Ligand chain length conveys thermochromism

Article abstract of DOI:10.1039/c4dt00769g

Thermochromic properties of a series of non-ionic copper compounds have been reported. Herein, we demonstrate that Cu(ii) ion with straight-chain primary amine (A) and alpha-linolenic (fatty acid, AL) co-jointly exhibit thermochromic properties. In the current case, we determined that thermochromism becomes ligand chain length-dependent and at least one of the ligands (A or AL) must be long chain. Thermochromism is attributed to a balanced competition between the fatty acids and amines for the copper(ii) centre. The structure-property relationship of the non-ionic copper compounds Cu(AL) ₂(A)₂ has been substantiated by various physical measurements along with detailed theoretical studies based on time-dependent density functional theory. It is presumed from our results that the compound would be a useful material for temperature-sensor applications. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4dt00769g

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Ligand chain length conveys thermochromism†
Mainak Ganguly,a Sudipa Panigrahi,a K. R. S. Chandrakumar,b Anup Kumar Sasmal,a
Cite this: Dalton Trans., 2014, 43,
11624
Anjali Palc and Tarasankar Pal*a
Thermochromic properties of a series of non-ionic copper compounds have been reported. Herein, we
demonstrate that Cu(II) ion with straight-chain primary amine (A) and alpha-linolenic (fatty acid, AL) co-
jointly exhibit thermochromic properties. In the current case, we determined that thermochromism
becomes ligand chain length-dependent and at least one of the ligands (A or AL) must be long chain.
Thermochromism is attributed to a balanced competition between the fatty acids and amines for the
copper(II) centre. The structure–property relationship of the non-ionic copper compounds Cu(AL)2(A)2
has been substantiated by various physical measurements along with detailed theoretical studies based
on time-dependent density functional theory. It is presumed from our results that the compound would
be a useful material for temperature-sensor applications.
Received 14th March 2014,
Accepted 5th June 2014
DOI: 10.1039/c4dt00769g
www.rsc.org/dalton
regard to their unusual properties such as thermochromic be-
havior; and selected copper(II) ethylenediamine complexes are
Introduction
Thermochromism is the reversible color change of a substance very interesting in this context.9,10 Thermochromism of ionic
that occurs with a change in temperature. This phenomenon Cu(II) species has been studied extensively for halide
occurs in solids1 and liquids.2 Continuous change in band compounds.4,11–13 Other ionic compounds of Cu(II) for which
position and/or absorption edge1,3,4 and discontinuous5 phase thermochromism has been studied in depth are with BF4− and
− 14–18
transition are well reported in the literature for observed the ClO4
.
Grenthe et al. have reported that bis-(N,N-diethyl-
thermochromic properties. Molecular-based inorganic solids ethylenediamine) copper(II) perchlorate is thermochromic
whose physical properties can be controlled by external per- while N-diethylethylenediamine copper(II) perchlorate is not.9
turbation are of particular interest.1,6,7 The study of phase Such anomalous features are of special interest in this field.
transitions provides an excellent opportunity to investigate
factors that affect compound properties without any perma- with the change in coordination geometry19 or change in the
nent change in the main skeletal arrangement.
effective ligand field strength of a coordinated ligand.20 Ther-
In most inorganic systems, thermochromism is associated
However, it is often the case that change in electronic struc- mally sensitive tetragonal distortion in the copper atom
ture vis-à-vis thermochromism due to thermal effects occurs at environment3 or change of the in-plane Cu–N coordination9,17
very low temperatures,8 and this limits applications of such are often considered to be mechanisms of thermochromism.
compounds. Hence, for practical purposes, synthesizing Lever et al.4,21 have suggested a temperature-dependent tetra-
materials for which thermochromic change occurs at higher gonal distortion of the Cu(II) coordination centre with an
temperatures is useful. In recent years a great deal of attention accompanying change in the d-orbital energies as the probable
has been focused on complexes formed among diamine, mechanism for thermochromism. Again, Gill and Nyholm
copper(II), and other transition metal salts, particularly with stated that the green to blue transition at higher temperatures
resulted from the decrease in solvation of the chloride ion sur-
rounding the metal atom.22 The thermochromism phenom-
enon may also occur due to alteration of the coordinated water
aDepartment of Chemistry, Indian Institute of Technology, Kharagpur 721302, India.
molecule.23
E-mail: tpal@chem.iitkgp.ernet.in
bTheoretical Chemistry Section, Chemistry Group, Bhabha Atomic Research Centre,
Because of the attention-grabbing overlap of solid-state and
inorganic chemistries, thermochromism of inorganic
Mumbai 400085, India
cDepartment of Civil Engineering, Indian Institute of Technology, Kharagpur 721302,
materials is studied extensively. Thermochromic materials are
widely used in the manufacture of color indicators for temp-
erature change, temperature-sensitive light filters, imaging
systems, smart dyes, optical switching, and so on.24–27 Greater
understanding of the various factors that influence color
changes is also very important.
India
†Electronic supplementary information (ESI) available: Synthesis of the thermo-
chromic compound, reflectance and absorption spectra, TD-DFT-simulated
absorption spectra, mass spectra, 1H NMR spectra. CCDC 859831. For ESI and
crystallographic data in CIF or other electronic format see DOI: 10.1039/
c4dt00769g
11624 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
For purely organic nitro compounds, the effect of alkyl (UK) equipped with an Mg Kα excitation source (1253.6 eV)
1
chain length on thermochromism was reported by Lee et al.28 and a five-channeltron detection system. The H NMR spectra
But for non-ionic copper compounds, investigation of such were recorded with a Bruker AC 200-MHz NMR Spectrometer.
phenomena is fairly new. Using straight-chain amines (A) and
Details of theoretical methods and models
straight chain fatty acids (AL), we have discovered a series of non-
ionic thermochromic copper compounds rendering violet/blue All theoretical calculations were performed with ORCA system
reversible color change with temperature. The thermochromic programs.29 The geometry of the copper complexes with acetic
behavior of the non-ionic compounds Cu(II)(AL)2(A)2 has been acid ligand as well as with amines of C10-decyl chain (DA) was
observed to be intriguingly related with the chain length of the optimized, without symmetry constraints. For the case of
ligands (fatty acids and amines). It has been recognized that closed- and open-shell systems, the RHF and ROHF types of
the minimum, necessary condition is to have a long chain of Kohn–Sham methods were applied, respectively. Herein, the
any one type of the participating ligands connected to the def2-TZVPP basis set was employed for copper and def2-SVP
Cu(II) ion center. The structure–property relationships of the used for all other atoms. Geometry optimization and time-
non-ionic compounds Cu(AL)2(A)2 (AL = alkanoate, A = amine) dependent density functional theory (TDDFT)-based calcu-
has been substantiated via DRS, UV-vis, 1H NMR, IR, mass lations30,31 for evaluating the electronic excitation spectrum
spectroscopy, elemental analysis and X-ray analysis. Further- were performed using density functional theory (DFT) based
more, the phenomenon has been interpreted theoretically. on BP86 exchange–correlation functional with Becke’s three-
Well-known copper(II) acetate or copper(II) stearate and long- parameter exchange functional and the Lee–Yang–Parr corre-
chain amines like hexadecylamine (HDA), tetradecylamine lation functional (B3LYP).32,33 The grid-based DFT was used in
(TDA), and dodecylamine (DDA), among others, were used in ORCA, which employs a typical grid quadrature to compute
our study.
the integrals. During the SCF procedure, the grid provides the
option “Lebedev 302 points.” The charge on each atom was
obtained via Mulliken population analysis. Regarding the effect
of temperature, as explained in the Experimental section,
herein we assumed that the copper–amine distance in the
complex is increased systematically due to relatively weaker
Experimental section
Chemicals, materials, and equipment
All reagents used were of AR grade. Copper acetate and stearic interaction compared to that of carboxylate group interaction
acid were obtained from Merck. All straight-chain amines were with the copper ion. Accordingly, we increased the Cu–N
obtained from Aldrich. All solvents used (methanol and heptane) distance by 0.2 Å at each step until they fall apart by 1 Å from
were of AR grade and obtained from Sisco Research Laboratory, their equilibrium geometry; these distorted geometries were
India, and were distilled before use. Triple distilled water was used for evaluating the electronic absorption spectra of these
used throughout the experiment. All glassware was cleaned with complexes.
freshly prepared aqua regia, subsequently rinsed with copious
amounts of distilled water and dried well before use.
Synthesis of copper stearate
The absorption spectra were recorded in a Perkin Elmer Long-chain copper alkanoates, such as copper stearate, were
Lambda 35 UK-vis spectrophotometer, and the solutions were synthesized from copper(II) acetate.34 Two solutions – 300 mg
taken in a 1 cm, well-stoppered quartz cuvette. The spectra of stearic acid dissolved in 20 mL methanol and 42 mg copper
were recorded against a reference solvent in a reference cell in acetate dissolved in 10 mL methanol – were mixed together
the double-beam spectrophotometer to subtract solvent back- under boiling condition and a greenish blue precipitate
ground. Elemental analyses have been done with a Perkin (insoluble in methanol) was obtained from the boiling reac-
Elmer CHN 2400. Fourier transform infra-red (FTIR) spectral tion mixture. The precipitate was washed thoroughly with hot
characteristics of samples were obtained in transmittance methanol in order to free it from the starting materials, copper
mode with a Perkin Elmer FTIR SPECTRUM 2000 instrument. acetate or stearic acid (both are soluble in methanol). The
The FTIR spectrum was recorded for over 512 scans of each greenish blue precipitate was dried (yield 93%) and character-
sample and the background spectrum was automatically sub- ized34 via powder XRD pattern, TLC, melting point (80 °C) and
tracted. The temperature control was achieved using the FTIR analysis to confirm sample purity.
GRASBEY SPECAC automatic temperature controller. Reflec-
tance spectra were measured using DRS (diffuse reflectance
Synthesis of thermochromic compound
spectra) mode with the Cary model 5000 UV-vis-NIR spectro- The amine complexes of the Cu(II) stearate were synthesized in
photometer. The single-crystal data were collected on Bruker- heptane under reflux conditions; 20 mg copper(II) stearate and
Nonius Mach3 CAD4 and CCD APEX-2 X-ray diffractometers 14 mg tetradecylamine amine (TDA) were added in 10 mL
that use graphite monochromated MoKα radiation (μ = heptane and refluxed for 30 min. The solution turned blue
0.71073 Å). The X-ray diffraction (XRD) patterns were recorded during the reflux. Upon cooling, a violet solid [Cu(St)2(TDA)2]
in an X’pert Pro Diffractometer with Cu (Kα = 1.54056) radi- gradually precipitated out from heptane. The violet solid (yield
ation. X-ray photoelectron spectroscopy (XPS) analysis was 94%) was then filtered, washed several times with heptane,
carried out with a VG Scientific ESCALAB MK II spectrometer and dried under vacuum.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 11624–11636 | 11625

View Article Online
Paper
Dalton Transactions
In similar fashion, all other compounds of copper alkano- recognized that a minimum requirement for thermochromism
ates, AL (acetate, propionate, etc.) were synthesized using is long-chain amines or long-chain fatty acids, that is, at least
amines (A) of various chain lengths.
any one of these two participating ligands should bear a long
To grow the single crystal, solid crude [Cu(Ac)2(TDA)2] product alkyl chain. Straight-chain primary amine (A) and fatty acid
(100 mg) was dissolved in 8 mL dichloromethane (DCM) and (AL) bind to the Cu(II) centre simultaneously to generate
kept in a test tube. Then, heptane was slowly added through the thermochromism. When both the ligands (fatty acids and
side of the test tube. As a result, a blue-colored DCM layer at the amines) are long, excellent pale violet-colored thermochromic
bottom and colorless heptane layer on DCM were observed. The compound results. Short-chain primary amines, including
test tube was then sealed with parafilm to allow slow evaporation ammonia, can also form non-ionic thermochromic com-
and kept undisturbed. After ∼15 days, a violet crystalline solid pounds but only with long-chain copper alkanoates such as
was obtained in colorless heptane. This crystalline solid was iso- palmitate, stearate (St), and so on. On the other hand,
lated and used for single crystal analysis.
short-chain amines, such as ammonia, methylamine, and ethyl-
amine, cannot form thermochromic compounds with short-
chain fatty acids; instead they produce non-thermochromic
deep blue tetraamine compounds.35 Lower alkanoates need
long-chain amines for engendering thermochromism. So, at
Results and discussion
The synthetic strategy yields a series of non-ionic thermo- least one of the two kinds of ligands has to bear a long
chromic compounds with a general formula Cu(II)(AL)2(A)2. chain to generate thermochromism (Scheme 1). The violet/
They exhibit violet/blue reversible color change with tempera- blue color reversal of all thermochromic compounds was
ture. For the first time, it is shown that thermochromism of exceptional and they showed no sign of degradation even after
Cu(II) complex can be ligand chain-length dependent. It is also 35 cycles.
Scheme 1 Ligand chain length and thermochromism.
11626 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Fig. 1 Temperature-dependent absorption spectra of the three representative thermochromic molten compounds (b–d) indicating red shifting
(>100 nm) of the λmax values compared to RT (25 °C) absorption spectra (a).
Thermochromic violet solids are obtained by refluxing change is not due to the solid–liquid phase change, but likely
copper(II) alkanoate and amine in heptane in a stoichiometric due to the change of geometry.
ratio of 1 : 2. The as-prepared violet non-ionic thermochromic
compounds are readily soluble in CHCl3, DCM, etc., with excel- chloroform as mentioned earlier. The UV-vis spectra (Fig. 2) in
lent blue color.9,36 But, the copper alkanoates are insoluble. chloroform
for Cu(St)2(DDA)2, Cu(St)2(TDA)2, and
Thus the selective solubility of the product helps to monitor Cu(St)2(HDA)2 show the absorption maxima at 645, 617, and
the progress of the reaction and isolation of the pure product. 601 nm at room temperature. In the case of the corresponding
It is worth mentioning that when the violet thermochromic acetate compounds [Cu(Ac)2(DDA)2, Cu(Ac)2(TDA)2,
Thermochromic compounds produce a blue solution in
compound is heated above a certain temperature, TBs, it turns Cu(Ac)2(HDA)2], the blue shift of the λmax values (628, 620, and
dark green and permanently loses its thermochromic property. 615 nm) with the increase of amine ligand chain length is
The reversible violet/blue color change is observed only below gradual. This is abrupt in the stearate cases at room tempera-
this TBs. Upon heating the violet solid (Fig. S1, ESI†) above its ture. The compounds with increased amine chain length
melting point (m.p.), a blue liquid is obtained. The blue liquid exhibit a stronger +I effect, rendering higher nucleophilicity.
transforms back to a violet solid under ambient temperature. This effect enhances the extent of overall d–d energy split-
At room temperature, above m.p., and above TBs, the represen- ting,37 resulting in the observable blue shift of the UV spectra.
tative thermochromic compounds show their absorption
In the mass spectrometer, the compounds show peaks
maxima at ∼540 nm, ∼660 nm, and ∼700 nm, respectively corresponding to the fatty acid (AL) and amine (A) ligands.
(Fig. 1). Red shift of the λmax values of the thermochromic Mass spectroscopy of the compound Cu(St)2(DDA)2 gives peaks
compounds with the increase in temperature is summarized in at 284, 227, and 115 due to stearate,38 and a very intense peak
Table 1. Surprisingly, the blue color is observed in the solution at 186 due to dodecyl amine. The peak at 145 (Cu + 2CH3CN)
or liquid phase while the violet color is observed for solids only. is indicative of copper. The mass spectrum gives information
The violet color is never observed in the liquid phase, indicating about the presence of copper, stearic acid and dodecylamine
that the violet color is characteristic of the thermochromic solid. (Fig. S2, ESI†). The molecular ion peaks, however, are not
An interesting exception is that the violet compound of obtained for the compounds, presumably due to the stepwise
copper(II) propionate/DDA melts to a blue liquid at 67 °C and loss of the coordinated ligands in the spectrometer. The
retains its blue color even in the solid state at room tempera- 1H NMR data (in CDCl3) supports AL : A = 1 : 1 in the thermo-
ture. No thermochromism is exhibited by the blue solid. But, chromic compounds (Fig. S3 and S4, ESI†). The elemental ana-
the blue solid turns violet and again brings back thermo- lyses and non-ionic nature of the thermochromic compounds
chromism if it is cooled to 0 °C. This compound is more
rigid and different from other thermochromic compounds of
Cu(II)-alkanoates. So, this compound confirms that the color
Table 1 Physical characteristics of Cu(St)2(A)2 compounds
λmax at different temperature (nm)
Compounds
m.p. (°C) TBs (°C) 25 °C 90 °C 120 °C 150 °C
Cu(St)2(HDA)2 93
Cu(St)2(TDA)2 83
Cu(St)2(DDA)2 79
133
128
134
538
532
534
647
638
653
665
654
667
715
709
692
Fig. 2 UV-vis absorption spectra of the representative compounds
in CHCl3.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 11624–11636 | 11627

View Article Online
Paper
Dalton Transactions
indicated that two fatty acids and two straight-chain ligands distorted geometry that deviates from the regular square pyr-
bonded to copper. Analysis of [Cu(St)2(DDA)2], [Cu(St)2(TDA)2], amidal one. The copper centre is mononuclear and sur-
[Cu(St)2(HDA)2] was done in the CHN analyzer and copper ana- rounded by the ligand atoms. Two amine N’s are in the
lysis carried out titrimetrically.39 The calculated values of the mutual trans position. Of the two alkanoates, one binds to the
C, H, N, and Cu percent corresponding to [Cu(St)2(DDA)2], copper center in a bidentate and the other in a monodentate
[Cu(St)2(TDA)2], [Cu(St)2(HDA)2] match well with the observed fashion. The solid compound forms a hydrogen-bonded layer
values of the complexes: found – C, 71.76%; H, 12.69%; structure (Fig. 3(c)), with the aliphatic chains hanging above
N, 2.99%; Cu, 6.47% requires C, 72.00%; H, 12.41%; N, 2.82%; and below the layer. Consequently, the compound formula is
Cu, 6.35% for [Cu(St)2(DDA)2]; found – C, 72.81%; H, 12.62%; Cu(Ac)2(TDA)2 [diacetatobis(tetradecylamine)copper(II)]. One
N, 2.67%; Cu, 5.95% requires C, 72.72%; H, 12.54%; N, 2.66%; water molecule helps to grow the single crystal. However, the
Cu, 6.01% for [Cu(St)2(TDA)2]; and found – C, 73.81%; anhydrous moiety is also thermochromic like the crystalline
H, 12.70%; N, 2.77%; Cu, 5.76% requires C, 73.45%; H, 12.6%; product. Single crystal could not be obtained for the thermo-
N, 2.52%; Cu, 5.71% for [Cu(St)2(HDA)2]. The data from chromic compounds where both the ligands (AL and A)
elemental analysis for acetate amine compounds also agree possess long chains. Consequently, it cannot necessarily be
well with the proposed structure.
assumed that the crystal structure of Cu(Ac)2(TDA)2 is repre-
Single crystal XRD (Fig. 3 and Table 2) revealed that the sentative of all compounds under investigation. However,
copper ion shows square pyramidal geometry having two elemental analyses and 1H-NMR support the molecular
straight-chain amine and two alkanoate groups. But this has a formula Cu(AL)2(A)2 of all compounds. Powder XRD (Fig. 4(a))
indicates that the compounds bear both the character of
alkanoate (AL) and amine (A). Fig. 4(b) shows that at higher
temperatures, the intensity of low angle peaks are decreased
and merged into a broader peak, indicating the destruction of
the layer structure. Peaks due to amines (θ 20°–30°) remain
unaltered even at higher temperatures.
Fig. 5 shows that the thermochromic compounds do not
undergo weight loss in the temperature range of 80–130 °C
where they show thermochromism. All three Cu(St)2(A)2 com-
pounds degrade oxidatively in two steps showing two slope
changes around 280 and 325 °C (evident from TG and DTG
plots). At 280 °C, two amine molecules are lost (∼40% weight
loss). After that, the stearate molecules are lost, one after the
other. A sharp endothermic peak in the DTA plot at ∼83 °C is
due to phase change of the thermochromic compounds. The
increased chain length of the amines in the compounds
(higher mol wt) is responsible for the increased melting point.
Interestingly, there is a rise of 6–9 °C in the melting point with
the increase in the chain length of the participating amines by
one –C2H5 unit (Table 3).
At room temperature, two peaks in the EPR spectra are
obtained (g = 2.043, 2.22) for Cu(Ac)2(TDA)2. With the increase
in temperature above its m.p., only one peak appears with
g value of 2.116 due to larger Cu–amine distance. The signal
constantly decreases with the increment of temperature
(Fig. 6). Depending on the energy gap between the ground
state and the first excited state in the Cu(II) unit, a noteworthy
decrease in spectral intensity is observed due to saturation
effects as reported by Winter et al.40 This effect is completely
reversible below the temperature TBs, indicating that the com-
pound is also stable at higher temperatures without any signs
of decomposition.
Fig. 3 (A) Complete single crystal structure. (B) Zoom crystal structure.
(C) Packing diagram of Cu(Ac)2(TDA)2·H2O.
Table 2 Crystallographic parameters for crystal structure of
Cu(Ac)2(TDA)2 compound (CCDC deposit no. 859831)
Formula
Molecular weight
T (K)
System
Space group
a (Å)
b (Å)
c (Å)
α (°)
β (°)
C32H70CuN2O5
626.44
293(2)
Triclinic
ˉ
P1
8.519(2)
9.763(3)
22.676(6)
93.901(8)
94.815(9)
99.624(9)
1846.2(8)
2
Temperature-dependent FTIR spectra (Fig. 7) of all three
representative Cu(St)2(A)2 indicate that the difference
γ (°)
V (Å3)
(<200 cm−1) between the asymmetric (νasCOO−, ∼1540 cm−1
)
and symmetric (νsymCOO−, ∼1340 cm−1) stretching frequency
Z
Dcalc (Mg m−3
R1 (I > 2σ(I))
)
1.127
0.0563
0.1760
of the COO− functionality at low temperature (below m.p.)
wR2(on P2, all data)
increases (>200 cm−1
) as the temperature is increased
11628 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Fig. 4 (a) Powder XRD of representative thermochromic compounds, Cu(St)2 and TDA at 25 °C; target copper. (b) Temperature-dependent powder
XRD of Cu(St)2(TDA)2; target cobalt.
Fig. 5 (a) TG plots of representative compounds. (b) TG/DTG/DTA plots of Cu(St)2(TDA)2.
Table 3 Melting point and TBs of representative thermochromic
compounds
Name
Melting point (°C)
TBs (°C)
Cu(Ac)2(DDA)2
Cu(Ac)2(TDA)2
Cu(Ac)2(HDA)2
Cu(St)2(DDA)2
Cu(St)2(TDA)2
Cu(St)2(HDA)2
83
88
95
79
83
93
134
138
138
134
128
133
(above m.p.) for the change from bidentate attachment of one
of the COO− functionality to monodentate.41 Thus heating pro-
duces monodentate binding of both COO− functionalities,
which eventually induce steric crowding. As a consequence,
–NH2 groups move apart from the Cu(II) centre by taking
thermal energy. Thus the –NH2 stretch at 3330 cm−1 becomes
prominent (like free –NH2 functionality36,42) above m.p. An
Fig. 6 Temperature-dependent EPR spectra of thermochromic com-
pound Cu(Ac)2(TDA)2.
increase in temperature causes a weakening of the Cu–N temperature. It can be spelled out in this context that the
bond43 and this weakening causes a perturbation44 in the short-chain amine or ammonia binds strongly to the Cu(II)
coordination sphere. Interestingly, at a temperature above TBs center containing short-chain fatty acids. Then the fatty acids
the amine stays apart and its re-joining the copper centre is remain outside the coordination zone forming ionic com-
hindered even when the temperature drops down to room pounds,35,45 and the thermochromic property is not observed.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 11624–11636 | 11629

View Article Online
Paper
Dalton Transactions
Fig. 7 Temperature-dependent FTIR spectra of copper compounds (a) Cu(St)2(HDA)2, (b) Cu(St)2(TDA)2 and (c) Cu(St)2(DDA)2. Black arrow indicates
the N–H stretch of amine (∼3330 cm−1), prominent at higher temperatures.
The absorption spectra supports relating the color change
The XPS analysis for all the copper compounds under
of the thermochromic compounds with the increase in Cu(II) investigation (the peaks at ∼933 eV and ∼953 eV, respectively,46
centre to amine distance with rising temperature. The λmax are responsible for 2p3/2, 2p1/2 of Cu(II); Fig. 8(b)) authenticates
value of Cu(II) alkanoate in THF blue shifts while long-chain the +2 oxidation state of copper. Interestingly, greenish-blue
amines coordinate (i.e., for thermochromic compounds). Again, Cu(St)2 shows a pronounced 933.52 eV peak. Upon TDA incor-
the blue-shifted λmax again red shifts when the compound is poration into Cu(St)2 (i.e., for thermochromic compounds),
dissolved in THF after being heated above TBs (Fig. 8(a)).
the 2p3/2 peak shifts to 934.10 eV. The heated thermochromic
Fig. 8 (a) UV-vis spectra in tetrahydrofuran of Cu(St)2, Cu(St)2(TDA)2 below m.p., and Cu(St)2(TDA)2 above TBs. (b) XPS spectra Cu(St)2, Cu(St)2(TDA)2
before melting, and Cu(St)2(TDA)2 above TBs for the element copper.
11630 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Scheme 2 Copper(II) is bonded to two long-chain fatty acids and amines through O and N donors, respectively. A balanced competition between
fatty acids and amines exhibiting thermochromism.
compound above the TBs shows the XPS peak shift from
934.10 eV towards a lower binding energy, indicating that the
thermochromic compound above TBs is closer to pure Cu(St)2.
This means amine is weakly interacting with the cooper
alkanoate above TBs.
Above TBs, the compounds become green and spectral fea-
tures (UV, Fig. 8(a), and XPS, Fig. 8(b)) are akin to Cu(II)-
alkanoates (devoid of amine). Structurally, these amines move
out from Cu(II) centre and find no way of return to it.
In brief (Scheme 2), copper(II) is bonded to two straight-
chain fatty acids and amines through O and N donors, respect-
ively. After heating, both the fatty acids probably become
monodentate and they come closer to Cu(II) centre. As a result
of steric crowding, the amines move apart from Cu(II) and the
complex becomes blue. At very high temperatures (above TBs),
N’s of the amines move further away from the Cu(II) centre
permanently.
Let us now discuss the TDDFT results on the thermo-
chromism properties of the selected compound. The opti-
mized geometry of the Cu(acetate)2(C10-decylamine)2, as
obtained by the DFT-based BP exchange–correlational func-
Fig. 9 Optimized geometry of (a) Cu(Ac)2(DA)2 and (b) Cu(Ac)2(methyl-
amine)2 compounds.
tional, is shown in Fig. 9; the calculated geometrical para-
meters, particularly the Cu–O and Cu–N of acid and amine
ligands, respectively, clearly demonstrate the correspondence
with experimental results and computations based on theory.
For instance, the computed bond distances of Cu–N as well as
Cu–O are 2.007, 2.044, and 1.977, 1.987 Å, respectively,
whereas the corresponding experimental values are 2.008,
1.991, and 1.971, 1.986 Å, respectively. In addition, one of the
carboxyl groups interacts with the copper ion through the
slightly elongated bidentate mode, and the interaction of other
carboxyl groups is monodentate. It is gratifying to note that
this general feature inferred from the present experimental
study is also captured by our computational results.
From the above discussions on the characterization of
thermochromic compounds via diverse physical techniques,
we can unambiguously state that the copper(II) ion and amine
distance gradually increases as temperature increases. In
addition, it is also evident that when the temperature of the
system is increased beyond TBs, the amine molecules leave the
copper ion and only the acid ligands are bound to the copper
ion. Based on this observation, in order to consider the effect
of temperature on the thermochromic systems, we gradually
increased the Cu–N distance by 0.2 Å until reaching 1 Å. These
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 11624–11636 | 11631

View Article Online
Paper
Dalton Transactions
Fig. 10 (a) TD-DFT-simulated absorption spectra of model complex Cu(Ac)2(DA)2. (b) TD-DFT-simulated absorption spectra of model complex
Cu(Ac)2(methyl amine)2.
11632 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
distorted geometries were considered for the simulation of the length is arising due to the fact that the thermal energy sup-
absorption spectra using the TDDFT method. plied to the system weakens the Cu–amine bond and thereby
The absorption spectrum calculated via TDDFT for the the bond length increases.
Cu(Ac)2(DA)2 compound at equilibrium as well as several other
By inspecting Table 4, it can be seen that for the equili-
distorted geometries is presented in Fig. 10(a) and (b). The cal- brium geometry of Cu(Ac)2(DA)2, the excitation spectra arises
culated dipole moment, charge on copper ion and the charge due to electronic transitions from βHOMO−6 and βHOMO−4
population at dx2−y2, oscillator strengths, maximum absorption to βLUMO, and analysis of these frontier orbitals indicates that
wavelengths (λ) along with their orbital assignments for the the electron density is transferred predominantly from the
complexes Cu(Ac)2(DA)2 and Cu(Ac)2(methylamine)2 are pre- ligand donor atoms, oxygen of acid ligands, and nitrogen of
sented in Tables 4 and 5, respectively. For the complex amine ligands to the copper ion, respectively. It implies that
Cu(Ac)2(DA)2 at equilibrium geometry, the spectra shows two the peaks result from the ligands-to-metal charge transfer
major peaks at 260.6 and 311.5 nm, corresponding to the exci- complexes. Interestingly, it can also be observed that the
tation of HOMO−10 and HOMO−8 to LUMO of the complex, second peak (βHOMO−4 to βLUMO) is mostly associated with
respectively. These two peaks undergo red shift systematically the dominant charge transfer from oxygen rather than nitro-
from 260.6 to 735.0 nm as the copper ion and amine ligand gen donor species. However, for the first peak (βHOMO−6 to
distance is gradually increased from 2.007 to 3.007 Å. It is βLUMO), the charge transfer originates from both the nitrogen
worthwhile noting that though such red shifts are also as well as oxygen donor species along with few percent from
observed for the case of Cu(Ac)2(methylamine)2, the shifts in the copper ion as well. This is illustrated in terms of the iso-
the maximum wavelength is only 245 nm whereas the corres- density molecular orbital plots as presented in Fig. 11. For the
ponding shift for decyl amine complex is 475 nm. It should be case of complexes distant from the equilibrium position,
noted that the experimentally observed λmax of copper complex another dominant peak with the largest significant oscillator
with the DDA ligand is 534, 653, 667, and 692 nm at 25, 90, strength starts emerging at higher wavelengths than that of
120, and 150 °C temperatures, respectively. Interestingly, the previous wavelengths. Close inspection of the results reported
experimentally observed red shift at higher wave length in Table 4 indicates that these peaks are attributable to the
(692 nm) closely matches the computed results (∼735 nm) for charge transfer from βHOMO−4, βHOMO−2, to βLUMO for the
the case of distorted Cu–amine thermochromic compound. geometries with the fixed Cu–N distance 2.607, 2.807, and
These results clearly demonstrate that the copper(II) com- 3.007 Å, respectively. These highly red-shifted peaks are the
pound with the shorter amine and acid chain length may not result of the charge transfer mostly from nitrogen atoms of the
be suitable for the observed thermochromic properties. It may amine groups (except for the complex with Cu–N distance of
be noted that the observed red shift in the absorption wave- 2.807 Å), while the oxygen from the carboxyl groups’ contri-
Table 4 Dominant optical transition characters for lowest excitation states of Cu(Ac)2(DA)2 complex, calculated by TDDFT-based B3LYP/def2-
DZVPP (TZVP for Cu) method
Cu–N distance
(Å)
Oscillator
strength
Maximum
wavelength λ (nm)
Dipole moment
(Debye)
Charge on
copper ion (a.u.)
Charge population
in dx2−y2 orbital (a.u.)
2.007
2.207
2.407
2.607
2.807
3.007
0.1860
0.2044
0.2024
0.1503
0.1535
0.1589
260.6
314.2
384.3
467.7
617.1
735.0
2.827
2.752
2.636
2.504
2.387
2.297
0.1348
0.2263
0.2743
0.3076
0.3377
0.3639
1.8021
1.7629
1.7443
1.7468
1.7619
1.7812
Cu–N distance
(Å)
Oscillator
strength
Maximum
wavelength λ (nm)
Orbital transitions and
weightage parameter
State
Atomic contributions
2.007
2.207
2.407
2.607
0.1860
0.1263
0.2044
0.0707
0.2024
0.0296
0.0803
0.0140
0.1535
0.0623
0.0515
0.1535
0.0451
0.0274
0.1589
260.6
311.5
314.2
316.9
384.3
422.9
461.4
488.1
467.7
512.0
561.8
617.1
556.3
634.6
735.0
10
8
10
8
10
8
10
8
9
10
8
7
10
8
128B–135B = 57%
130B–135B = 88%
128B–135B = 82%
130B–135B = 62%
129B–135B = 79%
131B–135B = 45%
129B–135B = 39%
131B–135B = 51%
130B–135B = 39.2%
130B–135B = 51%
131B–135B = 37.4%
132B–135B = 45.2%
130B–135B = 37%
131B–135B = 37.6%
132B–135B = 45.2%
13.2% Cu; 24.7% N; 16.5% O
5.9% Cu; 1.9% N; 75.2% O
33% Cu; 7.8% N; 31.9% O
6.9% Cu; 3.0 N; 69.0% O
12% Cu; 31% N; 32.4% O
8% Cu; 0.6% N; 84.2% O
11.3% Cu; 14.6 N; 56% O
6.8% Cu; 0.2 N; 88% O
7.2% Cu; 52.4% N; 15.3% O
10.9% Cu; 12.4% N; 58.1% O
6.3% Cu; 1.3% N; 88.2% O
5.6% Cu; 30.8% N; 37.4% O
18.8% Cu; 22.6% N; 38.4% O
8.3% Cu; 6.9% N; 75.9% O
5.6% Cu; 41.6% N; 19.4% O
2.807
3.007
6
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 11624–11636 | 11633

View Article Online
Paper
Dalton Transactions
Table 5 Dominant optical transition characters for the lowest excitation states of the Cu(Ac)2(methylamine)2 complex, calculated by TDDFT-based
B3LYP/def2-DZVPP (TZVP for Cu) method
Cu–N distance
(Å)
Oscillator
strength
Maximum
wavelength λ (nm)
Dipole moment
(Debye)
Charge on copper
ion (a.u.)
Charge population
in dx2−y2 orbital (a.u.)
1.993
2.193
2.393
2.593
2.793
2.993
0.1646
0.1238
0.1317
0.1258
0.1081
0.0671
314.3
265.0
322.2
392.8
477.3
559.5
2.73526
2.71833
2.69014
2.6711
2.6772
2.7042
0.1367
0.2190
0.2575
0.2865
0.3193
0.3527
1.7856
1.7467
1.7319
1.7404
1.7638
1.7872
Cu–N distance
(Å)
Oscillator
strength
Maximum
wavelength λ (nm)
State
Orbital transitions
Atomic contributions
1.993
2.193
2.393
2.593
2.793
2.993
0.1646
0.0417
0.1008
0.1238
0.0429
0.1317
0.0235
0.1258
0.0213
0.1081
0.0220
0.0671
314.3
220.8
365.9
265.0
429.1
322.2
496.7
392.8
558.6
477.3
609.1
559.5
8
15
8
11
8
10
8
10
8
50b–55b = 76%
55a–56a = 39%
50b–55b = 56%
48b–55b = 45%
50b–55b = 25%
49b–55b = 60%
50b–55b = 15%
49b–55b = 85%
52b–55b = 41%
49b–55b = 75%
52b–55b = 31%
50b–55b = 57%
71% O
25% Cu, 20.4% N, 40% O
57% O; 1.6% N; 33% Cu
46% Cu; 1.6% N; 45% O
6.0% Cu; 3.4% N; 70% O
20.1% Cu; 50% N; 19% O
7% Cu; 5% N; 68% O
6.5% Cu; 65.7% N; 9.9% O
5.5% Cu; 0.1 N; 88% O
12.8% Cu; 47.8% N; 29.5% O
5.1% Cu; 0.7% N; 89.2% O
8.1% Cu; 63.6% N; 17.8% O
10
8
10
bution is only one-third to that of nitrogen. In addition, it can to 0.353 a.u., as there is a less charge transfer from amine
also be noticed that the dipole moment is systematically ligands to the central copper ion. Similar observations were
reduced from 2.827 to 2.704 Debye, indicating the slight dis- made for Cu(Ac)2(methylamine)2. One of the primary peaks
appearance of relative trans-positions of the two amine groups. (at ∼314 nm) with significant oscillator strength (0.1646) for
Similarly, the charge on the copper ion is increased from 0.137 various Cu–N distances is attributable to the charge transfer
from oxygen of the carboxyl group exclusively with a much
smaller charge transfer from the nitrogen group. However, the
major difference is that the red shift due to the elongation of
the Cu–amine distance is significantly less than that of the
complex with the C10-decyl amine ligand groups. In summary,
the calculated transition gives rise to the energy absorption
band at 735 nm with oscillator strength 0.1589 of the complex
when the copper ion distance is 3.007 Å, which is remarkably
higher than that of the corresponding equilibrium geometry
(observed peak is at 260.6 nm with oscillator strength 0.1860).
Thus, the results obtained from the TDDFT-based methods
strongly supports the thermochromism properties of the
copper–amine complexes.
The UV absorption spectrum for the Cu(Ac)2(methyl
amine)2 and Cu(Ac)2(DA)2 complexes with one water molecule
were also evaluated and the results reveal a similar red-shift
pattern as discussed above. The corresponding spectrum for
these complexes with one water molecule is presented in the
ESI (Fig. S5 and S6†). Note also that the water molecule binds
with the acetate as well as amine ligands through hydrogen
bonding interaction (Fig. S7, ESI†).
It has been found that thermochromic compounds become
precursors for the synthesis of Cu2O nanorod (2–4 nm) in
mixed solvent (chloroform–toluene) via modified hydro-
thermal treatment (kept in a screw-capped test tube and
heated under a 100 W bulb for 6 h). Evolution of yellow solution
from blue solution indicates the formation of Cu2O organosol.
It is noteworthy that pure copper acetate/stearate under similar
Fig. 11 Isodensity molecular orbital plots of frontier orbitals (HOMO
and LUMO) as well as other lowest orbitals (isodensity value 0.03 a.u.).
11634 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Fig. 12 (A) Digital image, (B) absorption spectra, and (C) XPS spectra (for the element copper) under freeze-drying condition of Cu2O organosol
obtained from the thermochromic compounds. TEM image of Cu2O organosol obtained from the thermochromic compounds (D1) Cu(St)2(DDA)2,
(D2) Cu(St)2(TDA)2, (D3) Cu(St)2(HDA)2, (D4) Cu(Ac)2(DDA)2, (D5) Cu(Ac)2(TDA)2 and (D6) Cu(Ac)2(HDA)2. (E) HRTEM image and (F) SAED pattern of
Cu2O organosol obtained from thermochromic compounds Cu(Ac)2(TDA)2.
conditions does not produce Cu2O organosol. Such yellow- mation of thermochromic compounds for all alkanoates
colored Cu2O organosol obtained from thermochromic com- studied. Similarly, long-chain alkanoates are efficient for produ-
pounds is highly stable for months under ambient condition. cing thermochromic compounds with amine ligands of any
The peak at ∼429 nm for all the representative thermochromic chain length. Straight-chain amines and straight-chain satu-
compounds in the absorption spectra corroborates Cu2O par- rated fatty acids (but not the straight-chain thiols or unsaturated
ticle formation.47 Moreover, ∼932.80 and 952.81 eV binding fatty acids) show this remarkable property. The as-prepared
energy corresponds to Cu(I)2p3/2 and Cu(I)2p1/2, respectively.48 compounds should find applications as temperature sensors in
Finally, 0.245 nm lattice fringe for the (111) plane and SAED a reaction medium inside thin glass encapsulation as the com-
pattern confirm Cu2O synthesis (Fig. 12).49 Ultrasmall Cu2O pounds are thermodynamically very rugged. This is the first
nanorods in an organic medium, that is, as organosol, hitherto time that chain length-dependent thermochromism has been
unknown, may be a viable candidate for important applications. reported at higher temperatures, and the compounds may there-
fore find many practical sensor applications along with organo-
sol preparations.
Conclusions
We have demonstrated the design and synthesis of a series of
non-ionic thermochromic compounds of copper(II)-alkanoates
Acknowledgements
with alkyl amines. It has been shown that the reversible color The authors are grateful to the UGC, DST, NST, BRNS, and
change of non-ionic Cu(II) compounds relates to a balanced CSIR, New Delhi, India, and the Indian Institute of Techno-
competition between the incoming alkyl amine ligands with logy, Kharagpur, for financial assistance. The authors are also
the alkanoate ligand attached to a copper center. Long-chain thankful to Dr Abhishek Dey, Indian Association for the Culti-
amines were found to be excellent candidates for the for- vation of Science for temperature-dependent EPR analysis.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 11624–11636 | 11635

View Article Online
Paper
Dalton Transactions
24 J. H. Day, Chem. Rev., 1968, 68, 649–657.
25 L. Tansienhee, D. Lavabre, G. Levy and J. C. Micheau, New
J. Chem., 1989, 13, 227–233.
26 M. A. White and M. LeBlanc, J. Chem. Educ., 1999, 76,
1201–1205.
27 K. Sone and Y. Fukuda, Inorganic Thermochromism,
Springer-Verlag, Berlin, 1987.
28 S. C. Lee, Y. G. Jeong, S. H. Jang and W. H. Jo, Fibers
Polym., 2007, 8, 234–236.
29 F. Neese, The ORCA program system, Wiley Interdiscip. Rev.:
Comput. Mol. Sci., 2012, 2, 73–78.
30 T. Petrenko, S. Kossmann and F. Neese, J. Chem. Phys.,
2011, 134, 054116.
31 F. Neese and G. Olbrich, Chem. Phys. Lett., 2002, 362, 170–
178.
Notes and references
1 D. R. Bloomquist and R. D. Willett, Coord. Chem. Rev.,
1982, 47, 125–164.
2 A. Ouchi, T. Takeuchi and I. Taminaga, Bull. Chem. Soc.
Jpn., 1970, 43, 2609–2612.
3 M. Radjaipur and D. Oelkrug, Ber. Bunsenges. Phys. Chem.,
1979, 82, 159–163.
4 A. P. B. Lever, E. Mantovani and J. C. Donini, Inorg. Chem.,
1971, 10, 2424–2427.
5 P. Chandi, L. Fen-Ling, W. Sue-Lein and C. Chung-Sun,
Polyhedron, 1998, 17, 547–554.
6 K. Itaya, I. Uchida and V. D. Neff, Acc. Chem. Res., 1986, 19,
162–168.
7 P. Gütlich, A. Hauser and H. Spiering, Angew. Chem., Int.
Ed. Engl., 1994, 33, 2024–2054.
32 A. D. Becke, J. Chem. Phys., 1988, 88, 2547.
8 M. Kyuzou, A. Maeda, M. Wada, S. Watanabe, T. Takei, 33 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
T. Ohmura, Y. Kowaguchi, W. Mori and J. Tanaka, Bull.
Chem. Soc. Jpn., 2011, 84, 491–495.
9 I. Grenthe, P. Paoletti, M. Sandstroem and S. Glikberg,
Inorg. Chem., 1979, 18, 2687–2692.
Matter, 1988, 37, 785.
34 S. Panigrahi, S. Kundu, S. Basu, S. Praharaj, S. Jana,
S. Pande, S. K. Ghosh, A. Pal and T. Pal, J. Phys. Chem. C,
2007, 111, 1612–1619.
10 W. E. Hatfield, T. S. Piper and U. Klabunde, Inorg. Chem., 35 F. A. N. El-Dien, E. Y. A. Frag and G. G. Mohamed, Drug
1963, 2, 629–632. Test. Anal., 2010, 2, 234–242.
11 D. R. Bloomquist, M. R. Pressprich and R. D. Willett, J. Am. 36 M. Ganguly, A. Pal and T. Pal, J. Phys. Chem. C, 2012, 116,
Chem. Soc., 1988, 110, 7391–7398. 9265–9273.
12 K. L. Bray, H. G. Drickamer, E. A. Schmitt and 37 R. L. Harlow, W. J. I. Wells, G. W. Watt and S. H. Simonsen,
S. D. N. Hendrickson, J. Am. Chem. Soc., 1989, 111,
2849–2856.
13 J. Harada, T. Fujiwara and K. Ogawa, J. Am. Chem. Soc.,
2007, 129, 16216–16221.
Inorg. Chem., 1975, 14, 1768–1773.
38 M. Takayama, Int. J. Mass Spectrom. Ion Processes, 1995,
144, 199–204.
39 R. C. Brasted, Anal. Chem., 1952, 24, 1040–1041.
14 J. H. Fuhrhop and D. Fritsch, J. Am. Chem. Soc., 1984, 106, 40 A. Winter, A. Zabel and P. Strauch, Int. J. Mol. Sci., 2012, 13,
4287–4288. 1612–1619.
15 L. Menabue, G. C. Pellacani, L. P. Battaglia, A. B. Corradi, 41 Z. Puterova, J. Valentova, Z. Bojkova, J. Kozisek and
F. Sandrolini, A. Matori, R. J. Pylkki and R. D. Willett,
J. Chem. Soc., Dalton Trans., 1984, 2187.
16 J. Darriet, M. S. Haddad, E. N. Duesler and
D. N. Hendrickson, Inorg. Chem., 1979, 18, 2679–2682.
17 L. Antolini and L. Menabue, Transition Met. Chem., 1980, 6,
377–379.
F. Devinsky, Dalton Trans., 2011, 40, 1484–1490.
42 R. S. Clegg and J. E. Hutchison, Langmuir, 1996, 12, 5239–
5243.
43 G. M. Espallargas, M. Hippler, A. J. Florence, P. Fernandes,
J. Streek, M. Brunelli, W. I. F. David, K. Shankland and
L. Brammer, J. Am. Chem. Soc., 2007, 129, 15606–15614.
18 S. Hayami, N. Motokawa, A. Shuto, N. Masuhara, 44 V. T. Yilmaz, S. Guney and C. Kadak, Polyhedron, 2008, 27,
Y. Umebayashi, Y. Maeda, O. Sato, K. Inoue, R. Kawajiri, 1381–1386.
Y. Kumura, T. Okubo and T. Mitaniet, Chem. Lett., 2006, 35, 45 G. S. Jeon, G. Seo and J. S. Chung, Korean J. Chem. Eng.,
1114–1115. 1996, 13, 642–646.
19 G. Marcotrigiano, L. Menabue and G. C. Pellacani, Inorg. 46 S. Li, H. Z. Wang, W. W. Xu, H. L. Si, X. Tao, S. Lou,
Chem., 1976, 15, 2333–2336.
Z. Du and L. S. Li, J. Colloid Interface Sci., 2009, 330,
20 L. Fabrizzi, M. Michelani and P. Paoletti, Inorg. Chem.,
1974, 13, 3019–3021.
483–487.
47 V. Andal and G. Buvaneswari, Sens. Actuators, B, 2011, 155,
21 B. P. Kennedy and A. B. Lever, J. Am. Chem. Soc., 1973, 95,
6907–6913.
22 N. S. Gill and R. Nyholm, J. Chem. Soc., 1959, 3997–4007.
653–658.
48 J. J. Teo, Y. Chang and H. C. Zeng, Langmuir, 2006, 22,
7369–7377.
23 C. Pariya, F.-L. Liao, S.-L. Wang and C.-S. Chung, Poly- 49 C. H. Kuo, C. H. Chen and M. H. Huang, Adv. Funct. Mater.,
hedron, 1998, 17, 547–554.
2007, 17, 3773–3780.
11636 | Dalton Trans., 2014, 43, 11624–11636
This journal is © The Royal Society of Chemistry 2014