Spontaneous Resolution of Chiral 3-(2,3-Dimethylphenoxy)propane-1,2-diol under the Circumstances of an Unusual Diversity of Racemic Crystalline Modifications

Article abstract of DOI:10.1021/acs.cgd.7b00510

Depending on the conditions of crystallization from solutions, racemic 3-(2,3-dimethylphenoxy)propane-1,2-diol 1 forms three relatively stable crystalline modifications. Each of the crystalline forms, namely, two polymorphic racemic compounds and a racemi

Full text of DOI:10.1021/acs.cgd.7b00510

Subscriber access provided by AUSTRALIAN NATIONAL UNIV
Article
Spontaneous resolution of chiral 3-(2,3-
dimethylphenoxy)propane-1,2-diol under the circumstances
of an unusual diversity of racemic crystalline modifications
Alexander A Bredikhin, Dmitry V. Zakharychev, Zemfira A. Bredikhina,
Alexey V Kurenkov, Dmitry B. Krivolapov, and Aidar T. Gubaidullin
Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.7b00510 • Publication Date (Web): 10 Jul 2017
Downloaded from http://pubs.acs.org on July 17, 2017
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 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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
Graphical abstract
45x23mm (300 x 300 DPI)
ACS Paragon Plus Environment

Crystal Growth & Design
Page 2 of 31
1
2
Cover Page
3
4
Spontaneous resolution of chiral
5
6
7
8
3-(2,3-dimethylphenoxy)propane-1,2-diol under the circumstances
of an unusual diversity of racemic crystalline modifications
9
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
Alexander A. Bredikhin*, Dmitry V. Zakharychev, Zemfira A. Bredikhina, Alexey V. Kurenkov,
Dmitry B. Krivolapov, 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
Chiral phenyl glycerol ether bearing two methyl groups in the ortho and meta positions, 3ꢀ(2,3ꢀ
dimethylphenoxy)propaneꢀ1,2ꢀdiol 1, crystallizes from the racemic feed medium with the
formation of five crystalline modifications: αꢀrac, βꢀrac, γꢀrac, δꢀrac racemic compounds and
racemic conglomerate. The αꢀrac, βꢀrac and racemic conglomerate are formed during solutions
crystallization as good crystals, the nature of which was established by single crystal Xꢀray
diffraction. The γꢀrac and δꢀrac are formed during racꢀ1 crystallization from melt. The graphs of
the dependence of free Gibbs energy on temperature were constructed. The δꢀrac phase is
unstable and turns into γꢀrac, further the last spontaneously recrystallizes into βꢀrac. Stable in
dry state, in a solvent medium βꢀrac is converted to αꢀrac. Crystallization of racꢀ1 from
solutions at room temperature leads to the αꢀrac phase which is stable both in dry form and in
solutions. Taking into account that above room temperature the conglomerate is
thermodynamically more favorable than all polymorphic racemic compounds, and that this
particular form crystallizes from solutions under the condition of initial enantiomeric enrichment
of the sample, we realized a successful direct resolution of racꢀ1 into individual enantiomers.
* Corresponding author: Phone/fax: +7 843 2727393/ +7 843 2731872. Eꢀmail: baa@iopc.ru.
ACS Paragon Plus Environment

Page 3 of 31
Crystal Growth & Design
1
Spontaneous resolution of chiral 3-(2,3-dimethylphenoxy)propane-1,2-diol
under the circumstances of an unusual diversity of racemic crystalline
modifications
2
3
4
5
6
7
8
9
Alexander A. Bredikhin*, Dmitry V. Zakharychev, Zemfira A. Bredikhina, Alexey V. Kurenkov,
Dmitry B. Krivolapov, 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
ABSTRACT
Depending on the conditions of crystallization from solutions, racemic 3ꢀ(2,3ꢀdimethylphenoxy)
propaneꢀ1,2ꢀdiol 1 forms three relatively stable crystalline modifications. Each of the crystalline
forms, namely, two polymorphic racemic compounds and a racemic conglomerate, have been
characterized by single crystal XRD. Two more metastable racemic compounds crystallized
from the racemic melt have been found by DSC method. Additional thermochemical
investigations allowed to plot the dependence of the free Gibbs energy on temperature for all the
phases found. With the help of slurrying experiments, the nature of the transitions between solid
phases has been specified. It has been found that even a slight predominance of one of the
enantiomers in almost racemic 1 samples ensures the crystallization of the conglomerate. The
revealed features of racꢀ1 crystallization have been taken into account during the realization of
its resolution into individual enantiomers by the entrainment procedure.
INTRODUCTION
Polymorphism is an evergreen theme for researchers of molecular crystals.1,2 Current interest in
this problem is largely supported by the needs of pharmaceutical science and industry.3,4 In
recent years, chiral compounds are increasingly becoming objects of the research.5ꢀ11 An
additional feature of chiral systems that is absent in the case of achiral substances is the possible
interconversion of homochiral (scalemic, enantiopure) and heterochiral (mainly racemic)
crystalline forms.12ꢀ14 In the strict sense of the word, conglomerate homochiral crystals are not
polymorphic with respect to any racemic solid. However, like ordinary polymorphs, those and
others can be formed from a common feed medium (solution or melt) and can transform into
each other both as a result of solidꢀphase transitions and as a result of contact with the common
ACS Paragon Plus Environment

Crystal Growth & Design
Page 4 of 31
1
2
3
4
5
6
liquid phase.12 Lack of attention to this feature of a phase behavior of the chiral substance can
lead to withdrawal of the dosage form from the market and/or the need for remarketing of the
already registered drug.14 and references therein
The object of the present work is 3ꢀ(2,3ꢀdimethylphenoxy)propaneꢀ1,2ꢀdiol 1 (Chart 1), a
glycerol phenyl ether bearing two methyl groups in the ortho and meta positions of the phenyl
ring.
7
8
9
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
Chart 1. 3ꢀ(2,3ꢀDimethylphenoxy)propaneꢀ1,2ꢀdiol 1
and some structurally related compounds mentioned in the text
By its chemical structure, it belongs to a series of chiral terminal aromatic glycerol ethers
(TAGE) with the general formula ArꢀOꢀCH2CH(OH)CH2OH, which possess a number of
valuable and interesting properties for the researcher.15 In particular, in this series there is an
increased frequency of conglomerates,16 the existence of polymorphic modifications has been
proved for the racemic crystals.9,11 Finally, in this series the unique of its kind solidꢀphase
transition between the metastable anomalous and stable normal conglomerates has been
observed.9 We believe that the accumulation and systematization of information on the crystal
structure, phase diversity, quantitative characteristics of solid phases and transitions between
them in a series of structurally close compounds will broaden and deepen our understanding of
the relationship between the chemical structure of matter and its properties.
We have previously shown that the mono ortho methylꢀsubstituted phenyl glycerol ether 2
(Chart 1, API mephenesin) from the racemic feed material crystallizes as a normal
conglomerate,17 that is, it prones to spontaneous resolution during crystallization. Furthermore,
the appearance of additional methyl group in the second ortho position of the benzene ring in the
glycerol ether 4 changes dramatically its crystal structure in comparison with 2, but does not
change the type of crystallization, and this chiral diol remains a normal conglomerate.18 At the
same time, the racemic metaꢀsubstituted mephenesin analogue 3 crystallizes as a normal racemic
compound.17
Judging from the literature, (S)ꢀ1 (mp 104–105 oC)19 melts 21 degrees above than racꢀ1 (mp
82.5ꢀ83 oC).20 In this connection, compound 1 could also be expected to crystallize in the form of
ACS Paragon Plus Environment

Page 5 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
a conglomerate. As will be shown below, these expectations are to some extent justified, but the
presence of a second methyl substituent leads to rather unexpected consequences.
Diol 1 shares a common part and can be used as a precursor for the synthesis of API
xibenolol 5 (Chart 1). Xibenolol is the common name for 1ꢀ(tertꢀbutylamino)ꢀ3ꢀ(2',3'ꢀ
dimethylphenoxy)ꢀ2ꢀpropanol and is a selective β−adrenergic blocker.21 Its salt, xibenolol
hydrochloride, in racemic form is present at the market as the antiarrhythmic remedy under the
trade names Selapin and Rhythminal. Therefore, the effective production of 1 in the scalemic
(nonracemic, enantiopure) forms by stereoselective crystallization may not only be of theoretical
but also of practical interest, since synthetic routes exist for converting this diol to single
enantiomer forms of xibenolol.19
9
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
EXPERIMENTAL SECTION
Instrumentation. The IR spectra of the polycrystalline samples of racꢀ and (R)ꢀdiols 1 in
KBr pellets were recorded on a Bruker Tensor 27 spectrometer. Optical rotations were measured
on a Perkin–Elmer model 341 polarimeter (concentration c is given as grams per 100 milliliters).
Melting points for general purposes were determined using a Boëtius apparatus. HPLC analyses
were performed on a Shimadzu LCꢀ20AD system controller, using UV detector. The
thermograms were measured on a Netzsch DSC 204 F1 Phoenix differential scanning
o
calorimeter (τꢀsensor) in aluminum pans with the rate of heating and cooling of 5 C minꢀ1. The
mass of the samples amounted to about 1ꢀ5 mg for determining the characteristic temperatures of
the processes and measuring the enthalpies of phase transitions, and 5ꢀ10 mg for measurements
of heat capacities. The mass of the samples was controlled with Sartorius CPA 2P balance. The
heat capacities were measured in continuous mode at a heating rate of 5 °C/min. The standard
calibration of the instrument was performed using the temperatures and enthalpies of melting for
In, Sn, and Bi standards and the phase transition in adamantane. Verification of calibration
accuracy and measurement error estimation for temperature and enthalpy of melting values was
carried out by melting samples of naphthalene and benzoic acid, and for heat capacity
measurements by recording a thermogram of corundum disk (m = 50 mg). The reproducibility
for the measured temperature values was ~ 0.3 oC, for the melting enthalpies better than 1%, and
for the heat capacities ~ 2ꢀ5%.
Substances. Racemic and enantiopure 3ꢀ(2,3ꢀdimethylphenoxy)propaneꢀ1,2ꢀdiol were
prepared by analogy with published procedure.18 A detailed description of the experiment and
NMR spectra of the samples were given in the Supporting Information.
racꢀ3ꢀ(2,3ꢀDimethylphenoxy)propaneꢀ1,2ꢀdiol, racꢀ1.3: mp 81.5ꢀ90 oC (light petroleum ether/
EtOAc) [lit.20 mp 82.5ꢀ83 oC (ligroin; benzene)].
ACS Paragon Plus Environment

Crystal Growth & Design
Page 6 of 31
1
2
o
(R)ꢀ3ꢀ(2,3ꢀDimethylphenoxy)propaneꢀ1,2ꢀdiol, (R)ꢀ1: mp 101ꢀ102.5 C (light petroleum
3
4
5
6
7
8
9
20
20
25
ether/ EtOAc); [α]D = +1.4 (c 1.0, EtOH), [α]D = +13.4 (c 1.1, MTBE), {lit.22 [α]D = +1.44
(c 1.0, EtOH) 96.6 % ee}; 99.3% ee [chiral HPLC analysis; Daicel Chiralcel OD (0.46 x 25 cm)
column; eluent hexane/isopropanol = 8:2; flow rate 1.0 mL/min; UV detector 275 nm (tR = 11.2
min)].
o
20
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
(S)ꢀ3ꢀ(2,3ꢀDimethylphenoxy)propaneꢀ1,2ꢀdiol (S)ꢀ1: mp 101.5ꢀ103 C; [α]D = –1.0 (c 1.1,
EtOH), [α]D = –13.5 (c 1.0, MTBE), [α]D = +2.3 (c 1.1, CHCl3) {lit.19 mp 104–105 C
(EtOH); [α]D25 = +4.25 (c 1.0, CHCl3)}; 99.9 % ee (chiral HPLC analysis; tR = 13.0 min).
Solubility and nucleation boarder studies. In order to determine the solubility, the
polythermal laser monitoring last crystal disappearance method was used. Precisely controlled
samples [Shimadzu AUW 120D analytic balance (accuracy ± 0.01 mg)] of racꢀ1 and (S)ꢀ1 in 1
ml of solvent were placed in sealed glass vial, equipped with magnetic stir bar. For precise
temperature control another vessel equipped with a controlling thermometer, a stir bar and a
solvent was placed just near the test vial. All the vessels were placed in a thermostatic container.
The suspension of the sample was agitated at about 400 rpm, increasing the temperature
gradually (0.5 °C/min). The turbidity degree was registered to detect the “clear point”. The
temperature where the last crystals disappeared was taken as saturation temperature. After
complete dissolution of the crystals in the studied reactor the coolant was slowly cooled down
(0.3 °C/min), detecting the soꢀcalled ‘cloud points’ (the temperature where the first crystals were
reliably detected) in cooling cycles. Each "heatingꢀcooling" cycle was repeated three times. As
the limits of the range, the lowest value of the recorded endꢀofꢀdissolution temperature and the
highest value of the crystallization start temperature were selected.
20
20
o
Racemic 3-(2,3-dimethylphenoxy)propane-1,2-diol, rac-1 resolution by entrainment.
o
Racemic diol racꢀ1 (1.350 g) and (R)ꢀ1 (0.150 g) was dissolved in 176.5 ml of water at 50 C.
o
The solution was cooled to 39 C and then seeded with finely pulverized (R)ꢀ1 (0.008 g). After
stirring the mixture for 107 min at 37.5 oC, precipitated (R)ꢀ1 was collected by filtration (0.295 g
after drying; 87.4 % ee). The extra portion of racꢀ1 (0.287 g) was then dissolved in the mother
o
o
liquor at 50 C. The resulting solution was cooled to 39 C. After the addition of (S)ꢀ1 (0.008 g)
as seed crystals to the solution, and stirring the mixture for 90 min at 37.5 oC, (S)ꢀ1 (0.320 g after
drying; 53.2 % ee) was collected by filtration.
o
X-ray Analysis. The crystals of racꢀ1.1 (prisms, mp 79.5ꢀ82 C) and racꢀ1.2 (needles, mp
o
76.5ꢀ78 C) for single crystal Xꢀray diffraction analysis were prepared by slow evaporation of
the solution of racemic sample in a mixture of cyclohexane and EtOAc at ambient temperature.
A single crystal, selected at random from the racemic polycrystalline sample racꢀ1.3, which was
prepared by rapid recrystallization of the racemic sample from the hot solution in a light
ACS Paragon Plus Environment

Page 7 of 31
Crystal Growth & Design
1
2
3
4
petroleum ether and EtOAc mixture, was used for Xꢀray analysis; this crystal was designated
also as Sꢀ1.
5
6
7
8
The Xꢀray diffraction data for these crystals were collected on a Bruker AXS Smart Apex II
CCD diffractometer in the ωꢀscan and φꢀscan modes using graphite monochromated Mo Kα (λ
0.71073 Å) radiation at 296(2) K. The crystal data, data collection, and the refinement
parameters are given in Table 1.
9
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
Data were corrected for the absorption effect using SADABS program.23 The structures
were solved by direct method and refined by the full matrix leastꢀsquares using SHELX24 and
WinGX25 programs. All nonꢀhydrogen atoms were refined anisotropically. All hydrogen atoms
were inserted at calculated positions and refined as riding atoms except the hydrogen atoms of
OH groups which were located from difference maps and refined isotropically or as riding
atoms. All figures were made using Mercury program.26 Molecular structures and conformations
were analyzed by PLATON.27
Crystallographic data for the structures racꢀ1.1, racꢀ1.2 and racꢀ1.3 reported in this paper
were deposited with the Cambridge Crystallographic Data Centre as supplementary publication
number CCDC 1538660 – 1538662 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)1223ꢀ
336033 or eꢀmail: deposit@ccdc.cam.ac.uk).
ACS Paragon Plus Environment

Crystal Growth & Design
Page 8 of 31
1
2
3
Table 1
4
5
Crystallographic data for crystals picked from rac-1.1, rac-1.2, and rac-1.3 samples
6
7
сompound
rac-1.1
racꢀ1.2
rac-1.3 (S-1)
8
9
sample formula
M (g/mol)
C11 H16 O3
196.24
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
temperature, K
crystal class
space group
crystal size
Z, Z´
296(2)
296(2)
296(2)
Monoclinic
C2/c
triclinic
Pꢀ1
monoclinic
P21
0.50x0.50x0.2 mm3
0.4x0.01x0.01
0.36x0.10x0.10
8, 1
2,1
4, 2
cell parameters
a = 37.05(2) Å,
b = 5.859(4) Å,
c = 9.685(6) Å
102.786(8)°
4.808(3)
5.858(3)
18.302(10)
94.501(7)
93.993(8)
91.813(7)
12.831(16)
5.037(6)
17.50(2)
101.813(15)
V (Å3)
F(000)
2050(2)
848
512.3(5)
212
1107(2)
424
ρ
calcd (g/cm3)
1.271
1.272
1.177
ꢀ
θ
(cmꢀ1)
0.91
0.91
0.85
(deg)
2.255ꢀ29.504
12157
2.238ꢀ29.476
6293
1.621ꢀ29.329
9632
reflns measured
independent reflns / R(int)
2702 / 0.0334
137 / 0
2570/0.0350
135/2
4939/0.1152
262/1
Number of params /
restraints
Reflns [I > 2
σ(I)]
1866
1450
1472
R1 / wR2 [I > 2
σ
(I)]
0.0437 / 0.1155
0.0674 / 0.1311
0.985
0.0542/0.1459
0.1045/0.1708
1.034
0.0834/0.1605
0.2751/0.2376
0.877
R1 / wR2 (all reflns)
GOF on F2
ρmax/ρ
min (e Åꢀ3)
0.189 / ꢀ0.153
0.182 / ꢀ0.162
0.185 / ꢀ0.212
ACS Paragon Plus Environment

Page 9 of 31
Crystal Growth & Design
1
2
3
RESULTS AND DISCUSSION
4
5
6
7
Synthesis of 3-(2,3-dimethylphenoxy)propane-1,2-diol and diversity of its crystal forms.
The samples of racemic diol 1 required for the initial study were obtained by the interaction of
2,3ꢀdimethylphenol with racemic 3ꢀchloropropaneꢀ1,2ꢀdiol.
8
9
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
To obtain the scalemic diols we have used commercially available single enantiomeric (R)ꢀ
and (S)ꢀ3ꢀchloropropaneꢀ1,2ꢀdiols (see ESI for details).
Slow room temperature evaporation of solutions of racemic diol 1 in a mixture of
cyclohexane and a small amount of ethyl acetate leads to the formation of two types of crystals
on the bottom of the vessel: flat transparent prisms (Figure 1a, we will denote such samples as
racꢀ1.1) and characteristic beams of crystals forming around the common center (Figure 1b, racꢀ
1.2). Rapid cooling of heated concentrated solutions of racemic 1 in a mixture of light petroleum
and ethyl acetate results in the formation of loose polycrystalline samples in which the crystals
retain the acicular habitus (Figure 1c; hereinafter the appropriate crystalline samples will be
referred to as racꢀ1.3). The racꢀ1.3 samples melt over a wide temperature range of 81.5ꢀ90 °C,
which is typical for conglomerateꢀforming compounds. Crystallization of scalꢀ1 from a mixture
light petroleum/EtOAc leads to needles that differ markedly in size within the same sample
(Figure 1d). In their appearance, they resemble racꢀ1.3 crystals. The scalemic crystals melt in a
narrow range of 101ꢀ102.5 °C.
ACS Paragon Plus Environment

Crystal Growth & Design
Page 10 of 31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 1. Microscopic pictures of scalemic and racemic crystalline samples of diol 1: (a) prisms
racꢀ1.1 and (b) needles racꢀ1.2 after room temperature slow crystallization of racꢀ1 from
cyclohexane/ EtOAc; (c) racꢀ1.3 crystals after rapid racꢀ1 crystallization from hot solution in
light petroleum/EtOAc; (d) scalꢀ1 crystals from light petroleum/EtOAc.
The racꢀ1.1 crystals clearly differ from other diol 1 crystalline modifications in their
appearance. In order to clarify the relationships between all these modifications, we compared
their infrared spectra recorded in the KBr matrix in pairs. The comparison itself is illustrated in
Figure 2.
ACS Paragon Plus Environment

Page 11 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 2. IR spectra in KBr pellets (on left) and their correlation trajectories (on right): (a) racꢀ
1.3 (red curve) against scalꢀ1, in this case (R)ꢀ1, (blue curve); (b) (racꢀ1.1 (olive) against (R)ꢀ1
(blue); (c) racꢀ1.2 (magenta) against (R)ꢀ1 (blue); (d) racꢀ1.1 (olive) against racꢀ1.2 (magenta).
The left graph of each pair of drawings corresponds to the two compared IR spectra
ꢀꢁ ꢂ ꢃꢄꢅꢆ and ꢀꢇ ꢂ ꢃꢄꢅꢆ (A1, A2 are the relative absorption of the samples for the same
frequency ν), represented in the standard form. To quantify the degree of similarity, the standard
statistical parameter, Pearson correlation coefficient ꢈ ꢂ ꢉꢊꢋꢄꢌ ,ꢌ ꢆ, was used. The right graph in
ꢍ
ꢎ
ꢏ
ꢐ
ꢏ
ꢐ
ꢎ
ꢍ
the drawings (“correlation trajectory”) displays the parametric dependence ꢀꢁ vs ꢀꢇ, which is a
graphical representation of the correlation between the spectra and reflects the presence or
absence of mutual agreement in the variation of the absorption of different samples when
moving along the ν scale. Obviously, if the two spectra are equally normalized and completely
ACS Paragon Plus Environment

Crystal Growth & Design
Page 12 of 31
1
2
3
4
5
6
7
8
similar, the correlation trajectory should degenerate into a straight line with the angle tangent
equal to 1. If there are differences in the characteristic wavelengths and the extinctions of the
absorption bands of the spectra being compared, then characteristic loops appear on the graph.
Unlike the correlation coefficient R, which is an integral parameter, the graphical representation
of the correlation between the spectra demonstrates the nature of the differences between them in
compact and evident form, especially effective when it is necessary to quickly extract similar (or,
alternatively, different) fragments in a large data set. Details of such an analysis have been
described in our previous publications.28,16
9
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 can be seen from the figure, the spectra of polycrystalline samples of racꢀ1.3 and scalꢀ1
practically coincide, which indicates a significant sameness of their crystalline organization. For
chiral compounds in which functional groups (in this case, primary and secondary hydroxyl
groups of the glycerol fragment) capable of forming strong intermolecular hydrogen bonds are
present, the coincidence of the spectra of enantiopure and racemic samples is most likely
indicative that racꢀ1.3 represents a racemic conglomerate, that is, a mechanical mixture of
crystals, the organization of which coincides (or represents a mirror copy) with the (R)ꢀ1 crystal
packing. The coincidence of the spectra of scalꢀ1 and racꢀ1.3 allows us to use only the scalemic
sample for comparison purposes. As Figure 2 shows, all other pairs of spectra differ
significantly. Consequently, racꢀ1.1, racꢀ1.2 and racꢀ1.3 samples have different crystal packings.
Xꢀray analysis confirms this conclusion.
Single-crystal X-ray investigations. The unit cell of racꢀ1.1 crystals contains a single
symmetry independent molecule. This molecule and the numbering scheme adopted in the article
are shown in Figure 3a. The unit cell belongs to the monoclinic centrosymmetric space group
C2/c (Table 1). Already on this basis it is clear that racꢀ1.1 crystals represent a normal racemic
compound. The primary crystalꢀformative motif for this crystal packing is a 2D bilayer parallel
to the 0bc plane (Figure 3b). As it can be seen from the figure, the bilayer consists of infinite 1D
ribbons formed by enantiomers with the same configuration. Its 2D structure is consolidated by
infinite chains of intermolecular hydrogen bonds O1−H1···O'2 [d (O1···O'2) = 2.835(2) Å,
∠(O1H1O'2) = 172(2)°], O'2−H'2···O''1 [d (O'2···O''1) = 2.854(2) Å, ∠(O'2H'2O''1) = 160(2)°],
O''1−H''1···O'''2, etc.
ACS Paragon Plus Environment

Page 13 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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. (a) Geometry of the symmetry independent Sꢀmolecule in the crystal of racꢀ1.1 sample
and numbering scheme adopted in the text. Nonꢀhydrogen atoms are represented by probability
ellipsoids of thermal vibrations (p = 50%), hydrogen atoms – by spheres of arbitrary radii.
(b) Detail of the crystal packing for racꢀ1.1: 2D bilayer parallel to 0bc plane; Rꢀenantiomers are
colored in red, Sꢀenantiomers are colored in blue.
The racꢀ1.2 crystals also belong to the centrosymmetric space group (Table 1, Pꢀ1, Z' = 1).
That is, despite the fact that racꢀ1.2 crystals with their needle shape are to some extent similar to
the racꢀ1.3 crystals (Figures 1b and 1c), the racꢀ1.2 crystals represent the normal racemic
compound, and hence this form is a polymorphic version of racꢀ1.1. The primary motif in racꢀ
1.2 crystals is a 1D column oriented along the 0a axis (Figure 4). Two zigzag sequences of
intermolecular Hꢀbonds, O1
−
H1···O'2 [d (O1···O'2) = 2.782(3)Å,
∠(O1H1O'2) = 161(3)°] and
O'2 H'2···O''1 [d (O'2···O''1) = 2.775(3)Å,
−
∠(O'2H'2O''1) = 173(2)°], link the alternating
molecules of the opposite enantiomers to a single extended 1D system.
Figure 4. Detail of the crystal packing for racꢀ1.2: 1D column parallel 0a direction.
Xꢀray diffraction analysis results for a single crystal picked at random from the
polycrystalline racꢀ1.3 sample are shown in Table 1 (where this crystal is designated as Sꢀ1).
There are two independent molecules A and B in the asymmetric unit of the monoclinic P21 unit
ACS Paragon Plus Environment

Crystal Growth & Design
Page 14 of 31
1
2
3
4
5
6
cell of this crystal. Both molecules have common Sꢀconfiguration, practically coinciding lengths
of bonds and valence angles and differ only in the conformation of the glycerol fragment (see
below). These molecules are shown in Figure 5.
7
8
9
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 5. (a) Geometry of symmetry independent molecules in the crystal picked from racꢀ1.3
polycrystalline sample. Nonꢀhydrogen atoms are represented by probability ellipsoids of thermal
vibrations (p = 30%), hydrogen atoms – by spheres of arbitrary radii. (b) Conditional
superposition of the independent molecules; A molecule ꢀ red, B molecule ꢀ blue.
Crystallization of a racemate in the Sohncke group with any number of independent
molecules with equal configuration guarantees the enantiomeric purity of an ideal single crystal
and in this sense, is the proof of spontaneous resolution of a substance. Xꢀray diffraction analysis
of a single crystal selected from the enantiopure Rꢀ1 sample was brought to the establishment of
unit cell parameters, which, to within an experimental error, coincided with those listed in Table
1 for racꢀ1.3, which confirms their general nature.
The primary supramolecular crystalꢀformative motif in Sꢀ1 crystals is a 1D column formed
around the 21 axis in the 0b direction and consisting of alternating independent molecules A and
B. We draw the readers’ attention to the fact that oneꢀdimensional supramolecular motifs are
realized in the crystals racꢀ1.2 and racꢀ1.3 having an acicular habitus. The presence of a twoꢀ
dimensional motif in the case of racꢀ1.1 leads to a striking change in the habitus and the
formation of plateꢀlike crystals.
Let us return to the supramolecular motif in scalꢀ1 crystals. The molecules forming the
column are interconnected by a system of intermolecular hydrogen bonds (IMHB) O−H···O'.
The sequence of such IMHBs forms a leftꢀhanded Mꢀhelix in the case of a crystal formed by Sꢀ
molecules. Eight molecules (four A and four B) fall inside one helix pitch. By all these
indications, it coincides with the wellꢀknown for the TAGE family mephenesinꢀlike motif.15 In
particular, it is present in the scalemic crystals of monomethylated analogs of diol 1, orthoꢀ and
ACS Paragon Plus Environment

Page 15 of 31
Crystal Growth & Design
1
2
3
4
metaꢀtolyl glycerol ethers 2 and 3 (Chart 1).17 Details of the 1D supramolecular crystalꢀformative
motif in Sꢀ1 crystals are illustrated in Figures S1 and S2, ESI.
5
6
7
8
Spatial organization of the glycerol fragment in TAGE molecules is completely described by
enumerating the conformations of the H1O1C1C2, O1C1C2C3, C1C2C3O3, C2C3O3C4,
H2O2C2C3 и O2C2C3O3 fragments. In the nonꢀracemic crystals of all the just mentioned diols
1ꢀ3, conformational formulas for symmetry independent molecules A (ap,ꢀsc,ꢀsc,ap,ꢀsc,sc) and B
(sc,sc,sc,ap,ac,ap) turn out to be constant (angle signs are given for Rꢀenantiomers). The
numerical values of the corresponding torsion angles are also quite close, which indicates the
similarity of the spiral structures of the primary supramolecular motifs for the molecules being
compared. The main factor that destabilizes the mephenesinꢀlike motif in the case of a metaꢀ
substituted derivative 3 is the close contact between the methyl groups in the spiral stacks.17 This
factor leads to an increase in the spiral pitch and, as a consequence, to a decrease in the packing
density. It is easy to verify that in the case of 2,3ꢀdimethyl derivative 1 this tendency is
aggravated. The dimension of the full helix pitch of the supramolecular motif coincides with the
b parameter of the P21 unit cell, common for all diols 1ꢀ3. In the discussed series, this parameter
is 4.8558(3) Å for mephenesin 2, 4.9317(5) Å for metaꢀtolyl ether 3 and 5.037(6) Å for
disubstituted diol 1. The packing density of the crystals also decreases in the same sequence: the
value of PI is 67.5% for scalꢀ2, 65.5% for scalꢀ3 (68.6%) and 65.1% for Sꢀ1. (racꢀ1.1 69.9 %;
racꢀ1.2 70.0 %)
9
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
With this information alone, one would expect that the homochiral packing, which is the
only stable crystalline form in the case of mephenesin 2, loses its advantages in the case of metaꢀ
tolyl glycerol ether 3 and, especially, in the case of 2,3ꢀdimethylphenyl ether of glycerol 1. This
assumption is confirmed indeed in the case of compound 3, which crystallizes from the racemic
feed material in the heterochiral form of normal racemic compound. For racꢀ3 sample PI is equal
to 68.6% and the Gibbs energy of the racemic compound formation was found to be ꢀ2.70 kJ
molꢀ1.17 This means that for metaꢀmethyl derivative 3 an empirical regularity is valid, according
to which racemates are denser than the corresponding enantiopure samples, so called Wallach's
rule. (Or, if we comply with the historical justice, Liebisch's rule, as it is suggested in the
informative article of Ernst.29) However, in the same work in which the term “Wallach rule” first
appeared, it was shown that the rule is not absolute and contains numerous exceptions.30
Subsequently, this conclusion was repeatedly confirmed in the study of crystals of
homogeneous31,32 and dissimilar33 arrays of chiral organic compounds.
The crystallization of diol 1 discussed herein has much in common with the crystallization
of 3ꢀphenyllactic acid, studied in the Navare and MacDonald paper.13 In both cases, racemic
conglomerate and several polymorphs of racemic compound can crystallize from the racemic
ACS Paragon Plus Environment

Crystal Growth & Design
Page 16 of 31
1
2
3
4
5
6
feed material depending on the conditions. In both cases, Wallach's rule is formally observed,
and homochiral crystals are less dense than racemic crystals. But the firsts melt at higher
temperatures.
In order to analyze the reasons for preferences in the process of rac vs scal crystallization,
the series of compounds 2 – 3 ꢀ 1 is of particular interest since, with the general supramolecular
motif retaining for homochiral crystals throughout the entire series and the monotonic decrease
in density noted for them, the nature of the stable form of the crystals, scal ꢀ rac – scal, changes
during the crossꢀover from compound to compound. Such an analysis involves the use of highꢀ
level calculations and goes beyond the scope of this paper. But in any case, for its verification,
such analysis requires some reference points related to the energy of the crystals.
Generally speaking, taken by themselves, the data of the Xꢀray experiments do not provide a
basis for establishing energy preferences for one or another compound. We tried to do this in
other ways.
7
8
9
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
Thermal investigations of the stable and metastable crystalline diol 1 forms. Figure 6
shows the DSC thermograms of diol 1 samples obtained by crystallization from solutions. As it
o
can be seen from the figure, the melting of the scalemic sample (blue curve, ꢑ =102.8 C) is
ꢒ
ꢓ
characterized by a single peak of regular shape. Further experience has shown that the crystalline
enantiopure samples obtained under different conditions from the solution, as well as the
crystallization products of enantiopure melts, exhibit the same monomodal melting peak when
heated. Most likely, the scalemic diols 1 in the system under investigation are characterized by
the only described above type of crystal packing. Consequently, on the racemic samples
o
thermograms, the one and only one endothermic process, namely the peak ꢑ ꢓꢔꢕ= 81.1 C for
ꢒ
racꢀ1.3 sample (Figure 6, red curve), can be attributed to the eutectic (racemic conglomerate)
melting. All other observed endothermic peaks cannot belong to eutectics of enantiopure phases
with a different crystal structure. This means that the bimodal character of the melting curves of
racꢀ1.1 (olive) and racꢀ1.2 (magenta) samples indicates the presence of two racemic phases in
o
each sample. A phase with a lower melting point (ꢑ =76.5 C; henceforth referred to as αꢀrac)
ꢒ
∝
dominates in the racꢀ1.2 sample. Accordingly, the phase of the racemic polymorph with the
higher melting point (ꢑ =79.8 oC), prevailing in the racꢀ1.1 sample, will be denoted as βꢀrac.
ꢒ
ꢖ
ACS Paragon Plus Environment

Page 17 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 6. Experimental DSC traces of the polycrystalline diol 1 samples, obtained by solution
crystallization. The blue curve characterizes the melting of enantiopure Rꢀ1 sample, red curve ꢀ
racemic conglomerate, olive ꢀ racꢀ1.1, magenta ꢀ racꢀ1.2. Explanations in the text.
When the dry racꢀ1.1 crystals are stored at room temperature, the βꢀrac phase is stable for a
long time, but in the solvent (MTBE) medium at 20 °C, the plateꢀlike crystals break down fairly
quickly, and phase βꢀrac transforms to αꢀrac.
The melting thermograms are significantly more complicated for the racemic samples
obtained by melt crystallization. A full shape of the thermograms and a detailed description of
the experiments are given in ESI, Figures S3 and S4. Here we will outline the main results. Thus,
on the thermogram of racꢀ1 sample, which was previously melted and then cooled to 20 °C (Fig.
7, orange curve), a melting peak was observed at 79.8 °C, identical in its thermochemical
parameters to the melting peak of βꢀrac phase. It is preceded by fluctuations in the baseline,
indicating that this phase is formed during the recrystallization of some independent intermediate
phase (denoted as γꢀrac). If the melt, after cooling to 20 °C, was subjected to an additional
heating cycle of up to 62 °C to complete the recrystallization process, followed by cooling, then
on the thermogram of the sample with such history the only melting peak for βꢀrac phase was
observed (Fig. 7, olive curve).
ACS Paragon Plus Environment

Crystal Growth & Design
Page 18 of 31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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. Experimental DSC traces of the crystalline forms of diol 1 obtained by crystallization
from a melt. Green curve ꢀ a sample for which the melting cycle is commenced immediately
upon completion of crystallization at 50 °C; orange curve ꢀ a sample cooled after crystallization
to 20 °C; olive ꢀ a similar sample, additionally held at 60 °C to complete the recrystallization
before the melting cycle. Cyan curve ꢀ crystallization of the sample upon cooling, as well as a
fragment of the heating cycle corresponding to melting
The γꢀrac phase can be observed experimentally if the heating cycle begins immediately
after the crystallization of the racemic melt (Fig. 7, green curve). Under these conditions, its
melting peak at 71.5 °C becomes the main element of the thermogram, which makes it possible
to evaluate the thermochemical characteristics of this phase. The minor bimodal signal at higher
temperatures can be described by a combination of the βꢀrac and the conglomerate melting.
On the same thermogram, an additional minor endothermic peak (marked by cyan color) is
observed at 65.6 °C. It suggests that another relatively stable phase of the racemic composition,
δꢀrac, is present in the melting sample. Comparison of the crystallization enthalpy of the racemic
sample (the lowest cyan curve, ꢗꢘꢒ = 17.0 kJ molꢀ1) and the melting enthalpy of γꢀrac (ꢗꢘꢒ =
ꢙ
23.6 kJ molꢀ1) demonstrates a significant deficit in the enthalpy within crystallization/melting
cycle. This suggests that the δꢀrac phase is the first to form during the crystallization of the melt,
and then rapidly recrystallizes to the γꢀrac phase, which is accompanied by an exothermic effect.
The baseꢀline shift observed on the thermogram in the temperature range 55ꢀ65 °C confirms this
hypothesis. Then the observed minor endothermic peak can be attributed to the melting of the
ACS Paragon Plus Environment

Page 19 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
residual amounts of the δꢀrac phase, which did not have time to recrystallize into γꢀrac. Indirect
information obtained in additional DSC experiments (ESI, Figure S4) suggests that the δꢀrac
phase is a metastable ideal solid solution. Unfortunately, the short time of transformation of the
γꢀrac and δꢀrac phases makes it difficult to observe them under other conditions, except directly
in the DSC experiment, and accordingly the identification of their nature by independent
methods.
9
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
The final experimental thermochemical characteristics of all the identified crystalline forms
of compound 1 are shown in Table 2.
Table 2. DSC measured temperatures (ꢑ), enthalpies (ꢗꢘ), and entropies ꢗꢚ of fusion
for distinct crystalline diol 1 forms
ꢛꢜ
ꢝꢞꢜ
ꢝꢟꢜ
(°С) (kJ molꢀ1) (J molꢀ1 Kꢀ1)
Rꢀ1 102.8 35.1
93.2
90.5
94.1
82.3
68.3
50.2
(R+S)ꢀ1 81.1
αꢀracꢀ1 76.5
βꢀracꢀ1 79.8
γꢀracꢀ1 71.5
δꢀracꢀ1 65.6
32.1
32.9
29.1
23.6
17.0
Energetic preferences of the crystalline diol 1 forms. As noted above, the δꢀrac phase is
unstable and rapidly turns into the γꢀrac phase during the heating of racꢀ1 samples crystallized
from the melt. The γꢀrac phase also spontaneously recrystallizes into βꢀrac. The βꢀrac phase can
exist for a sufficiently long time in a dry state, but in a solvent medium at room temperature it
recrystallizes into αꢀrac. The relative thermodynamic stability of the αꢀrac phase and the racemic
conglomerate is not so obvious. In spite of the fact that the conglomerate has a higher melting
point, pure dry samples of the both phases at room temperature can exist for a long time (at least
a week) in an unchanged form. Even with active slurrying of the corresponding individual
samples in hexane medium during 24 hours at 20 °C they do not exhibit any appreciable
transformations. For example, the thermograms and IR spectra of the samples before and after
slurrying remain unchanged.
A quantitative representation about the relative thermodynamic stability of all the phases of
racemic composition revealed can be obtained by constructing the energy diagram of the system
under study, which reflects variations of the enthalpies, entropies, and Gibbs energies of these
phases as a function of temperature. For each phase the difference of the thermodynamic
ACS Paragon Plus Environment

Crystal Growth & Design
Page 20 of 31
1
2
3
4
functions by heating a crystalline sample from the initial temperature ꢛꢠ to the fusion
temperature ꢛꢜ and further to the final temperature of a melt ꢛꢡ can be represented as
ꢛꢡ
5
6
7
8
ꢜ
ꢦꢧꢨꢩꢪꢄ ꢆ
ꢣ ꢤꢥ
ꢛ ꢪꢛ , ꢛꢡ ꢫ ꢛ
ꢛꢠ
∆ꢞꢛꢠ/ꢛꢡ
ꢂ
ꢜ
ꢛ
ꢛꢡ
ꢢ
ꢜ
9
ꢦꢧꢨꢩꢪꢄ ꢆ
ꢨꢭꢄ ꢆ
ꢣ ꢤꢥ
ꢛ ꢪꢛ ꢬ ∆ꢞꢜ ꢬ ꢣ ꢤꢥ ꢛ ꢪꢛ , ꢛꢡ ꢮ ꢛ
ꢜ
ꢛꢠ
ꢛ
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
ꢛꢡ ꢤꢦꢥꢧꢨꢩꢪ
ꢜ
ꢄ ꢆ
ꢛ ꢪꢛ , ꢛꢡ ꢫ ꢛ
ꢣ
ꢛ
ꢛꢠ
∆ꢟꢛꢠ/ꢛꢡ
ꢂ
ꢜ ꢤꢦꢥꢧꢨꢩꢪ
∆ꢞꢜ
ꢛꢜ
ꢛꢡ ꢤꢨꢥꢭ
ꢛ
ꢢ
ꢜ
ꢄ ꢆ
ꢛ ꢪꢛ ꢬ
ꢄ ꢆ
ꢛ ꢪꢛ , ꢛꢡ ꢮ ꢛ
ꢣ
ꢬ ꢣ
ꢛ
ꢛ
ꢜ
ꢛꢠ
ꢛ
∆ꢯꢛꢠ/ꢛꢡ ꢂ ∆ꢞꢛꢠ/ꢛꢡ ꢰ ꢛꢡ ∙ ∆ꢟꢛꢠ/ꢛꢡ
.
Measurements of the effective heat capacities of the experimentally accessible samples by DSC
method were carried out with the use of samples of increased mass (~ 10 mg). The
corresponding thermograms are shown in Figure 8.
Figure 8. DSC curves presenting the effective heat capacity of racemic samples of different
phase compositions. Blue, red, olive, magenta and cyan lines correspond to Rꢀ1, racemic
conglomerate, βꢀrac, αꢀrac and δꢀrac phases. The dotted lines show the extrapolated values of
the heat capacity of individual phases in the region of solid/melt phase transitions. For δꢀrac
phase, the dashed line corresponds to the extrapolated values obtained from the crystallization
thermogram (see text).
Note that the heat capacities of the αꢀrac, βꢀrac, enantiopure phases and racemic
conglomerate in the temperature range from 20 °C to the melting temperatures turned out to be
close to each other within the experimental accuracy. For the δꢀrac phase, the heat capacity and
melting enthalpy were determined from the thermogram of crystallization of a preliminarily
ACS Paragon Plus Environment

Page 21 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
8
melted racemic sample. As noted above, γꢀrac phase was intermediate between δꢀrac and βꢀrac
and quickly turned into the latter. Therefore, it was not possible to measure experimentally its
heat capacity. To calculate the thermodynamic potentials of this phase and schematically display
them on the general energy diagram, the Cp value was used which was the mean between those
for δꢀrac and βꢀrac. We note that the variation of this parameter in the entire range between the
heat capacities of δꢀrac and βꢀrac does not introduce qualitative changes into the observed
picture.
9
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 absence of absolute values for standard thermodynamic potentials, the enthalpy and
the entropy of αꢀrac phase at 20 °С were taken as conventional zero. Obtained relationships are
graphically illustrated in Figure 9.
Figure 9. Dependences of the relative enthalpies (a) or entropies (b) of conglomerate (red), αꢀ
racꢀ1 (magenta), βꢀracꢀ1 (olive), γꢀracꢀ1 (green) and δꢀracꢀ1 (cyan) on temperature.
On the basis of these graphs, the dependences of the changes in the free Gibbs energies of
these phases on temperature have been plotted (Figure 10). It should be remembered that we
know only the relative changes in enthalpy and entropy. For this reason, the changes in Gibbs
energy were calculated as follows:
ꢧ
ꢧ
ꢿ
ꢿ
ꢹꢠ ꢤ
ꢹꢠ ꢤ
ꢛ
ꢇꢾ ꣀ
ꢱꢯꢛꢲ ꢂ ꢳꢞꢛꢲ ꢰ ꢞ
ꢺ ꢰ ꢛꢳꢟꢛꢲ ꢰ ꢟ
ꢄ
ꢛꢆ
ꢺ ꢂ ꢞꢲ ꢰ ꢛ ∙ ꢟꢲ ꢰ ꢳꢞαꢵꢻꢼꢽ ꢰ ꢛ ∙ ꢟαꢇꢵꢾꢻꢼꣀꢽꢺ,
ꢴꢵꢶꢷꢸ
ꢴꢵꢶꢷꢸ
which is equivalent to calculating the Gibbs energy of each phase relative to the hypothetical
system, the enthalpy and entropy of which coincides with the values of these parameters for the
ACS Paragon Plus Environment

Crystal Growth & Design
Page 22 of 31
1
2
3
4
5
6
7
8
αꢀrac phase at 20 °С. This does not affect the relative position of the curves and the
characteristic points on the graph, but, to avoid misinterpretation, it should be taken into account
that with such a transformation, the tangent of the tangent line for the ꢱꢯꢛꢲ plot is not reflecting
ꢧ
ꢹꢠ ꢤ
the proper entropy of this phase, but is equal to value ꢟꢛꢲ ꢰ ꢟ
. In particular, for ꢱꢯꢴꢛꢵꢶꢷꢸ
ꢴꢵꢶꢷꢸ
the tangent line at 20 °C is horizontal.
9
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
It is seen from Fig. 10 that in the entire accessible temperature range the γꢀrac and δꢀrac
phases are metastable and substantially inferior in energy to all other crystal modifications of
compound 1 known to us. Obviously, their formation during the crystallization process is due to
kinetic reasons, such as the ease and efficiency of the nucleation stage in the formation of δꢀrac
and the relatively low potential barrier of solidꢀphase recrystallization of δꢀrac into γꢀrac.
Further, γꢀrac in the solid phase is almost as easily transformed into βꢀrac. The ease of
implementation of such a cascade of solidꢀphase transformations at near to room temperatures
suggests that these three phases are characterized by a similar crystal structure. Under this
condition, the transformations do not need a significant rearrangement of the crystal packing, but
at each stage it is possible to obtain a noticeable gain in energy.
Figure 10. Dependences of the relative Gibbs free energies on temperature for different phases
of racemic composition. Red line – conglomerate, magenta ꢀ αꢀracꢀ1, olive - βꢀracꢀ1, green ꢀ γꢀ
racꢀ1, cyanꢀ δꢀracꢀ1, brown – stable (solid line) and metastable supercooled (dashed) racemic
liquid.
ACS Paragon Plus Environment

Page 23 of 31
Crystal Growth & Design
1
2
3
4
5
6
The αꢀrac and βꢀrac phases are energetically very close to each other and to racemic
conglomerate, so the energy relations of these three crystalline modifications are more complex.
Table 3 shows the calculated values of the free energies of the phase transitions ∆ꣁꣂꣂꢄꣃ꣄ꢆ
between the racemic conglomerate and other racemic phases at a temperature of 20 °C and at the
melting point of the corresponding phase. From these data and from Figure 10 it follows that the
thermodynamically preferred way of crystallization of racꢀ1 in the investigated temperature
range (from 20 °C to the melting temperature) is a conglomerate. At the melting temperature, the
αꢀrac phase is thermodynamically the least preferred among the three discussed ways of the
racemate crystallization. At the same time, this phase has the greatest value of the melting
enthalpy, and, accordingly, the lowest entropy in the crystalline state. An important consequence
of this fact is that as the temperature is lowered, the Gibbs potential of αꢀrac grows more slowly
than that of βꢀrac and also that of the conglomerate. At circa 53.5 °C this phase becomes more
stable than βꢀrac. Unfortunately, we failed to detect this formally enantiotropic transition
experimentally. We believe that the reason is the high potential barrier of the corresponding
solidꢀphase rearrangement of essentially different (Figures 3 and 4) crystal packings.
7
8
9
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
It also follows from Figure 10 that below the room temperature the αꢀrac phase should
become more stable than racemic conglomerate. It should be noted here that the very difference
in the calculated values of the free energy of αꢀrac and the conglomerate at the near room
temperatures (Table 3) is very small, ∆ꣁꢄꢇꣃꢾꣅ꣇꣄ꢆ/꣆< 100 J molꢀ1. This value is quite comparable
with the experimental error of DSC measurements, and therefore the question of the
thermodynamically stable state of the system at room temperature cannot be uniquely resolved
only on the basis of the data under consideration.
Table 3. Calculated values of the free energy of the phase transition ∆ꣁꣂꣂꢄꣃ꣄ꢆ between the
racemic conglomerate and other racemic modifications of compound 1
꣈ꢒ
ꢇꢾ꣇
∆ꣁ
,
∆ꣁ
,
ꣂꣂꢄꣃ꣄ꢆ
ꣂꣂꢄꣃ꣄ꢆ
Phase
kJ molꢀ1
0.097
kJ molꢀ1
0.41
αꢀrac
βꢀrac
γꢀrac
δꢀrac
0.497
1.66
2.70
0.12
0.85
1.35
To clarify the thermodynamically advantageous way of crystallization of racꢀ1 under these
conditions, we realized a direct experiment: equal amounts (~ 10 mg) of the racemic
conglomerate and the αꢀrac phase were mixed, suspended in hexane and subjected to vigorous
ACS Paragon Plus Environment

Crystal Growth & Design
Page 24 of 31
1
2
3
4
5
6
7
8
stirring for 4 hours. The initial state of each of the components and the final composition of the
system were controlled by DSC and IRS methods. It was found that, at the end of the mixing, the
precipitate consisted of a pure racemic conglomerate, there was no evidence of the presence of
the αꢀrac phase in it. This indicates that spontaneous recrystallization of the racemic αꢀrac
compound into a racemic conglomerate takes place under these conditions, and therefore it is the
conglomerate that remains the thermodynamically most advantageous state of the system at
20 °C and above.
9
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 a sense, the result of the slurrying experiment comes into conflict with the experimentally
established longꢀterm stability of dry αꢀrac samples. To clarify the situation, we undertook a
series of experiments on the crystallization of diol 1 samples of various enantiomeric
compositions. Figure 11 shows the results of these tests.
Figure 11. Thermograms of dry crystalline precipitates obtained by evaporating solutions of
compound 1 samples of different enantiomeric composition in MTBE. Blue color denotes an
enantiopure sample, x = 1.0; magenta stands for racemic sample, x = 0.50. Red dashed curve is
the thermogram of the sample of indubitable racemic conglomerate; solid red curves from top to
bottom correspond to the original compositions x = 0.55, 0.60, 0.70, 0.80, 0.95 (x are the mole
fractions of the predominant enantiomer).
Judging by the thermochemical parameters of the corresponding peak, the crystallization of
1 from a solution of a strictly racemic composition leads to the precipitation of a practically pure
αꢀrac phase. Obviously, such a result is determined by kinetic factors. But the presence of even a
small excess of one of the enantiomers in the system leads to the crystallization of the racemic
part of the precipitate exclusively in the form of a conglomerate. This is indicated by the
ACS Paragon Plus Environment

Page 25 of 31
Crystal Growth & Design
1
2
3
4
presence on the thermograms for x = 0.55÷095 of a characteristic peak of the corresponding
eutectic.
5
6
7
8
Racemic 3-(2,3-dimethylphenoxy)propane-1,2-diol, compound 1 could be resolved by
direct entrainment procedure. To a certain extent, our study of the phase behavior of
compound 1 was inspired by the propinquity of its chemical structure to the structure of the
known API xibenolol (Chart 1, structure 5). On the one hand, the ability of racꢀ1 to spontaneous
resolution during crystallization, discovered by us, opens the principle possibility of using direct
methods to obtain its individual enantiomers.34 In case of success, nonracemic 1 can serve as a
useful precursor in the synthesis of nonracemic 5. On the other hand, the complex phase
behavior of 1 and, in particular, the presence of several energyꢀclose (meta)stable crystalline
racemic compounds generally negatively affects the results of direct resolution.35,7 Nevertheless,
two facts, revealed as a result of our research, allowed us to hope for success. First, the racemic
conglomerate, at least above room temperature, proved to be thermodynamically more
advantageous than all polymorphic racemic compounds. Secondly, it is this form that crystallizes
primarily under the condition of initial enantiomeric enrichment of the sample.
9
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
With this in mind, we realized a successful demonstration experiment of racꢀ1 resolution by
entrainment. We used water as a solvent. The concentration of saturated racꢀ1 solution at 52 °C
determined by the polythermal method was Csat = 10.7 mg/ml, and for (R)ꢀ1 Csat = 5.3 mg/ml.
This means that under these conditions the Meyerhoffer rule is well observed. For these samples,
the inverse to dissolution spontaneous crystallization becomes noticeable at 39 °C in the first
case, and at 30 °C in the second.
For the demonstration experiment, we used slightly enriched (~ 10% ee) near racemic diol 1.
For such a sample, the concentration of the saturated solution was 8.5 mg/ml at 50 °C, and
spontaneous crystallization began at 29.5 °C. Crystallization, stimulated with enantiopure seed
crystals, was carried out at 37.5 °C. Other details of the process are given in the experimental
part. For monitoring the entrainment abilities of diol 1 during seed induced crystallization we
have used mother liquor enantiomeric composition measured by HPLC (Figure 12).
ACS Paragon Plus Environment

Crystal Growth & Design
Page 26 of 31
1
2
3
4
5
6
7
8
9
Figure 12. Mother liquor enantiomeric excess vs time in full cycle of preferential crystallization
of diol 1 (red curve ꢀ 1ꢀst run; blue ꢀ 2ꢀnd run). Solid circles indicate the values of the
enantiomeric excess at which the process was interrupted. Time for technical operations between
runs is omitted.
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 follows from the figure, during the crystallization process, the primary enantiomeric excess of
solute (9.8%) decreased to zero, changed sign, and raised up to 9.5% in the opposite direction. At
this point the process was interrupted. After separation of the precipitate, a missing amount of
the racemate was added to the mother liquor. In the second run, the enantiomeric excess of the
mother liquor crosses zero and reaches only 5.1% as maximum value. Therefore, for maintaining
the entrainment procedure as a robust multiꢀcyclic process, to restore the starting enantiomeric
enrichment of the nearꢀracemic diol to about 10% after separation of the enantiomeric
precipitate, a small amount of scalꢀ1 must be added to the mother liquor along with the racemate.
CONCLUSIONS
It has been established that the chiral phenyl glycerol ether bearing two methyl groups in the
ortho and meta positions of the phenyl ring, 3ꢀ(2,3ꢀdimethylphenoxy)propaneꢀ1,2ꢀdiol 1,
crystallizes from the racemic feed medium with the formation of at least five crystalline
modifications: αꢀrac, βꢀrac, γꢀrac, δꢀrac racemic compounds and racemic conglomerate (R+S).
The first two and the last modifications are formed during crystallization from solutions and
form sufficiently stable in dry form monocrystals of satisfactory quality. The nature of these
modifications has been revealed by Xꢀray diffraction analysis. Additional metastable phases γꢀ
rac and δꢀrac are formed during the crystallization of racꢀ1 from melt. This fact has been
detected by DSC method. The thermochemical characteristics (Tf, ꢁHf, ꢁSf and Cp(T)) have been
measured for all the modifications found and for the racemic liquid phase, and graphs of the
dependence of free Gibbs energy on temperature have been constructed on this basis.
It has been shown that the δꢀrac phase is unstable and rapidly turns into γꢀrac phase during
heating of racꢀ1 samples obtained from the melt. Further, γꢀrac phase also spontaneously
recrystallizes into βꢀrac. The βꢀrac phase can exist for a long time in a dry state, but in a solvent
medium it is converted to αꢀrac already at room temperature. From the dependences ꢁG.vs.T it
follows that the βꢀrac and αꢀrac phases are capable of enantiotropic transition at a temperature of
about 50 °C, however, we failed to detect that transition for dry samples.
At temperatures above 20 °C, the most thermodynamically stable crystal modification of
racꢀ1 is the racemic conglomerate (R+S). Nevertheless, in the case of the crystallization of racꢀ1
ACS Paragon Plus Environment

Page 27 of 31
Crystal Growth & Design
1
2
3
4
5
6
7
from solutions at room temperature, the racemic compound αꢀrac is formed, which is stable both
in dry form and in solutions. But even a small excess of one of the enantiomers in the mother
liquor leads to the fact that the entire racemic part of the crystallizing sample is precipitated in
the form of a racemic conglomerate.
8
9
Taking into account that above room temperature the conglomerate is thermodynamically
more favorable than all polymorphic racemic compounds, and that this particular form
crystallizes from solutions under the condition of initial enantiomeric enrichment of the sample,
we realized a successful demonstration experiment on the resolution of racꢀ1 into individual
enantiomers by the method of entrainment.
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
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI:
Synthetic procedures, details of crystal packing in Sꢀ1 crystals, and additional DSC experiments.
AUTHOR INFORMATION
Corresponding Author
* Tel.: +7 843 2727393; fax: +7 843 2731872.
Eꢀmail address: baa@iopc.ru (A.A. Bredikhin)
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval
to the final version of the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The authors thank the Russian Fund of Basic Research and the Government of the Republic of
Tatarstan for partial financial support (Grant No. 15ꢀ43ꢀ02238).
ACS Paragon Plus Environment

Crystal Growth & Design
Page 28 of 31
1
2
3
4
REFERENCES
(1) Bernstein, J. Polymorphism in molecular crystals; Oxford University Press: New York,
5
6
7
2002, 410 рp.
(2) Bernstein, J. Polymorphism − a perspective. Cryst. Growth Des. 2011, 11, 632–650.
(3) Polymorphism in pharmaceutical solids; Brittain, H. G., Ed.; Informa Healthcare USA:
New York, 2009, 480 pp.
8
9
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
(4) Polymorphism: in the pharmaceutical industry; Hilfiker, R., Ed.; WileyꢀVCH: New
York, 2006, 433 pp.
(5) Aitipamula, S.; Chow, P. S.; Tan, R. B. H. Conformational polymorphs of a muscle
relaxant, metaxalone. Cryst. Growth Des. 2011, 11, 4101–4109.
(6) Stephenson, G. A.; Kendrick, J.; Wolfangel, C.; Leusen, F. J. J. Symmetry breaking:
polymorphic form selection by enantiomers of the melatonin agonist and its missing polymorph.
Cryst. Growth Des. 2012, 12, 3964ꢀ3976.
(7) Brandel, C.; Amharar, Y.; Rollinger, J. M.; Griesser, U. J.; Cartigny, Y.; Petit, S.;
Coquerel, G. Impact of molecular flexibility on double polymorphism, solid solutions and chiral
discrimination during crystallization of diprophylline enantiomers. Mol. Pharm. 2013, 10, 3850–
3861.
(8) Braga, D.; Chelazzi, L.; Grepioni, F.; Nanna, S.; Rubini, K.; Curzi, M.; Giaffreda, S.;
Saxell, H. E.; Bratz, M.; Chiodo, T. Imazamox: a quest for polymorphic modifications of a chiral
and racemic herbicide. Cryst. Growth Des. 2014, 14, 1430–1437.
(9) Gubaidullin, A. T.; Samigullina, A. I.; Bredikhina, Z. A.; Bredikhin, A. A. Crystal
structure of chiral orthoꢀalkyl phenyl ethers of glycerol: true racemic compound, normal, false
and anomalous conglomerates within the single fiveꢀmembered family. CrystEngComm 2014,
16, 6716–6729.
(10) Carvalho, P. S.; Ellena, J.; Yufit, D. S.; Howard, J. A. K. Rare case of polymorphism in a
racemic fluoxetine nitrate salt: phase behavior and relative stability. Cryst. Growth Des. 2016,
16, 3875–3883.
(11) Fayzullin, R. R.; Zakharychev, D. V.; Gubaidullin, A. T.; Antonovich, O. A.; Krivolapov,
D. B.; Bredikhina, Z. A.; Bredikhin, A. A. Intricate phase behavior and crystal structure features
of chiral paraꢀmethoxyphenyl glycerol ether forming continuous and partial solid solutions.
Cryst. Growth Des. 2017, 17, 271–283.
(12) Levkin, P. A.; Torbeev, V. Y.; Lenev, D. A.; Kostyanovskii, R. G. Homoꢀ and
heterochirality in crystals. Top. Stereochem. 2006, 25, 81–134.
ACS Paragon Plus Environment

Page 29 of 31
Crystal Growth & Design
1
2
3
4
5
6
(13) Navare, P. S.; MacDonald, J. C. Investigation of Stability and Structure in Three
Homochiral and Heterochiral Crystalline Forms of 3ꢀPhenyllactic Acid. Cryst. Growth Des.
2011, 11, 2422–2428.
(14) Zhang, C. C.; Matzger, A. J. A newly discovered racemic compound of pioglitazone
hydrochloride is more stable than the commercial conglomerate. Cryst. Growth Des. 2017, 17,
414–417.
7
8
9
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
(15) Bredikhin, A. A.; Bredikhina, Z. A.; Zakharychev, D. V. Crystallization of chiral
compounds: thermodynamical, structural and practical aspects. Mendeleev Commun. 2012, 22,
171–180.
(16) Bredikhin, A. A., Bredikhina, Z. A. Stereoselective crystallization as a basis for singleꢀ
enantiomer drug production. Chem. Eng. Technol. 2017, 39, DOI: 10.1002/ceat.201600649
(17) Bredikhin, A. A.; Zakharychev, D. V.; Bredikhina, Z. A.; Gubaidullin, A. T.; Fayzullin,
R. R. Crystal structure and phase behavior of the tolyl glycerol ethers. From the conglomerate
former to the chiralityꢀdriven nanogelator. CrystEngComm 2012, 14, 211–222.
(18) Bredikhina, Z. A.; Kurenkov, A. V.; Krivolapov, D. B.; Bredikhin, A. A. Stereoselective
crystallization of 3ꢀ(2,6ꢀdimethylphenoxy)propaneꢀ1,2ꢀdiol: preparation of the singleꢀenantiomer
drug mexiletine. Tetrahedron: Asymmetry 2015, 26, 577–583.
(19) Sayyed, I. A.; Thakur, V. V.; Nikalje, M. D.; Dewkar, G. K.; Kotkar, S. P.; Sudalai, A.
Asymmetric synthesis of aryloxypropanolamines via OsO4ꢀcatalyzed asymmetric
dihydroxylation. Tetrahedron 2005, 61, 2831–2838.
(20) Ludwig, B. J.; West, W. A.; Currie, W. E. Muscleꢀparalyzing compounds related to
mephenesin. J. Am. Chem. Soc. 1952, 74, 1935–1939.
(21) The Merck Index, 14th ed., Ed. M. J. O’Neil, Merck and Co., Inc., Whitehouse Station,
NJ, USA, 2006, entry 10075.
(22) Kong, X.ꢀD.; Ma, Q.; Zhou, J.; Zeng, B.ꢀB.; Xu, J.ꢀH. A smart library of epoxide
hydrolase variants and the top hits for synthesis of (S)ꢀβꢀblocker precursors. Angew. Chem. Int.
Ed. 2014, 53, 6641–6644.
(23) Sheldrick, G. M. SADABS, Program for Empirical Xꢀray Absorption Correction; Brukerꢀ
Nonius: Delft, the Netherlands, 2004.
(24) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr. A 2008, 64, 112–122.
(25) Farrugia, L. J. WinGX suite for smallꢀmolecule singleꢀcrystal crystallography. J. Appl.
Cryst. 1999, 32, 837–838.
(26) Macrae, C. F.; Edgington, P.R.; McCabe, P.; Pidcock, E.; Shields, G.P.; Taylor, R.;
Towler, M.; van de Streek, J. Mercury: visualization and analysis of crystal structures. J. Appl.
Crystallogr. 2006, 39, 453–459.
ACS Paragon Plus Environment