Intermolecular interactions in (arene)chromium carbonyl compounds: Prediction of chiral crystal packing from racemate structure

Article abstract of DOI:10.1021/ja012141m

Six X-ray crystal structures are reported, all containing substituted triphenylmethanol derivative 4 either alone or as its mono or bis(chromium tricarbonyl) complexes. All four chromium complexes studied crystallize with two independent molecules in the

Full text of DOI:10.1021/ja012141m

Published on Web 04/12/2002
Intermolecular Interactions in (Arene)chromium Carbonyl
Compounds: Prediction of Chiral Crystal Packing from
Racemate Structure
Susan E. Gibson (ne´e Thomas),* Hasim Ibrahim, and Jonathan W. Steed*
Contribution from the Department of Chemistry, King’s College London, Strand,
London WC2R 2LS, U.K.
Received September 6, 2001
Abstract: Six X-ray crystal structures are reported, all containing substituted triphenylmethanol derivative
4 either alone or as its mono or bis(chromium tricarbonyl) complexes. All four chromium complexes studied
crystallize with two independent molecules in the crystallographic asymmetric unit. It is demonstrated that
from the X-ray crystal structure of the acentric racemic (()-(1pR,1′′R)(1pS,1′′S)-[Cr(CO)3(η6-t-Bu-
C6H3(CMeOMe)CPh2OH)], (()-3, it is possible to deduce the 4-fold helical structure of the chiral
(-)-(1pR,1′′R) isomer, (-)-3. The bimetallic derivatives demonstrate the ability to control intermolecular
interactions by the positioning of relative stereochemistry.
their predictive value, increases.5,9 Thus the predominance of
the strongly hydrogen-bonded carboxylic acid dimer allows the
Introduction
Crystal engineering is a science that aims to both predict and
control solid-state architecture: connectivity, mutual orienta-
tions, and symmetry.1-9 Successful crystal engineering has
implications in the fabrication of nonlinear optical (NLO)
materials (for which crystal polarity is a prerequisite), polymorph
control, control of solid-state reactivity, and the design of
composite materials.10-19 These goals have largely yet to be
realized;20 however, the identification and prediction of recog-
nizable, recurring solid-state packing motifs are a reality.3,21 In
cases where such motifs can be isolated from other interactions
of importance, the reproducibility of these motifs, and hence
design of numerous networks and topologies.22 Similarly, many
general rules can be laid down about crystal packing within a
wide variety of organic and metallo-organic systems.4 More
specific generalizations may often be made as well, as demon-
strated by Etter.23 The introduction of polarity into solid-state
systems is of particular interest in the design of NLO materials.24
While this is generally possible to engineer given a chiral
molecule (resolved chiral compounds must crystallize in polar
space groups in the absence of disorder), it is of much greater
economy to be able to engineer polar and/or chiral solid-state
structures from achiral or racemic building blocks. The occur-
rence of bulk crystal chirality (and its spontaneous resolution)
in the absence of molecular chirality is much more haphazard,
however, and often limited to very specific compounds.
Recently one of us has demonstrated the occurrence of crystal
chirality arising from helix formation as a result of weak CH‚‚‚
OdP interactions irrespective of the handedness of one of two
chiral centers in the phosphonate building block.25 We have also
recently carried out studies on nonglass solid-state structures
with a large number of crystallographically independent mol-
ecules (i.e., Z′ > 1)26,27 with a view to introducing “designer
asymmetry”.14,28-30 Independently, two of us have developed
a preparative route that allows the efficient enantio- and
(1) Desiraju, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 2311.
(2) Braga, D.; Maini, L.; Grepioni, F. Angew. Chem., Int. Ed. 1998, 37, 2240-
2242.
(3) Thalladi, V. R.; Goud, B. S.; Hoy, V. J.; Allen, F. H.; Howard, J. A. K.;
Desiraju, G. R. Chem. Commun. 1996, 401-402.
(4) Braga, D.; Grepioni, F.; Desiraju, G. R. Chem. ReV. 1998, 98, 1375.
(5) Desiraju, G. R. In The Crystal As a Supramolecular Entity; Lehn, J. M.,
Ed.; John Wiley & Sons: Chichester, 1996; Vol. 2.
(6) Desiraju, G. R. The Design of Organic Solids; Elsevier: Amsterdam, 1989.
(7) Burrows, A. D.; Harrington, R. W.; Mahon, M. F.; Price, C. E. J. Chem.
Soc., Dalton Trans. 2000, 3845-3854.
(8) Guliera, G.; Steed, J. W. Chem. Commun. 1999, 1563-1564.
(9) Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond; OUP/IUPAC:
Oxford, 1999.
(10) Sarma, J. A. R. P.; Allen, F. H.; Hoy, V. J.; Howard, J. A. K.; Thaimattam,
R.; Biradha, K.; Desiraju, G. R. Chem. Commun. 1997, 101-102.
(11) Zaworotko, M. J. Chem. Soc. ReV. 1994, 23, 283.
(12) Thalladi, V. R.; Boese, R.; Brasselet, S.; Ledoux, I.; Zyss, J.; Jetti, R. K.
R.; Desiraju, G. R. Chem. Commun. 1999, 1639-1640.
(13) Steed, J. W.; Atwood, J. L. Supramolecular Chemistry: An Introduction;
J. Wiley & Sons: Chichester, 2000.
(22) Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D.; Vo¨gtle, F. ComprehensiVe
Supramolecular Chemistry, 1st ed.; Pergamon: Oxford, 1996; Vol. 6.
(23) Etter, M. C. Acc. Chem. Res. 1990, 23, 120-126.
(24) Marder, S. R. Metal Containing Materials for Nonlinear Optics, 2nd ed.;
Wiley: Chichester, 1996; pp 121-169.
(14) Prince, P. D.; McGrady, G. S.; Steed, J. W. New J. Chem. 2001, in press.
(15) Dunitz, J. D.; Fillippini, G.; Gavezzotti, A. HelV. Chim. Acta 2000, 83,
2317-2335.
(16) Anwar, J. J. Pharm Pharmacol. 1998, 50, 51.
(25) Forristal, I.; Lowman, J.; Afarinkia, K.; Steed, J. W. CrystEngComm 2002,
457-461.
(17) Kordikowski, A.; Shekunov, T.; York, P. Pharm. Res. 2001, 18, 682-
688.
(26) Steiner, T. Acta Crystallogr., Sect. B 2000, 56, 673-676.
(27) Brock, C. P. J. Res. Natl. Inst. Stand. Technol. 1996, 101, 321-325.
(28) Hassaballa, H.; Steed, J. W.; Junk, P. C.; Elsegood, M. R. J. Inorg. Chem.
1998, 37, 4666-4671.
(18) Schmidt, G. M. J. J. Chem. Soc. 1964, 2014.
(19) Nichols, P. J.; Raston, C. L.; Steed, J. W. Chem. Commun. 2001, 1062-
1063.
(20) Gavezzotti, A. Acc. Chem. Res. 1994, 27, 309-314.
(21) Allen, F. H.; Howard, J. A. K.; Hoy, V. J.; Desiraju, G. R.; Reddy, D. S.;
Wilson, C. C. J. Am. Chem. Soc. 1996, 118, 4081.
(29) Steed, J. W.; Hassaballa, H.; Junk, P. C. Chem. Commun. 1998, 577-578.
(30) Steed, J. W.; Sakellariou, E.; Junk, P. C.; Smith, M. Chem.-Eur. J. 2001,
7, 1240-1247.
9
10.1021/ja012141m CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 5109-5116
5109

A R T I C L E S
Gibson et al.
Scheme 1. Synthesis of New Compounds
The X-ray crystal structure of (()-3 exhibits two crystallo-
graphically independent molecules (Z′ ) 2, see Table 1 for
crystallographic data for new compounds). Intriguingly, the two
molecules in the asymmetric unit are enantiomers (1pR,1′′R and
1pS,1′′S), and yet they do not give rise to a centrosymmetric
structure. In fact, the compound crystallizes in the noncen-
trosymmetric but achiral space group Pna21. Bond lengths and
angles within the two independent molecules are within normal
bounds and will not be discussed further. Consistent with Etter’s
rules,23 it would be anticipated that the crystal packing would
be dominated by hydrogen bonds involving the hydroxyl
functionality, and indeed this proves to be the case. Starting
with molecule number 1, based on Cr(1) (an SS enantiomer),
the OH group interacts with the strongest hydrogen bond
acceptor, the OMe functionality (O(2)‚‚‚O(6)), on the adjacent,
independent RR enantiomer, molecule 2, within the same
asymmetric unit. Molecule 2, in turn, hydrogen bonds to the
OMe group on a further RR enantiomer (O(7)‚‚‚O(1)′, generated
from molecule 1 by glide symmetry). This RR enantiomer then
interacts via an O(2)‚‚‚O(6) hydrogen bond to an SS enantiomer
generated from molecule 2. The result is an infinite hydrogen-
bonded chain comprising the following sequence: SS‚‚‚RR‚‚‚
RR‚‚‚SS‚‚‚SS‚‚‚RR‚‚‚ etc., that is, alternating homochiral and
heterochiral pairings. The angle between the heterochiral pair
(hydrogen bond O(2)‚‚‚O(6)) is ca. 30°(Figure 1a), whereas
between the homochiral pair the second independent hydrogen
bond induces a twist of ca. 90°, approximating to a 41 screw
operation (Figure 1b). Interestingly, when the homochiral pair
is subjected to the crystallographic 21 screw operation that forms
part of the space group symmetry, the result is a 4-fold helix
with pitch 15.28 Å. The helix comprises four molecules all of
the same handedness, held together by alternating homochiral
hydrogen bonds and edge-to-face π-stacking interactions, Figure
2.
Racemic (()-3 cocrystallizes with two molecules of dichlo-
romethane. As the next strongest hydrogen bond donors, the
CH protons of the CH2Cl2 molecules interact with the next
strongest hydrogen bond acceptors, the CO ligands, and fulfill
a space-filling role.
diastereoselective synthesis of a wide range of arene tricarbo-
nylchromium(0) complexes of type 1.31,32 We now report a
unique study in crystal engineering that brings together highly
stereoselective molecular and crystal synthesis allowing the
preparation and control of a range of resolved, chiral solid-
state compounds with Z′ > 1.
The structure of (()-3 is extremely interesting from a crystal
engineering viewpoint because it gives significant insight into
the interactions between pairs of identical enantiomers without
the need to prepare the enantiomerically pure material. This
enables us to make confident predictions about the structure of
a resolved enantiomer of 3, viz (-)-3. In the absence of the
opposite enantiomer, only homochiral pairings will result, giving
a hydrogen-bonded chain of type RR‚‚‚RR‚‚‚RR‚‚‚RR‚‚‚. This
pairing arises from a crystallographic glide operation in (()-3,
but, in a resolved material, it is not possible for the compound
to crystallize in an achiral space group (i.e., inversion, mirror,
and glide operations are ruled out). Thus, to maintain the same
homochiral hydrogen-bonded interaction in (-)-3 as observed
in (()-3 there must necessarily be two independent molecules
of the same chirality (Z′ ) 2). Note that while Z′ is also equal
to 2 in (()-3 this is for an entirely different reason, the presence
of two independent types of hydrogen bond (homochiral and
heterochiral). If we subtract glide operations from the achiral
space group of (()-3 (Pna21), we are left with the chiral space
group P21. Thus, a possible structure for (-)-3 (or its enanti-
omer) would be a 4-fold hydrogen-bonded helix comprising two
Results and Discussion
Treatment of the sterically hindered racemic arene tricarbo-
nylchromium(0) (()-2 with LiTMP (lithium tetramethylpiperi-
dine) at -78 °C followed by quenching with 3 equiv of
benzophenone leads to clean ortho-substitution to give the
racemic monoalcohol (()-3 in 93% yield and over 96%
diastereomeric excess, Scheme 1. This efficient diastereoselec-
tivity is the result of the preferred conformation in the lithiated
intermediate which minimizes steric interactions between the
benzylic group and the Cr(CO)3 fragment.33 The analogous
reaction with (+)-2, generated using chiral base methodology,31
affords almost enantiomerically pure (-)-3 (>96% de, >93%
ee).
(31) Cowton, E. L. M.; Gibson (ne´e Thomas), S. E.; Schneider, M. J.; Smith,
M. H. Chem. Commun. 1996, 839-840.
(32) Ibrahim, H.; Gibson, S. E. Chem. Commun. 2001, 1070-1071.
(33) Heppert, J. A.; Aube´, J.; Thomas-Miller, M. E.; Milligan, M. L.;
Takusagawa, F. Organometallics 1990, 9, 727.
9
5110 J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002

Interactions in (Arene)chromium Carbonyl Compounds
A R T I C L E S
Table 1. Crystallographic Data for New Compounds
(±)-4
(±)-4‚Me2CO
(±)-3‚2CH Cl2
(−)-3‚2CH Cl2
(±)-5a
(±)-5b‚CH Cl2
2
2
2
formula
formula weight
(g mol-1
C26H30O2
374.5
C29H36O3
432.58
C30H32Cl2CrO5
595.46
C30H32Cl2CrO5
595.46
C32H30Cr2O8
646.56
C
32.5H31ClCr2O8
689.02
)
temperature (°C)
wavelength (Å)
space group
unit cell dimensions
-153(2)
0.71073
P212121
-153(2)
0.71073
P21
-173(2)
0.71073
Pna21
-153(2)
0.71073
P41
-153(2)
0.71073
P212121
-153(2)
0.71073
P1h
(Å, deg)
a
b
c
10.0416(8)
11.4518(8)
18.8534(9)
8.6518(6)
10.0312(8)
14.6724(9)
27.4425(6)
13.8076(2)
15.2781(4)
19.7635(6)
19.7635(6)
15.1720(7)
10.4305(4)
18.1260(8)
30.9046(11)
12.7753(6)
13.2878(6)
18.6941(8)
77.804(3)
89.098(3)
89.761(2)
3101.4(2)
4
R
â
97.809(4)
γ
volume (Å3)
2168.0(3)
4
1.147
0.071
808
1261.58(15)
2
1.139
0.072
468
5789.1(2)
8
1.366
0.617
2480
5926.1(4)
8
1.335
0.603
2480
5842.9(4)
8
1.470
0.795
2672
Z
Dc (g cm-3
)
1.476
0.837
1420
µ (mm-1
F(000)
)
θ range (deg)
2.9-26.0
2.6-26.0
2.7-26.0
1.7-26.0
2.1-27.5
1.6-26.0
reflections collected
independent reflections
parameters
12 448
6995
27 315
36 245
21 498
18 869
4250
257
1.044
4276
285
1.059
10 486
696
1.041
11 437
696
1.018
12 750
776
1.010
12 120
792
1.034
GOF on F 2
final R indices,
I > 2σ(I)
R indices (all data)
R1 ) 0.0680,
wR2 ) 0.1446
R1 ) 0.1144,
wR2 ) 0.1640
0(3)
R1 ) 0.0680,
wR2 ) 0.1639
R1 ) 0.0775,
wR2 ) 0.1721
-1(2)
R1 ) 0.0598,
wR2 ) 0.1031
R1 ) 0.0908,
wR2 ) 0.1131
-0.01(2)
R1 ) 0.0663,
wR2 ) 0.1296
R1 ) 0.1224,
wR2 ) 0.1495
-0.04(2)
R1 ) 0.0494,
wR2 ) 0.1190
R1 ) 0.0667,
wR2 ) 0.1285
0.480(18)
R1 ) 0.0773,
wR2 ) 0.1775
R1 ) 0.1145,
wR2 ) 0.1975
absolute structure
parameter
largest diff peak
0.457
0.525
0.899
0.538
0.999
1.188
(e Å-3
)
Figure 1. (a) The two independent chromium complexes in the asymmetric unit of (()-3‚2CH2Cl2 showing the heterochiral SS‚‚‚RR hydrogen bond,
O(2)‚‚‚O(6) 2.806(4) Å. (b) The crystallographically independent homochiral RR‚‚‚RR hydrogen bond in (()-3‚2CH2Cl2, O(7)‚‚‚O(1)′ 2.721(4) Å (primed
1
1
1
atom generated by glide operation -x + /2, y - /2, z - /2). Both unique pairings are supported by CH‚‚‚OC interactions from the coordinated aryl rings
to adjacent carbonyl ligands, with C‚‚‚O distances in the range 3.177-3.331 Å.
pairs of crystallographically independent molecules (both RR
so all interactions along the helical chain are homochiral), related
by a crystallographic 21 axis.
interactions which occur along the 21 axis in (()-3 (Figure 2).
In (()-3 the π-π interaction is also homochiral and also results
in a noncrystallographic 41 operation (noncrystallographic
because it is between two crystallographically independent
molecules). Thus, both RR‚‚‚RR hydrogen bonding and RR‚‚‚
RR π-π interactions must result in 4-fold helix formation in
the resolved material, leading to a predicted space group of P41.
While this is necessarily the only way to go from the observed
homochiral interactions in (()-3 to a resolved structure contain-
ing only homochiral interactions, it is only part of the story,
however. We have not yet accounted for the edge-to-face π-π
9
J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002 5111

A R T I C L E S
Gibson et al.
Figure 2. View down the crystallographic 21 axis (c axis, 15.2781(4) Å)
showing the 4-fold helix resulting from the action of the 21 screw on the
approximate 41 symmetry of the RR‚‚‚RR pairing. Molecules related by 21
symmetry interact via edge-to-face π-stacking interactions.
Figure 4. The noncrystallographic 4-fold helix in (-)-3. Intermolecular
hydrogen-bonded distances: O(2)‚‚‚O(6) 2.732(5), O(7)‚‚‚O(1)′ 2.726(5)
Å (primed atom generated by 21 operation 2 - x, 1 - y, 1/2 + z). Note that
both distances are similar to the O(7)‚‚‚O(1)′ distance in (()-3 but distinct
from the heterochiral O(2)‚‚‚O(6) hydrogen bond.
OH)], by protection of one of the ortho sites using SiMe3. The
deprotonation reaction of the silylated material proved to be
relatively difficult, however, and the attempt was abandoned.
The predictability of the structure of (-)-3 is predicated upon
the isolation of the OH‚‚‚OMe interaction in particular as the
dominant crystal packing driving force. To verify whether this
was indeed the case, it was deemed of interest to also examine
the X-ray structure of the free ligand (()-4. The free substituted
triphenylmethanol derivative (()-4 can be isolated in 72% yield
by decomplexation of racemic (()-3 by irradiation with a 100
W light bulb over 7 days. Crystallization of 4 from a mixed
acetone/CH2Cl2/n-pentane solution unfortunately results in the
formation of a solvate 4‚OCMe2 in which the hydroxyl
functionality of the ligand interacts with the acetone carbonyl
oxygen atom (OH‚‚‚O 2.826(4) Å). However, crystallization
from n-pentane affords the solvent free material, which does
indeed display the expected OH‚‚‚OMe hydrogen bond, Figure
5. Remarkably both (()-4 and (()-4‚OCMe2 crystallize in chiral
space groups (P212121 and P21, respectively) despite being
present in solution in racemic form. Furthermore, both crystal
structures are disordered and contain varying proportions of both
enantiomers (83% of one enantiomer for (()-4 and 63% for
(()-4‚OCMe2; using molybdenum radiation it was not possible
to determine whether the R or S form predominates). In the
case of (()-4‚OCMe2, this disorder has essentially no effect
on the crystal packing since it involves two alternate positions
for just two atoms near the chiral center (plus associated t-butyl
group disorder). In solvent free (()-4, however, the OMe
acceptor atom is disordered, and in the minor enantiomer
(disordered component) it forms the shorter hydrogen bond with
the nondisordered OH group of an adjacent molecule, O‚‚‚O
distances 2.685(5) versus 2.900(4) Å. The structures of (()-4
and its acetone solvate clearly establish that the OH group does
indeed dominate the crystal packing and also highlight the
awkward shape of the ligand, resulting in a high propensity for
acentric crystal packing modes.
Figure 3. The crystallographic 41 helix in (-)-3 (c axis, 15.1720(7) Å).
Interactions are entirely of the edge-to-face π-π type.
The same is, of course, true for the SS enantiomer. Conversely,
the unobserved RS/SR diastereoisomer will have a significantly
different packing mode.
The predictions derived from (()-3 were experimentally
verified by the preparation of optically pure (-)-3 as detailed
above. It proved extremely gratifying to find that (-)-3 does
indeed crystallize in the chiral tetragonal space group P41 with
Z′ ) 2. In the structure there are two independent 4-fold helices.
One is situated on the crystallographic 41 axis in which the
molecules interact entirely by edge-to-face π-π interactions
(Figure 3). This results in a crystallographic c dimension very
similar to that found along the 21 axis in (()-3 (Table 1). The
other 4-fold helix results entirely from OH‚‚‚OMe hydrogen
bonding interactions, identical to the homochiral interactions
found in the racemate. The two independent hydrogen-bonded
distances are very similar to the homochiral RR‚‚‚RR combina-
tion found in (()-3, but distinct from the longer RR‚‚‚SS inter-
action. This helix is situated on a crystallographic 21 axis and
involves two pairs of two unique RR‚‚‚RR molecules, Figure 4.
Strident attempts were also made to prepare racemic and
resolved samples of the opposite diastereoisomer of 3, (()-
(1pR,1′′S)(1pS,1′′R)-[Cr(CO)3(η6-t-BuC6H3(CMeOMe)CPh2-
Reaction of (()-4 with 1 equiv of Cr(CO)6 results in a
complicated mixture of products from which the bimetallic
9
5112 J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002

Interactions in (Arene)chromium Carbonyl Compounds
A R T I C L E S
Figure 6. Intermolecular hydrogen bonding to CO in (()-5a between two
RSS enantiomers. Hydrogen-bonded distance 2.946(4) Å.
Figure 5. Twofold helical packing in (()-4 (solvent free) showing the
hydrogen-bonded chain to the major component. Hydrogen bonds to the
minor component are of the same form but shorter; OH‚‚‚O distances
2.685(5) and 2.900(4) Å.
complexes 5 may be isolated by column chromatography.
Complexes 5 possess three chiral elements (two centers and
plane chirality associated with the side of the trisubstituted aryl
ring complexed to chromium). The compound can, theoretically,
exist as a total of eight isomers comprising four diastereomeric
pairs denoted (()-(1pR,1′R,1′′R)-5, (()-(1pR,1′R,1′′S)-5, (()-
(1pR,1′S,1′′S)-5, and (()-(1pR,1′S,1′′R)-5 (and their enanti-
omers). In practice, only (()-(1pR,1′R,1′′S)(1pS,1′S,1′′R)-5, (()-
5a, and (()-(1pR,1′R,1′′R)(1pS,1′S,1′′S)-5, (()-5b, are observed
and isolated (as racemates) from the reaction mixture.
The X-ray crystal structures of diastereoisomers 5a and 5b
were determined, and both showed one Cr(CO)3 unit to reside
on the tert-butyl substituted ring (as in complexes 3), while the
second is coordinated to one of the phenyl substituents,
rendering the -C*Ph2OH central carbon atoms chiral. Remark-
ably, as for 3, both structures displayed Z′ ) 2 with the two
independent molecules being enantiomers of one another. (()-
5b also cocrystallizes with one disordered molecule of CH2-
Cl2. The view down the HO-C* axis shows both Cr(CO)3
groups to be pointing “up” toward the hydroxyl group in all
four unique molecules in the two structures, consistent with
steric arguments. While both structures contain two unique
enantiomers, their crystal packing modes are distinctly different
from one another. (()-5a crystallizes in the chiral space group
P212121 and spontaneously resolves. Thus, while the crystal
contains two enantiomers, it is chiral by virtue of the 2-fold
helical crystal packing arrangement. In contrast, (()-5b crystal-
lizes in the centrosymmetric space group P1h, and thus each
crystallographically independent enantiomer is related by inver-
sion symmetry to a molecule of the opposite handedness.
However, the two independent enantiomers are not symmetry-
related to one another.
Figure 7. Intramolecular hydrogen bonding in (()-5a by the RRS
enantiomer. Hydrogen-bonded distance 2.659(4) Å.
consistent with the weak hydrogen bond acceptor nature of the
CO ligand.34 This is particularly surprising given the propensity
in 3 for hydrogen bonding to the OMe group which should be
a stronger acceptor and suggests that the bimetallic species is
particularly sterically demanding. The independent RRS enan-
tiomer, on the other hand, contains an intramolecular OH‚‚‚
OMe interaction with the much shorter O‚‚‚O distance of
2.659(4) Å, Figure 7. This is also a surprising interaction given
the purely intermolecular nature of the hydrogen bonds observed
for compounds 3 and 4. Furthermore, it seems only to be
associated with the occurrence of opposite R/S handedness at
the first chiral carbon atoms and the plane chiral element. This
places the OMe group and OH group in close proximity. In
contrast, in RR and SS-3 they are orientated away from one
another. Similarly, the other diastereoisomer (()-5b shows only
intermolecular OH‚‚‚OC hydrogen bonds. From this observation,
we can therefore predict that the unobserved RS-3 should show
similar intramolecular hydrogen bonding.
Unlike its diastereoisomer, the two independent enantiomeric
molecules in (()-5b are directly hydrogen bonded together via
entirely heterochiral OH‚‚‚OC interactions to form an infinite
chain, Figure 8, with repeat RRR‚‚‚SSS‚‚‚RRR‚‚‚‚SSS‚‚‚. Interac-
tions between the two independent molecules in (()-5a are of
the edge-to-face π-π type observed in 3.
Conclusion
All compounds of type 3-5 are highly substituted monoal-
cohols, a class of compound known to commonly form
structures with Z′ > 1 (40% of monoalcohols have Z′ > 1, as
compared with 8.3% of the Cambridge Structural Database35,36
(CSD) as a whole27,37,38). It is likely that this behavior is a result
The chiral (()-5a shows two distinctly different types of
OH‚‚‚O hydrogen bond in the two independent molecules. The
SSR enantiomer (containing atoms Cr(1) and Cr(2)) exhibits a
homochiral intermolecular OH‚‚‚OC bond to one of the carbonyl
ligands (not OMe group, cf. 3) on an adjacent SSR enantiomer
(Figure 6) to give an infinite crystallographic 21 helix along
the a direction. The long O‚‚‚O distance of 2.946(4) Å is
(34) Braga, D.; Grepioni, F. Acc. Chem. Res. 1997, 30, 81-87.
(35) Fletcher, D. A.; McMeeking, R. F.; Parkin, D. J. Chem. Inf. Comput. Sci.
1996, 36, 746.
(36) Allen, F. H.; Kennard, O. Chemical Design Automation News 1993, 8, 31.
(37) Wilson, A. J. C. Acta Crystallogr., Sect. A 1993, 49, 795-806.
9
J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002 5113

A R T I C L E S
Gibson et al.
Figure 8. Infinite OH‚‚‚OC hydrogen-bonded chain in (()-5b comprising two independent heterochiral interactions. Hydrogen-bonded distances: O(10)‚‚‚
O(3)′ 2.840(7), O(2)‚‚‚O(11) 2.762(7) Å (primed atom generated by translation x - 1, y, z).
Experimental Section
of the fact that the hydrogen bonding requirements of the OH
group dominate the crystal packing, often resulting in frustra-
tion28-30 between the conflicting needs to achieve close packing
and maximize hydrogen bonding to the lone OH group.
Moreover, compounds 3-5 are all derivatives of triphenyl-
methanol. Triphenylmethanol along with its silicon and germa-
nium analogues exhibits eight crystallographically independent
molecules (Z′ ) 8). This is a consequence of the formation of
two independent sets of hydrogen-bonded tetramers. Even in
the absence of the OH group we have shown that compounds
of type YXPh3 (Y ) halogen, X ) group 14 element) exhibit
an extremely high incidence of Z′ > 1 as a consequence of weak
aryl CH‚‚‚Y interactions.14
In the particular case of (-)-3 we have shown that it is
possible to predict the precise nature of the crystal packing
geometry by consideration of the structure of the analogous Z′
) 2 racemate, which fortuitously crystallizes in an acentric space
group and demonstrates homochiral interactions. We have also
demonstrated that relative stereochemistry may be used to switch
on and off particular intermolecular and intramolecular interac-
tions. Furthermore, the sterically demanding nature of the system
limits the number of possible intermolecular interactions,
resulting in a high degree of reproducibility across many
structures.
Instrumental. All reactions and manipulations involving organo-
metallic compounds were performed under an inert atmosphere of dry
nitrogen, using standard vacuum line and Schlenk tube techniques.
Reactions and operations involving (arene)tricarbonylchromium(0)
complexes were protected from light. THF was distilled from sodium
benzophenone ketyl. The concentration of methyllithum was determined
by titration against diphenylacetic acid in THF. Tetramethylpiperidine
was stored over potassium hydroxide. Flash column chromatography
was performed using Merck silica gel 60 (230-400 mesh). All other
reagents were used as obtained from commercial sources. Melting points
were recorded in open capillaries on a Bu¨chi 510 melting point
apparatus and are uncorrected. IR spectra were recorded on Perkin-
Elmer 1600 FT-IR. NMR spectra were recorded at room temperature
on a Bru¨ker AM 360 and DRX 400, and J values are reported in hertz.
Mass spectra were recorded on a JEOL AX 505W and Kratos
MS890MS. Elemental analyses were performed by the University of
North London microanalytical service.
Preparations. (()-(1pR,1′′R)(1pS,1′′S)-η6-[2-(1-Methoxyethyl)-5-
tert-butyl-(diphenylhydroxymethyl) benzene]tricarbonylchromium-
(0), (()-3. Methyllithium (1.31 mL of a 1.76 M solution in diethyl
ether, 2.3 mmol) was added dropwise at -78 °C to a stirred solution
of tetramethylpiperidine (0.39 mL, 2.3 mmol) in THF (20 mL). The
solution was allowed to reach room temperature and recooled to -78
°C. Complex (()-2 (657 mg, 2 mmol) in THF (8 mL) was added via
a cannula, and the resulting orange solution was stirred for 1 h. A
solution of benzophenone (1.26 g, 6.9 mmol) in THF (5 mL) was added
via a cannula, and stirring was continued for a further 2 h at -78 °C.
The reaction vessel was removed from the acetone-dry ice bath, and
the reaction mixture was stirred at room temperature for 1 h. Methanol
was added, and the solvent was removed in vacuo. Flash column
chromatography (SiO2; hexane/diethyl ether, 10:0-5:5) of the residue
gave the complex (()-3 as a yellow solid (950 mg, 93%). mp: 158-
159 °C. IR (Nujol, cm-1): νOH 3386 (w), νCO 1953 (s), 1946 (m), 1879
By extending this approach, it should be possible to make
concrete predictions about the crystal packing mode of all
resolved compounds for which the racemate crystallizes in such
a way as to demonstrate the nature of homochiral interactions,
assuming those interactions to be isolated and dominant.
Furthermore, in selected cases, we can also use chirality to
predict and engineer the incidence of structures with Z′ > 1 (in
the absence of disorder). Thus, a racemate displaying a
centrosymmetric intermolecular interaction (e.g., formation of
a carboxylic acid dimer) must double Z′ upon resolution; that
is, a resolved chiral mono carboxylic acid dimer must exhibit
Z′ ) 2 if the racemate exhibits Z′ ) 1 with the two enantiomers
related by inversion symmetry.
1
(m), 1868 (s), 1860 (s). H NMR (360 MHz, CDCl3): δ 1.09 (s, 9H,
(CH3)3C), 1.46 (d, J ) 6.3 Hz, 3H, CH3CH), 2.47 (s, 3H, OCH3), 3.64
(s, 1H, OH), 4.52 (q, J ) 6.3 Hz, 1H, CH3CH), 4.72 (d, J ) 1.6 Hz,
1H, CCrHCCrCOH), 5.55 (d, J ) 6.8 Hz, 1H, CCr(t-Bu)CCrHCCrH), 5.60
(dd, J ) 6.8, 1.6 Hz, 1H, CCr(t-Bu)CCrHCCrH), 7.32-7.41 (m, 10H,
Har). 13C NMR (90 MHz, CDCl3): δ 22.0 (CH3CH), 30.8 ((CH3)3C),
33.7 ((CH3)3C), 56.0 (OCH3), 74.1 (CH3CH), 81.5 (COH), 91.5 (CCrH),
92.1 (CCrH), 94.9 (CCrH), 117.6 (CCr), 117.7 (CCr), 121.1 (CCr), 127.5
(2C, CarH), 127.6 (2C, CarH), 127.8 (CarH), 128.1 (2C, CarH), 128.2
(38) Brock, C. P.; Duncan, L. L. Chem. Mater. 1994, 6, 1307-1312.
9
5114 J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002

Interactions in (Arene)chromium Carbonyl Compounds
A R T I C L E S
(2C, CarH), 128.5 (CarH), 144.2 (Car), 144.4 (Car), 233.6 (CtO). MS
(EI): m/z (%) 510 (M+, 18), 426 (M+ - 3CO, 100), 394 (M+ - 3CO
- CH3OH, 62), 350 (27), 311 (37), 269 (50), 255 (53). Anal. Calcd
for C29H30CrO5 (510.55): C, 68.22; H, 5.92. Found: C, 68.21; H, 5.95.
(-)-(1pR,1′′R)-η6-[2-(1-Methoxyethyl)-5-tert-butyl-(diphenylhy-
droxymethyl)benzene]tricarbonylchromium(0), (-)-3. The procedure
for the synthesis of the racemic complex (()-3 was followed.
Methyllithium (0.37 mL of a 1.49 M solution in diethyl ether, 0.55
mmol) was added dropwise at -78 °C to a stirred solution of
tetramethylpiperidine (93 µL, 0.55 mmol) in THF (5 mL). Complex
(-)-2 (150 mg, 0.46 mmol) in THF (2 mL) was added followed by
the addition of benzophenone (255 mg, 1.4 mmol) in THF (5 mL).
Flash column chromatography gave (-)-3 as a yellow solid (171 mg,
73%). mp: 173-178 °C. Enantiomeric excess was determined by
CH), 31.0 ((CH3)3C), 33.7 ((CH3)3C), 57.1 (OCH3), 73.3 (CH3CH), 79.0
(COH), 87.1 (CCrH), 87.9 (CCrH), 88.2 (CCrH), 93.1 (CCrH), 94.7 (CCrH),
95.2 (CCrH), 95.9 (CCrH), 96.0 (CCrH), 114.5 (CCr), 115.9 (CCr), 117.1
(CCr), 118.5 (CCr), 127.2 (2C, CarH), 128.2 (CarH), 128.5 (2C, CarH),
142.1 (Car), 231.8 (C tO), 232.8 (CtO). MS (FAB): m/z (%) 669
(M+ + Na, 7), 646 (M+, 4), 562 (M+ - 3CO, 77), 478 (M+ - 6CO,
33), 426 (M+ - 3CO - Cr(CO)3, 100). Anal. Calcd for C32H30Cr2O8
(646.57): C, 59.44; H, 4.68. Found: C, 59.59; H, 4.57.
This was followed by complex (()-5b as a yellow solid (13 mg,
5%). mp: 160 °C (decomp). IR (Nujol, cm-1): νOH 3580 (w), νCO 1968
(s), 1960 (s), 1918 (m), 1879 (s), 1860 (s). 1H NMR (400 MHz,
CDCl3): δ 1.18 (s, 9H, (CH3)3C), 1.37 (d, J ) 6.3 Hz, 3H, CH3CH),
2.45 (s, 3H, OCH3), 3.71 (s, 1H, OH), 4.38 (q, J ) 6.3 Hz, 1H, CH3CH),
4.82 (d, J ) 6.5 Hz, 1H, o-CCrH), 5.00 (dt, J ) 6.5, 1.1 Hz, 1H,
m-CCrH), 5.06 (d, J ) 1.6 Hz, 1H, CCr(t-Bu)CCrHCCrCOH), 5.34 (dt, J
) 6.5, 1.1 Hz, 1H, m′-CCrH), 5.46 (d, J ) 6.8 Hz, 1H, CCr(t-Bu)CCr-
HCCrH), 5.60 (t, J ) 6.5 Hz, 1H, p-CCrH), 5.64 (dd, J ) 6.8, 1.6 Hz,
1H, CCr(t-Bu)CCrHCCrH), 6.31 (d, J ) 6.5 Hz, 1H, o′-CCrH), 7.40 (br
s, 5H, Har). 13C NMR (100 MHz, CDCl3): δ 22.1 (CH3), 30.9 ((CH3)3C),
33.8 ((CH3)3C), 55.8 (OCH3), 74.5 (CH3CH), 79.3 (COH), 87.3 (CCrH),
88.5 (CCrH), 90.8 (CCrH), 92.3 (CCrH), 93.9 (CCrH), 94.2 (CCrH), 96.0
(CCrH), 96.1 (CCrH), 115.3 (CCr), 115.7 (CCr), 117.5 (CCr), 120.1 (CCr),
127.3 (2C, CarH), 128.2 (CarH), 128.6 (2C, CarH), 141.5 (Car), 231.8
(CtO), 233.0 (CtO). MS (FAB): m/z (%) 669 (M+ + Na, 87), 562
(M+ - 3CO, 100), 533 (M+ - 3CO - CHO, 53), 478 (M+ - 6CO,
87), 446 (M+ - 6CO - CH3OH, 53), 426 (M+ - 3CO - Cr(CO)3,
80). Anal. Calcd for C32H30Cr2O8 (646.57): C, 59.44; H, 4.68. Found:
C, 59.52; H, 4.78.
i
HPLC analysis (Chiralcel ODH, 0.2% PrOH/n-hexane, 0.4 mL/min,
330 nm); (R) enantiomer tr ) 35.6 min (major); (S) enantiomer tr )
40.9 (minor): >93% ee. [R]21 ) -42.4° (c 0.75, CH2Cl2). All other
D
data were identical to those obtained for (()-3.
(()-2-(1-Methoxyethyl)-5-tert-butyl-(diphenylhydroxymethyl)-
benzene, (()-4. A solution of complex (()-3 (460 mg, 0.9 mmol) in
diethyl ether (180 mL) was irradiated with a 100 W light bulb with
slow stirring for 7 d. The resulting suspension was filtered through a
pad of Celite, and the filtrate was evaporated under reduced pressure.
Flash column chromatography of the residue (SiO2; hexane/diethyl
ether/NEt3, 90:9:1) gave compound (()-4 as a white solid (243 mg,
72%). mp: 124-125 °C. IR (Nujol, cm-1): νOH 3401 (s). IR (KBr,
cm-1): νOH 3444 (s), νCH 2965 (s), νCH 1446 (m), νCOC 1096 (m), νCOC
1068 (s), νar-CH 841 (m), νar-CH 763 (w), νar-CH 751 (m), νar-CH 700
Crystallography. Crystal data and data collection parameters are
summarized in Table 1. Crystals were mounted on a thin glass fiber
using silicon grease and cooled on the diffractometer to 100 K using
an Oxford Cryostream low-temperature attachment. Oscillation frames
each of width 1-2° in either φ or ω and of 10-60 s deg-1 exposure
time were recorded using a Nonius KappaCCD diffractometer, with a
detector to crystal distance of 30 mm. Crystals were indexed from five
preliminary frames each of 2° width in φ using the Nonius Collect
package.39 Final unit cell dimensions and positional data were refined
on the entire data set along with diffractometer constants to give the
final unit cell parameters. Integration and scaling (DENZO-SMN,
Scalepack40) resulted in data set corrected for Lorentz and polarization
effects and for the effects of crystal decay and absorption by a
combination of averaging of equivalent reflections and an overall
volume and scaling correction. Structures were solved by direct methods
(SHELXS-9741) and developed via alternating least squares cycles and
difference Fourier synthesis (SHELXL-9742) with the aid of the XSeed
interface.43 All non-hydrogen atoms were modeled anisotropically.
Hydrogen atoms were placed in calculated positions and allowed to
ride on the atoms to which they were attached with an isotropic thermal
parameter 1.2 times that of the parent atom (1.5 times for CH3 groups).
Hydroxyl protons were located experimentally and treated likewise if
refinement proved unfeasible. Hydrogen atom thermal parameters were
fixed at 1.2 times those of the parent atom. All calculations were carried
out either on a Silicon Graphics Indy workstation or an IBM-PC
compatible personal computer. Atomic coordinates, bond lengths and
angles, and thermal parameters have been deposited at the Cambridge
Crystallographic Data Centre. See Information for Authors, Issue No.
1.
1
(s). H NMR (400 MHz, CDCl3): δ 1.11 (s, 9H, (CH3)3C), 1.33 (d, J
) 6.3 Hz, 3H, CH3CH), 2.79 (s, 3H, OCH3), 4.19 (s, 1H, OH), 4.64
(q, J ) 6.3 Hz, 1H, CH3CH), 6.66 (d, J ) 2.1 Hz, 1H, CarHCarCOH),
7.23-7.24 (m, 11H, Har), 7.46 (d, J ) 8.1 Hz, 1H, Har). 13C NMR
(100 MHz, CDCl3): δ 21.3 (CH3CH), 31.0 ((CH3)3C), 34.4 ((CH3)3C),
55.1 (OCH3), 75.2 (CH3CH), 83.1 (COH), 124.5 (CarH), 127.0 (CarH),
127.3 (CarH), 127.5 (2C, CarH), 127.7 (2C, CarH), 127.8 (2C, CarH),
127.88 (2C, CarH), 127.92 (2C, CarH), 139.5 (Car), 144.5 (Car), 146.7
(Car), 147.9 (Car), 149.0 (Car). MS (EI): m/z (%) 356 (M+ - H2O,
100), 341 (M+ - H2O - CH3, 32), 324 (M+ - H2O - CH3OH, 52),
281 (M+ - H2O - CH3OH - C2H3O, 20), 265 (M+ - H2O - C7H7,
72). Anal. Calcd for C26H30O2 (374.52): C, 83.38; H, 8.07. Found: C,
83.47; H, 8.04.
(()-(1pS,1′S,1′′R)(1pR,1′R,1′′S)-η6-[2-(1-Methoxyethyl)-5-tert-bu-
tyl-(1-η6-phenyl-tricarbonylchromium(0)-1-phenyl-1-hydroxymeth-
yl)benzene]tricarbonylchromium(0), (()-5a, and (()-(1pR,1′R,1′′R)-
(1pS,1′S,1′′S)-η6-[2-(1-Methoxyethyl)-5-tert-butyl-(1-η6-phenyl-tri-
carbonylchromium(0)-1-phenyl-1-hydroxymethyl)benzene]tricar-
bonyl- chromium(0), (()-5b. A 25 mL round-bottomed flask fitted
with a Liebig air condenser with a water condenser on top was charged
with hexacarbonylchromium(0) (95 mg, 0.43 mmol), ligand (()-4 (162
mg, 0.43 mmol), dry THF (0.8 mL), and dry di-n-butyl ether (8 mL).
The suspension was thoroughly saturated with nitrogen, before being
heated to 130 °C, and refluxed under a slight nitrogen overpressure
(40 mbar). After 48 h, the orange reaction mixture was allowed to cool
to room temperature, and the solvent was removed in vacuo. Flash
column chromatography (SiO2; hexane/diethyl ether, 9:1) of the residue
offered first the dichromium complex (()-5a as a yellow solid (18
mg, 6%). mp: 161-164 °C (decomp). IR (Nujol, cm-1): νOH 3544
1
(w), νCO 1972 (s), 1955 (sh), 1904 (m), 1887 (s), 1867 (sh). H NMR
(39) Hooft, R. Collect; Nonius: Delft, 1998.
(40) Otwinowski, Z.; Minor, W. In Methods in Enzymology; Carter, C. W.,
Sweet, R. M., Eds.; Academic Press: London, 1997; Vol. 276, pp 307-
326.
(400 MHz, CDCl3): δ 0.68 (d, J ) 6.2 Hz, 3H, CH3CH), 1.18 (s, 9H,
(CH3)3C), 3.34 (s, 3H, OCH3), 4.09 (s, 1H, OH), 4.58 (q, J ) 6.2 Hz,
1H, CH3CH), 4.89 (d, J ) 6.4 Hz, 1H, o-CCrH), 4.98 (dt, J ) 6.4, 1.0
Hz, 1H, m-CCrH), 5.17 (d, J ) 1.6 Hz, 1H, CCr(t-Bu)CCrHCCrCOH),
5.30 (d, J ) 6.8 Hz, 1H, CCr(t-Bu)CCrHCCrH), 5.33 (dt, J ) 6.4, 1.0
Hz, 1H, m′-CCrH), 5.59 (t, J ) 6.4 Hz, 1H, p-CCrH), 5.69 (dd, J ) 6.8,
1.6 Hz, 1H, CCr(t-Bu)CCrHCCrH), 6.37 (d, J ) 6.4 Hz, 1H, o′-CCrH),
7.28-7.41 (m, 5H, Har). 13C NMR (100 MHz, CDCl3): δ 20.1 (CH3-
(41) Sheldrick, G. M. SHELXS-97; University of Go¨ttingen, 1997.
(42) Sheldrick, G. M. SHELXL-97; University of Go¨ttingen, 1997.
(43) Barbour, L. J. XSeed: University of Missouri, Columbia, 1999.
(44) Janiak, C. J. Chem. Soc., Dalton Trans. 2000, 3885-3896.
(45) Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112, 5525-
5534.
(46) Hunter, R.; Haueisen, R. H.; Irving, A. Angew. Chem., Int. Ed. Engl. 1994,
33, 566-568.
9
J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002 5115

A R T I C L E S
Gibson et al.
Acknowledgment. We thank the EPSRC and King’s College
London for funding of the diffractometer system. Grateful
acknowledgment is also given to the Nuffield Foundation for
the provision of computing equipment.
Supporting Information Available: X-ray crystallographic
file (CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
JA012141M
9
5116 J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002