Dominance of charge-assisted hydrogen bonding on short contacts and structures that crystallize with Z′ > 1

Article abstract of DOI:10.1021/cg200751m

An enantiotropic pair of polymorphs of (H₂-o-PDA)4(H ₅O₂⁺)(SO₄^2-) ₄(HSO₄^-)·2H₂O (1) interconvert at 140 - 142 K with a concomitant change in Z′. The details of the structural change are studied by single-crystal X-ray and neutron diffraction which shows that both forms are close packed with the difference resulting from the linearization of hydrogen bonds at low temperature as part of an extensive charge-assisted hydrogen-bonding network. The related (H ₂-o-PDA²⁺)₂(SO₄^2-) ₂· 3H₂O (2) has also been characterized by single-crystal neutron diffraction and exhibits a disordered "flip-flop" water chain and short nonbonded O · · · O contacts enforced by NH₃ · · · OSO₃^2- hydrogen-bonded bridges. A total of seven other related salts have been isolated and structurally characterized revealing a surprisingly high incidence of Z′ > 1 structures as a result of frustration in the charge-assisted hydrogen-bonded network.

Full text of DOI:10.1021/cg200751m

ARTICLE
pubs.acs.org/crystal
Dominance of Charge-Assisted Hydrogen Bonding on Short Contacts
and Structures that Crystallize with Z0 > 1
Kirsty M. Anderson,† Andrꢀes E. Goeta,† Jessica E. Martin,† Sax A. Mason,‡ Garry J. McIntyre,†,‡,#
Benedict C. R. Sansam,† Clive Wilkinson,† and Jonathan W. Steed*,†
†Durham University, Department of Chemistry, South Road, Durham, DH1 3LE, U.K.
‡Institut Laue-Langevin, 6 Rue Jules Horowitz, BP156, 38042, Grenoble, France
S Supporting Information
b
+
ABSTRACT: An enantiotropic pair of polymorphs of (H2-o-PDA2+)4(H5O2 )(SO42ꢀ)4(HSO4ꢀ) 2H2O (1) interconvert at
3
140ꢀ142 K with a concomitant change in Z0. The details of the structural change are studied by single-crystal X-ray and neutron
diffraction which shows that both forms are close packed with the difference resulting from the linearization of hydrogen bonds at
low temperature as part of an extensive charge-assisted hydrogen-bonding network. The related (H2-o-PDA2+)2(SO42ꢀ)2 3H2O
3
(2) has also been characterized by single-crystal neutron diffraction and exhibits a disordered “flip-flop” water chain and short
+
nonbonded O O contacts enforced by NH3
OSO32ꢀ hydrogen-bonded bridges. A total of seven other related salts have
3 3 3
3 3 3
been isolated and structurally characterized revealing a surprisingly high incidence of Z0 > 1 structures as a result of frustration in the
charge-assisted hydrogen-bonded network.
’ INTRODUCTION
second-order phase transition at 232 K to a higher-temperature
Z0 = 1 form accompanied by a small change in the host
conformation) to more subtle effects such as in 1-hydroxy-
1,1,3,3,3-pentaphenyldisiloxane16 where there is a phase transi-
tion observed between 100 and 150 K from Z0 = 2 to Z0 = 4 as
increased thermal motion of the coordinated phenyl groups
lowers the long-range symmetry of the structure. A range of
other subtle temperature-dependent structural transformations
have also been studied that do not necessarily result in a change
in Z0. Examples include proton transfer, as exemplified by the
work on acid-dimer systems by Wilson and co-workers17 where
temperature changes lead to movement of the proton across the
centrosymmetric acid-dimer motif.
We have shown that the combination of a molecule with a
resolved chiral center and a strongly directional supramolecular
synthon with a preference for centrosymmetry leads to frustrated
structures where in order to preserve the dimer motif more than
one crystallographically independent molecule must be present
in the asymmetric unit unless the system can adopt a less efficient
C2 packing mode.18 The percentage of structures crystallizing
with Z0 > 1 in systems like this is as high as 75%, compared to the
8.8% for the Cambridge Structural Database (CSD)19,20 as a
whole. In addition to conflicts between molecular properties
such as chirality and the formation of robust supramolecular
synthons, competing intermolecular interactions can also result
in crystallization with Z0 > 1. We have shown that there is a
marked tendency toward decreased-symmetry structures for
molecules engaged in more than one competing intermolecular
interaction and have observed the proportion of structures
Structures that crystallize with Z0 > 1, that is, more than one
molecule in the asymmetric unit, are of interest in many areas of
crystal engineering and crystal nucleation studies,1ꢀ7 not least in
the burgeoning field of crystal structure prediction where the
ability to predict whether a given molecule is likely to crystallize
with high Z0 can narrow the initial search criteria, offering large
computational advantages.8ꢀ10
Structures that crystallize with high Z0 are also of interest in the
field of crystal growth, and indeed there has been considerable
debate surrounding the phenomenon, particularly whether all
structures with Z0 > 1 are just metastable forms of known or
undiscovered thermodynamically stable Z0 = 1 polymorphs.3
This suggestion presupposes that for all structures with Z0 > 1
there is a lower energy Z0 = 1 form waiting to be discovered, but it
is hard to imagine that this postulate is true across all structures
particularly when considering concepts such as frustration,11 and
the fact that in some known polymorphic compounds Z0 > 1
forms have been shown to be more stable than their Z0 = 1
counterparts.12 It is also possible that enantiotropic systems exist
in which the stability order of Z0 = 1 and Z0 > 1 polymorphs
changes as a function of temperature.
Polymorphism where the polymorphs have different values of
Z0 are not common but do exist.13 In the majority of cases we
have identified, polymorphs with different Z0 values are obtained
using different crystal-growth conditions or media. There are also
some structures where the value of Z0 is dependent on tempera-
ture, although there has been some discussion on whether these
should be referred to as polymorphic transitions or modulated
structures.14 The structural effects arising from phase transfor-
mations where Z0 is modified range from changes in conforma-
tion such as in 9-phenylfluoren-9-ol 4-methylcyclohexylamine
clathrate15 (where the low-temperature Z0 = 3 form undergoes a
Received: June 14, 2011
Revised:
September 27, 2011
Published: September 27, 2011
r
2011 American Chemical Society
4904
dx.doi.org/10.1021/cg200751m Cryst. Growth Des. 2011, 11, 4904–4919
|

Crystal Growth & Design
ARTICLE
Figure 1. Portions of two X-ray diffraction patterns of 1 at 120 and 150 K clearly showing the additional diffraction spots at 120 K due to doubling
of the c axis.
Figure 2. DSC trace of 1 showing endotherm corresponding to the phase transition (138 K), glass transition, and large crystallization followed by
melting endotherm (180 K) due to surface sulfuric acid solution.
crystallizing with Z0 > 1 to be more than 50% in some such cases.11
In both of these examples, it is likely that any alternative Z0 = 1
structure would not be optimized in terms of the intermolecular
interactions involved and hence would be less stable.
related to 1 to see the effect of changing charge, hydrogen-
bonding strength, and molecular shape on the supramolecular
synthon formation in these systems.
We have studied the frustration concept extensively in our
investigations into a series of aromatic polyamine salts where
frustration between strongly hydrogen-bonded networks and
πꢀπ stacking leads to structures with higher Z0.21 This frustra-
tion is particularly evident in the structure of the mixed sulfate/
hydrogen sulfate salt of doubly protonated o-phenylenediamine,
’ RESULTS AND DISCUSSION
Polymorphism in (H2-o-PDA2+)4(SO42ꢀ)3(HSO4
) 4H2O.
2
ꢀ
3
Crystals of compound 1 were grown by recrystallizing o-pheny-
lenediamine from aqueous sulfuric acid solution. The resulting
crystals exhibit a structure with eight crystallographically inde-
pendent o-phenylene diammonium cations and formally Z0 = 2 at
120 K in which the sulfate, hydrogen sulfate. and water molecules
form an extended 3D network encapsulating the (H2-o-PDA2+)
cations.21 In more detailed studies, we find that compound 1
undergoes a first-order phase transition between 140 and 142 K
corresponding to halving of the c axis and hence halving of Z0
(from 2 to 1) as the temperature is increased. This phase change
was initially monitored by X-ray techniques. The doubling of the
c axis is shown by the increased number of diffraction spots at low
temperature in Figure 1.
The observed phase change was confirmed by differential
scanning calorimetry (DSC) (Figure 2) which shows an en-
dotherm with an onset temperature of 138 K correspondingꢀto1
the Z0 = 2 to Z0 = 1 transition (apparent ΔH = 1.38 kJ mol
without correction for sulphuric acid, vide infra) and correspond-
ing exotherm on the cooling scan. The fact that the phase change
(H2-o-PDA2+)8(SO42ꢀ)6(HSO4ꢀ)4 8H2O (1) which crystal-
3
lizes with Z00 = 26 (i.e., 26 independent molecules or ions in
the asymmetric unit)22 including eight independent cations.
Despite the large value of Z00, the value of Z0 is formally 2 as
the smallest integer formula unit is (H2-o-PDA2+)4(SO42‑)3-
(HSO4ꢀ)2 4H2O.23 In this structure the sulfate, hydrogen
3
sulfate, and water molecules form a strongly hydrogen-bonded
three-dimensional (3D) network around the H2-o-PDA mol-
ecules; however, the X-ray data collected on the structure were
not of sufficient quality to explore fully the hydrogen-bonded
network.
In this paper we explore the effect of temperature on the
hydrogen bonding observed in 1 and study the resulting proton
transfer in more detail using variable-temperature neutron
diffraction. We also explore the frustration concept further by
altering both the anion and the cation in seven new compounds
4905
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Table 1. X-ray Data for Compounds 1 (Before and after
Phase Change) and 2
1(LT)
1(HT)
2
T/K
120
140
120
a/Å
7.5186(3)
27.5580(12)
19.1638(8)
91.615(1)
P21
7.5904(4)
27.7714(15)
9.6134(5)
91.753(1)
P21
26.969(4)
9.5669(13)
7.5309(10)
93.165(2)
P21/c
b/Å
c/Å
β/°
space group
Z0/Z00
2, 26
1, 13
1, 7
R1, wR2 (I > 2σI)
0.0460, 0.1092
0.0265, 0.0650
0.0582, 0.1271
is therefore reversible suggests that the two forms of 1 represent
an enantiotropic system24 in which the Z0 = 1 polymorph is more
stable at high temperature and the Z0 = 2 form is the more stable
at lower temperature. Other interesting features in the DSC
include a glass transition (Tg = 149 K) and a large crystallization
exotherm followed by melting endotherm at 180 K which we
believe is due to the crystallization and subsequent melting of
surface sulphuric acid (concentrated H2SO4 has a melting point
of 276 K, but this value is very concentration dependent25 and
can go as low as ca. 180 K in ca. 35% aqueous solution) as the
crystals are highly deliquescent. The glass transition is also likely
to be associated with the sulphuric acid solution rather than the
crystalline product.
Figure 3. An overlay of 1(LT) (blue) and 1(HT) (red) viewed down
the a axis. Half of the 1(HT) structure has been generated by symmetry.
The crystals form as very large flat plates that are highly subject
to multiple twinning and cracking parallel to the plane of the
plate and hence the neutron data proved to be generally
extremely challenging to interpret as evidenced by the nonre-
gular spot shape seen in Figure 4. The resulting data are therefore
not of the highest quality but are more than sufficient to locate
the hydrogen atoms qualitatively, although some of the atomic
parameters are not particularly precise. The accurate location of the
hydrogen atoms gives rise to a slightly altered asymmetric unit
X-ray data were collected at 120 K (1(LT)) and 140 K
(1(HT)) on either side of the phase transition. Table 1 shows
crystallographic parameters for the two data sets, and Figure 3
shows an overlay of the non-hydrogen atom positions in the two
structure determinations.
formula for 1(LT), which is now formally (H2-o-PDA2+)8(H5O2 )2-
+
(SO42ꢀ)8(HSO4ꢀ)2 4H2O, and 1(HT), which is now formally
3
(H2-o-PDA2+)4(H5O2 )(SO42ꢀ)4(HSO4ꢀ) 2H2O. Figure 5
+
3
It is clear from Figure 3 that the positions of the majority of
non-hydrogen atoms do not change greatly in the two different
forms. The most significant change is around the sulfate anion
ringed in Figure 3 which is central to the hydrogen-bonded
network. Because of the short contact between one of the oxygen
atoms in the ringed sulfate ion and neighboring sulfate oxygen
atom as well as a careful study of the SꢀO bond lengths in
the ringed sulfate ion, we believe this is a hydrogen sulfate ion.
We therefore postulate that the phase change is related to the
strength and directionality of the hydrogen bonds within the
sulfate/hydrogen sulfate/water network. This distinction cannot
easily be made using X-rays as the data quality was not sufficient
to locate all the hydrogen atoms in the lower-temperature phase.
Neutron diffraction was therefore employed to ascertain the
exact hydrogen atom positions. Neutron experiments on a large
crystal of 1 were carried out using VIVALDI at the Institut Laue-
Langvin in Grenoble, France.26,27 The crystal was slow cooled to
120 K and a full data set was collected. The crystal was then
warmed to 130 K after which Laue patterns were collected at 2 K
intervals. The patterns were closely monitored for the disap-
pearance of diffraction peaks which would indicate that the phase
transition had occurred. According to the data collected, axis
halving occurred between 140 K and 142 K, which, allowing for
some hysteresis, is in agreement with the exotherm observed in
the DSC and with the X-ray data. Figure 4 shows parts of patterns
at the same sample orientation taken above and below the phase-
change temperature showing the disappearance of peaks corre-
sponding to the axis halving. The crystal was then warmed up to
144 K to ensure that the transition was fully complete and a full
data set was collected.
shows the unique part of the water/hydrogen sulfate/sulfate
hydrogen bonded network in the neutron structures of (a) 1(LT)
and (b) 1(HT).
Table 2 shows O O contact distances for the hydrogen
3 3 3
bonds shown in Figure 5 derived from the X-ray data (which are
more precise than the neutron data for the heavy atoms). The
distances for 1(LT) are presented in two parts, 1(LT)_1 and
1(LT)_2 corresponding to the two unique hydrogen-bonded
networks present in the structure (e.g., O16 in 1(LT)_1 corre-
sponds to O36 in 1(LT)_2).
Some O O distances in 1(HT) lie between the two
3 3 3
distances observed in the corresponding 1(LT) structure, sug-
gesting that some domains of the 1(HT) structure are averages of
the unsymmetrical parts of 1(LT). This averaging is not always
the case, however, as some of the 1(HT) O O distances lie
3 3 3
outside the range observed in the two 1(LT) parts.
Several of the O O distances in the two unsymmetrical
3 3 3
portions of 1(LT) (i.e., 1(LT)_1 and 1(LT)_2) differ signifi-
cantly; for example, there is nearly 0.1 Å difference in the two
O19 O12 distances (2.649(4) and 2.562(4) Å) compared
3 3 3
with the difference of just 0.001 Å in the two O3W O1W
3 3 3
distances (2.538(4) and 2.539(4) Å) which are identical within
experimental error. The fact that only some of these contact
distances differ between the unsymmetrical portions again
suggests that only part of the 1(LT) structure is affected by the
phase change. Figure 6 shows an overlay of the unique part of the
hydrogen-bonded network for the three independent networks,
namely, 1(LT)_1 (red), 1(LT)_2 (black), and 1(HT) (blue).
It is clear from Figure 6 that the major difference in the three
motifs is the hydrogen sulfate ion based on S(10) in 1(LT)_2.
4906
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Figure 4. Neutron Laue patterns recorded at 140 and 142 K showing the disappearance of supercell reflections (circled). The small squares indicate the
positions of the allowed reflections in 1(HT).
Table 2. O O Distances (Å) in the Hydrogen-Bonded
3 3 3
Networks of 1(LT) and 1(HT) from X-ray Dataa
1(LT)_1
1(LT)_2
1(HT)
O1W O16
2.764(4)
2.723(4)
2.733(4)
2.757(4)
2.538(4)
2.428(4)
2.753(4)
2.645(5)
2.649(4)
2.761(4)
2.688(4)
2.731(4)
2.786(4)
2.539(4)
2.425(4)
2.718(4)
2.598(4)
2.562(4)
2.772(3)
2.737(3)
2.717(3)
2.744(3)
2.554(3)
2.421(3)
2.704(3)
2.604(3)
2.590(3)
3 3 3
O1W O18a
3 3 3
O2W O5b
3 3 3
O2W O20c
3 3 3
O3W O1Wd
3 3 3
O3W O4W
3 3 3
O4W O2
3 3 3
O4W O2W
3 3 3
O19 O12
3 3 3
a a = ꢀ1 + X, +Y, ꢀ1 + Z, b = 1 + X, +Y, +Z, c = 2 ꢀ X, ꢀ0.5 + Y, 1 ꢀ Z,
d = 1 ꢀ X, ꢀ0.5 + Y, ꢀZ.
Figure 6. Overlay of the hydrogen-bonded portions of the three
independent networks, 1(HT) (blue), 1(LT)_1 (red), and 1(LT)_2
(black).
Figure 5. The unique part of the water/hydrogen sulfate/sulfate
hydrogen bonded network in the neutron structures of (a) 1(HT)
and (b) 1(LT).
Table 3. Selected OꢀH Bond Lengths from Neutron Data
1(LT)_1
1(LT)_2
1(HT)
O3WꢀH3W3
O4WꢀH3W3
1.06(3)
1.44(3)
1.19(3)
1.25(3)
1.11(2)
1.32(3)
The driving force for this change appears to be an increase in
hydrogen-bond strength in the hydrogen sulfate to sulfate
2ꢀ
hydrogen bond O19 O12 (SO4
HOSO3ꢀ) correspond-
3 3 3
3 3 3
+
Concomitantly, there is a proton shift around the H2O5 ion;
in 1(LT)_1 and 1(HT) hydrogen H3W3 is localized on O3W,
ing to a change in the O O distance from 2.649(4) in
1(LT)_1 and 2.590(3) in 1(HT) to 2.562(4) in 1(LT)_2.
3 3 3
4907
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Table 4. Hydrogen-Bond Parameters from Neutron Data in
the Water Chain of 2
DꢀH
H
A
D
A
DꢀH
3 3 3
A
3 3 3
3 3 3
Major Component
O1WꢀH1W1 O2W 0.922(14) 1.936(13) 2.822(6) 160.4(11)
3 3 3
O2W0ꢀH2W10
O3W 0.961(13) 1.798(13) 2.755(5) 173.3(14)
3 3 3
O3WꢀH3W2 O2W 0.925(12) 1.923(12) 2.807(6) 159.4(12)
3 3 3
O2WꢀH2W4 O1W 0.962(10) 1.803(10) 2.763(6) 176.0(12)
3 3 3
Minor Component
O1WꢀH1W2 O2W
0.86(3)
1.02(2)
0.94(2)
1.92(3) 2.763(6)
166(2)
3 3 3
O2WꢀH2W2 O3W
1.79(2) 2.807(6) 173.9(19)
3 3 3
O3WꢀH3W3 O2W0
1.82(2) 2.755(5)
1.87(3) 2.822(6)
171(2)
172(2)
3 3 3
O2W0ꢀH2W30
O1W 0.96(3)
3 3 3
Table 5. Hydrogen Bonds from Water to Sulfate Anions in 2
DꢀH
H
A
D
A
D-H
A
3 3 3
3 3 3
3 3 3
O1WꢀH1W3 O2
0.967(7)
0.951(8)
1.896(7)
1.890(8)
2.859(5)
2.828(5)
173.2(8)
168.5(8)
3 3 3
O3WꢀH3W1 O5
3 3 3
Figure 8. Short O O contact (in green) and associated hydrogen
3 3 3
bonds in 2. #1 = ꢀX, ꢀ1/2 + Y, 1/2 ꢀ Z.
is clear not only from the hydrogen atom positions but also
within the heavy-atom framework. We can thus say that in this
enantiotropic system lowering the temperature places stricter
directional constraints on the hydrogen-bonded network as the
proton thermal motion decreases, leading to symmetry breaking
and a more stable Z0 = 2 structure that can accommodate more
linear hydrogen bonds.28 At higher temperatures, the increased
proton thermal motion places fewer constraints on the hydro-
gen-bonding network allowing a higher symmetry structure that
minimizes unfavorable nonbonded contacts. The phase transi-
tion thus represents a conflict between the directional hydrogen
Figure 7. “Flip-flop” water chain in 2 showing (a) major (prime atoms
x, 0.5 ꢀ y, 0.5 + z) and (b) minor parts of the disorder (prime atoms x,
0.5 ꢀ y, z ꢀ 0.5).
but in 1(LT)_2 it moves to a position that is essentially halfway
between O3W and O4W (Table 3). Even though crystal quality
limits the Laue data precision, the qualitative nature of the change
4908
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
those where the location of protons is unclear (and hence likely
represent hydrogen-bonded hydrogen sulfate anions).32 To
study this suggested phenomenon further, we undertook a
neutron diffraction study of 2 using the four-circle diffractometer
D19 at the ILL in Grenoble.33,34 The neutron structure determi-
nation gave good agreement with the heavy-atom positions
found from the X-ray structure with the shortest sulfate
3 3 3
sulfate O O nonbonded contact being 2.932(9) Å; however,
3 3 3
location of the hydrogen atoms showed that there is disorder in
the “flip-flop” water chain which has two orientations of hydro-
gen atoms with a 63:37 ratio. Figure 7 shows the (a) major and
(b) minor components of the water chain. The water chain is
linked to the sulfate anions via two strong OꢀH O hydrogen
3 3 3
bonds (Table 5).
The neutron structure confirms that there are no hydrogen
atoms in between the sulfate oxygen atoms involved in the short
of their van der Waals radii (rvdW: O, 1.52 Å) indicating a strong
non-covalent O O interaction”. The proposed existence of
3 3 3
O
O contacts, and hence this contact is not a hydrogen bond.
attractive O O interactions between sulfate anions is surpris-
3 3 3
3 3 3
In view of the number of strong hydrogen bonds in this structure
and the connection between the strongly hydrogen-bonded
ing. Attractive O O interactions have been observed in neutral
3 3 3
aromatic nitro derivatives30 but are virtually unknown for anionic
oxygen atoms such as those in sulfate anions. A CSD search for
water chain and the sulfate anions, we suggest that the O
O
3 3 3
short contacts observed between sulfate anions in 2 are an
enforced repulsive short contact brought about by the strong,
charge-assisted hydrogen bonding from the water and o-pheny-
lenediammonium cations in other parts of the structure and
represent an unfavorable contribution to an overall stable lattice.
short (i.e., less than the van der Waals distance) O O contacts
3 3 3
between sulfate anions shows fewer than 10 hits, the majority of
which correspond to either highly disordered structures31 or
Table 6. Hydrogen Bonding in 2a
+
In particular, each short O O contact is bridged by one NH3
3 3 3
DꢀH
H
A
D
A
DꢀH
3 3 3
A
3 3 3
3 3 3
group and the neutron structure shows that the NH
O
3 3 3
N1ꢀH1A O41
1.030(9)
1.799(9)
1.887(9)
2.801(5)
2.878(5)
163.1(10)
160.4(12)
3 3 3
N1ꢀH1B O2
1.030(11)
a #1 = ꢀX, ꢀ1/2 + Y, 1/2 ꢀ Z.
3 3 3
Scheme 1. Amines Used to Produce Analogues of 1
Figure 9. Hydrogen-bonded network in 3.
Table 7. Summary of Structures 1ꢀ9b
face-face πꢀπ
edge-face πꢀπ
hydrogen bond
saturation
amine
anion
other
Z0
density
2D/3D
3D
stacking
stacking
2ꢀ
1
o-PDA
HSO4ꢀ/SO4
H2O/H3O+
4 or 8
1.631 (HT)
1.651 (LT)
1.604
Y
N
Y
ꢀ
2
o-PDA
SO4
H2O
2
3D
3D
2D
2D
2D
2D
2D
2D
2D
N
N
Y
N
N
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
ꢀ
3
2,3-DAN
1,8-DAN
3,4-DAT
TCHD
o-TOL
HSO4ꢀ
1/2
1
1.526
4
HSO4
1.800
2ꢀ
5
SO4
1
1.587
N
Y
2ꢀ
6
SO4
1
1.508
ꢀ
7
HSO4ꢀ
H3O+
ꢀ
1
1.558
Y
Y
Y
N
Y
Y
N
Y
8
PhNH3
o-TOL
HSO4
2
1.553
ꢀ
9a
9b
H2PO4ꢀ
H3PO4
1
1.541
o-TOL
H2PO4
4
1.390
4909
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
interactions across this bridge are short and linear (Figure 8,
Table 6). The existence of repulsive short contacts in charged
identical molecular shape to o-PDA but which is anticipated to
exhibit weaker hydrogen bonding, namely, o-toluidine (o-TOL).
Starting amines were recrystallized from a mixture of concen-
trated sulfuric (1ꢀ8) or phosphoric acid (9a,b) and water.
Table 7 shows a summary of the results obtained from the
crystallizations which will be discussed in detail below. Each
structure is examined against several criteria, first what the Z0
value is, whether there is hydrogen-bond saturation (where the
total number of donor hydrogen atoms exactly matches the
number of acceptor lone pairs),39 whether the cations are
involved in π-stacking, and finally what the overall dimension-
ality of the hydrogen-bonded network is, principally whether the
packing consisted of two-dimensional (2D) layers or 3D
networks.
hydrogen-bonded systems is well precedented.35ꢀ37
ꢀ
Analogues of (H2-o-PDA2+)4(SO42‑)3(HSO4
) 4H2O. It is
2
3
clear from the structures of 1 and 2 that the intermolecular
interactions play a large part in determining the overall crystal
packing arrangement, and hence we were interested in the effect
of varying the cation and, to a lesser extent the anion, on the
structure of chemically related compounds. A series of different
amines were chosen (Scheme 1), 2,3-diaminonaphthalene (2,3-
DAN), 1,8-diaminonaphthalene (1,8-DAN), and 3,4-diaminoto-
luene (3,4-DAT) representing different degrees of potential
πꢀπ stacking contributions while trans-diaminocyclohexane
(TCHD) removes the πꢀπ stacking capability of the cation
completely. We were also interested in the shape vs hydrogen
bonding argument38 so chose a derivative that has an almost
First, we were interested to see the effect of changing the πꢀπ
stacking ability of the cation on the packing so two naphthalene
derivatives were prepared using 1,2-diaminonapthalene (2,3-
DAN) and 7,8-diaminonapthalene (1,8-DAN), Scheme 1. 2,3-
DAN was recrystallized from a mixture of concentrated sulfuric
acid and water giving tiny crystals. These were not suitable for
determination using a laboratory X-ray source, so data were
collected using synchrotron radiation on Station 9.8 at Daresbury.
The crystals were very poor quality and gave split reflections
consistent with twinning, but we were able to solve the structure
2ꢀ
which crystallizes as (H2-2,3-DAN2+)(SO4 ) (3) with Z0 = 1/2.
We are confident of correct space-group choice despite space-
group absence violations, which are likely to be a result of the
overlap of multiple diffraction patterns; however, no suitable
twin law could be identified and as a result the residuals for the
refinement are relatively high. With this in mind detailed
structural parameters are not presented here, but we are con-
fident that the data are good enough to characterize the general
environment of the molecules. Despite the extended naphthenyl
π-surface, the naphthalene derived moieties do not π-stack;
instead, they are arranged in an offset manner with the closest
centroidꢀcentroid distance around 6.6 Å. The absence of π-
stacking seems to be a result of the optimization of the spacing
required for the hydrogen bonding within the structure, as
Figure 10. Asymmetric unit of 4 showing NꢀH O hydrogen bonds.
3 3 3
Figure 11. View down the a axis in 4 showing the hydrogen-bonded sheet (a) in detail and (b) an overview. The naphthalene-derived aryl backbone has
been removed for clarity in (a). #1 = +X, 0.5 ꢀ Y, ꢀ0.5 + Z, #2 = +X, 1 + Y, +Z, #3 = +X, 1.5 ꢀ Y, 0.5 + Z.
4910
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Figure 12. View down the b axis in 4 showing two layers.
Table 8. Hydrogen Bonding in 4a
DꢀHb
H
A
D
A
DꢀH
A
3 3 3
3 3 3
3 3 3
159.4
O41-H4A1 O8
0.97
1.03
0.91
0.91
0.91
0.91
0.91
0.91
1.60
1.52
1.92
1.91
2.12
1.99
2.01
2.04
2.529(2)
2.553(2)
2.769(2)
2.800(3)
2.850(2)
2.872(3)
2.865(3)
2.936(2)
3 3 3
O6ꢀH6A O22
175.8
154.3
167.3
136.0
162.6
156.6
167.8
3 3 3
N1ꢀH1B O1
3 3 3
N1ꢀH1C O5
3 3 3
N1ꢀH1A O31
Figure 13. πꢀπ stacking observed between different layers in 4 with
distances: A = 3.33 Å, B = 3.31 Å, C = 3.69 Å.
3 3 3
N2ꢀH2A O1
3 3 3
N2ꢀH2B O5
3 3 3
the c axis, but no suitable twin law could be found. Because of the
problems with the diffraction data the two hydrogen sulfate protons
could not be located in the difference map, but a study of the
hydrogen-bonding pattern and SꢀO bond lengths indicates that the
hydrogen atoms are located on O(3) and O(6). Compounds 4and 5
are not isomorphous but are essentially isostructural as 5also exhibits
hydrogen-bonded sheets (Figure 14 and Table 9).
In this structure there is no πꢀπ stacking between layers
(Figure 15) as this is presumably disrupted by steric hindrance
resulting from the methyl group in the backbone; there are
N2ꢀH2C O83
3 3 3
a #1 = +X, 0.5 ꢀ Y, ꢀ0.5 + Z, #2 = +X, 1 + Y, +Z, #3 = +X, 1.5 ꢀ Y, 0.5 +
Z. b OH hydrogen atoms located experimentally but not refined.
packing this way ensures that the donors and acceptors are fully
saturated and that the sulfate can accept six hydrogen bonds from
the naphthalene NꢀH donors (Figure 9).
Using 1,8-DAN as a starting ꢀmaterial gives a structure with
formula (H2-1,8-DAN2+)(HSO4 )2 (4) and Z0 = 1 (Figure 10).
The ammonium hydrogen atoms and hydrogen sulfate anions are
linked together via OꢀH O and NꢀH O hydrogen bonds
however some weak CꢀH π interactions (ca. 3.2 Å) invol-
3 3 3
ving only the minor position of the methyl group, and therefore
these are not thought to be particularly significant. In light of this,
it would appear that the hydrogen bonding in compounds 4 and
5 is the dominant (i.e., “structure determining”) interaction as 5
is isostructural despite the lack of π-stacking interactions and the
presence of the methyl group.
3 3 3
3 3 3
to form a sheet with the pendant naphthalene-derived aryl groups
all pointing in the same direction above the sheet (Figure 11).
Detailed hydrogen-bonding information is given in Table 8.
The hydrogen-bond donors and acceptors in the structure are
almost fully saturated with the exception of an SdO oxygen, O7,
which does not undergo any hydrogen bonding and instead
points perpendicular to the hydrogen-bonding plane into the
space between the hydrogen-bonded sheets. The naphthalene
molecules above the hydrogen-bonded sheet intercalate via π-
stacking with naphthalene molecules in the sheet above to form a
layer structure (Figure 12).
We were therefore interested to see how similar the packing
would be if the aromatic phenylenediamine was replaced by a
saturated analogue, namely, trans-1,2-cyclohexanediamine (TCHD).
Compound 6 crystallizes with formula (H2-TCHD2+)(SO42ꢀ) and
has Z0 = 1. Figure 16 shows the asymmetric unit viewed along the
b axis. With π-stacking constraints removed, the only interaction
of note is hydrogen bonding between the ammonium groups and
the sulfate counterion. Two of the three ammonium hydrogen
atoms participate in hydrogen bonds in the plane of the cyclo-
hexane ring, which is in the chair conformation, while the
remaining hydrogen participates in a hydrogen bond perpendi-
cular to the plane linking the layers together. It is interesting to
note that as in 5 the layer structure remains (Figures 17 and 18)
despite the lack of π-stacking interactions. Hydrogen-bond
parameters are shown in Table 10.
The doubly protonated 1,8-DAN molecules exhibit π-stacking
between layers in both edge-to-face and face-to-face interactions
(Figure 13), corresponding to a sandwich herringbone motif
similar to that observed in pyrene.40ꢀ42
We were also interested in what would happen if the π-stacking in
1 was disrupted by addition of an extra group to the backbone of the
aromatic ring so chose 3,4-diaminotoluene (3,4-DAT, Scheme 1) as
a starting amine. Structural determination of the resulting crystals
shows a Z0 = 1 structure with the same overall formula as 4, namely,
(H2-3,4-DAT2+)(HSO4ꢀ)2 (5). The dication in 5 has the backbone
methyl group disordered over two sites with an occupancy ratio of
56:44. The crystals were very small and plate-like leading to weak
diffraction. There was also some evidence of twinning affecting
As hydrogen bonding is clearly a key feature of the structures
observed thus far, we decided to leave the π-stacking component
constant and instead change the hydrogen-bonding capabilities
of the molecule. As we saw in 6, shape is also important and in order
4911
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Figure 14. View down the c axis in 5 showing the hydrogen-bonded sheet (a) in detail and (b) an overview. The phenyl backbone has been removed for
clarity in (a). #1 = 1 + X, +Y, +Z, #2 = ꢀ1 + X, +Y, +Z, #3 = 3 ꢀ X, 2 ꢀ Y, ꢀZ.
previously,43 but unlike the published structure we were able to
locate and refine all of the oxonium protons at 120 K thus
confirming the nature of the counteranion. Unsurprisingly in 7
the methyl hydrogen atoms do not participate in any intermole-
Table 9. Hydrogen Bonding in 4a
DꢀH
H
A
D
A
DꢀH
A
3 3 3
3 3 3
3 3 3
172.9
N1ꢀH1A O4
0.91
0.91
0.91
0.91
0.91
0.91
1.88
1.90
2.11
2.04
2.05
2.31
2.789(5)
2.807(5)
2.891(5)
2.837(5)
2.849(5)
2.895(5)
2.509(5)
2.551(5)
3 3 3
cular interactions and the packing is dominated by NꢀH O and
3 3 3
N2ꢀH2C O5
174.4
143.8
146.0
146.4
122.0
3 3 3
OꢀH O hydrogen bonds (Figure 19). Hydrogen-bonding
N1ꢀH1B O11
3 3 3
3 3 3
parameters are given in Table 11.
N1ꢀH1C O72
3 3 3
The packing in 7 is similar to the overall nature of the pack-
ing in 5 and 6 where the molecules are arranged in layers
(Figure 20) with successive layers linked together by π-stacking
interactions.
N2ꢀH2A O11
3 3 3
N2ꢀH2A O63
3 3 3
O6 O23
3 3 3
O33 O82
This layer structure with interlayer π-stacking is also seen in
the previously reported crystal structure of the aniline-derivative
anilinium hydrogꢀen sulfate which crystallizes with formula
3 3 3
a #1 = +X, 0.5 ꢀ Y, ꢀ0.5 + Z, #2 = +X, 1 + Y, +Z, #3 = +X, 1.5 ꢀ Y, 0.5 + Z.
+
(PhNH3 )(HSO4 ) (8) and Z0 = 2;21 however, the π-stacking
to keep as many parameters as consistent as possible, we decided to
concentrate on o-toluidine (o-TOL, Scheme 1) as a starting material.
This has essentially the same molecular shape as H2-o-PDA but with
a methyl substituent replacing an ammonium group, although it is
worth noting that the molecules are not directly comparable as the
charge of the protonated versions is different.
is slightly different between layers as shown in Figure 21.
Both 7 and 8 have one unique face-to-face π-stacking inter-
action and two equivalent CꢀH π stacking interactions. The
3 3 3
faceꢀface interaction in 8 is shorter than that in 7 which may be
due to steric hindrance by the methyl group in 7. However, the
Crystallization of o-TOL with a mixture of sulfuric acid and water
methyl group in 7 provides a much stronger CꢀH π inter-
3 3 3
leads to a structure with Z0 =1,namely,(H3O+)(H-o-TOL+)(SO42ꢀ
)
action which is nearly 1.5 Å shorter than the equivalent aryl-
(7). The room temperature structure of 7 has been determined
CH π interaction in 8.
3 3 3
4912
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Figure 15. Layer structure of 5.
Figure 18. View down the c axis showing hydrogen-bonded sheet in 6.
#3 = 1/2 ꢀ X, 1/2 + Y, Z, #4 = 1/2 ꢀ X, ꢀ1/2 + Y, Z.
Figure 16. View showing the asymmetric unit of 6 (symmetrically
generated molecules in gray). #1 = ꢀ1/2 + X, 1/2 ꢀ Y, 1 ꢀ Z, #2 = ꢀ1 +
X, Y, Z.
Table 10. Hydrogen-Bonding Parameters for 9a
hydrogen bond
H
A
D
A
DꢀH
A
3 3 3
3 3 3
3 3 3
174
N1ꢀH1A O2a
1.85
1.90
1.98
1.89
1.92
1.97
2.7586(18)
2.8104(17)
2.8701(17)
2.7866(17)
2.8226(18)
2.8656(18)
3 3 3
N1ꢀH1B O4b
177
164
168
172
168
3 3 3
N1ꢀH1C O1
3 3 3
N2ꢀH2A O1c
3 3 3
N2ꢀH2B O2b
3 3 3
N2ꢀH2C O3d
3 3 3
a a = 1/2 ꢀ X, ꢀ1/2 + Y, Z, b = ꢀ1/2 + X, 1/2 ꢀ Y, 1 ꢀ Z, c = 1/2 ꢀ X,
1/2 + Y, Z, d = ꢀ1 + X, Y, Z.
what would probably be predicted as the simplest form with just
the monoprotonated cation and a hydrogen sulfate anion in the
asymmetric unit; instead the H2-o-TOL cation crystallizes with a
sulfate/oxonium mixture, and the protonated aniline has two of
each cation and anion in the asymmetric unit, possibly suggesting
that the Z0 = 1 ArNH3 /HSO4ꢀ structure is not favorable.
+
Figure 17. View down the b axis showing hydrogen-bonded layers in 6.
As the anion clearly has a large effect on the hydrogen-bonding
interactions, we next decided to look at the effect changing the
anion has on the hydrogen bonding and hence the structure.
Previously we had shown that crystallization of o-PDA from
phosphoric acid gives a bis-dihydrogen phosphate salt with
Z0 = 1.21 The packing in this structure is similar to that in 1,
It is interesting to note the difference in composition of the
formula unit between 7 and 8 despite the fact that there is only a
small structural difference (a methyl group) between the two
compounds. Neither compound seems able to crystallize with
4913
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Figure 19. Hydrogen bonding in 7. #1 = +X, 0.5 ꢀ Y, 0.5 + Z, #2 = 1 ꢀ
X, ꢀY, 2 ꢀ Z, #3 = 1 ꢀ X, 0.5 + Y, 1.5 ꢀ Z.
Table 11. Hydrogen-Bond Parameters in 7
DꢀH
H
A
D
A
DꢀH
3 3 3
A
3 3 3
3 3 3
O1WꢀH1W O11
0.85(2)
0.82(3)
0.94(2)
0.91
1.64(3)
1.77(3)
1.59(2)
1.85
2.4905(14)
2.5737(14)
2.5282(14)
2.7391(15)
2.7367(15)
2.9990(16)
178(2)
171(2)
179(2)
3 3 3
O1WꢀH2W O32
Figure 21. π-stacking in (a) 7 and (b) 8, centroid centroid and
3 3 3
3 3 3
CꢀH centroid distances are A = 3.09 Å, B = 3.61 Å, C = 3.06 Å, D =
O1WꢀH3W O4
3 3 3
3 3 3
N1ꢀH1B O13
165.6
3.79 Å.
3 3 3
N1ꢀH1A O2
0.91
1.84
169.0
167.8
3 3 3
N1ꢀH1C O31
0.91
2.10
Table 12. Hydrogen Bonds in 9aa
3 3 3
DꢀH
H
A
D
A
DꢀH
3 3 3
A
3 3 3
3 3 3
N1ꢀH1D O42
0.91
2.03
2.13
1.88
2.932(2)
2.978(3)
2.770(2)
2.602(2)
2.627(2)
2.508(2)
2.619(2)
2.534(2)
172.7
155.0
163.7
3 3 3
N1ꢀH1E O63
0.91
3 3 3
N1ꢀH1F O1
0.91
3 3 3
O2ꢀH2O O54
0.82(3)
0.80(3)
0.78(3)
0.77(3)
0.79(3)
1.79(3)
1.84(3)
1.73(3)
1.85(3)
1.74(3)
172(3)
3 3 3
O4ꢀH4O O32
167(3)
174(3)
177(3)
175(4)
3 3 3
O6ꢀH6O O32
3 3 3
O7ꢀH7O O51
3 3 3
O8ꢀH8O O11
3 3 3
a #1 = 1 ꢀ X, ꢀY, 1 ꢀ Z, #2 = 1 ꢀ X, 0.5 + Y, 1.5 ꢀ Z, #3 = 1 ꢀ X, 1 ꢀ Y,
1 ꢀ Z, #4 = +X, 0.5 ꢀ Y, 0.5 + Z.
Figure 20. Two layers in 7 linked together by π-stacking interactions,
centroid centroid distance (in black) = 3.61 Å.
3 3 3
with the o-PDA molecules arranged in stacks surrounded by a 3D
network of hydrogen-bonded phosphates.
Recrystallization of o-TOL from an aqueous solution of
phosphoric acid gives two different crystal forms depending on
the initial crystallization conditions, specifically how concen-
trated the acid solutioꢀn is. The first form crystallizes with formula
(H-o-TOL)+(H2PO4 ) H3PO4 9a with Z0 = 1, whereas if a
3
lower concentration of phosphoricꢀacid is used then crystals with
the formula (H-o-TOL2+)(H2PO4 ) 9b with Z0 = 4 are obtained.
Compound 9a crystallizes as the dihydrogen phosphate salt
cocrystallized with a neutral phosphoric acid molecule. This
structure is known in the literature but only at room temperature.
There is extensive hydrogen bonding in the structure with each
donor and acceptor fully saturated (Figure 22 and Table 12).
The layer structure observed is very similar to the layer
structure already seen for compounds of this type (Figure 23)
Figure 22. Hydrogen bonding in 9a. Phenyl and methyl carbons and
hydrogens have been removed for clarity. #1 = 1 ꢀ X, ꢀY, 1 ꢀ Z, #2 = 1
ꢀ X, 0.5 + Y, 1.5 ꢀ Z, #3 = 1 ꢀ X, 1 ꢀ Y, 1 ꢀ Z, #4 = +X, 0.5 ꢀ Y, 0.5 + Z.
4914
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
with hydrogen-bonded layers linked together by π-stacking of
the aromatic rings. There is a difference to the previously
observed structures' however, in 9a the hydrogen-bond layers
are themselves hydrogen bonded together forming a continuous
motif rather than the arrangements seen earlier in structures 4ꢀ8
where the hydrogen-bonded layers did not interact.
Between the layers the H-o-TOL+ molecules are π-stacked in
the same way as 8 albeit with a much larger centroid centroid
3 3 3
distance of 4.90 Å (cf. 3.79 Å in 8); however, the CꢀH
3 3 3
centroid distance of 3.14 Å is shorter than in 8 (3.31 Å).
Compound 9b crystallizes as the dihydrogen phosphate salt with
an asymmetric unit consisting of four molecules of H-o-TOL+ and
four dihydrogen phosphate anions; hence Z0 = 4. Two distinct
hydrogen-bonded networks are formed, each containing two unique
o-TOL+ molecules and two dihydrogen phosphates (Figure 24).
Again the molecules are arranged in layers, but in this instance there is
no π-stacking between layers, only some weak CꢀH π interac-
3 3 3
tions (ca. 3.35ꢀ3.40 Å) involving methyl hydrogen atoms.
There is extensive hydrogen bonding within the layers with all
donors and acceptors fully saturated (Figure 25); hydrogen-bond
parameters are given in Table 13.
Figure 25. Hydrogen bonding in the two unique networks of 9b. #1:
+X, ꢀ1 + Y, +Z, #2: 1 ꢀ X, 1 ꢀ Y, ꢀZ, #3: 2 ꢀ X, ꢀY, ꢀZ, #4: 2 ꢀ X, 1 ꢀ
Y, ꢀZ, #5: 1 ꢀ X, 1 ꢀ Y, 1 ꢀ Z, #6: 1 ꢀ X, ꢀY, 1 ꢀ Z, #7: 1 + X, +Y, +Z,
#8: 2 ꢀ X, 1 ꢀ Y, 1 ꢀ Z.
Figure 23. Layer structure in 9a.
Figure 24. Layers in 9b. The two distinct networks are colored red and blue and consist of two unique o-TOL+ molecules and two dihydrogen
phosphates.
4915
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
Table 13. Hydrogen-Bond Parameters for 9ba
’ CONCLUSION
Our studies have shown that charge-assisted hydrogen bond-
ing is a very powerful tool in compounds containing protonated
ammonium groups. Ensuring that the system is fully saturated,47
that is, that all possible donors and acceptors participate in
hydrogen bonds, is key in structure formation and often results in
formation of structures with Z0 > 1 or cocrystallization with
neutral molecules in order to achieve it. Bilayer-type structures
with hydrophobic and hydrophilic regions are the most common
motif, often with aromatic stacking between the hydrophobic
part of the cations.
DꢀH
H
A
D
A
DꢀH
A
3 3 3
3 3 3
3 3 3
175.0
N4ꢀH4B O91
0.91
0.91
1.83
2.03
1.80
2.737(3)
2.857(3)
2.699(3)
2.650(3)
2.577(3)
2.647(3)
2.701(3)
2.731(3)
2.861(3)
2.571(3)
2.852(3)
2.719(3)
2.702(3)
2.648(3)
2.575(3)
2.704(3)
2.857(3)
2.716(3)
2.577(3)
2.644(3)
3 3 3
N4ꢀH4A O122
150.4
171.0
3 3 3
N4ꢀH4C O13
0.91
3 3 3
O11ꢀH11P O122
0.85(4)
0.74(5)
0.73(4)
0.91
1.83(4)
1.85(5)
1.93(4)
1.80
160(3)
166(5)
166(5)
167.8
3 3 3
O10ꢀH10P O162
3 3 3
O15ꢀH15P O163
3 3 3
N3ꢀH3B O9
3 3 3
N3ꢀH3A O13
0.91
1.82
176.2
3 3 3
Other consequences of the dominance of an extensive net-
work of charge-assisted hydrogen bonding within a structure are
short repulsive contacts between oxygen atoms, as in 2. We
N3ꢀH3C O164
0.91
2.02
153.5
3 3 3
O14ꢀH14P O124
0.88(4)
0.91
1.71(4)
2.01
166(4)
153.2
3 3 3
N1ꢀH1C O15
3 3 3
suggest that the short O O are merely a result of maximizing
3 3 3
N1ꢀH1B O4
0.91
1.81
178.6
3 3 3
the hydrogen-bonding strength and hence are an unfavorable
contribution to an overall favorable energy.35ꢀ37 We have also
shown that small changes in the hydrogen bonded network can
lead to symmetry breaking as in 1 where decreased proton
movement and the consequent linearization of the hydrogen
bond result in increased directional constraints and hence
symmetry breaking and an increase in Z0 from 1 to 2. The
example of 1 shows that it is oversimplistic to regard either Z0 = 1
or Z0 > 1 forms as having any intrinsically greater stability.
Instead, the crystal structure in general is a result of a balance in
structural features in combination with nucleation and growth
phenomena. The relative stability of polymorphs of a given
substance with different Z0 values is governed by the balance
(or conflict leading to structural frustration) between competing
factors. In this case, the Z0 = 2 structure arises from the
linearization of the strong, charge-assisted hydrogen bonds at
lower temperatures. Both forms are compatible with close
packing, as attested by the higher density of the Z0 = 2 form.
N1ꢀH1A O6
0.91
1.81
165.9
3 3 3
O8ꢀH8P O56
0.70(4)
0.93(5)
0.91
1.96(4)
1.66(5)
1.81
168(5)
168(5)
166.5
3 3 3
O7ꢀH7P O16
3 3 3
N2ꢀH2A O4
3 3 3
N2ꢀH2B O56
0.91
2.05
147.7
3 3 3
N2ꢀH2C O67
0.91
1.81
171.3
3 3 3
O2ꢀH2P O55
0.81(4)
0.88(4)
1.78(4)
1.77(4)
165(4)
170(4)
3 3 3
O3ꢀH3P O18
3 3 3
a #1: +X, ꢀ1 + Y, +Z, #2: 1 ꢀ X, 1 ꢀ Y, ꢀZ, #3: 2 ꢀ X, ꢀY, ꢀZ, #4: 2 ꢀ X,
1 ꢀ Y, ꢀZ, #5: 1 ꢀ X, 1 ꢀ Y, 1 ꢀ Z, #6: 1 ꢀ X, ꢀY, 1 ꢀ Z, #7: 1 + X, +Y,
+Z, #8: 2 ꢀ X, 1 ꢀ Y, 1 ꢀ Z.
The data in Table 6 suggest that hydrogen bonding plays a
much more important role in determining the structures than π-
stacking as witnessed by the fact that 9 of the 10 structures have a
fully saturated hydrogen-bonding system, whereas only 7 of the
10 structures have some form of π-stacking. Layer structures
feature heavily with only three of the compounds not featuring
hydrogen-bonded layers. The layer structures observed may be a
result of favoring the stacking of hydrophilic and hydrophobic
regions. Four of the 10 structures (40%) have Z0 > 1, which is a
considerably larger proportion than observed for both the CSD as
a whole (8.8%) and the value observed for ionic species (6.6%).21
Two of the remaining five structures with Z0 = 1 (7 and 9a) con-
tain additional molecules in the asymmetric unit, i.e., form
cocrystals.44,45 We have shown previously that molecules that
crystallize with Z0 > 1 are good candidates for cocrystal forma-
tion,46 and this seems to be the case here also. 9b crystallizes as the
dihydrogen phosphate salt with Z0 = 4. If we consider the overall salt
to be the “parent” molecule, then 9a is the “child” cocrystal where an
extra molecule, phosphoric acid, is included, and the value of Z0
drops from 4 to 1.46 None of the three remaining structures with Z0
= 1 have both faceꢀface πꢀπ stacking interactions and a fully
saturated hydrogen-bond system.
Compound 4 is interesting in that it is the only structure that
does not have a fully saturated hydrogen-bond system, as one of
the sulfate oxygen atoms does not participate in any hydrogen
bonds nor is there any extra water included in the structure to
compensate for this. It is clear that the interplay between
hydrogen bonding and π-stacking is important in determining
not only what the structure will actually look like but also what
form is crystallized out of solution, i.e., whether it crystallizes with
a dianion or two monoanions and/or whether there is any water
in the structure. However, the extensive, charge-assisted hydro-
gen-bonding networks are the dominant feature in each case.
’ EXPERIMENTAL SECTION
Samples were prepared by recrystallizing the relevant amine from an
acid solution using ∼0.1 g of amine and a 10 mL/2 mL mix of water/
concentrated acid unless otherwise stated. Specific experimental details
can be found below:
Compound 1. Crystals suitable for both X-ray and neutron
diffraction were grown from a solution containing 0.3159 g of o-
phenylenediamine dissolved in a solution containing 3 mL of conc.
sulfuric acid and 10 mL of water. Neutron diffraction studies were
carried out on a crystal of approximate volume 2.4 mm3 at the Institut
Laue-Langevin, Grenoble, using the diffractometer very intense vertical
axis Laue diffractometer (VIVALDI) with thermal neutron radiation
(λ = 0.8ꢀ5.2 Å). The VIVALDI instrument uses a cylindrical detector
whose inner surface is lined with neutron-sensitive image plates to
collect diffraction data over a solid angle of 8 sr and is especially suitable
for small-molecule samples (primitive unit cells with cell lengths less
than 25 Å). The sample was wedged between two wads of quartz wool
inside a thin-walled silica glass tube and housed within a He-flow
cryostat. Ten Laue patterns were recorded at each temperature at
typically 20° intervals of rotation around an axis perpendicular to the
incident neutron beam; the patterns at 120 K were recorded for 120 min
each, while the patterns at other temperatures were recorded for 45 min
each. The patterns were indexed using Lauegen of the CCP4 Laue
suite.48 Peak integration over the neutron Laue diffraction patterns was
based on a 2D minimum σ(I)/(I) routine,49 and the individual reflection
intensities were normalized to a common wavelength by comparison of
repeated observations including symmetry equivalents using an in-house
program (Rearrange_lam). X-ray diffraction studies were carried out on
4916
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
4917
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
a Bruker diffractometer equipped with a SMART 6K CCD area detector
and Oxford Cryostream N2 cooling device, using graphite-monochro-
mated MoꢀKα radiation (λ = 0.71073 Å)
were carried out on a Bruker diffractometer equipped with a SMART 1K
CCD area detector and Oxford Cryostream N2 cooling device, using
graphite-monochromated MoꢀKα radiation (λ = 0.71073 Å).
Compound 9b. Crystallization was set up using a 3:10:7 (v:v:v)
ratio of o-toluidine/water/conc. phosphoric acid. Structure determinations
were carried out on a Bruker diffractometer equipped with a SMART 6K
CCD area detector and Oxford Cryostream N2 cooling device, using
graphite-monochromated MoꢀKα radiation (λ = 0.71073 Å).
Compound 2. A clear crystal, of approximate volume 14 mm3
(3.4 ꢁ 2.8 ꢁ 1.4 mm), was mounted between two wads of quartz wool
inside a thin-walled silica glass tube. This was glued onto an aluminum
base, which was then mounted on a Displex cryorefrigerator on the ILL
thermal-beam diffractometer D19 equipped with a very large horizon-
tally curved position-sensitive detector. This detector is mounted
symmetrically about the equatorial plane, with sample-to-detector
distance of 76 cm, and subtends 30 deg vertically and 120 deg
horizontally. It is based on multiwire gas counter technology. Readout
of 256 ꢁ 640 pixels per frame gives a nominal resolution of 0.12 deg
vertically and 0.19 deg horizontally. The crystal was cooled slowly (2 K/
min), while monitoring the diffraction pattern, to 30 K. The space group
and the unit cell found by X-rays were confirmed at 30 K. The chosen
neutron wavelength was 1.246(1) Å from a Ge(115) monochromator in
reflection (takeoff angle 70°). The accessible intensities, up to 2θ e
122.3°, were measured, to preset monitor counts, in a series of typically
80° w scans, in steps of 0.07° and counting times of between 11.7 and
23.4 s per step. The average number of reflections per detector frame
(i.e., at any one orientation) was on the order of 35. A range of crystal
orientations (different j and χ positions) was used to explore as much
reciprocal space as time permitted. Because of its large horizontal
opening, only one detector position was required. Between the long
scans, three strong or medium reflections were monitored about every
6 h in shorter scans and showed no significant change; the total measure-
ment time at 30 K was 4.5 days. The unit-cell dimensions were calculated
(ILL program Rafd19) at the end of the data collection, from the
centroids in 3D of 6225 strong reflections (5.31 e 2θ e 122.29°).
Bragg intensities were integrated in 3D using a version of the ILL
program Retreat,48 modified for the new detector geometry. For the
6225 strongest reflections, the mean positional errors for the centroids
were 0.02°, 0.03°, and 0.07° (in the scan, horizontal and vertical
directions, respectively). A total of 20881 Bragg reflections was obtained,
of which 5538 were independent. The Bragg intensities were corrected
for attenuation by the crystal (minimum and maximum transmission
coefficients 0.4718 and 0.7327, ILL program D19abs) and for attenua-
tion by the cylindrical inner vanadium Displex heat-shield (minimum
and maximum transmission coefficients 0.9046 and 0.9675, ILL program
D19abscan).
’ ASSOCIATED CONTENT
S
Supporting Information. Crystallographic information
b
files. This material is available free of charge via the Internet at
http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: jon.steed@durham.ac.uk.
Present Addresses
#The Bragg Institute, ANSTO, New Illawarra Road, Lucas
Heights NSW 2232, Australia.
’ ACKNOWLEDGMENT
We would like to thank the EPSRC for funding and Mr. W.
Douglas Carswell for the DSC studies. We would also like to
thank Prof. William Clegg and the National Crystallographic
Service staff for collecting data on Compound 3.
’ REFERENCES
(1) Steed, J. W. CrystEngComm 2003, 5, 169.
(2) Anderson, K. M.; Steed, J. W. CrystEngComm 2007, 9, 328.
(3) Desiraju, G. R. CrystEngComm 2007, 9, 91.
(4) Bernstein, J.; Dunitz, J. D.; Gavezzotti, A. Cryst. Growth Des.
2008, 8, 2011.
(5) Chattopadhyay, B.; Hemantha, H. P.; Narendra, N.; Sureshbabu,
V. V.; Warren, J. E.; Helliwell, M.; Mukherjee, A. K.; Mukherjee, M.
Cryst. Growth Des. 2010, 10, 2239.
(6) Bond, A. D. CrystEngComm 2008, 10, 411.
Compound 3. Diffraction data were collected at 120 K (5) on a
Bruker SMART1000 CCD diffractometer equipped with an Oxford
Cryosystems open-flow cryostat, employing synchrotron radiation (λ =
0.6900 Å).
Compound 4. Structure determination was carried out on a Bruker
APEX diffractometer equipped with an Oxford Cryostream N2 cool-
ing device, using graphite-monochromated MoꢀKα radiation (λ =
0.71073 Å).
(7) Bishop, R.; Scudder, M. L. Cryst. Growth Des. 2009, 9, 2890.
(8) Cruz-Cabeza, A. J.; Karki, S.; Fabian, L.; Friscic, T.; Day, G. M.;
Jones, W. Chem. Commun 2010, 46, 2224.
(9) Day, G. M.; Cooper, T. G.; Cruz-Cabeza, A. J.; Hejczyk, K. E.;
Ammon, H. L.; Boerrigter, S. X. M.; Tan, J. S.; Della Valle, R. G.; Venuti,
E.; Jose, J.; Gadre, S. R.; Desiraju, G. R.; Thakur, T. S.; van Eijck, B. P.;
Facelli, J. C.; Bazterra, V. E.; Ferraro, M. B.; Hofmann, D. W. M.;
Neumann, M. A.; Leusen, F. J. J.; Kendrick, J.; Price, S. L.; Misquitta,
A. J.; Karamertzanis, P. G.; Welch, G. W. A.; Scheraga, H. A.; Arnautova,
Y. A.; Schmidt, M. U.; van de Streek, J.; Wolf, A. K.; Schweizer, B. Acta
Crystallogr. Sect. B 2009, 65, 107.
(10) Karamertzanis, P. G.; Kazantsev, A. V.; Issa, N.; Welch,
G. W. A.; Adjiman, C. S.; Pantelides, C. C.; Price, S. L. J. Chem. Theo.
Comput. 2009, 5, 1432.
(11) Anderson, K. M.; Goeta, A. E.; Steed, J. W. Cryst. Growth Des.
2008, 8, 2517.
(12) Babu, N. J.; Nangia, A. CrystEngComm 2007, 9, 980.
(13) Das, D.; Banerjee, R.; Mondal, R.; Howard, J. A. K.; Boese, R.;
Desiraju, G. R. Chem Commun. 2006, 555.
(14) Chandran, S. K.; Nangia, A. CrystEngComm 2006, 8, 581.
(15) Caira, M. R.; Nassimbeni, L. R.; Su, H.; Weber, E. CrystEng-
Comm 2003, 5, 351.
Compound 5. Structure determinations were carried out on a
Bruker diffractometer equipped with a SMART 6K APEX CCD area
detector and Oxford Cryostream N2 cooling device, using graphite-
monochromated MoꢀKα radiation (λ = 0.71073 Å).
Compound 6. Structure determinations were carried out on a
Bruker diffractometer equipped with a SMART 1K CCD area detector
and Oxford Cryostream N2 cooling device, using graphite-monochro-
mated MoꢀKα radiation (λ = 0.71073 Å).
Compound 7. Crystallization was set up using a 1:10:10 (v:v:v)
ratio of o-toluidine/water/conc. sulfuric acid. Structure determinations
were carried out on a Nonius Kappa CCD diffractometer and Oxford
Cryostream N2 cooling device, using graphite-monochromated
MoꢀKα radiation (λ = 0.71073 Å).
Compound 9a. Crystallization was set up using a 3:4:8 (v:v:v) ratio
of o-toluidine/water/conc. phosphoric acid. Structure determinations
(16) Coelho, A. C.; Amarante, T. R.; Klinowski, J.; Goncalves, I. S.;
Paz, F. A. A. Acta Crystallogr. Sect. E - Struct. Rep. Online 2008, 64, O237.
4918
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919

Crystal Growth & Design
ARTICLE
(17) Wilson, C. C.; Shankland, N.; Florence, A. J. J. Chem.
Soc.-Faraday Trans. 1996, 92, 5051.
(18) Anderson, K. M.; Afarinkia, K.; Yu, H. W.; Goeta, A. E.; Steed,
J. W. Cryst. Growth Des. 2006, 6, 2109.
(19) Allen, F. H.; Kennard, O. Chem. Des. Autom. News 1993, 8, 31.
(20) Allen, F. H. Acta Crystallogr., Sect. B 2002, 58, 380.
(21) Anderson, K. M.; Goeta, A. E.; Hancock, K. S. B.; Steed, J. W.
Chem. Commun 2006, 2138.
(22) van Eijck, B. P.; Kroon, J. Acta Crystallogr., Sect. B 2000, 56, 535.
(23) In our previous communication, this was recorded as a Z0 = 8
structure based on the number of independent cations as the location of
the hydrogen atoms was ambiguous.
(24) Bernstein, J. Polymorphism in Molecular Crystals; Oxford
University Press: Oxford, U.K., 2008.
(25) Rossa, L.; Vogtle, F. Top. Curr. Chem. 1983, 113, 1.
(26) Wilkinson, C.; Cowan, J. A.; Myles, D. A. A.; Cipriani, F.;
McIntyre, G. J. Neutron News 2002, 13, 37.
(27) McIntyre, G. J.; Lemꢀee-Cailleau, M. H.; Wilkinson, C. Physica B
2006, 385ꢀ86, 1055.
(28) Calleja, M.; Mason, S. A.; Prince, P. D.; Steed, J. W.; Wilkinson,
C. New J. Chem. 2003, 27, 28.
(29) Raghavaiah, P.; Supriya, S.; Das, S. K. Chem. Commun
2006, 2762.
(30) Wozniak, K.; He, H. Y.; Klinowski, J.; Jones, W.; Grech, E.
J. Phys. Chem. 1994, 98, 13755.
(31) Ishida, T.; Ibe, S.; Inoue, M. J. Chem. Soc., Perkin Trans. 2
1984, 297.
(32) Kim, J.; Jung, I. S.; Kim, S. Y.; Lee, E.; Kang, J. K.; Sakamoto, S.;
Yamaguchi, K.; Kim, K. J. Am. Chem. Soc. 2000, 122, 540.
(33) Mason, S. A.; Forsyth, V. T.; Howard, J. A. K.; Davidson, M. G.;
Fuller, W.; Myles, D. A. In Opportunities for Neutron Scattering in the 3rd
Millennium; Dianoux, J., Ed.; Institut Laue-Langevin: Grenoble, France,
2001, p 332.
(34) Thomas, M.; Stansfield, R. F. D.; Berneron, M.; Filhol, A.;
Greenwood, G.; Jacobe, J.; Feltin, D.; Mason, S. A. In Position-Sensitive
Detection of Thermal Neutrons; Convert, P., Forsyth, J. B., Eds.; Academic
Press: London, 1983, p 344.
(35) Braga, D.; D’Oria, E.; Grepioni, F.; Mota, F.; Novoa, J. J.;
Rovira, C. Chem.—Eur. J. 2002, 8, 1173.
(36) Wood, P. A.; McKinnon, J. J.; Parsons, S.; Pidcock, E.; Spackman,
M. A. CrystEngComm 2008, 10, 368.
(37) Price, S. L.; Stone, A. J.; Lucas, J.; Rowland, R. S.; Thornley,
A. E. J. Am. Chem. Soc. 1994, 116, 4910.
(38) Angeloni, A.; Crawford, P. C.; Orpen, A. G.; Podesta, T. J.;
Shore, B. J. Chem.-Eur. J 2004, 10, 3783.
(39) Etter, M. C. Acc. Chem. Res. 1990, 23, 120.
(40) Desiraju, G. R.; Gavezzotti, A. Acta Crystallogr. Sect. B 1989,
45, 473.
(41) Desiraju, G. R.; Gavezzotti, A. J. Chem. Soc., Chem. Commun.
1989, 621.
(42) Gavezzotti, A.; Desiraju, G. R. Acta Crystallogr. Sect. B 1988,
44, 427.
(43) Gubin, A. I.; Khakimzhanova, G. D.; Nurakhmetov, N. N.;
Erkasov, R. S.; Buranbaev, M. Z. Kristallografiya 1990, 35, 1568.
(44) Vishweshwar, P.; McMahon, J. A.; Bis, J. A.; Zaworotko, M. J.
J. Pharm. Sci. 2006, 95, 499.
(45) Bond, A. D. CrystEngComm 2007, 9, 833.
(46) Anderson, K. M.; Probert, M. R.; Whiteley, C. N.; Rowland,
A. M.; Goeta, A. E.; Steed, J. W. Cryst. Growth Des. 2009, 9, 1082.
(47) Loehlin, J. H.; Franz, K. J.; Gist, L.; Moore, R. H. Acta
Crystallogr. 1998, B54, 695.
(48) Campbell, J. W; Hao, Q.; Harding, M. M.; Nguti, N. D.;
Wilkinson, C. J. Appl. Crystallogr. 1998, 31, 496.
(49) Wilkinson, C.; Khamis, H. W.; Stansfield, R. F. D.; McIntyre,
G. J. J. Appl. Crystallogr. 1988, 21, 471.
4919
dx.doi.org/10.1021/cg200751m |Cryst. Growth Des. 2011, 11, 4904–4919