystal Structure and Magnetic Behaviour of a Five-Coordinate Iron(III) Complex
of Pyridoxal-4-Methylthiosemicarbazone Yemeli Tido, E.W.; Vertelman, E.J.M.;
Meetsma, A.; Koningsbruggen, P.J. van; Inorg. Chim. Acta 2007, 360, 3896
The Influence of Defects on the Electron-Transfer and Magnetic Properties of
RbxMn[Fe(CN)6]y·zH2O Vertelman, E.J.M.; Maccallini, E.; Gournis, D.; Rudolf, P.;
Bakas, T.; Luzon, J.; Broer, R.; Pugzlyz, A.; Lummen, T.T.A.; Loosdrecht, P.H.M.
van; Koningsbruggen, P.J. van; Chem. Mater. 2006, 18, 1951
155