Page 1
Antiferro-ferromagnetic transition in ultrathin Ni(OH)2 layer grown on
graphene surface and observation of interlayer exchange coupling in
Ni(OH)2/Graphene/Ni(OH)2 nanostructures
Shatabda Bhattacharya1, E. Mathan Kumar2, Ranjit Thapa2 and Shyamal K. Saha1,*
1Department of Materials Science,
Indian Association for the Cultivation of Science, Jadavpur, Kolkata-700032, India.2SRM Research Institute and Department of Physics and Nanotechnology, SRM University,
Kattankulathur - 603203, Tamil Nadu, India.
1. Characterizations
For Transmission Electron Microscope we used JEOL-2011 TEM. Powder X-ray diffraction
(XRD) with an X-ray diffractometer (RICH SEIFERT-XRD 3000P having X-Ray Generator-
Cu, 10 kV, 10 mA, λ 1.5418 Å). The X-ray photoelectron spectroscopy (XPS) is carried out
using OMICRON-0571 system. NICOLET MAGNA IR 750 system for Fourier transform
infrared spectroscopy spectra within the range 200-4000 cm-1. Atomic force microscopy is
done in the contact mode by PicoLE AFM equipment (Agilent Corp. USA). Electrical
measurements were carried out by a Keithley electrometer 6514 and for low temperature
setup JANIS close cycle cryostat has been used with Vacuum pressure >10 -7 bar in standard
four probe PPMS system. Magnetic measurements were done in a SQUID magnetometer
(Quantum design MPMS XL5).
Table S1. Parameters obtained after Rietveld analysis of XRD data for samples LC and MC
respectively.
1
Sample Phases
Lattice parameters (nm) Crystallite
size (nm)R.m.s
microstrain Standard deviationa c
HCgraphene 0.246 0.671 180.35 0.1045
0.011Ni(OH)2 0.311 0.461 13.03 0.038
MCgraphene 0.246 0.673 224.19 0.1082
0.012Ni(OH)2 0.310 0.464 12.2 0.067LC graphene 0.246 0.676 230.17 0.0931 0.011Ni(OH)2 0.312 0.465 8.05 0.203
Page 2
2. Charge transfer analysis by X-ray photoelectron spectroscopy
XPS study has been performed to characterize the ionic state of the elements present in the
samples. For the experimental verification of charge transfer from Ni to C at the interface
between graphene and Ni(OH)2, orbital energies of C and Ni atoms have been investigated in
detail by XPS. Fig. S1(a) shows the full range XPS spectrum of S1 and S3. Four intense
peaks are mainly observed for S1 around 112, 285, 533, and 854 eV for Ni 3s, C 1s, O 1s and
Ni 2p respectively. The magnified Ni spectra for S1 and S3 are given in Fig. S1(b). Two
intense peaks having binding energies 854.46 and 872.01 eV correspond to Ni 2p3/2 and 2p1/2
of the Ni(OH)2 phase for S1 and for S3, they appear at 855.70 and 874.18 eV respectively.
The C 1s spectrum appears at 285.01 eV for S1 and for S3 it is at 284.64 eV which is shown
in Fig. S1(c). In addition, the O 1s spectrum with a strong peak at 532.89 eV for S1 and
532.36 eV for S3 which are associated with hydroxide groups (OH-) of Nickel hydroxide.
Fig. S1(d) shows oxygen 1s spectrum.
Fig. S1. XPS spectra. (a) Full range spectra for S1 and S3. (b) comparison of Ni 2p peaks for S1
and S3. A blue shift is observed for S3. (c) C 1s spectra for S1 and S3. A red shift is observed
here for S3. (d) O 1s spectra.
Due to charge transfer from Ni to C at the interface, the orbital binding energy of carbon
decreases and that for nickel increases. As observed from the above analysis, C 1s energy
reduces from 285.01 to 284.64 eV and for Ni 2p orbital enegy increases from 854.46 to
855.70 eV. The shift in binding energy as indicated in Fig. S1(c) and S1(d) is due to charge
transfer process. As fractional charge is transferred from outermost orbital of nickel, after
charge transfer the binding energy on the orbitals due to coulombic nuclear force enhances on
2
Page 3
the rest of the charges (electron). The reverse process is occured in case of carbon where
these transferred charges fill up the orbitals of carbon thus reducing the coulombic nulcear
force on the outermost orbitals. The effect of this is observed in XPS analysis as discussed
above.
3. Raman and FTIR spectra
The raman spectra of the composite samples are shown in Fig. S2(a). For sample S1 and S3,
two prominent peaks arise at 1358 and 1601 cm-1 corresponding to D and G bands of
graphenes respectively. The D band signifies breathing modes of ring or K-point phonons of
A1g symmetry while G band represents the inplane bond stretching of the pairs of sp 2 bonded
C atoms (E2g phonons). Generally, the intensity ratio of D and G band (ID/IG) is used to
estimate the degree of disorder and defect present in the graphitic systems. It is also used to
determine the average size of sp2 domain. In case of graphene oxide, the ratio usually remains
less than unity but as it is reduced, D band becomes sharp and ID/IG ratio beomes greater than
unity. In this case the ratio is found to be 1.06 and 1.04 for S1 and S2 respectively.
Increament in the ratio is a signature of reduction of graphene oxide. In addition to these
peaks, two small peaks centered at 387 and 503 cm-1, which are identified as Eu(T) and A2u(T)
modes of Ni(OH)2 respectively. Thus Raman data successfully confirms Ni(OH)2 has been
anchored on the graphene sheets.
Fig. S2. Raman and FTIR spectra. (a) Raman images showing characteristics peaks of the
samples. (b) FTIR image of a typical sample.
FTIR technique was employed to further chracterize the graphene-Ni(OH)2 nanocomposites.
Fig. S2(b) shows the FTIR spectrum of a typical sample. The bands at ~3430 and ~1390 cm-1
are assigned to the stretching and deformation vibrations of O-H group. The band at 1058 cm-
1 is due to C–O stretching of the epoxides (C-O-C) present in graphene. The sharp band near
1631 cm-1 corresponds to O-H bending from absorbed H2O. Other bands near 2919 cm-1 and
3
Page 4
1247 cm-1 arise due to presence of residue acetate ions in the sample. Thus, from FTIR
analysis elemental groups that are present in the sample are identified.
4. Magnetic modeling of exchange constants
We have modeled our system to find the strength of coupling between the Ni2+ ions present in
the hexagonal lattice of Ni(OH)2 crystal. In two sublattice model using molecular field
approach, the standard Hamiltonian can be written as:
(S1)
Here Jij’s are the exchange coupling constants of the corresponding sites. Assuming Ising
type of ordering and considering the direction of applied magnetic field as parallel to c-axis,
the three exchange coupling constants J1, J2 and J3 can be equated to θ and TN by the
following expressions
(S2)
(S3)
Here kB is the Boltzmann constant and Xi are the number of nearest neighbours for the unit
cell of Ni(OH)2 as described in the inset of Fig. S3 and the values are 6, 2 and 12
respectively. The sign and magnitude of J1, J2 and J3 govern the strength and nature of
interactions between the interplanar (Out of plane) and intraplanar (in-plane) Ni2+ ions. Here,
we have considered exchange constants within the unit cell only.
Fig. S3. Unit cell of Ni(OH)2 showing interlayer and intralayer exchange constants
J1 is the first in-plane nearest neighbor which is the most significant for coupling between
the two spins. After that, J2 is the first out-of-plane nearest neighbor which connects the spins
of different hexagonal planes. Thus J2 describes the three dimensional ordering in the system
and J3 is the second out-of-plane next nearest neighbor. During evaluation of these exchange
constants for S3, we have taken θ=22.01 K from Curie-weiss fitting and TN=23.01 K obtained
4
Page 5
from the derivative plot of χT. Using S=0.76, we find J1/kB=4.20K, J2/kB=-0.18K and J3/kB=-
0.06K for sample S3. In case of S2, θ=17.29 K , TN=24.05 K and S=0.87. For S2 the values
come out as J1/kB=3.19K, J2/kB=-1.04K and J3/kB=-0.34K accordingly. While for the highest
concentration S1, J1/kB=2.96K, J2/kB=-1.07K and J3/kB=-0.35K using θ=17.42 K , TN=25.15 K
and S=0.93. After comparing J1 values for the three samples, it is to be mentioned that for S3
it is the highest which confirms the stronger ferromagnetic ordering for S3 other than the two
samples. Similarly for S1 case, the negative values of J2 and J3 signify, the presence of
antiferromagnetic ordering in the system which is equivalent to the bulk β-Ni(OH)2 phase.
Thus as the thickness of the Ni(OH)2 layer is decreased antiferromagnetic interaction
becomes weaker and ferromagnetic interaction becomes dominant and is the highest for S3.
5. Ab initio calculationTo investigate the charge transfer effect at the graphene/Ni(OH)2 interface, We have
performed the calculations using spin polarized density functional theory as implemented in
the Vienna ab-initio simulation package (VASP). The generalized gradient approximation
(GGA) – PBE functional was employed for the exchange and correlation effects of the
electrons. The potentials of the atoms were described by the projected augmented wave
(PAW) method. The cut-off energy is taken equal to 500 eV and the Brillouin zone
integration within the Monkhorst Pack scheme with 5×5×1 k-point mesh in reciprocal space.
All the structures were optimized until the total energy converged to less than 10-5 eV per
atom. Henkleman's Bader Charge analysis has been used to study the accumulation or
deficiency of charges on the given systems. The Ni(OH)2/graphene heterostructure is
constructed by stacking of (3×3×1) supercell of Ni(OH)2 on (4×4×1) supercell of graphene
shown in Fig. S4. The calculated lattice mismatch for this heterostructure is about 3.8% and
the interlayer distance between the Ni(OH)2 and graphene is around 2.763 Å. We have
considered 15 Å vacuum level along c-axis in the model geometry to avoid the spurious
interactions.
5
Page 6
Fig. S4. (a) Ni(OH)2/graphene heterostructure is constructed by stacking of (3×3) supercell of
NiOH2 on (4×4) supercell of graphene to minimize the lattice mismatch. Gray, Blue, Red and
white ball denote the carbon, nickel, oxygen and hydrogen atoms respectively. (b) Two layers of
Ni(OH)2 thin film of (3x3) supercell. Up and Down nickel atoms carried the equal and opposite
magnetic moment of about 0.159 μB. Gray, Blue, Red and white ball denote the carbon, nickel,
oxygen and hydrogen atoms respectively.
Our calculation result shows that, overall the layered Ni(OH)2 material is antiferromagnetic in
nature with ferromagnetic coupling along a plane. In this study we have consider two layer of
Ni(OH)2 thin film of (3x3) supercell shown in Figs. S4(a) and (b).
The Ni atoms in the system are bonded with four OH molecules to form the layered structure.
Each Ni atom exhibits magnetic moment of +1.59 μB corresponds to top layer and the bottom
layer Ni atom shows −1.59 μB magnetic moment and this value is exactly matched with other
result . In our calculations the effective magnetic moment µ is related to the spin value (S) via
the relation µ = gSμB. Here g value is considered as 2 for Ni(OH)2 phase. Using the relation
the spin S for each Ni atom is found to be about 0.795. This is similar to the observed
experimental value. In case of Ni(OH)2/graphene heterostructure, the magnetic moment of Ni
atom at the top layer is unaltered, but the magnetic moment of each Ni atom at down layer
(interact with graphene) is decreases to about 1.57 μB. For which the total magnetic moment
of the Ni(OH)2/graphene heterostructure is found to be about 0.14 μB and is shown in Fig.
S4(b).
6
Page 7
Fig. S5. (a) 3D plot of charge density difference isosurfaces of NiOH2-graphene heterostructure.
Blue lobs indicate charge depletion region and yellow lobs shows charge accumulation region,
here the isosurfaces value is taken as 2 ×10-3 e Bohr-3. Partial density of states (PDOS) of (b) Ni
atoms corresponds to top layer, (c) Ni atoms of down layer of the heterostructure. The red and
black arrow indicates the up and down spin. Fermi level (Ef) is indicated by the dashed vertical
line.
To understand the rise of ferromagnetic behavior of this heterostructure we have studied the
charge transfer and nature of bonding between the hero-layer. Charge Density Difference
(CDD) is calculated and it is shown in Fig. S5(a). From the CDD plot we have found that the
charge is accumulated (indicated by yellow region) on the carbon atoms correspond to
graphene layer and the charge depletion loop (indicated by blue region) has been formed on
the Ni and hydrogen atoms of the bottom layer of the Ni(OH)2 thin film. This observation
clearly explain the charge transfer from Ni(OH)2 layer to the graphene layer.
The charge transfer from the particularly Ni atoms to carbon atoms is also confirmed by the
Bader charge analysis. The net negative Bader charge value for graphene layer is found to be
-0.137 e compare to free graphene layer. For deeper insight we have studied the electronic
properties of the Ni(OH)2/graphene heterostructure. We observed that the asymmetry in the
DOS near the Fermi level is mainly due to Ni atoms. The PDOS for Ni atoms in the up and
down layers of NiOH2 is shown in Fig. S5(b) and (c). From the PDOS of both up and down
layer exhibit a magnetic semiconductor with a small bandgap. In case of Ni atoms in the top
layer the up spin electrons contribution is more in PDOS. Whereas for bottom layers the
down spin contribution is higher. Overall, due to charge sharing with the graphene the density
of electrons with up spin of Ni atoms at top layer is higher than the density of electrons with
down spin of Ni atoms at bottom layer.
However, we have also calculated the charge redistribution between 4s and 3d states of Ni.
Please see below the figure S6 and Table S2. It clearly shows the charge transfer from the Ni
states to graphene state. From Table S2, the integral value of 4s orbital is much less than 3d
orbital which indicates the stronger coupling between 3d and pz orbitals.
7
Page 8
Figure S6. Partial density of states (PDOS) of (a) 4s and (b) 3d states correspond to Ni atom
of top layer and PDOS of (c) 4s and (d) 3d states correspond to Ni atom of bottom layer of
Ni(OH)2/Graphene. The red and blue arrow indicates the up and down spin. Fermi level (Ef)
is indicated by the dashed vertical line.
Table S2. Demonstrate the occupancy of 4s and 3d states correspond to Ni atom of bottom
layer of bare Ni(OH)2 surface and Ni(OH)2/Graphene heterostructure. The calculated
occupancy is for single Ni atom, so per unit supercell the change in occupancy is higher. The
unit of occupancy is e. To be noted as we have considered plane wave approach for periodic
crystal, so little occupancy value has been observed for s electron also.
System∫−∞
0
ρ4 s dE ∫−∞
0
ρ3 d dE
Ni(OH)2 0.3242 8.1833
Ni(OH)2/Graphene 0.3170 8.1142
6. Magnetotransport measurement
Due to the charge transfer effect occurring at the interface of graphene/Ni(OH)2 and the
magnetic coupling among the ferromagnetic layers separated by nonmagnetic graphene
spacer, the overall magnetotransport property of the system is sensitive to the presence of
localized states. Fig. S7 shows the variation of magnetoconductance as a function of field at
8
Page 9
different temperatures for the two samples S1 and S2. For these two samples only an IEC is
observed in the hysteresis loop and during magnetotransport measurement a change in
conductance under application of magnetic field. At higher temperature region a positive
unsaturated magnetoconductance is detected and at low temperature a negative
magnetoconductance is perceived. The transition from positive to negative
magnetoconductance is occurred near about 150K for S1 and 80K for S2. In the present case
graphene/Ni(OH)2 is a disordered semiconductor system. In these systems, the extent of
disorder governs the relative importance of the roles played by localization effects and
electron-electron (e-e) interactions. At low temperature, e-e interaction is dominated giving
rise to negative magnetoconductance but as temperature rises, this interaction is destroyed
and weak spin orbit coupling is activated to give positive magnetoconductance at high
temperature. In case of sample S3, no effective magnetoconductance is observed throughout
the temperature because of the non-appearance of IEC due to strong ferromagnetic layers
grown on single side of graphene surface.
Since the magnetoconductance is sensitive to the extent of disorder, it serves as an especially
useful tool for identifying the microscopic transport mechanisms and scattering processes.
The magnetoconductivity is determined by contributions from localization and electron-
electron interaction. At low temperature, negative magnetoconductance is due to strongly
localized states that causes wave function shrinkage in a magnetic field.
Fig. S7. Magnetoconductance curves. (a) For S1. (b) For S2. Dotted points are the experimental
ones and red line for the fitted with the modified equations.
The magnetic field dependence for the contribution to the conductance due to electron-
electron interactions can be written as
9
Page 10
(S4)
Where a is the coefficient. This equation has been used to fit the observed
magnetoconductance data and it is well fitted at low temperature with a variation of H1/2 as
shown in Fig. S7(a).
At high temperature range this magnetic field variation could be ascribed to the weak spin-
orbit coupling acting at the interface as a result of charge transfer. In 2D systems spin orbit
coupling is greatly affected by the electric fields acting normally to the surface. The
fractional charge transfer from nickel to graphene at the interface creates a potential barrier
due to charge sharing. In the limit of weak spin orbit coupling affected by this charge
transfer, the magnetoconductance as a function of field can be expressed as
(S5)
This equation has been used to fit the experimentally obtained magnetoconductance at high
temperature region and fitted well with H2 variation depicted in Fig. S7(b). Hence it is seen
that for both the samples appearance of IEC is related to magneto transport in this case and
their field dependence are explained on the basis of charge localization model.
10