Flexibility – A Tool for Chirality
Control in Asymmetric Catalysis
Raivis Žalubovskis
Doctoral Thesis
Stockholm 2006
Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan i
Stockholm framlägges till offentlig granskning för avläggande av filosofie
doktorsexamen i kemi med inrikting mot organisk kemi, måndagen den 27
november, kl 10.00 i sal F3, KTH, Lindstedtsvägen 26, Stockholm. Avhandlingen
försvaras på engelska. Opponent är Professor Alexandre Alexakis, Université de
Genève, Schweiz.
ISBN 91-7178-491-8
ISRN KTH/IOK/FR--06/104--SE
ISSN 1100-7974
TRITA-IOK
Forskningsrapport 2006:104
© Raivis Zalubovskis
Universitetsservice US AB, Stockholm
Zalubovskis, R. 2006 ”Flexibility – A Tool for Chirality Control in
Asymmetric Catalysis”, Organic Chemistry, KTH Chemical Science and
Engineering, SE-100 44 Stockholm, Sweden.
Abstract
This thesis deals with the design and synthesis of ligands for
asymmetric catalysis: palladium catalyzed allylic alkylations, and rhodium and iridium catalyzed hydrogenations of olefins.
Chirally flexible phosphepine ligands based on biphenyl were
synthesized and their properties were studied. The rotation barrier for
configurationally flexible phosphepines was determined by NMR
spectroscopy. The ratio of the atropisomers was shown to depend on
the group bound to phosphorus. Only complexes with two homochiral
ligands bound to the metal center were observed upon complexation
with Rh(I). It was shown that one diastereomer of the flexible ligand
exhibits higher activity but lower selectivity than its diastereomer in
the rhodium catalyzed hydrogenation of methyl α-acetamidocinnamate. These ligands were also tested in nickel catalyzed silaborations.
Chiral P,N-ligands with pseudo-C2 and pseudo-CS symmetry
based on pyrrolidines-phospholanes or azepines-phosphepines were
synthesized and studied in palladium catalyzed allylic alkylations.
Semi-flexible azepine-phosphepine based ligands were prepared and
their ability to adopt pseudo-C2 or pseudo-CS symmetry depending on
the substrate in allylic alkylations was studied. It was shown on model
allyl systems with flexible N,N-ligands that the ligand prefers CSsymmetry in compexes with anti-anti as well as syn-syn allyl moieties,
but that for the latter type of complexes, according to computations,
the configuration of the ligand is R*,R* in the olefin complexes
formed after addition of a nucleophile to the allylic group.
A preliminary investigation of the possibilities to use a supramolecular approach for the preparation of P,N-ligands with
pseudo-C2 and pseudo-CS symmetry was made. An N,N-ligand with
C2 symmetry was prepared and its activity in palladium catalyzed allylic alkylation was studied.
Pyridine-based P,N-ligands were tested in iridium catalyzed hydrogenations of unfunctionalized olefins with good activities and selectivities. In order to attempt to improve the selectivity, ligands with a
chirally flexible phosphepine fragment were prepared and applied in
catalysis with promising results.
Keywords: flexibility, symmetry, dissymmetry, asymmetric catalysis, chiral
ligands, pseudo-C2, pseudo-CS, conformational study, rotation barrier, pyrrolidines,
phospholanes, azepines, phosphepines, supramolecular, hydrogenation, allylic alkylation, silaboration, palladium, rhodium, nickel, platinum, iridium.
"Ķīmija ir "nemierīga" zinātne, kas prasa, lai cilvēks tai ziedo sevi
visu"
“Chemistry is a “restless” science that requires man’s total dedication”
G. Vanags
List of Publications
I. Influence of Steric Symmetry and Electronic Dissymmetry on the
Enantioselectivity in Palladium-Catalyzed Allylic Substitutions
Vasse, J. L.; Stranne, R.; Zalubovskis, R.; Gayet, C.; Moberg, C.
J. Org. Chem. 2003, 68, 3258-3270
II. Stereochemical Control of Chirally Flexible Phosphepines
Zalubovskis, R.; Fjellander, E.; Szabó, Z.; Moberg, C. Eur. J. Org. Chem.,
in press
III. Self-Adaptable Catalysts - Substrate Dependent Ligand Configuration
Zalubovskis, R., Constant, S., Linder, D., Lacour, J., Privalov, T., Moberg,
C. manuscript
IV. Enantioselective Silicon-Boron Additions to Cyclic 1,3-Dienes Catalyzed by the Platinum Group Metal Complexes
Gerdin, M., Penhoat, M., Zalubovskis, R., Pétermann, C., Moberg, C.
preliminary manuscript
V. Appendix: Supplementary Material
Abbreviations
BINAP
BINOL
BSA
COD
DCM
DFT
DIAD
DMSO
dppe
dppp
MW
LAH
LDA
RT
TBAF
Tf
TMEDA
TMS
Δ
σ
2,2’-Bis(diphenylphosphino)-1,1’-binaphthyl
1,1’-Bi-2-naphthol
N,O-Bis(trimethylsilyl)acetamide
Cyclooctadiene
Dichloromethane
Density functional theory
Diisopropyl azodicarboxylate
Dimethyl sulfoxide
1,2-Bis(diphenylphosphino)ethane
1,3-Bis(diphenylphosphino)propane
Microwave
Lithium aluminum hydride
Lithium diisopropylamide
Room temperature
Tetrabutylammonium fluoride
Trifluoromethylsulfonic
N,N,N',N'-Tetramethylethylenediamine
Trimethylsilyl
Heating
Mirror plane
Table of contents
1. Introduction.....................................................................................................1
1.1 Asymmetric Catalysis................................................................................1
1.2 Aim of this Thesis .....................................................................................2
2. Flexibility: Atropos vs Tropos ........................................................................3
2.1 Definitions.................................................................................................3
2.2 Naturally Occurring and Synthetic Atropisomers .....................................4
2.3 Energy Barriers .........................................................................................5
2.4 Asymmetric Activation of Tropos Catalyst .............................................10
3. Chirally Flexible Phosphepines....................................................................15
3.1 Background .............................................................................................15
3.2 Synthesis of Phosphepine Ligands..........................................................15
3.3 NMR Investigation of Free Ligands........................................................17
3.4 NMR Investigation of Rh Complexes .....................................................19
3.5 Rh Catalyzed Hydrogenation ..................................................................20
3.6 Ni Catalyzed Silicon-Boron Addition to 1,3-Dienes...............................21
4. Steric Symmetry and Electronic Dissymmetry...........................................23
4.1 Background .............................................................................................23
4.2 Synthesis of P,N-ligands with pseudo-C2 and pseudo-CS Symmetry ......23
4.3 Pd Catalyzed Asymmetric Allylic Alkylation .........................................25
4.4 Mechanistic Considerations ....................................................................27
4.5 Synthesis of N,N-Ligands for π-Allyl Complexes ..................................30
4.6 Preparation of Pd Allyl Complexes .........................................................31
4.7 NMR Investigation of Pd Allyl Complexes ............................................33
5. Variation of Chelating Backbone .................................................................39
5.1 Background .............................................................................................39
5.2 Synthesis of P,N-ligands with Dibenzofuran Backbone..........................39
5.3 Synthesis of P,N-ligands – Supramolecular Approach ............................41
5.4 Pd Catalyzed Asymmetric Allylic Alkylation .........................................46
5.5 Synthesis of Pyridine Based Ligands ......................................................46
5.6 Ir catalyzed Hydrogenation of Unfunctionalized Olefins........................49
6. Concluding Remarks ....................................................................................51
Acknowledgements............................................................................................53
References ..........................................................................................................55
1. Introduction
1.1 Asymmetric Catalysis
The constantly growing demand for enantiopure substances from
different disciplines, especially from the pharmaceutical industry, requests development of new and highly efficient and also environmentally
friendly methods for the production of single enantiomers of the compounds required. Worldwide sales of single-enantiomer drugs reached
more than $159 billion in year 2002.1
There are three alternative paths to get access to enantiopure compounds: 1) enantiopure natural products (chiral pool); 2) resolution of
racemates into the pure enantiomers; 3) asymmetric synthesis.
The first path, as a rule, is limited to the preparation of only one
enantiomer. Often, but not always, it gives access to a limited amount of
the product due to low concentration in the natural source.
The second path deals with synthesis of racemates which must be
resolved into the enantiomers. The main methods for resolution are often
time and resource consuming. The major drawback of resolution (use of a
resolving agent, chiral chromatography and preferential crystallization) is
that a maximum of 50% theoretical yield of the desired enantiomer is
obtained. This problem might be solved by using dynamic kinetic resolution, where the slowly reacting enantiomer is continuously racemized, but
this technique is not generally applicable and therefore the problem of the
utilization of the unwanted enantiomer is still open, especially in a case of
large scale production.
The third path – asymmetric synthesis and especially asymmetric
catalysis – is the most attractive of all for academic laboratories as well as
for industry. The main advantage of asymmetric catalysis is that chiral
information from a small amount of the chiral catalyst is transferred to a
large amount of product. Under optimized conditions only one enantiomer is produced. This is of great significance since it lowers the amount
of waste and thereby makes the process environmentally more friendly.
However, there is still need for new catalytic systems to make existing
processes more efficient and to allow the synthesis of new substrates
which are becoming of importance.
1
1.2 Aim of this Thesis
The primary goal of this work was to develop new catalysts for
asymmetric synthesis and to extend our understanding of the factors influencing the selectivity of catalytic processes; such understanding is
crucial for further development of catalytic systems. One particular goal
was to design catalysts which may adopt to the substrate undergoing reaction and which therefore could be used for a broade range of substrates.
Another goal was to develop ligands for asymmetric catalysis with variable backbone of the ligands by using conventional as well as supramolecular approaches.
2
2. Flexibility: Atropos vs Tropos
2.1 Definitions
The term “atropisomerism”* was used for the first time by Kuhn
(1933), when referring to earlier work by Christie and Kenner2 (1922), in
which 6,6’-dinitro-2,2’-diphenic acid was resolved into two antipodes
and a new kind of enantiomerism was observed for the first time.
In general terms, atropisomerism is an axial chirality where rotation
around a single bond is hindered and two rotamers can be isolated. The
most common atropisomers are biaryls (Figure 1), where the groups A≠B
and C≠D are big enough to prevent rotation around the single bond and,
thereby, racemization. However, these terms do not provide any exact
definition, but leave some questions open: How stable do rotamers have
to be (how long is the half-life time) in order to be called atropisomers?
Does atropisomerism still exist where the isolation of enantiomers is
problematic or impossible but where the enantiomers can be observed by
spectroscopical methods, for instance NMR? It was later (1983) arbitrarily defined by Ōki3,4 that, in order to speak about atropisomerism, one
should be able to isolate isomers and their half-life time (t1/2) should at
least be 1000 sec (16.7 min). This is considered to be the minimum time
necessary for isolation of isomers. Speaking in terms of a free energy
barrier, which is temperature dependent, it has to be at least 22.3 kcal
mol-1 (93.3 kJ mol-1) at 300 K.
Atropos
A
C
B
D
X
A
C
B
D
Tropos
Figure 1. Enantiomeric chiral biphenyls.
For some time compounds, which did not fulfill the free energy
barrier criteria of atropisomers, were left without a proper name and were
simply named rotamers, until Mikami recently (2002) introduced the term
tropos.5
*
atropos from Greek a meaning not, and tropos meaning turn.
3
2.2 Naturally Occurring and Synthetic Atropisomers
Since the discovery of the first atropisomer by Christie and Kenner
many molecules with hindered rotation around a single bond have been
synthesized or discovered in Nature. Atropisomers were for a long time
considered as “academic curiosity” only. However, it has been demonstrated that the configuration of atropisomers may for example determine
the pharmaceutical properties of bioactive compounds, and that atropisomerism is crucial in asymmetric synthesis.6
Some naturally occurring atropisomers (Figure 2) have important
biological activities. For instance, knipholone exhibits good antimalarial
and antitumor activities7 and mastigophorene A stimulates nerve
growths.8 Naturally occurring murrastifoline-F demonstrates that natural
products are not restricted to C-C axial chiral biaryls but may also contain
chiral axes connected with heteroatoms.9
OH O
Me
Me
OH
Me
O
Me
OH
Me
OMe O
OH
OH
Me
Knipholone
N
*
OH
*
*
HO
Me
Me
OH
Me
Me
Me
Mastigophorene A
OMe
Me
N
H
OMe
Murrastifoline-F
Figure 2. Naturally occurring atropisomers.
Some of the most important examples of synthetic atropisomers are
derived from 1,1’-bi-2-naphthol (BINOL) (Figure 3). BINOL is broadly
used in enantiomerically pure form in asymmetric synthesis and in particular in asymmetric catalysis and also as a precursor of ligands like
BINAP, for the first time reported by Noyori10 (1980). BINAP was the
first successful atropligand used for Rh catalyzed hydrogenation of olefins. Another well-known biphenyl-based example is MeO-BIPHEP11
which is an analogue of BINAP proven to serve as an excellent ligand in
asymmetric catalysis.
4
OH
OH
BINOL
PPh2
PPh2
MeO
MeO
BINAP
PPh2
PPh2
MeO-BIPHEP
Figure 3. Important synthetic atropisomers.
2.3 Energy Barriers
Biphenyl exists in two twisted forms, 1 and 2, (Scheme 1) with a
torsion angle (ω) of 44° in the gaseous state.
3
2
1
ω
90o
4
Scheme 1. Interconversion of unsubstituted biphenyl.
The interconversion between the twisted forms occurs both through
the planar intermediate 3 (ω = 0°) and through the perpendicular form 4.
From electron diffraction experiments,12 rotation barriers# of 1.43 and
1.55 kcal mol-1, respectively, were determined for the two processes. Substitution of hydrogen atoms in biphenyl for deuterium leads to an increase
of the rotation barrier for 2.37 and 2.20 kcal mol-1 for the planar form 3
#
In this thesis energies are given in kcal. If in a reference energies are given in kJ, they are transferred into kcal counting 1 kcal = 4.184 kJ.
5
and the perpendicular form 4, respectively. Now the planar form has a
higher barrier than the perpendicular one.
The energy profile for tetraortho-substituted biphenyls is shown in
Figure 4.3 The maximal rotation barriers at 0° and ±180° correspond to
the planar forms, where steric repulsion between the ortho substituents is
dominating. The maxima at ±90° correspond to the complete absence of
π-orbital overlaps (electronic factors). π-orbital overlap bring the system
towards the planar form. Compromise between steric repulsion and πorbital overlap corresponds to the minima at ±44°, as we could see in
example with unsubstituted biphenyl 1.
E
ω
-90
-180
0
90
180
Figure 4. General energy profile for rotation in biphenyls.
Symmetrically o,o’-bisubstituted biphenyls 5 can exist in two enantiomeric forms, 5 (aS) and 5’ (aR), which interconvert via the planar (ω =
180°) form 6, where o-substituents are pointing in opposite directions.
This corresponds to a lower barrier for racemization compared to the
planar (ω = 0°) form where the o-substituents would have strong steric
interaction which would significantly increase the rotation barrier.
R
R
5
R
R
R
R
6
5'
Scheme 2. Interconversion of symmetric o,o’-biphenyls.
Some rotation barriers for biphenyls 5 are presented in Table 1.13,14,15
6
Table 1. Rotation barriers of o,o’-bisubstituted biphenyls 5.
Biphenyl
5a
5b
5c
5d
5e
5f
5g
a
ΔG* (kcal mol-1)
4.8a
17.6a
21.5b
23.2 b
19.3 b
24.6 b
22 b
R
F
Cl
Br
I
CH3
CF3
PPh2
Theoretically calculated values.13
b
Experimentally determined values.14,15
As we can see, there is a general tendency of increasing the rotation
barrier by increasing the size of the substituents in the o,o’-positions (biphenyls 5a-5f). For 5g however, one could expect a much higher barrier
than 22 kcal mol-1, considering that the diphenylphosphino group is a
bulky group. In order to explain the low rotational barrier, a sophisticated
cogwheel mechanism has been proposed (Figure 5), where the approaching o’-proton slips between the lone pair and the phenyl group attached to
phosphorus, whereby steric repulsion gets slightly reduced.15
Ph2P
H
Ph P
H
Ph
P
H
Ph
5g
Figure 5. Cogwheel motion in o,o’-bisubstituted biphenyl.
Considering more rigid systems, like binaphthyls 7 (Scheme 3), the
barrier for racemization strongly depends on the size of R, as one could
expect. It has been demonstrated by DFT calculations that racemization
occurs via the planar structure 8 for binaphthyl (R=H) and BINOL
(R=OH), where R groups are pointing in opposite directions. However,
this does not necessarily apply for other binaphthyl derivatives, where
racemization can also occur via the planar form, where the R groups are
pointing in the same direction. Racemization studies have also been performed and the rotation barriers were then found to be 22.1 kcal mol-1 for
binaphthyl in benzene at 44 °C (half racemization time τ1/2 = 68 min), and
37.8 kcal mol-1 for BINOL at 220 °C in diphenyl ether (τ1/2 = 60 min),16
7
demonstrating that BINOL in contrast to binaphthyl does not racemize at
room temperature and can be treated as a rigid system.
R
R
R
R
R
R
7
8
7'
Scheme 3. Interconversion of symmetric binaphthyls.
For bridged biphenyl systems (Scheme 4) there can obviously be
only one planar form involved in racemization, since the phenyl rings are
no longer free to rotate per 180°.
X
9
X
9'
Y
X
Y
10
X
10'
Scheme 4. Bridged biphenyls.
Experimentally determined rotation barriers for mono-bridged biphenyls 9 are listed in Table 2.17
Table 2. Rotation barriers for the mono-bridged biphenyls 9.
X
O
S
SO2
CH2
CH2CH2
C(CO2Et)2
C(CO2H)2
+
NMe2, BrNCOCHBrMe
NCOCHMePh
8
ΔG* (kcal mol-1)
9.0
16.2
18.9
12.0
24.0
14.4
23.0
13.4
10.2
11.2
As we can see from Table 2, lack of substituents in the 6 and 6’ positions leads to rather low rotational barriers and all listed derivatives, even
if they would be isolated in enantiopure form, would racemize at room or
slightly elevated temperature in a short period of time. The doublybridged biphenyls 10 (Scheme 4) exhibit higher stability (Table 3) than
mono-bridged systems.18 As Table 3 shows, the configurational stability
of the compounds increases along with the size of the substituents and,
for instance, dithiepin (X=Y=S) has a rotation barrier comparable with
that of BINOL, and therefore the compound does not racemize at room
temperature.
Table 3. Rotation barriers of doubly-bridged biphenyls 10.
X
O
C=O
O
S
-
ΔG* (kcal mol-1)
19.9
31.2
30.6
35.0
8
Y
O
C=O
S
S
-
Only a limited amount of experimental data on rotation barriers of
axially chiral ligands is available in the literature. The dioxaphosphepines
(phosphites) 11 and 12 (Figure 6) have rather low rotation barriers and
rapidly interconvert even on the NMR time scale, giving rise to one average signal at room temperature. Rotation barriers for compounds 11 and
12 have been found to be 10 (199 K) and 11.5 kcal mol-1 (259 K), respectively.19
MeO
But
O
P
O
MeO
But
O
O
But
OMe
But
O
P
O
But
11
But
But
But
O
O
P
O HO
OMe
But
But
But
But
12
Figure 6. Dioxaphosphepines.
In a recent example a rotational barrier of 13.0 kcal mol-1 (290 K)
was determined by a variable temperature NMR experiment for the dioxaphosphepine (phosphoramidite) Rh complex 13. 20 However, it is
worth to note that the coalescence temperature for the corresponding
9
phosphate Rh complex 14 was not reached (lower than 290 K), indicating
a rotational barrier lower than 13 kcal mol-1, which is close to the value
for free phosphate ligands like 11 or 12.
Ph
Ph
N
O
P
O
N
Ph Ph
Ph
Ph
O
O
O
P
O
O
O
P
O
P
O
Rh
(acac)
Rh
(acac)
13
14
Figure 7. Dioxaphosphepine Rh complexes.
2.4 Asymmetric Activation of Tropos Catalyst
In earlier experiments Mikami (1999)21 demonstrated that, by mixing Ru complex 15 (Scheme 5), which contains a rotationally flexible
biphenyl ligand, with the chirally rigid diamine 16 in CDCl3, the complex
17 formed as a 1:1 mixture of two diastereomers, and this ratio remained
over time. In the presence of (CD3)2CDOD at room temperature the diastereomeric ratio changed to 1:3 in 3 hours and remained at that value
over time.
Ar2
P
RuCl2(dmf)2
P
Ar2
15
Ph
H2N
H2N
Ph
16
Ar2
P Cl
Ru
P Cl
Ar2
H2
N
Ph
N
H2
Ph
2-propanol
Ar2
P Cl
Ru
P Cl
Ar2
17
Ar = 3,5-dimethylphenyl
Scheme 5. Isomerization of the flexible Ru complex 17.
10
17'
H2
N
Ph
N
H2
Ph
Almost at the same time (2000) Gagné22 prepared the Pt complex
18 (Scheme 6) which in the presence of pyridine exists as a 95:5 ratio of
the diastereomers 18 and 18’.
Ph2
P
O
Pyridine
Ph2
P
Pt
O
P
Ph2
O
Pt
40 oC
18
P
Ph2
O
18'
Scheme 6. Isomerization of flexible Pt complex.
Obviously inspired by experimental results, theoretical calculations
were performed in order to determine the mechanism of isomerization5 of
flexible ligands in a metal complex.
Three pathways (Scheme 7) were examined for isomerization of
the Ru complex 19, containing a flexible biphenyl unit. First of all path 3,
where complete dissociation of the diphosphinobiphenyl unit occurs, was
excluded, because experimental results showed that there was no ligand
exchange, when complex 19 and BINAP, which is a rigid equivalent of
diphosphinobiphenyl unit, were mixed. Theoretical calculations of path 1
showed that the activation energy for internal rotation through the planar
form of biphenyl was 36 kcal mol-1. By comparison with BINOL which
exhibits a similar rotation barrier, we can assume that there is no rotation
at room temperature. The one-arm-off mechanism (path 2) was examined
more closely. The activation energy for dissociation of the Ru-P bond was
calculated to be 55 kcal mol-1, which in fact was 19 kcal mol-1 higher than
path 1. However, calculations of path 2 performed including two MeOH
molecules (solvent supported dissociation) gave a dissociation energy of
only 22 kcal mol-1. Taking into account that the rotation barrier for the
free diphosphinobiphenyl unit 5g (Table 1) is 22 kcal mol-1, path 2 is the
most probable way of isomerization and moreover, it is in good agreement with experimental data regarding the complex 17.
11
Path 1
Ph2
H
P
Cl N2
Ru
P
Cl N
H2
Ph2
Ph
Ph
Path 2
Ph2
H
P Cl 2
N
Ru
P Cl N
H2
Ph2
Ph
Ph2P
Ph
P
Ph2
H
Cl N2
Ru
Cl N
H2
Ph
Ph
Ph2
H
Cl N2
P
Ru
Cl N
P
H2
Ph2
Ph
Ph
19
Path 3
Ph2P
H
Cl N2
Ru
Cl N
H2
PPh2
Ph
Ph
Scheme 7. Possible mechanisms of inversion of biphenyl unit in Ru complex.
The examples presented above, where flexible units adopt to the
one preferable configuration under the influence of chirally rigid moieties
(activators), bring us to logical conclusions about an advantage of having
flexible ligands in transition metal catalysis (Scheme 8).5
If a rigid racemic atropos ligand like BINAP is used for the preparation of a metal catalyst with a chiral enantiomerically pure activator,
then the result will unavoidably be a mixture of diastereomers and when
this mixture is used in asymmetric catalysis, the product will have a lower
ee. To obtain higher ee’s in catalytic reactions, one must either separate
the racemic atropos ligand or the corresponding diastereomeric mixture
of the metal catalyst, which is time-consuming and makes the entire
process expensive. However, if a flexible racemic tropos ligand is employed, then – under influence of an activator – the tropos ligand will
undergo self-adaptation in the metal complex, as was shown above
(Schemes 5 and 6), and the catalyst will be obtained as a single diastereomer or at least a diastereomerically enriched one without performing a separation, and asymmetric catalysis will provide a product with
higher ee.
12
L
L
L
L
rac-Atropos ligand
rac-Tropos ligand
M
M
X
X
*
*
X
L
Activator
X
X
L
*
M
L
X
Mixture of diastereomers
Substrate
Lower ee
Activator
X
*
M
L
X
Single diastereomer
Substrate
Higher ee
Scheme 8. Comparison of rac-atropos and rac-tropos ligands.
It is obvious that in catalysis it is a great advantage to have a flexible ligand, which undergoes self-adaptation and does not require resolution of the ligand or separation of a diastereomeric mixture of the catalyst
since this makes the process of catalyst preparation shorter and the catalysis cheaper than the use of rigid analogues.
13
14
3. Chirally Flexible Phosphepines
3.1 Background
Our interest in chirally flexible ligands led us to investigate
phosphepine-based ligands 20 (Figure 8) as flexible analogues of the
rigid binaphthyls 21, which for the first time were prepared by Gladiali23
and further explored by Stelzer24 and which recently found many applications in asymmetric catalysis.25 However, very little information can be
found about ligands based on the flexible phosphepines 20.26 We wanted
to investigate the ability of such compounds to adopt one preferable conformation depending on the chirally rigid substituents R on the phosphorus atom. Furthermore, we wanted to investigate the stability against racemization of the phosphepines. Finally, our goal was also to compare the
reactivities and selectivities of the flexible ligands 20 with those of the
corresponding rigid analogues 21 in catalysis.
P R
20
P R
21
Figure 8. Phosphepines.
3.2 Synthesis of Phosphepine Ligands
The flexible ligands 27 were synthesized according to the procedure shown in Scheme 9. Treatment of 2,2’-biphenol (22) with triflic
anhydride provided the bistriflic derivative 23 in good yield. A subsequent Ni catalyzed coupling with MeMgBr gave 2,2’-dimethylbiphenyl
(24). Lithiation of 24 and further treatment with Cl2PNEt2 gave the
aminophosphine 25 which was obtained with satisfactory purity and was
used as a racemic mixture. Treatment of 25 with dry HCl gas led to the
chlorophosphine 26, which was a key element in our ligand synthesis. A
simple treatment of 26 with MeOH in the presence of triethylamine gave
the phosphinite 27a in moderate yield (33%). However, when this procedure was used with L-(-)-menthol, some amount of chlorinated product
15
was obtained which made purification of the product complicated. A procedure where L-(-)-menthol was first treated with BuLi and then reacted
with phosphochloride 26 gave ligand 27b with satisfactory yield (60%).
(+)-Neomenthol, (-)-8-phenylmenthol and (-)-borneol reacted analogously to give 27c (90%), 27d (86%) and 27e (79%), respectively.
Treatment of 26 with (R)-2-phenylethylamine in presence of triethylamine gave aminophosphine 27f in 40% yield.
OH
OH
Tf2O
DCM
95%
22
1) BuLi, TMEDA
MeMgBr
OTf
OTf
P NEt2
NiCl2 dppp
95%
2) Cl2PNEt2
24%
23
25
24
HCl
P Cl
ROH or RNH2
P XR
base
39%
26
27
Ph
RX =
OMe ;
27a (33%)
O
;
27b (60%)
O
27c (90%)
; O
;
Ph
27d (86%)
;
HN
.
O
27e (79%)
27f (40%)
Scheme 9. Synthesis of flexible phosphepine ligands.
The synthesis of rigid ligands based on binaphthyl were performed
by reaction of L- or D-menthol with the phosphochloride 2827 by a route
analogous to that used for ligands 27b-e (Scheme 10) affording ligands
29 (36%) and 30 (39%).
Our choice of the structure of the rigid ligands was based on the
fact that ligand 29 is an equivalent of the flexible ligand 27b with aR
stereochemistry and 30 is a pseudo-enantiomer of 27b with aS
stereochemistry.
16
L- or D-menthol
P Cl
P O
or
P O
BuLi
28
30 (39%)
29 (36%)
Scheme 10. Synthesis of rigid phosphepine ligands.
3.3 NMR Investigation of Free Ligands
The stereochemical properties of the flexible phosphepines 27a-f
were studied by NMR. The phosphorus atom is not a stereogenic center
and therefore a pyramidal inversion at the phosphorus does not affect the
chirality of the molecules. The protons A-D and B-C (Figure 9) are interchanging via tropoisomerization in 27a, and protons A-B and C-D are
interconverting via pyramidal inversion at the phosphorus (Scheme 11).
R
P
P
R
P
R
Inversion at phosphorus
R
P
R
Tropoisomerization
Tropoisomerization
P
Inversion at phosphorus
R
P
P
R
P
R
Scheme 11. Pyramidal inversion at phosphorus and tropoisomerization in a
flexible phosphepine.
For all flexible ligands, except 27a, two signals were observed in
the P NMR spectrum, demonstrating that interconversion between diastereomers is slow on the NMR time scale. The four methylene protons
in 27a are nonequivalent and give four signals in 1H NMR spectrum and
31
17
eight signals for the compounds 27b-f due to the presence of two diastereomers, showing that both the atropisomerization and the pyramidal
inversion at phosphorus are slow. The four methylene protons in 27a
(Figure 9) gave signals at δ 2.97 (dd, JH,H = 11.9 Hz,, JP,H = 14.9 Hz, HA),
2.82 (d, JH,H = 15.0 Hz, HB), 2.51 (d, JH,H = 11.9 Hz, HC) and 2.34 (dd,
JH,H = 15.0 Hz, JP,H = 18.7 Hz, HD). From the H-H coupling constants we
concluded that A and C as well as B and D were geminal protons. Only
two out of four methylene protons coupled to phosphorus. From the NOESY spectrum we found that the protons A and B were aligned with the
phenyl rings, while the protons B and C gave cross peaks with the methoxy group. This assignment is in the agreement with the dihedral angle
dependence of the phosphorus proton coupling constant.28
Figure 9. Phosphepine 27a.
The diastereomeric ratios of compounds 27b-f were determined
from their 31P NMR spectra. Chirality transfer from the chirally rigid part
to the biphenyl moiety was found to be inefficient and no diastereomeric
ratio exceeded 1:1.43, as observed for 27d (Table 4).
Table 4. Diastereomeric ratio of atropisomers 27b-f.
Ligand
27b
27c
27d
27e
27f
Ratio
1:1.02
1:1.26
1:1.43
1:1.12
1:1.19
Our further NMR investigations were focused on the determination
of the tropoisomerization barriers of the compounds 27a and 27b by dynamic NMR spectroscopy. The large chemical shift difference between
the two phosphorus signals in 27b gave us an excellent opportunity to
18
determine the activation parameters of the exchange by NMR spectroscopy in a large temperature region. As was mentioned above, pyramidal
inversion at phosphorus does not change the chirality of the molecules.
Only flipping of the phenyl rings results in exchange of the isomers and
causes an increase of the line width of the phosphorus signals at elevated
temperatures. However, when we performed variable temperature NMR
experiments on compound 27b, the proton decoupled phosphorus spectra
showed rather low exchange rate and insignificant line width changes. By
increasing the temperature from 20 to 80 °C, the line width of the signals
changed only by approximately 10 Hz. The one-dimensional inversiontransfer experiments29 provided us with more accurate rate constants for
the exchange process between the diastereomers at different temperatures. From the values obtained, a Gibbs free energy of activation of 18.5
kcal mol-1 at 298 K was calculated for the tropoisomerization process of
27b.
For the compound 27a, only 1H NMR experiments could be used
for determination of the activation barrier for tropoisomerization. The
one-dimensional inversion-transfer experiments were run at variable temperatures by inverting the methylene proton signals and a Gibbs free energy of activation of 19.3 kcal mol-1 at 298 K was calculated for 27a. The
tropoisomerization barriers obtained for compounds 27a and 27b were
considerably higher than those of the corresponding N analogues (Table
2), but they were too low to permit isolation of the isomers at room temperature.
3.4 NMR Investigation of Rh Complexes
In order to study the influence of complexation to a metal ion on the
diastereomeric ratio of 27b-e, Rh(I) complexes of the ligands were prepared, using [Rh(COD)2]+BF4-. 31P NMR of a 2:1 mixture of the ligand
27b and [Rh(COD)2]+BF4- showed complete absence of free ligand signals but the appearance of two new doublets (JRh,P = 170 Hz, proton decoupled). No P-P couplings were observed. When excess ligand was
added, free ligand signals appeared in addition to those originating from
the complex. The behavior observed led us to conclude that two homochiral 2:1 complexes were formed: in case of an equilibrating mixture of
free and complexed ligands, a gradual addition of the ligand would lead
to a change of the chemical shift where a formation of a mixture homo-
19
and heterochiral complexes would give rise to additional signals as it had
been reported for complexes 13 and 14.20,30 31P NMR of a 2:1 mixture of
ligands 27c-e with [Rh(COD)2]+BF4– showed a similar behavior, where
only two doublets were observed.
When ligands 27b and 27c and [Rh(COD)2]+BF4– were mixed in
equimolar amounts, the only 31P NMR signals observed were those from
the separate complexes of Rh 27b and Rh 27c, demonstrating that mixed
complexes were not formed.
When rigid analogs 29 or 30 were mixed with [Rh(COD)2]+BF4–
one doublet was observed for each complex in 31P NMR. Comparison of
these spectra with those obtained from complex Rh 27b allowed us to
assign the isomers in Rh 27b. We found that the major complex contained aS,L-27b.
The diastereomeric ratios of Rh 27b-e complexes are dependent on
the alcohol group; the highest ratio, 1:1.94, was observed for Rh 27d
(Table 5).
Table 5. Diastereomeric ratio of Rh complexes of 27b-e.
Complex
Rh 27b
Rh 27c
Rh 27d
Rh 27e
Ratio
1:1.45
1:1.08
1:1.94
1:1.12
3.5 Rh Catalyzed Hydrogenation
In order to investigate the catalytic behavior of the flexible biphenyl ligands 27 and compare their properties with those of rigid analogues
29 and 30, the ligands were tested in Rh catalyzed hydrogenation of
methyl α-acetamidocinnamate (31) (Scheme 12).
CO2Me
Ph
NHAc
31
CO2Me
Rh(COD)2BF4 (1mol%)
L (2 mol%), MeOH
H2 (1 bar), RT
∗
Ph
NHAc
32
Scheme 12. Rh catalyzed hydrogenation of α-acetamidocinnamate (31).
20
The results of the catalytic reactions are presented in Table 6. All
ligands, except 27c, exhibited good reactivity with variable ee’s. We
found that the rigid ligands 29 and 30 both gave the (S)-isomer (entries 5
and 6), which showed that the binaphthyl moiety was responsible for the
absolute configuration of the product. Reaction using 29 gave full conversion within one hour with 63% ee. Ligand 30 exhibited approximately
three times higher reactivity, affording full conversion within 30 min
(68% conversion within 15 min) but with lower stereoselectivity (48%
ee). Assuming that the diastereomeric ratio of atropisomers in the catalytically active complex is the same as that in the Rh complexes of 27b
(1:1.45) (Table 5), and taking into account the difference in the reactivity
of the two complexes, assumed to be equal to those from 29 and 30, we
expected an ee of 27% of the (R)-isomer in a catalytic reaction employing
27b, providing that the biphenyl ligand provided the same enantioselectivity as the binaphthyl analogues. The expected (R)-isomer was indeed
observed, but with higher ee, 41%, i.e. close to the value expected from
the pure aS isomer of 27b, which is the pseudo-enantiomer of 30. The
major diastereomer was obviously more active but less selective than the
minor one. This demonstrated that the diastereomeric mixture of flexible
ligands exhibited properties similar to those of one of the rigid ligands in
the catalytic reaction.
Table 6. Rh catalyzed hydrogenation of 31.
Entry
1
2
3
4
5
6
Ligand
27b
27c
27d
27e
29
30
Time (h)
2
20
2
1
1
0.5
Convn (%)
100
0
23
100
100
100
Ee (%)
41 (R)
21 (S)
0
63 (S)
48 (S)
3.6 Ni Catalyzed Silicon-Boron Addition to 1,3-Dienes
Transition metal catalyzed silicon-boron additions to unsaturated
compounds are powerful synthetic tools. From 1,3-dienes olefins bearing
silyl and boryl groups at the 1- and 4-positions, which are important
building blocks for further synthetic transformations, are obtained.31
21
Inspired by the recent success in our group in the development of
stereoselective silicon-boron additions to 1,3-cyclohexadiene,32 we decided to test our flexible ligands in Ni catalyzed silaborations.
Silaboration
of
1,3-cyclohexadiene
(33)
with
2(dimethylphenylsilyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (34), using
a catalyst prepared from a 1:2 mixture of Ni(acac)2 and ligand in toluene
at 80 °C was performed (Scheme 13) and the results are presented in Table 7.
Ph
Si
O
B
O
Ph
Si
34
O
B
O
Ni(acac)2 (5 mol%)
L (10 mol%)
33
35
Scheme 13. Silaboration of 1,3-cyclohexadiene (33).
When the flexible racemic aminophosphine ligand 25 was employed, the best conversion obtained was 63% after 6 h. However, when
aminophosphine 27f with a chirally rigid amine moiety was employed,
only traces of product were observed after 16 h. In case of 27b the reaction proceeded with satisfactory conversion (Table 7, entry 2), but unfortunately low stereoinduction was observed. The cis-isomer, 35, was the
only product observed in all cases.
Table 7. Silaboration of 1,3-cylcohexadiene (33).*
Entry
1
2
3
Ligand
25
27b
27f
Time (h)
6
20
16
* Reactions were performed in toluene at 80 °C.
22
Convn (%)
63
60
<5
Ee (%)
4
-
4. Steric Symmetry and Electronic
Dissymmetry
4.1 Background
A traditional approach for ligand design for palladium catalyzed
asymmetric allylic alkylations where the ligands had C1 or C2 symmetry
was previously used in our group.33 The enantiodiscrimination of C2symmetric ligands relies only on the steric properties of the ligand. An
introduction of electronic dissymmetry destroys the rotational symmetry
and the electronic properties of the ligand have a more significant influence on the selectivity of the catalytic reaction. There are reports in the
literature where pseudo-C2 P,N-ligands have been prepared, however low
to moderate enantioselectivities in palladium catalyzed allylic alkylations
were observed.34 Therefore, we decided to focus our investigation of the
catalytic behavior of P,N-ligands with pseudo-C2 and pseudo-CS symmetry like 36 and 37 (Figure 10). It is important to note that the
enantioselectivity in catalysis in case of pseudo-CS ligand would mainly
depend on the electronic effects.
N
P
36
N
P
37
Figure 10. Phospholane-pyrrolidine ligands with pseudo-C2 (36) and pseudo-CS
(37) symmetry.
Another aspect of this project was to study the ability of semi-flexible
ligands to adopt pseudo-C2 and pseudo-CS symmetry in catalytic reactions.
4.2
Synthesis of P,N-ligands with pseudo-C2 and pseudo-CS
Symmetry
In order to study the catalytic properties of P,N-ligands with
pseudo-C2 and pseudo-CS symmetry, phosphepin-azepine ligands 38 and
23
39 (Figure 11) and phospholane-pyrrolidine ligands 36 and 37 (Scheme
14) were synthesized. The latter ligands were prepared by reaction of
(S,S)-2,5-hexanediol cyclic sulfate (40) with (2-aminoethyl)phosphonic
acid diethyl ester (41), which gave pyrrolidine 42 in acceptable yield.
Reduction of 42 by lithium aluminum hydride gave the air sensitive
phosphine 43 which immediately was used in the next step. Lithiation
and treatment with 40 or ent-40, followed by BH3⋅Me2S gave BH3 protected pseudo-C2 ligand 36 and pseudo-CS ligand 37, respectively.
O
O
O
S
H2N
PO(OEt)2
41
O
75
N
PO(OEt)2
47%
40
LAH, Et2O
N
PH2
oC
43
42
H3B
BuLi, 40
P
N
BH3 . Me2S
17%
BH3
36-BH3
BuLi, ent-40
BH3 . Me2S
21%
H3B
P
N
BH3
37-BH3
Scheme 14. Synthesis of phospholane-pyrrolidine ligands 36 and 37.
N
P
38
N
P
39
Figure 11. Phosphepin-azepine ligands with pseudo-C2 (38) and pseudo-CS (39)
symmetry.
24
4.3 Pd Catalyzed Asymmetric Allylic Alkylation
In order to compare the reactivity and selectivity of the P,N-ligands
they were assessed in Pd catalyzed asymmetric allylic alkylations.≠
The ligands were used in the alkylation of rac-1,3-diphenyl-2propenyl acetate (44) with dimethyl malonate as a nucleophile in the
presence of bis(trimethylsilyl)acetamide (BSA) and KOAc (Scheme 15).
Since Pd(OAc)2 serves to deprotect phosphines,35 reactions using this
palladium source were run using BH3-protected ligands. This is convenient, as the deprotected ligands are sensitive to oxidation.
O
OAc
Ph
Ph
44
2.5 mol% Pd(OAc)2 or 1.25 mol% [(η3 -C3H5)PdCl]2
2.5 mol% ligand, (MeOCO)2CH2, BSA, KOAc
O
MeO
OMe
*
Ph
Ph
45
Scheme 15. Pd catalyzed allylic alkylation of rac-1,3-diphenyl-2-propenyl acetate (44).
Pseudo-C2 symmetric ligands 36 and 38 (Figure 11) (Table 8, entries 1 and 3) exhibited higher enantioselectivity than the pseudo-CS analogs 37 and 39 (entries 2 and 4). In case of binaphthyl based ligand 38
also higher reactivity compared to 39 was observed. When the 1:1 mixture of 38 and 39 was used in catalysis an ee of 79% of the product with
(S) absolute configuration was obtained (entry 5), showing that the reactivity of pseudo-C2 ligand 38 was found to be seven times higher than
that of pseudo-CS 39.
≠
Results from the catalysis for the ligands 38, 39 are also included for comparison.
25
Table 8. Pd catalyzed allylic alkylation of rac-1,3-diphenyl-2-propenyl acetate
(44) with malonate.*
Entry
1
2
3
4
5
Ligand
36-2BH3
37-2BH3
38-BH3
39-BH3
38-BH3/39-BH3 1:1
Time (h)
7
7
6
72
4
Convn (%)
100
100
100
95
65
Ee (%)
81 (S)
67 (R)
98 (S)
37 (R)
79(S)
* Reactions were performed in DCM at 20 °C. Pd(OAc)2 as Pd source.
For evaluation of the ligands with substrates forming anti-anti-allyl complexes, alkylations of rac-3-cyclohexenyl acetate (46) with malonate
under condition similar to those used for 44 were performed (Scheme
16).
O
OAc
2.5 mol% Pd(OAc)2 or 1.25 mol% [(η3 -C3H5)PdCl]2
O
MeO
OMe
*
2.5 mol% ligand, (MeOCO)2 CH2, BSA, KOAc
46
47
Scheme 16. Pd catalyzed allylic alkylation of rac-3-cyclohexenyl acetate (46).
Although our ligands were not well suited for small cyclic substrates, the
pseudo-CS ligands 37 and 39 afforded product 47 with higher yields and
selectivities (Table 9, entries 2 and 4) than pseudo-C2 ligands 36 and 38
(entries 1 and 3).
Table 9. Pd catalyzed allylic alkylation of rac-3-cyclohexenyl acetate (46) with
malonate.*
Entry
1
2
3
4
Ligand
36-2BH3
37-2BH3
38-BH3
39-BH3
Time (h)
120
192
24
24
T (°C)
40
40
20
20
Convn (%)
0
73
40
70
* Reactions were performed in DCM, and Pd(OAc)2 was used as Pd source.
26
Ee (%)
24 (R)
12 (R)
26 (R)
4.4 Mechanistic Considerations
As it was demonstrated above, pseudo-C2-symmetric ligands are
more suitable for linear (syn-syn) substrates like 44 in Pd catalyzed
asymmetric allylic alkylations. Due to the stronger trans influence of
phosphorus, nucleophilic attack is expected to occur trans to phosphorus.36 Due to the dissymmetry of the ligand, two diastereomeric allyl
complexes I and III, with exo and endo configuration, respectively, can
form (Scheme 17). They should be close in energy due to the effective C2
symmetry of the ligand. Formation of the olefin complexes via nucleophilic attack should occur under Curtin-Hammett conditions, and since
the transition state to form olefin complex II should be lower in energy
than to form olefin complex IV, II should be the major product, which is
in agreement with catalytic results.
Ph N
P Ph
favored
Ph
P
N
Ph
I
Nu
Nu
II
Nu
Nu
Ph
Ph N
P Ph
III
N
P
Ph
IV
Scheme 17.
The electronic dissymmetry and the different trans influence of the
two donor atoms are the key features for the enantioselectivity employing
ligand 37 with pseudo-CS symmetry# which is geometrically more suitable for cyclic (anti-anti) substrates (Scheme 18). Two allyl complexes, V
and VII, with exo and endo configurations, respectively, can form. Due to
sterical reasons, the exo complex should be more stable, and formation of
olefin complex VI after nucleophilic attack trans to the phosphorus favored. Formation of the olefin complex VIII is expected to be suppressed
#
The corresponding N,N- or P,P-ligands with CS-symmetry are not chiral and would lead to racemic
product in catalysis.
27
due to the sterical repulsion between the Nu-group and the bulky moiety
of the ligand in the transition state.
N
P
favored
Nu
N
Nu
P
VI
V
Nu
N
Nu
N
P
P
VIII
VII
Scheme 18.
The conclusion that pseudo-C2 ligands are better for linear substrates like
1,3-diphenyl-2-propenyl acetate and pseudo-CS ligands are better for
cyclic substrates led us to the idea that a ligand containing on one side a
chirally rigid moiety and on the other side a flexible moiety (semiflexible ligand) might adopt either pseudo-C2 or pseudo-CS symmetry
depending on the geometry of the substrate (Scheme 19). It was expected
that a flexible ligand would adopt pseudo-C2 symmetry in the presence of
a linear (syn-syn) substrate, whereas the presence of a cyclic (anti-anti)
substrate would lead to pseudo-CS symmetry of the ligand.
LG
Ph N
P Ph
pseudo-C2
Scheme 19.
28
N
P
LG
Ph
Ph
syn-syn
anti-anti
N
N
P
P
pseudo-CS
This encouraged us to design and synthesize semi-flexible P,N-ligands 48
and 49 which on one side have a chirally rigid bisnaphthyl moiety and on
the other side a chirally flexible biphenyl moiety.
N
P
P
N
48
49
Figure 12. Semi-flexible P,N-ligands.
Semi-flexible ligands 48 and 49 were assessed in Pd catalyzed
allylic alkylation of rac-1,3-diphenyl-2-propenyl acetate (44) (Scheme
15). Ligand 49 with the chirally flexible moiety on the phosphorus part of
the ligand behaved as the 1:1 mixture of 38 and 39 (Table 8, entry 5; Table 10, entry 2). This suggests that interconvertion of the flexible part is
slower than the nucleophilic attack at the allylic complex. When ligand
48 containing a flexible nitrogen part was used, the same ee as that expected from a 1:1 mixture of 38 and ent-39 was observed (Table 10, entry
1).
Table 10. Pd catalyzed allylic alkylation of rac-1,3-diphenyl-2-propenyl acetate
(44) with malonate.*
Entry
1
2
Ligand
48-BH3
49-BH3
Time (h)
4
5
Convn (%)
60
55
Ee (%)
87 (S)
78 (S)
* Reactions were performed in DCM at 20 °C. Pd(OAc)2 was used as Pd source.
Semi-flexible ligand 48 was also tested in the allylic alkylation of
rac-3-cyclohexenyl acetate (46) (Scheme 16). Assuming that the reactivity of catalysts containing 38 and 39 is equal, an ee of 7% of the (S)product (47) would be expected for a catalytic system containing a 1:1
mixture of 38 and ent-39 (from Table 9). However, when ligand 48 was
employed in catalysis an ee value of 20% of the (S)-enantiomer of 47 was
observed, which is closer to that expected using only ent-39. This result is
consistent with the assumption that for cyclic substrates catalysts with
pseudo-CS symmetry are more reactive.
29
4.5 Synthesis of N,N-Ligands for π-Allyl Complexes
Since the results from the catalytic reactions did not provide conclusive evidence for the preferred configuration of the different metal
complexes, we decided to study the structure of model allyl complexes.
For this purpose we decided to use the corresponding N,N-ligands, which
are air stable and easy to handle. In addition, the high symmetry of N,Nligands makes interpretation of NMR spectra of the allyl complexes easier.
C2- and CS-symmetric ligands 52 and 53, respectively, have been
previously reported in the literature.37 The ligands were obtained by separation of a mixture of enantiomers and diastereomers. Chiral ligand 52
was also obtained by enantioselective synthesis. In order to avoid resolution, we decided to perform a stepwise synthesis of the ligands (Scheme
20). Reaction of (R)-2,2’-bis-(bromomethyl)-1,1’-binaphthyl (50) with
excess of 1,2-diaminoethane gave exclusively mono-substituted diamine
51. Subsequent reaction of 51 with another molecule of 50 gave C2symmetric ligand 52. CS-symmetric ligand 53 was obtained by reaction of
51 with ent-50.
NH2
Br H2N
(130 equiv.)
Br
NEt3, PhMe
60 oC
>95%
50
51
50
N
NH2
51
NEt3, PhMe
60 o C
59%
N
N
52
ent-50
N
NEt3, PhMe
60 oC
56%
N
53
Scheme 20. Synthesis of C2- and CS-symmetric N,N-ligands 52 and 53.
The same strategy was used for the synthesis of semi-flexible
ligand 56 where first reaction of 2,2’-bis-(bromomethyl)-1,1’-biphenyl
(54) with excess of 1,2-diaminoethane gave mono-substituted amine 55,
and subsequent reaction with 50 gave semi-flexible ligand 56 in good
yield (Scheme 21).
30
NH2
Br H2 N
(130 equiv.)
Br
54
50
N
NEt3, PhMe
60 oC
>95%
NH2
55
N
NEt3, PhMe
60 oC
75%
N
56
Scheme 21. Synthesis of semi-flexible ligand 56.
Flexible ligand 5738 was prepared in a microwave assisted one step
reaction by reacting 2 equivalents of 54 with 1 equivalent of 1,2diaminoethane (Scheme 22).
NH2
Br H2N
(0.5 equiv.)
Br
NEt3, PhMe
MW
50%
54
N
N
57
Scheme 22. Synthesis of flexible ligand 57.
4.6 Preparation of Pd Allyl Complexes
In order to investigate the ability of the semi-flexible or flexible
ligands to adopt pseudo-C2 or pseudo-CS symmetry in the intermediate πallyl complexes we prepared complexes from the corresponding N,Nligands.
The allyl complex 59 was prepared by treatment of the C2symmetric ligand 52 with bis[(η3-1,3-diphenylallyl)palladium chloride]
(58) in the presence of AgPF6 (Scheme 23). Complexes 60, 61 and 62
with the CS-symmetric, semi-flexible and flexible ligands, respectively,
were successfully prepared by the same route.
31
Ph
Ph
Pd
Cl
58
N
N
2
N
N
AgPF6, DCM, 50 oC
Pd
Ph
59
52
N
PF6 -
Ph
N
N
Pd
N
N
Pd
Pd
Ph
Ph
PF6 -
Ph PF6
Ph
60
N
Ph
Ph
PF6 -
62
61
Scheme 23. Synthesis of Pd allyl complexes with 1,3-diphenylallyl moiety.
We employed the same strategy for the preparation of the complexes with the cyclic allyl moiety. Treatment of 52 with bis[(η3cyclohexenyl)palladium chloride] (63) in the presence of AgPF6 gave
complex 64 (Scheme 24). Complexes 65, 66 and 67 with the CSsymmetric, semi-flexible and flexible ligands, respectively, were also
obtained by the same procedure.
Cl
Pd
2
63
N
N
N
N
AgPF6, DCM, 50 oC
Pd
N
N
N
N
N
Pd
Pd
PF665
PF6 -
64
52
N
Pd
PF666
PF667
Scheme 24. Synthesis of Pd allyl complexes with cyclic allyl moiety.
32
4.7 NMR Investigation of Pd Allyl Complexes
First the structure of π-allyl complexes derived from ligand 52
were studied. A single complex, 64, with the cyclic allyl system was observed due to the C2 symmetry of the ligand. As expected, the complex
was devoid of symmetry and showed three separate signals, at δ 5.63,
4.41, and 3.87 ppm, for the allylic protons. Two complexes, endo and
exo, may form from CS ligand 53. Only one compound, most probably
the exo isomer 65, was observed by NMR. In accordance with the expected CS symmetry, a symmetrical spectrum was observed at 298 K,
with signals at δ 5.52 for the central allylic proton and 3.83 ppm for the
two terminal allylic protons. The pairwise identity of the methylene protons also supported the mirror symmetry of the complex. At low temperature (193 K) separate signals were observed for the side allylic protons,
probably due to symmetry breaking as a result of a slow conformational
change of the five-membered chelate ring.
The 1H NMR spectrum of the π-allyl complex 60 based on the CS
ligand 53 with the 1,3-diphenylallyl moiety suggested the presence of a
symmetric complex. Signals from the central allylic proton appeared at
δ 6.13 and the terminal allylic protons at 4.59 ppm and, again, the methylene protons were pairwise identical. The large coupling constant of
11.7 Hz for the allylic protons strongly indicated a syn-syn geometry of
the allylic moiety. The spectrum of complex 59 based on the C2symmetric ligand 52 with the 1,3-diphenylallyl moiety indicated the presence of more than one isomer. Signals for the allylic protons in the major
complex were found at δ 6.01, 5.46 and 3.81 ppm. The large coupling
constants of 10.1 and 13.2 Hz are strong indication of a syn-syn geometry
of the allylic moiety. Signals from a minor isomer were also observed,
where the central allylic proton appeared at δ 5.91 and the terminal allylic
protons at 4.50 and 4.42 ppm. The coupling constants of 12.4 and 8.1 Hz
suggest a syn-anti configuration of the 1,3-diphenylallyl moiety.
The 1H NMR spectra of the rigid ligand based complexes 59, 60 and
64, 65 were compared to those obtained from the flexible ligand 57. The
NMR spectrum of complex 67 which contains the cyclic allyl moiety
exhibited high symmetry and resembled that of 65. Signals for the central
allylic proton appeared at δ 5.60 and for the terminal protons at 4.25 ppm,
and, as in complex 65, methylene protons were pairwise identical. The
spectrum is not compatible with a chiral complex C2-67 (Figure 13) or
with a mixture of two CS complexes, but represents a single CS structure
33
(CS-67). In analogy to the rigid complex 65 the symmetry of the complex
67 was broken at lower temperature. The spectrum of complex 62 with
the 1,3-diphenylallyl moiety was somewhat more complicated than that
of 67. A symmetric spectrum was obtained for the major isomer with two
signals for the allylic protons at δ 6.18 and 4.85 ppm with a coupling
constant of 11.8 Hz. This spectrum is consistent with the symmetric structure CS-62. However, a minor isomer was present, with an 1H NMR spectrum showing a dd at δ 5.94 ppm with coupling constants J = 12.6 and
8.2 Hz. This is consistent with a syn-anti configuration of the allylic moiety. It is not clear whereas the ligand in this minor isomer has CS or C2
symmetry.
N
N
N
Pd
N
Pd
Ph
Ph
PF6-
Ph
PF6Ph
CS-62
N
C2-62
N
N
N
Pd
Pd
PF6-
PF6-
CS-67
C2-67
Figure 13.
The 1H NMR spectrum of the π-allyl complex 66, based on the
semi-flexible ligand and the cyclic allylic system, indicated the presence
of only one compound. The central allylic proton appeared at δ 5.64 and
the terminal allylic protons at 4.31 and 3.78 ppm with a coupling constant
of 6.5 Hz. The presence of three allylic signals is due to the dissymmetry
of the ligand which contains binaphthyl and biphenyl moieties. Lowering
of temperature resulted in broadening of the signals, but only three signals for the allylic protons were observed at temperatures between 193
and 298 K. Most probably, by analogy to complex 67, a complex with
pseudo-CS stereochemistry (pseudo-CS-66) was formed (Figure 14). The
34
spectrum of 61, with the 1,3-diphenylallyl moiety, revealed the presence
of three complexes (two major and one minor). The central allylic protons for both major complexes were found at δ ca. 6.1 and the four signals for the terminal allylic protons appeared at 5.61, 5.25, 4.20 and 4.05
ppm with coupling constants of 13.3, 12.7, 10.8 and 9.6 Hz, respectively.
Signals from the minor isomer were observed at δ 5.89 for the central
allylic proton and 4.76 and 4.63 ppm for the terminal allylic protons with
coupling constants of 12.8 and 8.6 Hz. The major isomer with two large
coupling constants for allylic protons most probably has a syn-syn allyl
group and, by comparison of coupling constants with those from the
complexes 59 and 60 and taking in to account results for complex 62,
pseudo-CS configuration of the ligand (pseudo-CS-61 syn-syn) (Figure
14). The other major and the minor complex have, according to the coupling constants, syn-anti configuration of the 1,3-diphenylallyl moiety. In
both complexes the ligand most probably has pseudo-CS-symmetry and
the two complexes are therefore assumed to be either pseudo-CS-61 synanti and pseudo-CS-61 syn-anti’ or the corresponding endo isomers.
N
N
N
Pd
N
Pd
Ph
Ph
PF6-
Ph
PF6Ph
pseudo-CS-61 syn-syn
N
pseudo-CS-61 syn-anti
N
N
N
Pd
Pd
Ph PF 6
PF6-
Ph
pseudo-CS-61 syn-anti'
pseudo-CS-66
Figure 14.
In order to verify that the flexible complexes are CS symmetric,
PF6– in 62 and 67 was replaced with chiral anions (Δ)-TRISPHAT (68)39
35
and (Δ)-BINPHAT (69)40 (Figure 15).& (Δ)-TRISPHAT and (Δ)BINPHAT have proven to be effective NMR chiral solvating agents for
cationic organic and organometallic species.41
Cl
Cl
Cl
Cl
Cl
Cl
Cl
O
O
O
O
Cl
Cl
O
Cl
Cl
O
P
O
O
O
P
O
Cl
Cl
O
Cl
Cl
O
Cl
Cl
Cl
68
Cl
Cl
69
Figure 15. Chiral anions (Δ)-TRISPHAT (68) and (Δ)-BINPHAT (69).
In the CS structure containing 68 or 69, enantiotopic nuclei would
become diastereotopic and therefore would give rise to separate signals,
whereas atoms residing in the mirror plane, in our case only the central
allylic proton and carton atoms, are achirotopic and would not split (Figure 16). In a chiral C2 complex two enantiomers are present and all atoms
are chirotopic and doubling of all 1H NMR and 13C NMR should result.
σ
σ
N
N
Pd
CS
N
N
N
Pd
N
Pd
C2
Figure 16.
For complex 67, where PF6– was replaced with anion 68, little
change was observed in the NMR spectrum. An effective differentiation
&
Experiments were performed at the Department of Organic Chemistry, University of Geneva in
collaboration with Prof. Jérôme Lacour.
36
was observed when anion 69 was used in 1H NMR and 13C NMR, however. The signals for the protons of the methylene groups bridging the
two nitrogen atoms split into four signals, and the terminal allyl protons
became diastereotopic and nonequivalent. Only the central allylic proton
remained unsplit in salt [67][69]. Also in the 13C NMR spectrum the central allylic atom remained unsplit whereas the signal from the terminal
allylic carbon atoms gave rise to different signals. All this information is
consistent with a CS symmetry of the complex 67.
An analogous study was performed for complex 62 which contains
the 1,3-diphenylallyl moiety. In both cases where chiral anions 68 or 69
were used a splitting of signals in 1H NMR was observed except for the
central allylic proton. Also in the 13C NMR spectrum the signals from the
central allylic carbon atom and from the terminal allylic carbon atoms
remained unsplit. The signals from the methylene carbons were affected
by the presence of the chiral anion and splitting was observed in the spectrum. The information obtained from the combined studies of the 1H and
13
C NMR spectra is consistent with a CS symmetry of ligand 57 in complex 62.
By DFT calculations# we could demonstrate that in the olefin complexes obtained after addition of a nucleophile to the 1,3-diphenylallyl
complexes, the ligand preferred to adopt C2 configuration, while the
ligands in the cyclohexenyl complexes retained its mirror symmetry.
Since the transition state for the nucleophylic addition probably is product-like, this explains the results of the catalytic reactions (see page 31).
#
DFT calculations were performed by Dr. Timofei Privalov.
37
38
5. Variation of Chelating Backbone
5.1 Background
Although ligands with pseudo-CS symmetry proved to be preferred
over those having pseudo-C2 symmetry for palladium catalyzed allylic
alkylation of cyclic substrates, low enantioselectivities were observed.
One of the crucial parameters of an organometallic catalyst with a bidentate ligand is the bite angle (βn) (Figure 17).42 Changes in the backbone of
a bidentate ligand influence the bite angle, which could lead to better
outcome of a catalytic reaction.43 Good stereoselectivities in allylic alkylation of cyclic substrates were for example achieved by Trost using
ligand 70 (Figure 17),44 which has a larger bite angle than ligands 36-39.
bridge
O
P
βn
M
P
O
NH HN
PPh2 Ph2P
70
Figure 17.
Therefore we wanted to study the influence of the backbone of P,Nligands with pseudo-C2 and pseudo-CS symmetry on the outcome of catalytic reactions.
5.2 Synthesis of P,N-ligands with Dibenzofuran Backbone
We decided to use dibenzofuran as a bridge between the P and N
donor atoms. However, as it has also been previously observed in our
group, an N atom directly linked to an aromatic ring might have too low
electron density due to conjugation of the lone pair with the aromatic
system45 to coordinate to a metal ion, resulting in monodentate coordination. Taking this into account we designed ligands 71 and 72 (Figure 18)
with pseudo-C2 and pseudo-CS symmetry, respectively, where the heteroatoms are in benzylic position.
39
O
N
O
P
71
N
P
72
Figure 18.
The synthesis of ligands 71 and 72 is presented in Scheme 25.
Treatment of dibenzofuran (73) first with BuLi followed by paraformaldehyde gave exclusively dimethanol derivative 74; no other regioisomers
were observed. Treatment of 74 with aqueous concentrated HBr gave
bisbromo derivative 75 in good yield. Next followed the key step of the
synthesis, which is the desymmetrization of the molecule, where only one
Br is substituted with phthalimide, providing compound 76. Under optimized conditions in a microwave assisted reaction the product was obtained in 57% yield. Only traces of the disubstituted product were observed in the crude 1H NMR spectrum and unreacted 75 was separated by
column chromatography. Phosphorus was introduced by treatment of 76
with triethyl phosphite, affording compound 77 in good yield. Deprotection of the amine was performed by treatment of 77 with hydrazine hydrate, which gave amine 78 in an excellent yield. Chirality was first introduced on the N part. Reaction of amine 78 with (R)-2,2’-bis(bromomethyl)-1,1’-binaphthyl (50) gave binaphthyl derivative 79. Reduction of 79 followed by treatment with BH3.Me2S gave protected
phosphine 80. Unfortunately, degradation of compound 80 by release of
BH3 followed by oxidation by air oxygen was observed by 31P NMR. The
crude material was anyway treated with LDA followed by 50 to give BH3
protected ligand 71 with pseudo-C2 symmetry. Pseudo-CS ligand 72-BH3
was obtained by an analogous procedure where ent-50 was used instead
of 50.
40
1) BuLi, TMEDA, Et2 O
2) (CHO)n
HBr (48%)
reflux
84%
O
74%
O
OH
73
O
HO
Br
Br
74
75
N2H4 . H2O
EtOH, RT
Potassium Phthalimide
MeCN, MW
O
O
57%
N
O
100%
H2N
X
(EtO)2OP
O
P(OEt)3
50, NEt3
THF
O
N
(EtO)2OP
86%
78
76, X=Br
77, X=PO(OEt)2 , 92%
1) LAH / TMSCl
THF
O
N
2) BH3 · Me2S
BH3
12%
BH3
80
79
1) LDA, THF
2) 50
H2 P
O
N
P
H3B
71-BH3
1) LDA, THF
2) ent-50
20%
O
N
P
H3B
72-BH3
Scheme 25. Synthesis of ligands 71 and 72.
5.3 Synthesis of P,N-ligands – Supramolecular Approach
In order to make the synthesis of pseudo-C2 and pseudo-CS symmetric ligands shorter and easier we considered the possibility to synthesize the P and N moieties of the ligand separately and combine them in
the last step of the synthesis. A supramolecular approach attracted our
41
attention because of its relative simplicity. Only a few examples have
been reported so far in the literature where supramolecular catalysts have
been used in catalysis.46 Scheme 26 presents a general supramolecular
approach. By combination of a N(R) part A with a P(R) part B a pseudoC2 P,N-ligand would be obtained. Analogously, the assembly of a N(R)
part A and a P(S) part C would give a pseudo-CS symmetric ligand.
P
B
*
R
N
P
S
*
R
P
*
S
C
*
R
N
*
A
pseudo-CS
*
R
N
P
R
*
pseudo-C2
Scheme 26.
We found a structure where two alkynyl parts are assembled by
Pt(II) with cis stereochemistry at Pt particularly interesting.47 Ligand 81
with pseudo-C2 symmetry can be obtained by a stepwise assembly of the
two rather simple alkynyl parts 82 and 83 (Figure 19). In order to obtain
pseudo-CS ligand either 82 or 83 has to be replaced with the corresponding enantiomer.
PEt3
Et3P
Pt
+
P
N
81
Figure 19.
42
N
82
P
83
The synthesis of the N alkynyl part is presented in Scheme 27.
Coupling of 3-(hydroxymethyl)phenyl iodide 84 with trimethylsilylacetylene under Sonogashira conditions48 gave compound 85 in good yield.
Mitsunobu reaction49 of 85 with phthalimide provided compound 86.
Treatment of 86 with hydrazine hydrate gave amine 87. The moderate
yield in this reaction might be explained by the need for purification due
to the presence of by-products. Column chromatography on silica gel was
found to be the optimal purification method. Reaction of amine 87 with
(R)-2,2’-bis-(bromomethyl)-1,1’-binaphthyl (50) in the presence of base
gave chiral compound 88 with a trimethylsilyl-protected alkynyl moiety.
Deprotection was performed by treatment of 88 with tetrabutylammonium fluoride (TBAF) in moderate yield. Since the last two reactions
were performed in THF, we decided to perform a one-pot two step reaction. First 87 was reacted with 50 under conditions as above and then
TBAF was added to the reaction mixture to perform deprotection. Compound 82 was obtained in 60% yield based on 87.
TMS
TMS
I
Phthalimide
TMS
OH
Pd(PPh3)4, CuI
NEt3 , RT
83%
OH
TMS
N
88%
85
84
O
PPh3, DIAD
DCM, THF, RT
86
O
TMS
N2H4 . H2 O
50
EtOH, RT
79%
NEt3, THF, Δ
75%
NH2
N
88
87
TBAF
THF, RT
46%
N
82
Scheme 27. Synthesis of N alkynyl moiety 82.
43
The phosphorus moiety was synthesized by a two-step procedure
from compound 85 (Scheme 28). Treatment of alcohol 85 with phosphorus tribromide in the presence of pyridine gave bromide 89. Subsequent
reaction of 89 with chiral phosphepine 90 in the presence of excess of
sodium hydride provided BH3 protected phosphorus moiety 83-BH3 in
moderate yield. It is worth to note that excess of sodium hydride cleaved
the TMS group, providing the deprotected alkyne.
BH3
PH
TMS
90
85
PBr3, Py
2.5 eq NaH
CHCl3, RT
69%
Br
THF
46%
89
P
BH3
83-BH3
Scheme 28. Synthesis of P alkynyl moiety 83.
With P and N alkynyl units in hand we began the preparation of the
bidentate ligands.50 We attempted to prepare the mono-alkynyl platinum
complex 91 under established conditions by treatment of 82 with commercially available cis-PtCl2(PEt3)2 in the presence of CuCl (Scheme 29).
Our initial hope was to obtain the product with cis geometry on platinum47 although this isomer is known to be less stable than the trans isomer.51 To our disappointment the complex 91 was obtained with trans
geometry in 62% yield.
44
cis-PtCl2(PEt3 )2
N
N
PEt3
CuCl, NEt3
benzene, 80 oC
62%
Pt
Et3P
82
Cl
trans-91
Scheme 29. Synthesis of σ-alkynyl Pt complex 91.
In order to avoid cis-trans isomerization we decided to use a platinum source with a bidentate ligand (dppe). From reaction of 82 with
PtCl2 . dppe under the same conditions as above only bissubstituted alkynyl complex 92 was isolated (Scheme 30).
Ph
P Ph
Ph
Ph P
Pt
PtCl2 . dppe
N
CuCl, NEt3
benzene, 80 o C
42%
82
N
N
92
Scheme 30.
Literature52 conditions, where only mono-substituted alkynyl complex was obtained under copper free conditions in the presence of diethylamine in dichloromethane, were also applied for our substrate 92.
However, when the reaction was performed at room temperature or in
refluxing dichloromethane, no product formation was observed by 31P
NMR. Further optimizations of conditions are necessary in order to obtain pseudo-CS complexes and pseudo-C2 complexes like 81.
45
5.4 Pd Catalyzed Asymmetric Allylic Alkylation
The sterically bulky bisnaphthyl substituents are situated rather
close to each other in the palladium complex with ligands 71 and 72. This
might result in a small cavity containing the catalytically active centre,
which could be suitable for small cyclic substrates. Ligands 71 and 72
were tested in the alkylation of rac-3-cyclohexenyl acetate (48) (Scheme
16). To our disappointment, no formation of product was observed. In
order to compare the reactivity and selectivity of pseudo-C2 and pseudoCS symmetric ligands 71 and 72, respectively, with those of 36, 37, 46
and 47, they were submitted to palladium catalyzed allylic alkylation of
rac-1,3-diphenyl-2-propenyl acetate (44) with malonate (Scheme 15).
The results are presented in Table 11. Unfortunately both ligands, 71 and
72, exhibited low reactivity as well as low selectivity (entries 1 and 2).
Table 11. Pd catalyzed allylic alkylation of rac-1,3-diphenyl-2-propenyl acetate (44) with malonate.*
Entry
1
2
3
Ligand
71-BH3a
72-BH3a
92b
Time (h)
5
16
24 / 72
Convn (%)
2
3
12 / 30
Ee (%)
10 (S)
6 (S)
9 (S) / 9 (S)
* Reactions were performed in DCM at 20 °C. Pd source: a Pd(OAc)2, b [(η3-C3H5)PdCl]2.
In order to study the catalytic activity of the supramolecular ligand
92, it was used in the allylic alkylation of rac-1,3-diphenyl-2-propenyl
acetate (44) (Scheme 15, Table 11, entry 3). A conversion of 12% was
obtained after 24 hours and 30% after 72 hours (Table 11, entry 3). In
both cases ee’s of 9% were observed.
5.5 Synthesis of Pyridine Based Ligands
Previously in our group ligands 93-96 (Figure 20) were synthesized
and used in Pd catalyzed allylic alkylations.53
46
OMe
N
N
N
Ph
O
O
Ph2P
N
Ph
O
PPh2
94
93
OMe
Ph
Ph2P
95
O
Ph2P
96
Figure 20.
Structurally similar ligands, 97 (Figure 21), were successfully applied by
Pfaltz in Ir catalyzed hydrogenation of unfunctionalized olefins.54 ,55
∗ R
N
O
Ph2P
R= Me, Ph, tBu, Trityl
97
Figure 21.
In order to test ligands 93-96 in Ir catalyzed hydrogenations, we
decided to prepare the ligands. Chiral enantiopure 2-(1hydroxyalkyl)pyridines 101 and 102 were previously prepared in our
group.56,57 Treatment of 2-bromopyridine (98) with butyllithium followed
by nitrile 99 gave after acidic work-up ketone 100, which was reduced to
the diastereomeric alcohols 101 and 102 in a ratio of 83:17 (Scheme 31).
Treatment of the alcohols first with butyllithium followed by diphenylphosphine chloride and then BH3⋅Me2S gave BH3 protected ligands 93
and 94.
47
1) BuLi, Et2O
NC
2)
NaBH4
99
N
Br
41%
MeOH
N
+
N
N
98
HO
OH
O
100
102 16%
101 74%
43%
1) BuLi, THF
2)Ph2PCl
3)BH3 . Me2S
62%
1) BuLi, THF
2)Ph2PCl
3)BH3 . Me2S
N
N
O
O
Ph2P
BH3
93-BH3
H3B
PPh2
94-BH3
Scheme 31. Synthesis of pyridyl phosphate ligands 93 and 94.
Pyridyl phosphinite ligands 95 and 96 (Figure 20), which contain
moieties derivated from mandelic acid, were prepared according to previously reported procedures.53
Encouraged by results from iridium catalyzed asymmetric hydrogenation of unfunctionalized olefins using ligands 93 and 94 (see the next
paragraph), we designed ligands 103 and 104 where the diphenylphosphino moiety is replaced by chirally flexible phosphepine moiety
with a hope to improve the enantioselectivity. Treatment of the chiral
alcohol 101 or 102 first with butyllithium followed by phosphine chloride
26 and BH3.Me2S gave BH3 protected ligands 103 and 104, respectively
(Scheme 32).
48
1) BuLi, THF
2) 26, THF
101
N
O
.
3) BH3 Me2S, THF
H3 B
46%
P
103
102
1) BuLi, THF
2) 26, THF
N
O
3) BH3 . Me2S, THF
P
BH3
35%
104
Scheme 32. Synthesis of flexible pyridyl phosphate ligands 103 and 104.
5.6 Ir catalyzed Hydrogenation of Unfunctionalized Olefins
Hydrogenations of unfunctionalized olefins were performed in
collaboration with the group of Pfaltz& under previously developed conditions with 1 mol% catalyst loading (Scheme 33).54
R
R1
Ir catlyst (1 mol%)
2 h, 50 bar H2
R
1
∗ R
DCM, RT
Scheme 33. Ir catalyzed hydrogenation of unfunctionalized olefins.
The results for the selected substrates 105-110 (Figure 22) are presented
in Table 12. Ligands 93, 94 and 95 in general gave high conversion with
variable ee’s. The highest ee was observed for substrate 109 with ligand
93. Significant lowering of reactivity was observed for ligand 96. This
could be explained by the presence of the bulky phenyl substituent in 6position in pyridine ring.
&
Department of Chemistry, University of Basel, Switzerland.
49
MeO
MeO
106
105
CO2Et
107
OH
MeO
108
109
110
Figure 22. Substrates for Ir catalyzed hydrogenation.
Table 12. Ir catalyzed hydrogenation.*
93a
94a
95a
96a
104b
Convn (%) /
Convn (%) /
Convn (%) /
Convn (%) /
Convn (%) /
ee (%)
ee (%)
ee (%)
ee (%)
ee (%)
105
55/26(S)
72/17(R)
35/80(S)
0/-
97/26
106
99/15(R)
99/19(S)
99/65(S)
46/55(S)
-
107
99/60(R)
99/57(S)
97/56(R)
17/75(R)
99/56
108
21/17(S)
9/16(R)
3/rac
1/12(R)
-
109
99/92(+)
50/84(-)
95/81(+)
5/14(+)
99/79
110
-
99/74(S)
99/rac
12/34(R)
99/67
Ligand
Substrate
* Reactions were performed in DCM at RT, 1 mol% catalyst loading, 50 bar H2 pressure. Reaction
time: a 2 h;
b
3 h.
As mentioned in the previous paragraph, based on the results from
catalysis for ligands 93 and 94 we designed and prepared ligands with
chirally flexible phosphepine moieties 103 and 104. Preliminary results
for ligand 104 are presented in Table 12. For substrate 105 slightly higher
ee was obtained compared to the ligand 94. There was no improvement
observed in case of substrate 107 and slightly lower enantioselectivities
were observed for substrates 109 and 110 compared to the results from
ligand 94.
50
6. Concluding Remarks
Configurationally flexible monodentate phosphepine ligands have
been synthesized and their properties studied. Rotation barriers for the
biphenyl unit of phosphepines have been determined. It has been found
that flexible phosphepine ligands form homochiral complexes with Rh(I)
having a ligand:Rh ratio of 2:1. High selectivity for formation of complexes containing only the same type of ligand was observed from an
equimolar mixture of two different ligands and [Rh(COD)2]+BF4–. Flexible phosphepine ligands were tested in the Rh catalyzed hydrogenation of
α-acetamidocinnamate. Use of the flexible ligands resulted in a selectivity close to that of one of the corresponding rigid ligands. Flexible
phosphepine ligands were also tested in the Ni catalyzed silaboration of
1,3-cyclohexadiene, however a weak chirality transfer was observed.
P,N-ligands with pseudo-C2 and pseudo-CS symmetry have been
synthesised and used in Pd catalyzed asymmetric allylic alkylations. A
significant impact of electronic dissymmetry and steric symmetry on the
outcome of the catalytic reactions studied was demonstrated. Semiflexible P,N-ligands, which, depending on the structure of the substrate,
may adopt either pseudo-C2 and pseudo-CS configuration were synthesised and tested in allylic alkylations. The results obtained from the alkylation of rac-1,3-diphenyl-2-propenyl acetate showed that flexible ligands
act as a 1:1 mixture of rigid ligands, but that the activity and the selectivity of the complex containing a ligand with pseudo-C2 symmetry are
higher. In contrast, in the alkylation of rac-3-cyclohexenyl acetate the
complex containing the pseudo-CS isomer exhibited higher activity and
was more selective. Model allyl systems were studied and a preference of
a flexible ligand to adopt CS structure in complexes with both syn-syn
and anti-anti allylic systems was found. For the former type of system a
C2 structure was, however, found by computations to be more stable for
the olefin complex formed by additions of a nucleophile. Further studies
of model olefin complexes have to be performed in order to assess the
ability of flexible ligands to adopt C2 symmetry in olefin complexes obtained from syn-syn allyl complexes.
Ligands with different backbones were prepared and tested in catalytic reactions (allylic alkylation and Ir catalyzed hydrogenation) using
both conventional and supramolecular methods. A supramolecular approach to ligand preparation has a high potential for future rapid access to
51
ligand libraries. This method can give easy access to ligands with variable bite angles as well as allow easy preparation of ligands with different
donor atoms and different symmetries.
52
Acknowledgements
First of all, I would like to express my gratitude to my supervisor
Prof Christina Moberg for accepting me as graduate student, for giving
me an interesting project and teaching me the art of the science.
Then I would like to acknowledge The Swedish Research Council
for financial support and The Aulin-Erdtman Foundation, Knut och Alice
Wallenbergs stiftelse and Svenska Kemistsamfundet for making it possible to present this work on several conferences.
I also want to thank:
Dr. Brigita Vīgante for “opening my eyes” - without that I would
never have applied for a PhD student position abroad - and Pēteris Trapencieris for an encouragement to study abroad.
All present and former members of the Ki group, especially Robert
Stranne, Jean-Luc Vasse, Serghey Lutsenko and Fredrik Lake for giving
me good advice in the lab at the beginning of my studies at KTH; Anders
Frölander, Sari Paavola, Maël Penhoat, Dean Gillings, Antonio Fonduca,
Andrei Khalimon, Ester Fjellander and Mélanie Tilliet for being good
friends and making the environment in the lab friendly.
My dear colleagues Zoltán Szabó, Ester Fjellander, Timofei
Privalov, Andrei Khalimon, Martin Gerdin and Christian Böing for valuable help and excellent collaboration.
Everyone at the department of Organic Chemistry, especially Ulla
Jacobsson for all help with NMR; Krister Zetterberg, Olof Ramström and
Tobias Rein for valuable discussions about different topics; Henry Challis, Jan Siden, Jonas Hellberg, Lena Skowron and Ingvor Larsson for all
their help.
Gunārs Tirzītis and Arkadij Sobolev for just being who you are.
My family, especially my parents – Irēna and Arnolds for support
and encouragement.
Finally I would like to express my gratitude to Antra for endless
patience, understanding and love during the years.
53
54
References
1
Rouhi, A. M. Chem. Eng. News 2003, 81, 45-55.
2
a) Kuhn, R. ”Molekulare Asymmetrie,” in Freudenberg, K. Ed., Stereochemie, Franz Deutike: Leipzig, Germany, 1933; p 803. b) Christie, G. H.
and Kenner, J. H. J. Chem. Soc. 1922, 121, 614-620 (references are taken
from ref. 3).
3
Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Organic
Compounds. Wiley: New York, 1994; pp 1142-1147.
4
Ōki, M. Top. Stereochem. 1983, 14, 1-81.
5
Mikami, K.; Aikawa, K.; Yusa, Y.; Jodry, J. J.; Yamanaka, M. Synlett
2002, 1561-1578.
6
Bringmann, G.; Price Mortimer, A. J.; Keller, P. A.; Gresser, M. J.; Garner, J.; Breuning, M. Angew. Chem. Int. Ed. 2005, 44, 5384-5427.
7
a) Bringmann, G.; Menche, D.; Kraus, J.; Mühlbacher, J.; Peters, K.;
Peters, E. M.; Brun, R.; Bezabih, M.; Abegaz, B. M. J. Org. Chem. 2002,
67, 5595-5610. b) Kuroda, M.; Mimaki, Y.; Sakagami, H.; Sashida, Y. J.
Nat. Prod. 2003, 66, 894-897.
8
Fukuyama, Y.; Asakawa, Y. J. Chem. Soc. Perkin Trans. 1 1991, 27372741.
9
Bringmann, G.; Tasler, S.; Endress, H.; Kraus, J.; Messer, K.; Wohlfarth,
M.; Lobin, W. J. Am. Chem. Soc. 2001, 123, 2703-2711.
10
Miyashita, A.; Yasuda, A.; Takaya, H.; Toriumi, K.; Ito, T.; Souchi, T.;
Noyori, R. J. Am. Chem. Soc. 1980, 102, 7932-7934.
11
Schmid, R.; Foricher, J.; Cereghetti, M.; Schönholzer, P. Helv. Chim.
Acta 1991, 74, 370-389.
12
a) Almenningen, A.; Bastiansen, O.; Fernholt, L.; Cyvin, B. N.; Cyvin,
S. J.; Samdal, S. J. Mol. Struct. 1985, 128, 59-76. b) Bastiansen, O.;
Samdal, S. J. Mol. Struct. 1985, 128, 115-125.
13
Grein, F. J. Phys. Chem. A 2002, 106, 3823-3827.
14
Bott, G.; Field, L. D.; Sternhell, S. J. Am. Chem. Soc. 1980, 102, 56185626.
15
Desponds, O.; Schlosser, M. Tetrahedron Lett. 1996, 37, 47-48.
16
Meca, L.; Reha, D.; Havlas, Z. J. Org. Chem. 2003, 68, 5677-5680, and
references therein.
55
17
a) Sutherland, I. O.; Ramsay, M. V. J. Tetrahedron 1965, 21, 34013408. b) Walsh, P. J., Lurain, A. E.; Balsells, J. Chem. Rev. 2003, 103,
3297-3344. c) Superchi, S.; Bisaccia, R.; Casarini, D.; Laurita, A.; Rosini,
C. J. Am. Chem. Soc. 2006, 128, 6893-6902.
18
a) Mislow, K.; Glass, M. A. W.; Hopps, H. B.; Simon, E.; Wahl, G. H.
J. Am. Chem. Soc. 1964, 86, 1710-1733. b) Ōki, M.; Iwamura, H. Tetrahedron 1967, 24, 2377-2383.
19
a) Pastor, S. D.; Richardson, C. F.; Nabirahni, M. A. Phosphorus, Sulfur Silicon Relat. Elem. 1994, 90, 95-102. b) Whiteker, G. T.; Harrison, A.
M.; Abatjoglou, A. G. J. Chem. Soc., Chem. Commun. 1995, 1805-1806.
20
Monti, C.; Gennari, C.; Piarulli, U. Chem. Commun. 2005, 5281-5283.
21
Mikami, K.; Korenaga, T.; Terada, M.; Ohkuma, T.; Pham, T.; Noyori,
R. Angew. Chem. Int. Ed. 1999, 38, 495-497.
22
Tudor, M. D.; Becker, J. J.; White, P. S.; Gagné, M. R. Organometallics
2000, 19, 4376-4384.
23
Gladiali, S.; Dore, A.; Fabbri, D.; De Lucchi, O.; Manassero, M.
Tetrahedron: Asymmetry 1994, 5, 511-514.
24
Bitterer, F.; Herd, O.; Kühnel, M.; Stelzer, O.; Weferling, N.; Sheldrick,
W. S.; Hahn, J.; Nagel, S.; Rösch, N. Inorg. Chem. 1998, 37, 6408-6417.
25
a) Xiao, D.; Zhang, Z.; Zhang, X. Org. Lett. 1999, 1, 1679-1681. b)
Stranne, R.; Vasse, J. L.; Moberg, C. Org. Lett. 2001, 3, 2525-2528. c)
Chi, Y.; Zhang, X. Tetrahedron Lett. 2002, 43, 4849-4852. d) Junge, K.;
Oehme, G.; Monsees, A.; Riermeier, T.; Dingerdissen, U.; Beller, M.
Tetrahedron Lett. 2002, 43, 4977-4980. e) Junge, K.; Hagemann, B.;
Enthaler, S.; Spannenberg, A.; Michalik, M.; Oehme, G.; Monsees, A.;
Riermeier, T.; Beller, M. Tetrahedron: Asymmetry 2004, 15, 2621-2631.
f) Reetz, M. T.; Mehler, G.; Meiswinkel, A. Tetrahedron: Asymmetry
2004, 15, 2165-2167. g) Reetz, M. T.; Mehler, G.; Meiswinkel, A. Sell, T.
Tetrahedron: Asymmetry 2004, 15, 3483 (corrections). h) Kasák, P.;
Mereiter, K.; Widhalm, M. Tetrahedron: Asymmetry 2005, 16, 34163426.
26
a) Robinson, C. N.; Pettit, W. A. Tetrahedron Lett. 1972, 13, 49774978. b) Soulier, E.; Clément, J. C.; Yaouanc, J. J.; des Abbayes, H. Tetrahedron Lett. 1998, 39, 4291-4294. c) Soulier, E.; Yaouanc, J. J.; Laurent, P.; des Abbayes, H.; Clément, J. C. Eur. J. Org. Chem. 2000, 34973503.
27
Junge, K.; Oehme, G.; Monsees, A.; Riermeier, T.; Dingerdissen, U.;
Beller, M. J. Organomet. Chem. 2003, 675, 91-96.
56
28
Bentrude, W. G.; Setzer, W. N. Phosphorus-31 NMR Spectroscopy in
Stereochemical Analysis. (Verkade, J. G.; Quin, L. D. Eds.), VCH:
Deerfield Beach, 1987; pp 380-382.
29
Forsén, S.; Hoffman, R. A. J. Chem. Phys. 1963, 39, 2892-2901.
30
Monti, C.; Gennari, C.; Piarulli, U.; de Vries, J. G.; de Vries, A. H. M.;
Lefort, L. Chem. Eur. J. 2005, 11, 6701-6717.
31
Suginome, M.; Ito, Y. Chem. Rev. 2000, 100, 3221-3256.
32
Gerdin, M.; Moberg, C. Adv. Synth. Catal. 2005, 347, 749-753.
33
a) Nordström, K.; Macedo, E.; Moberg, C. J. Org. Chem. 1997, 62,
1604-1609. b) Wärnmark, K.; Stranne, R.; Cernerud, M.; Terrien, I.;
Rahm, F.; Nordström, K.; Moberg, C. Acta Chem. Scand. 1998, 52, 961963. d) Stranne, R.; Moberg, C. Eur. J. Org. Chem. 2001, 2191-2195. e)
Hallman, K.; Frölander, A.; Wondimagegn, T.; Svensson, M.; Moberg, C.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5400-5404.
34
a) Robert, F., Gaillard, N., Sinou, D. J. Mol. Cat. A: Chem. 1999, 144,
473-476. b) Kim, Y. K., Lee, S. J., Ahn, K. H. J. Org. Chem. 2000, 65,
7807-7813.
35
Riegel, N.; Darcel, C.; Stéphan, O.; Jugé, S. J. Organomet. Chem.
1998, 567, 219-233.
36
Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33, 336-345.
37
a) Mazaleyrat, J. P.; Cram, D. J. J. Am. Chem. Soc. 1981, 103, 45854586. b) Mazaleyrat, J. P. Tetrahedron: Asymmetry 1997, 8, 2709-2721.
38
Costa, A. M.; Jimeno, C.; Gavenonis, J.; Carroll, P. J.; Walsh, P. J. J.
Am. Chem. Soc. 2002, 124, 6929-6941.
39
Favarger, F.; Goujon-Ginglinger, C.; Monchaud, D.; Lacour, J. J. Org.
Chem. 2004, 69, 8521-8524.
40
Lacour, J.; Londez, A.; Goujon-Ginglinger, C.; Buss, V.; Bernardinelli,
G. Org. Lett. 2000, 2, 4185-4188.
41
Lacour, J.; Hebbe-Viton, V. Chem. Soc. Rev. 2003, 32, 373-382.
42
Van Leeuwen, P. W. N. M.; Kamer, P. C. J.; Reek, J. N. H.; Dierkes, P.
Chem Rev. 2000, 100, 2741-2769.
43
a) Dierkes, P.; van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans.
1999, 1519-1529. b) Van Haaren, R. J.; Druijven, C. J. M.; van
Strijdonck, G. P. F.; Oevering, H.; Reek, J. N. H.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans. 2000, 1549-1554.
44
Trost, B. M.; Bunt, R. C. J. Am. Chem. Soc. 1994, 116, 4089-4090.
57
45
Smith, M. B.; March, J. March’s Advanced Organic Chemistry. WileyInterscience: New York, 2001; pp 40-42.
46
Breit, B. Angew. Chem. Int. Ed. 2005, 44, 6816-6825.
47
Hua, J.; Lin, W. Org. Lett. 2004, 6, 861-864.
48
Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975, 16,
4467-4470.
49
Mitsunobu, O. Synthesis 1981, 1-28.
50
Jiang, H.; Lin, W. J. Am. Chem. Soc. 2004, 126, 7426-7427.
51
Long, N. J.; Williams, C. K. Angew. Chem. Int. Ed. 2003, 42, 25862617.
52
Furlani, A.; Licoccia, S.; Russo, M. V.; Villa, A. C.; Guastini, C. J.
Chem. Soc., Dalton Trans. 1984, 2197-2206.
53
Rahm, F.; Fischer, A.; Moberg, C. Eur. J. Org. Chem. 2003, 4205-4215.
54
Drury III, W. J.; Zimmermann, N.; Keenan, M.; Hayashi, M.; Kaiser,
S.; Goddard, R.; Pfaltz, A. Angew. Chem. Int. Ed. 2004, 43, 70-74.
55
Bell, S.; Wüstenberg, B.; Kaiser, S.; Menges, F.; Netscher, T.; Pfaltz, A.
Science 2006, 311, 642-644.
56
Adolfsson, H.; Nordström, K.; Wärnmark, K.; Moberg, C. Tetrahedron:
Asymmetry 1996, 7, 1967-1972.
57
Correction of the determination of absolute stereochemistry of the
pyridine alcohols: Cunningham, D.; Gallagher; E. T.; Grayson, D. H.;
McArdle, P. J.; Storey, C. B.; Wilcock, D. J. J. Chem. Soc., Perkin Trans.
1 2002, 2692-2698.
58