March 22, 2011 20:31 S1793-6047 00152

Functional Materials Letters

Vol. 4, No. 1 (2011) 1–4
© World Scientific Publishing Company
DOI: 10.1142/S179360471100152X

ROOM TEMPERATURE FERROMAGNETISM IN PURE AND

Cu DOPED ZnO NANORODS: ROLE OF COPPER OR DEFECTS

SHALENDRA KUMAR∗ , B. H. KOO and C. G. LEE†

School of Nano & Advanced Materials Engineering
Changwon National University, 9 Sarim Dong
Changwon 641-773, Republic of Korea
∗shailuphy@gmail.com
†chglee@changwon.ac.kr

SANJEEV GAUTAM and K. H. CHAE

Nano Analysis Center, Korea Institute of Science and Technology
Seoul 136-791, Republic of Korea
S. K. SHARMA and M. KNOBEL
Instituto de Fisica Gleb Wataghin, Universidade Estadual de Campinas
(UNICAMP) Campinas, 13.083-859, SP, Brazil

Received 7 October 2010; Revised 10 January 2011

We report structural, magnetic and electronic structure studies of pure and Cu doped ZnO nanorods with the aim to understand the
origin of ferromagnetism. A structural study indicates that all the samples exhibit single phase nature and rules out the formation
of secondary phase. NEXAFS measurements reveal that Cu atom exists in Cu2+ state. Magnetic hysteresis loop measurements
reflect that the pure and Cu doped ZnO nanorods exhibit room temperature ferromagnetism. The increase in the intensity of green
emission in photoluminescence study indicates that defects density increases with Cu doping.

Keywords: DMS; ZnO; x-ray diffraction; photoluminescence; NEXAFS; DC magnetization.

In the recent years dilute magnetic semiconductors (DMS) has
attracted the interest of researchers due to their potential appli-
cation in spintronics devices.1–3 Initially, it was assumed that
DMS and their properties can be tailored by partial replace-
ment of cations in semiconducting host matrix via transition
metal (TM) cations. But the discovery of the unexpected mag-
netism in HfO2 ,4 TiO2 ,5 ZnO6 etc. has triggered a debate on
the origin of ferromagnetism in the DMS. Several experimen-
tal as well as theoretical groups have reported the origin of
ferromagnetism in the nanoparticles or thin films of diamag-
netic oxides. Though, among the diamagnetic oxides showing
the d 0 ferromagnetism, zinc oxide (ZnO) has been recognized
as one of the most important materials for the application in
the multifunctional devices because of its various properties.
Even though ferromagnetism has been observed in a
number of systems, experimental studies on TM doped ZnO
have produced inconsistent results and the mechanism of
∗,†Corresponding Authors.
ferromagnetism (FM) in TM doped ZnO remains unclear. It
is considered that the TM dopants in ZnO form clusters or
secondary phase therefore having many controversies about
whether the observed FM is an intrinsic or extrinsic property
of the material. These facts motivated us to study the FM in
Cu doped ZnO. As it is well known that Cu is a potential mag-
netic ion only in +2 state with +1/2 spin and neither metallic
Cu nor its oxide (CuO and Cu2 O) is ferromagnetic. In these
circumstances, Cu doped ZnO has potential of showing mag-
netic behavior at RT because FM observed in Cu based sys-
tem will undoubtedly be the intrinsic property of the material.
Despite the above fact, Cu doped ZnO system is not far away
from controversies. There are many contradictory reports in
which some authors confirmed the presence of FM in this sys-
tem whereas others ruled out same.7–10 These controversial
results indicate that FM in DMSs is very sensitive to prepara-
tion method and conditions.
In this letter, we have reported room temperature ferro-
magnetism in Cu doped ZnO synthesized by co-precipitation

1
March 22, 2011 20:31 S1793-6047 00152
2 S. Kumar et al.
technique. The effect of Cu doping has been studied using
x-ray diffraction (XRD), transmission electron microscopy
(TEM), photoluminescence, DC magnetization and near edge
x-ray absorption fine structure spectroscopy (NEXAFS) mea-
surements. Zn1−x Cux O (0.0 ≤ x ≤ 0.05) nanorods were
prepared by co-precipitation method. Appropriate propor-
tion of analytical grade metal nitrate [Zn(NO3)2 ·6H2 O and
Cu(NO3 )2 · 3H2 O] were thoroughly mixed according to
desired stoichiometry and is reported elsewhere.11 XRD mea-
surements were carried out at room temperature using a
Phillips X’pert (MPD 3040) with a Cu K α source (λ =
1.54 Å). Magnetization measurements were performed at
room temperature using a commercial Quantum Design Phys-
ical properties measurement system (PPMS). TEM measure-
ments were performed using HR-TEM (JEM 2100F). The
NEXAFS measurements of these samples along with the ref-
erence compounds were performed at the soft x-ray beam line
at 7B1 XAS and 10B XRS of KIST-PAL. The data was nor-
malized and processed using Athena 0.8.056/IFEFFIT 1.2.11.
Figure 1 shows XRD patterns for Cu doped ZnO
nanorods. The patterns were analyzed and indexed using
Powder-X software which indicates that all samples exhibit
single phase nature with wurtzite lattice and exclude the pos-
sibility of secondary phase formation within the limit of XRD.
From the XRD analysis, it is observed that the peaks position
shifted towards lower 2θ values with Cu contents (see inset
in Fig. 1) which suggest that the lattice parameters increased
with Cu doping. Such a shift in XRD with Cu doping has
been reported in literature by Tiwari et al.12 for Cu doped
ZnO thin films and assumed to be an additional evidence for
Cu occupying the Zn sites in ZnO matrix. Moreover, the HR-
TEM and SAED pattern (due to the brevity of article, not
shown here) also indicate that there is no formation of nano
Fig. 1. (Color online) XRD pattern of Zn1−x Cux O (0.0 ≤ x ≤ 0.05). Inset
shows the extended view of (002) plane.
Fig. 2. (Color online) Photoluminescence spectra of Zn1−x Cux O (0.0 ≤
x ≤ 0.05). Inset shows M-H curve of Zn1−x Cux O (0.0 ≤ x ≤ 0.05).
size crystallite of Cu2 O and CuO. The inter-planer distance
(d = 2.63 = (002) plane) calculated from the HR-TEM
corresponds to the wurtzite phase of ZnO which suggests that
Cu ions are occupying the Zn position in ZnO matrix.
Inset in Fig. 2 represents the magnetic hysteresis loop of
pure and Cu doped ZnO measured at 300 K. Magnetization
of Zn1−x Cux O (0.0 ≤ x ≤ 0.05) performed as a function of
magnetic field shows that pure and Cu doped ZnO samples
display the ferromagnetic ordering at RT. The coercive field at
300 K for pure, 1 and 5% Cu doped ZnO is 88, 48 and 40 Oe
respectively. It is clear from the M–H curve that the magnetic
moment of Cu doped ZnO is higher compared to pure ZnO.
The origin of FM in Cu-ZnO is still not clear. Ye et al.13 pre-
dicted that FM in Cu–ZnO originates from the hybridization
of the Cu 3d and O 2 p bands. However, Hou et al.10 explained
FM in Cu-ZnO on the basis of indirect exchange interactions
using the carrier mediated model and Herng et al.14 eluci-
dated FM in the light of bound magnetic polarons (BMP)
model. By comparing our results to the reported in the lit-
erature, the saturation magnetization (M S ) values are smaller
than those obtained for Cu–ZnO.10–13 In the present case, this
observed magnetic behavior may be explained in the light
of BMP model. The NEXAFS measurements performed at
Cu K -edge clearly indicate that Cu has 3d9 electronic config-
uration, therefore, the FM may results due to the Cu doping
as well as defects.
The normalized room temperature photoluminescence
(PL) spectra for pure and Cu doped ZnO nanorods is shown in
Fig. 2. For pure ZnO, a high intensity sharp peak at 385 nm in
the UV-region is observed, while other peaks were detected in
blue-green region. Till now, it is agreed unanimously that the
March 22, 2011 20:31 S1793-6047 00152

Room Temperature Ferromagnetism in Pure and Cu Doped ZnO Nanorods: Role of Copper or Defects 3

UV emission originated from the near-band-edge emission

(NBE) transition in band gap of ZnO called recombination
of free excitons through an exciton-exciton collision process
while the mechanism of the blue-green emission is related to
oxygen vacancy or Zn interstitials.15,16 The origin of blue-
green emission in ZnO has remained controversial for the last
decades and it has been assumed that defects, TM doping, and
the oxygen vacancies are responsible for the green emission in
ZnO. Various groups have reported different theory of green
emission in ZnO. Reynold et al.17 proposed that the structured
green emission in ZnO is due to a transition from two shal-
low donors to a deep acceptor while Leiter et al.18 suggested
that the unstructured green emission in ZnO results from a
triplet excited state of neutral oxygen vacancy. However, in the
present case, we have observed two peaks with high intensity
in UV and green region as compared to pure ZnO. Previously,
Wang et al.19 reported that blue-green emission positions and
intensities vary with the annealing temperature and time. In
line with their findings, it is clear from our results that TM
doping can also vary the DLE position and relative intensi-
ties (see Fig. 2). The high intensity of the green emission in
Cu doped ZnO suggests that the copper impurity creates deep
level defect state in the band gap of ZnO. Dingle et al.16 sug- Fig. 3. (color online) (a) Cu K -edge spectra of Zn1−x Cux O (0.0 ≤ x ≤
gested that DLE in Cu doped ZnO is due to the charge transfer 0.05).

between Cu2+ ions and adjacent oxygen atoms (hole transfers

from d shell to sp3 orbitals). Graces et al.20 extended this the-
ory and proposed that Cu may exist in Cu1+ or Cu2+ state. If
Cu is in +1 state then it will give the structure less emission
that is due to donor-acceptor pair recombination involving the
copper acceptor and a shallow donor impurity whereas in pres-
ence of Cu2+ ions, a sharp emission peak similar to Dingle is
originated due to short lived transfer of an electron from neigh-
boring oxygen to the Cu ions. However, the room temperature
PL spectrum of Cu doped ZnO shows that green emission has
a blueshift with relatively high intensity compare to pure ZnO
which further support that Cu ions are substituted for tetrahe-
drally coordinated Zn2+ ions in ZnO matrix. Moreover it is
found that UV emission has blueshift for 1 and 5% Cu doping
in comparison to the pure ZnO. It can be seen from the PL
spectra that 1% Cu doped ZnO has the highest intensity of the
green emission peaks which indicate that the density of the
defects is higher in case of 1% Cu doped ZnO.
Figure 3 shows the Cu K -edge NEXAFS spectra of Cu
doped ZnO along with reference spectra of Cu metal, CuO
and Cu2 O. It can be noticed that the pre-edge spectral fea-
ture located ∼8985 eV exist for spectra except Cu2 O and Cu
metal foil (see inset of Fig. 3). This spectral feature arises
due to the quadrupole 1s → 3d transition and has a signif-
icant intensity because of 3d−2 p orbital mixing. In general,
pre-edge centroid position depends strongly on the Cu oxida-
tion state; however the pre-edge intensity is mostly influenced
by the Cu coordination geometry. The low intensity of the
pre-edge peak refer to the geometries with center of symme-
try (e.g., octahedral); whereas the high intensity refer to non-
centrosymmetric geometries (e.g., tetrahedral). However, this
transition is allowed only when the empty d state are available
such as in the 3d 9 electronic state of the Cu2+ ion and the struc-
tural unit around the absorbing atoms lacks centro-symmetry.
The Cu1+ ion prohibit such a transition due to 3d 10 electronic
configuration. The observed NEXAFS spectral features of Cu
of doped ZnO are very similar to those of CuO which indicates
that Cu ions are in +2 state.
Figure 4(a) displays O K -edge NEXAFS spectra for
pure and Cu doped ZnO nanorods. Based on the existing lit-
erature and band structure calculations, the spectral feature
in the energy range of 530–539 eV can be attributed to the
hybridization between O 2 p and Zn/Cu 3d/4s states, which
form the bottom of the conduction band. The spectral fea-
tures between 540 and 550 eV is attributed to the hybridization
of O 2 p and Zn/Cu 4 p orbitals. The spectral feature above
550 eV can be assigned to the hybridization between O 2 p
and Zn and Cu 4d states. By comparing the NEXAFS spec-
tra of pure and Cu doped ZnO, it is observed that two extra
spectral features evolve at ∼529 (marked by A1 ) and 532 eV
(marked by B1 ). It is evident from spectra that the intensity of
the peak at ∼529 increases with Cu doping which indicates
the strong hybridization of O 2 p orbital with Cu 3d orbitals
[see Fig. 4(b)]. The appearance of these new spectral features
(at ∼529 and 532 eV) and increase in their intensity with
March 22, 2011 20:31 S1793-6047 00152

4 S. Kumar et al.

in ZnO. Magnetization studies reflect that pure and Cu doped

samples show RTFM and magnetization increases with Cu
doping. Therefore, PL, NEXAFS and magnetization results
show that RTFM results from the defects and Cu ions.

Acknowledgment
This work was supported by BK21 project corp. SKS and MK
acknowledge the financial support from FAPESP and CNPq,
Brazil.

References
Fig. 4. (Color online) (a) O K -edge spectra of Zn1−x Cux O (0.0 ≤ x ≤ 1.
0.05). (b) Difference spectra of Zn1−x Cux O (0.0 ≤ x ≤ 0.05). Inset shows 2.
normalized O K -edge NEXAFS spectra of Cu doped ZnO (0.0 ≤ x ≤ 0.05). 3.
4.
Cu doping demonstrates that doping of Cu introduced unoc-
5.
cupied 3d states within the band gap of ZnO. Thakur et al.21
reported similar results in case of Mn doped ZnO. Singh 6.
et al.22 also studied the effect of defects and oxygen vacan- 7.
cies using Ar-ion implantation and Ar annealing on O K -edge
spectra of ZnO. They observed that the intensity of peak at 8.
∼532 eV increases with increase in number of defects in ZnO.
9.
In order to see the effect of Cu or defect, we further analyzed
the O K -edge spectra. Inset in Fig. 4(b) shows the normalized 10.
O K -edge NEXAFS spectra. The intensity of O K NEXAFS
spectra is proportional to the density of unoccupied state and 11.
the oxygen concentration. It can be seen from the inset in the
12.
Fig. 4(b) that the intensity of the spectral features increases
with Cu doping and has maximum intensity for 1% Cu doped 13.
ZnO. The increase in the intensity of the spectral features is a
direct indication of increase in the number of O 2 p holes in 14.
Cu doped samples. Krishnamurthy et al.23 reported that spec-
15.
tral features at 537 eV appears due to the presence of the O
vacancy. In the present work, we observed same features at 16.
537 eV, which intensity increases with Cu doping and has a 17.
higher value for 1% doping. In contrast, the Zn K -edge (due
to the brevity of the article not shown in the manuscript) of all 18.
the samples closely resembles each other, reflecting that the
19.
doping of Cu and oxygen vacancy barely influences the elec-
tronic state of Zn and thereby ruling out the existence of Zn 20.
interstitials. From NEXAFS study, it is clear that Cu doping
enhances the defects in the Cu doped ZnO, which is in good 21.
agreement with PL studies.
22.
In summary, we have successfully synthesized single
phase Cu doped ZnO nanorods using co-precipitation method. 23.
NEXAFS measurements show that Cu ions exist in +2 states.
PL study indicates that Cu doping enhanced the defects
G. A. Prinz, Science 282, 1660 (1998).
H. Ohno, Science 281, 951 (1998).
T. Dietl, Acta Phys. Pol. A 100, 139 (2001).
C. Das Pemmaraju and S. Sanvito, Phys. Rev. Lett. 94, 217205
(2005).
X. Zuo, S.-D. Yoon, A. Yang et al., J. Appl. Phys. 103, 07B911
(2008).
H. Ohno, J. Magn. Magn. Mater. 200, 110 (1999).
D. Chakraborti, J. Narayan and J. T. Prater, Appl. Phys. Lett. 90,
062504 (2007).
D. J. Keavney, D. B. Buchhold, Q. Ma et al., Appl. Phys. Lett.
91, 012501 (2007).
D. P. Buchholz, R. P. H. Chang, J. H. Song et al., Appl. Phys.
Lett. 87, 082504 (2005).
D. L. Hou, X. J. Ye, H. J. Meng et al., Appl. Phys. Lett. 90,
142502 (2007).
S. Kumar, Y. J. Kim, B. H. Koo et al., J. App. Phys. 105, 07C520
(2009).
A. Tiwari, M. Snure, D. Kumar et al., Appl. Phys. Lett. 92,
062509 (2008).
L. H. Ye, A. J. Freeman and B. Delley, Phys. Rev.B. 73, 033203
(2006).
T. S. Herng, S. P. Lau, S. F. Yu et al., J. App. Phys. 99, 086101
(2006).
K. Vanheusden, C. H. Seager, W. L. Warren et al., Appl. Phys.
Lett. 68, 403 (1996).
R. Dingle, Phys. Rev. Lett. 23, 579 (1969).
D. C. Reynolds, D. C. Look and B. Jogai, J. App. Phys. 89, 6189
(2001).
F. H. Leiter, H. R. Alves, A. Hofstaetter et al., Phys. Status Solidi
B 226, R4 (2001).
X. Wang, R. Zheng, Z. Liu et al., Nanotechnology 19, 455702
(2008).
N. Y. Garces, L. Wang, L. Bai et al., Appl. Phys. Lett. 81, 622
(2002).
P. Thakur, K. H. Chae, J. Y. Kim et al., Appl. Phys. Lett. 91,
162503 (2007).
A. P. Singh, R. Kumar, P. Thakur et al., J. Phys.: Condens.
Matter. 21, 185505 (2009).
S. Krishnamurthy, C. McGuinness, L. S. Dorneles et al., J. Appl.
Phys. 99, 08M111 (2006).