вход по аккаунту


Growth and Dissolution of Organic Crystals with УTailor-MadeФ InhibitorsЧImplications in Stereochemistry and Materials Science.

код для вставкиСкачать
Growth and Dissolution of Organic Crystals with “Tailor-Made”
Inhibitors- Implications in Stereochemistry and Materials Science
By Lia Addadi,* Ziva Berkovitch-Yellin,* Isabelle Weissbuch, Jan van Mil,
Linda J. W. Shirnon, Meir Lahav,* and Leslie Leiserowitz*
In memoriam Gerhard M . J. Schmidt
When crystals of organic compounds are grown in the presence of growth inhibitors, there
is a change in crystal morphology. A stereochemical correlation exists between the crystal
structure, its modified morphology, and the molecular structure of the inhibitor. This correlation has been successfully exploited for the efficient resolution of conglomerates, the engineering of organic crystals with desired morphologies, the direct and relative assignment
of the absolute configurations of chiral molecules and crystals, and for the design of a new
model for the spontaneous generation of optical activity. In an analogous way dissolution
of organic crystals in the presence of these growth inhibitors induces etch pits at preselected faces. The effect of solvent on crystal growth has been analyzed in some model systems. The experimental results are complemented by atom-atom potential energy calculations.
1. Introduction
The various and beautiful morphologies of crystals have
attracted the attention of natural philosophers since ancient times. Crystal formation from the smallest irreducible
units had already been conceived as long ago as 1611 by
Kepler;”]this concept played an important role in the development of molecular theory.
The relation between the morphology o f a crystal and its
internal symmetry at a molecular level was first demonstrated by Louis Pasteurizlin 1848. In his now famous experiment he separated, for the first time, the two enantiomers of sodium ammonium tartrate, utilizing the asymmetric
habit of their crystals. Over the years, however, it became
evident that the morphological characteristics of the crystals of a given compound depend not only on its crystal
structure, but also on external parameters of the overall
crystallizing system such as solvent, supersaturation, temperature, and specifically impurities present in the system”’.
In 1931 Milesi4’ demonstrated that crystals of PbCI2,
which normally crystallize with centrosymmetric point
symmetry (2/m,2/m,2/m), assume a chiral morphology of
symmetry 222 when grown in the presence of dextrin. With
the advent of new and efficient methods of crystal structure determination, most efforts were devoted to such determinations and less interest was focused on correlating
crystal structure and morphology. As a result, many important properties of materials which depend on this correlation were not exploited in a systematic way. On the other
hand, the use of dopants for inducing a particular crystal
morphology is a widespread technological tool, even
though the mechanism of the effect is scarcely understood.
Thus, for example, small amounts of ions such as Pb2+ or
[*] Dr. L. Addadi, Dr. Z. Berkovitch-Yellin, I. Weissbuch, Dr. J. van Mil,
L. J. W. Shimon, Prof. Dr. M. Lahav, Prof. Dr. L. Leiserowitz
Department of Structural Chemistry,
The Weizmann Institute of Science
Rehovot 76 100 (Israel)
0 VCH Veriagsgeself.rcha/f mbH. D-6940 Weinheim, 198s
SO$- have long been routinely used as crystallization
moderators in the industrial production of large NaCl single crystals for IR and Raman ~ p e c t r o s c o p y .Similarly,
sugar technologists recognized and studied the critical role
played by oligosaccharides in determining the rate of crystallization and habit of sucrose crystals on extraction from
In the course of investigating the packing and growth of
organic molecular crystals, we initiated a study on the systematics of the interactions between a crystallizing substrate and its environment, in general, and the solvent and
stereospecific “tailor-made” growth inhibitors in particular. We shall illustrate here some mechanistic aspects o f
the above interactions as they have emerged during these
studies and some applications thereof to the amplification
of chirality by crystallization, resolution of conglomerates,
the direct and relative assignment of absolute configuration of chiral molecules, the study of molecular interactions in crystals, the engineering of crystals with desired
morphologies, and the selective etching of crystal surfaces
upon dissolution.
2. Amplification of Chirality by Crystallization
with Chiral Additives; a New Method for the
Resolution of Conglomerates of Enantiorners
The synthesis of chiral enantiomerically pure materials
from nonchiral reagents has been accomplished by crystallization of the (in solution) nonchiral substrate in appropriately packed chiral single crystals, followed by a lattice
controlled reaction.”] This concept is illustrated in Scheme
1 for the generation of chiral cyclobutane polymers from
nonchiral dienes packing in “engineered” chiral crystals.@]
The molecules in each chiral crystal, (denoted by [ Id or
[ 1, in Scheme l), are related by translation such that the
neighboring C=C bonds are in appropriate proximity for
photodimerization. UV irradiation of a single crystal yields
dimers, trimers, and oligomers of a single chirality P, or P,.
0570-0833/85/0606-0466 $ 02.50/0
Angew. Chem. Inl. Ed. Engl. 24 (1985) 466-485
Scheme I . Lattice-controlled generation of chiral cyclobutane derivatives (enantiomers P, and PJ from non-chiral dienes
packing in “engineered” chiral crystals (d a n d I symbolize the respective enantiomeric packings). For examples of the
dienes, see Table I . The feedback leads to an inhibition of crystallization.
In a nonchiral solvent the d and 1 crystals are formed in
equal amounts, so that no net asymmetric induction of the
reaction products P, and P, is observed. O n the other hand,
a net asymmetric induction of one of the crystalline phases
is achieved when chiral dimers, trimers o r oligomers of a
single chirality (i.e. P, or P,) are present during the process
of crystallization. In all experiments it was observed that a
large excess of that phase was formed which was enantiomorphic to the one in which the additive itself had been
Induction with
Induction with
t - i g . I . tnantiunieric excc>s ( r e ) 01’ the dirnrrs obtained hy l a l l i c ~controlled
reaction after crystallization of the monomer in the presence of chiral dimers
as additives under various experimental conditions (see Scheme 1 and Table
I ).
Angew. Chem. Inr. Ed. Engl. 24 (1985)466-485
generated; namely, d crystals were obtained from a solution containing additives P, and 1 crystals from solutions
containing additives P, (Fig. 1).
The stereochemical similarity between the additive and
the respective enantiomorphous crystal of the monomer
was found to be of paramount importance,[’, l o ] while
parameters like temperature, concentration, or nature of
the medium had only a quantitative effect on the observed
asymmetric induction.
We suggested that, by virtue of this stereochemical similarity, a chiral dimer, trimer, or oligomer may substitute
two, three, or n monomer units at a growing site of that
crystal with the same absolute configuration as the parent
crystal of the photoproducts. The adsorption should cause
a drastic decrease in the rate of growth of this same crystal,
shifting the crystallization equilibrium towards the unaffected phase. This process of crystal growth and inhibition
is illustrated in Scheme 2, where the achiral monomer is
Scheme 2. (a) Spontaneous crystallization of a racematr ( R S ) to gibe lR],,/
conglomerate; (b) crystallization in presence of the additive “d” (latticecontrolled product of the (Rid phase); preferred crystallization of (S},;(c)
crystallization in presence of a chiral additive S’ (with stereochemical similarity to S); delayed crystallization of IS/,.
represented as a fast racemizing mixture of chiral conformations R and S in solution, and (Rid and (S],represent the
chiralities of the crystalline phases.
A natural extension of this scheme to any racemic mixture, fast racemizing or not, crystallizing in the form of a
conglomerate of (R)dand { S ) ,crystals, led to the suggestion
of a new method for kinetic resolution of such systems. It
is based on selective retardation of the growth of the crystals of one enantiomorph, say S , induced by small amounts
of chiral additives, say S’, with related stereochemistry
(Scheme 2 ~ ) . [ ’ ~ ~
In the light of this proposed mechanism we could rationalize various experimental results, describing kinetic
resolutions accidentally observed in the presence of chiral
additives, and where the fast crystallizing enantiomer had
absolute configuration opposite to that of the additive (Table
The predictive power of the hypothesis was further tested on other conglomerate systems including
threonine, glutamic acid. HCI, and asparagine.H,O. These
were resolved with the assistance of a variety of other amino acids used as growth retarders (Table l).~”I
The efficiency of the resolution was found to vary from
one system to another, and to depend upon the nature and
Table 1. Resolution of conglomerates of enantiomeric compounds or crystals. Top: previously described resolutions with chiral additives which are in
accord with our “rule of reversal”. Bottom: resolutions of conglomerates accomplished in our laboratory and designed on the basis of the same rule.
Sodium ammonium
Narwedine 1
p,p’-Dimethylchalcone 3
diacrylates 5 [c]
Chiral Additive [a]
(S)-Asp, (S)-Leu
(S)-Glu-y-Me ester
(S)-Glu, (S)-Ala
D-( +)-Malic acid
Enantiomer first
precipitating in excess
o-(-)-Sodium ammonium tartrate
I 1 11
(2R.3S)-2,3-Dibromo- 1,3-bis@-tolyl)I-propanone 4
from d-crystals [b]
cone, I-crystals [b]
Dimeric 3,3’-(p-phenylene) diacrylates
6 from d-crystals
diacrylates, I-crystals [b]
Glu. HCI
Asn H 2 0
His. HCI
acid 7
(S)-Glu, (S)-Gln,
(S)-Asn, (R)-Cys,
(S)-Phe, (S)-His,
(S)-Lys, (S)-Om,
@)-His, (S)-Ser, (S)Thr, (S)-Cys, (S)Tyr, (S)-Leu
(S)-Glu, (S)-Asp,
(S)-Ser, (S)-Gln,
(S)-Lys, (S)-Om,
(S)-Tyr, (S)-p-Methoxyphenylglycine,
(S)-Phe, (S)-Dopa,
(S)-Trp, (S)-Phe
acid 8
(R)-p- Hydroxyphenylglycine
(R)-3-Phenylhydracrylic acid 9
[a] All amino acids that have been used as chiral additives belong to the L-series, i.e. with the exception of Cys they are (S)-configurated.
d und 1 arbitrarily denote the different chirality of the crystals (without reference to the
absolute configuration). [c] Examples: R ’ =COOCHEt2, R 2 = COOMe,
COOEt, COOnPr; R’=(RS)-COOsBu, R2=COOEt, COOnPr.
concentration of the additive. In the best cases, [(RS)-glutamic acid. HCI (S)-lysine; (RS)-threonine (S)-glutamic acid ; (RS)-asparagine (S)-aspartic acid] complete resolution (100% ee) was achieved. The growth of the affected
crystals in these instances was delayed up to several days
with respect to that of the unaffected ones. In all the systems investigated we could demonstrate that the additive is
occluded throughout the bulk of the affected crystals in
amounts ranging typically from 0.05 to 1.5% w/w of substrate; it is found in much smaller amounts, if any, in the
crystals of the enantiomorph.
The inhibiting effect of the additive on the growth of the
affected enantiomorph has also been proven by direct
comparison of the size of (R)dand ( S ) ,single crystals grown
in parallel from seeds under conditions close to equilibrium (Fig. 2 ) . Although there are indications that the same
kind of inhibition may also influence the crystal nucleation
rate, such an effect is more difficult to isolate and quantify.
The most convincing demonstration of the adsorptiongrowth inhibition mechanism is provided by the changes in
the morphology of the affected crystals. The morphology
of a crystal is determined by the relative rates of growth of
its various faces. Since the growth rates of only those crystal faces where adsorption of inhibitor takes place are affected, crystals growing in the presence and in the absence
of additives are expected to display different habits. Indeed, in the resolution experiments it was consistently observed that the appearance of crystals of normal habit was
Angew. Chem. Int. Ed. Engl. 24 0 9 8 5 ) 466-485
glutamic acid, and (RS)-glutamic acid. HCI, grown in the
presence of (S)-lysine. HCI, the inhibition of growth is so
strong that the affected (S)-enantiomer appeared as a fine
powder coating the well-formed crystals of the (R)-enantiomorph (Fig. 4).[”]
tig 2. llirce ( H I -a n d three (S)-crystals of asparagine.H,U grown together
for 45 d‘iys under conditions near to equilibrium in the presence of (S)-serine. The large crystals are ( R ) -asparagine. H 2 0 and the small ones are (S)-asparagine. H:O.
followed (after variable time depending on the system) by
that of crystals of distinctly different morphology: when
the two types were separated by visual sorting, they were
found to consist of the unaffected and the affected enantiomers respectively. As expected, the crystallization of asparagine in the presence of different additives consistently
yielded crystals with different morphologies (Fig. 3). In the
case of (RS)-threonine, grown in the presence of say (S)-
Fig. 4. C r!stal> of (Sj-glulamic dcid. HCI grown in the prehence 0 1 increa.ring
amounts of additive lysine: (a) no additive or (R)-lysine; (b) + 2 mg/’mL (S)lysine; (c) + 5 mg/mL (S)-lysine; (d) +50 mg/mL (S)-lysine: ( e ) crystals of
( R ) - and (S)-glutamic acid.HCI grown in the presence of (S)-lysine: the
plates are the (R)-enantiomer while the powder is the (S)-enantiomer
t i g . i C r?\[dl\ 0 1 (S)-aaparaginc. H2U grown in the presence o f additivea:
(a) none, or (R)-additives; (b) (S)-glutamine as additive; (c) (S)-serine as additive; (d) (S)-ornithine as additive. In each case the crystals were obtained
from (RS)-asparagine as a mixture of (R)-asparagine crystals displaying a
morphology as in (a), and (S)-asparagine crystals displaying a morphology as
in (b)-(d) or others, depending on the additive.
Angew. C h e m . I n t . Ed. Engl. 24 11985) 466-485
This phenomenon offers the possibility of modifying
Pusteur’s classical experiment and of extending it to systems which undergo spontaneous resolution but d o not
display hemihedrism so that one cannot distinguish the enantiomorphous crystals by their morphologies.“x1 Another
kind of visual differentiation of the enantiomorphs was
obtained by crystallizing a conglomerate of colorless crystals in the presence of a resolved and colored “tailormade” inhibitor. When the plate-like crystals of (R)- and
(S)-glutamic acid.HC1 are grown in the presence of yellow
Nr-(2,4-dinitrophenyI)-(S)-lysine as additive, colorless
crystals of (R)-glutamic acid. HCI precipitate first, and
then yellow ones of the (S)-enantiomer (Fig. 5a-c).
The stereochemical correlations between the enantiomorphous crystals and their chiral inhibitors, namely that
the additive only affects the enantiomer of the same absolute configuration, provides us with a new method for the
determination of absolute configuration on a relative scale.
of molecular structure similar to those of the corresponding substrate molecule^.^"^ A stereochemical correlation
between the structures of the affected crystal surfaces and
the molecular structure of the inhibitor could be deduced
in each system. We could infer that the additive may be adsorbed only at those faces in which the part of the adsorbate that differs from that of the substrate points away
from the crystal interior. Once adsorbed, the additive inhibits the regular deposition of oncoming layers of substrate
molecules, slowing down the growth perpendicular to that
face and leading to a concomitant relative increase in its
surface area. A similar mechanism had been invoked by
Smythe[61in his elegant studies on the changes in habit of
the crystals of sucrose grown in the presence of oligosaccharides as additives.
Once this mechanism was established, it became possible to exploit it in order to systematically modify the morphology of crystals by tailoring additives which bind at a
preselected face and thus inhibit growth in a predictable
8"Fig. 5. Top: Crystallization of (R)- and (S)-glutamic acid. HCI in the presence of NL-(2,4-dinitrophenyI)-(S)-lysine.(a) First crop: colorless crystals of
(R)-glutamic acid. HCI; (h) second crop: mixture of colorless (R)-glutamic
acid. HCI and yellow (S)-glutamic acid. HCI ; (c) third crop primarily contains small, yellow crystals of (S)-glutamic acid. HCI.-Bottom: Crystals of
glycine grown at an air/water interface from aqueous solutions (concentration: 0.3 g/mL) in the presence of (d) (S)-leucine+ NC-(2,4-dinitrophenyI)(S)-lysine and (e) (R)-leucine N"-(2,4-dinitrophenyI)-(S)-lysine (see Section
o n faces 1
change of habit
normal growth
k, S k 2
This is revealed independently by morphological changes
and enantioselective occlusion of additives. The method is
related to the quasi-racemate technique of Fredga,[''l but
has the advantage of a wider applicability.
3. Morphological Crystal Engineering
The dramatic morphological changes associated with the
growth of organic crystals in the presence of additives reveal the high degree of specificity in the interaction of the
foreign material with the different structured surfaces of
the crystalline matrix. The morphological changes have
therefore a direct bearing on the mechanism of the adsorption-inhibition process on a molecular level.
In general, when growth is inhibited in a direction perpendicular to a given face, the area of this face is expected
to increase relative to those of other faces of the same crystal (Scheme 3). Differences in the relative surface areas of
the various faces of pure and affected crystals allow, therefore, identification of the affected faces and consequently
of the crystallographic directions involved. This type of
morphological analysis was carried out on a variety of organic compounds crystallized in the presence of additives
Scheme 3. Formation of crystal faces f with surface area S as a function of
their selative growth rates k. The change in habit results from selective adsorption of inhibitor o n faces of type 1. Left: Habit during normal growth.
Right: Change of habit by inhibition of growth of faces I .
This approach is illustrated here for benzamide, which
crystallizes from ethanol in the form of (001) plate-like
crystals elongated along the b axis (Fig. 6a). The main
packing feature of the structure (Fig. 7) consists of the typical ribbon m o t i ~ 2 0of
1 hydrogen bonded cyclic dimers interlinked by NH . . .O bonds along the b axis. Retardation
of growth of the (011) faces along b was achieved with benzoic acid as additive. The substitution of a molecule of
benzamide at the end of the ribbon motif by benzoic acid
results in replacement in the b direction of an N H . . . O
bond by a repulsive 0 . . .O interaction, thus retarding
growth in this direction and inducing the growth of barshaped crystals elongated in a (Fig. 6b). Inhibition of
growth along a was accomplished by adding o-toluamide
to the crystallization solution. The crystals consistently
grew as bars elongated in b (Fig. 6c). The additive o-toluamide can be easily adsorbed in the hydrogen-bonding
chain without interfering with growth in the b-direction.
The o-CH, group emerges, from the (104) side face, and
A n y e w . Chem. Inr. Ed. Engl. 24 (1985) 466-48s
thus interferes with growth along the a direction, in which
the dimers are stacked (Fig. 7b). Finally, thinner and thinner plates are obtained by adding increasing amounts of p toluamide, whose methyl substituent perturbs the already
weak van der waals interactions between the phenyl layers
in the c direction (Fig. 6d, 7b).[*"
Similar morphological changes were induced in crystals
of other primary amides and of other organic molecules,
including carboxylic acids, amino acids, dipeptides and
In accounting for the effect of a carboxylic acid additive
on the crystal habit of the corresponding primary amide
we assumed that the carboxyl group adopts the commonly
observed synplanar conformation lOa, as against the
rarely observed antiplanar conformation 10b.['21
Fig. 6. Crystals of benzamide: (a) Crystallization with no additive. (b) - (d)
Crystallization in the presence of increasing amounts (from left to right) o f
(b) benzoic acid; ( c ) o-toluamide; (d)p-toluamide.
benzoic acid
I direction
re p o L s ion
Fig. 7. (a) Schcmatic representation of part ot the benzamide structure showing a benzoic acid molecule inserted at the end of the ribbon motif of hydrogen bonded dimers along the b axis. (b) Packing arrangement of benzamide viewed along the h axis (b axis perpendicular to plane of paper).
Anqew. Chem. I n r . Ed.
Engl. 24 (1985)466-485
This postulate was eventually demonstrated by a diffraction experiment on the host-additive system asparagine-aspartic acid'231and confirmed by potential energy calculations o n benzamide-benzoic acid.1241
Crystals of asparagine. H 2 0 undergo a change in habit
when grown in the presence of aspartic acid"" (Fig. 8a).
We explain this change using arguments akin to those
given above for benzamide-benzoic acid. The affected
crystals are (010) platelets, whose formation can be rationalized in terms of adsorption of aspartic acid on an (010)
face with the condition that its hydroxyl oxygen atom
emerges from the face. Further growth at such a surface
site is inhibited by the replacement of an amide
hydrogen bond between asparagine moleN-H . . .Ocarhoxy
cules by an Ohydroxy.
. .Oc.,rhoxy
repulsion (Fig. 8b). The adsorption-inhibition is compatible only with a synplanar
conformation (see 10a) in the host crystal.
The concentration of occluded aspartic acid (15%) was sufficient to allow us to establish the conformation of the acid
by low-temperature X-ray diffraction (Fig. 8c).
Energy calculations were carried out o n the benzamidebenzoic acid system in order to quantify the substrate-additive interactions on the basis of the established mechanism of adsorption and inhibition. To ascertain on which
face the additive may be adsorbed with the least loss in energy, relative to the pure substrate molecule, we calculated
the binding energies of the additive and substrate molecules a t the different symmetry-related sites on the various
crystal faces (Fig. 6a). It was found that benzoic acid with
a synplanar O=C-0-H conformation is most easily adsorbed on the (01 1) face, with an energy loss of 7 kcal/mol.
This loss of energy relative to benzamide originates primarily from the fact that the amide oxygen atom is a much
stronger proton acceptor than the corresponding carboxy
oxygen.'"] Once bound, such a benzoic acid molecule interacts with an oncoming benzamide layer with an energy
loss as high as 10 kcal/mol relative to a corresponding
benzamide molecule adsorbed at the site of the additive.
An analogous calculation for the antiplanar conformer of
47 1
Asn+ Asp
I_ I
1 role)
The essence of the problem is that conventional X-ray
diffraction analysis of chiral crystals (i.e. assumption of
Friedel’s law) does not allow assignment of the absolute
orientation of the chiral molecule with respect to the crystal axes or crystal faces.1251
In other words, it is not possible
to distinguish between the polar structure of Scheme 4a
Scheme 4. Schematic representation o t the [wo poarible endntioineric crystal
structures ( a ) and ( b ) of the compound X-A with respect to the polar axis b
(space group PZ,);
interaction of the polar faces with the additives X - Y and
Z-A. a, b see text.
(where the chiral molecule is represented by X-A) and the
enantiomeric structure of Scheme 4b (in which the orientation of X-A with respect to the polar b axis is reversed). A
pictorial representation of a polar motif from a drawing by
Escher is given in Figure 9.
Fig. 8. (a) Crystal morphologies of pure (S)-asparagine’ H 2 0 (left) and of (S)asparagine.H20 grown in the presence o f (S)-aspartic acid (right); (b) crystal
structure (stereoview) of (S)-asparagine. H 2 0 with occluded (S)-aspartic acid
(full circles); (c) molecular structures (stereoviews) of asparagine and aspartic acid in the crystal structure of asparagine (OH g r o w full circle).
benzoic acid shows that it may be adsorbed and hinder
growth only at faces which are experimentally found to be
unaffected. This provides further support for the assignment of a synplanar conformation.
These theoretical and empirical considerations allow us
both to understand and to predict the induced changes in
crystal habit. We shall now show how this knowledge was
applied to the solution of certain fundamental problems in
4. Direct Determination of the Absolute
Configuration of Chiral Molecules
Ever since the recognition of chirality in molecules, the
direct assignment of the absolute configuration of a chiral
molecule in a chiral crystal which develops hemihedral
faces has challenged the chemical crystallographer.
tig. 9 Vrabing by M. C. C d l r r . displaying polar ayrninetrq. t h e fiah b \ \ i i i i i n
one direction while the birds fly in the opposite one. (Copyright M. C.
Escher heirs, c/o Cordon An-Baarn- Holland.)
In 1949, Wuser‘261
tried to establish the absolute configuration of (D)-tartaric acid by correlating the relative rates
of growth of the hemihedral (hkl) and (hkl) faces with the
ease of attachment of the “free” molecule at either face in
terms of intermolecular distances between the crystal and
the molecule to be attached. In fact, as Turner and Lonsdale pointed
given an asymmetric molecule X-A in
Angew. Chem. Inr. Ed. Engl. 24 (1985) 466-485
a polar crystal, as in Scheme 4, there is no a priori reason
why, on the basis of intermolecular distances only, the attachment of a n X to a n “A face” should take place more
readily than that of an A to an “X face”. The observed differences in development of hemihedral faces may be explained primarily in terms of surface-solvent interactions
or polarizability effects.
We developed a new method, based on controlled morphological modifications through adsorption of tailormade additives, for the direct assignment of the absolute
configurations of chiral molecules. First, we shall discuss
the case of polar crystals.[28’
4.1. Polar Crystals
Scheme 4 describes an appropriate polar crystal beThe unique axis of the crystal
longing to space group R1.
is parallel to the X - A direction. Faces f l and f2 delineate
the crystal in the polar +6 direction and f3, f4, f5 in the
-6 direction. The faces within each pair f l , f2, and f3,f5,
are homotopic since they are related by 2-fold symmetry.
Since the crystal is polar, the structures of the faces at the
+6 and -6 ends are different. By application of the twostep mechanism of adsorption-inhibition, in the crystal of
Scheme 4a an inhibitor X-Y will bind selectively at faces
f l and f2, and once bound, will retard growth along +6
and possibly other directions but not along - b . In an
analogous way an inhibitor Z-A will hinder the growth of
faces f3, f4 and f5 but not of f l and f2 (the opposite would
happen in the structure of Scheme 4b). Such retardation
will be associated either with an increase in the areas of
the inhibited faces with respect to the unaffected ones or
with the appearance of new faces on the affected side of
the crystal. The observed morphological differences between crystals grown in the presence and in the absence of
the additive thus establish the direction of the substrate
molecule X-A with respect to the polar axis. Consequently, the absolute configuration of the crystal and that of its
chiral molecular constituents can be derived.
We shall briefly illustrate here the case of lysine.HC1
and cinnamoyl alanine.[281(S)- and (R)-lysine. HCI crystallizes from water as a dihydrate in a monoclinic structure of
space group ml.
The packing arrangement delineated by
the crystal faces of the unaffected crystal is shown in Figure 10 (cf. Fig. 1 la). In the crystal the lysine molecules are
aligned parallel to the 6 axis with the @ H 3 N - C H - C 0 0 Q
moiety emerging from the f 6 end of the crystal and the E NHY pointing towards -6. In agreement with the analysis
presented above, additives with a modified carboxyl or aamino group, like lysine methyl ester, indeed inhibited
growth in the +6 direction, inducing development of the
(010) face (Fig. l l b ) . Conversely, additives which bear a
modified c-amino group, like norleucine or norvaline, inhibited growth along the -6 direction, with a concomitant
pronounced increase in the areas of the (li0) and (ii0)
faces (Fig. 1 Ic).
The observed changes in morphology implied that the
occluded inhibitor would be distributed anisotropically
along the polar 6 axis of the crystal. This expectation was
experimentally confirmed for the crystals of (S)-lysine. HCI grown in the presence of (5‘)-norleucine. AccordAnyew. Chem. I n t . Ed. En@. 24 11985) 466-485
Fig. 10. Packing arrangement of the crystals of (S)-lysine. HCI. 2 HIO viewed
along the c axis, as delineated by the observed [hkOl crystal faces.
Fig. 1 I. Computer-drawn pictures of typical crystals of (S)-lysine.HCI .2 H 2 0 viewed along the c axis, grown in the presence of: (a) no additives; (b) (Sf-lysine methyl ester; and (c) (S)-norleucine.
ing to a chromatographic analysis of the material taken
from the +6 and -6 ends of the crystal, the additive was
occluded preferentially at the - b end (Fig. 12).
I \
t [min]
Fig. 12. HPLC analysis oF(S)-norleucinc occluded at [he two opposite sides
of the polar b axis of crystals of (Qlysine. HCI ‘ 2 H 2 0 . In experiments @
and @ the analyses were performed on equal amounts of (S)-lysine. HCI. 2 H1O.
Fig. 13. Packing arrangement of the crystals of cinnamoyl-(S)-alanine as delineated by the observed {hko crystal faces.
Cinnamoyl alanine crystallizes in space group P 2 , (cinnamoyl = trans-3-phenylpropenoyl). In the crystal structure of the (S)-enantiomer all the OH groups form hydrogen bonds with a major component along -b, while
the C(chira1)-H bond is directed along +b. Figure 13
shows the orientations of the molecules with respect to the
crystal faces. The pure crystals are depicted in Figure
Fig. 14. Computer-drawn pictures of typical crystals of cinnamoyl-(Sj-alanine (viewed along the c axis) grown in the presence of: (a) no additive; (bj
methyl cinnamoyl-(S)-alaninate; and (c) cinnamoyl-(Rj-alanine.
We found, as expected, that the methyl ester of the substrate with the same absolute configuration induced large
(1 i I) and (i 1) faces (Fig. 14b), thus fixing the absolute direction of the 0 - H bond with respect to the polar axis and
hence also the absolute configuration of the substrate molecule. On the other hand cinnamoyl (R)-alanine, of configuration opposite to that of the host, induced, as expected,
formation of an (010) face (Fig. 14c), because adsorption
of this molecule can take place only from the + b side of
the crystal since the C-CH3 group of the inhibitor points
away from the surface along the + b direction.
Application of the present analysis to the morphological
differences observed by Srnythd6] in polar crystals of sucrose grown in the presence and in absence of raffinose
(and other di- and tri-saccharides), enabled us to assign the
absolute configurations of these sugars.
4.2. Enantiopolar Crystals
The above method of determining absolute configuration is not limited to the molecules in polar crystals, but is
also applicable to chiral molecules which induce morphological changes in centrosymmetric crystals composed of
racemic mixtures or meso compound^.[^^^'"^ The method is
based on the fact that, unlike chiral crystals, in centrosymmetric ones the orientations of the constituent molecules
with respect to the crystal axes are unambiguously assigned by conventional X-ray analysis. The known orientation of the enantiomers in such crystals can be exploited
for the direct assignment of the absolute configuration of
chiral resolved additive molecules, provided this structural
information can be transferred to such molecules. The absolute configuration of the resolved additives is determined through the morphological changes they selectively
induce on the enantiotopic faces of appropriately selected
Scheme 5 . Structure of an enantiopolar crystal and interdction o f t h e enantiotopic faces with enantiomeric additive molecules, schematic.
A prerequisite for application of this method is that
within the centrosymmetric racemic crystal a specific functional group (A) attached to an (R)-molecule will point towards the face f i but not towards fl (Scheme 5). Correspondingly, the same functional group attached to an ( S ) molecule will emerge at the enantiotopic face f l , but not at
f i . It is useful here to regard centrosymmetric crystals containing chiral molecules as enantiopolar. We define an enantiopolar crystal as comprising two enantiomeric sets of
intermeshed polar crystal structures related to each other
by symmetry elements of the second kind, e.g. a center of
Angew,. Chem. In!. Ed. Engl. 24 11985) 466-485
(S)-threonine are (2R,3S)- and (2S,3R)-configurated, respectively.
Fig. 15. Drawing by M. C. E.Tcher displaying enantiopolar symmetry: treating
the white and the black horses as chiral objects of opposite handedness, the
two enantiomeric ensembles of the white and the black horses are enantiopolar, pointing in opposite directions along the horizontal axis. (Copyright M.
C . Escher heirs, c/o Cordon Art-Baarn-Holland.)
Fig. 16. Packing of (RS)-serine viewed along the a axis (stereoview). The enantiomers are packed in homochiral bc layers: for clarity only half of each R
(open circles) or S (full circles) molecular layer is shown, delineated by the
four (0111 crystal faces.
inversion or a glide plane. This is also neatly illustrated in
a drawing by Escher (Fig. 15). Crystallization of such compounds in the presence of a chiral additive R‘, appropriately designed so that it will fit in the site of an (R)-molecule on the growing crystal faces f l or f2, but not on the
enantiotopic faces f l or f2, will hinder growth along the
- b direction but not along f b . By virtue of symmetry, the
enantiomeric additive (S’) will inhibit growth of faces f l
and f2, while the racemic additive (R’S’) will inhibit
growth along both directions, + b and - b .
These principles were applied to the monoclinic crystals
of (RS)-serine‘”] and g l y ~ i n e . [Racemic
serine crystallizes
in a molecular packing appropriate for this analysis, space
group P 2 , / a (Fig. 16). Within the structure, the C-H(Si)bond vector“] of the rigid methylene group of (R)-serine
has a major component along +b and, by symmetry, the
C-H(Re)-bond vector of the (S)-serine (S)-11 along -b.
H is)-C-OH
H (R)
H (S)
Thus, their replacement by a methyl, as in threonine, will
inhibit growth along the b direction. Namely an ( R ) threonine molecule with the side-chain P-carbon of chirality (S), will inhibit growth along + b , while the -CH3
group of (S)-threonine will replace the pro-(Re) hydrogen
of (S)-serine and hence inhibit growth along - b ; (R)- and
[*I H ( S ) denotes thepro-S H-atom and
H2N-C-H ( R )
H(Re) thepr<+R H-atom.
Anqew. Chem. i n [ . Ed. Engl. 24 (IPSS)466-485
Fig. 17. Photographs and computer-drawn pictures of cr)\tdl> uI (KSj-~crine
grown in the presence of: (a) no additives; (b) (R)-threonine; ( c ) (S)-threonine; (d) (RS)-threonine; and (e) (R)- or (S)-allothreonine ((2R.3R)- and
(2S,3S)-configurated, respectively).
(RS)-serine forms tabular crystals, with point symmetry
2 / m (Fig. 17a); the crystals affected by either (R)- or (S)-
threonine exhibit reduced morphological symmetry 2 (the
mirror plane is lost) and are enantiomorphous (Fig. 17b
and 17c, resp.). When (RS)-threonine is used as additive,
the morphological symmetry 2/m is left unchanged as a result of a combination of the effects induced by each additive separately. The crystals turn into rhombs (Fig. 17d).
The morphological changes, and our interpretation
thereof, imply that in this last experiment (RS)-threonine
must segregate along the 6-axis during crystal growth; occluded (R)-threonine will prevail in the + b half of the
crystal, whereas (S)-threonine will prevail in the -6 half.
This expectation was confirmed experimentally by HPLC
analyses (Fig. 18a, b, c) through work done in collaboration with S. Weinstein using the chiral mobile phase of
N. N-dimethyl-(S)-valine and cupric
The discriminative power of the various faces of a crystal of (RS)-serine is further manifested by the differences
t [rnin]
[RS 1-
(R)-aT hr
t [min]
Fig. 18. t n ~ n ~ i u i n c r di ici d l ) \ i \ uI Llirwriiiic occluded in the rhomb-like crystals of (RS)-serine (Fig. 17d) after separation of serine by cation exchange:
Material taken from: (a) tip of the crystal at the + h end; (b) tip of the same
crystal at the - b end; (c) whole crystal. Enantiomeric analysis of allothreonine (Fig. 17e) occluded in (RS)-serine crystals: (d) from the + h end:
(e) from the - h end.
in morphology induced by the additives threonine as
against allothreonine (cf. (S)-13, and Fig. 17d and 17e,
resp.). The established mechanism of adsorption-inhibition
allows a rationalization of this difference. The relative
orientation of molecular serine vis-a-vis its various crystal
faces, suggests that allothreonine can be adsorbed on the
homotopic (100) faces as well as on the enantiotopic (011)
faced'] (Fig. 17e).
This would imply that no distinct morphological differences can be expected between (RS)-serine crystals grown
in the presence of resolved or racemic allothreonine, as
was indeed observed (Fig. 17e). Furthermore, (R,S)-allothreonine, in contrast to (RS)-threonine, should not segregate enantiomerically along the b axis; this was confirmed
by HPLC analysis of tips of the crystal which were taken
from the opposite ends of the b axis (Fig. 18d, e). The effect of threonine and allothreonine on the crystals of (RS)serine is further independently explained by dissolution
experiments (see Section 6) and by atom-atom potential
energy c a h l a t i ~ n s . ~ ' ~ l
According to the surface binding energies of the additives threonine and allothreonine relative to pure serine at
several crystal faces, (R)-threonine may be easily adsorbed
only on faces (011) and ( O l i ) and, by symmetry, (S)threonine should preferably adsorb on (Oii) and (Oi 1); in
fact threonine is better bound than substrate serine by 3
kcal/mol. (S)-allothreonine may be easily adsorbed on
{loo},(01 1) and (01 i), and, by symmetry, (R)-allothreonine
on faces { IOO), (01 i) and (Oi I). The binding energies of allothreonine relative to serine are - 2. l kcal/mol at { 100)
and -0.7 kcal/mol at {Ol I}. These results are completely
compatible with experiment. In particular, we account for
the observation that the resolved and racemic allothreonine do not induce obvious changes in habit because they
can each be adsorbed on two principal faces. Moreover,
the possibility of occluding ( R ) - and (S)-allothreonine
through the large (100) faces masks any enantioselective segregation along the b axis in accordance with the HPLC
results (Fig. 18d, e).
The ability of the serine crystals to recognize the subtle
differences between diastereotopic hydrogens suggested
that the crystal of prochiral glycine may serve as a matrix
for the assignment of absolute configuration of all the chiral a-amino acids.[301The glycine molecule, although nonchiral, has two prochiral hydrogen atoms, H(R) and H(S).
In the a-modification of the crystal with space group P 2 , / n
(Fig. 19), the glycine molecules form two enantiopolar sets,
which are represented in Figure 19 by shaded and nonshaded atoms, respectively. All the molecules in the nonshaded set have the C-H(R) bond aligned with the b axis
and pointing towards +b, so that C-H(R) emerges from
the (010) face, while C-H(S) lies almost perpendicular to
it in the ac molecular layer. Only an (R)-amino acid additive, whose side chain will be parallel to the C-H(R)
bond, can be adsorbed at the (010) face and would subsequently hinder the crystal growth perpendicular to it. By
virtue of symmetry, an @)-amino acid can be adsorbed
only at the (OiO) face.
The symbol {/Anrepresents all the symmetry related faces, e.g in serine,
(01I / denotes (01 I). ( O i l ) , ( O l i ) and (Oii) faces. The symbol (hkl) represents only the specified (hkl) face.
Angew. Chem. Int. Ed. Engl. 24 (1985) 466-485
logical change at the - b side of the crystals, with the appearance of pyramids exhibiting a predominant (010) face
(Fig. 20b). All (R)-amino acids induce the development of
a racemic additive
a large (010) face (Fig. ~ O C ) while
causes crystallization of {OlO) platelets (Fig. 20d).
Amounts in the range 0.02 to 0.2% of the additives were
in the bulk of the glycine crystals. HPLC analysis
(Fig. 21) of these occluded a-amino acids inside single
Fig. I!, Picking 01 glycine crystals uith lakerb of glycine molecules parallel
uc plane (stereoview along the a axis). The two enantiopolar sets are
represented by shaded and nonshaded molecules.
to the
t [min]
Pure glycine crystallizes from water in the form of bipyramids, with the h axis perpendicular to the base of the
pyramid (Fig. 20a). It was observed that all natural (S)-amino acids, apart from proline, induce a dramatic morpho-
Fig. 21. Enantiomeric distribution of occluded (RSj-glutamic acid in the (010)
plate-like crystals of glycine by HPLC: (a) Only additive (S)-glutamic acid IS
present in material shaved from the (OiO) face of the crystal: (b) only additive
(R)-glutamic acid is present in material shaved from the (010) face: (c) relative distribution of (RS)-glutamic acid in the remaining whole crystal.
S = sensitivity in arbitrary units.
Fig. 20. Comparison between crystals of a-glycine grown in the presence of: (a) n o additives; (b) (S)-a-amino acids: (c) ( R )
a-amino acids: (d) racemic a-amino acids.
Angeu. < ' l i w r f . liii.
Ed. Engl. 24 I I Y R S ) 466-485
plate-like crystals of glycine demonstrated a total enantioselective segregation along the b axis; as expected, the (R)amino acids predominate in the b half of the crystal and
the (S)-amino acids in the - b half. This segregation suggests that the overall symmetry of the crystal has been lowered from P 2 , / n to P2,.
The changes in crystal habit of glycine as studied by potential energy caIcuIation~[~"]
were completely consistent
with the above-mentioned observations.
5. A New Model for Spontaneous Generation of
Optical Activity
The results reported on the glycine system acquire further interest as the (010) plate-like crystals of glycine,
grown in the presence of other (RS)-a-amino acids, have a
tendency to float at the air/water interface with the b-axis
perpendicular to the interface.[331Each such floating crystal will grow either along its +b or - b direction, depending upon whether face (010) or (010) is pointing towards
the water solution. Thus, if a single crystal of glycine exposes its (010) face to a solution of glycine containing a
mixture of (RS)-amino acids, only the (R)-acids will be occluded, if the (OiO) face is exposed to the solution, the (S)acids will be occluded. HPLC analysis of the enantiomeric
content of the occluded amino acids indeed confirmed
these expectations. Such resolutions of (RS)-leucine and
(RS)-valine are shown in Fig. 22a, b, c.
It is clear that in a racemic mixture there is equal probability for a given crystal to float on its (010) or (OiO) face.
We have observed, however, that if glycine is crystallized
in the presence of small amounts of resolved amino acids
bearing hydrophobic side chains, like (S)-leucine, (S)-phenylalanine, or (S)-a-aminooctanoic acid, the floating crystals of pyramidal shape always expose their (010) face to
the air. In fact, when crystals were grown in the presence
of equal amounts of two amino acids with side chains of
different character and opposite absolute configuration,
for example (S)-leucine with (R)-glutamic acid or (R)-alanine, or (R)-serine, all the resulting platelets had their (OiO)
faces directed to the air. This indicates that stereospecifically adsorbed leucine molecules are capable, even in tiny
amounts, of orienting the glycine crystals by virtue of the
hydrophobic character of their side chains.
The tendency of leucine molecules to orient floating glycine crystals and the enantioselective adsorption at the
face exposed to the solution were exploited to induce resolution of racemic amino acids by crystallization of glycine
at interfaces in the presence of resolved leucine; for example glycine, grown in the presence of racemic hydrophilic
amino acids plus small amounts of (S)-leucine, yielded
platelets all floating with their (OiO) faces pointing upwards, thus selectively occluding the (R)-enantiomers (Fig.
23a, b). Furthermore, even under conditions where polycrystalline crusts of glycine were formed (Fig. 23c, d) resolution occurred with an enantiomeric excess higher than
70% (Fig. 23d) in the examples studied. This effect was
also demonstrated visually by the formation of polycrystal478
Fig. 22. (a) Floating plate-like crqstals of glycine grown from a hater solution
(RS)leucine and I"'n-h/w (RS)-valine.
The crystals display both enantiomorphous morphologies, labelled rn and Ei;
(b) HPLC analysis of crystals m with the (010) face exposed to solution, after
washing with water: (c) HPLC analysis of glycine crystals ,Ti with the (0iO)
race r x p o x d to the solution after washing with water.
( 4 . 4 ~ )containing racemic I%w/w
line crusts of glycine grown in the presence of (S)-leucine N' -(2,4-dinitrophenyl)-(S)-lysine and (R)-leucine +
N' -(2,4-dinitrophenyl)-(S)-lysine. We obtained respectively colorless and colored crusts of glycine (Fig. 5d, e in
Section 2).
The experiments described above provide us with a possible model for the generation and amplification of optical
activity in an initially symmetrical world. Let's assume that
the first plate-like crystal of glycine formed from the 'prebiotic soup' floats with its (010) face upward. During
growth the crystal will occlude exclusively the (S)-amino
acids present in the "soup", thus leaving the solution enriched in the (R)-enantiomers. The excess of hydrophobic
(R)-amino acids so generated will in turn direct the floating of later-formed crystals with their (010) faces again upward, thus ever further enriching the solution with (R)-enantiomers.
6. "Tailor-Made" Etchants for Organic Crystals
Growth and dissolution of crystals are considered reciprocal processes that can be interchanged by altering the degree of saturation of the solution. This implies that an inhibitor or growth of a given crystal face must in principle
affect the rate of dissolution of that same face.['' In order to
Angew Cliewr Itir. Ed Engl. 24 (1985) 466-485
&I ' ' -0
Fig. 23. (a) Floating plate-like crystals of glycine grown in the presence of I"/o-w/w (S)-leucine and racemicp-hydroxyphenylglycine (0.5'L M./ w), racemic glutamic
acid (194,) and racemic methionine (O.Sa/o). All the crystals float with their (Oi0) faces exposed t o air and exhibit the same enantiomorphous morphology; (h) HPLC
analysis of the enantiomers of p-hydroxyphenylglycine (pHPG), glutamic acid, methionine and leucine occluded inside the Erystals; only the (R)-enantiomers of
pHPG, glutamic acid and methionine are found; (c) crystalline crust of glycine grown at the air/water interface. The aqueous solution ( 4 . 4 ~ contains
I"h-w/w (S)leucine and 3"b-w/w racemic glutamic acid; (d) HPLC analysis of the same crust, indicating an ee of 70% of occluded (R)-glutamic acid.
test this reciprocity, a systematic study was undertaken on
crystal dissolution in the presence of "tailor-made" growth
inhibitors. As models, we investigated some of the systems
we described above for crystal
Plate-like crystals of glycine (see Sections 4 and 5 ) with
well-developed (010) faces were submitted to partial dissolution in an undersaturated solution of glycine containing
variable amounts of other a-amino acids. When resolved
(R)-alanine was present in the solution, well-developed
etch-pits were formed only on the (010) face. These pits exhibit twofold morphological symmetry with surface edges
parallel to the a and c axes of the crystal (Fig. 24a). The enantiotopic ( O i O ) face dissolved smoothly (Fig. 24b), exactly
as it does when the crystal is dissolved in an undersatu-
Fig. 24. Optical microscope pictures ( x 5 0 ) of the (0101 faces of a glycine crystal after partial dissolution in the presence of (0-alanine: (a) the (010) face;
(h) ( O i O ) face
A n g e u . Cliern I n r . Ed. Engl. 24 il985J 466-485
rated solution of pure glycine. As expected, (S)-alanine induced etch-pits on the (OiO) face. Racemic (RS)-alanine
etches both (010) faces.
An explanation of these, and other similar results, takes
into consideration that dissolution from a given surface is
considered to begin at sites of emerging dislocation^.^^^^ At
these sites, more than one type of facial surface is exposed
to the solvent (Scheme 6 ) . The dissolution front propagates
Scheme 6. Formation o f etch pits at the sites of emerging dislocations as a result of selectjve binding of additive to the horizontal faces.
from these centers in various directions, at the same relative rates as for their growth. Under such circumstances
the overall shape of the crystal is preserved. On the other
hand, when an additive which binds selectively to a given
face of the crystal is present in the solution, the crystal will
dissolve in different directions with rates which differ from
those in pure solvent. As a result, the rate of dissolution in
a direction perpendicular to the affected face will decrease
relative to that perpendicular to the unaffected ones, and
etch pits will be formed at the dislocation centers on the
former faces.
The generality of the above stereochemical correlation
was tested for all the systems described above, including
serine/threonine and serine/allothreonine (see Section 4)
and for the enantiomorphous pairs of glutamic acid. HCI
and of asparagine. H,O. When rhomb-like crystals of (RS)serine are partially dissolved in the presence of (R)-threonine, well formed etch-pits develop at the (01 I ) and (01 i)
faces (see Figs. 16, 17). By virtue of symmetry, similar etchpits form at the (Oi 1) and (Oil) faces when (S)-threonine is
present in the solution. In neither experiment were etchpits formed at the (100) faces. On the other hand, both resolved and racemic allothreonine etch primarily the (100)
faces, in keeping with the results of the morphological and
segregation studies and with theoretical calculations.
The present method was also exploited for the manual
separation of a pair of enantiomorphous crystals. For example, when a crystalline conglomerate of asparagine . H 2 0 was partially dissolved in a solution of (S)-aspartic acid, well formed etch-pits were formed only on the
(010) faces of the (S)-crystals, but not on the (R)-enantiomorph (Fig. 25). The treated enantiomorphous crystals are
thus visually distinguishable and, therefore, separable.
In several ways the etching experiments yield results
complementary to those obtained from the studies on morphological changes and enantiomeric segregation. On the
other hand, etching provides a unique probe for differentiating between the two steps of binding of inhibitor and
retardation of growth. Furthermore, the etching experiments should provide a decisive insight into those systems
where the inhibitors appear to bind at one face but retard
growth at another. In addition, increase of the surface area
upon etching by the additives should influence the overall
rate of the crystal dissolution. This approach is currently
Fig. 25. Scanning electron micrograph of the (010) face of crystals of asparagine partially dissolved in the presence of (S)-aspartic acid; (a) (Sj-asparagine crystal; (b) (R)-asparagine crystal. White line: 100 urn.
being resorted to for increasing the rate of dissolution of
organic crystals in general, and of sparingly soluble drugs
in particular.
7. Simulation of Crystal Morphology by Energy
Atom-atom potential energy calculations were undertaken in order to gain an understanding of crystal morphology in terms of the intermolecular interactions between substrate molecules in the crystal bulk, and between
substrate and solvent or substrate and additive at the crystal solvent interface.
A controlling factor in crystal growth is the energy of the
interactions between neighboring molecules. According to
Hartman and Perdok,Li61
the crucial relationship is between
the attachment energy E,,,,, the energy per molecule released when a new layer is attached to the crystal face, and
the layer energy E , which is defined as the energy per molecule released when a new layer is formed. E, measures
the stability of a layer and E,,,, controls the growth rate perpendicular to this layer. Studies on the relationship between the internal structure of the crystal and its morphology led to the working hypothesis that the morphological
importance of a crystal face decreases with increasing atAnyeh?. Chem. Int. Ed. Enyl. 24 (1985) 446-485
tachment energy. From the known crystal structure it is
possible to calculate the corresponding values of E,,, for
various low-index faces. This information is used to derive
the theoretical form, i.e., the calculated morphology of the
crystal which best represents the crystal habit obtained by
the habit of the crystal grown
s ~ b I i m a t i o n . [ ~ ’To
from solution in the presence or absence of additives, substrate-solvent and substrate-additive interactions at the
crystal-solution interface must be taken into account.
habit is distinctly different from that of glycine grown from
aqueous solution. The most striking differences between
the two forms (Fig. 26 and Fig, 20, resp.), are the complete
disappearance in the latter of the {OOl) and { I O i } faces, the
formation of {Oll), and the drastic reduction in relative
area of (010). We account for these differences in terms of
the higher affinity of the water solvent for the crystal faces
(110) and (011) than for (010) and (OOl}.
7.1. The Theoretical Form of Glycine and the Solvent
The theoretical crystal form of glycine, depicted in Figure 26 (left), is in nice agreement with the morphology of
glycine crystals obtained by sublimation (Fig. 26, right). Its
L a
Fig. 28. Calculated crystal form of glycine incorporating the solvent effect.
Fig. 26. left. Theoretical crystal form of glycine; right: morphology of glycine
crystals obtained by sublimation.
To establish the relative hydrophilic affinities of these
faces the Coulomb potential energy was calculated at closest approach distances from the surface of each of the
four faces (Fig. 27). From these maps we may conclude
that the (OlO} face is the least polar and (01 I) the most polar face. Thus, preferential adsorption of water molecules
on the polar faces, primarily on (01I}, causes a reduction in
the growth rate normal to these faces relative to that of
(OlO}, with a concomitant increase in their surface areas.
When this solvent effect was simulated in the calculation
of the theoretical f ~ r m [ ~ ’ .we
~ *obtained
the habit shown in
Figure 28, which is in agreement with the observed morphology (Fig. 20a). Several other systems were studied, in-
. i .... :.. ....
. . . . .. . .. .. . .. . .
. . ~
, , < - %
Fig. 27. Electrostatic potential energy maps at closest approach distances of a unit positive charge from the surfaces of various glycine crystal faces. Solid contour
lines represent repulsive energies, and broken lines attractive energies. (a) (01 1) faces; (b) (110) faces; (c) (010)faces at the layer exposing the C-H bonds (Le. unshaded molecules at the (010) surface in Fig. 19). (d) (0101 faces at the layer exposing the N H bonds (i.e. the shaded molecules at the (010) surface in Fig. 19). (e)
Anqew. Chem. I n t . Ed. Engl. 24 /198Sj 466-485
48 1
cluding benzamide, ~ i n n a m i d e [ ~ ~and
. ~ *~~e ,r i n e [ ~ ’ and
yielded theoretical forms consistent with the observed habits.
7.2. The Riddle of Resorcinol
Polar crystals are ideal for elucidating the effect of solvent on crystal growth, because the difference in growth
rates of opposite faces (hkl) and (hkl)along a polar direction originate primarily from the differences in their solvent-surface interactions. However, due to the practical
difficulty in determining the absolute polarity of crystals,
in particular for crystals composed of nonchiral molecules
and containing light atoms only, this approach has been
somewhat neglected to date. Our technique of “tailormade” additives for the determination of the absolute configuration of crystals, allowed us to tackle such a problem
and clarify the influence of solvent on growth of the polar
crystal of a-resorcinol, 15.
In 1949, Wells[3y1found that in an aqueous solution the
a-form of resorcinol (space group Pna2,) grows unidirectionally along the polar c axis. The crystal (Fig. 29, left) exhibits “phenyl-rich” (01 1) and (Oi 1) faces at one end of the
c axis and “hydroxy-rich” (Oli) and (Oii) faces at the
other end (Fig. 29, right). The absolute direction of
growth along the polar axis with respect to the crystal
structure could not be fixed at that time, however, so that
Wells did not know which end of a given crystal was phenyl-rich and which end hydroxy-rich. Wells interpreted
this unidirectional growth along c to take place at the phenyl-rich faces as a result of stronger adsorption of water to
the hydroxy-rich (Oi i) faces, so inhibiting their growth relative to the phenyl-rich (011) faces.
Recently, D a ~ e y [ ~has
” ] proposed that strong surface-water interactions should enhance crystal growth. By use of
oriented growth on silica surfaces, D a v e y , Bourne, and Milisavljeoi~~~”
deduced that indeed the OH-rich (01l ] faces
are those that grow quicker in aqueous solution. The assignment by this technique is, however, not unambiguous.
In order to resolve the controversy between the two hypotheses we assigned the absolute direction of growth of
the resorcinol crystal in aqueous solution by employing the
additives 16 to 19, and independently by the Bijvoet meth0d.1~~1
17, X=OH;18, X=CH, ;
We found that the unidirectional growth in water takes
place primarily at the oxygen-rich faces in accordance with
the findings of D a v e y , Bourne, and Mili~avljeuic.~~‘~
tend to the paradoxical view that the unidirectional growth
at the hydroxy-rich (Oii) faces is due to inhibition of
growth at the phenyl-rich (011) faces because of a higher
affinity of water for the latter face. We base our arguments
on the relative structures and on the van der Waals and
Coulomb energy potentials of these surfaces.
Fig. 29. Left: Typical crystal of a-resorcinol grown from water. The end faces 101I ) , 1.e. (01 I ) and (Oil), and (Olil, i.e. (017) and (Oii), are marked; right: packing
arrangement of a-resorcinol, stereoview along the a axis. The planes parallel to the phenyl-rich 10111- and hydroxy-rich (Oii)-facesare denoted.
A n q e w Chem. Int. Ed. Engl. 24 (1985) 466-485
The (01 I} face is criss-crossed with grooves forming
pockets at their intersections, whereas the hydroxy-rich
( O i i ) face is relatively flat (Fig. 30). In order to obtain a
big. i u . Keaorcinol. Two diilerent stereoviews of an 01 I molecular layer
showing the surface structures at the (011) and (Oil) faces: (a) along the a
axis; (b) along the direction - b + c .
measure of the relative affinity of these two faces for water
in terms of van der Waals forces, their surface energy potentials were calculated as sensed by a pseudo van der
Waals’ water atom’ brought into best attractive contact
along each surface. The result indicates a higher (14%)affinity of the solvent for the (011) face than for (Oii}in accordance with their relative ‘roughness’. The relative polarity of these two faces in terms of their Coulomb potential
energy surfaces is shown in Figure 31.
The positive potential values at the (011) face indicate
that this face best interacts with negative ions or the oxygen of water, whereas (01i), which is characterized by negative potential values, interacts best with positive ions or
with the hydrogen of water. Thus, for solvent water, it is
preferable to consider the (011) and (Oii) faces as acidic
and basic faces, respectively, rather than as ‘phenyl-rich’
and hydroxy-rich’. The acidic nature of the (011) face is
also evident in so far as it exposes the 0 1 -H bond as well
as aromatic C-H bonds (see Fig. 30) which may participate in attractive C - H . . .O interactions.[431The proton-acceptor properties of the opposite (Oil) face arise from the
fact that three out of four resorcinol oxygen atoms 0 1 , 0 2 ,
OI’, 0 2 ’ are accessible to water (see Fig. 30). The contour
values of the two energy maps indicate that the basic face
has a deeper potential than the acidic face. Thus, the van
der Waals and Coulomb forces appear to balance out in
A n g e w . Chem. In,. Ed. Engl. 24 (1985) 466-485
Fig. 31. Electrostatic potential energy maps at closest approach distances of a
unit positive charge from the surfaces of: (a) a loll) face; (b) a { O i i ) face.
Solid contour lines represent repulsive energies and broken lines attractive
energies. Contour interval 5 kcal/mol.
terms of preferential binding of water; on the other hand,
it is clear that water can bind to both sides. In the light of
the experimental data we tend to believe that van der
Waals forces play the determining role. Thus, we maintain
that Well’s interpretation, that growth of a crystal face is
inhibited by strong adsorption of solvent, is essentially correct, although in the case of resorcinol he drew the incorrect conclusion about the growth direction, by assuming
that water is adsorbed preferentially on the hydroxy-rich
This approach in studying solvent-surface interactions is
encouraging. Indeed, it may eventually lead to a quantitative unraveling of the relative roles played by internal crys483
tal structure and solvent-surface interactions in determining crystal morphology.1381
8. Outlook
In the present article, we have attempted to show that it
is possible with the assistance of tailor-made additives to
control both growth and dissolution of organic crystals. A
remarkable aspect of this process is that it provides us with
the ability to navigate a given additive through a given face
into a selected subset of the symmetry-related sites of the
crystal. Thus, for example, in the case of benzamide, benzoic acid may be occluded at only one of the two centrosymmetrically related sites at the end of the ribbon motif
(Fig. 7a); the overall symmetry of the crystal is thus reduced from the centrosymmetric space group f l l / c to the
chiral space group EL1.Such “induced chirality” is currently being exploited for the performance of spontaneous
asymmetric syntheses. Furthermore, this type of chiral discrimination in centrosymmetric crystals is presumably
manifested also in inorganic crystals, including minerals,
which might have a bearing on the spontaneous separation
of enantiomers in Nature.
The present approach might further provide an effective
method for pinpointing fine intermolecular interactions in
crystals, (e.g. 0 . . .O repulsion in the benzamide-benzoic
acid system, (see Section 3 ) as revealed by the morphological changes they induce.
The role of the tailor-made additives is probably not
limited to the growth stage, but includes prenucleation aggregates which are expected to structurally resemble the
mature crystal. An extension in this direction might be of
importance in the resolution of enantiomers by crystallization.
So far we have confined ourselves to additives which resemble the host molecule. This was done in order to facilitate the extraction of detailed mechanistic information. We
can now reduce systematically this similarity, provided
strong specific interactions are preserved at interfaces between substrate and additive. This approach provides a
unique entry to the study of solvent effects on crystal morphology. It is of particular relevance for the better understanding of crystallizations taking place at structured biological interfaces, such as bones, teeth and shell^,^^^^^^ and
in the inhibition of ice formation in fish living at temperatures below the freezing point of water.14b1
In summary, we see that the old, but until recently empirical, technique of modifying crystal growth by addition
of impurities has a wide range of consequences in both
theoretical and practical chemistry and materials
science.l4’’ This field of research may be expected to expand rapidly in the future.
We thank the U.S.-Israel Binational Foundation, Jerusalem, the donors of the Petroleum Research Fund administered by the American Chemical Society, the Volkswagen
Stifrung, and the Israel Academy of Science and Humanities
forfinancial support of this work. One of us (L. A . ) is the recipient of the Helena Rubinstein Career Development Chair.
We are indebted to Prof. M. D. Cohen for a critical reading
of the manuscript, and are grateful to Edna Gati and Mar484
ianne Idelson for carrying a heavy load of the experimental
work. We are also grateful to Dr. S . Weinsteinfor his important cooperation in the HPLC analyses, and to Prof. A . van
Hook for bringing to our attention the work carried out on
the change of morphologies of sugars.
Received: August 9, 1984 [A 533 IE]
German version: Angew. Chem. 97 (1985) 476
[I] M. von Laue in “Historical Introduction”, lnrernational T a b l e f o r X-ray
Crysrallography. Vol. 1, The Kynoch Press, Birmingham 1965.
121 L. Pasteur, A n n . Phys. (Parisj 24 (1848) 442.
[3] H. E. Buckley: Crystal Growth. Chapman and Hall, London 1951: P.
Hartman in D. Fox, M. M. Labes, A. Weisberger: Physics and Chemisny
af the Organic Solid Stare. Interscience, New York 1963, Vol. 1, p. 369),
Vol. 2. p. 873; P. Hartman: Crystal Growth, North-Holland Publishing
Co., Amsterdam 1973, p. 367: R. Kern, Bull. Soc. Fr. Mineral. Cristullogr. 76 (1953) 391: A. F. Wells, Philos. M a g . 37(1946) 180, 217, 605.
[4] F. D. Miles, Proc. R . Soc. London Ser. A 132 (1931) 266.
[5] a) R. H. Doremus, R. W. Roberts, D. Turnbull: Growth and Perjection of
Crystals. Wiley, London 1958: b) N. Cabrera, D. A. Vermilyea in [5a], p.
393: c) F. C. Franck in [5a], p. 411; d) G. W. Sears in [Sa], p. 441; e) J .
Dugua, B. Simon, J . C r m t . Growth 44 (1978) 265.
[6] W. J. Dunning, N . Albon in [5a], p. 416; 9. M. Smythe, AUSI.J. Chem. 20
(1967) 1 1 15: Sugar Technol. Rev. I (1971) 191; G. Mantovani, G. Gilli,
F. Fagioli, C. R. X I I I . Assemb/,: C 115, Falsterbo (Sweden) 1967, p. 289:
A. van Hook, Kristal. Akad. Nauk S S S R I n l t . Kristallogr. 8 (1968) 45: J .
Am. Soc. Sugar Beer Technol. 21 (1981) 130.
171 B. S. Green, M. Lahav, D. Rabinovich,Acc. Chem. Res. I2 (1979) 19 I; L.
Addadi, S. Ariel, M. Lahav, L. Leiserowitz, R. Popovitz-Biro, C. P. Tang
in: Chemical Pliysicc ufSa1id.y and their Swfuce.7, Specialist Periodical Reports 8 (1980) 202.
[8] L. Addadi, M. Lahav, Pure Appl. Chem. 51 (1979) 1269; J . Am. Chem.
Soc. 100 (1978) 2838: 101 (1979) 152: L. Addadi, J. van Mil, M. Lahav,
ibid. lU4 (1982) 3422.
[9] J. van Mil, E. Gati, L. Addadi, M. Lahav, J . A m . Chem. Soc. 103 (1981)
1248; J. van Mil, L. Addadi, E. Gati, M. Lahav, ihid. 104 (1982) 3429: L.
Addadi, J. van Mil, E. Gati, M. Lahav, Macromol. Chem. Suppl. 4 (1981)
[lo] L. Addadi, J. van Mil, M. Lahav, J. Am. Chem. Soc. 103 (1981) 1249: L.
Addadi, S. Weinstein, E. Gati, 1. Weissbuch, M. Lahav, ibid. 104 (1982)
46 10.
[I I] J . P. Purvis, US-Pat. 2790001 (1957).
1121 H. L. Fike, US-Pat. 2937200 (1960).
1131 K. Harada, Nature iLondonj 205 (1965) 590; K. Harada, W. Tso, Bull.
Chem. Soc. Jpn. 45 (1971) 2859.
I141 L. Ostromisslensky, Ber. Dtsch. Chem. Ge.7. 4 / (1908) 3035.
1151 D. H. R. Barton, G. W. Kirby, J . Chem. Soc. 1962. 806.
1161 B. S. Green, L. Heller, Science 185 (1974) 527.
[I?] L. Addadi, E. Gati, M. Lahav, J . Am. Chem. Soc. 103 (1981) 1251; L.
Addadi, 2. Berkovitch-Yellin, N. Domb, E. Gati, M. Lahav, L. Leiserowitz, Naiure (London) 296 (1982) 21.
[lS] C. W. Bunn: Chemical Crystallography. Oxford University Press, Oxford
1961; 1. C. Paul, D. Y . Curtin, Chem. Rev. 81 (1981) 525.
[I91 A. Fredga, Tetrahedron 8 (1960) 126, and references cited therein
(201 L. Leiserowitz, A. T. Hagler, Proc. R . Soc. London Ser. A388 (1983) 133;
Z. Berkovich-Yellin, L. Leiserowitz, J . Am. Chem. Sor 102 (1980)
1211 Z. Berkovitch-Yellin, L. Addadi, M. Idelson, M. Lahav, L. Leiserowitz,
Angew. Chem. 94 (1982) 640: A n g e w Chem. Int. Ed. Engl. 21 (1982) 631 :
Angew. Chem. Suppl. 1982, 1336; L. Addadi, 2. Berkovitch-Yellin, 1.
Weissbuch, M. Lahav, L. Leiserowitz, Mol. Cry.sta. Liq. Cryst. 96 (1983) I .
1221 Z. Berkovitch-Yellin, L. Leiserowitz, J . Am. Chem. Soc. 104 (1982) 4052;
Z. Berkovitch-Yellin, S. Ariel, L. Leiserowitz, ibid. 105 (1983) 765.
[23] J. L. Wang, M. Sc. Thesis, Feinberg Graduate School, Rehovot (Israel)
(1983): J . L. Wang, Z. Berkovitch-Yellin, L. Leiserowitz, Acra Crystaltogr., in press.
[24] 2. Berkovitch-Yellin, J. van Mil, L. Addadi, M. Idelson, M. Lahav, L.
Leiserowitz, J . Am. Chem. Soc. 107 (1985) 3 I 1 1
[25] J. D. Dunitz: X-ray Analysis and the Structure of Molecules. Cornell University Press, Ithaca, NY 1979, p. 129-148: S. F. Mason: Molecular @tical Acriuiry and Chiral Discrimination, Cambridge University Press,
Cambridge 1982.
[26] J. Waser, J . Chem. Phys. 17 (1949) 498.
[27] E. E. Turner, K. Lonsdale, J . Chem. Phys. 18 (1950) 156.
[28] Z. Berkovitch-Yellin, L. Addadi, M. Idelson, L. Leiserowitz, M. Ldhav,
Nature lLondonJ 296 (1982) 27.
[29] L. Addadi, 1. Weissbuch, Z . Berkovitch-Yellin, M. Lahav, L. Leiserowitr, S. Weinstein, J . Am. Chem. Soc. 104 (1982) 2075.
Angen’. Chem. l n t . Ed. Engl. 24 il985) 466-485
1301 1. Weissbuch, L. Addadi, 2. Berkovitch-Yellin, E. Gati, S . Weinstein, M.
Lahav, L. Leiserowitz, J . Am. Chem. Soc. 105 (1983) 6615; 1. Weissbuch,
Ph. D. Thesis, to be submitted to the Feinberg Graduate School, Rehovot (Israel).
[31] S. Weinstein, M. H. Engel, P. E. Hare, Anal. Biochem. I 2 1 (1982) 370; S .
Weinstein, Anyen,. Chem. 94 (1982) 221: Angew. Chem. I n t . Ed. Engl. 21
(1982) 218: Angeiv. Chem. Suppl. 1982, 425.
[32] 1. Weissbuch, L. S h ~ m o nL.
, Addadi. Z. Berkovitch-Yellin, M. Lahav, L.
Leiserowitz, I v . J . Chem. Special Isrue. in press.
1331 1. Weissbuch, L. Addadi, Z. Berkovitch-Yellin, E. Gati, M . Lahav, L.
Leiserowitz, Nature /London)310 (1984) 16.
.I341 L. Shimon, M. Lahav, L. Leiserowitz, unpublished; L. Shimon, M. Sc.
Thesis, Feinberg Graduate School, Rehovot (Israel) 1982.
1351 W. G. Johnsteon, Prugr. Cemm. Sci. 2 (1962) 175; J. J. Gilrnan, W. G.
Johnsteon, C . W. Sears, J. Appl. Phys. 29 (1958) 749.
[36] P. Hartman, W. G. Perdok, Acta Cr.vstal1ngr. 8 (1955) 49.
[37] P. Hartman, J. Cryst. G?YJW~/I
49 (1980) 157, 166.
1381 Z. Berkovitch-Yellin, J . Am. Chem. Sor. 107 (1985) 3375.
[39] A. F. Wells, Discuss. Faraday Soc. 5 (1949) 197.
[40] R. .I. Davey, J . Cryrt. Growfh 34 (1976) 109; R. J. Davey in E. Kaldis:
Current Topics in Malerial Science, Yo/. 8 , North Holland Publishing
Co., Amsterdam 1982, Chapter 6.
1411 B. S. Milisavljevic, Dissertation, ETH Zurich 1982 (Dissertation No.
1421 F. Wireko, M. Sc. Thesis, Feinberg Graduate School, Rehovot (Israel);
F. Wireko, Z. Berkovitch-Yellin, F. Frolow, M . Lahav, L. Leiserowitz,
1431 Z. Berkovitch-Yellin, L. Leiserowitz, Acta Crystallogr. 8 4 0 (1984) 159.
[44] S. Weiner, W. Traub, H. A. Lowenstarn in P. Westbroek, E. W. de Johng:
Biomrneralizatron and Biological Metal Accumulation, Reidel, Dordrecht
1983. p. 205.
[45] L. Addadi, S. Weiner, Proc. Nut. Arad. Sci.. i n press.
[46] A. L. d e Vries in R. Gilles: Animals and Enoironmental Fitness, Pergarnon Press, Oxford 1980, p. 583.
[47] W. F. Berg, J . Photogr Sci. 31 (1983) 62.
Great Fear of Small Amounts of ElementsThe Significance of Analytical Chemistry in our Modern
Industrialized Community as Exemplified by Trace Element Analysis
By Giinther Tolg" and Rainer P. H. Garten
Dedicated to Professor Rudolf Bock on the occasion of his 70th birthday
Analytical chemistry is consolidating an important position within the framework of our
modern industrial community; the frontiers of trace (and ultra-trace) analysis have expanded into new territories, thus demanding a constant change in our mode of thinking in a
substance-related manner in analytical chemistry. An outline of the development of analytical chemistry during this century reveals a period of underdeveloped research and education followed by a current phase of impetuous advancement. However, as a result of increasingly antagonistic sectional convictions in the public mind concerning reservations
against, as well as efforts towards, efficient technological progress, this advancement evokes
new existential risks for analytical chemistry-viz. either to be used in an uncritical way or
to fall into discredit following slogans like 'high-performance analytical chemistry is to
blame for it all!' A much more constructive consideration says that risks can be estimated
and evaluated solely by means of a highly efficient analytical chemistry, when used with a
sense of responsibility. Analysts may help to clarify and to cope with the increasing fear of
decreasingly smaller amounts of trace elements-in both adverse groups in our community.
Strategies necessary to gain this end are outlined with regard to a methodological as well as
a political platform.
1. The Role of Analytical Chemistry in our Modern
Industrial Society
The chemical, biological, and physical properties of materials and-what is of interest to us in connection with
our environment-of complex systems of substances are
[+I, Dr. R. P. H. Garten
Laboratorium fur Reinststoffanalytik des Max-Planck-Instituts fur
Metaliforschung Stuttgart
Bunsen-Kirchhoff-Strasse 13, D-4600 Dortmund 1 (FRG)
[*] Prof. Dr. G. Tdlg
Other address:
lnstitut fur Spektrochemie und angewandte Spektroskopie
Runsen-Kirchhoff-Strasse 1 I , D-4600 Dortmund 1 (FRG)
Angew Chem. In/. Ed. Enyl. 24 (1985) 485-4414
often dependent upon very small alterations in composition. This dependence forms the basis of many fields of research, e.g. in the life sciences, in the earth sciences, and in
materials science; but, as we are learning more and more
each day, it also has direct serious consequences for man
and his environment in terms of health, working conditions, and other factors determining the quality of his life.
Thus, this research and public policy are entering progressively into closer relationships with which not only every
responsible, forward-looking scientist and politician, but
also every individual citizen must come to an understanding.
In an effort to comprehend these very complicated interrelationships, it is very easy to come to false conclusions
0 V C H Vrrlagsgesell.rchaJt mbH. 0-6940 Weinheini. 1985
0570-0833/85/0606-0485 $ 02.50/0
Без категории
Размер файла
4 326 Кб
crystals, inhibitorsчimplications, уtailor, growth, organiz, made, material, dissolution, science, stereochemistry
Пожаловаться на содержимое документа