Design, synthesis, and characterization of an Fe(ii)-polymer of a redox non-innocent, heteroatomic, polydentate Schiff's base ligand: negative differential resistance and memory behaviour

Article abstract of DOI:10.1039/c8nj04106g

In the present work, the unique negative differential resistance (NDR) and memory effect of an organic-metallic hybrid polymer based on the self-assembly of Fe(ii)-ions preceded by the synthesis of a newly designed multidentate Schiff's base ligand has been reported. The polymerization process was closely monitored in an absorbance study using a UV-vis spectrophotomer. The presence of sharp isosbestic points and saturation of absorbance at 1?:?1 ligand to metal concentration confirmed the stepwise growth of the polymer. A model monomer Fe-complex was also developed based on the model ligand, with the aim of understanding the properties of the polymer in a better way by comparative study but surprisingly both ligand and model ligand coordinate in a different fashion with the same Fe(ii)-metal ions under similar reaction conditions, though both of them have identical binding sites. It has been found that the monomer model complex is paramagnetic while the metallopolymer is diamagnetic in nature. FESEM images of thin films of the polymer revealed the presence of homogeneously distributed interparticulate mesopores which was further confirmed by surface area analysis. A long single strand of the polymer was found on the HOPG surface by AFM study. The cyclic voltammetry study indicated that both conjugated ligand and metallopolymers are electrochemically active. The bistable memory behaviour of the device fabricated on ITO has shown a negative differential resistance effect along with good RAM and ROM behaviour, showing the potential of this novel polymer as memristor.

Full text of DOI:10.1039/c8nj04106g

NJC
View Article Online
View Journal | View Issue
PAPER
Design, synthesis, and characterization of an Fe(II)-
polymer of a redox non-innocent, heteroatomic,
polydentate Schiff’s base ligand: negative
Cite this: New J. Chem., 2018,
42, 19090
differential resistance and memory behaviour†
a
a
Deepa Oberoi,a Parveen Dagar,b Uday Shankar,
Giriraj Vyas,b Anil Kumar,
a
Satyajit Sahub and Anasuya Bandyopadhyay
*
In the present work, the unique negative differential resistance (NDR) and memory effect of an organic–
metallic hybrid polymer based on the self-assembly of Fe(II)-ions preceded by the synthesis of a newly
designed multidentate Schiff’s base ligand has been reported. The polymerization process was closely
monitored in an absorbance study using a UV-vis spectrophotomer. The presence of sharp isosbestic
points and saturation of absorbance at 1 : 1 ligand to metal concentration confirmed the stepwise
growth of the polymer. A model monomer Fe-complex was also developed based on the model ligand,
with the aim of understanding the properties of the polymer in a better way by comparative study but
surprisingly both ligand and model ligand coordinate in a different fashion with the same Fe(II)-metal
ions under similar reaction conditions, though both of them have identical binding sites. It has been
found that the monomer model complex is paramagnetic while the metallopolymer is diamagnetic in
nature. FESEM images of thin films of the polymer revealed the presence of homogeneously distributed
interparticulate mesopores which was further confirmed by surface area analysis. A long single strand of
the polymer was found on the HOPG surface by AFM study. The cyclic voltammetry study indicated
that both conjugated ligand and metallopolymers are electrochemically active. The bistable memory
behaviour of the device fabricated on ITO has shown a negative differential resistance effect along with
good RAM and ROM behaviour, showing the potential of this novel polymer as memristor.
Received 12th August 2018,
Accepted 16th October 2018
DOI: 10.1039/c8nj04106g
rsc.li/njc
and behave anisotropically, are called ‘‘smart materials’’. In
recent decades, smart polymeric materials have gained deep
Introduction
The dream of realizing a ‘‘smart world’’ has been motivating interest; especially the domain of organic–inorganic hybrid
researchers to develop advanced materials based devices, which smart polymers has been the focal point of the research com-
have an ability to combine functionality with state-of-the-art munity. Some of the excellent properties have emerged from the
design and can sense and behave smartly. Realization of these synergistic effect of both organic and inorganic materials,
nano-devices will not only enhance the quality of human living through their fusion either at the nanoscale (composite) or at
standards but will also provide safety and comfort, for instance, as the molecular level (coordination polymer). Therefore, these
in the case of electrochromic materials,1–4 memristive devices,5–7 hybrid materials have the advantage of both of their counter-
pH responsive materials,8–10 piezoelectric polymers,11,12 photo- parts, viz. ease of processing, lightweight, flexibility, and extre-
chromic materials,13,14 thermochromic materials,15,16 self- mely high versatility in polymer’s design result from the organic
healing materials,17 nano-electromechanical switches,18 etc. component while the inorganic counterpart contributes towards
Such materials, which have been designed to mimic nature superior thermal and mechanical behaviour along with optical,
catalytic, electrical, and magnetic properties.19,20 The advantage
a Department of Polymer and Process Engineering, IIT Roorkee, Saharanpur-247001,
of these hybrid materials is that the metal ion is directly bonded
to the ligand, which results in a more organised electron cloud
distribution over the covalently bonded metallopolymer back-
bone as compared to doped hybrid materials.6
The successful performance of hybrid materials in different
fields14,21–25 led to an increased thrust to explore these materials
in the field of electronics26 in order to meet the increasing
India. E-mail: abptsfpt@iitr.ac.in
b Department of Physics, IIT Jodhpur, Jodhpur-342011, India
† Electronic supplementary information (ESI) available: The experimental syn-
thetic procedure of the model ligand and the Fe-monomer complex of the model
ligand and VSM data of the model ligand are provided. The SLS, FTIR, NMR,
COSY spectra, FESEM of powder sample of the ligand and AFM measurements
and calculations along with CV of the ligand. See DOI: 10.1039/c8nj04106g
19090 | New J. Chem., 2018, 42, 19090--19100 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
Paper
NJC
demand for large capacity data storage.27 Amongst these, non- passed on to other metal ions present in the same polymer
volatile memristive devices28 have received significant atten- chain at a faster rate through the same conjugated organic
tion as they offer low cost synthesis, easy processability, fast backbone. Each ligand is coordinated with two different Fe(II)-
switching speed, a large On/Off ratio along with longer stability centres at the same time and linear growth of the polymer is
and less power consumption.29,30 In the literature it has been possible by self-assembly of each Fe(II)-centre with one-half of
reported that Bandyopadhyay et al. observed the memristive two ligands. In this report, we are introducing a unique, redox
behaviour of a directly bonded metal–organic hybrid polymer non-innocent, diamagnetic, mesoporous Fe(II)-polymer with all
containing a redox active azo group.6 Later, Hu et al. reported required characteristics for future application as a multifunc-
the multilevel memories of a hybrid polymer of PMMA and tional material.
polyoxometallate (PMMA–MAPOM) with rewritable switching
properties and good retention.27 Sometimes, these hybrid
materials show a Negative Differential Resistance (NDR) effect,
Experimental section
Materials
a unique phenomenon which shows a decrease in current with
increasing voltage.31 NDR devices have tremendous potential in
4,4-Dihydroxybiphenyl, hydrazine monohydrate and N,N-dimethyl-
acetamide (N,N-DMAc), tetra-ethyl orthosilicate and 2-hydroxy-
1-naphthaldehyde were purchased from TCI Chemicals (India)
terms of low power memory devices and logic gate fabrication,
with the existence of the bistable states.32 However, intriguing
research is still underway to develop new strategies and
Pvt. Ltd and used as received. Fe(II) acetate and Pd/C were pur-
molecular architectures for these advanced functional hybrid
chased from Sigma-Aldrich, India. Celite, sodium bicarbonate
memory devices.
(NaHCO3) and sodium metal were purchased from Merck
An alternative way to enhance the desired functional proper-
Ltd, India. Nitric acid (HNO3) was purchased from Rankem,
ties of smart ‘tailor-made’ materials is to increase the area of
India. All the solvents were of AR grade and purified by
the interactive surface. This can be achieved by a uniform
standard procedures wherever specified. Solvents were pur-
porous structure of the material which provides a large surface
chased from Merck, India. Double distilled water was used
area to volume ratio. In this regard, mesoporous materials are
wherever required.
becoming promising candidates for the synthesis of materials
showing multifunctional properties.33,34 They have extensive
Instrumentation and characterization
ability to interact with ions/molecules/atoms/nanoparticles
not only through surface but also through bulk of the materials, The Fourier transform infrared spectra of the ligand and poly-
which helps in better conduction and movement of ions through mers were recorded on a PerkinElmer FT-IR C91158 spectro-
homogeneously distributed pores. These materials have been photometer using KBr pellets, in the range of 4000–400 cmꢀ1
.
extensively used over the years in adsorption, separation, catalysis, 1H NMR, 13C NMR and H–H COSY spectra were measured using
sensing, energy storage, biotechnology,35,36 etc. Some reports on a Bruker 400 Hz system with tetramethylsilane (TMS) as a
the non-volatile memory application of mesoporous structures of reference. The wide angle X-ray diffraction (WXRD) patterns
organic–inorganic hybrid perovskites also exist.37
of the sample were collected by using a RIGAKU ULTIMA IV X
With these considerations, the present study focusses on the ray diffractometer, in the angle range of 2y = 5–801 at a scan
design and synthesis of a novel redox non-innocent ligand38,39 rate of 41 minꢀ1, using Ni filtered CuKa as a radiation source.
by exploiting Schiff’s bases in the organic backbone. Such types The absolute molecular weight of the polymer sample was
of ligand can induct reversible electrochemical activity in both recorded by Static Light Scattering (SLS) experiment in batch
metal-monomers and metallopolymers.4,40,41 This reversible mode using a Brookhaven BI-200SM goniometer using filtered
electrochemical switching acts as a key to generate properties stock solution and solvent filtered through a 0.2 micron Nylon
like electrochromism, memristive property, ferroelectric property, filter. The UV measurements were made on a UV spectro-
etc. in hybrid materials.42–44 However, to the best of our knowl- photometer (UV-1800, Shimadzu Corp., Japan). Cyclic voltam-
edge, a comprehensive study of Schiff’s base hybrid polymers for metry (CV) measurements were performed using a cyclic
memory applications is scarce in the literature.27
voltammeter (CH1620D, CH Instruments). Atomic Force Micro-
Herein, we describe the synthesis and characterization scope (AFM) imaging of the polymer was carried out on a
method of a multidentate Schiff’s base ligand and its Fe(II)- freshly cleaved Highly Ordered Pyrolytic Graphene (HOPG)
polymer. The motivation behind this research is to generate substrate using a NT-MDT-INTEGRA Microscope in a tapping
novel properties from the strong metal–ligand interaction, mode at room temperature. Field Emission Scanning Electron
which is present in this kind of coordination polymer because Microscope (FESEM) images of the sample were recorded on
of the formation of a direct covalent bond between metal ions TESCAN MIRA 3, by pouring a 10ꢀ5 M solution of the sample on
and ligand instead of secondary interactions like van der Waals glass slides of area 1 cm2 mounted on stubs and sputtered with
forces present in nano-metal doped organic polymer matrices. gold using a vacuum sputter. Energy Dispersive X-Ray (EDX)
The backbone of the Fe(II) polymer is synthesized using the spectroscopy was used to determine the Fe content in the poly-
metal coordinated conjugated ligand, so that if any change is mer sample, attached to the FESEM unit. Surface area and pore
induced in the oxidation state of the metal ion by external size distribution analysis of the samples were performed using
stimuli in the metallopolymer, then that message would be a Quantachrome Autosorb IQ2 using N2. Vibrating Sample
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 19090--19100 | 19091

View Article Online
NJC
Paper
Magnetometer (VSM) Quantum Design PAR 155 was used to 8.50 (d, J = 8.3 Hz, 1H), 8.25 (d, J = 1.9 Hz, 1H), 7.79 (d, J = 9.4 Hz,
determine the magnetic moment of the samples.
1H), 7.66 (dd, J = 7.9 Hz, 1.0 Hz, 1H), 7.49 (dd, J = 8.4 Hz, 2.1 Hz,
1H), 7.44 (t, J = 7.6 Hz, 1H), 7.25 (t, J = 7.6 Hz, 1H), 7.05 (d, J =
8.3 Hz, 1H), 6.77 (d, J = 9.3 Hz). 13C NMR (400 MHz, d6-DMSO)
Synthesis of the ligand
3,30-Dinitrobiphenyl-4,40-diol (1). For the synthesis of the d: 178, 150, 147, 138, 134, 132, 129.3, 129, 128, 126, 125.5, 125,
dinitro derivative of 4,4-biphenol, a simple route has been 123, 120, 116, 115, 114, 108.
followed with the modification of the synthetic procedure.45
Fe(II) polymer [Fe(II)-poly]
Nitric acid (70%, 2 ml) was added dropwise to a solution of
4,40-dihydroxybiphenyl (5.4 mmol, 1.0 g) in 20 ml of glacial
acetic acid, with vigorous stirring over a period of an hour at
room temperature. After the complete addition, the stirring was
continued at room temperature for about three hours to
complete the nitration. Then 150 ml of water was added to
the above solution and stirring was continued for about three
hours at 50 1C. The reaction mixture was then cooled to room
temperature and filtered. Precipitates were then washed with
an adequate amount of water followed by ethanol and saturated
aqueous solution of NaHCO3. Then yellow colored precipitates
were collected. Yield: 90% (1.3 g). FTIR (KBr, cmꢀ1): 3431, 3251,
1626, 1528, 1470, 1309, 1241, 1173. 1H NMR (400 MHz, d6-DMSO)
d: 11.11 (s, 2H, OH), 8.10 (d, 2H, J = 2.4 Hz), 7.83 (dd, 2H, J = 8.7 Hz),
7.15 (dd, 2H, J = 8.7 Hz). 13C NMR (400 MHz d6-DMSO) d: 151.8,
137.9, 133.24, 129.62, 122.97, 120.15.
3,30-Diaminobiphenyl-4,40-diol (2). To a stirred solution of
3,30-dinitrobiphenyl-4,40-diol (1.38 g, 5 mmol) in ethanol, Pd/C
(0.12 g, 5 mmol) was added followed by the slow addition of
hydrazine monohydrate (5.5 ml).46 After the complete addition,
the mixture was heated and stirred at 85 1C. After 24 h, N,N-DMAc
was added slowly to the suspension until dark grey precipitates
appeared in the solution and addition of solvent was continued
until all the precipitates were completely dissolved.46 The final
hot solution was filtered through celite and then it was evapo-
rated in a rotary evaporator to reduce the volume of the solution
to its half and later on added into the ice-cold water. White
precipitates were washed thoroughly with distilled water and
dried in an oven. Yield: 70% (75 mg). FTIR (KBr, cmꢀ1): 3378,
3290, 1592, 1509, 1441, 1280, 1221, 904, 811. 1H NMR (400 MHz,
d6-DMSO) d: 8.89 (broad (br), s, 2H, OH), 6.71 (dd, 2H, J = 2.2 Hz),
6.59 (d, 2H, J = 8.1 Hz), 6.49 (dd, 2H, J = 8.1 Hz), 6.49 (dd, 2H),
4.47 (br, s, 2H, NH2). 13C NMR (400 MHz, d6-DMSO) d: 143, 137,
133, 115, 112, 114.
The polymer of the above ligand was prepared by the addition
of 300 mg (0.572 mmol) of Schiff’s base ligand in dry ethanol,
followed by the addition of 8–10 ml of NEt3 in a nitrogen
atmosphere. The mixture was stirred for about an hour. Then
100 mg (0.572 mmol) of Fe(OAc)2 was added to the above
mixture and was kept under reflux for about 24 h. The brown
precipitates so obtained were dried and washed several times
with hot ethanol yielding 0.32 g of the product (yield: 80%).
FTIR (KBr, cmꢀ1): 3438, 1614, 1514, 1400, 1344, 1257, 1123,
1085, 798. 1H NMR (400 MHz, d6-DMSO) d: 15.81 (br, 1H), 10.40
(br, 1H), 9.66 (br, 1H), 8.50 (br, 1H), 8.25 (br, 1H), 7.79 (br, 1H),
7.66 (br, 1H), 7.49 (br, 1H), 7.44 (br, 1H), 7.25 (br, 1H), 7.05 (br,
1H), 6.77 (br, 1H). Molecular weight as obtained by SLS was
68 000 g molꢀ1 (approx.) (for Berry plot vide Fig. S1, ESI†).
Degree of polymerization (molecular weight/repeating unit
weight), n E 120 (approx.).
Device fabrication
Indium tin oxide (ITO) coated substrates were used for device
fabrication. ITO was selectively etched to form a stripe by using
concentrated HCl and Zn dust. After etching, the substrate was
cleaned using a standard cleaning procedure, i.e. ultrasonica-
tion with soap solution, then rinsing with de-ionized (DI) water,
followed by cleaning the substrate with acetone, DI water, and
methanol using an ultrasonicator, thrice with each solvent and
each time for 15 minutes. The cleaned substrates were then
dried in a vacuum oven at 70 1C for 4–5 hours. These cleaned
substrates were then used to deposit the organic material on
them. A 25 mg mlꢀ1 Fe(II)-poly in DMSO solution was prepared
for the deposition of the film. The films were spin coated over
the substrate at a rotational speed of 1000 rpm. The spin coated
samples were kept under vacuum overnight. An 85 nm thick
aluminium (Al) layer was thermally evaporated on the spin
coated films. For the deposition of Al a shadow mask was used
which gave us a cross-bar architecture. The effective area of the
device was 0.25 cm2. So, for the device ITO acts as the bottom
contact and Al acts as the top contact. The I–V characterization
of the sample was carried out by using a two-probe measure-
ment technique where a Keithley 6430 Femto-ampere source
meter was interfaced with a vacuum probe station. Five such
devices were prepared and their electrical measurements were
analysed at room temperature, under ambient conditions.
Synthesis of ligand 3,30-bis(((2-hydroxynaphthalen-1-yl)-
methylene)amino)-[1,10-biphenyl]-4,40-diol (L1). The Schiff’s
base ligand (L1) has been synthesized by the reaction of 3,30-
dinitrobiphenyl-4,40-diol (2 mmol, 0.432 g) with 2-hydroxy-1-
naphthaldehyde (4 mmol, 0.744 g) in dry EtOH (20 ml) in a
three neck flask equipped with a condenser, under a N2 atmo-
sphere. Tetraethyl orthosilicate (6 ml) was added to this
solution mixture and then it was refluxed for about 6 h. After
the consumption of initial reactants, the solvent was removed
by using a rotary evaporator and the obtained crude mass was
dried in an oven for about an hour. Then the solid mass was
recrystallized from ethanol three times to get a pale-yellow
product. Yield 73% (1.5 g). FTIR (KBr, cmꢀ1): 3431, 1622, 1544,
Results and discussion
1350, 1236, 1265, 1236, 1121, 1157. 1H NMR (400 MHz, d6-DMSO) The ligand (L1) was designed with the aim of developing a new
d: 15.81 (d, J = 9.6 Hz, 1 H), 10.40 (s, 1H), 9.66 (d, J = 9.7 Hz, 1H), organic molecule with three coordination sites, viz. N, O, O,
19092 | New J. Chem., 2018, 42, 19090--19100 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
Paper
NJC
Scheme 2 (a) Schematic representation of the possible Fe(II)-poly
structure, (b) representation of the monomer unit of Fe(II)-poly with
n = 120 units.
Scheme 1 Schematic representation of synthesis of the ligand and
Fe(II)-poly.
and NMR studies. A detailed discussion of this study has been
provided separately in the NMR studies section.
in each half of the ligand and with the potential to bind with a
All the characterization studies have been summarized
variety of hard/soft metal ions for construction of different below:
organic–metallic hybrid polymers. Both ligand and Fe(II)-poly
FTIR studies
were characterized by various characterization techniques
(Scheme 1).
The overlay FTIR spectra of the ligand and polymer are pro-
When the metal and ligand were taken in 1 : 1 proportion for vided in ESI,† Fig. S2. The FTIR spectrum of the ligand shows a
polymerization reaction, the expectation was that all the –OH broad peak at 3431 cmꢀ1 owing to the presence of the –OH
groups in the ligand will first get deprotonated and then they group, a peak at 1630 cmꢀ1 due to –CQN stretching vibrations
will coordinate as –Oꢀ1 monoanionic donors, but the presence while another at 1236 cmꢀ1 due to –C–N stretching vibrations.
of the broad, characteristic signal of the –OH group in FTIR and However, a slight shift has been observed in these peaks upon
hydroxyl protons in NMR experiments was opposite to our coordination to the metal ion.47 The broad peak of the –OH
expectation and both these experiments confirmed the pre- group at 3431 cmꢀ1 in the ligand is shifted to 3438 cmꢀ1 in
sence of hydroxyl protons in Fe(II)-poly. As the obtained poly- Fe(II)-poly and another peak at 1622 cmꢀ1 of the ligand is
mer is a neutral coordination polymer and it is a diamagnetic shifted to 1614 cmꢀ1 in Fe(II)-poly, which is an indication of
material (NMR active), we propose the structure of the polymer stretching of the –CQN double bond on coordination of the
as per Scheme 2, which can support both the experimental imine moiety to the metal ion. A strong peak at 798 cmꢀ1 has
results.
been observed in the fingerprint region of the polymer, arising
Our developed organic ligand can be divided into two due to the bending vibration of the Fe–OH group.48
coordination halves and each half has three coordinating sites,
NMR studies
N, O, O. Then, to complete the coordination number of Fe(II)-
ions, two different ligands with one-half of each have to be NMR characterization of the ligand, deprotonated ligand and
coordinated to the same metal center, while the other half of polymer was performed as shown in Fig. 1. Because of the
both the ligands is coordinated to two different metal centres. presence of symmetry, all the compounds showed peaks for
This way a linear polymer chain will grow with continuous self- one-half of the ligand only. The conjugated ligand shows a
assembly between metal centers and organic ligand. If all four doublet at 15.82d due to the presence of the phenolic –OH (C2)
–OH groups will get deprotonated as per expectation, then the group, and a second hydroxyl group –OH0 (C-13) gives a sharp
neutral complex will only be made with the Fe(IV)-metal center, singlet at 10.40d. The difference in the nature and position of
which is the d4 system. Then, the metallopolymer would become two hydroxyl groups is originated from the hydrogen bonding
paramagnetic as Fe(IV) will make a low spin octahedral complex interaction of one (C-2) with the neighbouring imine N while
with the hard N,O,O donor ligand with two unpaired electron on the other is deprived of that interaction as two of them cannot
metal center. The same complex with the Fe(II) center (d6) metal participate in hydrogen bonding with the same imine group at
ion will only be neutral if each ligand will coordinate as mono- the same time. Again, hydrogen bonding interaction of the –OH
anionic hexacoordinated binding sites, where one hydroxyl group attached to C2 with imine N is through six-member ring
group of the ligand will coordinate as a neutral hydroxyl group formation while the same with a second hydroxyl group makes
and the other one will be a hard, monoanionic –Oꢀ1 site. The a five membered ring. As the five-membered ring is less stable
proposed structure is also diamagnetic in nature and can than the six membered ring, exclusively only –OH (C2) is taking
provide support for the presence of hydroxyl signals in FTIR part in hydrogen bonding. The effect of keto–enol tautomerism
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 19090--19100 | 19093

View Article Online
NJC
Paper
Fig. 1 Overlay 1H NMR spectra of the ligand, deprotonated ligand and Fe(II)-poly in DMSO-d6.
(vide Fig. S3, ESI†) is also observed in the imine proton peak.49 different aliquots of Et3N were added to completely deprotonate
As the imine proton (C11) has no neighbouring proton, only a the ligand first and then polymerization reaction with Fe(OAc)2
singlet was expected in the NMR signal but it gives a doublet was done, but each time NMR spectra remained unaltered,
because of keto–enol tautomerism. To confirm hydrogen bonding which supports that the proposed structure is the best adopted
interaction of the hydroxyl proton and the imine nitrogen, the coordinated form and Fe(II)-poly is only stabilized in the supra-
ligand was completely deprotonated with triethyl amine. The molecular structure with the conjugated backbone as the soft
NMR signal of the deprotonated ligand does not show any Fe(II) cation to minimize steric strain in the organic moiety.
hydroxyl proton peak and the imine proton shows a sharp When the monomeric Fe complex of the model ligand was
singlet as in the absence of hydroxyl proton it is not possible synthesized under identical conditions as for Fe(II)-poly pre-
to start keto–enol tautomerism. All the other peaks in the paration, then every time paramagnetic monomer complex was
deprotonated ligand are of the same type as found in the original obtained. This means that the deprotonated model ligand (L2)
ligand. The full designation of rest of the peaks has been done by has the best interaction and crystal field stabilization with hard
13C NMR spectroscopy (vide ESI,† Fig. S3) and H–H COSY NMR Fe(III) cations but in the case of the polymer to minimize steric
spectroscopy (vide ESI,† Fig. S4).
strain in the hexadentate ligand structure, the deprotonated
In the case of Fe(II)-poly, very broad peaks are observed in ligand (L1) accepts a proton from solvent during the reaction to
the NMR spectrum which are characteristics of the polymeric minimize the strain.
sample. As the –OH (C2) proton exclusively takes part in the
To further simplify the analysis of the ligand, a model ligand
hydrogen bonding interaction with imine N, the acidity of this has been synthesized and analyzed by NMR and VSM charac-
proton is more than that of the second hydroxyl proton. There- terization (vide ESI,† Scheme S1 and Fig. S5, S7). The NMR of
fore, when the ligand coordinates to the metal ion, first this the C-2, C-13 and C–H imine proton of the model ligand shows
hydroxyl group of one-half of the ligand is deprotonated and a similar nature to that of the ligand, with a small chemical
this half is donating as a tridentate, monoanionic binding site. shift. This small shift is arising as the model ligand has a
To complete the primary and secondary valency of Fe(II), one- structure which is half of the original ligand and thus lacks
half of the second ligand is also deprotonated and coordinates extensive conjugation which is present in the original ligand.
to the same center metal ion. Now, after coordination of one- The Fe-complex (monomer) of this model ligand has also been
half of one ligand to the centre metal ion as a monoanionic prepared by using identical reaction conditions (vide ESI,†
ligand, the acidity of –OH0 (C130) is increased more than that of Scheme S2 and Fig. S5). The NMR characterization of this
–OH (C2) in the second half of the ligand. So, the other half of monomeric iron complex shows sharp peaks in the range of
each coordinated ligand gets deprotonated through the –OH0 ꢀ10d to +35d thereby indicating that this model monomer
(C130) group (Scheme 2) and the second hydroxyl group (C20) in complex is paramagnetic in nature (vide ESI,† Fig. S6). These
the other half of the ligand is coordinated as the neutral site by sharp 1H peaks with a moderate paramagnetic shift signify the
donating its lone pair only. For this reason, both hydroxyl presence of low spin Fe(III).50 So, the model ligand can stabilize the
proton peaks are observed in the NMR spectrum of Fe(II)-poly metal in +3 oxidation state while Fe(II)-poly remains in +2 oxidation
instead of only one. To check the reproducibility of the result with the original ligand under identical reaction conditions.
19094 | New J. Chem., 2018, 42, 19090--19100 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
Paper
NJC
As the model ligand is only half portion of the original ligand,
so one ligand can coordinates to only one metal centre in the
monomer complex and this sole metal coordination makes it
easier for the model ligand to lose two/one proton(s) and coordi-
nate as tridentate, monoanionic from one side and tridentate,
bisanionic from another side to the Fe(III)-center to attain more
Crystal Field Stabilization Energy (CFSE). In this fashion the
basicity of the ligand is increased and bonding with the metal is
stronger because the metal is also present in a higher oxidation
state compared to Fe(II)-poly. In the case of Fe(II)-poly, the
hexacoordinated conjugated, bulkier ligand is simultaneously
coordinated to two different metal centres and because of that a
very close approach of the ligand to the metal is not possible as
that would increase steric strain in the totally conjugated back-
bone of the metallopolymer. So, in the case of Fe(II)-poly, the
ligand (L1) is coordinating in the hexadentate, bisanionic state,
while the model ligand is coordinated in the tridentate, tris
anionic state in the Fe-monomer metal complex.
Study of photophysical properties
After confirming the structure from IR and NMR studies, the
absorption spectra of 10ꢀ5 M solution (DMSO/water, 1/9 v/v) of
both ligand and Fe(II)-poly have been recorded at room tem-
perature and compared with the absorption spectrum of the
deprotonated ligand, by adding the required amount of triethyl-
amine to the ligand (Fig. 2), to understand the absorption
behavior of the ligand upon co-ordination with the metal ion.
The UV spectrum of the ligand shows four major bands at
wavelengths 243, 316, 453 and 480 nm. The band at 243 nm
arises due to the p–p* transition of naphthalene/aromatic rings.
The 316 nm and 343 nm bands arise due to the p–p* transition
of the –CQN group of the ligand. The bands at 453 and 480 nm
arise due to the n–p* transition of the –CQN bond. In the case
of the deprotonated ligand, a variation in the spectral peak is
observed compared to the spectrum of the free ligand because of
the lack of hydrogen bonding interactions upon deprotonation.
Fig. 3 (a) UV-vis titration of Fe(OAc)2 with the Schiff’s base ligand (10ꢀ5 M)
in DMSO/methanol with the Job’s plot in the inset, (b) and (c) sharp iso-
sbestic points at wavelength 417 nm and 496 nm respectively.
The spectral behavior of Fe(II)-poly is a combination of both
protonated and deprotonated ligand spectra with a slight shift
because of co-ordination to Fe(II)-ions. The band at 483 nm
disappeared in the polymer spectrum because of the involve-
ment of non-bonding electrons with the metal ions upon coor-
dination, and another broad low energy, low intensity transition
in the NIR region has been generated at 918 nm (vide ESI,†
Fig. S8) which is due to metal-to-ligand charge-transfer (MLCT)
transition (from the d-orbital of the metal to the p* orbital of
the ligand).
After the confirmation of the metal complex, the nature of
complexation between the metal ion and the ligand was deter-
mined by the UV-titration method (Job’s plot) (Fig. 3). To a
methanolic solution of the ligand, iron(II)-acetate solution was
consistently added until an equimolar stoichiometry of ligand/
metal ion (1 : 1) is observed. During the titration process, two
sharp isosbestic points at wavelengths 419 nm and 495 nm
appeared. The observed response at 500 nm wavelength was
plotted against metal to ligand proportion (Fig. 3, inset), which
clearly shows saturation of absorbance intensity at ligand/metal
equimolar concentration, i.e. 1 : 1 concentration. The band at
243 nm has been red shifted to 264 nm with the hypsochromic
shift. A new red shifted band at 345 nm appeared with a slightly
higher intensity during titration.
Fig. 2 Absorption spectra of 10ꢀ5 M solution of ligand, deprotonated
ligand and Fe(II)-poly in DMSO : methanol (1 : 9).
An additional increase in peak height after 500 nm wave-
length in UV titration spectra was also observed due to M–L
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 19090--19100 | 19095

View Article Online
NJC
Paper
The XRD pattern of the Fe(II)-poly shows similar peaks as that of
the ligand except that there is a huge drop in intensity for the
Fe(II)-poly. However, an additional peak at 29.571 is found in the
metal polymer, which is due to stacking of the conjugated aromatic
system in the polymer.52,53 This also clearly points towards the
existence of host–guest interactions between the Fe(II) units and the
multidentate organic ligand by direct bond formation between
them and the absence of formation of any aggregate because of
weak secondary interactions between the metal ion and the ligand.
As the XRD pattern of Fe(II)-poly shows only broad diffrac-
tion peaks, which indicate amorphous nature of polymer and it
has also been suggested that various p–p stacking structures are
expected in between the hybrid polymer chains.54 The XRD
pattern of Fe(II)-poly exhibits its amorphous nature, so our
various attempts to grow single crystals were in vain.
Fig. 4 Comparison of emission properties of the ligand and Fe(II)-poly
excited at 465 nm, with the inset showing the color of the ligand and
polymer in the UV chamber with L designating the ligand and P Fe(II)-poly.
Surface morphology
The morphological study of the polymer surface was performed
by various characterization techniques such as FE-SEM, AFM
and BET surface area analysis. A thin film of polymer was grown
on a glass substrate by pouring a well sonicated dilute (10ꢀ5 M)
solution prepared in methanol:DMSO solvent and later dried in
air for about four to five hours. The images obtained from
surface studies by different techniques are shown in Fig. 6
(images of the powdered Fe(II)-poly sample are provided in
ESI,† Fig. S9).
SEM micrographs of the Fe(II)-poly film show the porous
nature of the polymer and this was later confirmed by BET
surface area analysis using N2 gas at 77 K (Fig. 6(c)). The
polymer shows a reversible isotherm of type-IV, with a small
hysteresis and shows a Langmuir surface area of 39 m2 gꢀ1. The
pore size distribution plot is given in the inset of Fig. 6(c),
which further confirms the mesoporous nature of the polymer
with the pore size between 4 and 42 nm. This pore range closely
resembles the pore size obtained by FESEM analysis. Further,
the EDAX analysis of the Fe(II)-poly provides the percentage
distribution of its components and it has been found that Fe(II)
ions are present with 1% abundance with a regular distribution
throughout the organic framework (vide ESI,† Table S1 and
Fig. S10). However, an unexpected percentage distribution of
oxygen is shown by EDAX analysis, which is due to the fact that
the elemental analysis is performed on the glass substrate
which also adds its own energy dispersive spectra along with
the appearance of other components of the glass substrate
(Table S1, ESI†). The presence of pores throughout the polymer
film may allow better transportation of charge/ions through it
and thus may enhance the mobility of small molecules/ions
through the polymer thin film.
charge transfer transition, which also justifies the formation
of the complex. Shifting of the UV bands during the titration
shows the formation of a complex between the Fe(II)-ions and
the –OH group of the naphthalene ring and a –CQN bond.
The emission properties of the ligand and Fe(II)-poly are
shown in Fig. 4. Both the ligand and polymer were excited at
wavelength 465 nm and it has been found that the fluorescence
shown by the ligand has been significantly quenched in Fe(II)-
poly because of intermolecular interactions between the multi-
dentate Schiff’s base ligand and the Fe(II) ion,51 thereby causing
a radiation-less decay through internal conversion (IC).
XRD studies
In order to observe the crystalline behavior of both ligand and
polymer, the patterns of XRD were obtained by using an X-ray
diffractometer. The overlay results of the XRD pattern of both
ligand and polymer are shown in Fig. 5.
For AFM analysis an even dilute solution (10ꢀ6 M) of Fe(II)-
poly was used to grow a very thin film on a freshly cleaved
HOPG substrate. The dilute solution (DMSO:methanol) was
injected forcefully from one end of the HOPG surface, so that
the polymer chains will detangle and align along the line of force
on the HOPG surface, and then kept for drying in a vacuum oven
for 12 h. The image of a single polymer strand obtained by AFM,
which is shown in Fig. 6(d), shows a Fe(II)-poly thickness
Fig. 5 XRD graph of Fe(II)-poly and the ligand; the inset shows the XRD
peak due to stacking of the conjugated aromatic system.
19096 | New J. Chem., 2018, 42, 19090--19100 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
Paper
NJC
Fig. 6 (a and b) FESEM images of 10ꢀ5 M solution of Fe(II)-poly on a glass substrate (small pores have been highlighted by a circle with approx. size 44 nm
in (a)), (c) N2 adsorption isotherm of the Fe(II)-poly at 77 K with the pore size distribution in the inset and (d) the single strand of the Fe(II)-poly obtained
from AFM on the HOPG substrate.
of 1.4 nm, which closely resembles the calculation of the model the very strong bonding in the ligand and because of that
structure obtained by MOPAC calculations after energy mini- strong conjugated bonding, after reduction of the first imine
mization (vide ESI,† Fig. S11).
group, the message is passed to the second imine chromophore
very fast and it will get reduced at almost double potential than
that of the first imine group reduction.
Electrochemical properties
The metallopolymer showed two anodic responses and one
cathodic response in the range of ꢀ2.0 V to +1.5 V. Anodic
responses are due to stepwise oxidation of the Fe(II)–Fe(III)
couple and Fe(III)–Fe(IV) couple and the cathodic response is
originated from reduction of the organic ligand. As the imine
nitrogen is coordinated to the Fe(II)-metal ion in the polymer, a
right shift in the cathodic response is observed as the metal ion
pulls the electron cloud from the ligand and makes it a more
easily reducible substrate than the free ligand.
The redox properties of the free ligand and Fe(II)-poly were
studied by CV with glassy carbon as the working electrode, a Pt
wire as the counter electrode and Ag/AgCl as the reference
electrode in acetonitrile solution containing 0.1 M TBAP at a
scanning rate of 100 mV sꢀ1. The voltammograms of Fe(II)-poly
are shown in Fig. 7 and that of the ligand are shown in the ESI,†
Fig. S12.
The CV results of the ligand showed two reversible cathodic
responses in the range of 0 V to ꢀ2.0 V (vide ESI,† Fig. S12).
The reduction at ꢀ0.77 V is attributed to the reduction of the
first imine group, whereas the second imine group showed
I–V characterization
reduction at a higher potential (ꢀ1.85 V) as the neutral ligand The I–V curve for the device (vide Scheme 3) spun at 1000 rpm is
becomes negatively charged after the first reduction and repels shown in Fig. 8 and it has shown bistability in the conductance.
incoming electrons for further reduction. This also confirms When the voltage was scanned in the positive direction, it first
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 19090--19100 | 19097

View Article Online
NJC
Paper
Fig. 7 Cyclic voltammogram of Fe(II)-poly. (inset) Response of the
(a) Fe2+/Fe3+ redox couple and (b) Fe3+/Fe4+ redox couple in acetonitrile
solution containing 0.1 M TBAP at 100 mV sꢀ1 with glassy carbon as the
working electrode and Ag/AgCl as the reference electrode.
Fig. 8 Current–voltage (I–V) graph of Fe(II)-poly scanned between
ꢀ4.0 V and +4.0 V with the blue arrows showing the direction of the
voltage scan. The I–V curve for the device is recorded 5 times. The inset
shows the zoomed-in area present inside the green circle.
compliance remains there till it reaches ꢀ2.2 V. After ꢀ2.2 V the
I–V curve follows the high conducting path till it completes the
cycle at 0 V. We have collected hundreds of I–V curves and
found them to follow the same path reproducibly except in the
first few cycles where the second local maximum is shifting
towards the high voltage region to a maximum of 1.7 V. So, in
this whole process of recording the I–V curve, the device shows
two negative differential resistance (NDR) peaks on the positive
side and one on the negative side of the voltage scan. The maxi-
mum On–Off ratio for the device is 11 which is not a significant
figure to explain the bistability behaviour of the device, but
what is interesting is the NDR feature in the device, which in
turn will also reveal the non-volatile memory effect in the device
that can be written, erased and read repeatedly.32 To date only a
few reports of hybrid materials exhibiting the NDR effect have
Scheme 3 Device fabrication of Fe(II)-poly as active material.
traces the high current path as shown by the guided arrow been reported55 and our initial attempts at the study of our
(blue) and reaches a local maximum at 0.1 V. This region is synthesized economical novel hybrid polymer have shown this
zoomed in (green circle) and shown in the inset of Fig. 8. With a remarkable behaviour.
further increase in the voltage the current comes down and
We can explain the NDR behaviour of the device by under-
gives us a short lived local minimum. As the voltage increases standing the structure of the molecule. The molecule is doubly
further the current also increases till 1.65 V and beyond that the reduced by accepting two electrons and when the electron is
current again decreases till it reaches a valley at 2.1 V, which captured in one of the sites (nitrogen) of the molecule, that point
again is a local minimum. After this point the current again acts as a quantum well for the electron and unless a suitable
starts increasing till it reaches compliance at 2.95 V and stays resonant energy is provided to the molecule, the transport
there till the scan voltage reaches the maximum at 4 V. When through the molecule will be restricted and hence there is a
the curve was traced back from the positive voltage toward the decrease in the current through the device (after 0.1 V). But at
negative direction it maintains the compliance current till 2.8 V some higher voltage the electron will be released and the current
but after that it follows a high resistive path and continues to again starts to increase till two electrons are captured at two
follow the negative voltage scan except at ꢀ0.2 V where the different sites (nitrogen) of the molecule. This will again reduce
current goes to a local maximum. After this point the current the current through the device and after a particular voltage, it
goes to the low conducting state and remains in the low will start increasing again. This process also accompanies a
conducting state till it reaches compliance at ꢀ2.75 V and stays metal filament formation by reaching the compliance current.
in that state till it reaches the maximum voltage in the negative This is a process which follows the high conducting pathway, but
direction at ꢀ4 V. When the I–V curve was traced back from the as we retrace back toward the negative bias, it goes through the
negative to the positive bias the current corresponding to the low conducting path, because the metal filament starts breaking
19098 | New J. Chem., 2018, 42, 19090--19100 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
Paper
NJC
and erase voltage pulse of 4 and ꢀ4 V with a pulse width of 0.5 s
and at a gap of 0.5 s, and then the states were probed by using a
read voltage pulse of 1.4 V. It was found that the read voltage
pulse after the write and erase pulse retains the states efficiently
even after 2000 s. This long retention time of the device is
important for practical memory applications.
Conclusions
We have successfully synthesized and characterized the Schiff’s
base ligand L1. Fe(II)-poly has been fully characterized by NMR,
FTIR, and other techniques. The coordination capability of the
ligand and model ligand shows some remarkable differences
with the same metal ions although the model ligand is just one-
half of the original ligand but both have the same coordination
sites. Both ligand and polymer have shown reversible redox
activity. The memory behaviour of Fe(II)-poly has shown bist-
ability with a unique Negative Differential Resistance (NDR)
effect, which is crucial for state-of-the-art electronic devices.
The observed reproducible nature of the device with consistent
performance for more than 1000 cycles shows the potential
application of the device in RAM devices. The prolonged reten-
tion time of more than 2000 s shows ROM performance of the
device. Although the retention time and the switching cycle are
still far from satisfactory when compared to the current Si
technology, the research on organic memory is still in its infancy,
and its device performance is expected to increase by tuning of
the ligand structure and fabrication process. Therefore this novel
polymer has great potential for future memory devices.
Fig. 9 The RAM application of the device was studied by using a write,
erase, and read voltage pulse of 4 V, ꢀ4 V, and 1.4 V respectively. The pulse
width was of 1 second and was applied for 4, 4, and 6 s respectively.
out and the current starts decreasing. The current will increase
with the increase in negative bias but will follow a low con-
ducting pathway. The current will reach compliance and when
we retrace back from negative to positive side it follows a high
conducting path.
We studied the random access memory (RAM) feature for the
device where we used the write, read, erase, and read sequence
as 4 V, 1.4 V, ꢀ4 V, and 1.4 V respectively (vide Fig. 9). We have
chosen 1.4 V as the read voltage because the On–Off ratio is
maximum at that voltage. We found that the device is reading
the memory reproducibly at each read sequence after the write
and erase voltage. The write and erase sequence would run for
4 s at a gap of 0.5 s and the read sequence will run for 6 s at a gap
of 0.5 s. We run 1000 such cycles and found that the reading is
consistent and the states are reproducible. Here we have shown
only a part of the whole sequence.
Funding source
The read only memory (ROM) application was also tested for
the device by observing its retention characteristics and here
also the device showed consistent performance (vide Fig. 10).
The write and erase sequences were run for 20 s with a write
Financial support received from the Science and Engineering
Research Board, DST New Delhi (under the grant EMR/2016/
001994), is acknowledged. DO is grateful to the Council of
Scientific and Industrial Research, New Delhi, for her fellow-
ship. US is thankful to MHRD for financial support.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
Thanks are due to Dr Suman Mukhopadhyay of IIT Indore, and
Dr Kaushik Ghosh and Dr Paritosh Mohanty of IIT Roorkee for
their kind help. We thank Mr Bidyut Kumar Kundu of IIT
Indore for performing the cyclic voltammetry experiment. DO is
also highly thankful to Mrs Monika Chaudhary of IIT Roorkee.
References
Fig. 10 The ROM application of the device was studied by applying a
write voltage of 4 V of width one second for 20 s and probed with a read
voltage of 1.4 V for more than 2000 s. Similarly an erase voltage of ꢀ4 V of
width 1 s for 20 s followed by a read voltage of 1.4 V for more than 2000 s
to probe the erase state of the device.
1 S. K. Deb, Appl. Opt., 1969, 8, 192–195.
2 F. S. Han, M. Higuchi, Y. Akasaka, Y. Otsuka and D. G. Kurth,
Thin Solid Films, 2008, 516, 2469–2473.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 19090--19100 | 19099

View Article Online
NJC
3 M. Higuchi, Polym. J., 2009, 41, 511–520.
4 A. Bandyopadhyay and M. Higuchi, Eur. Polym. J., 2013, 49,
1688–1697.
Paper
30 C. Wang, P. Gu, B. Hu and Q. Zhang, J. Mater. Chem. C,
2015, 3, 10055–10065.
31 P. Routh, R. K. Layek and A. K. Nandi, Carbon, 2012, 50,
3422–3434.
5 L. O. Chua and S. M. Kang, Proc. IEEE, 1976, 64, 209–223.
6 A. Bandyopadhyay, S. Sahu and M. Higuchi, J. Am. Chem. 32 J. Lin, M. Zheng, J. Chen, X. Gao and D. Ma, Inorg. Chem.,
Soc., 2011, 133, 1168–1171. 2007, 46, 341–344.
7 S. Goswami, A. J. Matula, S. P. Rath, S. Hedstrom, S. Saha, 33 U. Tritschler, F. Beck, H. Schlaad and H. Colfen, J. Mater.
M. Annamalai, D. Sengupta, A. Patra, S. Ghosh, H. Jani, Chem. C, 2015, 3, 950–954.
S. Sarkar, M. R. Motapothula, C. A. Nijhuis, J. Martin, 34 P. Ferreira, R. Z. Hou, A. Wu, M. G. Willinger,
¨
`
S. Goswami, V. S. Batista and T. Venkatesan, Nat. Mater.,
2017, 16, 1216–1224.
P. M. Vilarinho, J. Mosa, C. Laberty-Robert, C. Boissiere,
D. Grosso and C. Sanchez, Langmuir, 2012, 28, 2944–2949.
ˇ
8 A. Serres, M. Baudys and S. W. Kim, Pharm. Res., 1996, 13, 35 B. J. Scott, G. Wirnsberger and G. D. Stucky, Chem. Mater.,
196–201. 2001, 13, 3140–3150.
9 P. Gupta, K. Vermani and S. Garg, Drug Discovery Today, 36 Y. S. Lin, S. H. Wu, Y. Hung, Y. H. Chou, C. Chang,
2002, 7, 569–579.
10 H. Ahn, J. Hong, S. Y. Kim, I. Choi and M. J. Park, ACS Appl.
Mater. Interfaces, 2015, 7, 704–712.
M. L. Lin, C. P. Tsai and C. Y. Mou, Chem. Mater., 2006,
18, 5170–5172.
37 K. Yan, B. Chen, H. Hu, S. Chen, B. Dong, X. Gao, X. Xiao,
J. Zhou and D. Zou, Adv. Electron. Mater., 2016, 2, 1–7.
11 E. Fukada, IEEE Trans. Ultrason. Eng., 2000, 47, 1277–1290.
12 K. S. Ramadan, D. Sameoto and S. Evoy, Smart Mater. 38 V. Lyaskovskyy and B. De Bruin, ACS Catal., 2012, 2,
Struct., 2014, 23, 033001.
270–279.
13 M. Irie, Chem. Rev., 2000, 100, 1685–1716.
14 R. Pardo, M. Zayat and D. Levy, Chem. Soc. Rev., 2011,
40, 672.
15 P. Kiria, G. Hyett and R. Binionsa, Adv. Mater. Lett., 2010, 1,
86–105.
39 C. Kaes, A. Katz and M. W. Hosseini, Chem. Rev., 2000, 100,
3553–3590.
40 S. Kumar, D. N. Dhar and P. N. Saxena, J. Sci. Ind. Res., 2009,
68, 181–187.
41 H. Niu, H. Kang, J. Cai, C. Wang, X. Bai and W. Wang,
Polym. Chem., 2011, 2, 2804.
42 B. Pradhan and S. Das, Chem. Mater., 2008, 20, 1209–1211.
16 Y. Wang, E. L. Runnerstrom and D. J. Milliron, Annu. Rev.
Chem. Biomol. Eng., 2016, 7, 283–304.
17 M. Burnworth, L. Tang, J. R. Kumpfer, A. J. Duncan, 43 T. L. Choi, K. H. Lee, W. J. Joo, S. Lee, T. W. Lee and Y. C. Mi,
F. L. Beyer, G. L. Fiore, S. J. Rowan and C. Weder, Nature,
2011, 472, 334–337.
18 F. Niroui, A. I. Wang, E. M. Sletten, Y. Song, J. Kong,
J. Am. Chem. Soc., 2007, 129, 9842–9843.
44 Y. Z. Tang, M. Zhou, J. Huang, Y. H. Tan, J. S. Wu and
H. R. Wen, Inorg. Chem., 2013, 52, 1679–1681.
´
E. Yablonovitch, T. M. Swager, J. H. Lang and V. Bulovic, 45 S. Guieu, Synth. Commun., 2012, 42, 3177–3186.
˜
´
´
ACS Nano, 2015, 9, 7886–7894.
46 B. Comesana-Gandara, M. Calle, H. J. Jo, A. Hernandez,
J. G. de la Campa, J. de Abajo, A. E. Lozano and Y. M. Lee,
J. Membr. Sci., 2014, 450, 369–379.
¨
19 S. Forster and M. Konrad, J. Mater. Chem., 2003, 13,
2671–2688.
20 B. M. Novak, Adv. Mater., 1993, 5, 422–433.
21 L. D. Carlos, R. A. S. Ferreira, V. De Zea Bermudez, B. Julian-
47 G. Bhagavannarayana, R. V. Ananthamurthy, G. C. Budakoti,
B. Kumar and K. S. Bartwal, J. Appl. Crystallogr., 2005, 38,
768–771.
´
´
Lopez and P. Escribano, Chem. Soc. Rev., 2011, 40, 536–549.
´
´
22 L. Wang, M. H. Yoon, A. Facchetti and T. J. Marks, Adv. 48 L. Legrand, M. Abdelmoula, a. Gehin, a. Chausse and J.-M.
´
R. Genin, Electrochim. Acta, 2001, 46, 1815–1822.
Mater., 2007, 19, 3252–3256.
23 C. R. Kagan, D. B. Mitzi and C. D. Dimitrakopoulos, Science, 49 J. Matijevi and M. Vinkovi, Croat. Chem. Acta, 2006, 79,
1999, 286, 945–947. 489–495.
24 J. Even, L. Pedesseau, J. M. Jancu and C. Katan, J. Phys. 50 P. B. Tsitovich, F. Gendron, A. Y. Nazarenko, B. N. Livesay,
Chem. Lett., 2013, 4, 2999–3005.
A. P. Lopez, M. P. Shores, J. Autschbach and J. R. Morrow,
´
25 P. Gomez-Romero, M. Chojak, K. Cuentas-Gallegos, J. A.
Inorg. Chem., 2018, 57, 8364–8374.
˜
Asensio, P. J. Kulesza, N. Casan-Pastor and M. Lira-Cantu´, 51 D. Sharma, S. K. Sahoo, R. K. Bera and R. Kamal, J. Fluoresc.,
Electrochem. Commun., 2003, 5, 149–153. 2013, 23, 387–392.
26 H. Kim, S. H. Jo, J. H. Jee, W. J. Han, Y. Kim, H. H. Park, H. J. Jin, 52 X. Wang, X. Chen, A. Thomas, X. Fu and M. Antonietti, Adv.
B. Yoo and J. K. Lee, New J. Chem., 2015, 39, 836–842. Mater., 2009, 21, 1609–1612.
27 B. Hu, C. Wang, J. Wang, J. Gao, K. Wang, J. Wu, G. Zhang, 53 R. Dholam, N. Patel, M. Adami and A. Miotello, Int.
W. Cheng, B. Venkateswarlu, M. Wang, P. S. Lee and
Q. Zhang, Chem. Sci., 2014, 5, 3404–3408.
28 C.-L. Liu and W.-C. Chen, Polym. Chem., 2011, 2, 2169.
J. Hydrogen Energy, 2009, 34, 5337–5346.
54 T. Sato, R. K. Pandey and M. Higuchi, Dalton Trans., 2013,
42, 16036–16042.
29 B.-B. Cui, Z. Mao, Y. Chen, Y.-W. Zhong, G. Yu, C. Zhan and 55 C. Wu, F. Li, Y. Zhang, T. Guo and T. Chen, Appl. Phys. Lett.,
J. Yao, Chem. Sci., 2015, 6, 1308–1315. 2011, 99, 2009–2012.
19100 | New J. Chem., 2018, 42, 19090--19100 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018