Directing self-assembly in solution towards improved cooperativity in Fe(iii) complexes with amphiphilic tridentate ligands

Article abstract of DOI:10.1039/c9dt00032a

An amphiphilic iron(iii) complex with a tridentate Schiff-base ligand was prepared by condensation of a hexadecyloxy functionalised salycylaldehyde with a diamine followed by complexation with FeCl₂ and anion methathesis with NaClO₄.

Full text of DOI:10.1039/c9dt00032a

Dalton
Transactions
View Article Online
View Journal
PAPER
Directing self-assembly in solution towards
improved cooperativity in Fe(III) complexes with
Cite this: DOI: 10.1039/c9dt00032a
amphiphilic tridentate ligands†
b,d
Ana I. Vicente, a,b Xinwey Wu,a Yannick Ortin, c Liliana P. Ferreira,
Maria de Deus Carvalho, a Sara Realista, a,b Andrew Barker,c Grace G. Morgan,c
Nuno Galamba, a,b Paulo J. Costa, a,b Maria José Calhorda a,b and
a,b,e
Paulo N. Martinho
*
An amphiphilic iron(III) complex with a tridentate Schiff-base ligand was prepared by condensation of a
hexadecyloxy functionalised salycylaldehyde with a diamine followed by complexation with FeCl2 and
anion methathesis with NaClO4. The complex shows spin crossover both in the solid state and solution.
However in solution self-assembly and consequently aggregation of individual molecules form concen-
tration dependent particles with sizes of 300 nm for higher concentrations, or 5 nm for lower concen-
trations. Aggregate formation was confirmed by NANO-flex 180° DLS Size, scan-rate dependent cyclic
voltammetry and scanning electron microscopy. Molecular simulations were used to investigate the self-
assembly of the complex in solution, including the role of residual water molecules. The simulations
showed the self-assembly of reverse micelle-like structures when a small water cluster is inserted in solu-
tion, whereas no large aggregates formed in dehydrated environments. The perchlorate anions were
found near the metal centres, stabilizing the aggregates around the water pool. Simulations of pre-
assembled structures further showed the lack of stability of large aggregates in the absence of water. The
larger aggregates promoted efficient communication between the iron(III) centres and the compound dis-
played spin crossover in solution at around 220 K with a 10 K hysteresis window, as measured by NMR
and SQUID magnetometry.
Received 3rd January 2019,
Accepted 27th February 2019
DOI: 10.1039/c9dt00032a
rsc.li/dalton
the counterion6–9 and the crystallisation solvents.10 Another
property that is known to influence solid state spin state switch-
Introduction
Spin crossover (SCO) is still a growing and important research ing is polymorphism. The switching profile can be altered or
topic, spanning topics as diverse as the role of metal ions in completely switched off in different polymorphs of the same
biology1 and magnetic devices.2,3 The phenomenon, mostly SCO compound.11,12 These constraints hinder the application
observed in the solid state, is normally correlated with intra of these compounds at the level of reliable device assembly.
and intermolecular interactions such as hydrogen-bonding, SCO in solution is less commonly investigated but overcomes
π–π stacking, π–anion or dipolar interactions, and their influ- some of the disadvantages presented above.13 Examples of SCO
ence on cooperative effects displayed in crystalline materials.4,5 in solution are scarce and recent literature includes bimetallic
When in the solid state, this phenomenon has to take into systems such as a pyrazolate bridged binuclear iron(II) complex
account factors such as sensitivity to the lattice contents, i.e. with two pincer-type {PNN} subunits showing spin state switch-
ing in solution through reversible ligand exchange,14 and the
heterometallic compound [{(Tp*)Fe(CN)3}2{Co(bpyR)2}2][OTf]2,
which shows thermally induced electron transfer control in
aCentro de Química e Bioquímica, Faculdade de Ciências, Universidade de Lisboa
Campo Grande, 1749-016 Lisboa, Portugal. E-mail: pnmartinho@ciencias.ulisboa.pt
solution.15 Spin transitions in solution are usually character-
ised by a gradual change in spin state population and follow a
Boltzmann distribution which is represented by a spin state
equilibrium and absence of cooperativity.16–18
bBiosystems and Integrative Sciences Institute (BioISI), Faculdade de Ciências,
Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal
cSchool of Chemistry, University College Dublin, Belfield, D4, Dublin, Ireland
dPhysics Department, University of Coimbra, 3004-516 Coimbra, Portugal
eDepartment of Inorganic Chemistry, Faculty of Science, Charles University, 128 43
Prague 2, Czech Republic
Mononuclear systems studied in solution include Co(II)19
and Fe(II)19 in [(tmpa)M(HL)](BF4) complexes with tmpa = tris
(2-methylpyridyl)amine, complexes employing hexadentate
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c9dt00032a
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Scheme 1 Synthesis of [Fe(OC16-SalEen)2]ClO4 (1).
The behaviour of 1 in dichloromethane at different concen-
trations was studied at room temperature by UV-vis spec-
troscopy and cyclic voltammetry (CV). The UV-vis spectra
were recorded at three different concentrations (2.2 × 10−4
,
ligands with cis/trans-1,2-diamino cyclohexanes as central
building blocks,20 and [Fe(H2Bpz2)2(bipy-NH2)] complexes that
exhibit ambient-temperature spin-state switching by protona-
tion of the amino group.21 Apart from these, other important
studies have also been made in solution for Fe(II) systems.12,18,22
In solution, however the molecule is not confined to a rigid
lattice environment and with suitable building blocks
the soft medium can promote the formation of locally
aggregated systems, which can confer cooperativity between
molecules.10–12,23 To induce self-assembly in solution, and
possible communication between metal centres through aggre-
gation of metallic centres, the synthesis of compounds functio-
nalised with long alkyl chains has been explored.24–26 In par-
ticular, we studied alkyl-functionalized iron(III) sal2trien
complexes23,27–29 and observed that one of these shows a 5 K
hysteresis in dichloromethane.27
Following our interest in tridentate ligands in Fe(III)
systems30,31 and our successful strategy of appending alkyloxy
groups onto the phenolate moieties of hexadentate
ligands,28,32,33 we decided to apply the same principle to design
an amphiphilic tridentate ligand. Thus, here we describe the
synthesis and characterisation of the amphiphilic compound [Fe
(OC16-SalEen)2]ClO4 (1) both in the solid state and in solution.
The magnetic properties of 1 in the solid state reveal a gradual
and incomplete SCO. However, in solution 1 exhibits a sharp
and hysteretic SCO, with a wider hysteresis window (10 K) and
higher T1/2 than previously reported in hexadentate analogues.23
4.4 × 10−5 and 4.4 × 10−6 M) and are shown in Fig. 1.
The UV-vis spectra for 1 at all concentrations display an
absorption band in the visible at 524 nm and several more
intense bands at higher energies. We performed time-depen-
dent density functional theory (TDDFT) calculations to assign
the more relevant bands and found, for the HS complex, one
band with two components, at 525 and 531 nm, respectively,
very close to the experimental value, which was assigned to a
ligand (phenolate)-to-metal transition (LMCT). This results is
in agreement with previous studies.28,34 The absorption in the
visible of the LS species has a similar nature, but lower energy
(639 nm).
The cyclic voltammograms were also recorded at three
different concentrations (4.4 × 10−5, 1 × 10−3 and 4.9 × 10−3 M).
However, at the lowest concentration, the cyclic voltammogram
is dominated by noise and is not shown in Fig. 2. Ohmic drop
was compensated for all experiments using a positive feedback
(>700 Ω) procedure implemented in the instrument.
The CVs show the FeIII/FeII reduction process, which is
quasi-reversible according to scan rate experiments (Fig. S1†).
Table 1 shows the electrochemical data for the different con-
centrations of 1. The potential for the FeIII/FeII reduction is
−491 and −526 mV vs. SCE for the 1 × 10−3 and 4.9 × 10−3
M
solutions, respectively. Additionally, the peak separations are
different for both concentrations, a value closer to a reversible
Results and discussion
Synthesis and characterisation
Complex 1 was prepared in a mixture of dichloromethane and
methanol by condensation of N-ethylethylenediamine (Een)
with 4-(hexadecyloxy)-2-hydroxybenzaldehyde, followed by reac-
tion with Fe(II) chloride and sodium perchlorate, yielding the
Fe(III) microcrystalline complex, after air oxidation and slow
evaporation of the solvent, Scheme 1.
The infrared spectrum of 1 shows a characteristic νCvN
stretching band, at 1607 cm−1, while the bands at 1087 and
1066 cm−1 were assigned to the stretching frequencies of the
ClO4− anion.
Fig. 1 Room temperature UV-vis spectra of 1 in dichloromethane at
three different concentrations.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
Fig. 2 Cyclic voltammograms of 1 in dichloromethane at two different
concentrations (1 × 10−3 and 4.9 × 10−3 M) at 0.1 V s−1. Pt was used as
working electrode (2 mm diameter), Pt foil and SCE as counter and
reference electrodes, respectively. NBu4PF6 was used as supporting
electrolyte.
Fig. 3 Variable temperature UV-vis spectra of [Fe(OC16-SalEen)2]ClO4
in a 4.4 × 10−5 M CH2Cl2 solution (cooling mode).
Table 1 Electrochemical data for 1 using different concentration (1 ×
10−3 and 4.9 × 10−3 M solutions)a
The UV-vis data for 1 were recorded between 193 and 298 K
both in the cooling (Fig. 3) and heating modes (Fig. S2–S4†).
The cooling mode spectra reveal isosbestic points at 596 and
398 nm confirming a spin equilibrium between the HS and LS
states. The same data show a small increase of the absorbance
at 670 nm and a proportional decrease in absorbance at
524 nm, which are indicative of LS population increase and
subsequent HS population decrease. It is worth mentioning
that although the absorbance of the HS band decreases, the
band is never fully converted into the LS band, thus indicating
that at this concentration the SCO is rather gradual and incom-
plete. The heating mode variable temperature UV-vis spectra
for 1 are not as well-behaved as the former. The LS band at
673 nm decreases proportionally with the temperature
increase. Fig S5† shows the variation of the extinction coeffi-
cients with temperature for the three different solutions at
640 nm confirming that the solution behaves better at the
Concentration/×10−3
M
E(FeIII/FeII)/mV vs. SCE
ΔEp/mV
1.0
4.9
−491
−526
59
115
a Dichloromethane solution, NBu4PF6 as supporting electrolyte. Pt was
used as working electrode (1 mm diameter), Pt foil and SCE as counter
and reference electrodes, respectively. ν = 100 mV s−1
.
process being observed for the 1 × 10−3 M solution (59 mV).
However, the peak separation value cannot be fully trusted, as
species adsorbed on the electrode surface upon reduction and
a thin film was deposited on the platinum working electrode
surface after performing the CVs.
Spin crossover in solution
The magnetic behaviour of 1 in solution was investigated at intermediate concentration.
several temperatures by UV-vis spectroscopy for three different
The magnetic behaviour of 1 was further studied by the
dilutions, 4.4 × 10−6, 4.4 × 10−5 and 2.2 × 10−4 M. The UV-vis Evan’s method35,36 using a 4.9 × 10−3 M deuterated dichloro-
spectra of the amphiphilic compound 1 for the three concen- methane (CD2Cl2) solution. Attempts to apply this method to
trations at 298 K exhibit a maximum absorbance at around solutions with concentrations below 1 × 10−3 M proved to be
524 nm (Fig. S2†), assigned by TDDFT to a LMCT absorption unsuccessful due to instrumental limitations. Variable temp-
1
characteristic of the HS compound. A new absorption band erature H NMR of 1 with a concentration of 4.9 × 10−3 M in
emerging at around 673 nm on cooling is indicative of a SCO CD2Cl2 with 0.03% TMS were recorded against a reference of
and accompanies the consequent increase of the percentage of CD2Cl2 with 0.03% TMS and the results of one cooling/heating
the LS component and the decrease of the HS band at 524 nm, cycle are shown in Fig. 4 (top).
Fig. S2.† The TDDFT calculated absorption spectrum of the LS
The χMT vs. T plot for 1 (Fig. 4 (top)) reveals an incomplete
form of the complex exhibits an absorption at ∼640 nm SCO with an approximately 10 K hysteresis loop in solution.
assigned to a ligand (phenolate)-to-metal transition. The The sharp χMT variation and the hysteresis observed are both
observed change reflects a very subtle thermochromic behav- indicative of the formation of a cooperative system in solution.
iour with a colour change from dark to light purple, thus This type of cooperativity was previously observed for similar
difficult to use as diagnostic for SCO in solution at these con- systems with hexadentate ligands.23,27–29
centrations. The difference in absorption between the HS and
At 200 K the magnetic plot shows a χMT of 0.6 cm3 mol−1 K,
the LS bands is more pronounced for the intermediate concen- corresponding to a mainly LS system, and on warming above
tration, 4.4 × 10−5 M and, therefore, we shall focus the discus- 250 K the χMT value reaches a value of 2.3 cm3 mol−1 K, i.e.
sion on this concentration, Fig. 3, S2, S3, S4 and S5.†
incomplete SCO transition (for Fe(III) HS, χMT = 4.38 cm3
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 5 χMT thermal variation of [Fe(OC16-SalEen)2]ClO4 (1): in the solid
state – black squares (field cooling (FC)) and red circles (field warming
(FW)); in the frozen 4.9 × 10−3 M CD2Cl2 solution (olive stars (FC) and
green diamonds (FW)).
The obtained spectra are shown (Fig. S6 and Table S1†) and
discussed in the ESI† supporting the assignments proposed
from the SQUID data.
Two series of SQUID measurements were performed for
solutions of 1, above the freezing temperature of the solvent,
using the same concentration (5 × 10−3 M CD2Cl2) in closed
quartz tubes, to prevent CD2Cl2 evaporation. For both series,
the samples were measured between 290 K and 185 K, in 5 K
steps, during cooling and warming at a rate of 5 K min−1. For
Fig. 4 Variable temperature χMT of [Fe(OC16-SalEen)2]ClO4 (1) in a 4.9 × the first series of measurements, three consecutive cycles were
10−3 M CD2Cl2 solution: (top) by 1H NMR using the Evan’s method;
(bottom) by SQUID magnetometry. Cycle 2 of the second series (blue
left triangles (FC) and orange right triangles (FW)). Cycle 5 of the second
series (cyan up triangles (FC) and magenta down triangles (FW)).
performed, and each data point was collected immediately
after temperature stabilization. For the second series, five con-
secutive cycles were performed, and each magnetisation
measurement was performed 10 minutes after temperature
stabilisation, as per the NMR experiments. All measured
thermal magnetisation cycles are very similar and reproduci-
mol−1 K). In the cooling mode the apparent T1/2 = 213 K and ble, and the difference between the two series of measure-
ments was only in the lower data dispersion observed for the
second series, which is understandable given the 10 minutes
of further waiting after thermal normalization. Also shown in
Fig. 4 (bottom) are the results of two thermal cycles on the
second set of SQUID experiments: cycle 2 and cycle 5. Two fea-
in warming the apparent T1/2 = 223 K, resulting in a 10 K hys-
teresis window. To the best of our knowledge, this is the
widest hysteresis loop attained to date in solution.
SQUID measurements of 1 were performed both in the
solid state and in solution (5 × 10−3 M CD2Cl2). The χMT versus
T data for a powder sample of 1 show a gradual and incom- tures can be underlined: (1) confirmation that the state
plete SCO profile (Fig. 5), with values of χMT = 2.9 cm3 mol−1
quenched in the frozen solution and measured below 150 K
K, at 300 K (mostly in the HS state), and χMT around 1.5 cm3
indeed corresponds to the high temperature χMT value of the
mol−1 K at about 100 K (predominantly LS). Gradual and liquid solution; (2) confirmation of the hysteretic magnetic
behaviour very similar to that revealed by the NMR experi-
ments, in the SQUID data this is evident between 210 K and
235 K, with a 10 K window.
incomplete spin crossover in the solid state for amphiphilic Fe
(III) complexes has been previously reported for a number of
complexes.27,29
A 5 × 10−3 M CD2Cl2 solution of 1 was then prepared, intro-
duced in a quartz tube, and quickly inserted into the SQUID
Aggregation studies
magnetometer at 150 K, i.e., at the temperature at which the Previous studies show that the introduction of alkyl chains in
solvent would freeze, and the solution would form a glass. The Schiff-base ligands can promote the formation of aggregates in
susceptibility of the frozen glass was measured in cooling solution.27–29 To confirm the interaction between molecules of
mode between 150 K and 10 K and over this temperature range 1 in solution and consequently the self-assembly and for-
the χMT value remains unchanged at 1.2 cm3 mol−1 K (Fig. 5).
mation of aggregates in dichloromethane, NANO-flex 180° DLS
Both the solid state sample and the frozen solution of 1 Size assays at the UV-vis and NMR concentrations, 4.4 × 10−5
were also analysed by 57Fe Mössbauer spectroscopy at 78 K. M and 4.9 × 10−3 M, respectively, were carried out and the
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
presence of particles confirmed by scanning electron obtained for the UV-vis concentration reveal spherical particles
microscopy (SEM). Interestingly the two different concen- with sizes below 100 nm.
trations gave different particle size in solution, Table 2, the
size being thus concentration dependent. The polydispersive
Molecular dynamics simulations
particles formed in a 4.9 × 10−3 M CD2Cl2 solution (Fig. S7†)
The self-assembly of complex structures like micelles and
reverse micelles has long been object of molecular simulation
studies.38–43 In this study, the self-assembly of a reverse
micelle-like aggregate of [Fe(OC16-SalEen)2]+ around a water
pool formed by nearly 100 molecules (some water molecules
diffused into the solvent), was observed (see Fig. 7a and b). A
similar number of perchlorate anions are distributed near the
metal centres and the water pool, whereas the other molecules
of the complex remain either isolated or forming short-lived
head-to-head dimers. Additional simulations showed a similar
behaviour with the number of molecules of complex varying
between 5 and 7; this is closely associated with the diffusion of
water molecules away from the initial cluster. Furthermore,
similar aggregates were observed at 230 K although longer
simulations were required to observe aggregation. Aggregation
was not observed at either temperature on a dehydrated solu-
tion. This is consistent with the simulations starting from pre-
assembled aggregates, where disaggregation occurred on a fast
time-scale, and with the non-aggregation of the complex mole-
measure about 300 nm, and the polydispersive particles
formed in a CH2Cl2 4.4 × 10−5 M solution (Fig. S8†) around
5 nm (Table 2).
The differences observed for the electrochemical data in
Table 1 suggest that the solution composition depends on the
concentration used. Therefore, chronoamperometric experi-
ments (Fig. S9†) were performed to calculate the diffusion
coefficient for 1 × 10−3 and 4.9 × 10−3 M solutions. Using the
Cottrell equation, the diffusion coefficients determined are
inversely proportional to the particle radius.37 The values
obtained were 3.9 × 10−5 and 2.9 × 10−6 cm2 s−1 for the lowest
and highest concentration, respectively. The smallest value of
the diffusion coefficient (2.9 × 10−6 cm2 s−1) indicates that
larger particles are present at the highest concentration
(4.9 × 10−3 M).
The particle size and shape were imaged by SEM, and are
shown in Fig. 6 and S10,† for the 4.9 × 10−3 M solution, and in
Fig. S11† for the 4.4 × 10−5 M solution.
SEM images for the two concentrations indicate differences
in the aggregation. They suggest that self-assembly might
occur for the 4.9 × 10−3 M solution with the formation of nm-
sized self-assembled ribbons which clamp together to form
hundreds of nm-sized aggregates. It is important to note that
the images obtained after evaporation of the solvent might not
be a clear image of the solution environment. The images
Table 2 Particle size of solutions of [Fe(OC16-SalEen)2]ClO4 (1)
Concentration/mM
Solvent
Particle size/nm
4.9
CD2Cl2
CH2Cl2
300
5
4.4 × 10−2
Fig. 7 (a) MD box with 25 molecules of [Fe(OC16-SalEen)2]+ and 25 per-
chlorate anions randomly distributed and a water cluster with 100 mole-
cules, after 100 ps in the NVT ensemble and 0.5 ns in the NPT ensemble.
The dichloromethane molecules (>21 000) are not shown. The iron
atoms are shown as black spheres. The water molecules form a cluster
near the centre of the MD box. A similar box was used to study the self-
assembly of the complex in the absence of water. (b) The MD box after
50 ns in the NPT ensemble showing a reverse micelle-like aggregate
comprised of 7 molecules of complex with the metal centres pointed to
the water molecules and 7 perchlorate anions near the metal centres. (c)
MD box showing a pre-assembled reverse-micelle-like with 10 mole-
cules of [Fe(OC16-SalEen)2], 10 perchlorate anions, and a water cluster
with 100 molecules, after 100 ps in the NVT ensemble and 0.5 ns in the
NPT ensemble. (d) The MD box after 50 ns in the NPT ensemble showing
a reverse micelle-like aggregate comprised of 7 molecules of complex
and 7 perchlorate anions.
Fig. 6 SEM images of aggregates of a 4.9 × 10−3 M CD2Cl2 solution of 1
(magnification 30000×).
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
cules in dehydrated regions of the solution, depicted in 2,4-hydroxybenzaldehyde, potassium bicarbonate and solvents
Fig. 7b. MD of pre-assembled reverse micelle-like structures were purchased and used without further purification.
with a water core, in turn, showed that these aggregates are Compound 4-(hexadecyloxy)-2-hydroxybenzaldehyde, was syn-
stable (see Fig. 7c). Again, different simulations showed thesized as described elsewhere.23 IR spectra were recorded on
different numbers of complex molecules, ranging between 5 a PerkinElmer FTIR spectrophotometer. Microanalyses (C, H
and 8, remaining in the pre-assembled aggregate in the time- and N) were measured by elemental analysis service at the
frame of the simulations. The perchlorate anions in these pre- University of Vigo, Spain.
assembled structures must be distributed near the head
Synthesis of compound 1 [Fe(OC16-SalEen)2]ClO4·H2O. To a
groups in the starting configuration to avoid disaggregation of solution of N-ethylethylenediamine (210 µL, 2 mmol) in
the solvated aggregates at room temperature and pressure. methanol (30 mL), 4-(hexadecyloxy)-2-hydroxybenzaldehyde
Thus, local neutrality is key for the stabilisation of the aggre- (402 mg, 2 mmol) was added and left stirring for 15 minutes
gates formed around the water pool. For the pre-assembled to give a yellow solution. A solution of iron(II) chloride
aggregates in the absence of water, disaggregation was (127 mg, 1 mmol) and sodium perchlorate (122 mg, 1 mmol)
observed for any starting aggregate size, regardless of the posi- in methanol was added to the mixture and left stirring for 1 h.
tion of the perchlorate anions in the initial configuration.
A precipitate was obtained and filtered to obtain a magenta
These results suggest that the aggregation process and the solid (95 mg, 20%). IR (KBr): νmax/cm−1 3249 (νNH, m), 1607
size of the aggregates depends on the water content in solu- (νCvN, s), 1532 (δCvC, m), 1264 (νC–N, s), 1122 (νClO4, s), 1087
tion, implying that the increase of the size with the concen- (νClO4, s). HR-MS (ESI): Calcd for C54H94FeN4O4 [M − ClO4]+
tration, assessed experimentally, although at much lower con- m/z = 918.66219, found m/z = 918.66202. Anal. calcd for
centrations than those that can be simulated, should be C54H96ClFeN4O9: C, 62.56; H, 9.33; N, 5.40. Found: C, 62.49; H,
associated to an increase of the number of crystallisation 9.23; N, 5.00.
waters in solution, at higher concentrations.
Magnetic susceptibility was recorded on a SQUID magnet-
ometer (Quantum Design MPMS). Data was obtained at 1000
Oe for temperatures ranging from 10 to 370 K and the molar
susceptibilities (χM) values were corrected for diamagnetism.
UV-vis measurements. UV-vis absorption spectra were
Conclusions
The amphiphilic compound [Fe(OC16-SalEen)2]ClO4 (1) displays recorded using a Shimadzu UV-1800 spectrometer with an
a gradual and incomplete SCO in the solid state, while in solu- Oxford Instruments Opistat DN2 cryostat linked to a tempera-
tion it behaves as a cooperative system, thus displaying a SCO ture control unit for temperature dependent measurements.
with a 10 K hysteresis window in deuterated dichloromethane.
NMR. Magnetic measurements of complex 1 in CD2Cl2 with
1
This behaviour is for the first time reported in a tridentate a concentration of 4.9 mM were performed by H NMR using
Schiff-base Fe(III) compound with wider hysteresis and higher the Evans’ method35,36 on a Bruker Avance 400 spectrometer
SCO temperature compared to our previously reported hexa- operating at 400.14 MHz at several temperatures in the temp-
dentate Fe(III) SCO amphiphilic compound.23 The SCO in solu- erature range of 190–300 K. The measurements were per-
tion has been shown to be concentration dependent by formed in standard 5 mm NMR tubes containing the paramag-
SQUID, NMR and UV-vis characterization and the different netic samples dissolved in CD2Cl2 with an inert reference of
SCO behaviours in solution are attributed to self-assembly and 0.03% TMS, against a reference insert tube filled with the
particle formation in dichloromethane.
same solvent (0.03% TMS in CD2Cl2) and their shift measured
NANO-flex 180° DLS Size measurements, complemented by in Hz.
scan-rate dependent cyclic voltammetry and scanning electron
Particle size measurements. The particle sizes of the
microscopy showed 300 nm particles for the higher concen- solutions of the complex at 4.9 × 10−3 M in CD2Cl2 and 4.4 ×
tration and 5 nm for the lower concentration. SEM images 10−5 M in CH2Cl2 were measured with a NANO-flex 180° DLS
show that although not reliable size-wise, molecules of 1 self- Size nanosizer from Particle Metrix.
assemble to form nano-ribbons, thus promoting the com-
FEG-SEM. A drop of the studied solutions was deposited on
munication between metal centres and therefore resulting in a an aluminium foil supported in a double-sided carbon tape
cooperative system. The formation and stability of the self- over the sample holder. All the samples were covered with a
assembled aggregates was further confirmed by molecular Au/Pd thin film (15–30 nm) in a sputter coater from Quorum
dynamics simulations emphasising the importance of water in Technologies, Q150 T ES model. The samples were analysed in
the aggregation process.
FEG-SEM from JEOL, JSM7001F model. The electron beam
tension accelerating voltage was 15 kV for all acquired images.
Electrochemical studies. Cyclic voltammetry and chronoam-
perometric studies were performed on an AUTOLAB PGSTAT
12 AUT71019 controlled by NOVA 2.0 software. A three-elec-
trode electrochemical cell with the reference electrode separ-
Experimental
Materials and methods
Synthetic procedures. N-Ethylethylenediamine, sodium per- ated by a glass frit was employed. Platinum (2 mm diameter,
chlorate monohydrate, iron(II) chloride, bromohexadecane, CHI instruments) and platinum wire were used as working
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
and counter electrodes, respectively, and a saturated calomel NPT ensemble for 50 ns; three trajectories, starting from
was used as reference electrode (SCE = 0.244 V vs. NHE). different initial velocities, were propagated for each system for
Previously distilled dichloromethane was used as solvent for comparison purposes. The T and p were controlled with the
solutions 4.4 × 10−5, 1.0 × 10−3 and 4.9 × 10−3 M of 1. Ohmic thermostat of Bussi et al.57 and the Parrinello–Rahman58 baro-
drop was compensated by a positive feedback present in the stat. Electrostatic interactions were computed with the par-
instrument. The working electrode was polished before each ticle-mesh Ewald59 (PME) method. A cut-off of 1 nm was used
experiment using 1, 0.5 and 0.03 μm alumina suspensions for non-bonded van der Waals and for the PME real space
(Al2O3, Bühler), washed thoroughly with Milli-Q water and electrostatic interactions. The equations of motion were solved
dried under a nitrogen flux. Tetrabutylammonium hexafluoro- with the Verlet leap-frog algorithm with a 1 fs time-step. Heavy
phosphate (NBu4PF6) was used as supporting electrolyte (Acros atom–hydrogen covalent bonds were constrained with the
Organics, 98%) and was recrystallised several times from hot LINCS algorithm.60
ethanol. Before measurements, the solutions were deoxyge-
nated by bubbling N2 5 minutes.
Further, simulations with and without water were carried
out starting from pre-assembled aggregates. The aggregates
were built with PACKMOL61 for Ncpx = 10, resembling reverse
micelles-like structures with the head groups pointing towards
Computational studies
Molecular dynamics (MD) simulations of the complex a central spherical volume either filled with water (Nw = 100) or
[Fe(OC16-SalEen)2]ClO4 (1) in dichloromethane were performed in vacuum. The systems were then solvated with dichloro-
at 230 K and 298 K and 1 atm with the program GROMACS methane. As discussed in the Results and discussion section
2016.44 The force field of the complex was derived from the the perchlorate anions must be packed next to the iron head
Generalized Amber Force Field (GAFF).45 The geometry optimi- groups in these pre-assembled structures to avoid fast
zations and the Merz–Kollman46,47 charges were calculated disaggregation.
with Gaussian 09.48 A capping approach, similar to that used
TDDFT62(unrestricted) calculations were performed on the
to derive the atomic charges of the AMBER lipid force field49 HS (S = 5/2) and LS (S = 1/2) optimized structures described
was used, separating the complex into head and tail groups above, using the Gaussian 09 implementation,48 with the
(see Fig S12†). The geometries of the capped head and tail B3LYP functional63,64 and the standard 6-311G** basis set for
were optimized at the opbe50/cc-pvtz51 theoretical level.52 The all the atoms. The SMD solvation model65 was used to account
charges of the head group were calculated at the HF/6-31G* for the dichloromethane solvent.
level with RESP,53 whereas those of the tail were obtained with
AM1-BCC.54 RESP charges were not used for the tail because of
the poor description of ESP charges for hydrocarbons;55 ESP
and RESP charges for the tail atoms were found to be similar
Conflicts of interest
when calculated for the complex and for the capped tail, exhi- There are no conflicts to declare.
biting negative and positive charges, respectively, for the H
and C atoms of the hydrocarbon tail. The AM1-BCC charges,
in turn, although developed to reproduce the HF/6-31G*
electrostatic potential (ESP) of a molecule, do not exhibit this
Acknowledgements
limitation, probably because bond charge corrections (BCC) We thank Fundação para a Ciência e a Tecnologia for financial
were parameterized to reproduce the free energies of solvation. support (PTDC/QEQ-QIN/3414/2014, UID/MULTI/00612/2019,
Further details on the force filed parameterization can be UID/MULTI/04046/2019) and fellowship to PNM (SFRH/BPD/
found in the ESI.†
73345/2010). GGM thanks Science Foundation Ireland for gen-
To investigate the self-assembly of the complex in solution erous support from Investigator Project Award 12/IP/1703. PJC
MD simulations were performed for Ncpx = 25 molecules of thanks the Investigador FCT Program IF/00069/2014 and
complex and NPCL = 25 perchlorate anions in dichloromethane Exploratory Project IF/00069/2014/CP1216/CT0006. The COST
in a cubic box with periodic boundary conditions. Further, to actions CM1305 (ECOSTBio) and CA15128 (MOLSPIN) are
assess the role of crystallisation waters and solvent residual gratefully acknowledged.
water molecules, in the aggregation process, simulations were
performed by inserting a small cluster of TIP4P-Ew56 water in
solution. For the latter a pre-equilibrated system of TIP4P-Ew
water with Nw = 100 molecules at 298 K and 1 atm, was
Notes and references
inserted in the box, forming a nearly spherical cluster. The
initial positions of the complex molecules and perchlorate
anions were randomly assigned for the systems with and
without water and then solvated with dichloromethane (over
21 000 molecules). After steepest descent energy minimization
and a 100 ps MD in the NVT ensemble and a 500 ps MD in the
NPT ensemble, the trajectories were further propagated in the
1 M. Nakamura and M. Takahashi, in Mössbauer
Spectroscopy, John Wiley & Sons, Inc., Hoboken, New Jersey,
2013, pp. 177–201.
2 O. Kahn and C. J. Martinez, Science, 1998, 279, 44–48.
3 E. J. Devid, P. N. Martinho, M. V. Kamalakar, I. Šalitroš,
U. Prendergast, J. F. Dayen, V. Meded, T. Lemma,
R. González-Prieto, F. Evers, T. E. Keyes, M. Ruben,
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
B. Doudin and S. J. Van Der Molen, ACS Nano, 2015, 9, 26 D. Rosario-Amorin, P. Dechambenoit, A. Bentaleb,
4496–4507.
M. Rouzières, C. Mathonière and R. Clérac, J. Am. Chem.
Soc., 2018, 140, 98–101.
27 P. N. Martinho, Y. Ortin, B. Gildea, C. Gandolfi, G. McKerr,
B. O’Hagan, M. Albrecht and G. G. Morgan, Dalton Trans.,
2012, 41, 7461–7463.
4 B. Weber, W. Bauer and J. Obel, Angew. Chem., Int. Ed.,
2008, 47, 10098–10101.
5 J. A. Real, A. B. Gaspar, V. Niel and M. C. Muñoz, Coord.
Chem. Rev., 2003, 236, 121–141.
6 B. A. Leita, S. M. Neville, G. J. Halder, B. Moubaraki, 28 C. Gandolfi, G. G. Morgan and M. Albrecht, Dalton Trans.,
C. J. Kepert, J. F. Létard and K. S. Murray, Inorg. Chem.,
2012, 41, 3726–3730.
2007, 46, 8784–8795.
7 Z. Ni and M. P. Shores, Inorg. Chem., 2010, 49, 10727–
10735.
29 C. Gandolfi, C. Moitzi, P. Schurtenberger, G. G. Morgan
and M. Albrecht, J. Am. Chem. Soc., 2008, 130, 14434–
14435.
8 M. Yamada, H. Hagiwara, H. Torigoe, N. Matsumoto, 30 P. N. Martinho, A. I. Vicente, S. Realista, M. S. Saraiva,
M. Kojima, F. Dahan, J. P. Tuchagues, N. Re and S. Iijima,
A. I. Melato, P. Brandão, L. P. Ferreira and M. de
Chem. – Eur. J., 2006, 12, 4536–4549.
D. Carvalho, J. Organomet. Chem., 2014, 760, 48–54.
9 M. Quesada, F. Prins, E. Bill, H. Kooijman, P. Gamez, 31 B. D. P. Cardoso, A. I. Vicente, J. B. J. Ward, P. J. Sebastião,
O. Roubeau, A. L. Spek, J. G. Haasnoot and J. Reedijk,
Chem. – Eur. J., 2008, 14, 8486–8499.
10 M. Hostettler, K. W. Törnroos, D. Chernyshov, B. Vangdal
F. V. Chávez, S. Barroso, A. Carvalho, S. J. Keely,
P. N. Martinho and M. J. Calhorda, Inorg. Chim. Acta, 2015,
432, 258–266.
and H. B. Bürgi, Angew. Chem., Int. Ed., 2004, 43, 4589– 32 C. Gandolfi, N. Miyashita, D. G. Kurth, P. N. Martinho,
4594.
G. G. Morgan and M. Albrecht, Dalton Trans., 2010, 4508–
11 M. A. Halcrow, Chem. Soc. Rev., 2011, 40, 4119–4142.
4516.
12 R. Pritchard, H. Lazar, S. A. Barrett, C. A. Kilner, 33 C. Gandolfi, T. Cotting, P. N. Martinho, O. Sereda, A. Neels,
S. Asthana, C. Carbonera, J. F. Létard, M. A. Halcrow,
H. S. Sheu, N. Yasuda, Y. Ozawa, K. Toriumi and Y. Wang,
Dalton Trans., 2009, 125, 6656–6666.
G. G. Morgan and M. Albrecht, Dalton Trans., 2011, 40,
1855–1865.
34 T. Boonprab, P. Harding, K. S. Murray, W. Phonsri,
S. G. Telfer, A. Alkaş, R. Ketkaew, Y. Tantirungrotechai,
G. N. L. Jameson and D. J. Harding, Dalton Trans., 2018,
47, 12449–12458.
13 M. F. Tweedle and L. J. Wilson, J. Am. Chem. Soc., 1976, 98,
4824–4834.
14 S. Samanta, S. Demesko, S. Dechert and F. Meyer, Angew.
Chem., Int. Ed., 2015, 54, 583–587.
35 L. A. Yatsunyk and F. A. Walker, Inorg. Chem., 2004, 43,
757–777.
15 D. Siretanu, D. Li, L. Buisson, D. M. Bassani, S. M. Holmes,
C. Mathonière and R. Clérac, Chem. – Eur. J., 2011, 17, 36 D. F. Evans, J. Chem. Soc., 1959, 2003–2005.
11704–11708.
37 W. T. Yap and L. M. Doane, Anal. Chem., 1982, 54, 1437–1439.
16 B. Weber and F. A. Walker, Inorg. Chem., 2007, 46, 6794– 38 M. L. Klein and W. Shinoda, Science, 2008, 321, 798–800.
6803.
39 M. Tarek, S. Bandyopadhyay and M. L. Klein, J. Mol. Liq.,
1998, 78, 1–6.
40 M. Jorge, Langmuir, 2008, 24, 5714–5725.
17 M. A. Halcrow, Spin-crossover materials : properties and
applications, J. Wiley and Sons, Inc, 2013.
18 L. J. Kershaw Cook, R. Kulmaczewski, R. Mohammed, 41 S. J. Marrink, D. P. Tieleman and A. E. Mark, J. Phys. Chem.
S. Dudley, S. A. Barrett, M. A. Little, R. J. Deeth and B, 2000, 104, 12165–12173.
M. A. Halcrow, Angew. Chem., Int. Ed., 2016, 55, 4327– 42 J. B. Maillet, V. Lachet and P. V. Coveney, Phys. Chem.
4331. Chem. Phys., 1999, 1, 5277–5290.
19 M. van der Meer, Y. Rechkemmer, F. D. Breitgoff, 43 S. Abel, J. Attia, S. Rémita, M. Marchi, W. Urbach and
S. Dechert, R. Marx, M. Dörfel, P. Neugebauer, J. van
Slageren and B. Sarkar, Dalton Trans., 2016, 45, 8394–8403.
20 H. Petzold, P. Djomgoue, G. Hörner, J. M. Speck, T. Rüffer
and D. Schaarschmidt, Dalton Trans., 2016, 45, 13798–
13809.
21 Y. H. Luo, M. Nihei, G. J. Wen, B. W. Sun and H. Oshio,
Inorg. Chem., 2016, 55, 8147–8152.
22 S. Rodríguez-Jiménez, M. Yang, I. Stewart, A. L. Garden and
S. Brooker, J. Am. Chem. Soc., 2017, 139, 18392–18396.
23 C. J. Johnson, G. G. Morgan and M. Albrecht, J. Mater.
Chem. C, 2015, 3, 7883–7889.
M. Goldmann, Chem. Phys. Lett., 2009, 481, 124–129.
44 D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof,
A. E. Mark and H. J. C. Berendsen, J. Comput. Chem., 2005,
26, 1701–1718.
45 J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman and
D. A. Case, J. Comput. Chem., 2004, 25, 1157–1174.
46 U. C. Singh and P. A. Kollman, J. Comput. Chem., 1984, 5,
129–145.
47 B. H. Besler, K. M. Merz and P. A. Kollman, J. Comput.
Chem., 1990, 11, 431–439.
48 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato,
A. Marenich, J. Bloino, B. G. Janesko, R. Gomperts,
B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov,
24 N. G. White, H. L. C. Feltham, C. Gandolfi, M. Albrecht and
S. Brooker, Dalton Trans., 2010, 39, 3751–3758.
25 S. Schlamp, P. Thoma and B. Weber, Chem. – Eur. J., 2014,
20, 6462–6473.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, 54 A. Jakalian, D. B. Jack and C. I. Bayly, J. Comput. Chem.,
F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, 2002, 23, 1623–1641.
D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, 55 D. E. Williams, J. Comput. Chem., 1994, 15, 719–732.
W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, 56 H. W. Horn, W. C. Swope, J. W. Pitera, J. D. Madura,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr.,
T. J. Dick, G. L. Hura and T. Head-Gordon, J. Chem. Phys.,
2004, 120, 9665–9678.
J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, 57 G. Bussi, D. Donadio and M. Parrinello, J. Chem. Phys.,
K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, 2007, 126, 014101.
J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, 58 M. Parrinello and A. Rahman, J. Appl. Phys., 1981, 52,
S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, 7182.
C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, 59 U. Essmann, L. Perera, M. L. Berkowitz, T. Darden,
K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox,
Gaussian 09. Revision D.01, Gaussian, Inc., Wallingford CT,
2016.
H. Lee and L. G. Pedersen, J. Chem. Phys., 1995, 103,
8577.
60 B. Hess, H. Bekker, H. J. C. Berendsen and J. G. E.
M. Fraaije, J. Comput. Chem., 1997, 18, 1463–1472.
49 C. J. Dickson, B. D. Madej, Å. A. Skjevik, R. M. Betz,
K. Teigen, I. R. Gould and R. C. Walker, J. Chem. Theory 61 L. Martínez, R. Andrade, E. G. Birgin and J. M. Martínez,
Comput., 2014, 10, 865–879. J. Comput. Chem., 2009, 30, 2157–2164.
50 Y. Zhang, A. Wu, X. Xu and Y. Yan, Chem. Phys. Lett., 2006, 62 C. Adamo and D. Jacquemin, Chem. Soc. Rev., 2013, 42,
421, 383–388.
845–856.
51 T. H. Dunning, J. Chem. Phys., 1989, 90, 1007–1023.
63 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
52 M. Güell, J. M. Luis, M. Solà and M. Swart, J. Phys. Chem. A, 64 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens.
2008, 112, 6384–6391. Matter Mater. Phys., 1988, 37, 785–789.
53 W. D. Cornell, P. Cieplak, C. I. Bayly and P. A. Kollman, 65 A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys.
J. Am. Chem. Soc., 1993, 115, 9620–9631.
Chem. B, 2009, 113, 6378–6396.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.