Insight into structural demand for cold crystallization of a small molecule. A case study for schiff base compounds that exhibit prototropic tautomerization

Article abstract of DOI:10.1246/bcsj.20170386

To elucidate the relationship between prototropic tautomerism and the cold crystallization phenomenon, o-hydroxy aryl Schiff base compounds derived from three different toluidine isomers were examined. Comparing their thermal behavior, we confirmed that c

Full text of DOI:10.1246/bcsj.20170386

Advance Publication Cover Page
Insight into a structural demand for cold crystallization of small molecule. A case study for
Schiff base compounds that exhibit prototropic tautomerization
Katsunori Iwase,* Yasuhiro Toyama, Isao Yoshikawa, Yasuhisa Yamamura,
Kazuya Saito, and Hirohiko Houjou*
Advance Publication on the web February 22, 2018
doi:10.1246/bcsj.20170386
© 2018 The Chemical Society of Japan
Advance Publication is a service for online publication of manuscripts prior to releasing fully edited, printed versions. Entire
manuscripts and a portion of the graphical abstract can be released on the web as soon as the submission is accepted. Note that the
Chemical Society of Japan bears no responsibility for issues resulting from the use of information taken from unedited, Advance
Publication manuscripts.

Insight into a structural demand for cold crystallization of small molecule. A case
study for Schiff base compounds that exhibit prototropic tautomerization
Katsunori Iwase,*1 Yasuhiro Toyama,1 Isao Yoshikawa,2 Yasuhisa Yamamura,3 Kazuya Saito3 and Hirohiko Houjou*2
1DENSO CORPORATION, 500-1 Minamiyama, Komenoki-cho, Nisshin, Aichi 470-0111, JAPAN
2Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, JAPAN
3Department of Chemistry, Faculty of Pure and Applied Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8571, JAPAN
E-mail: katsunori_iwase@denso.co.jp
Katsunori Iwase
Katsunori Iwase received Ph.D. from University of Tsukuba in 2015. He works for DENSO CORPORATION and
engaged in functional material research at Advanced Research and Innovation Center.
Hirohiko Houjou
Hirohiko Houjou received his Ph.D. degree from Tokyo Institute of Technology in 1998. He worked for National
Institute of Advanced Interdisciplinary Research, National Institute of Advanced Industrial Science and Technology,
and moved to the present affiliation in 2003. Specialized in physical organic chemistry, he studies the properties of
various molecular assemblies using solid-state spectroscopy and quantum chemical calculations.
Abstract
To elucidate the relationship between prototropic
maintaining the heat storage state at room temperature, and 3)
releasing heat when and where it is required.
tautomerism and the cold crystallization phenomenon,
o-hydroxy aryl Schiff base compounds derived from three
different toluidine isomers were examined. Comparing their
thermal behavior, we confirmed that cold crystallization
occurred to give a different crystalline phase compared to the
initial state. To elucidate the mechanisms of the cold
crystallization of the materials, their thermal, structural, and
energetic properties were investigated. The DFT calculation of
the molecules in their OH- and NH-forms suggested that the
NH-form has higher molecular flexibility, which may prevent
normal crystallization upon cooling, compared to the OH-form.
Spectroscopic analyses implied that the molecules are present
as a mixture of the OH- and NH-forms, and that they efficiently
crystallize when a preferable OH / NH ratio is achieved. For
this molecular system, we concluded that the equilibrium
between NH- and OH-forms controls the supercooling of the
liquid as well as nucleation and crystal growth.
R2
R2
R3
R3
R1
R1
f1
N
N
H
H
O
O
b) NH-form
: R1=Me R2=H R3=H
a) OH-form
1
2
3
:
:
R1=H
R1=H
R2=Me R3=H
R2=H R3=Me
Figure 1. Molecular structures of the Schiff base-compounds
studied.
Cold crystallization is an interesting phenomenon because
the crystallizing material shows exothermic anomalies during
the heating process. As a prerequisite, cold crystallization
materials should show supercooling during the cooling process
to maintain a molten or glass state, so that the crystallization
can occur in the subsequent heating process. Since these
prerequisites match the thermal behavior required from heat
storage materials, cold crystallization materials are potentially
applicable to heat storage. To date, there have been several
reports of polymers4-6), ionic liquids7), organic matter8-11), and
metal complexes12-16) as cold crystallization materials. However,
there are many aspects that remain unclear with respect to the
mechanism of cold crystallization.
1. Introduction
Primary energy sources like petroleum and natural gas are
ultimately discharged in the form of product gases and waste
heat after consumption in automobiles, factories, and
households. The effective use of this waste heat can contribute
to improving the energy efficiency of mechanical and electric
systems, and thus reducing their CO2 emissions. To achieve this
goal, technologies for heat storage,1) heat exchange,2) heat
extraction,3) and related processes must be improved. Among
these processes, especially for automobiles, heat storage is
highly desired because of the importance of eliminating the lag
time between the initial demand and eventual supply of heat.
The ideal behavior of heat storage materials would include 1)
storing heat when and where surplus heat is generated, 2)
Recently, Iwase et al. reported that certain Schiff base
complexes exhibit cold crystallization behavior, and attempted
to elucidate their cold crystallization mechanism. They
proposed two design guidelines for cold crystallization
materials: 1) to introduce a flexible moiety in the molecule to

increase the degree of freedom of molecular motion17,18), and 2)
allow several isomers that are interconvertible to coexist in the
1-(4-methylphenyl)iminomethylnaphthalene : 3
Yellow crystals (4.70 g, 90%), m.p. 137.6°C, IR(KBr)
1622 cm-1 (C=N); MS (FAB+) m/z 262.1 (calcd. for M+H+
262.13); Analysis calcd. for C18H15NO: C, 82.73; H, 5.79; N,
liquid (molten) state to disturb the homogeneity of the system19)
In accordance with these guidelines, Achira et al. reported that
a two-component molecular system can also cause cold
crystallization.20)
.
1
5.36; found: C, 82.57; H, 5.80; N, 5.20; H NMR (CDCl3)  =
2.40 (s, 3H), 7.06 (d (J = 9.8 Hz), 1H), 7.25-7.34 (mult, 5H),
7.51 (dt (J = 8.0 Hz, 1.3 Hz), 1H), 7.70 (d (J = 8.0 Hz), 1H),
7.78 (d (J = 8.8 Hz), 1H), 8.08 (d (J = 8.4 Hz), 1H), 9.31 (d (J
= 4.9 Hz), 2H), 15.62 (d (J = 4.9 Hz), 1H); 13C NMR (CDCl3) 
=21.1, 108.6, 118.8, 120.0, 122.7, 123.4, 127.2, 128.0, 129.4,
130.3, 133.3, 136.5, 136.7, 142.2, 153.7, 171.1.
In
the
present
study,
we
focused
on
2-hydroxy-1-naphthaldehyde Schiff bases (1, 2 and 3) of o-, m-,
and p-toluidine, respectively. These compounds show
prototropic tautomerism (Figure 1) by intramolecular proton
transfer between the oxygen and nitrogen atoms, resulting in
equilibrium between the keto-enamine (NH-form) and
enol-imine form (OH form).21,22) The OH/NH-form ratio varies
depending on molecular and crystal structures, temperature,
and photo-irradiation23). Consequently, these compounds follow
the second guideline, and are potential candidates for cold
crystallization. However, there have been no studies that have
focused the effect of tautomerization on cold crystallization.
The purpose of this study is to observe cold crystallization of
the three Schiff base compounds and to elucidate their
mechanisms from the viewpoint of structural chemistry.
2.2 Crystallography. For single crystal structure analysis,
1- and 3- were picked up from the as-prepared sample,
2·MeOH was recrystallized from methanol, 3- was
recrystallized from ethanol. X-ray measurements on single
crystals were carried out using diffractometers utilizing
imaging plate with monochromated MoKα radiation (0.71075
Å). A Rigaku diffractometer (Tokyo, Japan) used was VariMax
DW with Saturn. The structures were solved by the direct
method (SHELXS2013 for 1-, 2·MeOH and 3-, and il
Milione for 3-)25,26) and refined by the full-matrix
least-squares method on |F|2. The non-hydrogen atoms were
refined anisotropically. Hydrogen atoms were refined using the
riding model. Absorption correction was applied using an
empirical procedure. All calculations were performed using the
CrystalStructure crystallographic software package except for
refinement, which was performed using SHELXL2016 for 1-,
2·MeOH, and 3-, or SHELXL97 for 3-.25) Crystallographic
data have been deposited with Cambridge Crystallographic
Data Centre: Deposition numbers CCDC-1580131 to 1580134
to compounds 1, 2, and 3 (including polymorph). Copies of the
2. Experimental
2.1 Synthesis. The Schiff base compounds 1–3 were
prepared according to a reported procedure for condensation of
salicylaldehyde and aniline.17-19) In a typical example, to a hot
ethanol solution (50 mL) of 2-hydroxy-1-naphthaldehyde (3.44
g, 20.0 mmol) was added p-toluidine (2.14 g, 20.0 mmol). The
mixture was stirred at 70°C for 1 h, and then concentrated to
about 40 mL with vacuum evaporator, to afford yellow
crystalline precipitate. The product was collected by filtration,
washed with ethanol, and dried under reduced pressure. All the
chemicals and solvents were purchased from Tokyo Chemical
Industry and used without further purification.
The products were characterized by IR spectroscopy
using a JEOL FT/IR-420, elemental analysis (CHN) using a
Fusion Instruments EA1108, 1H and 13C NMR spectroscopy
using a JEOL ECS-400 (400 MHz for 1 H). As for the
compounds 1-3, the resonance signals were successfully
assigned by assuming their NH-form shown in Figure 1.24) The
melting points were determined as the onset temperature on the
first heating run of DSC.
data
can
be
obtained
free
of
charge
via
http://www.ccdc.cam.ac.uk/conts/retrieving.html (or from the
Cambridge Crystallographic Data Centre, 12, Union Road,
Cambridge, CB2 1EZ, UK; fax: +44 1223 336033; e-mail:
deposit@ccdc.cam.ac.uk).
1-: C18H15NO, Mw
=
261.31, yellow needle,
orthorhombic, space group P212121, a = 7.337(4), b = 12.402(6),
c = 14.367(7) Å, V = 1307.3(11) Å3, Z = 4, Dcalcd = 1.328 g
cm-3,
T = 93(2) K, 10334 reflections collected, 2996
independent (Rint = 0.0673), GOF = 1.024, R1 = 0.058, wR2 =
0.120 for all reflections.
1-(2-methylphenyl)iminomethylnaphthalene : 1
Yellow crystals (yield 4.24 g, 81%); m.p. 118.7°C;
IR(KBr) 1622 cm-1 (C=N); MS (FAB+) m/z 262.1 (calcd. for
M+H+ 262.13); Analysis calcd. for C18H15NO: C, 82.73; H,
5.79; N, 5.36; found: C, 82.53; H, 5.79; N, 5.20; 1H NMR
(CDCl3)  = 2.49 (s, 3H), 7.08 (d (J = 8.7 Hz), 1H), 7.19-7.22
(mult, 1H), 7.30-7.35 (mult, 4H), 7.52 (dt (J = 7.8 Hz, 1.6 Hz),
1H), 7.71 (d (J = 7.8 Hz), 1H), 7.80 (d (J = 9.1 Hz), 1H), 8.10
(d (J = 9.1 Hz), 1H), 9.31 (d (J = 5.4 Hz), 1H), 15.76 (d (J =
4.9 Hz), 1H); 13C NMR (CDCl3)  = 18.2, 108.8, 117.1, 118.7,
122.7, 123.5, 126.5, 127.1, 127.3, 128.1, 129.4, 130.9, 131.1,
133.3, 137.0, 143.5, 153.7, 171.8.
2·MeOH: C18H15NO·CH3OH, Mw = 293.35, yellow
needle, orthorhombic, space group P212121, a = 5.823(3) Å, b =
15.804(6) Å, c = 16.545(7) Å, V = 1522.6(12) Å3, Z = 4, Dcalcd
= 1.280 g cm-3, T = 93(2) K, 10670 reflections collected, 2766
independent (Rint = 0.0881), GOF = 1.073, R1 = 0.066, wR2 =
0.116 for all reflections.
3-: C18H15NO, Mw
=
261.32, brown needle,
orthorhombic, space group Pna21, a = 19.797(10) Å, b =
13.846(7) Å, c = 4.877(2) Å, V = 1336.8(11) Å3, Z = 4, Dcalcd
=
1.298 g cm-3, T = 293 K, 2870 reflections collected, 1266
independent (Rint = 0.0315), GOF = 1.058, R1 = 0.063, wR2 =
0.165 for all reflections.
1-(3-methylphenyl)iminomethylnaphthalene : 2
Yellow crystals (4.43 g, 85%), m.p. 92.4°C, IR(KBr) 1625
cm-1 (C=N); MS (FAB+) m/z 262.1 (calcd. for M+H+ 262.13);
Analysis calcd. for C18H15NO: C, 82.73; H, 5.79; N, 5.36;
found: C, 82.48; H, 5.77; N, 5.22; 1H NMR (CDCl3)  = 2.43 (s,
3H), 7.05 (d (J = 9.7 Hz), 1H), 7.10 (d (J = 7.5 Hz), 1H),
7.15-7.17 (mult, 2H), 7.30-7.36 (mult, 2H), 7.50 (dt (J = 7.6 Hz,
1.6 Hz), 1H), 7.69 (d (J = 7.5 Hz), 1H), 7.77 (d (J = 9.7 Hz),
1H), 8.07 (d (J = 8.6 Hz), 1H), 9.27 (d (J = 5.4 Hz), 15.53 (d (J
= 4.8 Hz), 1H); 13C NMR (CDCl3) = 21.5, 108.6, 117.1, 118.7,
120.7, 122.9, 123.5, 127.1, 127.3, 128.1, 129.4, 129.5, 133.4,
137.0, 139.7, 144.4, 153.7, 171.9.
3-: C18 H15NO, Mw = 261.31, yellow needle, monoclinic,
space group P21/n, a = 4.811(11) Å, b = 20.31(5) Å, c =
13.63(3) Å,  = 93.19(4)°, V = 1330(5) Å3, Z = 4, Dcalcd = 1.305
g cm-3, T = 93.19(4) K, 7806 reflections collected, 2399
independent (Rint = 0.1867), GOF = 1.188, R1 = 0.207, wR2 =
0.543 for all reflections.
2.3 Thermal Analyses. Thermal analyses were conducted
for 1–3 samples as prepared. The phase transition behavior of
the complexes was investigated by differential scanning
calorimetry (DSC) using a Mettler Toledo DSC-1. Open pans
of aluminum were used under a nitrogen gas flow (50 mL

min-1). The scanning rate was ±15.0˚C min-1. Transition
temperatures were determined as onset temperatures.
fiber optic spectrophotometer. Light from a xenon lamp was
guided into the condenser lens through a quartz optical fiber
and was focused on a selected 50-m -diameter area of the
sample.
Simultaneous analyses of powder X-ray diffraction and
DSC (XRD-DSC) were conducted for 1–3 samples as prepared,
using a measurement system composed of a Rigaku SmartLab
X-ray diffractometer and a Rigaku DSC. The sample was
heated/cooled at a rate of ±15 °C min-1 under N2 atmosphere．
During heating/cooling process, the XRD data were recorded
for 2θ range of 5–50° at a scan speed of 45° min-1. This means
that one strip of XRD corresponds to a data during the
temperature change of 15 °C, unless otherwise noted.
Monochromated CuKα radiation was used, and the voltage and
current of the generator were 45 kV and 200 mA, respectively.
2.4 Solid-state UV-Vis spectroscopic analysis.
Solid-thin-film absorption spectra were measured for the 1, 2
and 3 smeared onto glass slides at a sample thickness that
yielded a maximum absorbance of about one. Each sample was
placed on the stage of a conventional optical microscope
equipped with a UV-transmitting objective lens and a portable
Each spectrum, of which horizontal axis was converted to
wavenumber (region of 19–27 ×103 cm–1), was resolved to five
gaussian components by a non-linear curve-fitting method
(Marquardt method)27) coded into a home-made Wave-SPEC
program.28) The initial values of the centers of the curves were
set around 20, 21, 22, 24, 25 ×103 cm–1. The examples of the
resolution for 1–3 are given in the supplementary material
(Figure S5–S7), and the parameters for the fitting were
summarized in Table S1, S2, S3. After the peak resolution, the
subarea summed over three curves at higher wavenumber, and
that summed over the rest were calculated. The ratio of the
former to the total peak area was defied as ASW (area of
short-wavelength component), which was tentatively adopted
as a measure of the abundance of OH-form.
Heat flow / arbitrary unit
(A)
Endo.
Experimental
xiii
xii
28.0 °C
20
15
10
5
50.3 °C
86.1 °C
74.3 °C
98.2 °C
xi
x
122.4 °C
146.7 °C
160.6 °C
136.4 °C
111.9 °C
87.4 °C
ix
viii
vii
vi
v
118.7 °C
iv
iii
ii
62.5 °C
37.9 °C
19.9 °C
i
0
0
0
10
10
20
30
30
40 150 100 50
0
T / °C
(B)
Simulated
1-
1-
20
40
2q / °
Figure 2. XRD-DSC results (A) of 1 obtained in first heating and subsequent cooling processes compared with simulated patterns
(B) based on crystal structures (1- and 1-) determined independently. Left, XRD (5 ≤ 2q / ° ≤ 35); right, DSC (black line,
temperature; green line, heat flow).

Heat flow / arbitrary unit
(A)
Endo.
Experimental
20
24.8 °C
47.5 °C
71.2 °C
95.6 °C
xii
xi
x
15
ix
viii
vii
vi
118.8 °C
107.6 °C
10
72.0 °C
83.0 °C
58.0 °C
33.1 °C
v
5
iv
iii
ii
42.6 °C
7.7 °C
-17.1 °C
0
19.8 °C
i
0
0
10
10
20
30
40 100 50
0
T / °C
(B)
3-
Simulated
3-
20
30
40
2q / °
Figure 3. XRD-DSC results (A) of 3 obtained in second heating and subsequent cooling processes compared with simulated patterns
(B) based on crystal structures (3- and 3-) determined independently. Left, XRD (5 ≤ 2q / ° ≤ 35); right, DSC (black line,
temperature; green line, heat flow).
3. Results
pseudopolymorph including 1 eq. of a methanol molecule,
which are labeled as 2·MeOH. Upon heating, 2·MeOH melted
at 61-62 °C, which is near the boiling point of methanol,
recrystallized at 71 °C, and finally melted at 94 °C, the same
temperature as 2-. The single crystals of the two polymorphs
of 3, namely 3- and 3-, were obtained through the
concentration of the reaction mixture and by recrystallization
from ethanol, respectively. Each sample was composed of
yellow needles with various thickness, which we could not
judge as polymorphs with naked eyes. Through optical
microscopy we observed that the major crystals of the
as-prepared sample melted at 137 °C, while fine needles of the
latter sample exhibited solid-to-solid phase transition at
131–135 °C, and then melted at 137 °C along with thicker
needles. These observations cannot rule out the possibility that
the both samples are a mixture of 3- and 3- phases. As
discussed later, the powder XRD pattern of the as-prepared
sample was fairly reproduced from the single crystal structure
of 3- (m.p. 137 °C), and therefore we concluded that the
metastable 3- (m.p. 131 °C) was a crystallographic impurity
included in the as-prepared sample.
3.1 Crystal Structures of 1 – 3. The Schiff base
compounds 1–3 were readily obtained by simple condensation
reaction in an ethanol solution. Although these compounds
have been widely used for various metal complexes, we could
not find any comprehensive analytical data of their structure
and related properties in the literature. Also, there are
surprisingly few reports on the crystal structures of 1-3 as free
bases. As for 1, two polymorphs were independently reported,
CCDC11615929) (SIPFEC) and 99811830) (SIPFEC01).
According to those reports, the former adopts the OH-form and
melts at 128 °C, while the latter was assigned to the
zwitterionic NH-form (no data for its melting point). In the
present study, we found that product 1 (m.p. 118 °C) was
crystallized in a form identical to the previously reported
NH-form, and hereafter this phase is termed 1- We
tentatively call another phase 1-,29) although we were not able
to identify this type of crystal (Figure 2). As for 2, we were not
able to obtain a single crystal of good quality, without inclusion
of solvent molecules. As prepared, compound 2 was a powdery
microcrystalline solid that melts at 94 °C, which we call 2-.
Recrystallization of 2- with methanol afforded
a

For thermal analyses (see the section below), we used the
1-, 2-, 3- samples as the initial states. In spite of our
attempts to pick up crystal chips from the aluminum sample
pans that underwent appropriate heating/cooling history in the
DSC apparatus, the crystal structures of the other phases were
not clarified. The crystal data, together with other relevant
information regarding the structural determination, are listed in
Table S4.
3.2 Phase Assignment by Simultaneous XRD-DSC. First,
the thermal behavior of the as-prepared crystalline samples 1-,
2- and 3- were examined by DSC (for more details see the
Supporting Information). These samples exhibited totally
different behavior from each other, especially during
cooling/heating processes after their melting: 1 recrystallized in
the cooling process (Figure S1); 2 remained in liquid or glass
identical to the pattern calculated for 3-. It is worth noting that
the cold crystal showed double melting peaks around 132 and
136 °C (as peak top temperatures). By careful inspection of the
XRD-DSC results with a scan rate of 1 °C min–1 (Figure S4),
we found that the sample underwent another solid-solid
transformation at around 131–135 °C to form 3-, which
finally melted at 137 °C.
3.3 Solid-state UV-Vis spectroscopic analysis. The
UV-Vis absorption spectra of the initial crystals are shown in
Figure 4. The UV-Vis spectra of both 1 and 3 gave peaks at
around 380, 410, 440, and 470 nm, whereas that of 2 had peaks
at about 450 and 490 nm, with weak shoulder-peaks in the short
wavelength region between approximately 380 and 430 nm.
This difference in the absorption profile suggests a difference
in the tautomer ratio since the peaks around 380–420 nm arise
from the OH-form, while the peaks between 420–500 nm can
be attributed to the NH-form. The small table inserted in Figure
4 shows the values of ASW, the ratio of the short-wavelength
components (defined in the experimental section), which can be
a measure of the abundance of OH-form to the total amount of
the tautomers.
state without crystallization (Figure S2);
3
exhibited
supercooling and subsequent cold crystallization (Figure S3).
Because our present interest is in examining potential heat
storage materials, we hereafter focus on the comparison
between the phase transition behavior of 1 and 3.
Figure 2 shows the XRD-DSC results for 1, along with a
powder X-ray diffraction (PXRD) pattern calculated from the
single crystal data of 1-. The PXRD patterns i–iv, before
melting at 118.7 °C, are almost identical to the calculated
pattern of 1-. During the cooling process, the DSC curve
showed an exothermic anomaly at 86.6 °C, from which the
PXRD patterns (xi–xiii) showed different peaks from that of
1-. These patterns were also different form the patterns
calculated for the 1- crystal, so we tentatively denote this
recrystallized phase as 1-, although its crystal structure is still
unknown. After cooling to -10 °C, the sample was reheated.
The 1- crystal melted at 115.2 °C, and thereafter the DSC
curves showed essentially iterative profiles.
ASW
0.61
0.29
0.68
1
2
3
1
2
3
400
500
600
Figure 3 shows the XRD-DSC results for 3, along with
PXRD patterns calculated for 3- and 3-. The initial state was
identified as 3-, based on the comparison of its PXRD pattern
(i) with the calculated patterns Although there were only slight
differences between the patterns of 3- and 3-, they are
discriminable from the weak reflections at 11 and 14° of 2q.
The sample was heated above its melting point (137.6 °C), and
then cooled to -20 °C to be supercooled, showing a glass
transition around -1 °C on the DSC curve. During the
subsequent heating process, the PXRD patterns of the glass
(below -5 °C, pattern ii) and supercooled liquid (up to 40 °C,
pattern iii) both showed a halo corresponding to amorphous
structure. Upon further heating, an exothermic anomaly was
observed at 42.6 °C (Tcc-1), and simultaneously a set of sharp
peaks emerged in the XRD pattern (iv–vi), which clearly
demonstrates cold crystallization. The reflection peaks of the
first cold crystal were not consistent with 3- or 3-. We
tentatively denote this crystalline phase as 3-, although its
crystal structure is still unknown. Upon further heating, at
approximately 85 °C (Tcc-2), the sample underwent solid-solid
transformation to another crystalline phase (pattern vii),
showing an exothermic anomaly with a broad peak over
70–100 °C. During this transformation, changes including the
weakening of a peak at 6°, and unification of two peaks around
15–16° occurred. The PXRD patterns (viii–xii) were
substantially unchanged until the sample was cooled to 25 °C
after being raised to 125 °C at the maximum processing
temperature. The PXRD pattern of the obtained crystal was
l / nm
Figure 4. Result of UV-Vis spectra in 350-650 nm region in the
initial crystals for 1 (1-), 2 (2-) and 3 (3-) compounds.
Similarly, we measured the solid-state UV-Vis absorption
spectra of the samples using a temperature-controlled stage
placed across the optical path of the microscope. The
temperature was changed at the same scanning speed as that of
DSC (15 °C min-1). The spectra of 1 and 3 at various
temperatures are shown in Figures 5 and 6, respectively．The
melting (Tm), crystallization (Tc) and cold crystallization (Tcc)
temperatures were indicated with arrows, while the glass
transition temperature (Tg) was indicated at the temperature
determined by the DSC measurement.
In the first heating process of compound 1, a slight
decrease of the shoulder at 380 nm was observed when the
temperature was increased from 25 to 50 °C. Upon further
heating to the Tm (approximately 118 °C), another shoulder
arose at 490 nm. After fusion, the shoulder rapidly grew to give
a sharp onset at approximately 510 nm. During the subsequent
cooling process, the spectrum gradually became fine-structured.
After recrystallization, the peaks in the long wavelength region
(450–490 nm) were remarkable. During the second heating
process, the short wavelength region (370–450 nm) gradually
increased. In contrast to the first heating process, the spectrum
showed no change during fusion.

2nd heating
1st heating
1st cooling
1
0.8
0.6
0.4
0.2
0
1
0.8
0.6
0.4
0.2
0
1
0.8
0.6
0.4
0.2
0
0
50
100
0
50
100
100
50
0
T / C°
T / C°
130
110
-10
10
30
50
Tm
128
122
118
115
90
50
25
90
70
50
30
10
-10
Tm
70
Tc
90
110
118
128
400 500 600
400
500
l / nm
600
400
500
l / nm
600
l / nm
Figure 5.Temperature-dependent UV-Vis spectra in 350-650 nm region in first heating - cooling process and second heating process
for 1 compound. Tm, melting; Tg, glass transition; Tc, crystallization; Tcc, cold crystallization.
2nd heating
1st heating
1st cooling
1
0.9
0.8
0.7
0.6
0.5
0.4
1
0.9
0.8
0.7
0.6
0.5
0.4
1
0.9
0.8
0.7
0.6
0.5
0.4
0
50
100
150
0
50
100
150
150
100
50
0
T / C°
T / C°
T / C°
137
135
90
70
50
30
10
-10
-10
10
30
50
140
134
110
90
70
50
24
Tg
Tm
Tcc-2
Tcc-1
Tg
Tm
70
90
110
130
140
400
500
600
400
500
600
400
500
600
l / nm
l / nm
l / nm
Figure 6. Temperature-dependent UV-Vis spectra in 350-650 nm region in first heating - cooling process and second heating process
for 3 compound. Tm, melting; Tg, glass transition; Tc, crystallization; Tcc, cold crystallization.
For compound 3, the spectrum did not change until
approximately 70 °C in the first heating process. We observed
the growth of a shoulder peak at approximately 490 nm prior to
fusion. During the subsequent cooling process, the spectrum
was almost unchanged, while the sample underwent
supercooling and transformed into a glassy state. During the
second heating process, the onset at the long-wavelength edge
moderated after the glass transition temperature (~30 °C), and
immediately the spectrum sharpened upon cold crystallization
at Tcc-1. The spectrum became slightly broader with
intensification of the shoulder peak at approximately 490 nm,
while the second cold crystallization was observed at Tcc-2.
After melting of the cold crystal, the spectrum became
less-structured, and the shoulder at the long-wavelength edge
disappeared. The insets in Figures 5 and 6 show the ASW values.
derivatives used in this study were only differed with the
position of their methyl group, which may be responsible for
the observed differences in the thermal behavior and crystal
structures. The UV-Vis absorption spectra of the initial crystals
(1-, 2-, and 3-) were noticeably different from each other,
indicating a difference in the tautomer ratio of the Schiff
bases.21,22) However, the spectra of these Schiff base samples
varied in a complicated fashion depending on the temperature
and thermal history, suggesting that the spectral profile is not
solely determined by the tautomer ratio. Some intermolecular
interactions may influence the electronic transition, while other
minor factors including -conjugation effect is similar among
these compounds. It may be expected that a precise peak
decomposition analysis could reveal the relative abundance of
the tautomers. Regrettably, each spectrum is considerably broad
and fused so the unique solution of the peak decomposition
could not be achieved. Therefore, we confined the number of
4. Discussion
4.1 Relationship between UV-Vis absorption and
tautomerism. The o-hydroxy aryl aldehyde Schiff bases are
the peak components to five, and introduced a parameter, ASW,
to characterize the spectral profiles (see the Experimental
section). According to reported studies on the tautomerism of
Schiff bases, the NH-form gives an absorption band in the
longer wavelength region compared to the OH-form. Based on
this conventional view, in this study, we tentatively regard this
parameter as a measure of the abundance of the OH-form
relative to all the tautomers present in the solid.
known to show thermochromism, originating from
a
temperature-dependent change in the relative abundance of the
prototropic tautomers, i.e. the OH- and NH-forms. The
thermochromic behavior changes depending on the relative
stability of the tautomers, which may be influenced by
π-electron delocalization, intramolecular hydrogen bonding,
substituent effects, and temperature.31) The three toluidine
4.2 Potential surfaces of tautomers with respect to

phenyl rotation. In previous studies, the authors highlighted
features observed for the potential surface of materials that
exhibit cold crystallization. For example, the potential surfaces
that the OH- and NH-forms are in equilibrium in the
temperature range of the present study. Furthermore, because
the flat-bottomed minima lead to a large density of states, the
ratio of the NH-form increases rapidly as the temperature
increases if the phenyl ring freely rotates or is liberated. Thus,
it is suggested that the NH-form is present as a mixture of
various rotamers in the molten state. On the other hand, during
the cooling process, the OH-form easily falls into the
energy-minimum structure, which is suitable for crystal
packing along the relatively steep well, whereas the molecules
in NH-form stay as a mixture of several rotamers. Based on
these computational results, we can postulate the mechanism of
supercooling for salicylideneaniline derivatives as follows: (1)
at ambient temperatures, the tautomerization equilibrium is
biased towards the OH-form; (2) as the temperature rises, the
proportion of the NH-form increases; (3) in the molten state,
the structural variety (torsion angle of phenyl ring) of the
NH-form increases; (4) during the cooling process, some
molecules in the OH-form begin to crystalize, while those in
the NH-form interfere with the crystallization process.
Salicylideneanilines are regarded as a “pure compounds” the
crystalline state, but can actually be a mixture of different
compounds in the molten state. If the proportion of the
OH-form becomes sufficiently large for crystallization before
the temperature reaches Tg, the melt of the compound
crystalizes but otherwise gets supercooled.
4.3 The influence of tautomerism on cold crystallization.
The DFT results described above suggest that the molecules in
the OH-form crystallize more easily than those in the NH-form.
The variation in ASW values estimated for the initial crystals
(Figure 4) indicate that the optimal distribution of the tautomers
depends on the crystal packing structure. These findings
suggest that the crystallization would proceed more efficiently
if the ratio of the tautomers is closer to an optimal value for
crystallization. Following this hypothesis, we hereafter give a
comprehensive explanation for the cold crystallization of 3.
During the first heating process of 1- (Figure 5, left), the
ASW value rapidly decreased from 0.6 to 0.2 at 50 °C, and then
exhibited a mild recovery to 0.5 when approaching the melting
temperature. This change corresponds to the equilibrium shift
upon temperature increase, and the latter change suggests the
stabilization of the OH-form accompanied by alteration of the
potential curve. In the subsequent cooling process, ASW value
increased to 0.7 at 70 °C, where sample crystallized to 1-, and
thereafter remained relatively constant. This coincidence of
crystallization and saturation point of ASW is consistent with our
hypothesis. During the second heating process, the ASW value
remained almost constant along with the overall spectral profile
until the sample melted. The spectral insensitiveness of the 1-
crystal to changes in temperature suggests an overwhelming
stability of the OH-form in this crystal packing environment.
Namely, the 1- crystal is primarily a non-thermochromic phase,
in contrast to the 1- crystal.
calculated
for
bis[1-(toluyliminomethyl)naphthalene-2-olato]nickel(II)s had
flat-bottomed regions around the energy minima with respect to
the rotation of the phenyl groups.17-19) This suggests that the
phenyl groups can rotate without noticeable energy change,
indicating that the molecules have high conformational
flexibility. Such flexibility prevents molecules from falling into
a conformation appropriate for crystallization upon cooling,
resulting in supercooling. Based on this assumption, we
attempted to explain the thermal behavior of compounds 1 and
3. In particular, compound 1 recrystallized upon cooling,
whereas 3 showed no anomalies relevant to crystallization in
the cooling process, and finally underwent a glass transition.
We hypothesized that the potential surface of 3 has a shallow
well around an energy minimum. To quantify this effect, we
conducted quantum chemical calculations of the potential
surfaces of the OH- and NH-forms, by varying the twist angle
of phenyl group (f1, Figure 1-(a)). The geometric data of the
molecules were obtained by full optimization using from their
crystal structures (1- and 3-) as initial data. All calculations
were performed at the level of B3LYP/6-311G** using the
Ganssian09 program package32). Hereafter, we denote the
calculated structures of the ideal OH- and NH-forms of 1, as
1OH and 1NH, respectively, to discriminate from their observed
forms. The potential surface of 1OH (Figure 7(A)) has two
minima located at 150 and 210°, while that of 1NH (Figure
7(B)) has one shallow well spreading from 150 to 210°. This
difference may be ascribed to the -electron delocalization and
intramolecular hydrogen bonding. At the angle of 180°, a
bending structure of the N-phenyl group appears to contribute
to stabilization, compensating the steric repulsion between the
-OH and methyl groups. For both 1OH and 1NH, the obvious
energy increase at f1 < 90° and 270° < f1 regions resulted from
the steric repulsion between the methyl group and the
azomethine and 8-naphthyl protons. In contrast, the potential
surface of 3OH (Figure 7(C)) has four minima at 30, 150, 210,
and 330°, owing to the rotational symmetry of the p-toluyl
group. Similarly, the potential surface of 3NH (Figure 7(D)) has
two shallow wells at approximately 0 and 180°.
(C)
(A)1OH
3OH
25
20
15
10
5
25
20
15
10
5
0
0
-5
-5
(D)
3NH
(B)
1NH
25
20
15
10
5
25
20
15
10
5
As for the 3- crystal in heating process, the ASW value
decreased from 0.7 to 0.5 when the temperature was increased
from 70 °C to the melting point. Upon cooling, the ASW value
suggested relatively a high proportion of the NH-form in the
temperature range from 140 to 50 °C, and gradually recovered
to 0.7 by –10 °C. This distribution of tautomers may have
prevented the molecules from crystalizing due to the structural
diversity with respect to the phenyl ring rotation, as suggested
from the potential curve (Figure 7(C) and (D)). Upon further
cooling, even though the ASW value recovered to the level
sufficient for crystallization, the sample did not allow for the
nucleation and growth processes33,34) due to the rise of its
viscosity, leading to vitrification. During the second heating
0
0
-5
-5
0
90
180
270
0
90
180
270
360
f
1 / °
f
1 / °
Figure 7. Difference in molecular energy of 1 and 3 in OH-
form and NH-form between the fully optimized structure and
that with varying ϕ1. (A), 1OH; (B), 1NH; (C), 3 OH; (D), 3 NH
.
The calculated energy at the potential bottom of the
OH-form was lower than that of the NH-form by 2.1 kJ mol-1
for 1 and by 3.0 kJ mol-1 for 3. These energy differences are
comparable to RT at ambient temperature (~25 °C), suggesting

process, the absorption spectrum irregularly changed. The ASW
value showed a sudden increase between 50–70 °C (Tcc-1), and
thereafter followed the trajectory similar to that of the first
heating process until melting. Along this temperature course,
we observed crystallization around 30 °C and solid-to-solid
transition between 90–130 °C (Tcc-2) by microscopic analysis.
These events were identified as a two-step cold crystallization
process with thermal anomalies at 43 and 72 °C that were
observed in the XRD-DSC results for 3 (Figure 3). The
discrepancies in temperature between the two methods may
have arisen from a difference in experimental conditions. For
example, the thermal conductivity of the sample container
varies between the glass for the microscope and Al for the DSC
analysis. Therefore, we assigned the sharp increase of ASW to
the formation of the 3- phase, and the following gentle
decrease to the equilibrium shift of the tautomers in the 3-
the homogeneity of the system.
Supplementary Material
DSC charts for analyses of phase relation of 1-, 2- and
3-, XRD-DSC chart of cold crystal (3) obtained in the heating
processes at 125.0—139.6 °C (scan rate: 1 °C min-1) compared
with simulated pattern based on crystal structures (3- and 3-)
determined independently. An example of the resolution of a
UV-Vis spectrum into five gaussians through a curve fit.
Atomic coordinates of fully optimized 1 and 3 molecules in
vacuum.
Acknowledgement
The authors are grateful to professor emeritus Hiroshi
Nakamura for his helpful instruction of Marquardt method and
distribution of the home-made program.
phase. Because the 3- phase appears over
a
narrow
A part of this work was conducted at Advanced
Characterization Nanotechnology Platform of the University of
Tokyo, supported by "Nanotechnology Platform" of the
Ministry of Education, Culture, Sports, Science and
Technology (MEXT), Japan.
temperature range (135–137 °C), it is difficult to detect the
change in the ASW value transitioning from 3- to 3-. After the
melt of the 3- phase, the ASW remains at a relatively low value
(0.5–0.6), suggesting that the NH-form-biased distribution is
probable in the molten state, even if the liquid is supercooled.
The above considerations validate our hypothesis that the
crystallization of the Schiff base samples is influenced by the
balance between the distribution of the tautomers and the
viscosity of their supercooled liquid state. The melt sample
undergoes crystallization in the cooling process if the OH-form
ratio reaches an appropriate value before the viscosity
interferes with the crystallization. In contrast, the melt sample
forms a supercooled glass if the OH-form ratio remains at a low
level until the viscosity suppresses the nucleation and
subsequent crystal growth. Even for the supercooled sample,
cold crystallization occurs if the molecules recover sufficient
mobility and appropriate tautomer ratio as the temperature
rises.
References
1.
2.
3.
4.
5.
V.V. Tyagia, S.C. Kaushika, S.K. Tyagib, T. Akiyamac,
Renew Sustain Energy Rev. 2011, 15, 1373-1391.
K. Manjunath, C. S. Kaushik, Renewable & Sustainable
Energy Reviews 2014, 40, 348.
H. Ogura, T. Yamamoto, H. Kage, Energy 2003, 28,
1479-1493.
D. Mileva, I. Kolesov, R. Androsch, Colloid Polym Sci
2012, 290, 971-978.
M. Di Lorenzo, M. Gazzano, M. Righetti,
Macromolecules 2012, 45, 5684-5691.
6.
7.
H. Tsuji, L. Bouapao, Polym Int 2012, 61, 442-450.
T. Endo, T. Kato, K. Tozaki, K. Nishikawa, J. Phys. Chem.
B 2010, 114, 407–411.
5. Conclusion
In this study, we compared thermal behavior of three
2-hydroxy-1-naphthaldehyde Schiff base compounds derived
from o-, m-, and p-toluidines by a series of simultaneous
XRD-DSC measurements. We found that the subtle difference
in molecular structure provided the compounds with
remarkable changes in their crystal structure and phase
transition behavior. Only the p-toluidine derivative exhibited
supercooling followed by cold crystallization, while the
o-isomer showed ordinary melting-crystallization cycles, and
the m-isomer showed irreversible vitrification. These Schiff
base compounds showed distinct absorption spectra, depending
on the distribution of their prototropic tautomers, i.e. the OH-
and NH-forms. DFT calculations suggested that the NH-form
has a flat-bottomed potential well depending on phenyl rotation,
which prevents the molecule from rapidly crystallizing. The
solid-state, variable-temperature absorption spectra enabled the
discussion of changes in the distribution of the tautomers along
with the phase transition behavior. The results were consistent
with our hypothesis that the rate of crystallization is influenced
by the ratio of the tautomers and viscosity of the sample.
Specifically, the melt sample forms a supercooled glass if the
OH-form ratio remains low until the viscosity suppresses the
nucleation and subsequent crystal growth. This hypothesis
gives a reasonable explanation for the cold crystallization of the
supercooled samples. Considering structural demand for
developing cold crystallization, it is notable that compound 3
satisfied the two guidelines proposed in previous studies:17-19)
1) nearly-free rotation of the phenyl ring increases the degree
of freedom of molecular motion, and 2) coexistence of OH/NH
tautomers interconvertible in the liquid (molten) state disturbs
8.
9.
K. Piela, I. Turowska-Tyrk, M. Drozd, M. M. Szostak, J.
Mol. Struct. 2011, 991, 42–49.
S. Sampath, A. A. Boopathi, A. B. Mandal, Phys. Chem.
Chem. Phys. 2016, 18 (31), 21251-21258.
10. A. Honda, Y. Takahashi, Y. Tamaki, K. Miyamura,
Chemistry Letters 2016, 45 (2), 211-213.
11. S.-J. Kim, T. E. Karis, J. Mater. Res. 1995, 10(8),
2128-2136.
12. Q. Wang, A. Takeuchi, Y. Yamamura, K. Saito, W. Mori,
M. Sorai, J. Phys. Chem. B 2008, 112, 11039–11048.
13. N. Arai, M. Sorai, S. Seki, Bull. Chem. Soc. Jpn. 1972,
45(8), 2398–2406.
14. Z. Rezvani, A. R. Abbasi, K. Nejati; M. Seyedahmadian,
Polyhedron 2005, 24, 1461–1470.
15. Y. Shirota, J. Mater. Chem. 2000, 10, 1-25.
16. S. P. Das, R. Ganguly, Y. Li, H. S. Soo, Dalton
Transactions 2016, 45 (34), 13556-13564.
17. K. Iwase, H. Houjou, Y. Yamamura, K. Saito, Chem. Lett.
2013, 42, 1040–1042.
18. K. Iwase, Y. Nagano, I. Yoshikawa, H. Houjou, Y.
Yamamura, K. Saito, J. Phys. Chem. C 2014, 118,
27664–27671.
19. K. Iwase, I. Yoshikawa, H. Houjou, Y. Yamamura, K.
Saito, Bull. Chem. Soc. Jpn. 2015, 88, 989–995.
20. H. Achira, I. Yoshikawa, H. Houjou, Chem. Lett. 2016,
45(12), 1415-1417.
21. K. Amimoto, T. Kawato, J. Photochem. Photobiol. C:
Photochemistry Reviews 2005, 6(4), 207-226.
22. E. Hadjoudis, I. M. Mavridis, Chem. Soc. Rev. 2004, 33,
579-588.

23. H. Joshi, F. S. Kamounah, G. van der Zwan, C. Gooijera,
L. Antonov, J. Chem. Soc., Perkin Trans. 2 2001,
2303-2308.
24. S. R.Salman, J. C. Lindon, R. D. Farrant, T. A. Carpenter,
Magn. Reson. Chem. 1993, 31, 991-994.
25. G. M. Sheldrick, Acta Crystallogr. 2008, A64, 112−122.
26. M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G.
L. Cascarano, L. D. Caro, C. Giacovazzo, G. Polidori, D.
Siliqi & R. Spagna, J. Appl. Cryst. 2007, 40, 609–613.
27. D. W. Marquardt, J. Soc. Indust. Appl. Math. 1963, 11,
431–441.
28. R. Tahara, T. Morozumi, H. Nakamura, J. Phys. Chem. B
1997, 101, 7736–7743.
29. B. Kaitner, E. Meštrović, G. Pavlović, J. Chem.
Crystallogr. 1998, 28, 77–82.
30. A. K. Baghdouche, S. Mosbah, Y. Belhocine, L.
Bencharif, Acta Cryst. 2014, E70, o676.
31. T. M. Krygowski, H. Szatylowicz, O. A. Stasyuk, J.
Dominikowska, M. Palusiak, Chem. Rev. 2014, 114,
6383−6422.
32. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G.
A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A.
Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B.
Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J.
L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini,
F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D.
Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng,
W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J.
Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H.
Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J.
E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R.
Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.
C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M.
Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski,
R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman,
and D. J. Fox, Gaussian 09, Revision D.01, D. J.
Gaussian, Inc., Wallingford CT, 2009.
33. R. Boistelle, J. P. Astier, J. Cryst. Growth 1988, 90,
14−30.
34. D. Turnbull, J. Appl. Phys. 1950, 21, 1022−1027.