Packing incentives and a reliable N-H...N-pyridine synthon in co-crystallization of bipyridines with two agrochemical actives

Article abstract of DOI:10.1039/c1ce05730h

The co-crystallization of agrochemical actives thiophanate-methyl and thiophanate-ethyl with 2,2′-bipyridine, 4,4′-bipyridine and 1,2-bis(4-pyridyl)ethane was investigated with conventional crystallization, the slurry method and liquid-assisted grinding. Co-crystals of both thiophanates with all bipyridines were found and the structures solved with single crystal X-ray diffraction. Whereas the 2,2′-bipyridine co-crystals seem to form because of a combination of weak interactions, and in the case of the thiophanate-methyl, partly because of close packing incentives, the 4,4′-bipyridine and 1,2-bis(4-pyridyl)ethane co-crystals form mainly because of a favourable N-H...N-pyridine hydrogen bonding synthon.

Full text of DOI:10.1039/c1ce05730h

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
CrystEngComm
Cite this: CrystEngComm, 2011, 13, 6531
www.rsc.org/crystengcomm
PAPER
Packing incentives and a reliable N–H/N–pyridine synthon in
co-crystallization of bipyridines with two agrochemical actives†
*
Elisa Nauha, Erkki Kolehmainen and Maija Nissinen
Received 15th June 2011, Accepted 5th August 2011
DOI: 10.1039/c1ce05730h
The co-crystallization of agrochemical actives thiophanate-methyl and thiophanate-ethyl with 2,20-
bipyridine, 4,40-bipyridine and 1,2-bis(4-pyridyl)ethane was investigated with conventional
crystallization, the slurry method and liquid-assisted grinding. Co-crystals of both thiophanates with
all bipyridines were found and the structures solved with single crystal X-ray diffraction. Whereas the
2,20-bipyridine co-crystals seem to form because of a combination of weak interactions, and in the case
of the thiophanate-methyl, partly because of close packing incentives, the 4,40-bipyridine and 1,2-bis(4-
pyridyl)ethane co-crystals form mainly because of a favourable N–H/N–pyridine hydrogen bonding
synthon.
now decided to examine co-crystallization with common co-
Introduction
crystal formers, namely 4,40-bipyridine (44bp)12–17 and 2,20-
bipyridine (22bp)18 (Scheme 1) as pyridine solvates of TM and
TE are also known. Due to the success with the first bipyridines,
crystallizations with the similar 1,2-bis(4-pyridyl)ethane (44bpe)
(Scheme 1) were also attempted. An important goal was to see
how much the difference of one carbon at the end of the chains in
TE, in comparison to TM, affects the crystallization behavior
and packing of the molecules.
The phenomenon of co-crystallization1,2 is important for phar-
maceutical,3–5 agrochemical6,7 and other industrial applications
because of the different and sometimes improved properties of
co-crystals of active ingredients compared to those of the parent
molecules. For pharmaceutical applications the goal of co-crystal
forms is usually to increase the solubility and bioavailability of
the active ingredient while maintaining the physical stability of
the dosage form, but for agrochemical actives, which are often
administered as slurry formulations in water, the goal is to lower
the solubility of the active to make a more stable formulation and
stop the active being washed away too soon once administered.
In a simplified model co-crystallization can be seen to arise for
two different reasons, either close packing or favorable hydrogen
bonding interactions.1,8 The strategy to find co-crystals that form
because of close packing requires a trial and error approach since
it is very hard to predict if a pair of molecules can pack tightly.
For co-crystals caused by favorable hydrogen bonding interac-
tions one can make educated guesses about possible co-crystal
formers based on the knowledge about hydrogen bonding syn-
thons9 with, for example, the help of the CCDC database. In
practice, however, packing in crystals is influenced by an intricate
combination of close packing incentives, hydrogen bonding and
other interactions, and therefore there is as yet no reliable way to
design co-crystals.
Experimental
Materials
TM (IUPAC: dimethyl 4,40-(o-phenylene)bis(3-thioallophanate))
of 99.8% purity from BASF, TE (IUPAC: diethyl 4,40-(o-phe-
nylene)bis(3-thioallophanate)) of 99% purity from Chem Service,
2,20-bipyridine from Merck, 4,40-bipyridine from Sigma-Aldrich,
Earlier, we investigated the polymorphism and versatile
solvate formation of two agrochemical actives, thiophanate-
methyl (TM)10 and thiophanate-ethyl (TE)11 (Scheme 1), and
€
€
University of Jyvaskyla, Department of Chemistry, Nanoscience Center,
P.O. Box 35, FI-40014 Jyva€skyla€, Finland. E-mail: maija.nissinen@jyu.
fi; Fax: +358 14 260 4756; Tel: +358 14 260 4242
Scheme 1 Molecular structures of thiophanate-methyl (TM), thio-
phanate-ethyl (TE), 2,20-bipyridine (22bp) and 4,40-bipyridine (44bp),
and 1,2-bis(4-pyridyl)ethane (44bpe).
† CCDC reference numbers 821123–831127. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/c1ce05730h
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6531–6537 | 6531

View Online
1,2-bis(4-pyridyl)ethane (99%) from Aldrich, distilled water and
solvents of analytical purity (min 99%) were used in the crys-
tallization, slurry and milling experiments.
atoms were placed in idealized positions, or in the case of
hydrogen bonding hydrogens, found from the electron density
map and included in structure factor calculations. The N–H
hydrogen atoms found in the electron density map were
ꢀ
restrained to a distance of 0.91 A to give the best fit to the X-ray
Slurries
data and to ensure stable refinement. Pictures of the structures
were drawn with Mercury.22 Crystal data and collection
parameters of the structures are presented in Table 1.
Slurries of TM with the 22bp and 44bp were carried out in
a acetonitrile : water solution. The slurries were mixed for three
days at 50 ꢀC, filtered with suction and let dry in open vessels.
13C CP/MAS NMR
Crystallizations
The 13C CP/MAS NMR spectra of the slurried TM samples were
measured with a Bruker Avance 400 FT NMR spectrometer with
a dual 4 mm CP/MAS probehead. The sample was packed in a 4
mm diameter ZrO2 rotor, which was spun at 10 kHz rate at 296
or 373 K. The contact time for CP was 4 ms, pulse interval 4 s,
time domain 2 K, which was zero filled to 8 K in the frequency
domain. Exponential window function with 5 Hz line broadening
was used. 20 000 scans were acquired. 13C shifts are referenced to
the CO-signal of glycine (176.03 ppm) measured before the TM
sample.
Equimolar (1 : 1) amounts (<40 mg in total) of the compounds
were weighed and dissolved in acetonitrile; the solutions were
transferred to clean crystallization test tubes, which were covered
with parafilm with a few holes and the solvent was let evaporate
in room temperature. Crystallization of TM with 22bp and TE
with 44bp was also attempted in ethanol with and without
seeding by the grinded samples, but the crystallizations produced
single crystals of the ethanol solvate of TM10 and form I of TE,11
respectively.
Grinding
Results and discussion
Liquid-assisted grinding was carried out on a Retsch MM 200
ball mill in 10 ml milling vessels for 20 min with an oscillation
frequency of 20 sꢁ1. 50 mg of TM/TE and a molar amount of
22bp or 44bp, according to the ratio in the co-crystal structures,
were weighed and a few drops of 1 : 1 water : ethanol solution
were added to the milling vessels. All but the TE–44bp mixture
produced PXRD patterns similar to the calculated patterns from
the structures. For the TE–44bp mixture and the TM–22bp
mixture, milling was also done with 3 drops of 25% MeCN in
water to confirm the results.
The formation of the TM co-crystals with 4,40-bipyridine and
2,20-bipyridine was first observed in slurry experiments, with the
results confirmed by PXRD and 13C CP/MAS NMR, after which
they were also found to crystallize from solution. Co-crystals for
TE were acquired directly with solution crystallization and also
used in the liquid-assisted grinding and thermomicroscopy
experiments. The 1,2-bis(4-pyridyl)ethane co-crystals were later
acquired by solution crystallization for structural comparison.
13C CP/MAS NMR results
High resolution solid state NMR and especially 13C CP/MAS is
nowadays often used as an aid or complementary technique in
polymorph, solvate and co-crystal characterization.10,23 In this
work 13C CP/MAS spectra were measured for the TM–22bp and
TM–44bp co-crystals, acquired by the slurry method, as well as
for pure TM alone (Fig. 1). The assignment of the spectra is
based on the comparison with liquid state NMR spectra.
Thermomicroscopy
The behavior of the 22bp and 44bp co-crystals during heating
was studied under polarized light with a Mettler FP82HT hot
stage connected to a Mettler FP90 central processor with an
Olympus BH-2 microscope. The primary heating rate used was
10 ꢀC minꢁ1 from 30 ꢀC to melting/decomposition of the sample.
In the TM spectrum its C]O resonance is at 153.8 ppm and
the aromatic carbons give three resonances as expected by
symmetry reasons at 135.8, 131.9 and 126.3 ppm, respectively.
C]S resonances are at 180.9 and 174.2 ppm, respectively. The
tentative assignments of 2,20-bipyridine resonances in the TM–
22bp co-crystal spectrum are at 154.7 ppm (carbons 1,10), 149.4
(carbons 3,30), 137.1 ppm (carbons 5,50), 126.1 ppm (carbons
4,40) and 120.8 ppm (carbons 6,60), respectively. In the spectrum
of TM–44bp, 4,40-bipyridine signals are easy to find and assign at
149.9 ppm (carbons 3,30, 5,50), 143.7 ppm (carbons 1,10) and
121.3 ppm (carbons 2,20, 6,60). Differing from the TM sample in
this co-crystal TM’s CH3O-signals exist as a doublet at 55.0 and
53.7 ppm. In TM–22bp spectrum the CH3O-group shows two
resonances but differs from TM–44bp with nonequal intensities.
Comparison of TM–44bp and TM–22bp spectra reveals that
the chemical shift differences of C]S resonances (at 182.5 and
180.5 ppm) in TM–44bp co-crystal are clearly smaller than in
TM–22bp where those are at 183.3 and 175.9 ppm, respectively.
PXRD
For powder X-ray diffraction analyses the samples were pressed
to a zero background silicon plate sample holder and measured
on a PANalytical X’Pert Pro system in reflection mode with non-
monochromated Cu-radiation. A 2q angle range of 5–35ꢀ and
a step time of 38 s were used with step resolution of 0.0167ꢀ.
Figures were drawn with X’Pert HighScore Plus.19
Single crystal X-ray diffraction
The X-ray diffraction data were collected on a Nonius Kappa
CCD diffractometer with an Apex II detector, using graphite-
ꢀ
monochromated CuKa radiation (l ¼ 1.54178 A) at 173 K. The
structures were solved by direct methods, refined, and expanded
by Fourier techniques with the SHELX-97 software package.20
Absorption correction was performed with Denzo 1997.21 All
non-hydrogen atoms were refined anisotropically. Hydrogen
6532 | CrystEngComm, 2011, 13, 6531–6537
This journal is ª The Royal Society of Chemistry 2011

View Online
Table 1 Crystal data and collection parameters for the co-crystal structures
Compound reference
Chemical formula
TM–22bp
TE–22bp
TM–44bp
TE–44bp
14H18N4O4S2$
C10H8N2
526.63
Triclinic
11.8193(4)
12.8098(5)
26.3527(10)
91.574(2)
95.655(2)
100.640(3)
3897.9(2)
P-1
TM–44bpe
TE–44bpe
2(C12H14N4O4S2)$
1.5(C10H8N2)$C2H3N 0.5(C10H8N2)
960.11
C
14H18N4O4S2$ C12H14N4O4S2$
0.5(C10H8N2)
420.48
C
2(C12H14N4O4S2)$ 2(C14H18N4O4S2)$
C12H12N2
869.02
C12H12N2
925.12
Formula Mass
Crystal system
448.54
Triclinic
10.6193(3)
15.0562(4)
16.2601(4)
108.694(4)
95.323(3)
109.312(4)
2265.87(10)
P-1
2
10 805
7733
0.0585
0.0517
Monoclinic
8.3860(7)
17.2950(12)
15.4566(11)
90.00
98.93(3)
90.00
2214.6(3)
P2(1)/c
4
Triclinic
9.0382(2)
9.5723(2)
12.5370(3)
100.183(3)
104.203(3)
101.444(3)
1001.16(4)
P-1
2
4322
3298
0.1194
0.0657
Triclinic
9.0972(2)
9.5762(2)
23.8891(3)
86.224(1)
88.273(1)
88.214(1)
2074.79(7)
P-1
2
10 604
7144
0.1069
0.0604
Triclinic
11.9780(10)
14.1211(12)
14.9536(12)
107.572(4)
106.503(5)
91.062(4)
2296.8(3)
P-1
2
9982
7030
0.1410
0.0688
ꢀ
a/A
ꢀ
b/A
ꢀ
c/A
a/ꢀ
b/ꢀ
g/ꢀ
ꢁ3
ꢀ
Unit cell volume A
Space group
Z
6
18 442
13 511
0.1463
0.0802
Meas. reflns
Indep. reflns
Rint
Final R1 values
(I > 2s(I))
Final wR(F2) values
(I > 2s(I))
6504
3778
0.0975
0.0713
0.1257
0.0722
0.1394
1.032
0.1812
0.1077
0.2117
1.039
0.1788
0.0735
0.1882
1.069
0.2103
0.1094
0.2389
1.071
0.1459
0.0832
0.1628
1.043
0.1382
0.1494
0.1758
1.018
Final R1 values
(all data)
Final wR(F2) values
(all data)
Goodness of fit on F2
ꢀ
Taking into account the X-ray data (see below) it is quite clear to
assign the TM–22bp resonance at 175.9 ppm to that one involved
in hydrogen bonding. According to X-ray data C]S groups do
not form hydrogen bonds in TM–44bp co-crystal, which explains
the small shift difference. In pure TM sample the C]S shift
difference is very close to that of TM–22bp suggesting similar
hydrogen bonding behavior.
with ring centroid-C distances of 3.71 and 3.94 A, and subse-
ꢀ
quent ring centroid-H distances of 2.78 and 3.03 A.
The TE co-crystal with 2,20-bipyridine crystallized in space
group P21/c with a molecule of TE and half a molecule of 22bp in
the asymmetric unit, giving the ratio of 2 : 1. The TE molecules
build up hydrogen bonded chains (Fig. 2b) with an R2,2(8) motif
consisting of two N–H/S hydrogen bonds. There are addi-
tionally only intramolecular N–H/O]C hydrogen bonds in an
S(6) motif. The chains of TM pack parallel to each other with p–
p stacking interactions to neighboring chains with ring distances
Surprisingly no trace from MeCN is visible in TM–22bp
spectrum although the powder X-ray pattern of the sample is
similar to the calculated pattern of a structure that contains also
MeCN. This, together with the ability of the co-crystal to form
also in ethanol solution, suggests that the co-crystal is still intact
even without the MeCN.
ꢀ
of approximately 3.4 A. The 22bp molecules are located in
between the arms of the TE molecules in channels running
through the structure in the direction of the crystallographic
a axis (Fig. 2b). There are weak hydrogen bonds between the
22bp nitrogens and the ethyl hydrogens of TE molecules in
neighboring chains. The arrangement of TE molecules in the
structure of this co-crystal is interestingly the same as that in the
previously obtained isomorphous acetone, DCM, chloroform
and dioxane solvates of TE11 with the 22bp taking up the space of
two solvent molecules.
Co-crystal structures with 2,20-bipyridine
The MeCN solvate of the TM co-crystal with 2,20-bipyridine
ꢁ
crystallized in space group P1 with a TM : 22bp : MeCN ratio of
4 : 3 : 2, containing half of this in the asymmetric unit. The TM
molecules build up double chains composed of rings of four TM
molecules hydrogen bonded to each other (Fig. 2a). The rings are
made with two hydrogen bonding arrangements, one of which is
a R2,2(8) motif24 composed of two N–H/S hydrogen bonds and
the other a second order R2,2(8) motif with one N–H/S and one
N–H/O–C hydrogen bond. There is also a very weak C–H/O–
C bond supporting the connection with this motif. The rings of
TM arrange into chains with a R2,2(12) motif of two N–H/O]
C hydrogen bonds.
Co-crystal structures with 4,40-bipyridine
In the TM$4,40-bipyridine co-crystal the TM to 44bp ratio is
2 : 1, while in the TE$4,40-bipyridine co-crystal the TE to 44bp
ratio is 1 : 1. However, there are similarities in the hydrogen
bonding patterns of these co-crystals (Fig. 3a and b). The similar
hydrogen bonded unit consists of two molecules of TM/TE
connected with an R2,2(12) motif of two N–H/O]C hydrogen
bonds and two molecules of 44bp which are hydrogen bonded to
this pair. The difference of the structures lies in the hydrogen
bonding of these units into chains.
One of the 22bp molecules resides right inside the ring of TM
molecules while the other 22bp molecule along with the MeCN
molecules is located between the TM double chains. There are
edge-to-face aromatic interactions between the benzene rings of
the TM molecules and the 22bp that are located outside the rings
The TM co-crystal with 4,40-bipyridine crystallized in space
group P-1, with one molecule of TM and half a molecule of 44bp
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6531–6537 | 6533

View Online
Fig. 2 (a) Rings of TM with 22bp in the MeCN solvate of the TM co-
crystal with 22bp and (b) the channels of 22bp in the TE co-crystal with
22bp. MeCN molecules and C–H hydrogen atoms removed for clarity.
Fig. 1 CPMAS NMR of the TM–22bp co-crystal (top), TM form I
(middle) and the TM–44bp co-crystal (bottom).
in the asymmetric unit. The R2,2(12) hydrogen bonded pairs are
connected with hydrogen bonds to the 44bp molecules making
up infinite parallel chains where a pair of TM molecules and
a 44bp molecule alternate (Fig. 3a). Interestingly, in this struc-
ture the sulfur atoms are not involved in hydrogen bonding and
one N–H hydrogen is also left without a hydrogen bond
acceptor.
The TE co-crystal with 4,40-bipyridine crystallized in space
group P-1, with three molecules of TE and three molecules of
44bp in the asymmetric unit. The symmetrically different mole-
cules of TE are fairly similar in conformation, with differences
mostly in the orientation of the ethyl groups at the ends of the
arms. The torsion angles of the ethyl group from the C]O
groups are 170.2ꢀ, ꢁ156.1ꢀ, 165.1ꢀ, 179.2ꢀ, ꢁ88.9ꢀ and 155.1ꢀ.
The symmetrically different molecules also have the same
Fig. 3 Hydrogen bonding and packing in the (a) TM and (b) TE–44bp
co-crystals with C–H hydrogen atoms removed for clarity.
hydrogen bonds to neighboring molecules. The reason for so
many molecules in the asymmetric unit seems to be the combined
influence of the flexibility of conformation of the ends of the arms
of the TE molecules and in the variable conformations of the
6534 | CrystEngComm, 2011, 13, 6531–6537
This journal is ª The Royal Society of Chemistry 2011

View Online
44bp molecules. The torsion angles between the rings of the 44bp
molecules are 41.3ꢀ, 43.4ꢀ and 44.9ꢀ.
The R2,2(12) hydrogen bonded pairs are connected with an
R2,2(8) motif of two N–H/S hydrogen bonds to neighboring
pairs forming hydrogen bonded chains of TE molecules in the
structure (vertical direction in Fig. 3b). These chains pack
parallel to each other with the 44bp molecules of the neighboring
chains located partly in between the arms of the TE molecules.
Unlike in the TM co-crystal, the other pyridine acceptor in the
44bp does not take part in hydrogen bonding.
Co-crystal structures with 1,2-bis(4-pyridyl)ethane
Both the 1,2-bis(4-pyridyl)ethane co-crystals crystallized in space
group P-1, with a ratio of 2 : 1 (TM/TE : 44bpe). Also in these
co-crystals the TM/TE molecules make pairs connected with an
R2,2(12) motif of two N–H/O]C hydrogen bonds, and two
molecules of 44bpe are hydrogen bonded to this pair.
In the TM co-crystal with 1,2-bis(4-pyridyl)ethane there are
two unconnected domains in the structure consisting of the
symmetry unequivalent TM molecules and half molecules of
44bpe on an inversion center. In one case the pairs are only
connected with hydrogen bonds to the bipyridine, similarly to the
TM co-crystal with 44bp, making up chains, but in the other case
the hydrogen bonded pairs are connected with a D(2) N–H/S
motif to other TE molecules to build up double chains which
then connect into sheets via the bipyridine molecules (Fig. 4a).
The two domains make up sheets that stack up on each other in
the crystal.
In the TE co-crystal with 1,2-bis(4-pyridyl)ethane the
hydrogen bonding between the TE molecules is identical to that
in the TE co-crystal with 44bp, but additionally the TE chains are
connected via the 44bpe molecules like in the TM co-crystal with
44bp (Fig. 4b). In the chains of TE pairs of the two symmetry
unequivalent TE molecules connected with the R2,2(12) motif
alternate.
Fig. 4 Hydrogen bonded pairs in (a) the two domains of the TM$bis(4-
pyridyl)ethane co-crystals and (b) the TE$1,2-bis(4-pyridyl)ethane co-
crystals with C–H hydrogen atoms removed for clarity.
22bp inside the ring and the MeCN molecule, which suggests that
bigger guests could fit better in the ring cavity and in place of the
MeCN. The packing coefficient of the 22bp co-crystal with TE,
however, is the lowest of all the co-crystal structures, which
suggest that favorable weak interactions must play a major role.
As the amount of hydrogen bonding is the same as in the other
structures, this structure is stabilized by the aromatic stacking
interactions between the TE molecules.
Packing incentives
The packing coefficients25 (Fig. 5) for co-crystal structures were
calculated using the formula C(k) ¼ Z ꢂ V(mol)/V(cell), where Z
is the number of molecules in the unit cell, V(mol) is the
molecular volume of the molecules,26 and V(cell) is the volume of
the unit cell. When compared to the packing coefficients of the
two polymorphs of TM and the three polymorphs of TE,11 the
results indicate that the packing of the co-crystals is in general
more loose-fitting than that of the polymorphs with the TM–
22bp co-crystal being the exception. Form III of TE is left out of
consideration as the structure has voids and is thus very loosely
packed.
As the 2,20-bipyridines are not hydrogen bonded to the TM/TE
molecules in either of the co-crystal structures, the reason for the
formation of structures is in the packing of the thiophanates and
the template effect of the 22bp. The 22bp co-crystal with TM
forms most likely because of the several simultaneous strong and
weak hydrogen bonds between the TM molecules as well as
favorable close packing. It is likely that the 22bp molecules
template the formation of the rings of TM enabling this sort of
clathrate packing to take place. There is some movement of the
Fig. 5 Packing coefficients of the polymorphs of TM and TE and the
bipyridine co-crystals.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6531–6537 | 6535

View Online
The 4,40-bipyridine and 1,2-bis(4-pyridyl)ethane co-crystals
have low packing coefficients and are formed because of the
strong hydrogen bonds between the 44bp/44bpe and thiophanate
molecules. These strong hydrogen bonds are also the reason why
the packing of these co-crystals of TM and TE is more similar to
each other than that of the 2,20-bipyridine co-crystals. The 44bpe
co-crystal with TM is the most densely packed of all the hydrogen
bonded co-crystals, likely because of the formation of the two
slightly different layers. Because of the efficient use of hydrogen
Table 2 Approximate melting points (MP) of the compounds and co-
crystals as determined by hot stage microscopy
Compound MP (pure) MP (22bp co-crystal) MP (44bp co-crystal)
TM
TE
22bp
44bp
167
192
70
110–120
120–130
—
160–170
140–150
—
110
—
—
3
ꢀ
bonds, the 44bpe co-crystal with TE has voids of 50 A and a lower
packing coefficient than the other hydrogen bonded co-crystals.
crystal could be determined with the thermomicroscope. The
melting ranges for the co-crystals in Table 2 are very broad due to
poor repeatability of the melting observations caused by inter-
ference from the unco-crystallized pure compounds and due to
the changes in the crystallite size when the co-crystals were
produced with different methods.
Grinding
Having acquired such a group of similar co-crystals, we decided
to test whether the known co-crystals could be produced with
liquid-assisted grinding. Compounds were weighed in the ratios
of the known co-crystals and water : ethanol was used as solvent.
The TM–22bp co-crystal is known to contain MeCN, but this
was disregarded partly to see if a co-crystal that does not contain
acetonitrile exists.
The melting points of the TM–22bp co-crystal and both the
TE co-crystals, however, do show a clear difference to the pure
compounds, all melting at temperatures in between the melting
points of the pure compounds, and the formation of these co-
crystals can thus be recognized with thermomicroscopic obser-
vations. The TM–44bp co-crystal, however, melts only at the
temperature where TM starts to decompose and cannot be
recognized in this manner from pure TM. If single crystals were
not obtained to confirm the co-crystal formation, thermomi-
croscopy, which requires very little material, could still have been
used together with powder diffraction to help verify at least three
of the co-crystals.
The TM–44bp and TE–22bp mixtures formed the known co-
crystals when grinded with water : ethanol, but the TE–44bp and
TM–22bp mixtures did not (Fig. 6). The TE–44bp mixture
PXRD does not have peaks from pure 44bp or any of the known
TE polymorphs. This could indicate another co-crystal that has
better packing than the original co-crystal. The crystallization of
such a co-crystal was attempted from ethanol without success.
The TM–22bp PXRD is similar to the calculated pattern from
the MeCN solvate of the co-crystal, indicating that the formation
of the co-crystal may not depend on the presence of MeCN, and
that there may be a co-crystal in which ethanol takes the place of
MeCN. Attempts to crystallize this were also unsuccessful. New
grinding experiments of TE–44bp and TM–22bp with 3 drops of
25% MeCN in H2O were performed to confirm the formation of
the original co-crystals. Both the samples gave the pattern of the
known co-crystal structure.
The melting points of the found co-crystals are mostly lower
than those of the pure actives, which is the opposite of what is
generally preferred for agrochemical co-crystals. For TM and
TE, making the melting point higher with co-crystals is unlikely
because of the thermal decomposition of the actives upon
melting. The co-crystals could still be viable forms for processing
of formulations because of possible solubility differences and
crystallite shape preferences.
Conclusions
Thermomicroscopy
With the discovery of these six co-crystals, we have acquired
further proof that TM and TE can include a number of guest
molecules in their crystal lattices. The reasons for this behavior
seem to be the large amount of hydrogen bond donors coupled to
weak acceptors, the flexibility of the molecules, as well as the
possibility for aromatic interactions. The two arms of the
molecule being ortho-substituted in the benzene ring may also
make the packing of the molecules somewhat difficult via forcing
the sulfur and oxygen atoms quite close to each other unless guest
molecules are included between the arms.
Thermomicroscopy was used to investigate the melting behavior
of the co-crystals. As both TM and TE decompose upon
melting,10,11 we were unable to use melt film methods to produce
the co-crystals. We wanted to know if the formation of a co-
The N–H/N–pyridine hydrogen bond was found to be
a reliable synthon for co-crystal design with TM and TE. The
inherent weakness of the N–H/S]C hydrogen bond commonly
used in the polymorphs and solvates of TM and TE is likely the
cause, as was found for a more simple dithiooxamide by Piotr-
kowska et al.27
An important goal was to see how much the size difference of
TM and TE affects the crystallization behavior and packing of
the molecules. For the hydrogen bonded co-crystals of TM and
TE the size difference of the molecules was found to affect the
Fig. 6 PXRD patterns of the EtOH and MeCN grinding experiments of
TM with 22bp (left) and TE with 44bp (right) compared to the patterns
calculated from the structures.
6536 | CrystEngComm, 2011, 13, 6531–6537
This journal is ª The Royal Society of Chemistry 2011

View Online
6 Pat., WO 2008/117037, 2008.
7 Pat., WO 2008/117060, 2008.
8 G. R. Desiraju, Angew. Chem., Int. Ed., 2007, 46, 8342–8356, DOI:
10.1002/anie.200700534.
9 G. R. Desiraju, Angew. Chem., Int. Ed. Engl., 1995, 34, 2311–2327,
DOI: 10.1002/anie.199523111.
10 E. Nauha, H. Saxell, M. Nissinen, E. Kolehmainen, A. Schafer and
R. Schlecker, CrystEngComm, 2009, 11, 2536–2547, DOI: 10.1039/
b905511h.
11 E. Nauha, A. Ojala, M. Nissinen and H. Saxell, CrystEngComm,
2011, 13, 4956–4964, DOI: 10.1039/c1ce05077j.
12 I. D. H. Oswald, W. D. S. Motherwell and S. Parsonsa, Acta
Crystallogr., Sect. B: Struct. Sci., 2005, 61, 46–57, DOI: 10.1107/
S0108768104028605.
13 H. W. Roesky and M. Andruh, Coord. Chem. Rev., 2003, 236, 91–119,
DOI: 10.1016/S0010-8545(02)00218-7.
14 I. D. H. Oswald, D. R. Allan, P. A. McGregor, W. D. S. Motherwell,
S. Parsons and C. R. Pulham, Acta Crystallogr., Sect. B: Struct. Sci.,
2002, 58, 1057–1066, DOI: 10.1107/S0108768102015987.
15 A. Ranganathan, V. R. Pedireddi, G. Sanjayan, K. N. Ganesh and
C. N. R. Rao, J. Mol. Struct., 2000, 522, 87–94, DOI: 10.1016/
S0022-2860(99)00356-7.
16 S. Varughese and V. R. Pedireddi, Chem.–Eur. J., 2006, 12, 1597–
1609, DOI: 10.1002/chem.200500570.
17 T.-F. Tan, J. Han, M.-L. Pang, H.-B. Song, Y.-X. Ma and
J.-B. Meng, Cryst. Growth Des., 2006, 6, 1186–1193, DOI: 10.1021/
cg060009y.
structures in a subtle manner. The main hydrogen bond pairing
was always the same, but the size of the molecule influenced the
usage of the rest of the donors and acceptors. In the TM co-
crystals with 4,40-bipyridine and 1,2-bis(4-pyridyl)ethane one N–
H donor and all or most of the C]S acceptors are unused, while
in the TE co-crystal with 4,40-bipyridine only a pyridine acceptor
is unused and in the TE co-crystal with 1,2-bis(4-pyridyl)ethane
all donors and acceptors are in use. This is because the larger
molecules have more conformational freedom and can pack well
even with more hydrogen bonds in use. For the 2,20-bipyridine
co-crystals the difference is not clear while TE makes a channel
structure known to also fit other molecules and TM builds
a clathrate type structure templated by the 2,20-bipyridine.
The stoichiometric diversity in the co-crystals with the same
bipyridines raises the question of whether there could be more
co-crystals with different molar ratios like in the work of Trask
et al.28 on caffeine co-crystals. The TE–4,40-bipyridine sample
grinded with EtOH could be an example of this, but further
investigation is required. We believe there are also numerous
other possibilities for co-crystals as well as solvates of TM and
TE. Similar molecules with flexible chains attached ortho on
a benzene ring could also be reliable co-crystal formers and we
plan on further investigating these.
€
18 B. Ye, M. Tong and X. Chen, Coord. Chem. Rev., 2005, 249, 545–565,
DOI: 10.1016/j.ccr.2004.07.006.
19 PANalytical B.V., 2006, 2.2b.
20 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,
64, 112–122, DOI: 10.1107/S0108767307043930.
Acknowledgements
21 Z. Otwinowski, D. Borek, W. Majewski and W. Minor, Acta
Crystallogr., Sect. A: Found. Crystallogr., 2003, 59, 228–234, DOI:
10.1107/S0108767303005488.
22 C. F. Macrae, P. R. Edgington, P. McCabe, E. Pidcock, G. P. Shields,
R. Taylor, M. Towler and J. van de Streek, J. Appl. Crystallogr., 2006,
39, 453–457, DOI: 10.1107/S002188980600731X.
We would like to thank the academy of Finland (proj. no.
116503) for partly funding the work and Dr Heidi Saxell at
BASF SE in Ludwigshafen, Germany for fruitful discussions.
23 R. K. Harris, Analyst, 2006, 131, 351–373, DOI: 10.1039/b516057j.
24 M. C. Etter and J. C. MacDonald, Acta Crystallogr., Sect. B:
Struct. Sci., 1990, B46, 256–262; J. Bernstein, R. E. Davis,
L. Shimoni and N.-L. Chung, Angew. Chem., Int. Ed. Engl., 1995,
34, 1555–1573.
Notes and references
1 A. D. Bond, CrystEngComm, 2007, 9, 833–834, DOI: 10.1039/
b708112j.
€
2 C. B. Aakeroy and D. J. Salmon, CrystEngComm, 2005, 7, 439–448,
DOI: 10.1039/b505883j.
ꢂ
ꢂ
25 A. I. Kitaıgorodskiı, Organic Chemical Crystallography, Consultants
Bureau, New York, 1961.
3 O. Almarsson and M. J. Zaworotko, Chem. Commun., 2004, (17),
1889–1896, DOI: 10.1039/b402150a.
26 Calculation
Molinspiration, http://molinspiration.com/.
of
Molecular
Properties
and
Drug-likeness,
4 N. Blagden, M. de Matas, P. T. Gavan and P. York, Adv. Drug
Delivery Rev., 2007, 59, 617–630, DOI: 10.1016/j.addr.2007.05.011.
5 D. McNamara, S. Childs, J. Giordano, A. Iarriccio, J. Cassidy,
M. Shet, R. Mannion, E. O’Donnell and A. Park, Pharm. Res.,
2006, 23, 1888–1897, DOI: 10.1007/s11095-006-9032-3.
27 B. Piotrkowska, A. Wasilewska, M. Gdaniec and T. Polonski,
CrystEngComm, 2008, 10, 1421–1428, DOI: 10.1039/b806061d.
28 A. V. Trask, J. van de Streek, W. D. S. Motherwell and W. Jones,
Cryst. Growth Des., 2005, 5, 2233–2241, DOI: 10.1021/cg0501682.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6531–6537 | 6537