4-Benzoylamino-3-hydroxybutyric acid, historically first "anomalous racemate": Reinvestigation

Article abstract of DOI:10.1021/cg5017615

Chiral 4-benzoylamino-3-hydroxybutyric acid (1) was recognized in 1930 as the first example of "anomalous racemates" (correct to say, anomalous conglomerates), that is, specific addition compounds formed by different enantiomers in unequal ratio. Through the comparative (racemic against homochiral samples) inspection of the IR spectra, single crystal X-ray diffraction, PXRD analysis, and solubility data we have found that this substance forms normal racemic compound in the solid state, and must be excluded from the very short list of anomalous conglomerates. At the same time homo-1 is dissolved in 25 times better than rac-1, and this feature belongs to another interesting and rare type, namely, "anticonglomerates". Some of the reasons for this behavior are discussed.

Full text of DOI:10.1021/cg5017615

Subscriber access provided by NEW YORK UNIV
Article
4-Benzoylamino-3-hydroxybutyric acid – historically
first “anomalous racemate”: reinvestigation
Alexander A Bredikhin, Zemfira A Bredikhina, Dmitry V
Zakharychev, Aida I Samigullina, and Aidar T. Gubaidullin
Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/cg5017615 • Publication Date (Web): 12 Feb 2015
Downloaded from http://pubs.acs.org on February 18, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Crystal Growth & Design is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
Cover Page
4-Benzoylamino-3-hydroxybutyric acid – historically first “anomalous
racemate”: reinvestigation
Alexander A. Bredikhin*, Zemfira A. Bredikhina, Dmitry V. Zakharychev, Aida I. Samigullina,
and Aidar T. Gubaidullin
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A.E. Arbuzov Institute of Organic and Physical Chemistry of Kazan Scientific Center
of Russian Academy of Sciences, Arbuzov St., 8, Kazan 420088, Russian Federation
The chiral 4ꢀbenzoylaminoꢀ3ꢀhydroxybutyric acid (1) was recognized in 1930 as the first example
of “anomalous racemates” (correct to say – anomalous conglomerates), i.e. specific addition
compounds formed by different enantiomers in unequal ratio. Through the comparative (racemic
against homochiral samples) investigation using IR spectrometry, single crystal Xꢀray diffraction,
PXRD analysis, and solubility studies, we have found that this substance forms a normal racemic
compound in the solid state, and thus must be excluded from the very short list of anomalous
conglomerates. At the same time homoꢀ1 is dissolved in 25 times better than racꢀ1, and this feature
belongs to another interesting and rare type, namely “anticonglomerates”. We discuss some of the
reasons for this behavior.
*
Corresponding author: Phone/fax: +7 843 2727393/ +7 843 2731872. Eꢀmail: baa@iopc.ru.
ACS Paragon Plus Environment

Crystal Growth & Design
Page 2 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
ꢀBenzoylaminoꢀ3ꢀhydroxybutyric acid, historically
first “anomalous racemate”: reinvestigation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Alexander A. Bredikhin,* Zemfira A. Bredikhina, Dmitry V. Zakharychev, Aida I. Samigullina,
Aidar T. Gubaidullin
A.E. Arbuzov Institute of Organic and Physical Chemistry of Kazan Scientific Center of Russian
Academy of Sciences, Arbuzov St., 8, Kazan 420088, Russian Federation
ABSTRACT
Chiral 4ꢀbenzoylaminoꢀ3ꢀhydroxybutyric acid (1) was recognized in 1930 as the first example of
“
anomalous racemates” (correct to say – anomalous conglomerates), i.e. specific addition
compounds formed by different enantiomers in unequal ratio. Through the comparative (racemic
against homochiral samples) inspection of the IR spectra, single crystal Xꢀray diffraction, PXRD
analysis, and solubility data we have found that this substance forms normal racemic compound in
the solid state, and must be excluded from the very short list of anomalous conglomerates. At the
same time homoꢀ1 is dissolved in 25 times better than racꢀ1, and this feature belongs to another
interesting and rare type, namely “anticonglomerates”. Some of the reasons for this behavior are
discussed.
INTRODUCTION
Chiral 4ꢀbenzoylaminoꢀ3ꢀhydroxybutyric acid (1) is structurally resemblant to the important
neurotransmitter γꢀaminoꢀnꢀbutyric acid (GABA) and may serve as a precursor in the synthesis of
carnitine, an essential cofactor of fatty acid metabolism (Scheme 1). In this latter capacity the amino
acid 1 attracts the attention of chemists for over a century.
ACS Paragon Plus Environment

Page 3 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
Scheme 1. Benzoylaminohydroxybutyric acid 1 and related biologically important compounds
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In 1930, when attempting to obtain enantiopure acid 1 by crystallisation of the samples of
intermediate enantiomeric composition, Bergman and Lissitzin found that the enantiomer ratio in
1
the thus formed crystalline precipitate is close to an integer ratio of 2:1. Based on this fact, the
authors concluded that the molecular addition compound of suitable composition was formed in the
solid phase during crystallization. To emphasize the difference from typical for chiral substances
"racemic compound", i.e., addition compound with a ratio of enantiomers of 1:1, for the so detected
1
new crystal type Bergman and Lissitzin coined the term "anomalous racemate". Under this name
the phenomenon of crystallization of chiral substances, which form a regular (not caused by
disorder of the crystal) addition compounds with unequal ratios of enantiomers, have been
2
discussed in the wellꢀknown monograph by Jacques et al.
2
Amino acid 1 opens the list, given in the monograph, which consists of only 9 compounds.
Subsequently, on the one hand, a number of novel compounds have been attributed to this class. On
the other hand, new studies were published in which the phase behavior of compounds, previously
ranked as "anomalous racemates", received other interpretation. Analysis of the current state of the
3
problem is presented in our review. Virtually, if one considers only configurationally stable chiral
compounds, the existence in the solid phase of the addition compounds with unequal ratios of
4ꢀ6
enantiomers is reliably proven only for compounds 2, 3 and 4 (Scheme 2).
Scheme 2. Proven examples of “anomalous racemates”
ACS Paragon Plus Environment

Crystal Growth & Design
Page 4 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Before continuing the discussion, it seems appropriate to make a point about the terminology
used. The existence in the unit cell of the crystal of unequal amounts of enantiomers excludes the
presence of the symmetry elements of the second kind. Consequently, such compounds should
crystallize into one of the 65 Sohncke space groups. This condition inevitably leads to formation of
two enantiomorphic crystal modifications during the equilibrium crystallization of the chiral
7
ꢀ9
compound from its racemate. A mixture of two nonꢀidentical crystalline phases should be
considered as a conglomerate.
We offered to classify the all possible for a chiral substance conglomerates according the
9
enantiomeric excess of the asymmetric unit of the unit cell, eeas.u.. Following this classification,
the name of normal or common conglomerates refers to those ones, in which each single crystal is
formed by a single enantiomer; of course, eeas.u = 1 for these cases. Such conglomerates are most
known; they are formed during crystallization of racemic chiral substances prone to spontaneous
resolution. For the second group of conglomerates eeas.u = 0. This means that the asymmetric unit is
composed of equal amounts of opposite enantiomers, and the single crystal in whole has a racemic
composition. In this case, the corresponding sites of the enantiomorphous crystal are occupied with
the opposite enantiomers. For the first time this type of conglomerates was noticed by Bishop and
10
Scudder, they proposed the name of false conglomerates for them. In more detail this
ACS Paragon Plus Environment

Page 5 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
phenomenon (under the name of kryptoracemates) was described by Fábián and Brock;
kryptoracemates were found to account for 0.1% of all organic structures containing either a
11
racemic compound, a meso molecule, or some other achiral molecule.
The third and the last group of conglomerates comprise those, in which the asymmetric unit is
formed by molecules of various enantiomers, but in unequal amounts. Thus the enantiomeric purity
of the asymmetric unit may be in principle any rational number between 1> eeas.u > 0. For obvious
reasons, Z' ≥ 3 for this class of crystals. We proposed to call this class of addition compounds with
unequal ratio of enantiomers as anomalous conglomerates, instead of an inadequate name of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
anomalous racemates.
Apparently anomalous conglomerates are the most rare crystal types for a chiral substance, and
are represented in nature by exceptional examples. Therefore, each member of this class deserves
close investigation. Meanwhile, the compound that opens the list, i.e. 4ꢀbenzoylaminoꢀ3ꢀhydroxyꢀ
butyric acid 1, remains in fact unexplored nor in terms of crystallography, nor in terms of phase
behavior. In this paper we have attempted to fill this gap, at least partially.
EXPERIMENTAL SECTION
Instrumentation. The NMR spectra were recorded on a Bruker Avanceꢀ400 spectrometer (399.9
1
13
6 3
МHz for H and 100.5 MHz for C) in DMSOꢀD or CDCl with TMS or the signals of the solvent
as the internal standard. The IR spectra of the polycrystalline samples of racꢀ and scalꢀcompounds
in KBr pellets were recorded on a Bruker Tensor 27 spectrometer. Optical rotations were measured
on a PerkinꢀElmer model 341 polarimeter (concentration c is given as g/100 mL). Melting points for
general purposes were determined using a Boëtius apparatus and are uncorrected. HPLC analyses
were performed on a Shimadzu LCꢀ20AD system controller, using UV detector.
Substances. Racemic and enantiopure 4ꢀbenzoylaminoꢀ3ꢀhydroxybutyric acid, racꢀ1 and (R)ꢀ1,
were prepared from racꢀ and (S)ꢀepichlorohydrin as shown in Scheme 3. A detailed description of
the experiment and the physicochemical characteristics of the intermediates are given in the
Electronic Supplement.
ACS Paragon Plus Environment

Crystal Growth & Design
Page 6 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
o
20
(
R)-4-Benzoylamino-3-hydroxybutyric acid, (R)-1. Mp 114–116 C; [α]
D
= ꢀ11.1 (c 1.7, H
2
O).
12
o
20
20
D
1
{
lit. : mp 112ꢀ113 C, [α]
D
= ꢀ11.2 (c 1.4, H
2
O)}; [α]
= ꢀ22.5 (c 1.0, 0.5 N aq. NaOH). {lit. :
2
0
[
α]D = ꢀ22 (in 0.5 N aq. NaOH)}; 98 % ee [chiral HPLC analysis of the methyl ether of (R)ꢀ1,
which was obtained by the reaction between (R)ꢀ1 and CH ; Daicel Chiralcel OJ (0.46 x 25 cm)
2
N
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
column; eluent: hexane/isopropanol = 9:1; flow rate: 1.0 mL/min; UV detector 254 nm (t = 17.4
R
1
min (major), t
R
= 23.4 min (minor)]. H NMR (DMSOꢀD
6
):
δ
= 2.24 (dd, J = 8.6, 15.3 Hz; 1H,
N), 4.07 (m, 1Н, CH),
4.98 (br.s, 1Н, OH), 7.44ꢀ7.54 (m, 3H, Ph), 7.85ꢀ7.87 (m, 2H, Ph), 8.42 (t, J = 5.7 Hz; 1H, NH),
CHHC=O), 2.45 (dd, J = 4.2, 15.3 Hz; 1H, CHHC=O), 3.30 (m, 2Н, CH
2
1
3
1
6
2.04 (br.s., 1H, COOH). С NMR (DMSOꢀD ): δ = 41.6 (Cꢀ2), 46.4 (Сꢀ4), 67.4 (Cꢀ3), 128.1
2,6
3,5
4
1
(C
Ar), 129.1 (C Ar), 132.0 (C Ar), 135.4 (C Ar), 167.4 (C=O), 173.7 (Cꢀ1).
o
13
o
rac-4-Benzoylamino-3-hydroxybutyric acid, rac-1. Mp 172–173 C. Lit. : mp 176–177 C.
NMR spectra were identical with those cited above for (R)ꢀ1.
Solubility investigations. Filters Millipore 0.45 ꢁm PTFE hydrophilic were used for the samples
filtering. Micro syringes Hamilton (precision within ±1%) were used for analytical sampling and for
adding fixed volumes of solvent. The mass of the samples was controlled with Sartorius CPA2P
balance (accuracy ±1 ꢁg).
The solubility of racemic and enantiopure samples was determined by chromatographic control of
the concentration of saturated solution, equilibrious with the corresponding crystalline phase. For
this purpose to pure acetonitrile (1mL) the corresponding solid phase was sequentially added by
weighted portions with stirring until a stable turbidity of the liquid and the sediment at the bottom
were observed. The total amount of the added substance was not less than 2 mg for racemate and 38
mg for enantiopure 1. The vial was hermetically sealed by stopper and the slurry was stirred at a
constant temperature (20 °C) during the day; the continual presence of the large excess of solid
phase was monitored visually. After stopping the stirring and sedimentation, the liquid phase was
sampled by syringe and filtered. An aliquot of the filtered solution (0.5 mL) was transferred to a test
tube, and was diluted with methanol (0.5 mL). A solution of diazomethane in ether was added to the
ACS Paragon Plus Environment

Page 7 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
aliquot until a stable light yellow coloration of the mixture. This is done to convert the free acid to
the methyl ester, which is more convenient for subsequent chromatographic analysis. After 10
minutes the sample was evaporated to dryness in vacuo, the residue was dissolved in a fixed volume
(1 mL) of isopropanol followed by a chromatographic determination of the substance concentration
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
in the thus obtained sample. For the preꢀcalibration a series of accurately weighted samples of acid
1 (0.5ꢀ5 mg) was used, each of which, after dissolution in methanol, was subjected to a similar
treatment (i.e. reaction with diazomethane, drying, dissolution in the precise volume of
isopropanol).
Experiments were repeated twice; 2ꢀ3 chromatographic measurements were performed for each
ꢀ1
ꢀ3
ꢀ1
system. The absolute values of the solubility totaled 1.00(0.03) g⋅L (or 4.49(0.15) ·10 mole⋅L )
ꢀ1
ꢀ1
for racꢀ1 and 25.4(1.1) g⋅L (or 0.114(0.005) mole⋅L ) for (R)ꢀ1.
To study the equilibrium solubility of the points of intermediate enantiomeric composition (0 < ee
<
1) the series of (R)ꢀ1 solutions in acetonitrile was prepared in first. The concentrations of
enantiopure component within the series were a fortiori lower than its maximum solubility in
ꢀ1
acetonitrile, and ranged from 0.13 to 4 mg⋅mL . The solutions were prepared by serial dilution in
half of aliquots of peak solution. In what follows, a fixed volume (1.5 mL) of one of thus prepared
(R)ꢀ1 solution was added to the weighed portion of racꢀ1 (2 mg), the vials were sealed, and the
mixture was stirred overnight at 20 °C until equilibrium was reached. Visual observation indicated
the constant presence of a large excess of the solid phase in all samples. Selected values of the
racemate mass, as well as the concentration and solution volume of the enantiopure substance,
makes it possible to crystallize the solid phase in a wide range of enantiomeric compositions,
including a hypothetical "anomalous conglomerate" 2:1.
After reaching equilibrium a liquid phase aliquot of each sample was filtered, treated with
diazomethane and subjected to chromatographic analysis as described above. In addition to the total
concentration of acid 1, the equilibrium concentrations of individual stereoisomers in the liquid
phase were determined in each of the experiments. To control the accuracy of the dilution
ACS Paragon Plus Environment

Crystal Growth & Design
Page 8 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
procedure, the preselected aliquots of stock solutions of all concentrations of enantiopure 1 were
similarly analyzed.
The amount and composition of the equilibrium solid phase for each sample was calculated by
difference in mass of the individual enantiomers in the racemate weighed portion and in the source
and equilibrium solutions. It should be noted that for the determination of the composition of the
solid phase, this technique proved to be more reproducible and accurate than the direct analysis of
solids, as the latter inevitably entrains with itself a noticeable amount of solution, which
significantly differs from solid by composition.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
X-ray analysis. Both enantiopure and racemic samples of 4ꢀbenzoylaminoꢀ3ꢀhydroxybutyric acid
1
were studied by single crystal Xꢀray analysis. The crystals of (R)ꢀ1 for this purpose were grown
by slow evaporation of the corresponding saturated solution in ethyl acetate; racꢀ1 crystals were
grown by two ways. Crystals having the form of flat prisms (see Figure 1a), were prepared by slow
crystallization from concentrated waterꢀfree methanol solution on the interphase boundary with
diethyl ether, taken in 5ꢀfold excess. In the sequel, such crystals will be denoted as racꢀ1.1.
The second type of crystals, long needles (Figure 1b), were prepared by slow evaporation of
concentrated solutions of the racemate in water. These crystals are referred to as racꢀ1.2.
The Xꢀray diffraction data of the investigated crystals were collected on a Bruker AXS Smart
Apex II CCD ((R)ꢀ1, racꢀ1.1) and Bruker AXS Kappa Apex II CCD (racꢀ1.2) diffractometers in the
ωꢀ and θꢀscan modes using graphite monochromated MoK (λ 0.71073 Å) radiation. The crystal
α
data, data collection, and the refinement parameters are given in Table 1. Data were corrected for
14
the absorption effect using SADABS program. The structures were solved by direct method and
15
16
refined by the full matrix leastꢀsquares using SHELXTL and WinGX programs. All nonꢀ
hydrogen atoms were refined anisotropically. All hydrogen atoms were inserted at calculated
positions and refined as riding atoms except the hydrogen atoms of NH and OH groups which were
located from difference maps and refined isotropically.
ACS Paragon Plus Environment

Page 9 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
All three crystals (racꢀ1.1. and racꢀ1.2 samples were investigated at different temperatures) have
very similar unit cell parameters, do not contain solvent molecules, and were solved in related
spatial groups. Both racemic samples were refined in the centrosymmetric group C2/c. The
structure of the enantiopure sample was solved and refined in Sohncke space groups C2. For
reasons which we’ll discuss later (see Results and Discussion) the attempts were made to solve the
structures in the primitive triclinic cells P1 and Pꢀ1, which, however, did not lead to an
improvement in the convergence parameters. Conversely, the model leads to disorder fragments in
which the carboxyl groups are deployed at 180 degrees. In this regard, we believe that the option of
solving and refinement the structures in baseꢀcentered monoclinic cells is more correct, as
evidenced by the very good convergence parameters (see Table 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
7
Data collections: images were indexed, integrated, and scaled using the APEX2 data reduction
18
package. All Figures were made using Mercury program. Molecular structures and conformations
19
were analyzed by PLATON. Crystallographic data (excluding structure factors) for the structures
of (R)ꢀ1, racꢀ1.1 and racꢀ1.2 reported in this paper has been deposited with the Cambridge
Crystallographic Data Centre as supplementary publication numbers CCDC 1036966ꢀ1036968
respectively. Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union
Road, Cambridge CB2 1EZ, UK, (fax: +44ꢀ(0)1223ꢀ336033 or eꢀmail: deposit@ccdc.cam.ac.uk).
The Xꢀray phase studies for the compound 1 polycrystalline samples were performed on a Bruker
D8 Advance diffractometer equipped with Vario attachment and Vantec linear PSD, using Cu
radiation (40 kV, 40 mA) monochromated by the curved Johansson monochromator (λ Cu Kα1
1.5406 Å). Roomꢀtemperature data were collected in the reflection mode with a flatꢀplate sample.
The samples were loaded on a silicon plate reducing background scattering. Patterns were recorded
o
o
o
in the 2Θ range between 3 and 60 , in 0.008 steps, with a step time of 0.1 ꢀ 0.5 s.
ACS Paragon Plus Environment

Crystal Growth & Design
Page 10 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. Crystallographic Data for Crystalline Compounds rac-1 and (R)-1
Compound, sample
Formula
racꢀ1.1
racꢀ1.2
(R)ꢀ1
C
11
H
13NO
4
C
11
H
13NO
4
11 4
C H13NO
M (g/mol)
223.22
151(2)
223.22
296(2)
223.22
150(2)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Temperature, K
Crystal class
Space group
Crystal size
Z, Z´
Monoclinic
Monoclinic
Monoclinic
C2/c
C2/c
C2
3
3
3
0.22×0.20×0.10 mm
8, 1
0.38×0.18×0.14 mm
8, 1
0.58×0.53×0.04 mm
8, 2
Cell parameters
a = 24.794(5) Å,
b = 11.355(2) Å,
c = 7.7615(14) Å
a = 24.64(1)
b = 11.524(5)
c = 7.871(4)
a = 24.355(6) Å,
b = 11.626(3) Å,
c = 7.783(2) Å
β
= 97.727(2)°
β
= 98.24(1)
β = 98.371(3)°
3
V, Å
2165.4(7)
2212(2)
944
2180(1)
944
F(000)
944
3
ρ
calc, g/cm
1.369
1.05
1.340
1.03
1.360
1.04
ꢀ
1
ꢀ
, cm
θ
, deg
3.07≤
θ
≤ 30.42
4.26≤
9457
2583
θ
≤ 27.93
1.94≤
10932
5006
θ ≤ 29.38
Reflections measured
Independent reflections
10479
2752
[R(int) = 0.0226]
[R(int) = 0.0199]
[
R(int) = 0.0725]
Number of parameters
156/1
151/0
314 / 1
/
restraints
Reflections [I > 2
σ(I)]
2335
1405
4688
R
1
1
/ wR
2
2
[I > 2
σ
(I)]
0.0517/ 0.1305
0.0626/ 0.1506
0.1225/ 0.1830
0.979
0.0332/ 0.0778
0.0367 / 0.0802
1.032
R
/ wR
(all reflections)
0
.0604 / 0.1374
.029
2
Goodnessꢀofꢀfit on F
1
ꢀ3
ρ
max
/
ρ
min (eǺ )
0.542 / ꢀ0.432
0.309 / ꢀ0.277
0.352 / ꢀ0.276
ACS Paragon Plus Environment

Page 11 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
Synthesis. Preparation of the necessary for our research enantiopure and racemic samples of 4ꢀ
benzylaminoꢀ3ꢀhydroxybutyric acid 1 is shown in Scheme 3.
Scheme 3. Preparation of racemic and enantiopure samples 1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reagents and conditions: (i) (S)ꢀepichlorohydrin (1 eq, 96 % ee), benzaldehyde (1 eq), aqueous NH
3
o
(
26 wt%, 1.7 eq), ethanol; (ii) PhC(O)Cl (1 eq), Py (1 eq), CHCl
3
, ꢀ40 C; (iii) NaCN (1.6 eq), KI
2
o
(0.019 eq), aqueous ethanol (67%), >99% ee; (iv) 3N NaOH, H O , 63ꢀ65 C, 98% ee.
2
As a source of chirality for (R)ꢀ1 an enantiopure epichlorohydrin (S)ꢀ5 was used. Condensation
of the latter with benzaldehyde and ammonia and subsequent treatment of the condensation
products with benzoyl chloride in dry pyridine yielded the chloromethyl derivative (S)ꢀ6, which in
turn was converted to enantiopure nitrile (R)ꢀ7. This synthetic route is based on the modified
20
literature procedures for the synthesis of the racemic products reported by Mazetti and Lemmon.
However, we found that the original literature procedure utilizing hydrolysis of boiling nitrile 7
with aqueous HBr leads to a complete hydrolysis of the product to 3ꢀaminoꢀ2ꢀhydroxybutyric acid.
Therefore, we developed here the optimized procedure using a soft alkaline hydrolysis in the
2
1
presence of hydrogen peroxide, as proposed in reference for labile nitriles. The same sequence of
synthetic transformations of racemic epichlorohydrin racꢀ5 was used in this work to obtain a sample
of racꢀ1.
ACS Paragon Plus Environment

Crystal Growth & Design
Page 12 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Crystal habit and IR spectra comparison for (R)-1 and rac-1 crystals. Amino acid 1 is
composed of hydrophobic (phenyl ring) and hydrophilic (carboxy, hydroxy and amino) fragments.
Therefore, it is soluble in organic solvents and in water. Figure 1 shows micrographs of crystals of
racemic and enantiopure amino acids 1 grown by us from an organic solvent (see Experimental
Section) and water.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Microscopic pictures of the racemic and homochiral samples of compound 1. (a) Thin
prisms of racꢀ1.1 from organic solvent; (b) needleꢀlike crystals of racꢀ1.2 from water; plates of (R)ꢀ
both from organic medium (c) and water (d).
1
1
The paper states that the homochiral sample is crystallized from water as a solvate with one
molecule of water. According to the authors, this crystal modification has mp = 78–81 °C. At the
same time the racemate crystals do not form hydrates. Figure 1 show that the essential difference
between the appearance of the crystals, prepared from organic and aqueous medium, is
ACS Paragon Plus Environment

Page 13 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
characteristic only for the racemate. The crystals grown at the interface of methanol and ether (see
Experimental Section) comprise flat transparent good faceted prisms (Figure 1a), whereas the ones,
precipitated from water, have elongated needleꢀlike form (Figure 1b). Crystals of the acid (R)ꢀ1
grown from ethyl acetate constitute poorly faceted plates (Figure 1c). The enantiopure sample
crystallizes from water as thin trapezoids (Figure 1d). We note here that the crystals obtained from
different medium have (in pairs) similar melting points, namely 114–116 °C for (R)ꢀ1 and 172–173
°C for racꢀ1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
When the crystal forming molecules are identical (up to mirroring), the simplest procedure to
detect similarities and differences of the crystal structure is a comparison of their vibrational
spectra. For the purpose of establishing of the reliable criteria of the similarities and differences of
complex spectral curves, we have previously proposed to use the standard Pearson correlation
coefficient r between two digital arrays (A
i
,ν
i
), where ν
i
stands for the vibration frequency (usually
corresponds to the extinction at this
ꢀ1
expressed in wave numbers in increments of 1 cm ), and A
i
2
2
wave number. To allow for a more detailed visual comparison of the vibration spectra, we
proposed to use a graphical representation of the correlation between the two spectra, that is,
1
i
2
visually display them in the coordinates A versus A . In the case of perfect coincidence (identity)
i
of the spectra, the correlation trajectory degenerates into the first diagonal. Partial mismatch
between the spectra in frequencies, intensity, baseline drift, the presence of impurities in the sample
and/or matrix, or the like, results in a deviation from the diagonal trajectory, and is described in a
very characteristic way for each type of distinctions. The details of this approach have been
2
3
described in our recent paper.
Comparison of the IR spectra for compound 1 crystalline samples is shown in Figure 2. As can
be seen from Figures 2a and 2b, the spectra of different (R)ꢀ1 crystalline samples perfectly
correlated (r = 0.998), correlation trajectory is compact and practically coincides with the main
diagonal. To grow the crystals depicted in Figure 1c, the anhydrous solvent was used, and the
melting point of both types of (R)ꢀ1 crystals coincides with the literature data for the anhydrous
ACS Paragon Plus Environment

Crystal Growth & Design
Page 14 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
samples. This means that the Bergman’s and Lissitzin’s information that the enantiopure acid
1
precipitates from water in the form of the monohydrate (R)ꢀ1
⋅
H
2
O, is most likely erroneous, and
enantiopure samples of 1 grown from the organic and aqueous medium are identical. For this
reason, only one crystal grown from the organic solvent was examined by a single crystal Xꢀray
analysis (see below).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Correlation trajectories (a,c,e) for IR spectra in KBr pellets (b,d,f) of the pair of crystals:
(R)ꢀ1 from water (red curve) against (R)ꢀ1 from ethyl acetate (blue curve) (a,b); racꢀ1 from water
(red curve) against racꢀ1 from methanol/ether (blue curve) (c,d); racꢀ1 from methanol/ether (red
curve) against (R)ꢀ1 from ethyl acetate (blue curve) (e,f).
The vibrational spectra of the racemic crystals also correlate quite well with each other (Figure
2c,d), however the correlation coefficient, r = 0.991, decreases somewhat. Nevertheless, the
ACS Paragon Plus Environment

Page 15 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
correlation trajectory is not as compact as for a pair of (R)ꢀ1 crystals. The most distinct differences
ꢀ1
are manifested in the region of νOH ~ 3450 cm . It can be assumed that these differences are not
associated with the internal crystal structure, but with the crystal habit. However, to avoid
ambiguity, both types of crystals have been studied by Xꢀray diffraction; the samples were tested at
various temperatures during the analysis (Table. 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As seen from Figure 2f, there are significant differences between the spectra of racemic (red
curve) and enantiopure (blue curve) samples of 1 (for comparison were used samples grown from
R
i
A
the organic medium); and the correlation coefficient r = 0.984 between the arrays A and A (here
i
and hereinafter R and A upper indices stand for the racemic and enantiopure samples) is relatively
poor. The corresponding correlation trajectory for the two spectra is shown in Figure 2e. In general,
the “fuzzy” image presented in Figure 2e is typical for poorꢀcorrelated spectra.
The discrepancy between the IR spectra of crystals of racemic and homochiral samples allows
to reject the possibility of 1 to crystallize in the form of normal conglomerate. But the rest of the
crystallographic possibilities at this stage remain open. Further investigation of the phase behavior
of amino acid 1 was performed by means of single crystal Xꢀray investigations.
Single crystal X-ray investigations. Before proceeding to describe the crystal structure of the
investigated samples, we note one important fact. As mentioned in the experimental part, the
hydrogen atoms of NH and OH groups were located from difference maps and refined isotropically
with sufficiently good convergence parameters (Table 1). Within the all structures the carboxyl
moieties of adjacent molecules form intermolecular hydrogen bonds, leading to dimeric structures,
which are typical of carboxylic acids in crystals. In principle, such a Hꢀdimer may be (a) located in
the general position; (b) formed around the twofold rotation axis, orthogonal to the Hꢀbonds plane;
(c) formed around a center of symmetry lying in the hydrogen bonds plane; (d) located on the
twofold rotation axis, lying inꢀplane of Hꢀbonds. Typically, variants (b) and/or (c) are realized in the
crystals of carboxylic acids, which results in the wellꢀknown cyclic supramolecular synthon denoted
as a “fork” Hꢀdimer (or "antiꢀdimer" on the Scheme 4). In our case only the option (d) does not lead
ACS Paragon Plus Environment

Crystal Growth & Design
Page 16 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to conflicts with the existing C2 or C2/c lattice symmetry elements. For this reason, for all the
studied samples the primary structure solution (in which all atoms are in general position) led to the
dimers, indicated as "synꢀdimers" in Scheme 4.
Scheme 4. “Antiꢀ” and “synꢀdimers” of carbonic acids
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Our turning to Cambridge Crystallographic Database showed that "synꢀdimers" of carboxylic
acids with noncentroꢀsymmetrical arrangement of the hydroxyl groups are fairly common, though,
of course, do not dominate. Our selective analysis of both "right" and "wrong" structures showed
that the problem of the correct determination of the position of the hydrogen atoms of the hydroxyl
group in the carboxyl moiety of carboxylic acid dimers is a serious problem, which requires a
special consideration. However, it is beyond the scope of the problems addressed in this paper. We
hope that soon we will present the results of our analysis of this phenomenon in a separate
publication.
In this work, we prefer to deal with a more conventional "antiꢀdimer" ("fork” Hꢀdimer). For this
purpose, the hydrogen atoms of hydroxyl groups in the carboxyl fragments during the refinement
were fixed in special position on the twofold axis, and the distances OꢀH and O⋅⋅⋅H in "antiꢀdimers"
are aligned in pairs, i.e. receive the same value in each triad O⋅⋅⋅H⋅⋅⋅O but different for those
fragments which are not connected by symmetry. The characteristics obtained for this model of
refinement are discussed below. Note that, except for the carboxyl moieties themselves, they remain
substantially constant for the other mentioned refinement modes.
Racemic sample rac-1.1, grown in an organic medium (Figure 1a) crystallizes with one molecule
in the asymmetric unit of a baseꢀcentered monoclinic cell. The geometry of the molecule and partial
numbering scheme (common to all subsequent molecules studied in this work) are shown in Figure
ACS Paragon Plus Environment

Page 17 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
3
. The bond lengths and angles are within the range of expected average values for organic
compounds. The most important result in the context of our study is that the sample crystallizes in
the "nonꢀSohncke" space group C2/c. Consequently, racꢀ1.1 crystals present a normal racemic
compound which contains an equal amount of both enantiomers in the unit cell.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. ORTEP plot of the molecule in the crystals of rac-1.1 and the partial numbering scheme
adopted for the molecules 1 in the present work. Displacement ellipsoids are drawn at the 50%
probability level, hydrogen atoms are represented as fixedꢀsize spheres.
The crystal structure of the sample rac-1.2 (Figure 1b) is very close to those for the sample rac-
1.1 (Table 1), although they were obtained at significantly different temperatures. This confirms the
above conclusion about the similarity of the crystal structure of racꢀ1 samples grown in aqueous
and organic medium. Simultaneously, the data obtained allow making some assumptions about such
a pronounced environment effect on the habit of racemic crystals.
Figure 4 shows the Miller indices for some crystal faces. As follows from the figure, for all the
studied crystals the faces of the smallest growth are the base planes [100] and [ꢀ100], thus a crystal
growth is carried out mainly in the tangential directions. Although the quality of the faceting does
not always allow to accurately determine the preferential direction among them. The significant
difference of crystals of the racemic sample, prepared from an aqueous solution, is to block the
ACS Paragon Plus Environment

Crystal Growth & Design
Page 18 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
growth of faces belonging to the second basic direction 010, so that the main growth occurs along
the directions close to the third base vector 001. It is of interest to note, that in racꢀ1 crystals the
centrosymmetric intermolecular hydrogen bonds N–H⋅⋅⋅OH, through which oneꢀdimensional
supramolecular associates (columns) are linked into 3D supramolecular packing (see below), are
oriented approximately in the same 010 direction. Based on these considerations, the competitive
participation of water molecules in the formation of hydrogen bonds with the amino groups likely
disables the growth of one of the faces of the crystals. Of course, due to recombination processes,
the growth of the corresponding faces does not stop completely, but noticeably slows down, which
in the long run may cause the formation of needleꢀshaped crystals.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Miller indices of crystal faces of the racꢀ1.1 (a), racꢀ1.2 (b), and (R)ꢀ1 (c) crystals.
ACS Paragon Plus Environment

Page 19 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
Supramolecular structure in the rac-1.1 crystal (the same as for crystals of rac-1.2) is determined
by the combined action of the classical intermolecular hydrogen bonds O–H⋅⋅⋅O and N–H⋅⋅⋅
O
(
Table 2). Crystals also include weaker C–H ⋅⋅⋅O interaction acting primarily in the same directions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Parameters of the classical intermolecular H-bonds for rac-1.1 crystals
о
D–H⋅⋅⋅
A
D–H, Å
1.421(2) 1.421(2)
H
⋅⋅⋅A, Å
D
⋅⋅⋅A, Å
∠
DHA,
Symmetry operation
2ꢀx,y,1/2ꢀz (2ꢀfold axis)
2ꢀx,y,1/2ꢀz (2ꢀfold axis)
x,ꢀy,1/2+z
O2ꢀH2⋅⋅⋅O2”
O1ꢀH2’⋅⋅⋅O1”
O3ꢀH3⋅⋅⋅O6”’
N5ꢀH5⋅⋅⋅O3””
2.841(4)
2.597(2)
2.736(1)
2.969(1)
179(2)
132 (2)
165(2)
154(2)
1.42(2)
0.86(2)
0.82(2)
1.42(2)
1.90(2)
2.21(2)
3/2ꢀx,1/2ꢀy,ꢀz
In the analysis of the crystal structure of rac-1 as the initial zeroꢀdimensional supramolecular
unit one can take the localized "fork" dimer, the molecules of which are physically linked through
intermolecular Hꢀbonds between the carboxyl groups and are symmetrically related by simple
rotary axes (Figure 5a). These dimers can be formed only by molecules with the same configuration
of the chiral centers (Figure 5a).
The molecules, constituent of the nearby enantiomeric "fork" dimers, are linked through the pairs
of classical hydrogen bonds O3–H3⋅⋅⋅O6=C6 obeying the central symmetry. The arising
supramolecular structure presents a oneꢀdimensional column oriented along 0c axis (Figure 5b,c). In
the crystal, each such column (e.g., column, indicated in gray in Figure 5d) is surrounded by six the
same ones (these columns are colorꢀcoded in the figure). In addition, each 1D column is linked to
four adjacent columns through the system of the spreading along 0c direction centrosymmetric
intermolecular bonds N–H⋅⋅⋅OH (in Figure 5d they are marked with red dashed lines). Thus
constructed packing of the racemic compound is sufficiently dense (PI = 70.4%) and contains no
voids, potentially available for the molecules of a solvent.
Enantiopure (R)ꢀ1 sample crystallizes with two molecules per asymmetric unit of baseꢀcentered
monoclinic cell (space group C2). Increasing the number of symmetrically independent molecules
(parameter Z') when moving from a racemic to the enantiomeric sample is common for carboxylic
ACS Paragon Plus Environment

Crystal Growth & Design
Page 20 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
acids. As shown by Steed et al., this fact is related to the dimeric nature of their supramolecular
24
synthon. Comparative analysis reveals a striking similarity between the unit cell parameters for
(R)ꢀ1 and rac-1 (Table 1). Taking into account that the group C2 is a subgroup of C2/c, we can
expect that the packing nature of homoꢀ and heterochiral crystals would be close to one another.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Supramolecular crystal structure of rac-1. (a) Zeroꢀdimensional enantiomeric "fork"
dimers; view along the 0b axis. Hereinafter dotted lines denote the intermolecular hydrogen bonds
(IMHB); (R)ꢀenantiomers colored in red, (S)ꢀenantiomers – in blue. (b) The 1D column of "fork"
dimers joined by IMHB O–H⋅⋅⋅O=C (labeled green); view along the 0b axis. (c) The same column;
ACS Paragon Plus Environment

Page 21 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
view along the 0c axis. (d) Merging of the oneꢀdimensional columns into a threeꢀdimensional
supramolecular structure through IMHB NH⋅⋅⋅OH (marked in red); view along the 0c axis.
Symmetrically independent molecules A and B have an identical configuration and close values
of bond lengths and bond angles. Visually, the structures of these molecules are compared in Figure
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
a.
Figure 6. Conditional superposition of the molecules existing in the amino acid 1 crystals. (a) Two
independent molecules, A (orange) and B (light blue), in the crystal of (R)ꢀ1. (b) Molecule A
(
(
orange) in (R)ꢀ1 crystal and molecule of Rꢀenantiomer (red) in the crystals of rac-1. (c) Molecule B
light blue) in (R)ꢀ1 crystal and molecule of (S)ꢀenantiomer (blue) in the crystals of rac-1.
More specifically, the conformations of the molecules 1 in crystals are characterized by torsion
angles listed in Table 3. For ease of comparison, this table also shows torsion angles for the two
enantiomers in the rac-1.1 crystals. The table shows that the torsion angles of molecule A in the
(R)ꢀ1 crystals, by the absolute values and by the signs, are practically identical with those of the
independent molecules having Rꢀconfiguration in rac-1.1 crystal. (The virtual distinction for torsion
angles N5C4C3C2 must not be misleading: in both cases we are dealing with antiperiplanar
ACS Paragon Plus Environment

Crystal Growth & Design
Page 22 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
orientation of N5ꢀC4 and C3ꢀC2 bonds, and the actual difference in the dihedral angles of only
about 5 °.) Visually, this similarity can be clearly seen in Figure 6b.
Table 3. Main torsion angles for the molecules 1 in (
R
)-1 and rac-1.1 crystals
rac-1.1 sample
(R)ꢀ1 sample
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Torsion angle (°)
A molecule
ꢀ155.22
24.88
B molecule
156.02
ꢀ24.59
164.90
91.42
Rꢀenantiomer
ꢀ157.05
22.97
Sꢀenantiomer
157.05
ꢀ22.97
C8C7C6N5
C8C7C6O6
C7C6N5C4
C6N5C4C3
N5C4C3C2
N5C4C3O3
C4C3C2C1
C3C2C1O1
C3C2C1O2
ꢀ175.00
ꢀ84.14
ꢀ174.93
ꢀ79.25
174.93
79.25
ꢀ177.81
ꢀ60.75
ꢀ84.40
42.00
177.35
ꢀ64.33
ꢀ177.35
64.33
ꢀ179.30
ꢀ119.95
60.69
ꢀ176.24
ꢀ162.06
23.14
ꢀ179.89
ꢀ118.00
59.70
179.89
118.00
ꢀ59.70
At the same time, the torsion angles for independent B molecule, which naturally has the Rꢀ
configuration in the crystal of (R)ꢀ1, are more similar to the torsion angles of Sꢀenantiomer in the
racꢀ1.1 crystal. The impression arises (see Figure 6c) that the independent molecule B mimics the
molecule of (S)ꢀenantiomer, which is absent in the real homochiral crystal. This impression is
enhanced in the analysis of the intermolecular hydrogen bonds (Table 4) and supramolecular
packing details in the crystals of (R)ꢀ1.
Table 4. Parameters of the intermolecular H-bonds for (
R
)-1 crystals
о
D–H⋅⋅⋅
O2AꢀH2A⋅⋅⋅O2A”
O1AꢀH2A’⋅⋅⋅O1A” 1.289(7) 1.289(7) 2.567(2)
O2BꢀH2B⋅⋅⋅O2B”’
1.308(6) 1.308(6) 2.609(3)
O1BꢀH2B’⋅⋅⋅O1B”’ 1.390(6) 1.390(6) 2.778(2)
A
D–H, Å
1.341(5) 1.341(5) 2.674(2)
H
⋅⋅⋅A, Å
D
⋅⋅⋅A, Å
∠
DHA,
Symmetry operation
ꢀx,y,ꢀz (2ꢀfold axis)
ꢀx,y,ꢀz (2ꢀfold axis)
1ꢀx,y,1ꢀz (2ꢀfold axis)
1ꢀx,y,1ꢀz (2ꢀfold axis)
1/2ꢀx,ꢀ1/2+y,ꢀz
1/2ꢀx,1/2+y,1ꢀz
ꢀ
170(2)
169(2)
172(2)
175(2)
167(2)
153(2)
159(2)
144(1)
O3AꢀH3A⋅⋅⋅O6B
*
0.83(2) 1.91(2)
0.82(2) 2.06(2)
0.85(2) 2.09(1)
0.86(2) 2.21(2)
2.719(1)
2.811(2)
2.903(2)
2.947(2)
O3BꢀH3B⋅⋅⋅O6A**
N5AꢀH5A⋅⋅⋅O3B
N5BꢀH5B⋅⋅⋅O3A
ꢀ
ACS Paragon Plus Environment

Page 23 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
Thus, independent molecules A and B in the (R)ꢀ1 crystals are merged in "fork" Hꢀdimers around
simple rotary axes of order 2. These Hꢀdimers, A⋅⋅⋅A and B⋅⋅⋅B, are analogues of homochiral dimers
R
⋅⋅⋅R and S⋅⋅⋅S in the rac-1 crystals. Next, the dimers A⋅⋅⋅A and B⋅⋅⋅B in the (R)ꢀ1 crystals are bound
through the system of IMHB O3–H3⋅⋅⋅O6=C6 in oneꢀdimensional columns (Figure 7) which are
very similar with the oneꢀdimensional columns in the rac-1 crystals (Figure 5c). Here the role of Rꢀ
enantiomer is played by the molecule A, and the molecule B plays the role of Sꢀenantiomer. Finally,
four of the six contiguous columns are merged in 3D crystal structure with the particular 1D column
through the channel system of IMHB N–H⋅⋅⋅OH (indicated by red dashed rays in Figure 7).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. 1D column formed by independent molecules A and B in the crystals of (R)ꢀ1. View
along 0c axis.
On the whole, judging from the calculated value of the packing index equal 69.3%, so constructed
crystal packing is sufficiently dense. Nevertheless, it is possible to name at least three reasons,
according to which the homochiral crystal packing should be less advantageous than heterochiral
one. Firstly, two symmetry independent molecules participate in the formation of homochiral crystal
instead of one in the case of racemic crystal. Secondly, the conformation sequence for the main
chain of nonhydrogen atoms C8C7C6N5C4C3C2C1 in Sꢀenantiomer in a racemic crystal is
described by the formula ap,ap,sc,ap,ap, while the conformation of independent B molecule, which
simulates the Sꢀenantiomer in homochiral crystal, is described by formula ap,ac,ort,ꢀort,ap, where
at least two sterically optimal antiperiplanar moieties are replaced by two less favorable anticlinal
and orthogonal fragments. Third, for some crystal forming intermolecular classical hydrogen bonds
ACS Paragon Plus Environment

Crystal Growth & Design
Page 24 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
in the (R)ꢀ1 crystal (Table 4), the angular characteristics (
comparison with the same in the racꢀ1.1 crystal (Table 2).
∠DHA) are changed for the worse in
Before leaving the section associated with the Xꢀray analysis of amino acid 1 crystals, we would
like to make one more remark. Based on our previous studies of a series of glycerol ethers, we
established the relationship between the orientation of supramolecular structures and unit cell
parameters and found that the direction of the strongest crystal forming interaction corresponds to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
9
the smallest unit cell parameter. In the case of compound 1, the smallest parameter is the c axis, –
followed by the b axis. Therefore, we can assume that the strongest interactions in our systems are
formed by paired hydrogen bonds O–H⋅⋅⋅O=C propagating along the direction of 0c. Next in force
are hydrogen bonds NꢀH⋅⋅⋅OH oriented along 0b. Finally, the localized IMHB are oriented along the
longest a axis, forming "fork" dimers; these interactions have a minimal impact on the formation of
a general supramolecular structure of the crystals.
Solubility and ternary phase diagram. Thus, Xꢀray analysis showed that, at least from a
racemic solution, the amino acid 1 is crystallized in the form of a normal racemic compound. In
addition, the crystals of the racemic compound (as well as homochiral crystals) do not contain voids
in their structure available for solvent molecules. Finally, we examined a quantitative dependence
of the solubility of compound 1 on the enantiomeric composition of the solid phase, which is
present in equilibrium with the solution. This kind of study allowed us to construct the ternary
phase diagram of the solubility of the system and thus to establish the number and nature of the
phases formed by it and to identify the areas of their (phases) stability and some thermodynamic
parameters of the interfacial equilibria.
We used an achiral acetonitrile as a solvent in our study, since it is chemically stable, does not
form crystal solvates with the system components and provides acceptable solubility measurement
intervals. In our work, we restricted ourselves to isothermal measurements at 20 ± 0°C.
The equilibrium solubility of racemic and enantiopure samples of 1, i.e. saturated solution
concentration in equilibrium with the corresponding crystalline phase, Сsat, was determined by
ACS Paragon Plus Environment

Page 25 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
liquid chromatography as described in the Experimental section. The values obtained were equal
R
ꢀ3
ꢀ1
A
ꢀ1
С
sat = 4.49·10 mole
⋅
L for racꢀ1 and Сsat = 0.114 mole
⋅
L for (R)ꢀ1. Generally speaking, the
1
difference in solubility of the homochiral and racemic samples was observed by our predecessors.
Moreover, the solubility of an enantiopure substance typically exceeds the solubility of a racemate
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5,26
2ꢀ10 times.
Here we emphasize that for the amino acids 1 these differences (about 25 times!) are
abnormally large.
When studying the equilibrium solubility of solid samples of intermediate enantiomeric
composition (0 < ee < 1), it is necessary to control not only the concentration and enantiomeric
composition of the obtained solution, but also the resulting composition of the solid phase, which
was retained after reaching equilibrium, because it may differ from the initial composition (before
addition of solvent). The details of quantitative chromatographic measurements are given in the
Experimental section. The measurement results are shown in Table 5 and Figure 8. In Figure 9 the
data are represented as a fragment of the ternary phase diagram of the system.
Table 5. Equilibrium composition of the solid and liquid phases of saturated solutions of
4-benzoylamino-3-hydroxybutyric acid 1 in acetonitrile
sample
index
[C
R
] in liquid
[C
S
] in liquid
mole fraction of
(R)ꢀ1 in liquid (R)ꢀ1 in solid phase
mole fraction of
ꢀ1
ꢀ1
phase, mole
⋅L , phase, mole
⋅L ,
3
3
×
10
×10
phase
1
2
3
4
5
6
7
8
2.25
2.18
2.65
3.47
4.85
9.97
18.0
114
2.25
1.63
1.54
1.27
0.80
0.49
0.18
ꢀ
0.50
0.57
0.63
0.73
0.86
0.95
0.99
1.00
0.50
0.50
0.50
0.51
0.53
0.48
0.53
1.00
ACS Paragon Plus Environment

Crystal Growth & Design
Page 26 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. The dependence between the amino acid 1 solubility in acetonitrile and the enantiomeric
excess of 1 in solution. Red circles ꢀ the experimental solubility values, blue curve ꢀ the values
calculated under the assumption that the single racemic compound is formed in the system.
Figure 9. Ternary phase solubility diagram (fragment) in the system CH CN/(R)ꢀ1/(S)ꢀ1. Red
3
circles ꢀ the experimental values of the equilibrium composition of the liquid phase.
ACS Paragon Plus Environment

Page 27 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
As evidenced by Figure 8, the experimental points corresponding to the equilibrium composition
of the liquid phase (red circles) are visually situated on a monotonic curve, which does not contain
discontinuities or extrema, when the enantiomeric composition of the liquid phase lies within the
range of ee = 0.5÷0.98. Therewith, in this range of the compositions of the liquid phase, which is in
equilibrium with the solid phase, the enantiomeric composition of the solid remains almost
unchanged and corresponds to racemate within an experimental error (see Table 5, 5th column).
These features reliably indicate that in the investigated range of compositions the only equilibrium
that exists in this system is between the solution and a crystalline 1:1 racemic compound. Thus the
investigated interval belongs to the branch of the ternary phase diagram contained between pure
racemic compound and its eutectic with enantiopure substance. If this is true, then the liquidus line
in this range should be described by the solubility product (SP) for normal racemic compound
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(1).
Numerical SP value may be found on the basis of the solubility of pure racemic compound. Since
for such a system
1
ꢂ
ꢂ
ꢀ
ꢁ ꢃ = ꢀꢄ ꢃ =
ꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢊ2ꢋ,
ꢂ
2
ꢅ
ꢆꢇꢈ
then
ꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢄꢌ = ꢊ1/2ꢅ ꢆꢂ ꢇꢈꢋꢍꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢊ3ꢋ.
The blue curve in Figure 8 is the liquidus line, calculated on the basis of equation (1) after the
ꢀ6
2
ꢀ2
substitution of the numerical value of SP = 5.04⋅10 mole ⋅L , found from equation 3, after the
R
substitution of the experimental value of Сsat . Figure 8 shows that the points, corresponding to the
experimentally observed values of the equilibrium composition of the liquid phase, fit well the
theoretical curve. This suggests that the phase diagram of the system is exhaustively described by
the equilibrium between a solution, a racemic 1:1 compound, and an enantiopure substance. The
phase diagram has no zones of solid solutions, and no molecular addition compounds of
intermediate composition are formed in the system. Consequently, in the equilibrium conditions the
ACS Paragon Plus Environment

Crystal Growth & Design
Page 28 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
existence of the crystalline molecular compound 2:1, postulated by Bergman and Lissitzin, is
impossible.
Normal racemic compounds refer to simple eutectic systems, in which the solid eutectic is formed
by individual phases of the racemic compound and the predominant enantiomer (for clarity,
arbitrarily denoted by symbol R). In the solid state, the activity of the pure substance forming the
individual phase is unity. For the solution in equilibrium with the solid phase, the activities of both
(liquid and solid) components are equal. Consequently, the activity of the dominant enantiomer in
solution in equilibrium with the eutectic is also equal to one. For a dilute solution the activity can be
replaced of the concentration, and then the enantiomeric composition of the eutectic could be easily
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
27,28
found according to the following scheme:
ꢏꢐ
ꢏꢐ
ꢀꢁ ꢃ − ꢀꢄ ꢃ
ꢀ
ꢏꢐ
ꢎꢎꢏꢐ
ꢏꢐ
=
ꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉ
ꢏꢐ
ꢏꢐ
ꢁ ꢃ + ꢀꢄ ꢃ
ꢂ
ꢍ
ꢉꢀꢁ ꢃꢀꢄ ꢃ = ꢄꢌ = ꢊ1/2ꢅ ꢋ ꢉꢉꢉ
ꢆꢇꢈ
ꢀꢁ ꢃ = ꢅꢑ
ꢏꢐ
ꢆꢇꢈ
ꢂ
ꢍ
ꢊ1/2ꢅ ꢋ
ꢆꢇꢈ
ꢑ
ꢉꢉꢉꢉꢉꢀꢄꢏꢐ
ꢃ
=
ꢉꢉ
ꢅ
ꢆꢇꢈ
ꢍ
ꢂ
ꢅ
2
ꢆꢇꢈ
ꢑ
1
− ꢒ
ꢓ
ꢅ
ꢆꢇꢈ
ꢎꢎꢏꢐ
=
ꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢉꢊ4ꢋ
ꢍ
ꢂ
ꢅ
2
ꢆꢇꢈ
ꢑ
1 + ꢒ
ꢓ
ꢅ
ꢆꢇꢈ
Calculated by eq. (4) the value eeeu = 0.999. This means that for the chiral amino acid 1 the
eutectic composition is virtually indistinguishable from the composition of the enantiopure
compound. Accordingly, 4ꢀbenzoylaminoꢀ3ꢀhydroxybutyric acid 1 can be classified as a
pronounced "anticonglomerate", i.e. a substance that is capable of strong enantiomeric enrichment
of the predominant chiral component in solution, even in the case where the enantiomeric excess of
the entire sample, distributed between the solution and the solid phase, turns out to be very small. It
is possible that this behavior of substances present in the prebiotic brines is directly related to the
29
origin of homochirality of life.
ACS Paragon Plus Environment

Page 29 of 56
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
Experiment on the nonequilibrium crystallization. As it was just demonstrated, the equilibrium
phase behavior of amino acid 1 corresponds to a simple eutectic system "normal racemic compound
↔
pure enantiomer." But it cannot be denied, however, that the principal possibility exists of
formation of metastable anomalous conglomerate in far from equilibrium conditions. Bergman and
Lissitzin have attributed the structure of the anomalous molecular compound to the solid phase
obtained in such conditions, namely during a rapid crystallization of the highly supersaturated
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
aqueous solution of acid 1 sample with an enantiomeric excess of about 33%.
We have repeated the experiment described in this work. For which purpose 25 mg of (R)ꢀ1 was
dissolved in 1 ml of water; then, 50 mg of racꢀ1 was added to the solution, and the mixture was
heated until complete dissolution. Upon reaching room temperature, the vial was filled with acicular
crystals, the crystal habit of which is the same as in Figure 1b. The soaked with the mother liquor
(i.e. almost saturated solution of pure enantiomer) precipitate was washed on the filter with a
minimum amount of ice water. After drying in air 44 mg of the acid 1 was collected; mp 174 °С,
2
0
[
α
]D = ꢀ 4 (c 1.0, 0.5 N aq. NaOH). The precipitate was studied by powder diffractometry.
In Figure 10 the experimentally obtained diffraction pattern for the above prepared
polycrystalline sample (black curve) is compared with the experimental diffraction patterns of pure
racꢀ1 (red curve) and (R)ꢀ1 (blue curve). The black and red curves are practically identical, and this
gives the reason to believe that the unit cell parameters and other characteristics of the crystal
packing in both samples are the same. The coincidence of the diffraction patterns, in combination
with the similar crystal habits and close to one another melting points, allows identifying the
crystalline component of the precipitated in the nonequilibrium conditions solid as normal racemic
compound, i.e. racꢀ1.
ACS Paragon Plus Environment