An iron(II) spin crossover grafted cyclotriphosphazene

Article abstract of DOI:10.1016/j.poly.2013.02.065

The synthesis of the new cyclotriphosphazene (CTP) ligand substituted with a pendant 2,6-bis(benzimidazole-2-yl)pyridine (bbp), namely (pentaphenoxy)[4-{2, 6-bis(benzimidazole-2-yl)pyridine-4-yl}phenoxy]cyclotriphosphazene L is reported. The single crystal structure of L shows that the bbp group is attached to the CTP via the oxygen. L reacts with FeX₂ (X = ClO ₄^- or BF₄^-) salts forming the [FeL₂]X₂ complexes 1 and 2 respectively. For [FeL ₂](BF₄)₂ (2), the single crystal structure shows an 'N₆' coordination sphere around the iron atom. UV-Vis, resonance Raman and M?ssbauer spectroscopies and magnetic susceptibility measurements, aided by density functional theory (DFT) calculations, determine the complexes are low spin below 300 K but display spin crossover (SCO) behavior above this temperature, hence showing that the addition of a phosphazene to a SCO moiety does not prevent SCO.

Full text of DOI:10.1016/j.poly.2013.02.065

Polyhedron 55 (2013) 37–44
Contents lists available at SciVerse ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
An iron(II) spin crossover grafted cyclotriphosphazene
a,
a
a
Ross J. Davidson a, Eric W. Ainscough a, , Andrew M. Brodie , Geoffrey B. Jameson , Mark R. Waterland ,
⇑
⇑
Boujemaa Moubaraki b, Keith S. Murray b, Keith C. Gordon c, Raphael Horvath c, Guy N.L. Jameson c
a Chemistry – Institute of Fundamental Sciences, Massey University, Private Bag 11 222, Palmerston North 4442, New Zealand
b School of Chemistry, Building 23, Monash University, Clayton, Victoria 3800, Australia
c MacDiarmid Institute for Advanced Materials and Nanotechnology, Department of Chemistry, University of Otago, Dunedin 9054, New Zealand
a r t i c l e i n f o
a b s t r a c t
Article history:
The synthesis of the new cyclotriphosphazene (CTP) ligand substituted with a pendant 2,6-bis(benzimid-
azole-2-yl)pyridine (bbp), namely (pentaphenoxy)[4-{2,6-bis(benzimidazole-2-yl)pyridine-4-yl}phen-
oxy]cyclotriphosphazene L is reported. The single crystal structure of L shows that the bbp group is
attached to the CTP via the oxygen. L reacts with FeX2 (X = ClO4ꢀ or BF4-) salts forming the [FeL2]X2 com-
plexes 1 and 2 respectively. For [FeL2](BF4)2 (2), the single crystal structure shows an ‘N6’ coordination
sphere around the iron atom. UV–Vis, resonance Raman and Mössbauer spectroscopies and magnetic sus-
ceptibility measurements, aided by density functional theory (DFT) calculations, determine the com-
plexes are low spin below 300 K but display spin crossover (SCO) behavior above this temperature,
hence showing that the addition of a phosphazene to a SCO moiety does not prevent SCO.
Ó 2013 Elsevier Ltd. All rights reserved.
Received 11 October 2012
Accepted 18 February 2013
Available online 5 March 2013
Keywords:
Cyclotriphosphazene
Iron(II)
Spin crossover
1. Introduction
time phosphazene systems containing these imidazolyl-pyridyl
species have been reported and it was of interest to determine if
For many years, spin crossover (SCO) materials have been sug-
gested as having potential for use in quantum computers and mas-
sive data storage systems for example [1,2]. However, because
these materials are often crystalline, they are difficult and expen-
sive to deposit and process. Previously we attempted to improve
the processing properties by attaching pendant 2,20:60,200-terpyri-
dine (terpy) substituents to both cyclotri- and polyphosphazene
(CTP and PP) (Scheme 1) and forming the subsequent iron(II) com-
plexes [3]. The repeating nitrogen-phosphorus backbone provided
an electronically mute scaffold for the metal–ligand centers; how-
ever, it was determined that only the start of SCO was observed at
very high temperatures, preventing most practical applications [3].
The present study aims to improve on the terpy-based design
by using a 2,6-bis(1H-benzimidazol-2-yl)pyridine (bbp) moiety at-
tached to a CTP. The physical properties of the resultant iron(II)
complexes were investigated using a variety of spectroscopic tech-
niques, such as electronic absorbance, solid state resonance Raman
and Mössbauer spectroscopy as well as magnetic susceptibility.
DFT calculations were employed to obtain insight into the behavior
of the compounds. Bbp was chosen, as iron(II)-bis-bbp
([Fe(bbp)2]2+) complexes are well known for their SCO behavior
[4–8], thus should provide an improvement over terpy-based
phosphazene complexes previously reported [3]. This is the first
and how the substitution of the phosphazene alters their SCO
behavior.
2. Experimental
Analytical grades of solvents were used. 2,6-bis(1H-ben-
zimidazol-2-yl)pyridine-4(1H)-one (HObbp) [9] and 1,2,2,3,3-pen-
takis(phenoxy)-1-chlorocyclotriphosphazene, N3P3(OPh)5Cl, [10]
were synthesized by literature methods. K2CO3, Fe(ClO4)2ꢁ6H2O
and Fe(BF4)2ꢁ6H2O were sourced from Aldrich–Sigma. To avoid ir-
on(II) oxidation, all manipulations were carried out under an argon
atmosphere, using standard Schlenk techniques.
Caution! Perchlorate salts with organic ligands are potentially
explosive and should be handled with the necessary precautions.
2.1. Synthesis of the compounds
2.1.1. [N3P3(OPh)5(Obbp)] (L)
N3P3(OPh)5Cl (195 mg, 0.31 mmol) was added to a solution con-
taining HObbp (100 mg, 0.31 mmol) and K2CO3 (45 mg, 0.33 mmol)
in acetone (50 mL). After stirring at reflux for five days the solvent
was removed under reduced pressure, leaving a pink solid that was
washed with CHCl3/water and dried over MgSO4. A minimal
amount of hexane was added to the solution to form a white precip-
itate. The precipitate was filtered and dried under vacuum for ele-
mental analysis. Diffractable crystals were grown via a slow
evaporation of an acetone/water solution. Yield: 140 mg (51%).
⇑
Corresponding authors. Tel.: +64 63569099; fax: +64 63505682.
E-mail addresses: e.ainscough@massey.ac.nz (E.W. Ainscough), a.brodie@
massey.ac.nz (A.M. Brodie).
0277-5387/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.poly.2013.02.065

38
R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
Table 1
R
R
R
Crystal and Refinement Data for Lꢁ3H2OꢁC3H6O and 2ꢁC3H6OꢁC5H12O.
R
P
N
N
N
P
N
Compound
Lꢁ3H2OꢁC3H6O
52H49N8O10P3
2ꢁC3H6OꢁC5H12
C106H92B2F8FeN16O14P6
2229.26
O
R
R
P
P
N
P
Molecular formula
C
M (g molꢀ1
T (K)
)
1038.91
163(2)
R
n
R
R
R
123(2)
Cyclotriphosphazene (CTP)
Polyphosphazene (PP)
Crystal system
Space group
a (Å)
triclinic
P1
triclinic
P1
ꢁ
ꢁ
Scheme 1. Generic polyphosphazene (PP) and cyclotriphosphazene (CTP)
11.8419 (4)
14.0379 (4)
15.3918 (11)
91.665 (7)
90.322 (6)
96.231 (2)
2542.4 (2)
2
1.634
1.357
117.88
7193
12.880(3)
17.690(4)
24.070(5)
85.16(3)
83.21(3)
75.32(3)
5259.6(21)
2
2.736
1.407
102.26
11111
structures.
b (Å)
c (Å)
a
(o)
b (o)
ESMS: m/z 927 [N3P3(OPh)5(Obbp)H]+, 965 [N3P3(OPh)5
c
(o)
1
(Obbp)K]+.31P{1H} NMR (CDCl3): d 9.6 ppm (s). H NMR (CDCl3): d
V (Å3)
8.02 ppm (s, 2H), 7.40 (s, 2H), 7.05–7.00 (m, 10H), 6.95 (d
(J = 8 Hz), 10H), 6.81 (t (8), 8H), 6.77 (d (8), 5H). Anal. Calc. for C49
H37N8O6P3ꢁ /3CHCl3: C, 59.28; H, 3.77; N, 11.13. Found: C, 59.18;
Z
l
(Cu K
a
) (mmꢀ1
)
qcalc (g cmꢀ3
2hmax (°)
)
2
H, 3.76; N, 11.06%.
Number of unique reflections
Data/restraints/parameters
7193/60/685
R1 = 0.0756
wR2 = 0.1775
R1 = 0.1347
wR2 = 0.2407
1.074
11111/2766/1494
R1 = 0.0916
wR2 = 0.2155
R1 = 0.1137
wR2 = 0.2296
0.967
2.1.2. [FeL2](ClO4)2 (1)
Final R indices [I > 2
r(I)]
To a stirred solution of L (100 mg, 0.11 mmol) in MeOH:CHCl3
(1:1, 8 mL), Fe(ClO4)2ꢁ6H2O (4.0 mg, 0.066 mmol) was added,
immediately turning the solution purple. Stirring was continued
for 30 min before the solvent was removed under reduced pres-
sure. The solid was dissolved in CHCl3 and filtered through Celite.
The filtrate was dried under reduced pressure, leaving a purple so-
lid. Weakly diffractable crystals were grown by dissolving the solid
in acetone and diffusing in tert-butyl-methyl ether. Yield: 72 mg
R indices (all data)
Goodness-of-fit (GOF) on F2
CHN [14]. 31P{1H} NMR measurements were carried out on a Bruker
Avance 400 spectrometer and 1H NMR on a Bruker Avance 500
spectrometer. Electrospray mass spectra were obtained from aceto-
nitrile solutions on a Micromass ZMD spectrometer run in the posi-
tive ion mode. Listed peaks correspond to the most abundant
isotopomer; assignments were made by a comparison of observed
and simulated spectra. All ground state vibrational measurements
were made using KBr disks on a Nicolet 5700 FT-IR spectrometer.
Continuous wave excitation was used for all Raman measurements.
UV–Vis absorbance spectra were recorded using an Oceans Optics
USB2000 + UV–Vis spectrophotometer on solutions prepared by
dissolving the solid complexes in acetonitrile. Cyclic voltammetry
was obtained using a glassy carbon working electrode, Pt counter
(62%). ESMS: m/z 955 [Fe(L)2]2+
.
31P{1H}NMR (CD3CN):
d
1
11.28 ppm (d (93 Hz), 4P), ꢀ44.09 (t (93), 2P). HNMR (CD3CN): d
14.48 ppm (4H), 9.00 (4H), 8.06 (4H), 7.85 (20H), 7.67 (10H),
7.57 (20H), 6.73 (4H), 6.32 (4H), 6.17 (4H). Anal. Calc. C98H74Cl2
FeN16O20P6ꢁC5H12Oꢁ2H2O: C, 55.41, H, 4.06; N, 10.04. Found: C,
55.26; H, 4.01; N, 10.01%.
2.1.3. [FeL2](BF4)2 (2)
The same procedure as for 1 was used, except Fe(BF4)2ꢁ6H2O
was used in place of Fe(ClO4)2ꢁ6H2O. Diffractable crystals were
grown by dissolving the solid in acetone and diffusing in tert-
butyl-methyl ether. Yield: 70 mg (63%). ESMS: m/z 955 [Fe(L)2]2+
.
electrode and Ag/AgCl reference electrode;
m
= 0.1 V sꢀ1 recorded
31P{1H}NMR (CD3CN): d 11.6 ppm (d (93 Hz), 4P), ꢀ41.9 (t (93),
2P). 1HNMR (CD3CN): d 14.60 ppm (4H), 8.94 (4H), 8.04 (4H),
7.84 (20H), 7.61 (10H), 7.57 (20H), 6.84 (4H), 6.64 (4H), 6.11
in acetonitrile, 0.1 M TBAClO4 at 10ꢀ3 for the complexes 1; 0.1 M
TBABF4 at 10ꢀ3 for 2.
Resonance Raman measurements were carried out based on a
modified version of a system described previously [15–17]. Spectra
were acquired of solid-state samples at 79, 298 and 362 K with a
number of excitation wavelengths. Temperature control was
achieved by utilizing a variable temperature cell (Specac, Wood-
stock, GA, USA) and a high stability temperature controller (Specac,
Woodstock, GA, USA). Vacuum-purging of the cell was used to
minimize condensation and frosting of the quartz window of the
variable temperature cell. For excitation at 350.7 and 568.2 nm, a
continuous-wave Innova I-302 krypton-ion laser (Coherent, Inc.)
was used. For excitation at 457.9 and 514.5 nm an Innova Sabre
DBW argon-ion laser (Coherent, Inc.) was used. The beam was
passed through either a Pellin-Broca prism (for 350.7, 457.9 and
514.5 nm) or a holographic laser band pass filter (Kaiser Optical
Systems, Inc.) and subsequently two irises in order to remove un-
wanted laser-lines. The beam-power was adjusted between 20 and
40 mW at the sample, depending on the wavelength used. The
sample and collection lens were arranged in a 180ꢀbackscattering
geometry where the collection lens also served to focus the excita-
tion beam on the sample. The Raman photons were focused on the
entrance slit of an Acton Research SpectraPro500i spectrograph
(Princeton Instruments, Inc.) with a 1200 grooves mmꢀ1 grating.
3
(4H). Anal. Calc. C98H74Cl2FeN16O20P6ꢁ /4C5H12OꢁH2O: C, 56.39; H,
3.95; N, 10.34. Found: C, 56.34; H, 4.02: N, 10.39%.
2.2. Crystallographic studies
The X-ray data were collected at low temperature with a Riga-
ku-Spider X-ray diffractometer, comprising a Rigaku MM007
microfocus copper rotating-anode generator, high-flux Osmic
monochromating and focusing multilayer mirror optics (Cu K radi-
ation, k = 1.5418 Å), and a curved image-plate detector. CrystalClear
was utilized for data collection and FSProcess in PROCESS-AUTO was
used for cell refinement and data reduction. Crystal refinement
data are given in Table 1. The structures were solved by direct
methods and refined using both the SHELXTL [11] and OLEX2 [12] pro-
grams. Hydrogen atoms were calculated at ideal positions.
For the solution of 2ꢁC3H6OꢁC5H12O, the electron density (15 eꢀ
per cell in a void volume of 118.5 Å3) of the occupationally and
positionally disordered acetone molecule was removed by using
PLATON/SQUEEZE [13].
2.3. Physical measurements
The slit width was set to 50 l .
m, giving a resolution of ca. 2 cmꢀ1
Microanalyses were performed by the Campbell Microanalytical
Laboratory, University of Otago, New Zealand and assigned using
Radiation from Raleigh and Mie-scattering was attenuated using
a notch filter (Kaiser Optical Systems, Inc.) for 568.2 nm and Razor

R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
39
Edge filters (Semrock, Inc.) for other wavelengths. The dispersed
photons were detected using a Princeton Instruments liquid nitro-
gen cooled 1152-EUV charge-coupled detector controlled by a
Princeton Instruments ST-130 controller. WinSpec/32 software
(Roper Scientific, Inc.) was used to control the CCD, and spectra
were analyzed using GRAMS/32 (Galactic Industries Corp.) soft-
ware. Wavelength calibration was achieved using a reference sam-
ple made from 1:1 toluene:acetonitrile and general alignment was
done using a solid sample of carbamazepine.
pellet of accurately known susceptibility provided by Quantum De-
sign Inc. and checked against the temperature-dependent behavior
of CuSO4ꢁ5H2O.
2.4. Computational
Density functional theory (DFT) calculations were carried out
using the Gaussian09 package (Gaussian, Inc) [18]. Frequency
and time-dependent (TD) calculations were performed on opti-
mized ground-state structures, all results were displayed using
GaussView [19]. All calculations used the 6-31G(d) basis set
employing two different levels, OLYP and B3LYP. The vibrational
spectra generated by both levels were compared to the measured
vibrational data determining OLYP to be the most accurate level.
An unambiguous assignment of vibrational modes from visual
comparison of spectra was possible for most peaks. TD-DFT calcu-
lations were carried out in an acetonitrile solvent field using the
SCRF-PCM method which creates the solvent cavity via a set of
overlapping spheres. Geometry optimizations were not carried
out in a solvent field, as these are difficult to achieve for molecules
this size; however, correlation is found to be better than for calcu-
lations where solvent contributions have been completely
neglected.
57Fe Mössbauer spectra were recorded using approximately
30 mg of sample placed in a nylon sample holder (12.8 mm diam-
eter, 1.6 mm thickness) with Kapton windows. The spectra were
measured on a Mössbauer instrument from SEE Co. (Science Engi-
neering & Education Co., MN) equipped with a closed cycle refrig-
erator system from Janis Research Co. and SHI (Sumitomo Heavy
Industries Ltd.). Data were collected in constant acceleration mode
in transmission geometry with an applied field of 47 mT parallel to
the
c
-rays. The zero velocity of the Mössbauer spectra refers to the
m metallic
centroid of the room temperature spectrum of a 25
l
iron foil. Analysis of the spectra was conducted using the WMOSS
program (SEE Co., formerly WEB Research Co., Edina, MN).
Magnetic susceptibilities were determined using a Quantum
Design Inc. Squid MPMS5 magnetometer with the ꢂ25 mg samples
held in a calibrated gel capsule that was placed in the center of a
drinking straw that was fixed to the end of the sample rod. A DC
field of 1 T was used. The instrument was calibrated against a Pd
The two different levels for computational models (B3LYP and
OLYP) with a 6-31G(d) basis were compared to vibrational data
collected to determine which was the most accurate in addition
N
N
NH
N
N
H
O
N
P
OPh
N
Cl
OPh
N
P
N
P
N
i
N
PhO
OPh
PhO
OPh
P
P
P
OPh OPh
OPh OPh
(L)
ii
HN
HN
NH
NH
X-
PhO
N
N
N
N
N
PhO
OPh
P
OPh
PhO
P
P
N
N
N
O
N
Fe
O
N
N
P
P
OPh
PhO
PhO
P
N
OPh
OPh
X-
-
X = ClO4 (1)
-
X = BF4 (2)
Scheme 2. Synthesis scheme for L, 1 and 2. (i) HObpp, K2CO3; (ii) 1 Fe(ClO4)2ꢁ6H2O or 2 Fe(BF4)2ꢁ6H2O.

40
R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
Fig. 1. Crystal structure of L. Hydrogen atoms and solvate species removed for clarity.
Fig. 2. Crystal structure of 2. Hydrogen atoms, disordered atoms, solvate species and anions removed for clarity.
to validating the models. Higher level basis sets were not explored
due to the size of the molecule making such calculations unfeasi-
ble. The B3LYP frequencies were scaled by 0.9613 [20] and OLYP
by 0.9782 [21]. The assignments of the spectra were made using
each level and the MAD values determined for all of the assigned
peaks (Supplementary data Table S2.1). These values indicated that
there was little difference between the levels, but a comparison be-
tween the measured and calculated bond lengths indicated that

R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
41
Table 2
In addition, to prevent the reactions with the imidazole proton, a
carbonate base was used in place of a hydride. As a final factor,
HObpp had limited solubility in THF, therefore it was necessary
to use acetone, a solvent with a much lower boiling point.
The iron(II) complexes, 1 and 2, were synthesized by reacting
two equivalents of the L with one equivalent of the appropriate iron
salt, Fe(ClO4)2ꢁ6H2O or Fe(BF4)2ꢁ6H2O, dissolved in CHCl3/CH3OH
(Scheme 2). The reactions occurred rapidly at room temperature
to produce purple solutions from which the air stable, solid prod-
ucts were isolated. NMR coupling data were unavailable for the iron
complexes due to the complexes being slightly paramagnetic in
solution even below room temperature.
Electrochemical data for compounds 1 and 2.
Complex
E
(V) (
D
Ep, mV)
E
(V) (
D
Ep, mV)
E
(V) (DEp, mV)
½ox
½red
½red
1
2
0.575 (135)
0.550 (108)
ꢀ0.930 (317)
ꢀ0.912 (190)
ꢀ1.474 (irr)
ꢀ1.556 (irr)
the OLYP models were an order of magnitude more accurate than
the B3LYP models (Supplementary data Table S2.2).
3. Results and discussion
3.1. Synthesis
3.2. Molecular structures
The phosphazene ligand L was synthesized by reacting the
potassium salt of HObpp with N3P3(OPh)5Cl (Scheme 2) for 5 days.
As with the 2,6-bis(2-pyridyl)-4(1H)-pyridone substituted phos-
phazene [3], HObpp is most stable in the enone tautomer; this con-
sequently lowers the acidity of the phenol, reducing the reactivity.
The crystal structure of L (Fig. 1) shows the Obbp group attached
to a phosphorus atom via the oxygen atom consistent with the
spectroscopic data collected. In addition, no interactions are ob-
served between any of the substituent groups and the phosphazene
ring; this is an indication that if the polymer were synthesized it
Fig. 3. Left: the visible spectrum of 1 from 253–343 K. Right: eMCLT vs. temperature for (1) (black squares) and 2 (red circles).
Fig. 4. Molecular orbitals associated with the 548 nm transition (phenoxy groups and phosphazene ring removed for clarity).

42
R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
Fig. 5. rR collected for 2. Left: kex 568 nm. Right: kex 514 nm.
would not decay via substituent effects; see Supplementary data
(Table S1.1) for bond lengths and angles. The iron(II) compound 2
shows an ‘N6’ coordination sphere, consistent with the spectro-
scopic data collected (Fig. 2). The iron–nitrogen bond lengths are
Fe(1)–N(1A) 1.920 (6), Fe(1)–N(2A) 2.005(5), Fe(1)–N(3A) 2.002
(6), Fe(1)–N(1B) 1.903 (6), Fe(1)–N(2B) 1.995 (6) and Fe(1)–
N(3B)2.003 (6) Å, consistent with a low spin(LS) iron(II) bbp com-
plex [4]. The phosphazene rings of 2 also exhibit little change as a
result of the metal coordination (see Supplementary data Table
S2). The tetrafluoroborate anions only display weak bonding with
the hydrogens of the cationic complex, filling in the cavities; there-
fore, it can be assumed that the perchlorate anion of the isomor-
phous but poorly diffracting 1 will reside in an analogous position.
(ClO4)2(0.415 V) [22] by 160–133 mV respectively, indicative of
the phosphazene being an electron-withdrawing group.
3.4. Electronic spectroscopy
The temperature dependent electronic spectra for 1 are shown
in Fig. 3. Both 1 and 2 display identical MLCT bands, with kmax at
560 and 561 nm respectively, typical of LS t2g6 forms of [Fe(bbp)2]2+
complexes [23]. There are red shifts of 8 and 9 nm, respectively, rel-
ative to the unsubstituted complex [Fe(bbp)2](ClO4)2 (kmax 552 nm
in acetonitrile) [23], which can be explained by the slight electron-
withdrawing effects of the phosphazene ether unit, as previously
reported [3,17,24,25]. Beyond this, the first ten calculated transi-
tions with nonzero oscillator strengths can be found in the Supple-
mentary Data (Table S3.1). The lowest-energy peaks have been
attributed to MLCT transitions. Fig. 4 shows some of the molecular
orbitals of 1 and 2 involved in the transition with the largest oscil-
lator strength (548 nm, see Supplementary data), which can be de-
scribed as Hꢀ1 and Hꢀ2 ? L, L+1, L+2 and L+3. As the transition
involves a net electron shift from the iron to the bbp ligand, this
is consistent with the assignment of MLCT with no observable orbi-
tal overlap with the phosphazene ring.
3.3. Electrochemistry
Cyclic voltammetry measurements determined that each of the
complexes had one reversible oxidation potential associated with
the iron center and two reductions associated with the ligand
(see Table 2). The oxidation potentials of 1 and 2 differed by
25 mV providing further evidence that the anions were interacting
with the complex. In addition, the oxidation potentials are
cathodically shifted relative to unsubstituted complex, [Fe(bbp)2]
The extinction coefficient (e) of the MLCT band decreased as the
temperature was increased as expected for SCO behavior. How-
ever, a plot of the eMCLT versus temperature revealed a significant
difference in the behavior of 1 and 2 (see Fig. 3). Baitalik and co-
workers demonstrated that ruthenium-bpp complexes could
hydrogen bond with the anions via imidazole protons resulting
in different physical properties based on the anion present [26]
which possibly explains the difference. In addition, neither plot
shows the two plateaux expected for a Boltzmann distribution of
HS and LS species. This has previously been observed and ex-
plained by Linert et al. [23] and explained as arising from an equi-
librium formed between the complex and a coordinating solvent
(e.g. acetonitrile or benzonitrile).
½FeL2ꢃ2þ ꢀ ½FeLðPhCNÞ3ꢃ2þ þ L ꢀ ½FeðPhCNÞ6ꢃ2þ þ 2L
The combination of each of these processes means that the ther-
modynamic parameters of the SCO process could not be deter-
mined even if it was occurring.
3.5. Vibrational spectroscopy
FT-IR spectra are dominated by aromatic ring deformations and
anion vibrations therefore solid-state resonance Raman (rR) was
employed to specifically investigate the behavior of the iron-bbp
Fig. 6. Vibrational mode assigned to 1556 cmꢀ1 for HS-[Fe(L)2]2+, red arrows
indicate displacement vectors.

R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
43
Fig. 7. Magnetic moment leff vs. temperature. Left: 1, Right: 2. Black squares (1st run), red triangles (2nd run), blue diamonds (3rd run).
center as a function of temperature. rR is well suited to this as it
shows selective enhancement of modes within active chromoph-
ores, where the excitation wavelength (kex) determines which par-
ticular chromophore is being probed [17]. The spectra were
collected at kex 568.2 and 514.5 nm (exciting the MLCT transition)
(Fig. 5). Assignments of modes from the other excitation wave-
lengths kex are included in the Supplementary Data (Table S4.1).
Spectra were recorded at 80, 298 and 362 K to observe the pres-
ence of both the HS and LS species. By comparison to DFT models,
the enhanced vibrational modes were found to primarily consist of
the benzimidazole and pyridine ring distortions. Significant activ-
ity of the chelating nitrogen atoms was observed, which is consis-
tent with the assignment of the electronic transitions as MLCT.
As the temperature was increased, changes in the rR spectra
were observed; for example, additional peaks were observed at
1556 cmꢀ1 (Fig. 5). Comparison to a high spin (HS) DFT model al-
lowed a tentative assignment of the differences to the formation
of the HS species (Fig. 6).
3.6. Magnetic susceptibility
Shown in Fig. 7 are plots of the magnetic moments (
temperature for 1 and 2. Each of the complexes displayed an in-
crease in eff as the temperature was increased. Due to the danger
leff) versus
Fig. 8. Mössbauer spectra of 1 and 2 recorded at 4.6 K.
l
of heating a perchlorate salt, 1 could not be heated beyond 300 K.
As a result only the start of the spin transition curve was observed.
The LS form gave moment values, below ꢂ250 K, of close to zero, as
expected. However, for 2 the complex could be heated to 400 K
(the limitation of the SQUID). The complex displayed a full SCO
Table 3
57Fe Mössbauer data for complexes 1 and 2.
T (K)
d (mm sꢀ1
)
D )
EQ (mm sꢀ1
1
2
4.6
0.35
0.32
0.29
0.34
0.29
0.47
0.46
0.47
0.45
0.48
curve with a T at 336 K (Fig. 7, 1st run). In addition, after heating
200
294
4.6
½
2 beyond 350 K solvent was lost from the lattice, greatly altering
the magnetic behavior of the complex, which does not completely
return to the LS state upon cooling (Fig. 7, 2nd run). Successive
heating resulted in a further loss of solvent, further altering the
magnetic behavior (Fig. 7, 3rd run), trapping it in a spin state that
has more HS character than has the parent. The precise make-up of
295
acceptable signal-to-noise ratio and therefore isotope enrichment
was not employed. Spectra were recorded at 4.7, 200 and 293 K.
Each of the complexes (1 and 2) displayed a quadrupole doublet
the spin state(s) of the partially desolvated forms, below T , re-
½
with parameters (d and
D
EQ) characteristic of LS [Fe(bbp)2]2+ com-
quires further detailed studies such as Mössbauer spectra and
powder X-ray diffraction vs. temperature.
plexes (see Fig. 8 and Table 3) which is consistent with the crystal
structure and magnetic data obtained. Neither complex displayed a
significant change in its Mössbauer parameters as the temperature
was increased (see Table 3); the small change in isomer shift is due
to the second-order Doppler effect [27]. The slight asymmetry in
the line widths is most likely a texture effect caused by preferred
orientation of the crystallites in the sample [28]. The data are
3.7. Mössbauer spectroscopy
As a final means of investigating the coordination mode and
spin state of the complexes, variable temperature Mössbauer spec-
troscopy was employed. Natural abundance iron provided an

44
R.J. Davidson et al. / Polyhedron 55 (2013) 37–44
therefore consistent with the magnetic data, as all measurements
were below the SCO temperature, due to the limitations of the
Mössbauer instrument.
can be found, in the online version, at http://dx.doi.org/10.1016/
j.poly.2013.02.065.
References
4. Conclusions
[1] O. Kahn, C.J. Martinez, Science 279 (1998) 44.
[2] P. Gutlich, H.A. Goodwin, Spin crossover – an overall perspective, in: Spin
Crossover in Transition Metal Compounds I, Springer-Verlag, Heidelberg. 2004,
pp. 1–47.
[3] R.J. Davidson, E.W. Ainscough, A.M. Brodie, G.B. Jameson, M.R. Waterland, H.R.
Allcock, M.D. Hindenlang, B. Moubaraki, K.S. Murray, K.C. Gordon, R. Horvath,
G.N.L. Jameson, Inorg. Chem. 51 (2012) 8307.
[4] R. Boca, F. Renz, M. Boca, H. Fuess, W. Haase, G. Kickelbick, W. Linert, M.
Vrbova-Schikora, Inorg. Chem. Commun. 8 (2005) 227.
[5] M. Enamullah, W. Linert, Thermochim. Acta 388 (2002) 401.
[6] M. Hasegawa, F. Renz, T. Hara, Y. Kikuchi, Y. Fukada, J. Okubo, T. Hoshi, W.
Linert, Chem. Phys. 277 (2002) 21.
[7] R. Boca, M. Boca, L. Dlhan, K. Falk, H. Fuess, W. Haase, R. Jarosciak, B.
Papankova, F. Renz, M. Vrbova, R. Werner, Inorg. Chem. 40 (2001) 3025.
[8] S. Hayami, N. Motokawa, A. Shuto, R. Moriyama, N. Masuhara, K. Inoue, Y.
Maeda, Polyhedron 26 (2007) 2375.
[9] M.G.B. Drew, C. Hill, M.J. Hudson, P.B. Iveson, C. Madic, L. Vaillant, T.G.A.
Youngs, New J. Chem. 28 (2004) 462.
[10] E.T. McBee, K. Okuhara, C.J. Morton, Inorg. Chem. 5 (1966) 450.
[11] G.M. Sheldrick, SHELXL, Suite of Programs for Crystal Structure Analysis, in,
Tammanstrasse 4, Gottingen, 1998.
[12] O.V. Dolomanov, L.J. Bourhis, R.J. Gildea, J.A.K. Howard, H. Puschmann, J. Appl.
Crystallogr. 42 (2009) 339.
A new CTP ligand L containing a bbp moiety was synthesized by
reacting the potassium salt of HObbp with N3P3(OPh)5Cl. Iron(II)
complexes were synthesized by the reaction of two equivalents
of L with one equivalent of the appropriate iron(II) salts, Fe(ClO4)2-
ꢁ6H2O or Fe(BF4)2ꢁ6H2O, to afford 1 and 2. The single crystal X-ray
structure of 2 showed an ‘N6’ coordination sphere typically formed
for [Fe(bbp)2]2+ complexes. The iron–nitrogen bond lengths corre-
sponded to an LS complex. In addition, as both tetrafluoroborate
and perchlorate anions are weakly coordinating tetrahedral anions,
it is reasonable to assume that 1 has an isomorphous structure to 2.
UV–Vis spectroscopy of the Fe(II) complexes displayed an MLCT
band that is slightly red-shifted by about 8 nm with respect to
the discrete [Fe(bbp)2]2+ complex. Through the use of rR, aided
by TD-DFT calculations, it was determined that the phosphazene
did not contribute to this transition, rather it originated from the
[Fe(bbp)2]2+ core. In addition, the MLCT band extinction coefficient,
e, decreased as the sample was heated in solution, and although
[13] P. Vandersluis, A.L. Spek, Acta Crystallogr., Sect. A 46 (1990) 194.
[14] O.V. Dolomanov, CHN, first ed., The University of Nottingham, Nottingham,
2004.
[15] N.J. Lundin, P.J. Walsh, S.L. Howell, J.J. McGarvey, A.G. Blackman, K.C. Gordon,
Inorg. Chem. 44 (2005) 3551.
typical for SCO, analysis of this was complicated by a series of equi-
libria associated with solvent coordination. The combination of
variable temperature solid-state rR, magnetic and Mössbauer spec-
troscopy confirmed the crystallographic data, that at low tempera-
tures the complexes are LS but beyond 300 K they crossover to HS.
This investigation proves that the addition of a phosphazene to a
SCO moiety does not prevent SCO and provides a new route for
the creation of workable SCO materials based on the polyphospha-
zene scaffold.
[16] S.L. Howell, B.J. Matthewson, M.I.J. Polson, A.K. Burrell, K.C. Gordon, Inorg.
Chem. 43 (2004) 2876.
[17] R. Horvath, K.C. Gordon, Coord. Chem. Rev. 254 (2010) 2505.
[18] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,
O.K.Y. Honda, H. Nakai, T. Vreven, J.A. Montgomery, Jr., J.E. Peralta, F. Ogliaro,
M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M.
Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth,
P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman,
J.V. Ortiz, J. Cioslowski, D.J. Fox, GAUSSIAN 09, in, Gaussian, Inc., Wallingford, CT,
2009.
[19] R. Dennington, T. Keith, J. Millam, GaussView, in: Semichem Inc., Shawnee
Mission, KS, 2009.
[20] O. Alver, C. Parlak, M. Senyel, J. Mol. Struct. 923 (2009) 120.
[21] Y. Tantirungrotechai, K. Phanasant, S. Roddecha, P. Surawatanawong, V.
Sutthikhum, J. Limtrakul, J. Mol. Struct. 760 (2006) 189.
[22] A.W. Addison, S. Burman, C.G. Wahlgren, O.A. Rajan, T.M. Rowe, E. Sinn, J.
Chem. Soc. Dalton (1987) 2621.
[23] W. Linert, M. Konecny, F. Renz, J. Chem. Soc. Dalton (1994) 1523.
[24] E.W. Ainscough, H.R. Allcock, A.M. Brodie, K.C. Gordon, M.D. Hindenlang, R.
Horvath, C.A. Otter, Eur. J. Inorg. Chem. (2011) 3691.
Acknowledgements
We thank the Massey University Research Fund for financial
support and a PhD Scholarship (to R.J. D.) We also thank Associate
Professor S.G. Telfer and the MacDiarmid Institute for Advanced
Materials and Nanotechnology for the use of the Spider Diffractom-
eter, Dr M. Lein for the use of the double helix super computer and
assistance with calculations, Associate Professor A.C. Partridge for
assistance with calculations and G. Weal (University of Otago) for
help in collecting Mössbauer data. K.S. M. acknowledges the sup-
port of an Australian Research Council Discovery Grant.
Appendix A. Supplementary data
[25] R. Horvath, C.A. Otter, K.C. Gordon, A.M. Brodie, E.W. Ainscough, Inorg. Chem.
49 (2010) 4073.
[26] C. Bhaumik, S. Das, D. Saha, S. Dutta, S. Baitalik, Inorg. Chem. 49 (2010) 5049.
[27] P. Gütlich, E. Billon, A.X. Trautwein, Mössbauer Spectroscopy and Transition
Metal Chemistry, Springer Verlag, Heidelberg, 2011.
CCDC 904683 and 904684 contains the supplementary crystal-
lographic data for this paper. These data can be obtained free of
charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html, or
from the Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336 033; or e-mail: depos-
it@ccdc.cam.ac.uk. Supplementary data associated with this article
[28] J.A. Kitchen, N.G. White, C. Gandolfi, M. Albrecht, G.N.L. Jameson, J.L. Tallon, S.
Brooker, Chem. Commun. 46 (2011) 6464.