Isoquinoline-based Werner clathrates with xylene isomers: Aromatic interactions vs. molecular flexibility

Article abstract of DOI:10.1039/c5dt00084j

The crystal structures of the Werner clathrates Ni(NCS)₂(isoquinoline)₄ (H) with para-xylene (px), meta-xylene (mx) and ortho-xylene (ox) have been elucidated. The kinetics of thermal decomposition of the three inclusion compounds were performed using the isothermal technique of Flynn and Wall. Selectivity of H for the xylene isomers was determined for both the liquid and vapour phase binary mixtures of the xylenes. The chosen ligand has a larger aromatic system to improve the possible π interactions between H and the selected guests. The planarity of the isoquinoline ligand causes H rigidity and its selectivity was compared to a related Werner complex containing the more flexible 4-phenylpyridine. This journal is

Full text of DOI:10.1039/c5dt00084j

Dalton
Transactions
PAPER
Isoquinoline-based Werner clathrates with xylene
isomers: aromatic interactions vs. molecular
flexibility†
Cite this: Dalton Trans., 2015, 44,
6863
Merrill M. Wicht,a,b Nikoletta B. Báthori*b and Luigi R. Nassimbenia
The crystal structures of the Werner clathrates Ni(NCS)2(isoquinoline)4 (H) with para-xylene (px), meta-
xylene (mx) and ortho-xylene (ox) have been elucidated. The kinetics of thermal decomposition of the
three inclusion compounds were performed using the isothermal technique of Flynn and Wall. Selectivity
of H for the xylene isomers was determined for both the liquid and vapour phase binary mixtures of the
xylenes. The chosen ligand has a larger aromatic system to improve the possible π interactions between H
and the selected guests. The planarity of the isoquinoline ligand causes H rigidity and its selectivity was
compared to a related Werner complex containing the more flexible 4-phenylpyridine.
Received 8th January 2015,
Accepted 4th March 2015
DOI: 10.1039/c5dt00084j
www.rsc.org/dalton
Soldatov4 revisited the host compound [Ni(NCS)2(4-methyl-
pyridine)4] and determined three crystal structures: dense (α),
Introduction
Werner clathrates are inclusion compounds of general formula
MX2L4·nG, where M is a divalent metal cation (typically Ni(II),
Co(II), Fe(II), Cu(II) and Mn(II)), X is an anionic ligand (NCS−,
microporous (β) polymorphs and γ-inclusion phases. The
adaptability of this host towards guest compounds is attribu-
ted to the fact that its β-phase volume expands by 5–14% upon
guest inclusion. Recently there has been renewed interest in
Werner clathrates5,6 and Lusi and Barbour7 have described
polymorphism associated with an order–disorder phase tran-
sition of the guest in the [Ni(NCS)2(4-phenylpyridine)4] clath-
rate dependent on the thermal range investigated. The host
preferentially discriminated in favour of one of the three
isomers of xylene in solid–vapour competition experiments8
and Batisai and co-workers9 explored the preparation of
phases by mechanochemical techniques which resulted in a
series of solid solutions. Different sorption properties were
exhibited by these solid solutions compared with the pure
Werner complexes. Recently the structures of three Werner
complexes have been analysed in terms of their packing, and
the results of crystal densities were justified by Hirshfeld
surface analysis and density functional theory calculations.10
In this work we present the structures, thermal decompo-
sition kinetics and the selectivity of the Werner host [Ni-
(NCS)2(iso-quinoline)4, H] with para-, meta- and ortho-xylene
(px, mx and ox, respectively). The ligand isoquinoline has a
fused ring structure which according to the Cambridge Struc-
tural Database11 has previously been encountered only once as
a ligand in Werner clathrates12 but its selectivity was not inves-
tigated. This ligand was chosen because its two fused rings
give it a large aromatic surface area which is important for
C–H⋯π and π⋯π host–guest interactions. The structural line
diagrams and atomic numbering scheme for the H and the
xylene guests are shown in Scheme 1.13
NCO−, CN−, NO3 or halide), L is a substituted pyridine or
−
α-arylalkylamine, and G is a guest, usually an organic aromatic
compound. They have the ability to absorb organic com-
pounds reversibly, and the first experimental studies of this
phenomenon were carried out by Schaeffer and co-workers in
1957.1 Lipkowski has studied various physico-chemical aspects
of these compounds, including their selectivity, crystal struc-
tures, and the thermodynamics of sorption and kinetics of
desorption, and these have been summarised in reviews.2,3 In
the well-studied Werner complex [Ni(NCS)2(4-methylpyridine)4]
the two anionic ligands are in trans positions to each other
and steric hindrance between the pyridine ligands results in a
‘four-blade propeller’ arrangement around the central metal
ion. Noted in these studies is the dependence of lattice expan-
sion on the volume and shape of the guest molecules. In most
cases the host : guest ratio is 1 : 1. Physical properties of these
structures are dependent on both the packing of the hosts as
well as the kind of properties of the guest. The enthalpy of
clathration is related to the cavities in the host lattice.3
aUniversity of Cape Town, Rondebosch, Cape Town, South Africa.
Fax: +27 21 6505419; Tel: +27 21 6505893
bCape Peninsula University of Technology, Zonnebloem, Cape Town, South Africa.
E-mail: bathorin@cput.ac.za; Tel: +27 21 4608354
†Electronic supplementary information (ESI) available: The non-isothermal TG
curves for H·mx and H·ox and the activation energies for H·mx and H·ox. CCDC
1041252–1041254. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/c5dt00084j
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 6863–6870 | 6863

Paper
Dalton Transactions
cyanate groups were ordered and refined uneventfully. The
atomic positions of the px guest were located unequivocally
from the difference electron density map. However, conver-
gence of the refinement could only be achieved with the guest
treated isotropically with appropriate bond length constraints.
The ordered px molecules are situated in cavities (Fig. 1b, px is
presented with the space fill model, red) and form a helical
arrangement along the c axis.
The H·mx structure displays similar features with the H·px
and crystallised in the P65 (no. 170) space group. The asym-
metric unit contains one host and one guest and the host posi-
tioned in Wyckoff position a. In a similar manner to H·px, the
host has two ordered and two disordered isoquinoline ligands
which were refined to 50% S.O.F.s (Fig. 1c). The mx guest is
ordered and positioned in cavities (Fig. 1d) and the packing is
enantiomeric with H·px.
Scheme 1 Structural line diagrams and atomic numbering scheme of
the host, Ni(NCS)2(isoquinoline)4, and the xylene isomers.
The structure of H·ox was solved in the monoclinic C2/c
(no. 15) space group. One host molecule is in the general posi-
tion (Wyckoff f) and one is located in a diad (Wyckoff e). Simi-
larly one guest is in the general position and the second guest
is at the centre of inversion (Wyckoff b), thus the asymmetric
Results and discussion
Crystal structures of H·px, H·mx and H·ox
The structure of H·px was solved in the hexagonal space group unit is composed of 1.5 host and 1.5 guest molecules (Fig. 1e).
P61 (no. 169) and the details of the data collection and refine- The host molecule in the general position has one of its four
ment are summarised in Table 1. The asymmetric unit con- isoquinoline ligands disordered and refined to S.O.F.s 77%
sists of one host and one guest molecule (Fig. 1a). The host is and 23%. The host molecule is in a special position with two
positioned in Wyckoff position a and is severely disordered. of its ligands situated on a diad. Both the ordered (Fig. 1f,
Two of the isoquinoline ligands lying trans across the Ni centre molecule A, yellow) and the disordered guest (Fig. 1f, molecule
are disordered and eventually were modelled with 50% site B, green) are located in cavities. The structure shows remark-
occupancy. The final model of the ligand had refined the able similarity with H·px and H·mx from the [100] direction.
C and N atoms anisotropically and the H atoms with the usual
The intermolecular interactions between the xylene guests
riding model. The other two isoquinoline ligands and the thio- and the host framework were analysed using the program
Table 1 Crystallographic data for H·px, H·mx and H·ox
H·px
H·mx
H·ox
Chemical formula
Host : guest ratio
Formula weight
Temperature/K
Crystal system
Space group (no.)
a/Å
b/Å
c/Å
α/°
Ni(NCS)2(C9H7N)4·C8H10
1 : 1
797.65
Ni(NCS)2(C9H7N)4·C8H10
1 : 1
797.65
Ni(NCS)2(C9H7N)4·C8H10
1 : 1
1199.50
173(2)
173(2)
173(2)
Hexagonal
P61 (no. 169)
10.7810(15)
10.7810(15)
59.780(12)
90.00
Hexagonal
P65 (no. 170)
10.8390(15)
10.8390(15)
59.167(12)
90.00
Monoclinic
C2/c (no. 15)
10.671(2)
19.132(4)
59.609(12)
90.00
β/°
90.00
120.00
6017.3(17)
1/6
1.321
MoKα
2496
90.00
120.00
6019.9(17)
1/6
1.320
MoKα
2466
91.49(3)
90.00
12166(4)
1.5/12
1.310
MoKα
2016
γ/°
V/Å3
Z′/Z
Dcalc./Mg m−3
Radiation type
F(000)
Crystal size/mm
Colour, crystal form
No. of total reflections
No. of unique reflections
Θmin–max/°
0.41 × 0.45 × 0.51
Blue, octahedral
8794
0.24 × 0.26 × 0.53
Blue, octahedral
5150
0.080 × 0.190 × 0.340
Blue, octahedral
9660
7094
4048
7315
2.04/27.12
0.0692, 0.1963, 1.017
637/8794
1.0102/−0.723
12.00/26.89
0.0461, 0.1055, 1.028
667/5150
0.229/−0.166
2.13/27.26
0.0686, 0.1639, 1.061
893/9660
0.796/−0.622
R[F2 > 2σ(F2)], wR(F2), S
No. of parameters/data
Res. peak (max/min)/e Å−3
6864 | Dalton Trans., 2015, 44, 6863–6870
This journal is © The Royal Society of Chemistry 2015

Dalton Transactions
Paper
Fig. 1 Structural comparison of H·px ((a) molecular structure, (b) packing viewed down a, px is presented with the space fill model, red), H·mx ((c)
molecular structure, (d) packing viewed down a, mx is presented with the space fill model, orange) and H·ox ((e) molecular structure, (f) packing
viewed down a, ox is presented with the space fill model, yellow and green).
Crystal Explorer14 which calculates the Hirshfeld surface15 of a The C⋯H contacts are highlighted in blue, and comprise ca.
target molecule in a crystal structure and depicts all its inter- 48% of the interactions. The upper lobe, labelled ①, rep-
actions with the neighbouring molecules. Fig. 2a shows the px resents close contacts between the px hydrogens and the aro-
guest covered by its Hirshfeld surface when surrounded by matic system of the host, while the lower lobe ② shows the
host molecules and the generated fingerprint plot,16 the 2D contacts between the hydrogens of the ligands and the aro-
representation of the 3D surface. The red areas on the surface matic region of the px. The structure also displays a number of
indicate C–H⋯π close contacts between the px and the hosts. C–H⋯π interactions between the host molecules. Fig. 2b dis-
The fingerprint plot was generated using one pair of the two plays the mx guest in its Hirshfeld surface and the corres-
disordered isoquinoline moieties. The resulting plot, gener- ponding fingerprint plot shows a similar environment for the
ated by using the alternative pair of ligands, is very similar. guest to the H·px structure with ca. 47% C⋯H contacts. In the
Fig. 2 Hirshfeld surfaces and their fingerprint plots of px (a), mx (b) and ox (c, ordered guest only); ① indicates close contacts between the guest
hydrogens and the aromatic system of the host, while ② shows the contacts between the ligand hydrogens and the aromatic region of the guest.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 6863–6870 | 6865

Paper
Dalton Transactions
Table 3 Activation energy ranges for thermal decomposition (kJ mol−1
case of H·ox, the disordered guest was not analysed. The
Hirshfeld surface of the ordered guest shows less intensive
interactions with the nearby hosts (smaller area in Fig. 2c).
This corresponds to the amount of observed C⋯H contacts
(ca. 44%) and their generally longer nature. Interestingly there
are significant (ox)C–H⋯S(host) contacts with (di + de) = 2.7 Å.
This type of hydrogen bonding has previously been identified
by Đaković et al. in Ni(II) complexes.17
)
Reaction step
H·ox
H·mx
H·px
Step 1
Step 2
Step 3
138–142
93–100
112
170–177
103–116
118–120
178–192
105–108
115–120
ranges for each of the three inclusion compounds and are
indicated in Table 3. The extent of reaction is defined as
Kinetics of thermal decomposition
α ¼ ðmt ꢀ m0Þ=ðm1 ꢀ m0Þ
The kinetics of thermal decomposition of all three inclusion
compounds were carried out by the non-isothermal technique
of Flynn and Wall.18 Desorption curves were recorded at fixed
heating rates β = 2, 4, 8, 16 and 32 K min−1 for all three clath-
rates. A set of curves for H·px is shown in Fig. 3. The decompo-
sition takes place in three distinct steps. Step 1 represents the
disintegration of the inclusion compound via loss of the guest.
This step is calculated as a 13.3% loss in mass. Steps 2 and 3
correspond to the decomposition of the host, in each step the
mass loss of two isoquinoline ligands at a percentage loss of
32.4% each. The decomposition curves of H·mx and H·ox are
similar and have been deposited in the ESI (Fig. S1 and S2†).
The calculated and experimental TG results for H·px, H·mx
and H·ox are summarised in Table 2.
with m0 = initial mass, mt = mass at time t and m∞ = final
mass. Plots of log β/β0 vs. 1/T for the H·px where β is the
heating rate are shown in Fig. 4. The activation energies for α
positions 0.2, 0.5 and 0.75 of Step 1 are illustrated in Fig. 4a.
Similarly, Step 2 at α positions 0.2, 0.5 and 0.8 and Step 3 at α
positions 0.3, 0.45 and 0.6 are demonstrated in Fig. 4b and 4c,
respectively. Activation energies were calculated from
the slopes:
Ea
slope ¼ ꢀ0:457
R
The activation energies of the three thermal decomposition
steps for each compound are shown in Table 3. Graphical
information for H·mx and H·ox has been deposited in the ESI
(Fig. S3 and S4†).
For Step 1 the mass loss corresponding to the loss of
1 mole of guest and activation energy were calculated over α
Step 1 decomposition reaction required the largest acti-
vation energy. The activation energy for Step 2 is the lowest of
the three steps and is 105–108 for H·px, 103–116 for H·mx and
93–100 kJ mol−1 for H·ox. The higher values of the activation
energies associated with the loss of the volatile xylene guests
can be justified in terms of the topologies of the structures of
the clathrates. In each case the xylene guest is trapped in a
cavity, thus requiring a severe disruption of the host frame-
work in order to release the guest. The effect of topology on
the thermal stability and kinetics of decomposition of
Fig. 3 Non-isothermal TG curves for H·px.
Table 2 Thermal analysis results for H·px, H·mx and H·ox
H·px exp (calc.)
%
H·mx exp (calc.)
%
H·ox exp (calc.)
%
Compound
Mass loss
Step 1
Step 2
13.1 (13.3)
32.3 (32.4)
32.1 (32.4)
13.1 (13.3)
32.1 (32.4)
31.8 (32.4)
13.2 (13.3)
31.8 (32.4)
32.1 (32.4)
Fig. 4 Activation energies for (a) Step 1, (b) Step 2 and (c) Step 3 for
H·px.
Step 3
6866 | Dalton Trans., 2015, 44, 6863–6870
This journal is © The Royal Society of Chemistry 2015

Dalton Transactions
Paper
inclusion compounds has been reviewed19 and, in general, procedures (crystallisation and solid–vapour sorption) is
structures that may be described as intercalates are less stable shown graphically and no preference for either of the two
than tubulates, which in turn are less stable than cryptates. xylenes is observed. That is, in all the three cases poor selecti-
The latter often have higher activation energies of desorption.
vity is found for one of the isomers over the other two.
Selectivity experiments
Experiments to determine the selectivity profile of this host
were carried out by two methods. Firstly by dissolving the host
in liquid mixtures of the guests with known proportions and
harvesting the crystals of the ensuing inclusion compounds
for analysis; secondly by exposing the powdered apohost to
mixtures of the guest vapours (solid–vapour sorption) and sub-
sequent analysis by headspace gas chromatography. These
competition experiments between pairs of xylene isomers (ox/
px, ox/mx and mx/px) were performed at different ratios (0 : 1;
0.2 : 0.8; 0.4 : 0.6; 0.6 : 0.4; 0.8 : 0.2 and 1 : 0). The crystals were
harvested, dried and lightly crushed for analysis.
Discussion
The question to be addressed is why the isoquinoline host
under discussion is not selective towards any of the xylene
isomers either from liquid or vapour mixtures, while the host
bis(isothiocyanato)tetrakis(4-phenylpyridine)nickel(II) has been
shown to discriminate efficiently between ortho-, meta- and
para-xylenes.8 Competition experiments between equimolar
pairs of two isomers and an equimolar mixture of all three
xylenes in the vapour state show the preference for enclathra-
tion to be in the order ox > mx > px. We have therefore ana-
lysed the non-bonded interactions which occur between a
given guest and this host. We have chosen to compare the
structures containing the ortho- and meta-xylene guests,
because they are similar in that they both have a host : guest
ratio of 1 : 1, while the structure of para-xylene has a host :
guest ratio of 1 : 3. The program Crystal Explorer was employed
to calculate the Hirshfeld surfaces of the ox and mx molecules
and their fingerprint plots were generated (Fig. 6).
Fig. 6a shows the fingerprint plot of the ox guest. The two
peaks labelled ① are associated with H⋯H interactions at
(di + de) ≈ 2.48 Å (57% of the close contacts) while those
labelled ② are due to C⋯H contacts with (di + de) ≈ 2.87 Å
(37%). In contrast, the mx structure, shown in Fig. 6b, has
H⋯H contacts labelled ① at 2.43 Å (68%) and C⋯H contacts
labelled ② at 2.85 Å (28%). The Hirshfeld surface analysis sup-
ports the experimental results and explains why the host
favours the ox guest. Also it demonstrates that the H⋯H and
C⋯H interactions occur at similar distances, slightly over the
sum of the van der Waals radii of 2.40 and 2.90 Å20 respect-
ively, but there is a greater percentage of C⋯H interactions in
the ox structure and the circled area in Fig. 6b shows that
there are many long H⋯H contacts, resulting in a less efficient
packing of the mx structure.
When a host compound is crystallised from a mixture of
two guests A and B to form a crystal containing a ratio of the
two guests (H·n′A·m′B), the selectivity coefficient of this com-
petition experiment can be determined from the formula
KA:B ¼ ðKB:AÞꢀ1 ¼ ZA=ZB ꢁ XB=XA ðXA þ XB ¼ 1Þ
where XA and XB are the mole fractions of the guests in the
mother liquor and ZA and ZB are their mole fractions in the
crystal. The results of the competition experiments are shown
in Fig. 5 in which the mole fraction of a given guest in solution
(Xguest) is plotted against its mole fraction in the solid state
(Zguest). In Fig. 5a a small amount of preference is shown for
ox over px for the composition mixture of ox/px. Fig. 5b indi-
cates the selectivity of ox compared with mx in the crystal
structure with the host showing no preference for one over the
other. Similarly in Fig. 5c there is no preference for the host to
enclathrate one of mx or ox over the other. Selectivity for both
In addition to the host–guest non-bonded interactions,
however, the host with 4-phenylpyridine ligands possesses
Fig. 6 Fingerprint plots for (a) the ox guest and (b) the mx guest in their
inclusion compounds with bis(isothiocyanato)tetrakis(4-phenylpyridine)
nickel(II).
Fig. 5 Selectivity curves for (a) ox/px; (b) ox/mx and (c) mx/px.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 6863–6870 | 6867

Paper
Dalton Transactions
considerable torsional flexibility in its pyridine and phenyl
rings. This fact has been discussed in detail previously.21 The
labelling of the four torsion angles (τ1, τ3, τ5 and τ7) between
the pyridine ring and the thiocyanato ligand (N–Ni–N–C) was
carried out cyclically in a manner to yield the minimum in the
sum of the square of their differences. The eight torsion
angles of each of the three structures have been tabulated and
labelled as before21 (Fig. 7) and their values are summarized
in Table 4. It was noted that the differences in the torsion
angles τ3 and τ4 as well as τ7 and τ8 observed in the mx struc-
ture indicated large rotations of the phenyl ring (Table 4 for
mx, Δ values in bold) compared with τ1/τ2 and τ5/τ6 differences.
The selectivity for mx is less than ox but greater than px. In the
case of ox, alternating positive and negative torsion rotations
are observed (τ1/τ2 and τ5/τ6 negative; τ3/τ4 and τ7/τ8 positive).
This attests the fact that this is a highly flexible ligand, which
easily adapts to the requirement of a given guest, rendering it
selective.
Table 5 Torsion angles τ1, τ3, τ5 and τ7 for H and some typical nickel
Werner clathrates where the ligands are pyridine (W1), 4-methyl- (W2),
4-vinyl- (W3), and 4-phenyl- (W4) pyridine
Compound
τ1 (°)
τ3 (°)
τ5 (°)
τ7 (°)
H·px
H·mx
H·ox
W1
W2
W3
16.5
24.5
24.2
15.7
30.4
33.3
29.2
49.9
25.1
50.8
21.6
39.8
36.8
36.0
17.9
48.3
26.4
51.6
35.7
44.5
43.9
50.3
50.3
59.3
34.5
40.6
38.5
48.0
W4
The torsion angles describing the pyridine moieties in the
currently investigated host H (τ1, τ3, τ5 and τ7) are remarkably
similar to previously analysed compounds. They correspond to
the ‘four-blade propeller’ configuration found in the structures
of analogous compounds containing pyridine (W1), 4-methyl-
(W2), 4-vinyl- (W3), and 4-phenyl- (W4) pyridines. This corre-
sponds to the ++++ configuration described by Lipkowski.2,22
Table 5 and Fig. 8 show the range of torsion angles for these
Fig. 8 Plot of torsion angles of Werner complexes based on values pre-
sented in Table 5.
apohosts, as well as the torsion angles for the currently dis-
cussed three structures H·px, H·mx and H·ox. Despite present-
ing four positively rotated ligands in the isoquinoline host H,
the torsion angles also indicate two large and two smaller
angles. This shows compensation for the interaction of the
host with the guest, whereas in the apohost structures no
guest is present.
Fig. 7 Torsion angles in (a) Ni(NCS)2(4-phenylpyridine)4 and (b) H.
Experimental section
Preparation of Werner clathrate
Table 4 Torsion angles of Ni(NCS)2(4-phenylpyridine)4 in its crystals of
xylene isomers
The host compound, H, bis(isothiocyanato)tetrakis(isoquino-
line)nickel(II), was prepared by adding stoichiometric quan-
tities of an ethanolic solution of isoquinoline (20 ml, 0.01 M)
to an ethanolic solution of nickel isothiocyanate (5 ml, 0.01
M). Violet crystals of Ni(NCS)2(C9H7N)4 were formed immedi-
ately and were filtered and allowed to air dry overnight.
Enclathration of any given xylene guest was carried out by
dissolving the host in benzene, adding the xylene isomer (px,
mx and ox) dropwise and stirring at 60 °C for 30 minutes,
cooling and filtering. Crystallisation occurred within 24 hours.
Deep blue crystals of H·px, H·mx and H·ox were formed.
Guest
Ox
Δ
τ1/τ2
32.4
45.1
–12.7
τ3/τ4
τ5/τ6
τ7/τ8
39.7
25.7
14.0
34.1
44.4
–10.3
43.9
31.3
12.8
mx
Δ
20.3
21.8
−1.5
20.8
–29.7
50.5
46.6
15.5
31.1
34.4
–30.1
64.5
px
Δ
23.0
34.7
−11.7
34.6
19.2
15.4
40.2
40.0
0.2
39.7
27.1
12.6
6868 | Dalton Trans., 2015, 44, 6863–6870
This journal is © The Royal Society of Chemistry 2015

Dalton Transactions
Paper
Single crystal X-ray analysis
Conclusions
Intensity data of a selected single crystal for compounds H·px,
H·mx and H·ox were collected on a Bruker DUO APEX II diffr-
actometer23 with graphite monochromated Mo Kα1 radiation
(λ = 0.71073 Å) at 173 K using Oxford Cryostream 700. Data
reduction and cell refinement were performed using Saint
Plus.24 The space group was determined from systematic
absences by using XPREP.25 The structure was solved using
SHELXS-9726 and refined using full matrix least squares
methods in SHELXL-9728 with the aid of the program X-Seed.27
The hydrogen atoms bound to carbon atoms were placed at
idealized positions and refined as riding atoms. Diagrams and
publication material were generated using PLATON,28 X-Seed
and Mercury (3.1).29 Crystal data and structure refinement
parameters are given in Table 1. CCDC 1041252–1041254
contain the supplementary crystallographic data for structures
H·px, H·mx and H·ox.
A Werner clathrate [Ni(NCS)2(isoquinoline)4] has been syn-
thesised and its properties were elucidated. Due to the impor-
tance of the xylene isomers in the petroleum industry, para-,
meta- and ortho-xylenes were considered as the guests in this
study. Single crystal structures of the host·guest compounds
were analysed and the packing was established. The analysis
of the fingerprint plots revealed the importance of the C–H⋯π
interactions and unique C–H⋯S interactions were observed in
the H·ox structure.
The non-isothermal technique of Flynn and Wall was used
to record the kinetics of thermal decomposition of the three
inclusion compounds. Similar plots for the three compounds
show initial loss of the guest, followed by two steps, each
denoting the mass loss of two isoquinoline ligands.
The selectivity of H towards the xylene isomers was deter-
mined using two methods, viz. solid–vapour sorption and crys-
tallisation from a liquid solution of the host and a bimolecular
mixture of the guests. The results were analysed using head-
space gas chromatography. No significant preference for one
of the xylene isomers over the other two was found, showing
poor selectivity of this host.
The poor selectivity of this host H was compared to that of
the related host [Ni(NCS)2(4-phenylpyridine)4]. The success of
the latter host was attributed to the torsional flexibility of the
phenyl moieties in the ligands. In contrast, the isoquinoline
ligands, although containing larger aromatic systems, have no
such flexibility and their relative conformation is largely con-
trolled by their ortho-hydrogens.
Powder X-ray diffraction
Powder X-ray diffraction experiments were carried out on a
Bruker D8 diffractometer using Cu Kα radiation. The sample
was ground to a fine powder and loaded into an aluminium
tray. Where available these spectra were compared with those
determined from the single crystal structures.
Thermogravimetric analysis
Thermal analyses were performed on a TA Q500 instrument
from 30 to 350 °C at heating rates ranging from 2, 4, 8, 16 and
32 °C min−1 with dry nitrogen as purge gas flowing at 60 ml
min−1. All samples were dried on filter paper and placed in an
open pan for thermogravimetric analysis. Sample masses
varied from 2 to 5 mg.
Acknowledgements
Authors thank Mr David Kok for assistance with the GC analy-
sis, Dr Eustina Batisai, the National Research Foundation (Pre-
toria, South Africa) and the Cape Peninsula University of
Technology for financial support. The authors declare that
they have no conflict of interest.
Competition experiments
The selectivity of the host for a particular isomer was evaluated
using two different procedures and analysing both by head-
space gas chromatography. The first was a solid–vapour exper-
iment in which the crushed host was exposed to a mixture of
two xylene guests in an evacuated chamber at room tempera-
ture for 18 hours. The compound was removed from the
chamber, dried and placed in a headspace vial for GC analysis.
The second method involved crystal formation of the host with
a guest mixture using the same procedure mentioned above.
Crystals were harvested, dried and placed in headspace vials
for GC analysis.
Notes and references
1 W. D. Schaeffer, W. S. Dorsey, D. A. Skinner and
C. G. Christian, J. Am. Chem. Soc., 1957, 79, 5870.
2 J. Lipkowski, in Inclusion Compounds, Academic Press,
New York, 1984, vol 1, ch. 3.
3 J. Lipkowski, in Comprehensive Supramolecular Chemistry,
ed. D. D. MacNicol, F. Toda and R. Bishop, Elsevier
Science, Oxford, 1996, vol. 6, ch. 20.
4 D. V. Soldatov, G. D. Enright and J. A. Ripmeester, Cryst.
Growth Des., 2004, 4, 1185.
5 C. J. Adams, M. F. Haddow, D. J. Harding, T. J. Podesta and
R. E. Waddington, CrystEngComm, 2011, 13, 4909.
6 S. Wohlert, I. Jess, U. Englert and C. Nather, CrystEng-
Comm, 2013, 5326.
Gas chromatography
GC analysis was performed on an Agilent 7890A instrument
with a Varian CP Wax capillary column (30 m × 250 μm ×
0.25 μm), nitrogen carrier gas and FID detector with inlet and
detector temperatures of 280 °C. Vials were incubated at 50 °C
for 10 minutes before injection; oven temperature at 35 °C for
1 minute, followed by a gradient at 10 °C min−1 to 120 °C for
2 minutes.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 6863–6870 | 6869

Paper
Dalton Transactions
7 M. Lusi and L. J. Barbour, CrystEngComm, 2014, 16, 36.
18 J. H. Flynn and L. A. Wall, Polym. Lett., 1966, 4, 323.
8 M. Lusi and L. J. Barbour, Angew. Chem., Int. Ed., 2012, 51, 19 L. R. Nassimbeni, Acc. Chem. Res., 2003, 36, 631.
3928. 20 A. Bondi, J. Phys. Chem., 1964, 68, 441.
9 E. Batisai, M. Lusi, T. Jacobs and L. J. Barbour, Chem. 21 L. R. Nassimbeni, M. L. Niven and M. W. Taylor, Inorg.
Commun., 2012, 12171. Chim. Acta, 1987, 132, 67.
10 S. M. Soliman, Z. B. Elzawy, M. A. M. Abu-Youssef, 22 J. Lipkowski, J. Mol. Struct., 1981, 75, 13.
J. Albering, K. Gatterer, L. Öhrström and S. F. A. Kettle, 23 Bruker 2005. APEX2. Version 1.0-27, Bruker AXS Inc.,
Acta Crystallogr., Sect. B: Struct. Sci., 2014, 70, 115.
11 F. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 380.
12 P. Baran, M. Boča, R. Boča, A. Krutošíková, J. Miklovič,
J. Pelikán and J. Tituš, Polyhedron, 2005, 24, 1510.
13 MarvinSketch 14.8.11.0, 2014, ChemAxon (http://www.
chemaxon.com).
14 S. K. Wolff, D. J. Grimwood, J. J. McKinnon, M. J. Turner,
D. Jayatilaka and M. A. Spackman, CrystalExplorer (Version
3.1), University of Western Australia, 2012.
Madison, Wisconsin, USA.
24 Bruker 2004. SAINT-Plus (including XPREP). Version 7.12,
Bruker AXS Inc., Madison, Wisconsin, USA.
25 Bruker 2003, XPREP2. Version 6.14, Bruker AXS Inc.,
Madison, Wisconsin, USA.
26 G. M. Sheldrick, SHELXS-97 and SHELXL-97 Programs for
crystal structure determination and refinement, University of
Göttingen, 1997.
27 L. J. Barbour, J. Supramol. Chem., 2001, 1, 189.
15 M. A. Spackman and D. Jayatilaka, CrystEngComm, 2009, 28 A. L. Spek, PLATON,
A Multipurpose Crystallographic
11, 19.
Tool, Utrecht University, Utrecht, The Netherlands,
16 (a) M. A. Spackman and J. J. McKinnon, CrystEngComm,
2008.
2002, 4, 378; (b) J. J. McKinnon, D. Jayatilaka and 29 C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington,
M. A. Spackman, Chem. Commun., 2007, 3814.
17 M. Đaković, D. Vila-Viçosa, M. J. Calhorda and Z. Popović,
CrystEngComm, 2011, 13, 5863.
P. McCabe, E. Pidcock, L. Rodriguez-Monge, R. Taylor,
J. van de Streek and P. A. Wood, J. Appl. Crystallogr., 2008,
41, 466.
6870 | Dalton Trans., 2015, 44, 6863–6870
This journal is © The Royal Society of Chemistry 2015

Copyright of Dalton Transactions: An International Journal of Inorganic Chemistry is the
property of Royal Society of Chemistry and its content may not be copied or emailed to
multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.