Covalent assistance in supramolecular synthesis: In situ modification and masking of the hydrogen bonding functionality of the supramolecular reagent isoniazid in co-crystals

Article abstract of DOI:10.1039/c1ce05152k

The supramolecular reagent isonicotinic acid hydrazide (isoniazid) is a promising molecule in the supramolecular synthesis of multi-component molecular complexes. Due to the covalent reaction of the carbohydrazide functional group with simple ketones and

Full text of DOI:10.1039/c1ce05152k

CrystEngComm
www.rsc.org/crystengcomm
Volume 13 | Number 19 | 7 October 2011 | Pages 5599–5964
COVER ARTICLE
Lemmerer et al.
Covalent assistance in supramolecular synthesis

Dynamic Article Links <
C
CrystEngComm
Cite this: CrystEngComm, 2011, 13, 5692
www.rsc.org/crystengcomm
PAPER
Covalent assistance in supramolecular synthesis: in situ modification and
masking of the hydrogen bonding functionality of the supramolecular reagent
isoniazid in co-crystals†
a
b
Andreas Lemmerer, Joel Bernstein‡ and Volker Kahlenbergc
*
*
Received 31st January 2011, Accepted 14th March 2011
DOI: 10.1039/c1ce05152k
The supramolecular reagent isonicotinic acid hydrazide (isoniazid) is a promising molecule in the
supramolecular synthesis of multi-component molecular complexes. Due to the covalent reaction of the
carbohydrazide functional group with simple ketones and aldehydes, the hydrogen bonding
functionality of isoniazid can be modified, where two of the hydrogen bond donors are replaced with
hydrogen bonding ‘‘inert’’ hydrocarbons. The ketones used are propanone, 2-butanone,
cyclopentanone, cyclohexanone, cycloheptanone, cyclooctanone, 40-methylacetophenone and
benzophenone and the aldehyde 40-methylbenzaldehyde. The ‘‘modifiers’’ bonded to the isoniazid then
give a measure of control of the outcome of the supramolecular synthesis with 3-hydroxybenzoic acid
depending on the identity and steric size of the modifier used. The steric size itself can be used to shield
or to ‘‘mask’’ the remaining hydrogen bonding functionality of isoniazid such that common homomeric
and heteromeric interactions are prevented from taking place. This process of covalent assistance to
supramolecular synthesis has been carried out in a one-pot covalent and supramolecular reaction to
make six unhydrated co-crystals 2–7 and five hydrated co-crystals 8–12.
arrangements.4 Similar studies that try to rationalise the
Introduction
observed crystal structure through the molecular structure have
been noted for the amides.5
How does one control and modify the self-assembly process of
organic molecules towards a desired solid state? This question is
one of the tenets of crystal engineering and the work done so
far1,2 has fallen under the umbrella of supramolecular synthesis,
which looks at understanding and then (hopefully) controlling
how the molecules crystallize. Work that has achieved some
measure of success in unimolecular complexes is in carboxylic
acids. Leiserowitz set the scene and described the two common
hydrogen bonding arrangements of carboxylic acids, dimers and
catmers.3 Das et al. showed subsequently how the dimers, which
are the dominant case, can be reduced to form catemers by
altering the electronic and steric effects of the acid molecule as
well as to generate a number of different types of catemeric
When assembling multimolecular complexes, or more accu-
rately, a multi-component molecular complex,6 the key is to
identify the intermolecular interactions that will operate to bring
together two or more different molecules. Predominant in terms
of strength and directionality is the hydrogen bond.7 Work in this
field has benefitted from identifying first the heteromeric inter-
actions8 that are reliably doing this, and then identifying the
molecules that contain these functionalities. The intermolecular
interaction is then called a robust heterosynthon, which is
defined as a supramolecular synthon between unlike but
complementary functional groups,9 and the molecule becomes
a supramolecular reagent.6b,10 The carboxylic acid/pyridine
hydrogen bond is one of the leading candidates to date.11 It has
been shown by Shattock el al. that this hydrogen bond occurs
very robustly, even in the presence of other hydrogen bonding
aMolecular Sciences Institute, School of Chemistry, University of the
Witwatersrand, Johannesburg, 2050, South Africa. E-mail: Andreas.
lemmerer@wits.ac.za; Fax: +27 11 717 6749; Tel: +27 11 717 6711
bDepartment of Chemistry, Ben-Gurion University of the Negev, Beer-
Sheva, Israel 84105. E-mail: yoel@bgu.ac.il; Fax: +972 8 647 2943; Tel:
+972 8 647 2455
functionalities.12 The work by Aakeroy and co-workers has
€
cInstitute of Mineralogy and Petrography, University of Innsbruck,
Innsbruck, Austria 6020. Fax: +43 512 507 2926; Tel: +43 512 507 5503
† Electronic supplementary information (ESI) available: PXRD patterns
of 4–12. See DOI: 10.1039/c1ce05152k
‡ Present address: Faculty of Natural Sciences, New York University
Abu Dhabi, P. O. Box 129188, Abu Dhabi, United Arab Emirates.
E-mail: joel.bernstein@nyu.edu, Tel: 971 2 628 4190.
5692 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Scheme 1 The proof of concept carried out for covalent assistance to supramolecular synthesis using niazid: conventional supramolecular synthesis of
co-crystal between niazid and adipic acid forms 2-D sheets, whereas the covalent-assisted supramolecular synthesis features a covalent reaction between
niazid and the modifier acetone, and subsequent co-crystallization with adipic acid to form 1-D ribbons. See ref. 24.
Scheme 2 An overview of the different molecules used to demonstrate the covalent assisted supramolecular synthesis. There are nine modifiers used,
eight ketones and one aldehyde, which are then reacted with the supramolecular reagent isoniazid to produce nine modified isoniazid supramolecular
reagents.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5693

Scheme 3 The nine modified reagents shown in Scheme 2 are then co-crystallized with the co-crystallizer mHBA to produce six unhydrated co-crystals
and five hydrated co-crystals. The process of modification and co-crystallization all occurs in one pot. A conventional supramolecular synthesis not
involving the modifying process is shown schematically and contrasted to the covalent assisted supramolecular reaction.
shown how this heterosynthon can be modified and controlled to
design specific assemblies.13 By increasing the basicity of the N
atom, the strength of the resulting O–H/N hydrogen bond
drives the co-crystallization with a number of different acids and
increases the supramolecular yield.14 The same principle can be
used to bring together three different molecules and form
Fig. 1 The contents of the asymmetric unit for unhydrated co-crystals 2–7, showing the atomic numbering scheme. Displacement ellipsoids are shown
at the 50% probability level. The disorder in the cyclopentane ring of 4 is discussed in the text. Atoms with superscript i are at the symmetry position
[ꢀx + 3, ꢀy + 1, ꢀz + 1].
5694 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Fig. 2 The contents of the asymmetric unit for hydrated co-crystals 8–12, showing the atomic numbering scheme. Displacement ellipsoids are shown at
the 50% probability level. The disorder in the methyl ring of 12 is discussed in the text.
a ternary co-crystal. For example, by synthesizing a molecule
with two different basic N atoms, two carboxylic acids with
different acidities in principle can interact with the two different
N atoms,15 on the principle of best acceptor/best donor
hierarchies.16
groups to interact with different functional groups.23 From
a Cambridge Structural Database (CSD) analysis, and limited
work with acids, this supramolecular reagent was found to form
seven types of homosynthons with itself, and five heterosynthons
with acids.23 In addition, it is also a molecule that allows one
to modify its hydrogen bonding functionality in situ during the
co-crystallization process, by adding simple ketones to the co-
crystallizing solution. This process of modifying was demon-
strated in a proof of concept experiment with its isomer nicotinic
acid hydrazide (niazid) (Scheme 1).24 Co-crystallization of niazid
with adipic acid resulted in the utilization of the C(4) homo-
synthon of the O]C–N–H– group to connect the niazid mole-
cules and the R22(7) heterosynthon to connect the acid molecules
to the niazid to form a hydrogen bonded sheet. These sheets are
connected by the hydrogen bonding interactions of the amine
group –NH2 to the amide C]O and acid C–OH using N–H/O
hydrogen bonds. However, if one adds acetone to the crystalli-
zation (by changing the solvent from methanol to acetone), then
the NH2 group is replaced by the acetone carbons and the
N–H/O hydrogen bonds between the sheets are not able to
form. This changes the crystal structure of the co-crystal to
a ribbon arrangement, while the C(4) and R22(7) synthons are
retained. In summary, we have removed the hydrogen bonding
functionality by replacing the two hydrogens by ‘‘inert hydrogen
bonding’’ hydrocarbons, and subsequently altered the overall
Further examples, and by no means an exhaustive list, of
steering the crystal structure by modifying the molecular struc-
ture in both unimolecular and multimolecular complexes include
changing the identity of the substituents,17 the charge on the
functional groups,18 the spacers between functional groups,19 the
steric size adjacent to the functional group,20 by removing
hydrogen bonding sites through coordination chemistry,21 and
the relative position of the hydrogen bonding groups on the
molecule.22 Most of these examples require the synthesis of the
potential reagents or selection of available moieties, and then
carrying out the supramolecular reaction in a subsequent
sequential step. The reagent preparation/modification and co-
crystallization are often two distinct and physically separated
events.
The molecule we have chosen to modify is isonicotinic acid
hydrazide (isoniazid). Isoniazid is a versatile molecule. It is an
active pharmaceutical ingredient that helped cure tuberculosis as
part of a triple therapy cocktail, it co-crystallizes with carboxylic
acids to form pharmaceutical co-crystals, and is also a versatile
supramolecular reagent as it has multiple donor and accepting
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5695

Table 1 Crystallographic data for unhydrated co-crystals 4–7
4
5
6
7
Formula
Mr
Temperature/K
(C12H13N3O)$(C7H6O3)
341.36
173(2)
0.71073
0.50 ꢂ 0.40 ꢂ 0.10
Monoclinic
P21/c
19.7268(4)
8.3772(2)
10.1372(2)
90
98.644(2)
90
1656.20(6)
4
1.369
0.098
720
3.17 to 25.50
10 130
3073 [0.0235]
2285
0.0333
(C14H19N3O)$(C7H6O3)
383.44
100(2)
(C15H15N3O)$(C7H6O3)
391.42
173(2)
(C19H15N3O)2$(C7H6O3)
740.80
173(2)
ꢀ
Wavelength/A
1.5418
1.5418
1.5418
0.42 ꢂ 0.33 ꢂ 0.20
Crystal size/mm3
Crystal system
Space group
0.40 ꢂ 0.40 ꢂ 0.20
Monoclinic
P21/c
21.3902(15)
7.8621(5)
11.2995(7)
90
97.986(3)
90
1881.8(2)
4
1.353
0.744
816
4.17 to 67.40
11 997
3383 [0.0366]
2940
0.0459
0.36 ꢂ 0.32 ꢂ 0.10
Monoclinic
P21/c
24.4670(16)
7.3712(5)
10.8211(6)
90
99.666(6)
90
1923.9(2)
4
1.351
0.775
824
3.67 to 67.49
11 105
3452 [0.0425]
2909
0.0527
Triclinic
ꢁ
P1
ꢀ
a/A
9.8338(9)
10.2475(9)
10.8863(10)
104.295(8)
112.784(8)
103.370(8)
912.17(18)
1
1.350
0.727
389
4.75 to 62.65
4365
ꢀ
b/A
ꢀ
c/A
a/ꢁ
b/ꢁ
g/ꢁ
V/A
Z
3
ꢀ
r (calcd)/Mg mꢀ3
m/mmꢀ1
F(000)
q Range for data collection/ꢁ
Reflections collected
No. of unique data [R(int)]
No. of data with I > 2s(I)
Final R(I > 2s(I))
Final wR2 (all data)
2793 [0.0202]
2516
0.0568
0.0855
0.1329
0.1631
0.1618
architecture in what we describe as a covalent assisted supramo-
lecular synthesis. What makes this concept work is that we are
modifying the molecular structure of the supramolecular reagent,
yet still achieving our goal of making a co-crystal with control on
the co-crystal architecture.
(N0-(Butan-2-ylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid), 3
ꢁ
See ref. 23 for synthesis. Colourless plates. Mp 133–135 C.
(N0-(Cyclopentylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid), 4
We report here the systematic investigation of the modification
of the supramolecular reagent isoniazid with specific modifiers
(eight ketones and one aldehyde, Scheme 2), which primarily
removes the –NH2 functionality, and also the utilization of the
steric size of the modifiers to further control the supramolecular
architecture of co-crystals formed with 3-hydroxybenzoic acid
(mHBA) (Scheme 3).
0.200 g of isonicotinic acid hydrazide (1.46 mmol) and 0.213 g of
3-hydroxybenzoic acid (1.54 mmol) were added into a sample
vial. The solids were dissolved in AP-grade hot cyclopentanone
(6 ml), refluxed in a closed vial for six hours, and left to stand at
room temperature open to the atmosphere. Colourless plates
ꢁ
were afforded after 10 days. Mp 163–164 C.
(N0-(Cyclooctylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid), 5
Experimental
Compounds. All reagents were purchased from commercial
sources and used without further purification. Melting points
were determined on a hot stage microscope.
0.150 g of isonicotinic acid hydrazide (1.09 mmol), 0.155 g of 3-
hydroxybenzoic acid (1.12 mmol) and 0.142 g of cyclooctanone
(1.12 mmol) were added into a sample vial. The solids were
dissolved in AP-grade hot methanol (5 ml), refluxed for four
hours in a closed vial, and left to stand at room temperature open
to the atmosphere. Colourless plates were afforded after 5 days.
(Isonicotinic acid hydrazide)$(3-hydroxybenzoic acid), 1
ꢁ
0.370 g of isonicotinic acid hydrazide (2.70 mmol) and 0.380 g of
3-hydroxybenzoic acid (2.75 mmol) were added into a sample
vial. The solid was dissolved in AP-grade hot methanol (5 ml)
and left to stand at room temperature. Colourless needles were
afforded after 3 days. Mp 172–173 ꢁC, corresponding to the
literature value of isoniazid.
Mp 118–120 C.
(N0-(1-p-Tolylethylidene)isonicotinohydrazide)$(3-
hydroxybenzoic acid), 6
0.200 g of isonicotinic acid hydrazide (1.46 mmol), 0.213 g of 3-
hydroxybenzoic acid (1.54 mmol) and 0.21 g of 40-methyl-
acetophenone (1.59 mmol) were added into a sample vial. The
solids were dissolved in AP-grade hot methanol (5 ml), refluxed
for 3 hours in a closed vial, and left to stand at room temperature
open to the atmosphere. Colourless plates were afforded after
(N0-(Propan-2-ylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid), 2
ꢁ
ꢁ
See ref. 23 for synthesis. Colourless plates. Mp 133–134 C.
4 days. Mp 151–152 C.
5696 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Table 2 Crystallographic data for hydrated co-crystals 8–12
8
9
10
11
12
Formula
(C9H11N3O)$(C7H6O3) (C10H13N3O)$
(C12H15N3O)$
(C7H6O3)$(H2O)
373.40
100(2)
1.5418
(C13H17N3O)$(C7H6O3) (C14H13N3O)$(C7H6O3)
$(H2O)
333.34
173(2)
0.71073
(C7H6O3)$(H2O)
347.37
173(2)
$(H2O)
387.43
100(2)
$(H2O)
395.41
173(2)
1.5418
Mr
Temperature/K
ꢀ
Wavelength/A
0.71073
0.35 ꢂ 0.20 ꢂ 0.15
1.5418
Crystal size/mm3
Crystal system
Space group
0.40 ꢂ 0.24 ꢂ 0.12
0.24 ꢂ 0.24 ꢂ 0.2
0.40 ꢂ 0.36 ꢂ 0.1
Monoclinic
P21/n
0.44 ꢂ 0.24 ꢂ 0.20
Triclinic
Triclinic
Triclinic
Triclinic
ꢁ
P1
ꢁ
P1
ꢁ
P1
ꢁ
P1
ꢀ
a/A
6.3511(6)
10.5853(8)
12.8026(12)
75.694(7)
78.613(8)
81.087(7)
812.47(12)
2
6.3699(2)
10.9498(3)
12.8120(3)
81.009(1)
80.566(1)
83.859(1)
867.67(4)
2
6.3632(6)
10.8918(10)
13.3842(12)
87.434(7)
80.241(8)
83.083(8)
907.26(14)
2
1.367
0.829
396
3.35 to 67.45
6.3127(4)
11.4293(6)
26.1598(15)
90
91.109(5)
90
1887.07(19)
4
1.364
0.817
824
3.38 to 67.36
6.1506(5)
10.5426(6)
16.2041(12)
73.603(6)
90.403(6)
79.320(6)
988.57(12)
3
1.328
0.797
416
2.85 to 67.28
ꢀ
b/A
ꢀ
c/A
a/ꢁ
b/ꢁ
g/ꢁ
3
ꢀ
V/A
Z
r (calcd)/Mg mꢀ3
m(Mo-Ka)/mmꢀ1
F(000)
1.363
0.103
1.330
0.099
352
3.29 to 25.50
368
1.63 to 25.50
q Range for data
collection/ꢁ
Reflections collected 5230
No. of unique data [R 2990 [0.0242]
(int)]
14 210
3232 [0.0256]
8274
3240 [0.0326]
11 880
3403 [0.0381]
9145
3545 [0.0374]
No. of data with I > 1906
2s(I)
2702
2666
2858
2833
Final R(I > 2s(I))
Final wR2 (all data)
0.0335
0.0741
0.0408
0.1144
0.0487
0.1338
0.0453
0.1257
0.0501
0.1430
(N0-(Diphenylmethylene)isonicotinohydrazide)2$(3-
hydroxybenzoic acid), 7
(N0-(Cyclohexylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid)$(H2O), 10
0.200 g of isonicotinic acid hydrazide (1.46 mmol), 0.213 g of 3-
hydroxybenzoic acid (1.54 mmol) and 0.266 g of benzophenone
(1.46 mmol) were added into a sample vial. The solids were
dissolved in AP-grade hot methanol (5 ml), refluxed for 3 hours
in a closed vial, and left to stand at room temperature open to the
atmosphere. Colourless plates were afforded after 7 days. Mp
0.190 g of isonicotinic acid hydrazide (1.39 mmol), 0.190 g of 3-
hydroxybenzoic acid (1.38 mmol) and 1 ml of cyclohexanone
were added into a sample vial. The solids were dissolved in AP-
grade hot methanol (5 ml), refluxed for 30 minutes in a closed
vial, and left to stand at room temperature open to the
atmosphere. Colourless plates were afforded after 7 days. Mp
ꢁ
ꢁ
167–168 C.
145–146 C.
(N0-(Propan-2-ylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid)$(H2O), 8
(N0-(Cycloheptylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid)$(H2O), 11
0.200 g of isonicotinic acid hydrazide (1.46 mmol), 0.213 g of 3-
hydroxybenzoic acid (1.54 mmol) and 1 ml of acetone were
added into a sample vial. The solids were dissolved in AP-grade
methanol (6 ml) and stirred with no heating for 12 hours.
Thereafter, the crystals were allowed to grow by slow evapora-
tion at room temperature. Colourless plates were afforded after
0.150 g of isonicotinic acid hydrazide (1.09 mmol), 0.150 g of 3-
hydroxybenzoic acid (1.09 mmol) and 1 ml of cycloheptanone
were added into a sample vial. The solids were dissolved in AP-
grade hot methanol (5 ml), refluxed for 2 hours in a closed vial,
and left to stand at room temperature open to the atmosphere.
ꢁ
Colourless plates were afforded after 6 days. Mp 139–140 C.
ꢁ
2 days. Mp 113–115 C.
(N0-(Butan-2-ylidene)isonicotinohydrazide)$(3-hydroxybenzoic
acid)$(H2O), 9
(N0-(4-Methylbenzylidene)isonicotinohydrazide)$(3-
hydroxybenzoic acid)$(H2O), 12
0.200 g of isonicotinic acid hydrazide (1.46 mmol), 0.213 g of 3-
hydroxybenzoic acid (1.54 mmol) and 1 ml of 2-butanone were
added into a sample vial. The solids were dissolved in AP-grade
methanol (6 ml) and stirred with no heating for 12 hours.
Thereafter, the crystals were allowed to grow by slow evapora-
tion at room temperature. Colourless plates were afforded after
0.200 g of isonicotinic acid hydrazide (1.46 mmol), 0.213 g of 3-
hydroxybenzoic acid (1.54 mmol) and 0.175 g of 40-methyl-
benzaldehyde (1.46 mmol) were added into a sample vial. The
solids were dissolved in AP-grade hot methanol (6 ml), refluxed
for 6 hours in a closed vial, and left to stand at room temperature
open to the atmosphere. Colourless plates were afforded after
ꢁ
ꢁ
2 days. Mp 113–115 C.
6 days. Mp 169–170 C.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5697

Single crystal X-ray diffraction
Single crystal diffraction data were collected in u scan mode on
an Oxford Diffraction Gemini R Ultra diffractometer equipped
with a Ruby CCD-detector with Mo-Ka radiation (l ¼
ꢀ
0.71073 A, graphite monochromator, mono-capillary colli-
ꢀ
mator) for compounds 4, 8 and 9, and with Cu-Ka (l ¼ 1.5418 A,
multi-layer optics) for compounds 5, 6, 7, 10, 11 and 12 at 100 or
173 K using an Oxford Cryostream 700 cooler. Data reduction
and cell refinement were carried out using CrysAlisPro.25 Space
group assignments were made using XPREP226 and empirical
absorption corrections25 on all compounds. In all cases, the
structures were solved in the WinGX27 suite of programs by direct
methods using SHELXS-9728 and refined using full-matrix least-
squares/difference Fourier techniques on F2 using SHELXL-97.28
All non-hydrogen atoms were refined anisotropically. There-
after, all hydrogen atoms attached to N and O atoms were
located in the difference Fourier map and their coordinates
refined freely with isotropic parameters 1.5 times (O) or 1.2 times
(N) those of the heavy atoms to which they are attached. All C–H
hydrogen atoms were placed at idealized positions and refined as
riding atoms with isotropic parameters 1.2 times those of the
heavy atoms to which they are attached.
The positional disorder around the carbon atom C(17) in 4 was
resolved by finding alternate positions from the difference
Fourier map for the atom. These atoms were then refined
anisotropically together with their site occupancy such that the
sum of the occupancies for the two alternate atom positions
equalled one. The ratio of major to minor component is 81(1) to
19(1)%. The bond lengths and bond angles were restrained using
ꢀ
the SADI instruction in SHELX to be within 0.01 A of each
other for the two parts. Hydrogen atom positions were then
calculated for the respective atoms using a riding model.
The positional disorder around the four carbon atoms, C(9),
C(10), C(11) and C(12), in 7 was resolved by finding alternate
positions from the difference Fourier map for the respective
atoms. These atoms were then refined anisotropically with site
occupancies of 50% such that the sum of the occupancies for the
two alternate atom positions equalled one. The bond lengths
were restrained using the DFIX instruction in SHELX to be
ꢀ
1.39 A. Hydrogen atom positions were then calculated for the
respective atoms using a riding model.
The methyl group in 12 is disordered and the disorder
modelled using the AFIX 123 instruction in SHELX.
ORTEP-style diagrams of the thermal displacement ellipsoids
at the 50% probability level and molecular numbering scheme are
given in Fig. 1 and 2. Diagrams and publication material were
generated using ORTEP-3,29 PLATON30 and DIAMOND.31
Experimental details of the X-ray analyses are provided in Tables
1 and 2. Crystallographic data for the co-crystals 2 and 3 were
reported previously.23
Powder X-ray diffraction
Powder X-ray diffraction data for compounds 1, 2, 4–8 were
collected on a Philips powder diffractometer using CuKa1 radi-
ation, graphite monochromator on the diffracted beam and
operating at 40 kV, 30 mA at 293 K. Powder X-ray diffraction
data for compounds 3 and 9 were collected on a Stoe STADI MP
Scheme 4 The two different homosynthons expected for modified
isoniazid, as well as the two heterosynthons used to co-crystallize the
modified isoniazid with carboxylic acids.
5698 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Table 3 Geometrical parameters for hydrogen bonds in unhydrated co-crystals 2–7
:(D–H/A)/ꢁ
Symmetry transformations
ꢀ
d(D–H)/A
ꢀ
ꢀ
Compound
d(H/A)/A
d(D/A)/A
2
O(1)–H(1)/N(1) (a)
C(11)–H(11)/O(2)
O(3)–H(3)/N(3) (b)
N(2)–H(2)/O(4) (c)
3
0.98(2)
0.95
0.86(2)
0.89(2)
1.73(2)
2.42
2.07(2)
1.99(2)
2.711(2)
3.123(2)
2.881(2)
2.879(2)
178(2)
131
158(2)
174(2)
—
—
ꢀx, ꢀy + 1, ꢀz + 1
x, ꢀy + 1/2, z ꢀ 1/2
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/N(3) (b)
N(2)–H(2)/O(4) (c)
4
1.03(3)
0.95
0.91(3)
0.85(3)
1.62(3)
2.67
1.97(3)
2.28(3)
2.636(2)
3.317(2)
2.836(2)
3.103(2)
172(3)
126
159(2)
163(3)
—
—
ꢀx + 1, y ꢀ 1/2, ꢀz + 3/2
x, ꢀy + 3/2, z ꢀ 1/2
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/N(3) (b)
N(2)–H(2)/O(4) (c)
5
1.00(2)
0.95
0.87(2)
0.85(2)
1.63(2)
2.71
1.94(2)
2.357(2)
2.630(1)
3.346(2)
2.785(2)
3.192(2)
176(2)
125
162(2)
169(1)
—
—
ꢀx + 1, y ꢀ 1/2, ꢀz + 3/2
x, ꢀy + 3/2, z ꢀ 1/2
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/N(3) (b)
N(2)–H(2)/O(4) (c)
6
1.08(2)
0.95
0.88(2)
0.83(2)
1.55(2)
2.63
2.02(2)
2.95(2)
2.627(2)
3.288(2)
2.872(2)
3.740(2)
177(2)
127
161(2)
160(2)
—
—
ꢀx + 1, y ꢀ 1/2, ꢀz + 3/2
x, ꢀy + 3/2, z ꢀ 1/2
O(1)–H(1)/N(1) (a)
O(3)–H(3)/O(4) (b)
N(2)–H(2)/O(4) (c)
7
1.08(2)
0.87(2)
0.85(2)
1.55(2)
1.93(2)
3.08(2)
2.622(2)
2.782(2)
3.736(2)
171(2)
168(2)
135(2)
—
ꢀx + 1, ꢀy + 1, ꢀz
x, ꢀy + 3/2, z + 1/2
O(1)–H(1)/N(1) (a)
C(10A)–H(10A)/O(2)
O(1)–H(1)/N(1) (b)
N(2)–H(2)/O(4) (c)
0.87(3)
0.95
0.88(3)
—
1.87(3)
2.84
1.88(3)
—
2.744(3)
3.498(11)
2.744(3)
—
177(4)
127
177(4)
—
—
—
ꢀx + 3, ꢀy + 1, ꢀz + 1
—
diffractometer using flat-plate bisecting transmission geometry
and strictly monochromatic CuKa1-radiation operating at 40 kV,
40 mA. Powder X-ray diffraction confirmed that the single
crystal structures were representative of the bulk material (see
ESI†).
In a process akin to the one used for the carbohydrazide func-
tional group, we searched the CSD for homosynthons of the N-
acylhydrazone, however this time by including the pyridine
functional group.34 Of the 55 well determined structures found,
the C(7) (Scheme 4a) heterosynthon is observed 28 times, while
the homosynthon C(4) is observed 23 times (Scheme 4b). The
difference between the two synthons lies in the identity of the
acceptor site atom, being the single lone pair of the pyridine N
atom in C(7) and one of the two lone pairs on the O atom in C(4).
The C(4) homosynthon is not unexpected as it is one of the most
commonly observed hydrogen bonding patterns for amides.35
The expected heterosynthon is unchanged, the hydrogen bond
from the carboxylic acid to the pyridine nitrogen. This hetero-
synthon exists in two geometric variants. The carboxylic acid
group and the pyridine ring can be co-planar, to form a R22(7)
hydrogen bonded ring with a supporting weak C–H/O
hydrogen bond (Scheme 4c). If the two groups are not co-planar,
the heterosynthon forms a discrete D hydrogen bond (Scheme
4d). The predictability of the carboxylic acid/pyridine hetero-
synthon is supported by evidence in the literature. In the previous
work with isoniazid co-crystals, all 9 of the co-crystals feature
this heterosynthon.23 In the related co-crystal systems involving
isonicotinamide (39/40)36 and nicotinamide (20/22)37 together
with carboxylic acids, either of the two heterosynthons, D or
R22(7), repeatedly appears, with relative frequency shown in
brackets.
Results
The results are divided into two sets of compounds, those co-
crystals that are non-hydrates (Fig. 1) and those that are hydrates
(Fig. 2). The source of the water is the condensation reaction,
which means that there will always be a stoichiometric amount of
water present in the crystallization solution, apart from any
moisture that enters from the atmosphere or is present in the
solvent.
Expected synthons
In our previous work, by looking at only the carbohydrazide
functional group on the unaltered isoniazid, we searched the
CSD32 for the most common homosynthons exhibited by this
functional group in the absence of any other molecules or
functional groups in 47 structures. There, we found two chain
synthons, C(3) and C(4), and four ring synthons, R22(6), R22(10),
R33(10) and R24(10).23 These synthons are in agreement with
a more detailed study of hydrogen bonding synthons in organic
hydrazides.33 Upon conversion of the NH2 group to a N]C
group, to create a N-acylhydrazone functional group, the
number of hydrogen bond donors is reduced from three to one.
By combining the modified isoniazid with the acid mHBA, the
total number of hydrogen bond donors is three if the co-crystal
does not contain water (2–7, Table 3). For clarity, the hydrogen
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5699

Table 4 Geometrical parameters for hydrogen bonds in hydrated co-crystals 8–12
:(D–H/A)/ꢁ
Symmetry transformations
ꢀ
d(D–H)/A
ꢀ
ꢀ
Compound
d(H/A)/A
d(D/A)/A
8
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/O(4) (b)
N(2)–H(2)/O(1W) (c)
O(1W)–H(1W)/N(3) (d)
O(1W)–H(2W)/O(2) (e)
9
1.01(2)
0.95
0.88(2)
0.87(1)
0.87(2)
0.86(2)
1.62(2)
3.48
1.85(2)
2.12(2)
2.15(2)
1.90(2)
2.619(2)
3.929(2)
2.718(2)
2.978(2)
2.999(2)
2.767(2)
169(2)
112
171(2)
172(2)
163(1)
178(2)
—
—
ꢀx + 1, ꢀy, ꢀz + 1
—
x + 1, y, z
ꢀx + 2, ꢀy + 1, ꢀz + 1
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/O(4) (b)
N(2)–H(2)/O(1W) (c)
O(1W)–H(1W)/N(3) (d)
O(1W)–H(2W)/O(2) (e)
10
0.91(3)
0.95
0.85(2)
0.85(2)
0.87(2)
0.85(2)
1.73(3)
3.34
1.88(2)
2.15(2)
2.18(2)
1.93(2)
2.621(2)
3.808(2)
2.721(2)
2.985(2)
3.020(2)
2.778(2)
165(2)
113
170(2)
170(2)
163(2)
177(2)
—
—
ꢀx + 1, ꢀy, ꢀz + 1
—
x + 1, y, z
ꢀx + 2, ꢀy + 1, ꢀz + 1
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/O(4) (b)
N(2)–H(2)/O(1W) (c)
O(1W)–H(1W)/N(3) (d)
O(1W)–H(2W)/O(2) (e)
11
1.04(3)
0.95
0.92(3)
0.89(2)
0.86(3)
0.89(2)
1.57(3)
3.49
1.81(3)
2.06(2)
2.16(3)
1.93(3)
2.605(2)
3.919(2)
2.717(2)
2.936(2)
3.017(2)
2.809(2)
172(2)
110
169(2)
168(2)
173(2)
173(2)
—
—
ꢀx + 1, ꢀy, ꢀz + 1
—
x + 1, y, z
ꢀx + 2, ꢀy + 1, ꢀz + 1
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/O(4) (b)
N(2)–H(2)/O(1W) (c)
O(1W)–H(1W)/N(3) (d)
O(1W)–H(2W)/O(2) (e)
12
1.11(2)
0.95
0.88(2)
0.95(2)
0.87(2)
0.88(2)
1.50(2)
3.14
1.80(2)
2.62(2)
2.14(2)
1.92(2)
2.602(2)
3.663(2)
2.679(2)
3.179(2)
2.991(2)
2.786(2)
168(2)
116
172(2)
118
166(2)
170(2)
—
—
ꢀx + 1, ꢀy, ꢀz + 1
—
x + 1, y, z
ꢀx + 2, ꢀy + 1, ꢀz + 1
O(1)–H(1)/N(1) (a)
C(10)–H(10)/O(2)
O(3)–H(3)/O(4) (b)
N(2)–H(2)/O(1W) (c)
O(1W)–H(1W)/N(3) (d)
O(1W)–H(2W)/O(2) (e)
0.88(3)
0.95
0.95(3)
0.89(2)
0.85(2)
0.87(2)
1.77(3)
3.61
1.80(3)
1.96(2)
2.21(2)
1.88(2)
2.632(2)
4.025(2)
2.726(2)
2.840(2)
3.027(2)
2.741(2)
167(2)
110
165(2)
172(2)
161(2)
171(2)
—
—
ꢀx + 1, ꢀy, ꢀz + 1
—
x + 1, y, z
ꢀx + 2, ꢀy + 1, ꢀz + 1
bond formed by the carboxylic acid H(1) is marked with a bold a,
the phenol H(3) marked with a bold b, and the amide H(2) by
a bold c. For the hydrated co-crystals 8–12 (Table 4), the letters
d (H(1W)) and e (H(2W)) are used in addition for the two H’s on
the water molecule. When assigning the graph set notations, we
will only use the a, b, c, d and e hydrogen bonds. Any C–H/O
hydrogen bonds are shown but are not included in the
designation.
crystals 2 and 6. Group 1’s all have the same hydrogen bonding
patterns as well as similar unit cell parameters. The structural
description for this group will be based on co-crystal 3. Co-
crystal 3, which was described briefly before,23 consists of two
hydrogen bond patterns: it has a C22(15) and a C(4) chain (Fig. 3).
The larger chain is made up of the COOH/Npyr (a) and the
OH/N]C (b) hydrogen bond in between, and extends along
the b-axis (molecules are related by a two-fold screw axis). This
chain, using the two heterosynthons, forms the intermolecular
attraction between the two different molecules being co-
crystallized. The third hydrogen bonded interaction is the
C(4) N(2)–H(2)/O(4) (c) homosynthon between the modified
isoniazid (N0-(butan-2-ylidene)isonicotinohydrazide), forming
a chain along the c-axis (molecules are related by a c-glide)
(Fig. 4).
Attempted co-crystallization of unmodified isoniazid with mHBA
The preparation of the co-crystal between isoniazid and mHBA
(1) was attempted from a large number of solvents, yet only the
individual components ever crystallized out. Attempts employ-
ing liquid-assisted grinding were also unsuccessful (see ESI†).
Co-crystal 4 has the same hydrogen bonded interactions (see
Table 3 for comparison). However, in co-crystal 5 with the same
C22(15) pattern, one can see the steric influence of the chosen
modifier (cyclooctane) as it sterically hinders the amide proton
H(2) from approaching the carbonyl O(4). The donor–acceptor
The unhydrates 2–7 using modified isoniazid
Co-crystals 2 and 3 were reported previsously,23 but are here
included together with relabelled Ortep style diagrams to aid in
comparison with the related co-crystals reported in this work.
The unhydrated co-crystals 2–6 also share common structural
features (co-crystal 7 will be described separately). The 5 co-
crystals 2–6 can be subdivided into two structural groupings.
Group 1 consists of co-crystals 3, 4 and 5, and group 2 of co-
ꢀ
distance for this C(4) chain is 3.740(2) A but one can still clearly
see the amide groups pointing at each other to form a C(4) chain
with an angle of 160(2)ꢁ (Fig. 5). One might describe this situa-
tion as the modifying group ‘‘masking’’ the other hydrogen
bonding functionality of the modified isoniazid molecules. Note
5700 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Fig. 4 Two views of the C(4) homosynthon formed by the N(2)–H(2)/
O(4) (c) hydrogen bond in co-crystal 3. Note that the butan-2-ylidene
modifier has no influence on the C(4) homosynthon. All it has done is
remove the two amine H atoms found on the original isoniazid molecule.
Fig. 3 Two views of the C22(15) chains formed by the COOH/Npyr (a)
and the OH/N]C (b) hydrogen bonds in co-crystal 3 that has butan-2-
ylidene as the modifier group. Note that the butan-2-ylidene modifier has
no influence on the :N(3) lone pair.
one modified isoniazid molecule (N0-(diphenylmethylene) iso-
nicotinohydrazide) and one half a mHBA acid molecule. This is
unusual, as the mHBA acid does not have a molecular centre of
inversion. However, in this case, the mHBA acid molecule
crystallizes with static disorder around a centre of inversion. The
mHBA molecule has the carboxylic acid group and four C atoms
of the aromatic ring in the asymmetric unit. The carboxylic acid
functions as both the carboxylic acid group as well as the phenol
group. The two missing C atoms of the aromatic ring are
generated by the O(2) atom and the C(7) atom of the carboxylic
acid functional group. In other words, the mHBA molecule exists
in two different orientations in between the modified isoniazid,
shown clearly in Fig. 10a and b. The cause of this unusual
arrangement is the modifying group. The benzophenone has
made the N(3) atom completely inaccessible for the phenol as an
acceptor group. Now, the carbonyl O(4) of the amide is still
accessible, but instead the phenol H(3) hydrogen bonds to the
pyridine ring, in an identical fashion to the carboxylic acid. This
requires the H atom of the phenol to point away from the
carboxylic acid ring, thus creating an eight-membered chain
H(1)–O(1)–C(7)–C(1)–C(1)i–C(7)i–O(1)i–H(1)i with inversion
symmetry [(i): ꢀx + 3 , ꢀy + 1 , ꢀz + 1]. In addition, the amide
H(2) has now been completely masked and hence no C(4)
homosynthon is formed. Hence, we now have a zero-dimensional
crystal packing due to the steric bulk of the modifying
group masking the H(2) donor and the lone pair of :N(3) of the
amide group. The carbonyl O(4) has two lone pairs, which are
accessible but due to packing effects, they are not used.
the relative arrangement of the modified isoniazid molecules
(N0-(cyclooctylidene)iso-nicotinohydrazide), which have over-
lapping pyridine rings but with the modifiers apart (Fig. 6). This
arrangement is also seen in 4 and 5.
Group 2 contains co-crystals 2 and 6. The former has an
isolated R44(30) ring formed by the COOH/Npyr (a) and the
OH/N]C (b) hydrogen bonds, instead of a C22(15) chain
(Fig. 7). Nonetheless, it still features the C(4) chain as it has the
smallest modifying group (acetone). Co-crystal 6 is similar. The
ring is R44(28) instead as the phenol hydrogen bonds to the amide
oxygen atom O(4) instead of the hydrazone N(3) (Fig. 8). In this
case, the bulky modifying group, 40-methylacetophenone, exerts
a steric influence, which prevents the phenol group from
hydrogen bonding with N(3) (Fig. 8). Co-crystal 6 also has a C(4)
homosynthon, but as in the case of co-crystal 5, the steric bulk of
the modifying group hinders the approach of the amide H(2) to
the carbonyl O(4), resulting in a non-linear close contact of 135
(2)ꢁ. The modified isoniazid molecules in the chain have a parallel
orientation to each other, different to the overlapping orienta-
tion as seen in 4 and 5 (Fig. 9).
Isoniazid modified with benzophenone (7)
Co-crystal 7 has the bulkiest modifying group, benzophenone,
and this has a significant effect on the hydrogen bonding and the
crystal packing. The asymmetric unit of this co-crystal contains
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5701

Fig. 7 View of the R44(30) ring formed by the COOH/Npyr (a) and the
OH/N]C (b) hydrogen bonds in co-crystal 2 that has propan-2-ylidene
as the modifier group. Note that the propan-2-ylidene modifier has no
influence on the :N(3) lone pair.
group (N(2)–C(13)–C(8)–C(9A): 2.6ꢁ) and the mHBA molecule
to form the R22(7) heterosynthon; this is the scenario when the
carboxylic acid hydrogen bonds to the pyridine. When the
phenol instead hydrogen bonds to the pyridine, the alternative
conformation (B) of the pyridine ring is adopted, where the
torsion angle is 30.9ꢁ (N(2)–C(13)–C(8)–C(9B)), and the pyridine
ring is not co-planar with either the N-acylhydrazone group or
the mHBA molecule.
The hydrates 8–12 using modified isoniazid
The remaining five co-crystals fall into the hydrated co-crystal
category. In these structures the reaction side product of the
modifying group reaction was incorporated into the co-crystal-
lization reaction and the crystal lattice. Co-crystals 8–12 can all
be grouped together. They all show the same hydrogen bonding
Fig. 5 View of the C(4) homosynthon in ball and stick (a) and space
filling (b) representations of co-crystal 5. The yellow circle shows the
increased separation between the O(4) acceptor and the H(2) donor due
to the modifier group cyclooctylidene. The DHA acceptor angle is
160(2)ꢁ and the D/A distance 3.740(2) A.
ꢀ
The conformation of the pyridine ring is also affected by the new
hydrogen bonding. The ring exists in two different conforma-
tions relative to the rest of the modified isoniazid molecule, and
in addition, to the acid molecule. In one conformation (A), the
pyridine ring is essentially co-planar with the N-acylhydrazone
Fig. 8 Ball and stick (a) and space filling (b) views of the R44(28) ring
formed by the COOH/Npyr (a) and the OH/N]C (b) hydrogen bonds
in co-crystal 6 that has 1-p-tolylethylidene as the modifier group. Note
that 1-p-tolylethylidene modifier now has a steric or masking influence on
the :N(3) lone pair, such that the H(3) hydrogen atom from mHBA now
hydrogen bonds to O(4).
Fig. 6 The relative arrangement of two modified isoniazids in co-crystal
5, where the steric bulk of the cyclooctylidene modifiers has the pyridine
rings overlapping, similarly to the modified isoniazid shown in Fig. 4.
5702 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Fig. 10 Two equivalent representations (a and b) of the hydrogen
bonding interactions of the modified isoniazid with mHBA. (a) shows
conformation A and B, where the heterosynthon COOH/Npyr (a) is part
of the asymmetric unit and hydrogen bonds to the modified isoniazid.
The heterosynthon OH/Npyr (b) hydrogen bonds to the symmetry
equivalent modified isoniazid [symmetry operator: ꢀx + 3 , ꢀy + 1 , ꢀz +
1]. (b) shows the reverse scenario, where the OH/Npyr (b) heterosynthon
is part of the asymmetric unit. (c) The C(4) homosynthon is not observed,
as the modifier diphenylmethylene masks both the :N(3) lone pair and the
H(2) donor.
b hydrogen bonds. The second, smaller ring, R22(26), uses the
O(1)–H(1)/N(1) (a), the N(2)–H(2)/O(1W) (c) and
O(1W)–H(1W)/O(2) (d) hydrogen bonds. These two rings
alternate to form a 1-D ribbon (Fig. 11a). Adjacent ribbons are
connected by the O(1W)–H(2W)/O(2) (e) hydrogen bond
(Fig. 11b). There are no C(4) homosynthons in the hydrated co-
crystals, with the amide H(2) hydrogen bonding in each instance
to the water molecule.
Fig. 9 Two views (a) and (b) of the C(4) homosynthon in ball and stick,
and space filling (c) representations of co-crystal 6. The yellow circle
shows the increased separation between the O(4) acceptor and the H(2)
donor due to the modifier group 1-p-tolylethylidene. The DHA acceptor
Discussion
The aim of this work was to prepare co-crystals with isoniazid
and mHBA, while concurrently modifying the hydrogen bonding
functionality of the isoniazid molecule by adding stoichiometric
amounts of simple ketones to the co-crystallization solution. The
amount of ketone added was not constant but at least a stoi-
chiometric 1 : 1 ratio to the isoniazid was present each time. The
ketones functioned as the co-crystallization solvent in the prep-
aration of 2, 3 and 4, respectively acetone, 2-butanone and
cyclopentanone, and were used in excess. These are examples in
which the modifying group acts as a solvent and no other
angle is now 135(2)ꢁ and the D/A is 3.736(2) A.
ꢀ
features and crystal packing. Except for 11, all crystallize in the
ꢁ
triclinic space group P1 and have similar unit cell parameters.
Co-crystal 11 has similar a and b cell axes, but with a doubling of
c in the monoclinic space group P21/n (see Table 2). The
hydrogen bonded interactions consist of two rings. The larger
ring is identical to the R44(28) ring seen in 6 and uses the a and
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5703

employed, the :N(2) lone pair is unmasked and acts as an
acceptor for the phenol of the mHBA. The N(2)–H(2) donor is
further away from the ketones but is masked by the modifiers
cyclooctylidene (5), 1-p-tolylethylidene (6) and diphenyl-
methylene (7). This observation is based on the increase in
distance from the H(2) donor to the O(4) carbonyl acceptor of
adjacent modified isoniazids, such that the C(4) homosynthon
becomes a close contact in 5 and 6 and is completely absent in 7.
The C(4) homosynthon is still clearly observed in co-crystals with
less bulky ketones, as in 2, 3 and 4. The C(7) homosynthon is
never observed, possibly because of the greater acidity and donor
ability of the carboxylic acid hydrogen of mHBA over the amide
H. Our most successful attempt to modify all the donors of the
isoniazid and the acceptor lone pair on :N(3) is co-crystal 7.
Fig. 10c clearly demonstrates how the positions of the two phenyl
rings on the modifying diphenylmethylene group fill much of the
space around the H(2) donor and the :N(3) lone pair, making
them inaccessible and completely masking their potential
hydrogen bonding interactions. A summary of the modifying
and masking effect that best illustrates the concept of covalent
assistance to supramolecular synthesis is presented in Scheme 5.
When making these observations on co-crystals 2–7, where the
position of the modifying group relative to the H(2) donor and
the :N(3) acceptor has a masking influence, that influence is
dependent on the conformation of the modifying group relative
to the rest of the isoniazid molecule. There is free rotation around
the N(2)–N(3) single bond. In the solid state, we find that the
modifying group is rotated such that the N]C double bond is
syn to the amide H(2) proton (see torsion angles H(2)–N(2)–
N(3)–C(14) in Table 5). As a consequence, the torsion angle
C(13)–N(2)–N(3)–C(14), listed in Table 5, has its modulus around
180ꢁ. This is in fact a very desirable scenario for modifying the
hydrogen bonding, as itallows the steric influenceto be maximised
when attempting to mask the H(2) donor. The reason for this
conformation, and the reason for not observing the anti case, is
most likely the presence of the carbonyl O(4) group, which is
bulkier than the H atom. Molecular modelling calculations done
on the N0-(propan-2-ylidene)isonicotinohydrazide molecule as
a model system show that the anti conformation is much more
stable than the syn conformation.x Nonetheless, there is no
apparent reason to rule out other possible conformations, i.e. it is
theoretically possible to obtain polymorphs of these co-crystals
that could have other conformations due to packing effects. These
polymorphs could conceivably show different hydrogen bonding
patterns due to the different steric consequences.
Fig. 11 (a) and (b) show the generic hydrogen bonding and packing
architecture for the hydrated co-crystals 8–12; shown here is co-crystal 8.
reagents are added. This was possible as the ketones have
a sufficiently low boiling point to act as crystallization solvents
for slow evaporation, and the isoniazid and mHBA molecules are
both soluble in them. For the co-crystals 5, 6 and 7, the ketones
or aldehydes used were cyclooctanone, 40-methylacetophenone
and benzophenone respectively. In these three cases, the modi-
fiers are not suitable as solvents due to the lack of solubility and
high boiling points (benzophenone is a solid and hence
completely unsuitable). Stoichiometric amounts of the modifiers
were added and all the components dissolved in methanol. In all
crystallizations, the condensation reaction was induced by
keeping the solutions hot for a few hours. Fortunately, the
modified isoniazids remain sufficiently soluble to co-crystallize
out with the mHBA, even when large groups are added e.g.,
benzophenone in 7. Co-crystals 2–7 all have the two amine
hydrogen bond donors of the carbohydrazide group removed
from the isoniazid molecule. In this sense, the hydrogen bonding
ability of the isoniazid is reduced, i.e. modified. This modification
was already achieved by using acetone in 2. The reason for using
increasingly larger modifiers was to observe the influence of the
chosen modifying groups on the third hydrogen bond donor, the
amide H(2), as well as the lone pair of the :N(3) atom. The :N(3)
lone pair is adjacent to the site where the modifiers are added,
and hence can be easily masked by a suitably sized molecule. This
occurs by using a bulky phenyl ring such as in 40-methyl-
acetophenone and benzophenone. In all the other modifiers
x Geometry optimizations were performed with Gaussian 0338 using the
a density functional theory (DFT) type of
B3LYP method,39
calculation with hybrid functionals, and the 6–31+g(d,p) basis set.40
These show that the barrier to rotation around the N(2)–N(3) bond is
32.9 kJ molꢀ1, with the syn-conformation (C(13)–N(2)–N(3)–C(14) ¼
0ꢁ) 25.0 kJ molꢀ1 higher in energy than the anti conformation (180ꢁ)
found in the crystal structure. According to the Boltzmann equation
this difference in energy would result in only 0.0040% of the
syn-conformation being found in the gas phase. Interestingly the lowest
energy conformations are staggered ones, with C(13)–N(2)–N(3)–C(14)
torsion angles of ꢃ70ꢁ, 7.8 kJ molꢀ1 lower in energy than the crystal
structure conformation. However, this corresponds to a conformation
around N(2) slightly distorted from planarity, which may lead to
a reduction in the strength of the N–H/O hydrogen bond, and is thus
not energetically favorable within the crystal structure.
5704 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

Scheme 5 Summary of the modifying and masking effect of four chosen co-crystals that best illustrate the covalent assistance to supramolecular
synthesis concept. The increasing steric size of the modifying group allows for the selective masking of the H(2) donor and N(3) acceptor in co-crystals 3,
5, 6 and 7.
The masking of the lone pair on the :N(3) atom is less
influenced by the conformation of the modifying group than by
the actual condensation reaction when the modifying group is
added. In this case, the unsymmetrical ketones always add such
that the lone pair is syn to the bulkiest part of the modifying
group, i.e. the ethyl part in 2-butanone, or the aromatic ring in
40-methylacetophenone. Hence, the masking of the lone pair can
be achieved immediately by adding bulky unsymmetrical
ketones.
The hydrate co-crystals were not intended or desired results.
The nine ketones were chosen such that the steric effects could be
observed for all the different modifiers. The attempts with
cyclohexanone, cycloheptanone and 40-methylbenzaldehyde
could therefore not be included in the series of unhydrated co-
crystals. Nonetheless, all five hydrated co-crystals contribute to
the in situ modification concept proposed here.
Co-crystals 8 and 9, which are the related hydrated versions of
the unhydrated co-crystals 2 and 3, were prepared by carrying
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5705

Table 5 Summary of modifications in unhydrated co-crystals 2–7
2
3
4
5
6
7
Modifying group
Propan-2-ylidene Butan-2-ylidene Cyclopentylidene Cyclooctylidene 1-p-Tolylethylidene Diphenylmethylene
Removal of NH2 donor
Masking of N(2)–H(2) donor
Masking of :N(3) lone pair
C(4) homosynthon observed
Yes
No
No
Yes
Yes
No
No
Yes
No
No
Yes
Yes, R22(7)
N(3)
C22(15)
11.7(2)
Yes
A bit
No
Close contact
Yes, R22(7)
N(3)
C22(15)
1.1(2)
Yes
A bit
Yes
Close contact
Yes, D
O(4)
R44(28)
1.9(2)
Yes
Yes
Yes
No
Yes, R22(7)
N(1)
None
ꢀ2.5(2)
Yes
O(1)–H(1)/N(1) heterosynthon Yes, R22(7)
Yes, R22(7)
N(3)
C22(15)
8.9(2)
O(3)–H(3)/X
Hydrogen bond patterns
N(3)
R44(30)
34.3(2)
Mod. isoniazid:
H(2)–N(2)–N(3)–C(14)/ꢁ
Mod. isoniazid:
ꢀ167.3(1)
ꢀ33.0(2)
177.7(2)
176.2(1)
ꢀ179.9(1)
ꢀ10.4(2)
ꢀ177.1(1)
ꢀ6.3(2)
179.5(2)
C(13)–N(2)–N(3)–C(14)/ꢁ
Mod. isoniazid:
ꢀ29.4(3)
ꢀ26.0(2)
3.3(6) C(9A),
N(2)–C(13)–C(8)–C(9)/ꢁ
31.7(6) C(9B)
175(1)a
mHBA: O(1)–C(7)–C(1)–C(2)/ꢁ ꢀ173.8(1)
Estimated from O(1)–C(7)–C(1)–C(1)i angle, where (i): 3 ꢀ x, 1 ꢀ y, 1 ꢀ z.
12.2(3)
13.5(2)
16.2(2)
7.2(2)
a
out the crystallization without heating to test if it is possible to
obtain co-crystals 2 and 3 (prepared originally by heating the
vial). The condensation reaction did occur but led to hydrated
co-crystals 8 and 9 of the modified isoniazid and mHBA. The
hydrated co-crystals 10–12 were prepared similarly (by heating)
in the same way as 2–7, yet water co-crystallized with them. As
mentioned above, water is a by-product of the aptly named
‘‘condensation’’ reaction and hence is unavoidably present when
the co-crystallization process starts. One possible strategy for
preventing the inclusion of water in the co-crystal structure is the
addition of molecular sieves to the crystallization vial as was
done previously when preparing both hydrated and unhydrated
ammonium sulfonate salts.41 The crystal packing when there is
water present has so far led to isostructurality of the five
hydrated co-crystals. Even in the presence of the water molecule,
our primary heterosynthon, the O(1)–H(1)/N(1) hydrogen
bond, is observed and connects the mHBA with the modified
isoniazid. Additionally, the heterosynthon observed in 6,
O(3)–H(3)/O(4), is observed in all five hydrated co-crystals.
Hence, all five hydrated co-crystals have the identical R44(28) ring.
The similarity ends after that. The H(2) donor, which in the
unhydrated co-crystals was used in the C(4) homosynthon, now
hydrogen bonds to the water. In these hydrated co-crystals, the
modifying group is not able to prevent (by masking) the
hydrogen bond to the water molecule, mainly as the small size of
the water molecule allows it to fit in. The water molecule plays
the same hydrogen bonding role in all five structures regardless
of the type of modifying group used (see Table 4). The two
hydrogen atoms on the water hydrogen bond to N(3) and O(2). It
is to be seen in future work if the same architecture is observed
with a larger variety of modifying groups. Conformationally, the
geometry of the modified isoniazids is identical to the unhydrated
co-crystals, where the N]C is syn to the amide H(2) (see Table 6
for relevant torsion angles). Curiously, in all five hydrated co-
crystals, the :N(3) lone pair is not used as an acceptor; instead the
carbonyl O(4) acts as an acceptor to the phenol H(3). This non-
use cannot be said to be due to the modifiers, as we have seen in 2
and 3 that it is possible for N(3) to act as an acceptor for the H(3)
hydrogen bond donor.
The conformation of the mHBA acid molecule in the co-
crystals is consistently planar in terms of the rotation around the
C(1)–C(7) bond. The modulus of the torsion angle O(1)–C(7)–
C(1)–C(2) is syn-periplanar in all co-crystals except for 2 and 7.
This means that the acid H(1) is on the same side as the phenol
H(3). In 7, the two H’s are related by a centre of inversion and are
necessarily anti to each other.
Table 6 Summary of modifications in hydrated co-crystals 8–12
8
9
10
11
12
Modifying group
Removal of NH2 donor
Propan-2-ylidene
Yes
Yes
No
Yes, R22(7)
Yes
R44(28), R66(26)
ꢀ24.9(2)
161.3(2)
ꢀ4.9(2)
ꢀ7.4(2)
Butan-2-ylidene
Yes
Yes
No
Yes, R22(7)
Yes
R44(28), R66(26)
ꢀ27.3(2)
157.8(2)
ꢀ9.6(2)
1.2(2)
Cyclohexylidene
Yes
Yes
No
Yes, R22(7)
Yes
R44(28), R66(26)
ꢀ32.7(2)
150.9(2)
ꢀ10.3(2)
0.6(2)
Cycloheptylidene
Yes
Yes
No
Yes, R22(7)
Yes
R44(28), R66(26)
ꢀ37.7(7)
150.3(1)
ꢀ22.1(2)
7.8(2)
4-Methylbenzylidene
Yes
Yes
No
Yes, R22(7)
Yes
R44(28), R66(26)
ꢀ3.7(2)
176.1(1)
ꢀ4.9(2)
ꢀ4.5(2)
Masking of :N(3) lone paira
C(4) homosynthon
O(1)–H(1)/N(1) heterosynthon
O(3)–H(3)/O(4)
Hydrogen bond patterns
Mod. isoniazid: H(2)–N(2)–N(3)–C(14)/ꢁ
Mod. isoniazid: C(13)–N(2)–N(3)–C(14)/ꢁ
Mod. isoniazid: N(2)–C(13)–C(8)–C(9)/ꢁ
mHBA: O(1)–C(7)–C(1)–C(2)/ꢁ
a
Not clear if this is due to packing effects or due to the water molecule preferentially hydrogen bonding to the carbonyl O atom.
5706 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011

6 (a) G. P. Stahly, Cryst. Growth Des., 2009, 9, 4212–4229; (b)
€
Conclusion
C. B. Aakeroy and D. J. Salmon, CrystEngComm, 2005, 7, 439–448;
(c) M. J. Zaworotko, Cryst. Growth Des., 2007, 7, 4–9; (d)
S. L. Childs and M. J. Zaworotko, Cryst. Growth Des., 2009, 9,
4208–4211.
An additional concept and tool in supramolecular synthesis has
been introduced with which to engineer (modify) the assembly of
multi-component molecular complexes. By making use of
modifying groups, we have exerted some measure of control on
how two molecules co-crystallize, by substantially reducing the
number of hydrogen bonding interactions, primarily by a cova-
lent reaction and secondarily by the steric influence, masking the
remainder of the hydrogen bonding donors and acceptors on the
supramolecular reagent isoniazid. What makes this modifying
process distinct from previous ones is that it is an in situ process,
where the modifying process takes place in the same vessel as the
co-crystallizing process, and finally, where even the solvent
molecules are used as modifying groups.
7 G. R. Desiraju, Acc. Chem. Res., 2002, 35, 565–573.
€
8 (a) C. B. Aakeroy, J. Desper and B. A. Helfrich, CrystEngComm,
2004, 6, 19–24; (b) J. A. McMahon, J. A. Bis, P. Vishweshwar,
T. R. Shattock, O. L. McLaughlin and M. J. Zaworotko, Z.
Kristallogr., 2005, 220, 340–350; (c) J. A. Bis, O. L. McLaughlin,
P. Vishweshwar and M. J. Zaworotko, Cryst. Growth Des., 2006, 6,
2648–2650.
9 (a) R. D. B. Walsh, M. W. Bradner, S. Fleischmann, L. A. Morales,
ꢂ
B. Moulton, N. Rodrıguez-Hornedo and M. J. Zaworotko, Chem.
Commun., 2003, 186–187; (b) J. A. Bis and M. J. Zaworotko, Cryst.
Growth Des., 2005, 5, 1169–1179.
€
10 C. B. Aakeroy, A. M. Beatty and B. A. Helfrich, J. Am. Chem. Soc.,
2002, 124, 14425–14432.
€
11 C. B. Aakeroy, I. Hussain, S. Forbes and J. Desper, CrystEngComm,
2007, 9, 46–54.
12 T. R. Shattock, K. K. Arora, P. Vishweshwar and M. J. Zaworotko,
Cryst. Growth Des., 2008, 8, 4533–4545.
Acknowledgements
€
13 C. B. Aakeroy, J. Desper, E. Elisabeth, B. A Helfrich, B. Levin and
J. F. Urbina, Z. Kristallogr., 2005, 220, 325–332.
This work was supported in part by Grant No. 2004118 from
the United States—Israel Binational Science Foundation
(Jerusalem). A.L. thanks the South African National Research
Foundation for a postdoctoral scholarship (SFP2007070400002)
and the Oppenheimer Memorial Trust for financial support. The
authors thank Dr Dmitry Mogilyanski (Ben-Gurion University)
€
14 C. B. Aakeroy, J. Desper and M. E. Fasulo, CrystEngComm, 2006, 8,
586–588.
15 C. B. Aakeroy, J. Desper and J. F. Urbina, Chem. Commun., 2005,
€
2820–2822.
16 M. C. Etter, Acc. Chem. Res., 1990, 23, 120–126.
17 K. N. Lehane, E. J. A. Moynihan, N. Brondel, S. E. Lawrence and
A. R. Maguire, CrystEngComm, 2007, 9, 1041–1050.
18 S. Eppel and J. Bernstein, Cryst. Growth Des., 2009, 9, 1683–1691.
19 (a) B. R. Bhogala, S. Basavoju and A. Nangia, CrystEngComm, 2005,
7, 551–562; (b) P. Vishweshwar, A. Nangia and V. M. Lynch, Cryst.
Growth Des., 2003, 3, 783–790.
20 C. M. Reddy, L. S. Reddy, S. Aitipamula, A. Nangia, C.-K. Lam and
T. C. W. Mak, CrystEngComm, 2005, 7, 44–52.
21 L. Rajput and K. Biradha, CrystEngComm, 2009, 11, 1220–1222.
22 V. R. Thalladi, M. Debros, A. Gehrke, H.-C. Weiss and R. Boese,
Cryst. Growth Des., 2007, 7, 598–599.
€
and Dr Daniel Tobbens (University of Innsbruck) for powder
diffraction work, Dr Vincent Smith (University of Stellenbosch)
for help with refining co-crystal 7, Dr Catharine Esterhuysen
(University of Stellenbosch) for modelling assistance, and Prof.
Gautam Desiraju for insightful comments at a poster presenta-
tion of some of this work.
23 A. Lemmerer, J. Bernstein and V. Kahlenberg, CrystEngComm, 2010,
12, 2856–2864.
Notes and references
24 A. Lemmerer, J. Bernstein and V. Kahlenberg, CrystEngComm, 2011,
13, 55–59.
25 CrysAlisPro, Oxford Diffraction Ltd., Version 1.171.33.34d (release
27-02-2009 CrysAlis171.net).
26 XPREP2, Bruker, Version 6.14, Bruker AXS Inc., Madison,
Wisconsin, USA, 2003.
27 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837–838.
28 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2007,
64, 112–122.
29 L. J. Farrugia, J. Appl. Crystallogr., 2003, 30, 565.
30 A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13.
31 K. Brandenburg, J. Appl. Crystallogr., 1999, 32, 1028–1029.
32 F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 380–388.
33 R. Centore, A. Carella, A. Tuzi, A. Capobianco and A. Peluso,
CrystEngComm, 2010, 12, 1186–1193.
34 The database searches were done on the Cambridge Structural
Database, Ver. 5.30, November 2008 release, with Feb and May
2009 updates.32 The search query: molecular fragment isohydrazone
molecule. Filters: 3D coordinates determined, R factor # 0.075, no
powder structures, only organics, resultedin 123 hits which were
reduced to 55 by manual inspection by removing duplicate entries,
structures with solvents, and any host–guest complexes.
35 G. T. R. Palmore and J. C. MacDonald, The Role of Amides in the
Noncovalent Synthesis of Supramolecular Structures in Solution, at
Interfaces, and in Solids, in The Amide Linkage: Structural
Significance in Chemistry, Biochemistry, and Materials Science, ed.
A. Greenberg, C. M. Breneman and J. F. Liebman, Wiley-
Interscience, New Jersey, 2003, pp. 291–337.
1 (a) Organic Crystal Engineering: Frontiers in Crystal Engineering, ed.
E. R. T. Tiekink, J. J. Vittal and M. J. Zaworotko, John Wiley &
Sons, West Sussex, 2010; (b) Crystal Engineering: from Molecules
and Crystals to Materials, ed. D. Braga, F. Grepioni and A. G.
Orpen, Kluwer, Dordrecht, 2008; (c) Making Crystals by Design, ed.
D. Braga and F. Grepioni, Wiley-VCH, Weinheim, 2007; (d)
Frontiers in Crystal Engineering, ed. E. R. T. Tiekink and J. J.
Vittal, John Wiley
& Sons, Chichester, 2005; (e) Crystal
Engineering: the Design and Application of Functional Solids, ed. K.
R. Seddon and M. J. Zaworotko, Kluwer, Norwell, 1999; (f)
Crystal Engineering: the Design of Organic Solids, G. R. Desiraju,
Elsevier Scientific Publishers, Amsterdam, 1991.
€
2 (a) C. B. Aakeroy, N. R. Champness and C. Janiak, CrystEngComm,
2010, 12, 22–43; (b) R. J. Davey, K. Allen, N. Blagden, W. I. Cross,
H. F. Lieberman, M. J. Quayle, S. Righin, L. Seton and
G. J. T. Tidy, CrystEngComm, 2002, 4, 257–264; (c) G. R. Desiraju,
Chem. Commun., 2005, 2995–3001; (d) J. D. Wuest, Chem.
Commun., 2005, 5830–5837; (e) P. Erk, H. Hengelsberg,
M. F. Haddow and R. van Gelder, CrystEngComm, 2004, 6, 474–
483; (f) D. Braga, L. Brammer and N. R. Champness,
€
CrystEngComm, 2005, 7, 1–19; (g) C. B. Aakeroy and A. M. Beatty,
Aust. J. Chem., 2001, 54, 409–421; (h) C. B. Aakeroy, Acta
€
Crystallogr., Sect. B: Struct. Sci., 1997, 53, 569–586; (i)
L. R. MacGillivray, CrystEngComm, 2002, 4, 37–41; (j)
C. V. K. Sharma, Cryst. Growth Des., 2002, 2, 465–474.
3 L. Leiserowitz, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst.
Chem., 1976, 32, 775–802.
4 (a) D. Das and G. R. Desiraju, CrystEngComm, 2006, 8, 674–679; (b)
D. Das and G. R. Desiraju, Chem.–Asian J., 2006, 1, 231–244; (c)
D. Das, R. K. R. Jetti, R. Boese and G. R. Desiraju, Cryst. Growth
Des., 2003, 3, 675–681.
36 N. B. Bathori, A. Lemmerer, G. A. Venter, S. A. Bourne and
M. R. Caira, Cryst. Growth Des., 2011, 11, 75–87.
37 A. Lemmerer and J. Bernstein, CrystEngComm, 2010, 12, 2029–2033.
38 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr, T. Vreven,
5 S. S. Kuduva, D. Blaser, R. Boese and G. R. Desiraju, J. Org. Chem.,
2001, 66, 1621–1626.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 5692–5708 | 5707

K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi,
V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich,
A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz,
Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov,
G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson,
W. Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian 03,
Revision B.05, Gaussian, Inc., Pittsburgh, PA, 2003.
39 (a) A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652; (b) C. Lee,
W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785–789.
40 (a) W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys., 1972,
56, 2257–2261; (b) P. C. Hariharan and J. A. Pople, Theor. Chim.
Acta, 1973, 28, 213–222.
41 A. Lemmerer, S. A. Bourne and M. A. Fernandes, Cryst. Growth
Des., 2009, 9, 2265–2279.
5708 | CrystEngComm, 2011, 13, 5692–5708
This journal is ª The Royal Society of Chemistry 2011