Final Report: Microporous Magnets for the Room Temp Separation of Oxygen from Air
T. David Harris, Department of Chemistry
Overview
The ISEN booster award has enabled my research group to synthesize ligand radical-supported molecular complexes
that serve as models for microporous frameworks. We have demonstrated that record magnetic coupling between
paramagnetic metal ions and the azophenine radical ligand can be attained. These results are now guiding our efforts to
incorporate similar metal-ligand combinations into frameworks, with the goal of employ these compounds in the magnetic
extraction of dioxygen from air.
Introduction
The separation of oxygen from air represents a major industrial process, with 100
million tons of O2 extracted from air annually. 1 Moreover, this number is expected
to rapidly grow, owing in large part to the need for pure O2 to increase energy
efficiency in both pre- 2 and post-combustion 3 CO2 capture in coal-burning power
plants. Currently, cryogenic distillation is by far the most widely used process for air
separation, and it is the only process capable of high-capacity production and
generation of high-purity gases. 4 Despite the high performance of cryogenic
distillation, an enormous amount of energy must be expended in order to reach
temperatures necessary to refrigerate the air. One possible alternative to distillation
is to separate the components air through porous magnetic media. Here, as the air
passes through the pores of the magnet, the paramagnetic S = 1 O2 molecules are
attracted to the pore walls, while the diamagnetic components, such as N2 and argon,
are repelled through the material. For magnetic air separation to be energy Figure 1. Top: MOF comprised of 1D
economical, we must first synthesize porous materials that behave as magnets at or inorganic chains (black) connected by
near room temperature. Several experimental 5,6 and theoretical 7 studies have focused organic bridges (blue). Bottom: Ligand
redox chemistry can be used to generate
on membranes embedded with powdered neodymium5 or superparamagnetic strong direct magnetic coupling between
6
nanoparticles, however this strategy suffers from a number of inherent problems, paramagnetic inorganic chains and radical
such as dilute and inhomogeneous distribution of magnetic particles, in addition to a organic bridging ligands.
lack of synthetic control at the atomic level.
An ideal solution would be to simply synthesize a homogenous material that is both permanently porous and magnetic
at room temperature. Currently, however, no porous magnet has been shown to exhibit an ordering temperature above 219
K, 8 with the vast majority ordering well below 100 K. 9 These low ordering temperatures result from the fact that porous
magnetic materials generally feature paramagnetic inorganic nodes connected to one another by many-atom, diamagnetic
bridging ligands. As such, the paramagnetic inorganic centers can then only couple with one another through the bridging
ligand via weak, indirect superexchange. Since magnetic ordering temperature is directly correlated to the strength of
magnetic coupling between spin centers, the resulting compound only behaves as a magnet at low temperature. As an
alternative, the use of paramagnetic organic linkers can lead to strong, direct coupling pathways between inorganic and
organic units. In particular, metal-organic framework (MOF) materials are ideally suited for this challenge, as they
combine the attributes of permanent porosity and facile synthetic tunability. We are working to synthesize MOFs
comprised of paramagnetic inorganic components connected to one another via redox-active organic bridging ligands.
Here, simple redox chemistry can then be carried out on the bridging ligand to give an S = 1/2 organic radical unit (Fig 1).
Direct through-bond interactions between inorganic and organic components will lead to extremely strong magnetic
coupling and thus ordering temperatures at unprecedented high temperatures. Our initial work has focused on the
synthesis and study of molecular complexes that serve as models for these magnetic MOFs.
Results
Our preliminary studies have focused on model complexes of bulky 2,5-diamino-1,4-diiminobenzoquinone ligands
(NRLH2), wherein one NRL2− ligand bridges two Scheme 1
[L′MII]n+ units, to give the dinuclear complex
[(L′MII)2(NRL2−)]n+ (Scheme 1). Here, ′ Lis a
capping ligand that serves to block the formation
of an extended chain. This general synthetic
strategy is established, having previously been
employed in the formation of dinuclear complexes bridged by
dihydroxybenzoquinone 10- 15 derivatives and by NRL2− 16- 19 itself. Note,
however, that NRL2− has only been studied as a bridging ligand between
diamagnetic metal ions or S = 1/2 CuII centers, with no investigations of
coupling between metal and ligand radical.
Using the phenyl-substituted molecule NPhLH2 20,21 as a bridging
ligand12,13,22 and tris(2-pyridylmethyl)amine (TPyA) as a capping ligand, we
have synthesized the series of dinuclear complexes [(TPyA)2MII2(NPhL2−)]2+
(M = Mn (1), Fe (2), Co (3), Ni (4)) according to the synthetic procedure
depicted in Scheme 1. The crystal structures of 1-4 have been determined for
all complexes using single-crystal X-ray diffraction, and they are
isostructural, as exemplified by complex 2 (Fig 2 upper). In the case of M =
V, Cr, Cu, and Zn, analogous reactions invariably produce insoluble
precipitates upon introduction of NPhL2− to “[(TPyA)M]2+”. Solid-state IR
studies of these products indicate the absence of TPyA and protonated amines
and the presence of NPhL2−, suggesting the formation of extended solids. Work
is underway both to vary capping ligand and to crystallize these insoluble
products. Finally, we have observed that subtle modification of the bridging
ligand aryl substituent can lead to important electronic differences in the
resulting dinuclear complex. For instance, carrying out the complex forming
reaction for M = Fe, with the 2-methylphenyl (MePh) groups in place of
phenyl groups on the bridging ligand, leads to crystallization of the complex
[(TPyA)2FeIII2(NMePhL2−)]4+ (5). The one-electron oxidation of each FeII center
to FeIII is unexpected, given the very similar redox potentials of 2 and 5
(Table 1). Likely, the oxidation is occurring upon crystallization and is the
result of crystal-packing effects associated with the MePh substituents, which
may slightly shift the redox potential and enable a spontaneous, solventassisted oxidation.
To assess the presence of magnetic coupling interactions in 1-5, variabletemperature dc magnetic susceptibility data (Fig 5 lower) were collected and
used to construct plots of χMT vs. T, where χM is the molar dc magnetic
susceptibility and χMT is proportional to effective magnetic moment. As
expected, these data reveal only weak antiferromagnetic (1,2-5) or
ferromagnetic (3) superexchange coupling in each complex, occurring
between paramagnetic metal centers through the diamagnetic bridging ligand,
with fits to the data giving exchange coupling constants of J = −1.95 (1),
−2.84 (2), −2.59 (4), and −1.01 (5) cm−1 (Table 1). These values are
consistent with coupling constants previously obtained for dinuclear
complexes supported by substituted benzoquinonoid bridging ligands.10-15
Cyclic voltammograms collected for 1-4 reveal a rich electrochemistry, with
the complexes exhibiting up to three fully reversible redox events (Fig 3
upper; Table 1). Of particular interest is the couple found at ca.−1.65 V vs.
Fc/Fc+ for all complexes. The invariance of potential as a function of metal
ion for this process suggests that it is ligand-centered, corresponding to the
NPh 2− NPh 3−•
L / L couple. Accordingly, this result suggests that a MII2(NPhL3−•)
composition should be chemically accessible for all complexes. Toward this
end, the Fe congener 2 was treated with one equivalent of the strong reductant
(C5Me5)2Co (−1.91 V vs. Fc/Fc+ in MeCN) 23 in MeCN to give an intense blue
solution. Subsequent diffusion of diethyl ether vapor into this solution at−3 5
°C gave crystals of the one-electron reduced complex [(TPyA)2M2(NPhL)]+ (6;
Fig 2 lower). Close comparison of the bond lengths in 2 and 6 reveals several
key features (Table 2). First, the average benzoquinone C-C distance
decreases from 1.431(5) to 1.417(5) Å in moving from 2 to 6. In contrast, the
average benzoquinone C-N distance increases from 1.331(4) to 1.362(4) Å.
These differences reveal a net increase in C-C bond order and net decrease in
C-N bond order, consistent with the additional electron occupying a molecular
Figure 2. Reduction of 2 with (C5Me5)2Co to 6.
Orange, blue, and gray ellipsoids represent Fe, N,
and C atoms, respectively; H atoms are omitted for
clarity. See Table 2 for bond lengths.
Figure 3. Upper: Cyclic voltammograms for 1-4
(Table 1). CV for 5 is similar to that of 2. Lower:
Dc susceptibility data for 1-5. Black lines
correspond to fits to the data (Table 1).
Table 1. Selected Electrochemicala and Magnetic Datab for 1-6
∆E1/2 (1)
∆E1/2 (2)
∆E1/2 (3)
S
J
Ueff
1
0.12
−1.75
−
0
−1.95(1)
−
2
0.32
−0.19
−1.65
0
−2.84(2)
−
3
0.02
−0.16
−1.64
3
4
0.19
−1.80
−
0
−2.59(1)
−
5
0.42
−0.15
−1.69
0
−1.01(1)
−
6
−
−
−
≤900
50(1)
a
7
/2
+
Potentials are referenced to the Cp2Fe/Cp2Fe couple and are expressed in V. Only
reversible waves are reported. bAll energies are expressed in cm-1.
Figure 4. Dc magnetic susceptibility data for 6.
orbital of ligand character, i.e. a ligand based reduction to give NPhL3−•. In further support of this hypothesis, the average
Fe-NL distance decreases from 2.110(3) to 2.074(3) Å, consistent with a stronger Fe-N interaction resulting from a more
reduced bridging ligand. Moreover, these structural changes are similar to those previously observed for a ligand-centered
reduction in a dinuclear dihydroxybenzoquinone complex.15 Finally, the average Fe-NTPyA distance changes only slightly,
increasing from 2.234(3) to 2.249(3) Å, support against a metal-centered redox event. Taken together, these observations
lead us to propose a configuration for 6 of [(TPyA)2Fe2(NPhL3−•)]+. Mössbauer spectroscopic measurements will confirm
this assignment and are underway.
Having established the ability to generate a paramagnetic bridging ligand for at least one member of the series of
dinuclear complexes, we turned our attention to examining the magnetic behavior engendered through interactions
between the metal centers and ligand radical in 6. The plot of χMT vs. T (Fig 4) reveals two key pieces of information.
First, the near constant value of χMT is close to that expected for an S = 7/2 ground state (g = 2.00), the result of direct
antiferromagnetic coupling between two high-spin FeII centers (2 × S = 2) and a ligand radical (S = 1/2). In addition, the
lack of temperature dependence of the data above 100 K demonstrates remarkably strong magnetic coupling. Here, with
increasing temperature, no decrease is observed in the data up to 300 K, indicating that the S = 7/2 ground remains isolated
from excited spin states to at least 300 K. In other words, the coupling between Fe centers and ligand radical is so strong
that it cannot be overcome by thermal energy even at 300 K. Due to the lack of temperature dependence, it is impossible
to fit the data to obtain a coupling constant J in order to quantify the interaction. However, we can simulate a curve for
this spin system to obtain a lower bound on the magnitude of J ≥ 900 cm−1 (Fig 4). This magnitude of J is at least 50×
larger than that observed in a similar dihydroxybenzoquinone-bridged FeII2 complex15 and by far the largest ever
observed for a single-molecule magnet. 24
In order to assess the presence of magnetic anisotropy in 6, low-temperature magnetization data were collected at
selected dc fields (Fig 6, lower right). The splitting of the resulting isofield curves, along with their saturation well below
the expected M = 7 µB for an S = 7/2 ground state with g = 2, demonstrates
qualitatively the presence of significant magnetic anisotropy. To quantify
this effect, the data were fit using the program ANISOFIT 2.0 25 to give
parameters of D = −8.4 cm −1, |E| = 0.6 cm−1, and g = 2.11. To our
knowledge, this value of D is the largest yet observed in a multinuclear
single-molecule magnet, surpassing a cyano-bridged CuII3FeIII2 cluster
with an S = 5/2 ground state and D = −5.7 cm−1, 26 and likely stems from a
combination of high-anisotropy high-spin FeII centers and relatively small
spin ground state. This value of D implies that 6 could behave as a singlemolecule magnet with a maximum relaxation barrier of U = (S2 − 1/4)|D| =
101 cm−1.
Following the observation of a large negative D, variable-frequency ac
susceptibility data under zero applied dc field were collected in order to
probe single-molecule magnet behavior. The frequency-dependent out-ofphase component (χM′′) of the ac susceptibility (Fig 5, left) demonstrates
that 6 is indeed a single-molecule magnet. The corresponding Arrhenius
plot of relaxation time (Fig 5, upper right) exhibits a linear region at high Figure 5. Left: Variable-frequency ac susceptibility
temperature, indicative of a thermally activated relaxation process. 27 A fit data for 2. Upper Right: Arrhenius plot of relaxation
to the data in the temperature range 5.0-6.8 K provides an effective time, with Ueff = 50(1) cm−1. Lower Right: Lowrelaxation barrier of Ueff = 50(1) cm−1, with τ0 = 2.7(2) × 10−10 s. temperature magnetization data for 6 at selected fields.
Considering the value of D = −8.4 cm −1 obtained from magnetization data, this relaxation barrier is in exact agreement
with the energetic separation between ground MS = 7/2 levels and first-excited MS = 5/2 levels, given as 6|D| = 51 cm−1.
This observation suggests that thermally assisted quantum tunneling between MS = ±5/2 levels acts as the dominant
relaxation pathway for 6 in this temperature range. As temperature is further decreased, the data begin to deviate from
linearity, denoting the presence of additional fast relaxation processes, such as quantum tunneling and/or spin-spin
relaxation, that shortcut the barrier.
External Funding Resulting from this ISEN Funding
Source: National Science Foundation
Title:
CAREER: Synthesis and Studies of One-Dimensional Magnets Supported by Bulky, RedoxActive Benzoquinonoid Bridging Ligands (DMR-1351959)
Role:
PI
Period:
3/1/14-2/28/19
Amount: $575,000
Publications Resulting from this ISEN Funding
Jeon, I.-R.; Park, J. G.; Xiao, D. J.; Harris, T. D. “An Azophenine Radical-Bridged Fe2 Single-Molecule Magnet
with Record Magnetic Exchange Coupling” J. Am. Chem. Soc. 2013, 135, 16845-16848.
Invited Presentations Resulting from this ISEN Funding
(1) Missouri University of Science and Technology - Rolla, MO, October 2013
(2) ACS “New Trends in Molecular Magnetic Materials” Symposium - Indianapolis, IN, September 2013
(3) National High Magnetic Field Laboratory - Tallahassee, FL, August 2013
Notes and References
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
Ermsley, J. Oxygen, Nature’s Building Blocks: An A-Z Guide to the Elements; Oxford University Press: Oxford, England, 2001.
Descamps, C.; Bouallou, C.; Kannich, M. Energy 2008, 33, 874.
Hadjipaschalis, I.; Kourtis, G.; Poullikkas, A. Renewable Sustainable Energy Rev. 2009, 13, 2637.
(a) Nandi, S. P.; Walker, P. L., Jr. Sep. Sci. Technol. 1976, 11, 441. (b) Castle, W. F. Int. J. Refrig. 2002, 25, 158.
(a) Strzelewicz, A.; Grzywna, Z. J. J. Membr. Sci. 2007, 294, 60. (b) Rybak, A.; Grzywna, Z. J.; Kaszuwara, W. J. Membr. Sci. 2009, 336, 79.
(a) Gwak, J.; Ayral, A.; Rouessac, V.; Cot, L.; Grenier, J.-C.; Choy, J.-H. Microporous Mesoporous Mater. 2003, 63, 177. (b) Gwak, J.;
Ayral, A.; Rouessac, V.; Kim, K. H.; Grenier, J.-C.; Cot, L.; Choy, J.-H. Sep. Purif. Technol. 2005, 46, 118. (c) Madaenia, S. S.; Enayatia, E.;
Vatanpour, V. Polym. Adv. Technol. 2011, 22, 2556.
Borys, P.; Pawelek, K.; Grzywna, Z. J. Phys. Chem. Chem. Phys. 2011, 13, 17122.
Kaye, S. S.; Choi, H. J.; Long, J. R. J. Am. Chem. Soc. 2008, 130, 16921.
(a) Kurmoo, M. Chem. Soc. Rev. 2009, 38, 1353. (b) Dechambenoit, P.; Long, J. R. Chem. Soc. Rev. 2011, 40, 3249.
Pierpoint, C. G.; Francesconi, L. C.; Hendrickson, D. N. Inorg. Chem. 1977, 16, 2367.
Kitagawa, S.; Kawata, S. Coord. Chem. Rev. 2002, 224, 11.
Tao, J.; Maruyama, H.; Sato, O. J. Am. Chem. Soc. 2006, 128, 1790.
Min, K. S.; Rheingold, A. L.; DiPasquale, A.; Miller, J. S. Inorg. Chem. 2006, 45, 6135.
Guo, D.; McCusker, J. K. Inorg. Chem. 2007, 46, 3257.
Min, K. S.; DiPasquale, A. G.; Golen, J. A.; Rheingold, A. L.; Miller, J. S. J. Am. Chem. Soc. 2007, 129, 2360.
Siri, O.; Taquet, J.-p.; Collin, J.-P.; Rohmer, M.-M.; Bénard, M.; Braunstein, P. Chem. Eur. J. 2005, 11, 7247.
Taquet, J.-p.; Siri, O.; Braunstein, P. Inorg. Chem. 2006, 45, 4668.
Huang, Y.-B.; Tang, G.-R.; Jin, G.-Y.; Jin, G.-X. Organometallics 2008, 27, 259.
Kumbhakar, D.; Sarkar, B.; Das, A.; Das, A. K.; Mobin, S. M.; Fiedler, J.; Kaim, W.; Lahiri, G. K. Dalton Trans. 2009, 9645.
Wenderski, T.; Light, K. M.; Ogrin, D.; Bott, S. G.; Harlan, C. J. Tetrahedron Lett. 2004, 45, 6851.
Khramov, D. M.; Boydston, A. J.; Bielawski, C. W. Org. Lett. 2006, 8, 1831.
Tyeklar, Z.; Jacobson, R. R.; Wei, N.; Murthy, N. N.; Zubieta, J.; Karlin, K. D. J. Am. Chem. Soc. 1993, 115, 2677.
Gennett, T.; Milner, D. F.; Weaver, M. J. J. Phys. Chem. 1985, 89, 2181.
Delfs, C.; Gatteschi, D.; Pardi, L.; Sessoli, R.; Wieghardt, K.; Hanke, D. Inorg. Chem. 1993, 32, 3099.
Shores, M. P.; Sokol, J. J.; Long, J. R. J. Am. Chem. Soc. 2002, 124, 2279.
Wang, C.-F; Zuo, J.-L.; Bartlett, B. M.; Song, Y.; Long, J. R.; You, X.-Z. J. Am. Chem. Soc. 2006, 128, 7162.
(a) Cole, K. S.; Cole, R. H. J. Chem. Phys. 1941, 9, 341-341. (b) Boettcher, C. J. F. Theory of Electric Polarisation; Elsevier: Amsterdam,
1952. (c) Aubin, S. M.; Sun, Z.; Pardi, L.; Krzystek, J.; Folting, K.; Brunel, L.-J.; Rheingold, A. L.; Christou, G.; Hendrickson, D. N. Inorg.
Chem. 1999, 38, 5329-5340.