Dual-supramolecular contacts induce extreme Hofmann framework distortion and multi-stepped spin-crossover

Article abstract of DOI:10.1039/d0dt04007j

An extended nitro-functionalised 1,2,4-triazole ligand has been used to induce considerable lattice distortion in a 2-D Hofmann framework materialviacompeting supramolecular interactions. Single crystal X-ray diffraction analyses on [Fe₃(N-cintrz)₆(Pd(CN)₄)₃]·6H₂O (N-cintrz: (E)-3-(2-nitrophenyl)acrylaldehyde) reveal a substantial deviation from a regular Hofmann structure, in particular as the intra- and inter-layer contacts are dominated by hydrogen-bonding interactions rather than the typical π-stacking arrays. Also, the 2-D Hofmann layers show an assortment of ligand conformations and local Fe^IIcoordination environments driven by the optimisation of competing supramolecular contacts. Temperature-dependent magnetic susceptibility measurements reveal a two-step spin crossover (SCO) transition. Variable temperature structural analyses show that the two crystallographically distinct Fe^IIcentres, which are arranged in stripes (2?:?1 ratio) within each Hofmann layer, undergo a cooperative HS ? HS/LS ? LS (HS = high spin, LS = low spin) transition without periodic spin-state ordering. The mismatch between crystallographic (2?:?1) and spin-state (1?:?1) periodicity at the HS?:?LS step provides key insight into the competition (frustration) between elastic interactions and crystallographically driven order.

Full text of DOI:10.1039/d0dt04007j

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Dual-supramolecular contacts induce extreme
Hofmann framework distortion and multi-stepped
spin-crossover†
Cite this: Dalton Trans., 2021, 50,
1434
d
Manan Ahmed,a Helen E. A. Brand,b Vanessa K. Peterson, c Jack K. Clegg,
Cameron J. Kepert,e Jason R. Price,a,b Benjamin J. Powellf and
a
Suzanne M. Neville
*
An extended nitro-functionalised 1,2,4-triazole ligand has been used to induce considerable lattice distor-
tion in a 2-D Hofmann framework material via competing supramolecular interactions. Single crystal
X-ray diffraction analyses on [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O (N-cintrz: (E)-3-(2-nitrophenyl)acrylalde-
hyde) reveal a substantial deviation from a regular Hofmann structure, in particular as the intra- and inter-
layer contacts are dominated by hydrogen-bonding interactions rather than the typical π-stacking arrays.
Also, the 2-D Hofmann layers show an assortment of ligand conformations and local FeII coordination
environments driven by the optimisation of competing supramolecular contacts. Temperature-dependent
magnetic susceptibility measurements reveal a two-step spin crossover (SCO) transition. Variable temp-
erature structural analyses show that the two crystallographically distinct FeII centres, which are arranged
in stripes (2 : 1 ratio) within each Hofmann layer, undergo a cooperative HS ↔ HS/LS ↔ LS (HS = high spin,
LS = low spin) transition without periodic spin-state ordering. The mismatch between crystallographic
(2 : 1) and spin-state (1 : 1) periodicity at the HS : LS step provides key insight into the competition (frustra-
tion) between elastic interactions and crystallographically driven order.
Received 23rd November 2020,
Accepted 4th January 2021
DOI: 10.1039/d0dt04007j
rsc.li/dalton
be equivalent.9 Comprehensively describing multi-stepped
Introduction
transitions is a substantial challenge10,11 as it must encompass
The conversion between low-spin diamagnetic (t2g6eg ; LS) and both ferro- and anti-ferro-elastic interactions. In complex
0
high-spin paramagnetic (t2g4eg ; HS) states in FeII-based material structures not all interactions can be simultaneously
2
systems has attracted interest over a wide range of scientific minimised – this results in elastic frustration,10 and multistep
fields
for
fundamental
and
application-driven transitions.10,11 Further, it has been shown that strong
considerations.1–3 Such materials display an assortment of through-bond (covalent) interactions lead to anti-ferro-elastic
thermal behaviours encompassing gradual transitions, first- interactions, whereas weak through-space interactions often
order switching with hysteretic effects and multi-stepped pro- result in ferro-elastic interactions.11,12
files. Central to communication in spin transitions are elastic
To this end, many studies utilise the intrinsic long-range
interactions due to structural changes between HS and LS connectivity of coordination networks to both enhance and
phases. First-order hysteretic transitions can be described by disrupt spin-state conversion communication.13–16 Recent
elastic models4–7 or Ising-like models,8 which can be shown to reports have shown that Hofmann-type frameworks provide an
excellent platform to design, synthesise and develop a deeper
understanding of elastic frustration.11,17–28 Since these frame-
works are constructed by a modular (metal + ligand) approach,
aSchool of Chemistry, The University of New South Wales, Sydney, 2052, Australia.
E-mail: s.neville@unsw.edu.au
supramolecular interaction networks which act to induce and
bThe Australian Synchrotron, 800 Blackburn Road, Clayton, Victoria, Australia
stabilise elastic frustration10 can be readily included via ligand
cANSTO, Lucas Heights, New South Wales, Australia
dSchool of Chemistry and Molecular Biosciences, The University of Queensland,
functionalisation. Furthermore, as these systems often encap-
sulate guest molecules in the interlayer spacing, antagonistic
host–guest interactions can also be utilised to subtly modify
the magnitude of directionality of frustration effects. The 2-D
Hofmann framework [Fe(bztrz)2Pd(CN)4]·guest (bztrz = (E)-1-
phenyl-N-(1,2,4-triazol-4-yl)-methanimine)21 exemplifies this
approach, whereby utilising a functionalised 1,2,4-triazole
St Lucia, Queensland 4072, Australia
eThe School of Chemistry, The University of Sydney, Sydney, 2006, Australia
fSchool of Mathematics and Physics, The University of Queensland, St Lucia,
Queensland 4072, Australia
†Electronic supplementary information (ESI) available: Additional structural
and magnetic data. CCDC 2045470–2045472. For ESI and crystallographic data
in CIF or other electronic format see DOI: 10.1039/d0dt04007j
1434 | Dalton Trans., 2021, 50, 1434–1442
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
4H-1,2,4-triazole. The 2-D Hofmann framework [Fe(N-
cintrz)2Pd(CN)4]·3H2O was formed as single crystals by slow
diffusion of Fe(ClO4)2·6H2O, N-cintrz ligand and K2[Pd(CN)4]
in ethanol and water (50 : 50). Red plate-shaped single crystals
were produced over a period of 8 weeks. IR spectroscopy shows
a strong CN stretching band (2113 cm−1) from the Pd(CN)4
ligand and an azomethine (CvN) stretching band from the
N-cintrz ligand. The bulk phase purity was confirmed by
powder X-ray diffraction (Fig. S9†) and elemental (CHN) ana-
lysis. Thermogravimetric analysis (RT-600 °C; 1 °C min−1
)
shows a mass loss of ca. 2.25% up to 180 °C associated with
the loss of 3 water molecules per FeII site; above 180 °C frame-
work decomposition occurs (Fig. S1†).
Fig. 1 (a) (E)-3-(2-Nitrophenyl) acrylaldehyde (N-cintrz) and (b) (E)-2-
(((4H-1,2,4-triazol-4-yl)imino)methyl)phenol (saltrz). The secondary
hydrogen bonding functional groups are coloured red.
Overall structure description
The single crystal structure shows the material is composed of
two crystallographic distinct FeII sites in a 21Fe1 : 1 Fe2 ratio
(due to Fe1 residing on an inversion centre), one and a half
[Pd(CN)4]2− groups, three N-cintrz ligands (L1, L2 and L3) and
three water molecules (2 × A type with hydrogen bonds and 1 ×
B type with van der Waals interactions; the later are disordered
over two positions at a 50 : 50 occupancy). Thus, the overall
formula is [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O. The asymmetric
unit at 100 and 150 K are the same (Fig. S2 and 3†).
ligand results in Hofmann layer distortion and the low temp-
erature stabilisation of a mixed spin-state species (HS : LS).
Guest variation in this framework allows a range of multi-
stepped transitions to emerge (half-, two- and three-stepped)
through changes to supramolecular interactions and steric
effects. This concept opens a huge number of possibilities to
produce and develop a deeper understanding the origins of
multi-stepped transitions.
Judicious ligand functionalisation has been shown to be a
key tool in tailoring multi-stepped SCO behaviour. In particu-
lar, ligands with hydrogen bonding and/or aromatic inter-
action capability have been extensively utilised to drive a com-
petition between long-range ferro- and anti-ferro-elastic inter-
actions. While there are a range of conceivable synthetic
methods available to induce such features, we and others have
been focused on utilising the unique metal ion binding mode
of 1,2,4-triazole ligands (Fig. 1) in Hofmann frameworks, and
facile functionalisation of such ligands as an effective
approach to generate competition between elastic
interactions.17–22,26,28 Here, alongside utilising a functiona-
lised 1,2,4-triazole ligand to construct a 2-D Hofmann frame-
work, we have appended a nitro-group to the periphery of the
ligand to supply supramolecular interaction capacity beyond
the established 1,2,4-triazole interactions (Fig. 1a). This
approach was applied to the related 2-D Hofmann framework
[Fe3(saltrz)6(Pd(CN)4)3]·8H2O (Fig. 1b),22 where by ligand the
hydroxyl-function groups induced a wide array of competing
host–host and host–guest interactions resulting in substantial
structural distortion and a four-stepped spin-state transition
pathway. Therefore, here our overall aim is to examine the dis-
tortive influence of a combination of 1,2,4-triazole and nitro-
group supramolecular contacts on the structure and SCO pro-
perties in this new 2-D Hofmann framework material.
Each FeII centre (Fe1 and Fe2) adopts a distorted octahedral
geometry with six N-donor atoms, consisting of four [Pd
(CN)4]2− anions at the equatorial positions and two N-cintrz
ligands in the axial positions. The Fe1 and Fe2 sites differ crys-
tallographically due to their N-cintrz ligand connectivity, with
Fe1 coordinated to L3 and the inversion centre equivalent (L3′)
and Fe2 coordinated to L1 and L2. The N-cintrz ligands are co-
ordinated through N1 of the 1,2,4-triazole group (Fig. 2).
The three crystallographically distinct N-cintrz ligands (L1,
L2 & L3) differ significantly in their geometry and relative
arrangement in the interlayer spacing (Fig. 2). Notably, a trans-
orientation is observed for L1 and L2 with respect to the rela-
tive position of the unbound N2 of the triazole group and the
peripheral NO2 group, whereas a cis-arrangement is present in
L3 (Fig. 2). Furthermore, L1 and L2 show greater torsion
between the triazole and nitro-benzene rings compared to L3
(Table S2†). The Fe–N–N(trz) ‘bite’ angle differs for each ligand
and FeII site (Table S2†).
The framework structure is formed by the bridging of FeII
sites by [Pd(CN)4]2− anions to produce neutral 2D Hofmann
grid-type layers, [FeIIPd(CN)4] (Fig. 3(a)), which lie approxi-
mately in the ac-plane. Within each layer, the Fe1 and Fe2 sites
are arranged in stripes along the a-axis, with the Fe1 arrays
spaced by pairs of Fe2 stripes (Fig. 3(b and c)). The Fe1 and
Fe2 sites show different degrees of octahedral tilting with
respect to an idealised planar layer (Fig. 3(b); Table S2†),
resulting in layer undulation. Further highlighting the layer
distortion, the Fe–NuC angles deviate significantly from 180°
(Table S2†).
Results and discussion
Synthesis and characterisation
The interlayer spacing is derived from the head-to-tail inter-
The N-cintrz ligand (Fig. 1a) is readily prepared by a conden- digitation of N-cintrz ligands from neighbouring layers
sation reaction between 2-nitrocinnamaldehyde and 4-amino- (Fig. 4). Adjacent Hofmann layers are stacked along the crystal-
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1434–1442 | 1435

View Article Online
Paper
Dalton Transactions
Fig. 2 Thermal ellipsoid plot (50% probability) of the asymmetric unit
of [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O at 250 K showing the two FeII sites
(Fe1 and Fe2), the three distinct N-cintrz ligands (L1, L2 and L3), the Pd
(CN)4 ligands and the water molecules (2 × type A hydrogen bonded
and type B with van der Waals interactions and disordered over two
sites).
Fig. 4 Structural view of two adjacent 2-D layers (red & blue) highlight-
ing the head-to-tail ligand interdigitation and array of nitro-group place-
ments within the interlayer spacing (the NO2 groups are coloured green
for clarity).
Unlike the majority of 2-D Hofmann frameworks, the intra-
and inter-molecular interactions are dominated by host–host
hydrogen bonding interactions (Fig. 5).22 The nitro-groups
from L1 and L2 form various intra-layer C(H) contacts with
neighbouring ligands, as shown in Fig. 5(a). There are also
various intra-layer N⋯C(H) interactions involving L2 and L3
(Fig. 5(a)). The nitro-group from L3 forms the only inter-layer
contact (O⋯Ntrz; Table S3†). The water molecules are engaged
in an array of host–guest interactions, as depicted in Fig. 5(b).
lographic b-axis, with a relatively large interlayer spacing of ca.
15 Å (Fig. 4; Table S2†). The nitro-groups from the N-cintrz
ligands are directed into the interlayer spacing between adja-
cent ligands as depicted in Fig. 4. Despite the substantial layer
distortion, the layers are efficiently packed with a near-com-
plete head-to-tail overlap of N-cintrz ligands from layers above
and below (Fig. 4).
Fig. 3 Structural representation of [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O at 100 K. (a) A single 2-D Hofmann layer, highlighting the Fe1 (red) and Fe2
(orange) distribution in stripes along the a-direction. (b)–(c) One 2-D layer illustrating the subtle layer undulation and respective N-cintrz ligands
protruding above and below the layers and array of ligand geometries.
1436 | Dalton Trans., 2021, 50, 1434–1442
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Fig. 5 Structural representation of [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O (100 K) showing the array of (a) host–host (green) and (b) host–guest (purple)
hydrogen bonding interactions present within a single Hofmann layer. The ligands (L1, L2 & L3) are labelled, and hydrogen atoms have been omitted
for clarity.
Spin crossover properties
Temperature-depended magnetic susceptibility data were col-
lected on a bulk crystalline sample of [Fe3(N-cintrz)6(Pd
(CN)4)3]·6H2O over the range of 300–50–300 K (Fig. 6(a)). At
300 K, the observed χMT values are ∼3.5 cm3 K mol−1, consist-
ent with HS FeII sites. With cooling, the χMT values remain
constant and then gradually decrease between 210 and 100 K,
with the χMT values below 100 K (∼0.07 cm3 K mol−1) consist-
ent with LS FeII sites. A subtle two-step character is evident in
the SCO profile as can be seen in the δχMT/δT plot (Fig. 6(a):
inset). Data were collected at a variety of thermal scan rates
(0.5, 1, 2 and 4 K min−1; Fig. S5†) to determine the nature of
the thermal SCO transition without kinetic effects. Differential
scanning measurements were also conducted and show endo-
and exo-thermic peaks at temperatures consistent with those
obtained by magnetic susceptibility (Fig. S6†).
Variable temperature structural analysis
Variable temperature powder X-ray diffraction data were col-
lected using synchrotron radiation to monitor the structural
evolution over the SCO temperature range (280–100–280 K).
The thermal dependence of the peak positions shows a clear
two-step SCO behaviour (Fig. 6(b): inset; Fig. S13†). A plot of
the peak position versus temperature maps the thermal depen-
Fig. 6 (a) Temperature-dependent magnetic susceptibility plot (χMT vs.
dence of the magnetic susceptibility trace in SCO character
T). Data recorded at
300–50–300 K in continuous mode. Other scan rates (4, 1, 0.5 K min−1
a scan rate of 2 K
min−1 over the range
and temperature (Fig. 6(b)). Fine analysis of the individual pat-
terns did not reveal a phase transition over the entire SCO
temperature range.
Single crystal data were collected for the HS (250 K), inter-
mediate plateau (150 K), and LS (100 K) states, respectively.
The average Fe–N bond distance at 250 K, for both Fe1 and Fe2
)
are provided in the ESI.† (b) Variable temperature (280–100–280 K) syn-
chrotron powder diffraction peak position evolution (2 0 −1 reflection; λ
= 0.5903 Å).
are consistent with HS FeII sites (Table 1), as anticipated from both Fe1 and Fe2 are consistent with LS FeII sites (Table 1).
the magnetic susceptibility data. The degree of octahedral dis- The Σ parameter at Fe1 is seen to increase with transition from
tortion (Σ) at Fe1 is slightly reduced compared to that of Fe2 HS to LS, whereas the Σ parameter at Fe2 reduces slightly
(Table 1). Similarly, the average Fe–N bond lengths at 100 K for (Table 1). Despite the Fe1 and Fe2 sites being in quite different
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1434–1442 | 1437

View Article Online
Paper
Dalton Transactions
Table 1 Selected variable temperature structural parameters
To probe these ligand design features further, we investigate
here the coexistence of tailored hydrogen-bonding capacity
(i.e., the nitro-group) which should provide enhanced anti-
ferro-elastic character when combined with the known propen-
sity of the 1,2,4-triazole group to form various antagonistic
host–host and host–guest contacts17–22,26,28 and that of guest
molecules.
Temperature/K
250
150
100
d〈Fe1–N〉 a/Å
d〈Fe2–N〉 a/Å
Æ©Fe1 b/°
Æ©Fe2 b/°
2.161(2)
2.162(2)
17.6
2.050(2)
2.031(3)
18.0
1.959(5)
1.953(5)
19.3
21.3
19.9
20.6
a Average bond length of the FeII octahedron. b Octahedral distortion
parameter (sum of | 90 − θ | for the 12 cis angles of the FeII
octahedron).
From a structural perspective, the 2-D Hofmann framework
[Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O closely resembles that of
[Fe3(saltrz)6(Pd(CN)4)3]·8H2O.22 In particular, the two distinct
FeII sites (Fe1 and Fe2, Fig. 2), their relative ratio (1 : 2) and
their distribution within the Hofmann layers (i.e., in 1 : 2
stripes, Fig. 3a and b) is equivalent in both materials.
Furthermore, even with the distinction in ligand length and
hydrogen bond donor group (i.e., NO2 vs. OH) in N-cintrz and
saltrz (Fig. 1), the range of ligand conformations (i.e., cis vs.
trans and planar vs. twisted, see Fig. 2) are identical. Moreover,
whilst there is a ∼2 Å difference in layer separation between
the two due to the longer N-cintrz ligand (N-cintrz: 14.8 Å;
saltrz: 12.4 Å in HS state), the 2-D layer packings achieved in
these structures are each dominated by hydrogen-bonding
rather than aromatic contacts typical in Hofmann frameworks.
Given the irregularity of this overall structure, it is somewhat
surprising that the same Hofmann distortions are observed in
both and suggests that it is an energetically preferred struc-
tural arrangement. Together, these two structural studies
clearly illustrate the fruitful strategy of including secondary
hydrogen bonding capacity to drive substantially distorted
Hofmann structures.
local coordination environments, and the presence of a two-
step feature in the SCO transition profile, the average Fe–N
bonds at 150 K for both Fe1 and Fe2 are intermediate between
HS and LS (Table 1). This suggests that at the intermediate
plateau region there is no three-dimensional long-range order-
ing of HS and LS sites. Careful inspection of precession
images at 150 K did not uncover any diffuse features consistent
with site ordering at a shorter length scale or at a lower dimen-
sion (Fig. S4†).
Aside from the spin-state variation to the FeII sites, the
Hofmann framework structure remains relatively unchanged
over the SCO temperature range. There is a ca. 4.5% overall
decrease in unit cell volume, consistent with a complete HS to
LS transition. Hirshfeld surface analysis (fingerprint
decomposition),30,31 highlights the domination of NO⋯HC
and N⋯HO interactions irrespective of temperature (Fig. S7†),
and the fingerprint plots (Fig. S8†) indicate that temperature
(and hence spin-state) do not drastically affect the inter-
molecular interactions.
The primary motivation for introducing lattice distortion
into 2-D Hofmann frameworks was to further probe and
develop a deeper understanding of elastic competition and the
relationship to multi-stepped SCO transitions.10,11 Broadly
speaking the lattice distortive effects of [Fe3(N-cintrz)6(Pd
Discussion
A series of targeted 2-D Hofmann materials incorporating sys- (CN)4)3]·6H2O and [Fe3(saltrz)6(Pd(CN)4)3]·8H2O22 in the HS
tematically functionalized 1,2,4-triazole ligands exist;17–22,26,28 state are equivalent. This equivalent distortive effect at the FeII
these have provided unique insight into the structural and sites is readily quantified by comparing the Σ parameter at
spin-state switching effect of ferro- and anti-ferro-elasticity each of the HS FeII sites, whereby Σ values of ∼18° and ∼21°
generated by competing supramolecular interactions.10–12 All are observed for Fe1 and Fe2, respectively, in both materials.
of the [Fe(R-trz)2M(CN)4]·guest examples thus far reported Despite this seemingly equivalent inherent frustration of the
show SCO properties dictated by elastic frustration arising HS lattices, with temperature perturbation very different SCO
from the unique monodentate 1,2,4-triazole binding mode traces are observed: [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O shows a
which intrinsically supports antagonistic supramolecular inter- gradual two-step SCO and [Fe3(saltrz)6(Pd(CN)4)3]·8H2O22
actions (host–host and host–guest). An important facet of this shows a cooperative four-step SCO.
ligand design is that supplementary supramolecular inter-
One of the primary distinctions between [Fe3(N-cintrz)6(Pd
action capacity can be readily integrated into the host network (CN)4)3]·6H2O and [Fe3(saltrz)6(Pd(CN)4)3]·8H2O22 is the inter-
via the inclusion of ligand functional groups with specific layer communication pathways. There are two interlayer com-
hydrogen-bonding interaction capacity and guest molecules. munication pathways via hydrogen bonding in [Fe3(N-
The effectiveness of this approach is illustrated in the example cintrz)6(Pd(CN)4)3]·6H2O, NO2⋯Ntrz and NO2⋯C(H), as
[Fe3(saltrz)6(Pd(CN)4)3]·8H2O (Fig. 1b),22 which shows substan- depicted in Fig. 7. These interactions act as direct connections
tial Hofmann layer distortion and a four-stepped SCO tran- between Fe1 and Fe2 on distinct layers (Fig. 7). There are also
sition driven by ligand hydroxyl group interactions. A further various interlayer aromatic contacts between head-to-tail over-
important feature of this 2-D structural platform is that the lapped N-cintrz ligands (Fig. 4). On the other hand, for
intermolecular communication pathways, which are important [Fe3(saltrz)6(Pd(CN)4)3]·8H2O,22 there are no direct interlayer
for the long-range transmission of spin-state change, can be hydrogen bonding interactions, but longer communication
tailored by other factors such as ligand length and aromaticity. pathways do exist via the water molecules (host⋯water⋯host).
1438 | Dalton Trans., 2021, 50, 1434–1442
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
prevent long-range spin-state ordering from occurring – in
which case the two peaks observed in δχMT/δT must indicate
spin-state transitions rather than true thermodynamic phase
transitions.11,34 Alternatively, the peaks in δχMT/δT could
signal true phase transition with long-range order, but the
defects could prevent this from being detectable crystallogra-
phically. Secondly, it may be that long-range ordering of
HS : LS sites occurs in less than three-dimensions, yielding
diffuse scattering too weak to observe due to the subtlety of
the modulation. A possible scenario here is the formation of
two-dimensionally ordered HS : LS Hofmann layers, e.g., with
alternation of HS/LS sites in one (striped) or two (checkboard)
directions within the layers, with there being no long-range
correlation in the third, interlayer direction, due to the com-
parative weakness of the interlayer coupling and, again, to the
mismatch of the crystallographic periodicity. Notably, in both
scenarios the elastic coupling inherent within the Hofmann
Fig. 7 The intermolecular contacts (red dashes) between adjacent
layers (red and blue). Fe1 and Fe2 sites noted.
The primary mode of interlayer communication is via aromatic lattice to some large extent overrides any tendency the FeII
contacts between head-to-tail overlapped saltrz ligands. There sites may have to transition based on the inequivalence of
is clear theoretical evidence that frustrated elastic interactions their local geometries. This situation is similar in one sense to
can lead to disordered states in both two-12 and three-dimen- the behaviour seen in [Fe3(saltrz)6(Pd(CN)4)3]·8H2O,22 in which
sional31 models. Therefore, the enhanced intermolecular inter- inequivalent FeII sites transition within single steps of the mul-
actions in [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O may mean that this tistep transition, but different in another in that in
framework is more strongly frustrated than [Fe3(saltrz)6(Pd [Fe3(saltrz)6(Pd(CN)4)3]·8H2O22 some partial degree of three-
(CN)4)3]·8H2O.22 Rather than leading to additional spin-tran- dimensional coherence of HS and LS sites was observed, albeit
sition steps, this probable enhanced frustration and emerges with some crystallographic sites being mixed spin-state arrays.
in [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O as short-range spin-state
order at the HS : LS intermediate plateau. This picture is sup-
ported by recent calculations that suggest that long-range
Experimental
interactions are important for the spin-state ordering arrays
General
observed in [Fe3(saltrz)6(Pd(CN)4)3]·8H2O.11,22
Chemical and solvents were an analytical grade, commercially
available and used without further purifications.
Caution: Iron(II) perchlorate is potentially explosive, care-
fully handle and the minimum amount was used to avoid any
explosion.
The observation of a 50 : 50 HS : LS plateau in a parent
system that contains a 2 : 1 ratio of inequivalent FeII sites
raises a number of interesting scenarios regarding the local
stabilisation of this mixed spin-state.32 Both elastic inter-
actions and crystallographically distinct metal sites have been
proposes as driving forces for antiferro-elastic order.33 Firstly,
we note that this 50% plateau is incompatible with the parent
Preparation of nitrocinnamaldehyde
symmetry of the structure, and that any 50% ordering necess- Nitrocinnamaldehyde was synthesized using
a modified
arily requires reduction in the local symmetry; this is consist- reported method.35 A round bottom flask was charged with
ent with the crystallographic observation of intermediate Fe–N 1.32 g (10 mmol) cinnamaldehyde and 5.35 mL acetic anhy-
distances, with site-averaging of local HS and LS states at each dride and cooled in an ice bath while stirring. When the temp-
of the inequivalent FeII sites at the plateau temperature, 150 K. erature reached 0 °C, a solution of concentrated nitric acid
The absence of any discernible evidence for symmetry (0.42 mL) in glacial acetic acid (1.19 mL) was added via a drop-
reduction in the X-ray diffraction data at the SCO plateau, ping funnel over 2 hours with vigorous stirring and ensuring
either in the form of Bragg modulation peaks or diffuse scat- the temperature remained below 5 °C. After the addition, the
tering features, suggests two possible scenarios regarding the resulting mixture was stirred for an additional 30 min and
50% spin state stabilisation. Firstly, it may be that the length then allowed to warm up gradually to room temperature and
scale of any spin-state ordering is below the observable limit, allowed to stand for 24 h. Finally, 40% HCl was cautiously
leading to extreme smearing of X-ray scattering associated with added keeping the temperature below 5 °C until a precipitate
these features; given the stoichiometry mismatch between appeared. The solution was cooled (refrigerated) overnight to
HS : LS fraction and inequivalent FeII sites (requiring, for completely precipitate the solid. The crude product was
example, ABAB-type ordering to impose itself upon an AABAAB washed with excess water, filter and recrystallize in hot
lattice) there appears a strong possibility that long-range order- ethanol. The bright-yellow needles were collected, dried in air
1
ing of any 50 : 50 HS : LS patterning is disrupted by low energy and store at 4 °C. H NMR (400 MHz, d-CDCl3, ppm): δ = 9.77
defects (for example, of type ABAABAB). Such defects could (d, 1H), 8.10–8.04 (dd, 1H), 7.99–7.56 (m, 4H), 6.65–6.57 (q,
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1434–1442 | 1439

View Article Online
Paper
Dalton Transactions
1H). 13C NMR: δ 193.31, 147.48, 134.01, 132.86, 131.33, 129.27, borosilicate capillary tube. Thermogravimetric analysis was
125.42. IR (cm−1): 1678 (s), 1569 (w), 1504 (s), 1435 (w), 1340 carried out on
a Mettler-Toledo TGA/SDT851e analyser
(m), 1312 (w), 1119 (s), 975 (m), 863 (w), 828 (w), 786 (m), 734 (30–600 °C at 1 °C min−1) using an alumina sample holder (N2
(s), 694 (m), 672 (m), 597 (w). ESI-MS (ESI+, m/z): calculated for flow of 10 mL min−1). Differential scanning calorimetry
[M + H]+ C9H7O3N 177.16, found 178.38. Melting point: measurements were performed using a Netzsch DSC 204
128.0 °C (lit. 127–129 °C).
instrument (100 to 250 K; rate of 10 K min−1; 20 mL min−1 N2
atmosphere). The low temperature was obtained with a liquid
nitrogen cryostat cooling system attached with an instrument.
The measurement was carried out with 12.2 mg of crystalline
sample sealed in an aluminium pan with a mechanical crimp
and an empty pan used as a reference. The data were analysed
with the Netzsch Proteus Analysis 2019. The Hirshfeld surface
of the asymmetric unit was generated by crystal Explorer 17.5
software which comprised 3D dnorm surface plots and 2D histo-
gram of de and di fingerprints plot.29,30 Single crystal CIFs
were used as input file for analysis.
Preparation of (E)-3-(2-nitrophenyl)acrylaldehyde (N-cintrz)
A general method was adopted to synthesize the Schiff base of
(E)-3-(2-nitrophenyl)acrylaldehyde. An equal amount of nitro-
cinnamaldehyde 1.77 g (10 mmol) and 4-amino 1,2,4-triazole
0.840 g (10 mmol) were dissolved in 30 mL of ethanol and
refluxed with a catalytic amount of conc. sulfuric acid (2–3
drops) at 80 °C for 5 h. The yellow precipitate was collected by
filtration. The crude product was washed with cold ethanol
and recrystallized from hot ethanol yielded yellow needle-
shaped microcrystals of N-cintrz (yield: 2.04 g, 83.8%). 1H
NMR (400 MHz, d6-DMSO, ppm): δ = 8.78 (s, 2H), 8.61–8.59 (d,
1H), 8.08–8.06 (ddd, 1H), 7.85–7.54 (m, 4H) 6.98–6.92 (q, 1H).
13C NMR: δ 195, 154.77, 149.69, 148.77, 147.75, 134.40, 132.61,
132.09, 129.78, 129.47, 125.27, 109.27. IR (cm−1): 1510 (s),
1340 (s), 1248 (w), 1171 (s), 1051 (s), 995 (w), 979 (w), 935 (w),
862 (m), 848 (w), 832 (w), 759 (w), 731 (s), 695 (w), 614 (s), 516
(m). Elemental analysis calculated for C11H9N5O2 (%): C 54.32;
N 28.79; H 3.73; found (%): C 54.13; N 29.28; H 3.56. ESI-MS
(ESI+, m/z): calculated for [M + H]+ C11H9N5O2 243.22, Found
244.08. Melting point: 230.5 °C.
Magnetic susceptibility
Variable temperature magnetic susceptibility data were col-
lected using a Quantum Design Versalab equipped with a
vibrating sample mount (VSM) under a constant applied mag-
netic field of 3000 Oe. The crystalline sample was retained in a
small amount of mother liquor and sealed in a polyethylene
tube design for measurement to present solvent loss. The data
were collected continuously (rate of 2 K min−1, 300–50–300 K).
Single crystal X-ray diffraction
Synthesis of [Fe3(N-cintrz)6(Pd(CN)4)3]·6H2O
A single crystal was mounted on a cryoloop with the paratone-
N oil. Data collection was carried out on a Bruker D8 Quest
diffractometer equipped with a graphite monochromatic Mo-
Kα (λ = 0.71073 Å) X-ray generator and PHOTON II CPAD detec-
tor. The crystal was positioned at 40 mm from the detector
and the spots were measured using 10 s exposure time. The
data were collected on a 0.152 × 0.75 × 0.025 mm3 red block
crystal at 250 K, 150 K and 100 K. An Oxford cryo-stream con-
troller was used. The APEX3 crystallographic software
(v2018.1.0, Bruker AXS Inc.) was used for data integration and
scaling and XPREP software was employed for space-group
determination.36,37 All structures were solved by the direct
method using ShelXT and refined on F2 by full-matrix least-
square with ShelXL using the X-Seed 4000 and OLEX2
interfaces.38–41 All non-hydrogen atoms were refined with an-
A bulk crystalline sample was prepared by vial-in-vial slow
diffusion methods. A large vial was prepared containing a
mixture of N-cintrz (38.9 mg, 0.16 mmol) and K2[Pd(CN)4]
(23.0 mg, 0.08 mmol) and small vial was prepared containing
Fe(ClO4)2·xH2O (20.1 mg, 0.08 mmol). The small vial was
inserted carefully into a large vial and then both vials were
filled with a mixture of an EtOH : H2O (50 : 50) solution. Over a
period of 8 weeks red square-shaped crystals were formed.
Elemental anal. calcd for FePdC26H18N14O4·3H2O (%): C:
38.67; H: 2.97; N: 24.29. Experimental (%): C: 38.60; H: 2.50;
N: 25.54. IR bands (cm−1): 2170(m), 1631(w), 1513(s), 1338(s),
1175(s), 1053(s), 978(m), 864(s), 620(s), 504(w).
Characterisation
Elemental analysis was performed at the Mark Wainwright isotropic thermal parameters and hydrogen atom in the
analytical centre (UNSW) or at The Campbell Microanalytical ligands were placed at idealized positions and refined using a
Laboratory (University of Otago). Fourier transformation (FTIR) riding model with isotropic thermal parameters, except the
were recorded on a Thermo-Fisher Nicolet-iS5 spectrometer in hydrogen atoms of the disordered water molecules (O9A and
the range of 400–4000 cm−1 which was equipped with a smart O9B). Details of the crystallographic data collection and refine-
iD7 diamond attenuated total reflectance (ATR) window. NMR ment parameters are summarised in Table S1,† selected struc-
spectra were recorded on a Bruker 400 MHz NMR equipped tural parameters are given in Table S2,† and hydrogen
with an autosampler in d6-DMSO as a solvent. Mass spec- bonding interactions are given in Table S3.† The crystallo-
troscopy was performed on a Thermo-Fisher mass spectro- graphic information files (CIF) have been deposited in the
meter (m/z 50–2000) under the positive ion mode with the ESI Cambridge Crystallographic Database (CCDC) and are freely
source using methanol as eluent. Melting points were deter- available
under
the
CCDC
reference
numbers:
mined using an OptiMelt (SRS MPA 100) apparatus in an open 2045470–2045472.†
1440 | Dalton Trans., 2021, 50, 1434–1442
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Powder diffraction using synchrotron radiation
Conflicts of interest
Variable temperature X-ray powder diffraction data were col-
lected on the Powder Diffraction beamline (BL-10; 20.99 keV
and 0.5903 Å) at the Australian Synchrotron.42 A polycrystalline
sample was ground gently into a slurry using a mortar and
There are no conflicts to declare.
pestle, packed into borosilicate glass capillary tube (0.5 mm Acknowledgements
diameter) and mounted onto the powder diffraction beamline.
JKC, BJP, CJK and SMN acknowledge support from Fellowships
and Discovery Project funding from the Australian Research
Council (ARC). Access and use of the facilities of the
Australian Synchrotron was supported by ANSTO. JKC acknowl-
edges the support of the ARC through LE170100144.
The capillary was rotated at ca. 1 Hz during data collection to
aide powder averaging. The wavelength was determined accu-
rately using NIST SRM LaB6 660b standard. Temperature-
dependent data were collected upon cooling and warming in
steps of 120 K h−1 from 100–280–100 K. Data were collected
using the Mythen microstrip detector43 from 1.5 to 75° in 2θ.
To cover the gaps between detector modules, two data sets,
each of 60 seconds in duration, were collected with the detec- References
tor set 5° apart and these were then merged to give a single
1 P. Gütlich, H. A. Goodwin and Y. Garcia, Spin crossover in
data set. A slit size of 2 mm was used, to ensure that the frac-
tion of the capillary illuminated by the X-ray beam is the same
as the isothermal zone on the cryostream. The obtained raw
data were merged using in-house software PDViPER and
further process with the Surfer software package. Le-Bail ana-
lysis and peak fitting were performed using TOPAS44 software
package.
transition metal compounds I, Top. Curr. Chem, 2004.
2 P. Gütlich, A. B. Gaspar and Y. Garcia, Beilstein J. Org.
Chem., 2013, 9, 342–391.
3 M. A. Halcrow, Spin-crossover materials: properties and appli-
cations, John Wiley & Sons, 2013.
4 H.-Z. Ye, C. Sun and H. Jiang, Phys. Chem. Chem. Phys.,
2015, 17, 6801–6808.
5 M. Nishino, K. Boukheddaden, Y. Konishi and
S. Miyashita, Phys. Rev. Lett., 2007, 98, 247203.
6 M. Nishino, K. Boukheddaden and S. Miyashita, Phys. Rev.
B: Condens. Matter Mater. Phys., 2009, 79, 012409.
7 C. Enachescu, M. Nishino, S. Miyashita, L. Stoleriu and
A. Stancu, Phys. Rev. B: Condens. Matter Mater. Phys., 2012,
86, 054114.
8 J. Wajnflasz and R. Pick, J. Phys. Colloq., 1971, 32, C1-91–
C1-92.
9 G. Ruzzi, J. Cruddas, R. H. McKenzie and B. J. Powell, 2020,
arXiv:2008.08738.
Conclusions
The targeted incorporation of an organic ligand with two
modes of supramolecular contacts (1,2,4-triazole and nitro-
groups) has led to a 2-D Hofmann framework which exhibits
characteristics of extreme elastic frustration. Highlighting
this are the two distinct FeII sites in a rare 2 : 1 ratio, an
array of ligand conformations and layer distortion such that
interlayer contacts are dominated by hydrogen-bonding
rather than aromatic contacts observed in classic Hofmann 10 M. Paez-Espejo, M. Sy and K. Boukheddaden, J. Am. Chem.
frameworks. There are many ligand–ligand and ligand–water Soc., 2016, 138, 3202–3210.
interactions involving the 1,2,4-triazole and nitro-groups 11 J. Cruddas and B. J. Powell, Inorg. Chem. Front., 2020, 7,
resulting in a structural complexity where all the interactions 4424–4437.
cannot be simultaneously minimised, resulting in elastic 12 J. Cruddas and B. J. Powell, J. Am. Chem. Soc., 2019, 141,
competition between ferro- and anti-ferro elastic interactions. 19790–19799.
The resultant temperature-induced SCO behaviour is multi- 13 O. Kahn and M. C. Jay, Science, 1998, 279, 44–48.
stepped, supporting the expanding body of theoretical and 14 K. S. Murray and C. J. Kepert, Top. Curr. Chem., 2004, 233,
experimental evidence showing that elastically frustrated
structures drive and support multi-stepped transitions and 15 M. C. Muñoz and J. A. Real, Coord. Chem. Rev., 2011, 255,
hence mixed spin-state species. Of additional importance in 206–2093.
this study is the finding that the elastic competition, which 16 Z.-P. Ni, J.-L. Liu, N. Hoque, W. Liu, J.-Y. Li, Y.-C. Chen and
includes 3-D intermolecular communication channels, over- M.-L. Tong, Coord. Chem. Rev., 2017, 335, 28–43.
rides the periodic ordering of mixed spin-state species at 17 Y. M. Klein, N. F. Sciortino, F. Ragon, C. E. Housecroft,
195–228.
intermediate temperatures leading to a disordered intermedi-
ate phase. The disparity between structural and spin-state
C. J. Kepert and S. M. Neville, Chem. Commun., 2014, 50,
3838–3840.
periodicity in this structure, which we hypothesise underlies 18 E. Milin, V. Patinec, S. Triki, E.-F. Bendeif, S. Pillet,
the absence of long-range spin-state ordering, is a new
feature of multi-stepped transition materials, hence future
M. Marchivie, G. Chastanet and K. Boukheddaden, Inorg.
Chem., 2016, 55(22), 11652–11661.
work on this system will include detailed calculations that 19 N. F. Sciortino, F. Ragon, K. A. Zenere, P. D. Southon,
test this premise.
G. J. Halder, K. W. Chapman, L. Pineiro-Lopez, J. A. Real,
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1434–1442 | 1441

View Article Online
Paper
Dalton Transactions
C. J. Kepert and S. M. Neville, Inorg. Chem., 2016, 55, 29 C. F. Mackenzie, P. R. Spackman, D. Jaytilaka and
10490–10498. M. A. Spackman, IUCrJ, 2017, 4(5), 575–587.
20 K. A. Zenere, S. G. Duyker, E. Trzop, E. Collet, B. Chan, 30 M. A. Spackman and J. J. McKinnon, CrystEngComm, 2002,
P. Doheny, C. J. Kepert and S. M. Neville, Chem. Sci., 2018,
9, 5623–5629.
4, 378–392.
31 J. Cruddas and B. J. Powell, 2020, arXiv:2007.13983.
21 M. J. Murphy, K. A. Zenere, F. Ragon, P. D. Southon, 32 B. Weber, C. Carbonera, C. Desplances and J.-F. Létard,
C. J. Kepert and S. M. Neville, J. Am. Chem. Soc., 2017, 139,
1330–1335.
22 N. F. Sciortino, K. A. Zenere, M. E. Corrigan, G. J. Halder,
Eur. J. Inorg. Chem., 2008, 10, 1589–1598.
33 E. Collet and P. Guionneau, C. R. Chim., 2018, 21, 1133–
1151.
G. Chastanet, L.-F. Létard, C. J. Kepert and S. M. Neville, 34 B. J. Powell, Contemp. Phys., 2020, 61, 96–131.
Chem. Sci., 2017, 8, 701–707. 35 R. E. Buckles and M. P. Bellis, Org. Synth., 1963, 4, 60.
23 L. Fu-Ling and J. Tao, Chem. – Eur. J., 2017, 23, 18252– 36 APEX3 Crystallography Software Suite, Bruker Inc., Madison,
18257. USA, 2018.
24 Y. Meng, Q.-Q. Sheng, M. N. Hoque, Y.-C. Chen, S.-C. Wu, 37 SADABS, University of Göttingen, 1996.
J. Tucek, R. Zboril, T. Liu, Z.-P. Ni and M.-L. Tong, Chem. – 38 G. Muller-Buschbaum, Acta Crystallogr., Sect. C: Struct.
Eur. J., 2017, 23, 10034–10037.
Chem., 2015, 71, 3–8.
25 W. Liu, Y.-Y. Peng, S.-G. Wu, Y.-C. Chen, M. N. Hoque, 39 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. Howard
Z.-P. Ni, X.-M. Chen and M.-L. Tong, Angew. Chem., Int. Ed.,
2017, 56, 14982–14986.
and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339–341.
40 L. J. Barbour, J. Appl. Crystallogr., 2020, 53(4), 1141–1146.
26 N. F. Sciortino, F. Ragon, Y. M. Klein, C. E. Housecroft, 41 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Adv., 2015,
C. G. Davies, G. N. L. Jameson, G. Chastanet and
S. M. Neville, Inorg. Chem., 2018, 57, 11068–11076.
27 F.-L. Liu, D. Li, L.-J. Su and J. Tao, Dalton Trans., 2018, 47,
1407–1411.
28 A. T. Brennan, K. A. Zenere, H. E. A. Brand, J. R. Price,
M. M. Bhadbhade, G. F. Turner, S. A. Moggach,
71, 3–8.
42 K. S. Wallwork, B. J. Kennedy and D. Wang, AIP Conf. Proc.,
2007, 879–882.
43 B. Schmitt, C. Brönnimann, E. Eikenberry, F. Gozzo,
C. Hörmann, R. Horisberger and B. Patterson, Nucl.
Instrum. Methods Phys. Res., Sect. A, 2003, 501, 267–272.
F. J. Valverde-Muñoz, J. A. Real, J. K. Clegg, C. J. Kepert and 44 A. Coelho, TOPAS, Version 4.1, Bruker AXS GmbH,
S. M. Neville, Inorg. Chem., 2020, 59(19), 14296–14305. Karlsruhe, Germany, 2007.
1442 | Dalton Trans., 2021, 50, 1434–1442
This journal is © The Royal Society of Chemistry 2021