S-position isomers of BEDT-TTF and EDT-TTF: Synthesis and influence of outer sulfur atoms on the electrochemical properties and crystallographic network of related organic metals

Article abstract of DOI:10.1002/1521-3765(20011203)7:23<5168::AID-CHEM5168>3.0.CO;2-S

The synthesis and characterization of new modified tetrathiafulvalenes (TTF), the S-position isomers of BEDT-TTF and EDT-TTF, are described. The synthetic strategy presented in this work is based on an efficient and unprecedented two-step sequence for the conversion of a vicinal bis(hydroxymethyl) functionality into a disulfide ring. Different routes are discussed in terms of efficiency for the synthesis of the symmetric S-position isomer of BEDT-TTF and that of EDT-TTF. Their electrochemical properties are combined with data obtained from UV/Vis spectroscopy and orbital calculations, and the electronic influence of peripheral sulfur atoms on the neutral and oxidized species is discussed. The introduction of these outer sulfur atoms at the periphery of the TTF core gives rise to specific intermolecular S ··· S interactions in the corresponding organic materials. Crystallographic studies of radical cation salts synthesized upon electrocrystallization clearly showed that the network obtained is dictated by the outer sulfur atoms, which are responsible for a characteristic and unprecedented windmill array.

Full text of DOI:10.1002/1521-3765(20011203)7:23<5168::AID-CHEM5168>3.0.CO;2-S

FULL PAPER
Solid-State Transformations of Zinc 1,4-Benzenedicarboxylates Mediated
by Hydrogen-Bond-Forming Molecules
Mark Edgar, Robert Mitchell, Alexandra M. Z. Slawin,
Philip Lightfoot, and Paul A. Wright*[a]
Abstract: The zinc 1,4-benzenedicar-
upon the loss or addition of water.
data (compound 6, monoclinic: C2/c,
boxylates [Zn3(bdc)3(H2O)3] ¥ 4D MF Removal of all solvent molecules in-
a ˆ 17.979(1),
b ˆ 6.352(1),
c ˆ
(1;
bdc ˆ 1,4-benzenedicarboxylate),
cluded during crystallisation gives poor-
ly crystalline [Zn(bdc)] (4), which is
readily converted to highly crystalline
solids upon contact with hydrogen-
bond-forming molecules such as water,
DMF and small alcohols. The crystal
structures of the mono- and dihydrates
[Zn(bdc)(H2O)] (6) and [Zn(bdc)-
(H2O)2] (7) have been determined ab
initio from powder X-ray diffraction
7.257(1) ä, b ˆ 91.477(1)8; compound 7,
monoclinic: C2/c, a ˆ 14.992(1), b ˆ
[Zn(bdc)(H2O)] ¥ DMF
(2),
and
[Zn(bdc)] ¥ DMF (3) crystallise at room
temperature from mixtures of toluene/
dimethylformamide (DMF) under con-
centrated, dilute and dry condi-
tions, respectively. The structure of
5.0303(2),
c ˆ 12.098(1) ä,
b ˆ
103.82(1)8). The methanol adduct
[Zn3(bdc)3] ¥ 6CH3OH (5) is the same
as that prepared previously by direct
crystallisation. Comparison of these ad-
duct structures with those prepared
directly reveal that they are formed by
in situ recrystallisations. Subsequent re-
moval of included molecules gives amor-
phous [Zn(bdc)], which can be recrys-
tallised again when placed in contact
with hydrogen-bond-forming molecules.
phase
1
(monoclinic: P21/c, a ˆ
13.065(1), b ˆ 9.661(1), c ˆ 18.456(1) ä,
b ˆ 106.868(2)8) consists of layers con-
taining stacks of three zinc cations
linked by mono- and bidentate bdc
groups. Structure 1 converts to the
known phase 2 by an irreversible, re-
constructive phase transformation,
whereas 2 and 3 interconvert reversibly
Keywords: carboxylate ligands ¥ in-
clusion compounds ¥ organic inor-
ganic hybrid composites ¥ structure
elucidation ¥ zinc
metal centres are linked by doubly or triply coordinating
bridging ligands (such as diamines[3] or di-[4] or tricarboxy-
lates[5]) the thermal stability is lower, and this aspect of their
properties requires particular attention.
Zinc carboxylates have been shown to crystallise with a
range of two- and three-dimensionally connected lattices.[4, 5]
Zinc benzenetricarboxylates[6] have been shown to form well
defined and structurally stable three-dimensional frame-
works. It has been shown that adsorption of H2O and NH3
was possible for the desolvated [Zn(bdc)] but significant
structural rearrangement took place for the cobalt and nickel
forms,[5c] indicating an inclusion process rather than adsorp-
tion.
Zinc benzenedicarboxylates[7] crystallise at room temper-
ature when bases such as amines are added to solutions that
contain zinc salts and benzenedicarboxylic acids. The solvent
mixes that have been employed include molecules that form
strong hydrogen bonds such as dimethylformamide (DMF),
methanol and thiourea.[8] For thiourea adducts the resulting
metal carboxylates may occur as clusters, chains or sheets,
depending on the isomer of the dicarboxylic acid used. Zinc
benzenedicarboxylates crystallise in the presence of DMF and
Introduction
There is currently great interest in the synthesis of porous
organic inorganic hybrid frameworks, as the chemistry of
their internal surfaces differs markedly from that of inorganic
solids such as microporous aluminosilicates or aluminophos-
phates.[1] In principle, it is possible to imagine a wide range of
different organic groups being used to make up the pore walls.
These solids would be expected to display distinctive adsorp-
tion and separation behaviour. Organic inorganic hybrid
solids such as aluminium and other metal phosphonates, in
which the organic moieties (methyl and aryl groups) line the
internal surfaces, possess encouraging thermal stabilities in
the absence of oxygen, and retain their crystallinity and
porous structure upon loss of molecules included upon
crystallisation.[1] For metal organic frameworks,[2] in which
[a] Dr. P. A. Wright, Dr. M. Edgar, R. Mitchell,
Dr. A. M. Z. Slawin, Dr. P. Lightfoot
School of Chemistry, The University of St Andrews
St Andrews, Fife, KY16 9ST (UK)
Fax : (‡44)1334-463808
E-mail: paw2@st-and.ac.uk
5168
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5168 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 23

5168 5175
methanol to form the two-dimensional lattice struc-
tures [Zn(bdc)(H2O)] ¥ DMF[9] and [Zn3(bdc)3] ¥ 6CH3OH,[10]
respectively. The former is reported to give a solid that
adsorbs small molecules from the vapour phase, whereas the
latter is found to show reversible desorption and adsorption of
some of the included methanol molecules.
As part of a research program to investigate the sorption
behaviour of porous organic inorganic hybrids, we have
examined the structural changes associated with the loss and
subsequent uptake of small volatile molecules capable of
forming strong hydrogen bonds in zinc 1,4-benzenedicarbox-
ylate solids. The latter are found to be reactive solids, which
upon uptake of hydrogen-bond-forming molecules undergo
structural changes typical of inclusion compounds rather than
exhibiting the adsorption behaviour within fixed frameworks
that is observed for conventional molecular sieves.
Results and Discussion
Direct crystallisation of zinc 1,4-benzenedicarboxylate hybrid
materials: The procedure of Yaghi[9] was followed in an
attempt to prepare [Zn(bdc)(H2O)] ¥ DMF (2) for subsequent
study. Direct crystallisation from toluene/DMF solution
was found to give [Zn3(bdc)3(H2O)3] ¥ 4D MF (1),
[Zn(bdc)(H2O)] ¥ DMF (2) or [Zn(bdc)] ¥ DMF (3) depending
on the reaction conditions. To determine the conditions under
which each solid might be formed directly, a flow experiment
was devised in which a nitrogen-gas stream, saturated in
triethylamine vapour, was permitted to bubble through or
over the surface of the Zn/bdc solution. The crystals that
formed after a few hours were determined to be compound 1
(Figures 1 and 2, top). A second crop of crystals was taken
from the filtrate after 14 days and single-crystal and powder
X-ray diffraction analyses indicated that they were compound
2 (Figures 3 and 2, bottom). Compound 3 was prepared from
mixtures of dry solvents. The behaviour of compounds 1 and 2
was examined by thermogravimetric analysis (TGA), X-ray
powder diffraction and solid-state NMR spectroscopy.
Compound 1 is a two-dimensional layered structure, with
three zinc atoms forming a ™stack∫ that is connected through a
framework of bidentate and monodentate benzenedicarbox-
ylates. The three-zinc-atom ™stack∫ contains an octahedral
zinc sandwiched between two tetrahedral zinc atoms (Znoct
Zntet ˆ 3.252(2) ä), with the tetrahedrally coordinated zinc
atoms binding to a terminal water ligand. Crystallographically
distinct DMF molecules separate the [Zn(bdc)] layers and
these molecules display Owater ODMF distances of 2.614(7) ä
that are typical of H-bonds.
Figure 1. Structure of 1 from single-crystal data. Top: Viewed along the a
axis through the lamellae, showing that DMF molecules appears to fill
voids created by the framework. Bottom: Viewed along the b axis along the
lamella, showing DMF molecules forming
between the [Zn(bdc)] framework lamellae.
a hydrogen-bonded layer
Solid-state transformations on loss of included solvents:
Compound 1 was found to transform to compound 2 with
loss of one DMF molecule per unit formula, upon standing in
air over a period of days to weeks depending on the crystal
quality and the storage conditions. The transformation can be
readily followed in the bulk of the sample by X-ray powder
diffraction (Figure 4, top) and in single crystals, by using
crossed polarisation under the optical microscope. A reaction
front can be seen to move rapidly through a crystal of 1 with
attendant loss of birefringence, indicating the transformation
is not topotactic (Figure 5). The rate of interconversion was
not accurately determined, but was observed to be related to
crystal quality, with poorly formed crystals converting within a
few hours and larger single crystals remaining stable for many
days. Once initiated, the conversion process occurs within
minutes. There is no evidence to support a drying-out process,
which would be evident by a change in the outer surface of
each crystallite. Comparison of the crystal structures (Fig-
Chem. Eur. J. 2001, 7, No. 23
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5169 $ 17.50+.50/0
5169

P. A. Wright et al.
FULL PAPER
Figure 2. Top: Calculated (upper trace) and experimental X-ray powder
diffraction pattern of 1. Bottom: Calculated (upper trace) and experimen-
tal X-ray powder diffraction pattern of 2.
Figure 3. Structure of 2 from single-crystal data. Top: Viewed along the a
axis through the lamellae, showing that DMF molecules appears to fill
voids created by the framework. Bottom: Viewed along the b axis along the
lamellae, showing hydrogen-bonded DMF molecules forming an integral
part of the framework.
ures 1 and 3) confirms that the transformation is not topo-
tactic. The layered structures are quite different, for example,
in the precursor structure each ™layer∫ contains units of three
zinc polyhedra, whereas upon loss of DMF, the layers in
compound 2 possess units of two pentahedrally coordinated
zinc atoms (Znpent Znpent ˆ 2.94 ä). The layers are interdigi-
tated with terminal water ligands from one layer protruding
into voids of the neighbouring layer; each DMF molecule is
arranged such that oxygen oxygen distances, typical of
hydrogen bonds, are present (Owater ODMF ˆ 2.60 ä, O-H-
O ˆ 172.48).
solvents are thoroughly dried prior to use (Figure 4c). Dry
solvents readily yield crystalline 3 (triclinic: a ˆ 9.055(12), b ˆ
8.921(16), c ˆ 7.940(13) ä, a ˆ 99.90(8)8, b ˆ 100.91(6)8, g ˆ
103.28(8)8) as determined by powder X-ray diffraction. These
crystals undergo a solid-state recrystallisation and convert to
compound 2 on exposure to air for several weeks as
determined by X-ray powder diffraction.
Thermogravimetric analysis of the 2 indicates that solvent
molecules are lost in two discrete events: between 658C
À1158C, 5.39 wt%, and 120 2208C, 21.48 wt%. The X-ray
diffraction patterns of samples, heated to 908C and 1908C to
isolate the resulting solids, show that crystallinity is retained,
but reduced (Figure 4, middle). 13C {1H}CP-MAS NMR
spectroscopy (Figure 6) indicated that DMF is retained in
the intermediate solid (908C), but is absent from the sample
heated to 1908C, and analysis of the mass losses is consistent
with water and DMF being lost in separate events.
Recrystallisation of the solvent-free solid upon exposure to
hydrogen-bond-forming molecules–an examination of the
reversibility of these transformations: It was not possible to
determine the structure of the sample produced by the loss of
the solvent molecules included upon crystallisation, but it was
chosen as a starting point for adsorption studies. Although a
similar material, prepared by heating a sample of compound 2
that had been prepared directly from solution, was reported to
possess microporosity (270 m2 gÀ1 with micropore volume of
0.11 cm3 mLÀ1 (0.094 cm3 gÀ1)),[9] our material showed no
Type I adsorption of N2 at 77 K. This significant difference
between the physical properties of the two materials (they
As an aside, it has been shown that it is possible to form
compound 3 directly by crystallisation from solution by using
an almost identical experimental preparation as described
above, with the additional precaution to ensure that all
5170
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5170 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 23

Organic Inorganic Hybrid Composites
5168 5175
Figure 5. Optical microscopy (under crossed-polars) showing the temporal
evolution of freshly prepared 1 on conversion to 2 on loss of a single DMF
molecule.
Figure 6. 13C CP-MAS NMR spectra of compunds 1 7 (™ssb×s∫ indicate
spinning side bands).
have identical cell parameters) might be attributed to the
microcrystalline nature of the starting material presented
here. Also, our own N2 adsorption measurements on large
single crystals of directly synthesised 2 heated in vacuum at
1208C (and 1408C) gave no evidence for a Type I isotherm.
Compound 2 can be described as an inclusion compound, with
water and DMF molecules providing critical hydrogen bonds
as well as acting as spacers. These results suggest that the
[Zn(bdc)] structure is strongly rearranged, readily losing
microporosity upon removal of solvent molecules.
Upon first exposing the [Zn(bdc)] (4; solvent-free) solid
(prepared from both single crystals and from polycrystalline
sample) to moist air and subsequently to DMF, it was possible
to fully restore solid compound 2, (in contrast to the reported
behaviour of this solid).[9] Exposure of the freshly desolvated 4
to a range of small molecules that exhibit strong hydrogen
bonding, resulted in a series of highly crystalline solids. The
samples prepared by exposure to methanol gave a powder
pattern essentially identical to that reported by Yaghi[10] for
directly synthesised [Zn3(bdc)3] ¥ 6CH3OH (5), with identical
cell parameters. Examination of the structure of 5 indicated
Figure 4. Top: Conversion of 1 (upper trace) to 2 (lower trace) on loss of
DMFat room temperature over a period of 31 hours monitored using X-ray
powder diffraction. The intermediate conversion (middle trace) was taken
after 24 hours. Middle: Desolvation of 2 monitored using in situ variable
temperature X-ray powder diffraction (lower trace 308C, middle trace
908C, upper trace 1908C). The sample was loaded in an open ended 0.5 mm
external diameter quartz capillary and heated in a furnace in 1 hour steps of
208C from 308C to 2108C. Bottom: Comparison of 3 directly crystallised
from solution (lower trace) with 3 formed by dehydration of 2 at 908C
(upper trace).
Chem. Eur. J. 2001, 7, No. 23
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5171 $ 17.50+.50/0
5171

P. A. Wright et al.
FULL PAPER
that the layers contain units of three zinc polyhedra, with the
zinc units linked by dicarboxylate ligands. Methanol acts as an
interconnecting ligand between the zinc centres and also as an
individual terminal ligand. There is, therefore, considerable
structural rearrangement upon the transformation from
compound 2 via 4 to compound 5. Methanol is an integral
component of this structure and not a guest molecule within a
pore defined by a stable fixed framework.
The role of the included solvent in the framework structure
is demonstrated by allowing the solid to lose methanol and
take up water in moist air. This results in a solid-state
recrystallisation and the formation of a highly crystalline
monohydrate [Zn(bdc)(H2O)] (6). This material can be
dehydrated to form 4, which can then be treated with
methanol to form compound 5; this in turn can reform the
monohydrate on loss of methanol. This interconversion loop
between compounds 4, 5 and 6 has been repeated up to five
times without loss of crystallinity. The structure of this
monohydrate was solved from the powder X-ray diffraction
data (Figure 7, bottom) and was shown to consist of ™zig-zag∫
chains of zinc bdc units (Figure 7). All zinc centres exhibit
trigonal bipyramidal coordination; monodentate bdc units
and a bound water account for the trigonal arrangement, with
À
ˆ
longer Zn O bonds to C O functional groups of neighbour-
À
ing chains binding the structure together (Zn Obdc ˆ 2.01,
À
À
Zn Owater ˆ 1.99, Zn OC ˆ 2.25 ä). There is hydrogen
ˆ
O
bonding between the water molecule and the neighbouring
chains, and zinc oxygen bonds link one chain to two
neighbouring ones. The structural integrity is insufficient to
allow removal of the water without collapse of the framework.
The monohydrate solid is a rather compact and dense
structure without micropores. Dehydration of the monohy-
drate induced collapse of the ™zig-zag∫ chain structure to yield
amorphous 4, which displayed no microporosity.
Upon addition of water to the fully desolvated 4, or
addition of water to the monohydrate, a highly crystalline
dihydrate solid, [Zn(bdc)(H2O)2] (7), forms, the structure of
which was also determined from powder X-ray diffraction
data (Figure 8, bottom). The structure is shown to consist of
chains of distorted tetrahedral zinc centres linked through
monodentate bdc groups to produce a ™zig-zag∫ arrangement
(Figure 8). The remaining two coordination sites around each
zinc centre are occupied by water ligands that serve to provide
hydrogen bonds within arrays of stacked chains and across to
neighbouring chains. The distorted tetrahedra have typical
À
À
bond lengths of Zn Obdc ˆ 2.01 ä, Zn Owater ˆ 2.02 ä, with
Figure 7. Structure of 6. Top: Viewed along the b axis, highlighting the
three-dimensional interconnected network. Middle: Viewed along the a
axis, highlighting the ™zig-zag∫ arrangement of the chains. Bottom: Final
Rietveld refinement for 6.
the hydrogen bonds within each stack of chains and between
À
stacks of chains displaying O O distances of 2.74 ä and
2.71 ä, respectively. The arrangement of the chains and the
coordination of the zinc in the dihydrate solid is such that a
three-dimensional network is not possible. If either the
monohydrate or dihydrate again comes into contact with
methanol, at room temperature, the methanol does not
displace the water molecules from the structure. This materi-
al, unlike the monohydrate, cannot be readily dehydrated
and reformed on contact with water for a second time
(Scheme 1).
added to desolvated intermadiate 4. In each case recrystalli-
sation to a new phase is followed, upon standing in moist air,
by replacement of the alcohol by water to give the mono-
hydrate. 2-tert-Butylphenol is a comparatively large molecule
compared to any micropores available within the framework
and would be expected to be excluded if the framework
remained unchanged. However, it is the hydrogen-bonding
ability of the solvent that is key in the recrystallisation and
Similar behaviour to that observed upon addition of
methanol is also seen if ethanol or 2-tert-butylphenol are
5172
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5172 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 23

Organic Inorganic Hybrid Composites
5168 5175
Scheme 1. Reaction scheme summarising the chemistry of zinc benzene-
dicarboxylate.
A similar type of behaviour to that observed for these
Zn(bdc) compounds, that is, the loss of crystallinity upon
solvent removal followed by recrystallisation in the presence
of hydrogen-bond-forming molecules, has been observed by
Kepert et al. for a series of nickel 1,3,5-benzenetricarbox-
ylates in the presence of alcohols.[5c] They conclude on the
basis of NMR and IR spectroscopy, that the poorly crystalline
intermediate possessed a porous framework that readily re-
adsorbed alcohols.
Solid-state 13C NMR spectra of the series of zinc 1,4-
dicarboxylates studied in this work show the loss of DMF and
the changes in environment of the carboxylate units. In
particular, the spectra are consistent with compound 2
possesing only bidentate carboxylate units, whereas the
chemical shift of intermediate 4 is closer to that observed
for the dihydrate, in which the carboxylate unit is mono-
dentate.
Conclusion
These studies indicate that the zinc dicarboxylates which are
prepared by crystallisation at room temperature are highly
reactive. Removal and replacement of strongly hydrogen-
bonding molecules (DMF, methanol, ethanol, water) result in
solid-state, reconstructive transformations, which can be
considered as being mediated in situ by these ™solvent∫
molecules. Observation of one such transformation under the
optical microscope reveals a reaction front moving through a
single crystal, and leaves a polycrystalline ™pseudomorph∫.
Careful examination of the crystal structures of the reactant
and product show that there are no common structural
features, and there must be considerable structural rearrange-
ment.
The [Zn(bdc)] that is prepared by removal of all DMF and
water from the transformation product of compound 1 was
not found to adsorb N2. The difference between this solid and
that reported by Yaghi may lie in pore blockage resulting from
the polycrystalline nature of compound 2, formed from the
highly crystalline precursor 1. Also, directly-synthesised large
Figure 8. Structure of 7. Top: Viewed along the b axis, highlighting the
™chain-like∫ nature. Middle: Viewed along the c axis, highlighting the ™zig-
zag∫ arangement of the chains. Bottom: Final Rietveld refinement for 7.
subsequent formation of the monohydrate, and not the
presence (or absence) of micropores.
Taken as a series, the increase in density of the zinc 1,4-
benzenedicarboxylates on going from as-prepared
1
(1.515 gcmÀ3), 2 (1.678 gcmÀ3), 3 (1.679 gcmÀ3), 4 (unknown),
5 (1.731 gcmÀ3), 6 (1.936 gcmÀ3) to 7 (1.930 gcmÀ3) is
illustrative of reconstructive changes of the structure, as
hydrogen-bonding molecules are first lost and then subse-
quently taken up, rather than of a fixed framework with
variable guest content.
Chem. Eur. J. 2001, 7, No. 23
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5173 $ 17.50+.50/0
5173

P. A. Wright et al.
FULL PAPER
38.90 H 2.10; TGA: calcd weight loss ˆ 18.00 amu; experimental weight
loss ˆ 18.33 amu.
(0.1 mm) single crystals of 2 did not give Type I adsorption
behaviour when fully desolvated. The fully desolvated zinc
1,4-benzenedicarboxylate (4) was found to be highly reactive,
and to take up DMF and water to reconstitute the original
crystal structure of 2. Furthermore, the desolvated solid
recrystallised when placed in contact with small alcohols and
water, giving highly crystalline solids, including compound 5,
which had previously only been prepared by direct crystal-
lisation.
Our experiments show that suitable solvents enable the
[Zn(bdc)] framework to connect in a variety of different
forms; however, the structural integrity of the frameworks
was not sufficient to enable removal of the solvent without
concurrent recrystallisation or collapse of the framework.
Subsequent inclusion of solvent molecules can be highly
selective, and should be considered in light of framework
formation rather than inclusion into a pre-existing framework.
[Zn(bdc)(H2O)2] (7): Compound 7 was formed directly from compound 4
on addition of water, and from compound 6 on addition of water followed
by drying in air. Solid-state 13C NMR (125 MHz, 258C, adamantane): d ˆ
112.53 118.1 (bdc aromatic), 156.0 159.4 (bdc carbonyl); elemental
analysis calcd (%) for C8H8O6Zn (265.4): C 36.18, H 3.02; found: C 36.26
H 2.70; TGA: calcd weight loss ˆ 36.0 amu; experimental weight loss ˆ
36.46 amu.
X-ray crystallography: The single-crystal X-ray diffraction intensities were
collected over a half hemi-sphere at 293 K on a Bruker CCDdiffractometer
employing graphite monochromated MoKa radiation (l ˆ 0.71073 ä). The
structures were solved by using direct methods and refined by full-matrix
least-squares on F 2 using SHELXTL program. High-resolution powder
X-ray diffraction data were collected by using a STOE Stadi/p diffrac-
tometer in transmission mode, scanning 2q between 58 and 908, with
monochromatic CuKa radiation (curved germanium (111)) with l ˆ
1
1.54056 ä). Samples were mounted between two sheets of mylar foil.
Diffractograms were indexed by using the Visser and Louer algorithms
available within the STOE WINXPOW software. For ab initio structure
solution, pattern decomposition using the Lebail method[11] was carried out
over the range 58 < 2q < 608, with the GSAS suite.[12] For both 6 and 7, Zn
positions were determined by Patterson methods by using SHELXS97.[13]
The remaining non-H atoms were located from difference Fourier maps
derived from Rietveld refinement of the full powder diffraction profile. All
non-H atoms were refined isotropically, with no restraints being required.
H atoms were not located. Further details are given in Tables 1 6.
Crystallographic data (excluding structure factors) for the structures
reported in this paper have been deposited with the Cambridge Crystallo-
graphic Data Centre as supplementary publication no. CCDC-171974,
CCDC-171975, CCDC-171952 and CCDC-171953. Copies of the data can
be obtained free of charge on application to CCDC, 12 Union Road,
Cambridge CB21EZ, UK (fax: (‡44)1223-336-033; e-mail: deposit@ccdc.
cam.ac.uk).
Experimental Section
[Zn3(bdc)3(H2O)3] ¥ 4DMF (1): Triethylamine was allowed to come into
contact, through vapour diffusion, with
a solution of zinc nitrate
(Zn(NO3)2 ¥ 6H2O, 0.95 g, 5.0 mmol), and benzene dicarboxylic acid
(0.41 g, 2.5 mmol) in a mixture of toluene (70 mL) and dimethylformamide
(35 mL). Solid crystalline material formed at the air solvent interface
within hours (typically 24 hours was required to provide sufficient yield),
which was filtered and washed with toluene. Solid-state 13C NMR
(125 MHz, 258C, adamantane): d ˆ 13.6 and 18.5 (CH3)2NCHO), 144 and
146 (CH3)2NCHO), 113.2 120.0 (bdc aromatic),153.8 154.7 (bdc carbon-
yl).
Solid-state NMR spectroscopy: 13C {1H} CP-MAS NMR spectra were
recorded by using a Bruker MSL 500 MHz instrument with spinning speeds
of up to 9.5 kHz, dipolar-dephasing filters of between 100 ms and 500 ms
were employed at the beginning of the FIDacquisition in order to assign
aromatic C and CH groups, and dicarboxylate and DMF carbonyl groups.
Processing was carried out with Bruker WINNMR software and spectra
were fit and analysed with WINMAS and WINFIT.
[Zn(bdc)(H2O)] ¥ DMF (2): Triethylamine was allowed to come into
contact, through vapour diffusion, with
a solution of zinc nitrate
(Zn(NO3)2 ¥ 6H2O, 0.095 g, 0.5 mmol), and benzene dicarboxylic acid
(0.041 g, 0.25 mmol) in a mixture of toluene (70 mL) and dimethylforma-
mide (35 mL). Solid crystalline material formed at the air solvent
interface after five days; this was filtered and washed with toluene. Solid-
state 13C NMR (125 MHz, 258C, adamantane): d ˆ 13.7 and 18.5
(CH3)2NCHO), 144.8 (CH3)2NCHO), 111.2 120.0 (bdc aromatic), 154.8
(bdc carbonyl).
Optical microscopy: The transformation of single crystals of 1 was followed
under the optical microscope. Photographs were taken with a Kodak
Table 1. Single-crystal data for compounds 1, 2 and 5.[10]
[Zn(bdc)]DMF (3): The synthesis of this
compound was identical to that of compound
1 with the added provision of using dry
solvents. Solid-state 13C NMR (125 MHz,
258C, adamantane): d ˆ 11.6 and 20.6
1
2
5
crystal dimensions [mm]
0.11 Â 0.1 Â 0.01
monoclinic
P21/c
13.065(1)
9.661(1)
0.12 Â 0.1 Â 0.01
monoclinic
P21/n
crystal system
space group
a [ä]
triclinic
P1
≈
(CH3)2NCHO),
148.9
(CH3)2NCHO),
110.5 120.5 (bdc aromatic), 154.7 158.8
(bdc carbonyl).
6.733(1)
15.516(1)
12.453(1)
9.726(4)
9.911(5)
10.450(5)
99.72(4)
111.11(4)
108.42(4)
844.9(7)
2
b [ä]
c [ä]
a [8]
b [8]
g [8]
V [ä3]
Z
1calcd [MgmÀ3
l [ä]
q range [8]
T [K]
measured reflections
independent reflections
refinement
18.456(1)
[Zn(bdc)] (4): Compound 4 was formed by
heating compound 2 for three hours at 1908C.
Solid-state 13C NMR (125 MHz, 258C, ada-
mantane): d ˆ 113.0 117.1 (bdc aromatic),
159.5 (bdc carbonyl).
106.87(1)
102.80(1)
2229.2(3)
2
1.515
0.71073
1.63 23.27
293 (2)
1268.7(1)
4
[Zn3(bdc)3] ¥ 6CH3OH (5): Compound 5 was
formed form compound 4 by steeping the
solid in methanol for three hours. This
material can also be formed directly by
crystallisation for solution at room temper-
ature.[10]
]
1.678
1.731
0. 71073
2.13 23.25
293 (2)
0. 71073
2.20 25.05
293 (2)
12342
3204 (Rint ˆ 0.0949)
5472
3577
1819 (Rint ˆ 0.0538)
2998 (Rint ˆ 0.0423)
[Zn(bdc)(H2O)] (6): Compound
6 was
full-matrix least-squares on F 2
formed upon the loss of methanol (at room
temperature) from compound 5. Solid-state
13C NMR (125 MHz, 258C, adamantane): d ˆ
113.3 117.1 (bdc aromatic), 158.7 (bdc car-
bonyl); elemental analysis calcd (%) for
C8H6O5Zn (247.4): C 38.81, H 2.42; found: C
absorption correction/max/min Sadabs/1.00000/0.560220 Sadabs/1.00000/0.789698 Psi scan/0.786/0.670
data/restraints/parameters
R (obs)
3154/2/286
0.0441
1773/3/181
0.0347
2998/0/236
0.0642
Rw (obs)
F 2 (goodness of fit)
0.0944
0.873
0.0667
1.035
0.1395
1.024
5174
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5174 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 23

Organic Inorganic Hybrid Composites
5168 5175
Table 2. Powder X-ray data for compounds 3, 6 and 7.
Table 6. Selected bond lengths [ä] and bond angles [8] for compound 7.
3[a]
6
7
Zn1 O2
2.03(1) Zn1 O3
2.03(1) C6 C7
1.42(2)
1.41(2)
1.29(1)
À
À
À
À
À
À
Zn1 C8
2.64(1) O3 C8
1.27(1) C4 C6
crystal system
space group
a [ä]
b [ä]
c [ä]
triclinic
monoclinic
C2/c
17.979(1)
6.353(3)
7.257(3)
monoclinic
C2/c
14.992(8)
5.030(2)
12.098(5)
À
À
À
C4 C7
1.41(2) C4 C8
1.59(2) O5 C8
≈
P1
O2-Zn1-O2
92.9(5)
O2-Zn1-O3 135.3(3)
O3-C8-C4 118.1(1)
C4-C8-O5 120.1(2)
C4-C7-C6 112.2(1)
O3-C8-O5 121.7(1)
9.055(1)
8.921(2)
7.940(1)
99.90(8)
100.91(8)
103.28(8)
597.5 (27)
2
1.679
1.54056
5
293
O2-Zn1-C8 100.8(4)
Zn1-O3-C8 103.6(9)
C6-C4-C8
C4-C6-C7
O3-Zn1-C8
C6-C4-C7
C7-C4-C8
98.7(5)
133.8(2)
112.9(2)
113.3(2)
113.8(1)
a [8]
b [8]
g [8]
V [ä3]
Z
91.47(4)
103.82(3)
828.6 (9)
2
1.936
1.54056
5
293
886.0 (9)
2
1.930
1.54056
5 90
293
Thermogravimetric analysis: The interconversion from 1 to 3 occurs
through the loss of DMF from the crystal lattice. Thermo-gravimetric
analysis (TGA) was performed between 308C and 6008C under both
nitrogen and oxygen gas flow, with a TA Instrument thermal balance and
analysed by using TA Instrument Universal Analysis software.
1calcd [MgmÀ3
l [ä]
]
2q range [8]
T [K]
60
90
reflections used for solution
101
131
reflections used for refinement
total parameters
353
41
397
44
Acknowledgement
Rwp
0.1147
27.21
0.1146
12.88
The EPSRC are thanked (ME) for financial support, Humphrey Yiu for
assistance with the TGA measurements, Sylvia Williamson for CHN
analysis, and Milanja Smith and Frank Riddell for NMR expertise.
c2
[a] Indexed from powder diffraction data collected from directly crystal-
lised material.
[1] a) K. Maeda, A. Sasaki, K. Watanabe, Y. Kiyozumi, F. Mizukami,
Chem. Lett. 1997, 9, 879; b) K. Maeda, Y. Kiyozumi, F. Mizukami, J.
Phys. Chem. B 1997, 101, 4402 ; c) K. Maeda, J. Akimoto, Y. Kiyozumi,
F. Mizukami, J. Chem. Soc. Chem. Commun. 1995, 1033; d) V. J.
Carter, P. A. Wright, J. D. Gale, R. E. Morris, E. Sastre, J. Perez-
Pariente, J. Mater. Chem. 1997, 7, 2287; e) G. Alberti, R. Vivani, F.
Marmottini, P. Zappelli, J. Porous Mater. 1998, 5, 205; f) A. Clearfield,
D. Wang, Y. Tian, E. Stein, C. Bhardwaj, J. Solid State Chem. 1995, 117,
275.
[2] a) O. M. Yaghi, C. E. Davis, H. Li, D. Richardson, T. L. Groy, Acc.
Chem. Res. 1998, 31, 474; b) M. Eddaoudi, D. B. Moler, H. Li, B. Chen,
T. M. Reineke, M. O×Keeffe, O. M. Yaghi, Acc. Chem. Res. 2001, 34,
319.
Table 3. Fractional atomic coordinates for compound 6.
x
y
z
Ui (Â100)
Zn1
O2
O3
O4
C5
C6
C7
C8
0.000000
0.6863(6)
0.250000
0.1599(14)
0.250000
0.5343(11)
0.0668(23)
0.0329(18)
0.0904(20)
2.29(1)
1.72(3)
1.72(3)
1.72(3)
1.37(3)
1.37(3)
1.37(3)
1.37(3)
À 0.0982(5)
0.5659(17)
1.0048(19)
0.7067(15)
0.3875(23)
0.2942(33)
0.4592(21)
0.1080(28)
0.000000
À 0.0476(4)
À 0.1126(7)
À 0.1839(8)
À 0.2410(1)
À 0.1892(9)
À 0.0538(21)
[3] a) C. J. Kepert, M. J. Rosseinsky, Chem. Commun. 1998, 31; b) C.
Livage, C. Egger, M. Nogues, G. Ferey, J. Mater. Chem. 1998, 8, 2743;
c) F. Serpaggi, G. Ferey, Microporous Mesoporous Mater. 1999, 32,
311; d) O. M. Yaghi, H. L. Li, T. L. Groy, Inorg. Chem. 1997, 36, 4292;
e) S. Subramanian, M. J. Zaworotko, Angew. Chem. 1995, 107, 2295;
Angew. Chem. Int. Ed. Engl. 1995, 34, 2127.
[4] a) C. J. Kepert, T. J. Prior, M. J. Rosseinsky, J. Solid State Chem. 2000,
152, 261; b) M. J. Plater, M. R. St J. Foreman, E. Coronado, C. J.
Gomez-Garcia, A. M. Z. Slawin, J. Chem. Soc. Dalton Trans. 1999,
4209.
[5] a) T. M. Reineke, M. Eddaoudi, M. O×Keeffe, O. M. Yaghi, Angew.
Chem. 1999, 111, 2712; Angew. Chem. Int. Ed. 1999, 38, 2590; b) T. M.
Reineke, M. Eddaoudi, M. Fehr, D. Kelly, O. M. Yaghi, J. Am. Chem.
Soc. 1999, 121, 1651; c) C. J. Kepert, T. J. Prior, M. J. Rosseinsky, J.
Am. Chem. Soc. 2000, 122, 5158.
Table 4. Selected bond lengths [ä] and bond angles [8] for compound 6.
À
À
À
Zn1 O2
2.02(1) Zn1 O3
2.02(1) O2 Zn1
2.02(1)
1.34(1)
1.53(2)
1.34(2)
À
À
À
O4 C5
1.34(1) O2 O4
2.17(1) O2 C5
À
À
À
C6 C8
1.34(2) C5 C6
1.43(2) C6 C7
À
À
À
C8 C7
1.34(3) C7 C8
1.34(2) C8 C6
O2-Zn1-O2 135.5(6)
O2-Zn1-O3 112.3(3)
C5-C6-C8 119.9(2)
C6-C7-C8 110.9(1)
C5-C6-C7 105.9(2)
C7-C6-C8 133.6(2)
C6-C8-C7 115.3(2)
O2-Zn1-O4
O3-Zn1-O4
O4-O2-C5
O2-C5-O4
O4-C5-C6
88.2(4)
86.7(3)
129.7(2)
108.0(1)
124.6(2)
O2-Zn1-O4
O4-Zn1-O4 173.4(6)
O2-O4-C5
O2-C5-C6
94.3(3)
119.2(2)
126.7(2)
[6] a) O. M. Yaghi, C. E. Davis, G. M. Li, H. L. Li, J. Am. Chem. Soc. 1997,
119, 2861; b) O. M. Yaghi, C. E. Davis, G. M. Li, H. L. Li, J. Am.
Chem. Soc. 1997, 119, 2861; c) O. M. Yaghi, H. L. Li, T. L. Groy, J. Am.
Chem. Soc. 1996, 118, 9096; d) O. M. Yaghi, R. Jernigan, H. Li, C. E.
Davis, T. L. Groy, J. Chem. Soc. Dalton. Trans. 1997, 2383.
[7] H. Li, M. Eddaoudi, M. Keefe, O. M. Yaghi, Nature 1999, 402, 276.
[8] A. D. Burrows, R. W. Harrington, M. F. Mahon, C. E. Price, Dalton
2000, 3845.
[9] H. Li, M. Eddaoudi, T. L. Groy, O. M. Yaghi, J. Am. Chem. Soc. 1998,
120, 8571.
[10] H. L. Li, C. E. Davis, T. L. Groy, D. G. Kelley, O. M. Yaghi, J. Am.
Chem. Soc. 1998, 120, 2186.
Table 5. Fractional atomic coordinates for compound 7.
x
y
z
Ui (Â100)
Zn1
O2
O3
C4
O5
C6
C7
C8
0.500000
0.0643(7)
0.250000
4.18(2)
2.96(2)
2.96(2)
1.80(3)
2.96(2)
1.80(3)
1.80(3)
1.80(3)
0.5527(6)
0.3944(7)
0.3094(11)
0.4378(7)
0.2407(13)
0.3241(10)
0.3849(11)
À 0.2134(16)
0.3271(17)
0.5809(28)
0.2617(16)
0.6237(28)
0.6778(30)
0.3712(28)
0.1620(8)
0.2121(8)
0.0464(14)
0.0496(8)
0.1063(11)
À 0.0572(14)
0.1063(12)
[11] A. Le Bail, J. Solid State Chem. 1989, 83, 267.
[12] A. Larson, R. Von Deele, Los Alamos National Laboratory, Report
No. LA-UR-86 748, 1987.
[13] G. M. Sheldrick, SHELXTL, Program for the solution of crystal
structures, University of Gˆtingen, 1993.
800 ASA colour film in a Centron DF-300 camera mounted on a Prior
optical microscope employing crossed-polars and filter. The freshly
prepared crystalline precursor material 1 displayed birefringence when
viewed under crossed polars; this allowed the transformation to compound
2 to be monitored.
Received: May 7, 2001 [F3243]
Chem. Eur. J. 2001, 7, No. 23
¹ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0723-5175 $ 17.50+.50/0
5175