Solid State Communications 152 (2012) 1625–1629

Contents lists available at SciVerse ScienceDirect

Solid State Communications

journal homepage: www.elsevier.com/locate/ssc

Room-temperature ferromagnetism in N þ -implanted ZnO:Mn thin films

H.B. Ruan a,b, L. Fang a,n, G.P. Qin b, T.Y. Yang b, W.J. Li b, F. Wu a, M. Saleem a, C.Y. Kong b
a
Department of Applied Physics, Chongqing University, Chongqing 400044, China
b
Optical Engineering Key Lab, Chongqing Normal University, Chongqing 400030, China

a r t i c l e i n f o abstract

Article history: Mn–N co-doped ZnO films with wurtzite structure were fabricated by RF magnetron sputtering
Received 22 November 2011 together with the ion-implantation technique. Then a post-annealing at 650 1C for 10 min in a N2
Received in revised form atmosphere was performed to activate the implanted N þ ions and recover the crystal quality, and a
11 April 2012
p-type ZnO:Mn–N film with a hole concentration of about 2.1  1016 cm  3 was obtained. It is found
Accepted 14 April 2012
that the Mn mono-doped ZnO film only exhibits paramagnetic behavior, while after N þ -implantation, it
by H. Akai
Available online 16 May 2012 shows ferromagnetism at 300 K, and the magnetization of the ZnO:Mn–N films can be further enhanced
by thermal annealing due to the activation of the N acceptors. Our experimental results confirm that
Keywords: the codoping N acceptors are favorable for ferromagnetic ordering of Mn2 þ ions in ZnO, which is
A. Mn–N co-doped ZnO
consistent with the recent theoretical calculations.
B. N þ -implantation
& 2012 Elsevier Ltd. All rights reserved.
C. Bound magnetic polarons
D. Ferromagnetism

1. Introduction N and severe compensation of donors in ZnO, the reported Mn–N

co-doped ZnO samples are usually of n-type or semi-insulating
In the quest for materials involving the charge and spin conductivity. The synthesis of p-type Mn–N co-doped ZnO is still
degrees of freedom of electrons in a single substance, since the an experimental challenge.
theoretical study predicted that the Mn doped ZnO (ZnO:Mn) Among many methods of attempting nitrogen doping,
would show ferromagnetism with Curie temperature (TC) well ion-implantation was reported to be a promising technique to
above room temperature (RT) [1], ZnO based diluted magnetic enhance the solubility of the N acceptors in ZnO. Moreover, this
semiconductors (DMSs) have attracted much attention due to method offers a precise doping through the control of dose, area
their potential applications in ‘‘spintronics’’. However, recent localization, and dopant depth. Several groups have obtained
experimental studies on ZnO:Mn materials resulted in very good p-type conductivity in ZnO using this method [18–20]. In
controversial results, i.e. paramagnetism [2,3], spin glass [4], the present work, we fabricated the Mn–N co-doped ZnO films
and anti-ferromagnetism to ferromagnetism at low temperatures using RF magnetron sputtering in combination with N þ -implan-
or even at high temperatures have been reported in ZnO:Mn tation, and then annealed the samples in an N2 ambient to
system [5–14]. According to first principles calculations [15–17], activate the N acceptors and recover the induced lattice disorder.
if carriers are not doped into ZnO:Mn, the superexchange inter- It is found that the incorporation of N acceptors is favorable for
action between the neighboring Mn2 þ ions is antiferromagnetic the ferromagnetic ordering of doped Mn2 þ ions in ZnO.
in character while ferromagnetism can be stabilized only by hole
doping, e.g. through N substitution for O. However, ferromagnet-
ism has also been observed in heavily n-type ZnO:Mn [8,9]. Thus 2. Experimental
the circumstances under which ZnO:Mn can be ferromagnetic is
still debatable. Nitrogen has been considered to be a promising ZnO:Mn thin films were deposited on quartz glass substrates
acceptor impurity for making p-type ZnO. Actually, several by radio-frequency (RF) magnetron sputtering. A high-purity
researchers have obtained N doped ZnO:Mn materials by various ceramic ZnO (99.99%) disk (60 mm in diameter), along with some
techniques, such as chemical synthesis [10], sputtering [11,12], smaller Mn slices (  5 mm in diameter) placed equidistantly on
plasma enhanced chemical vapor deposition [13] and thermal ZnO disk, was used as the sputtering target. The Mn doping
oxidation of Zn3N2:Mn [14]. However, due to the low solubility of concentrations can be adjusted by varying the number of Mn
slices. In this experiment, we have controlled the Mn-doping level
at  8 at%. The sputtering chamber was first evacuated to a
n
Corresponding author. Tel.: þ86 23 65678369. base pressure below 8  10  4 Pa with a turbo molecular pump.
E-mail address: fangliangcqu@yahoo.com.cn (L. Fang). Then ultrapure Ar was introduced into the chamber. During film

0038-1098/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ssc.2012.04.059
1626 H.B. Ruan et al. / Solid State Communications 152 (2012) 1625–1629

1.8x1021 Table 1
Electrical properties of ZnO:Mn, as-implanted and annealed ZnO:Mn–N films.

sum Samples Carrier Mobility Resistivity Seebeck Type

1.5x1021
concentration (cm2 V  1 s  1) (Ocm) coeff.
(cm  3) (mV/K)
N concertration (cm-3)

1.2x1021
ZnO:Mn 2.0  1018 13.76 0.22  174 n
100keV 200keV as-implanted 1.7  1017 2.50 14.33  40 n
ZnO:Mn–N
9.0x1020 annealed 2.1  1016 2.17 134.90 þ11.6 p
ZnO:Mn–N

6.0x1020

X-ray Intensity (a.u.)

3.0x1020
pure ZnO

0.0 a
0 100 200 300 400 500 600 700
b
Depth (nm)

X-ray Intensity (a.u.)

c
Fig. 1. (Color online) N concentration profiles simulated with the SRIM software
for the as-implanted sample.
33 34 35 36
ZnO:Mn 2θ (degree)
deposition, the sputtering pressure and RF power were kept at
2 Pa and 120 W. The N ions were implanted into these samples.
as-implanted ZnO:Mn-N
Ion implantation was performed at 300 K with N þ ions of
energy 200 keV (dose of 5  1016 cm  2) plus 100 keV (dose of
2.6  1016 cm  2). The distribution of the implanted N þ ions in
ZnO simulated by TRIM code was shown in Fig. 1. To activate the
implanted N þ ions and recover the crystal quality, the implanted annealed ZnO:Mn-N
samples were annealed in an N2 ambient.
The thickness of the films was about 700 nm measured by a
step profiler. The crystal structure and the surface morphology of 30 33 36 39 42
the samples were characterized by X-ray diffraction (XRD) (Cu Ka 2θ (degree)
radiation source) and high-resolution field-emission scanning
Fig. 2. (Color online) XRD spectra of ZnO:Mn, as-implanted and annealed
electron microscopy (FE-SEM) respectively. Electrical properties
ZnO:Mn–N films. The inset shows the diffraction peak position of the doped
of the films were investigated by a standard four-probe measure- ZnO films in comparison with that of pure ZnO film.
ment. Chemical bonding states and chemical compositions of the
films were analyzed by X-ray photoelectron spectroscopy (XPS).
The magnetic properties of all the films were measured using a the (002) plane of ZnO and no phases related to manganese oxide
commercial physical property measurement system (PPMS) from or nitride can be observed, which indicate that all the films are
Quantum Design fitted with a vibrating sample magnetometer. single phase and have wurtzite structure with c-axis preferential
All the samples were handled very carefully to avoid any trace orientation. Since N þ -implantation induced severe microscopic
magnetic contamination. The magnetic field was applied parallel structure disorder and lattice strain in the ZnO:Mn films, the ZnO
to the film plane. (002) diffraction intensity was reduced and the full width at half
maximum (FWHM) was considerably broadened. However, when
the as-implanted film was annealed, to a great extent, the ZnO
3. Results and discussion (002) diffraction peak was recovered, implying an improvement
in the crystal quality of the codoped ZnO layer. Moreover, for
The resistivity and Hall measurements of the doped ZnO films doped ZnO films, a slight shift of XRD peaks toward the lower
were performed and the experimental data were listed in Table 1. diffraction angle was also observed in comparison with that of the
The Mn mono-doped ZnO films were of n-type conductivity pure ZnO film, implying a significant increase in lattice constant
with an electron concentration of about 2.0  1018 cm  3. After for the Mn mono-doped and Mn–N co-doped ZnO films, as shown
N þ -implantation, the electron concentration of the films was in the inset of Fig. 2. It is understandable that both the ionic radii
decreased by about one order of magnitude. However, due to the of Mn2 þ (80 pm) and N3  (146 pm) are relatively larger than that
significant compensation defects and/or localization of holes of Zn2 þ (74 pm) and O2  (138 pm). Similar observations have
occurring in the films, the as-implanted ZnO:Mn–N films were also been reported in Refs. 11 and 12.To further evaluate the
still n-type conductive. Post-annealing is needed to remove these structure of the ZnO:Mn–N films, the surface morphology of the
defects and activate N atoms by moving them to the right lattice films were studied by FE-SEM. Fig. 3 shows the surface morphol-
sites. When the samples were post-annealed at 650 1C for 10 min, ogy of the co-doped ZnO films before and after annealing. It is
we obtained one p-type ZnO:Mn–N film with a hole concentra- seen that the surface of the both films is smooth and the crystal
tion of about 2.1  1016 cm  3 [21]. To further confirm the p-type grains are uniformly distributed, which suggest that the films we
conduction, we also measured the Seebeck coefficients of the obtained are of acceptable crystallinity.
films in the temperature range of 20–40 1C, and the correspond- In order to identify the composition and the chemical states
ing values are listed in Table 1. of the ZnO:Mn–N films, XPS measurements were conducted and
Fig. 2 shows the XRD patterns of the ZnO:Mn, as-implanted the binding energy scale was calibrated using the adventitious C
and annealed ZnO:Mn–N films. Only one peak corresponding to 1s peak at 284.6 eV. The core level peaks of Zn 2p, O 1s, Mn 2p
 H.B. Ruan et al. / Solid State Communications 152 (2012) 1625–1629 1627

Fig. 3. Surface morphologies of (a) as-implanted and (b) annealed ZnO:Mn–N films.

2p 3/2
530.7eV
1021.7eV
Intensity (a.u.)

Intensity (a.u.)
2p 1/2

1010 1020 1030 1040 1050 1060 526 528 530 532 534 536 538 540
Binding energy (eV) Binding energy (eV)

2p 1/2 397.2eV 399.5eV

2p 3/2
641.0eV
Intensity (a.u.)

Intensity (a.u.)

630 635 640 645 650 655 660 665 392 394 396 398 400 402 404 406 408 410
Binding energy (eV) Binding energy (eV)

Fig. 4. XPS spectra of (a) Zn 2p, (b) O 1s, (c) Mn 2p, and (d) N 1s core levels for the annealed p-type ZnO:Mn–N film.

and N 1s XPS spectra for the annealed ZnO:Mn–N film were which is the desired location for p-type doping acting as
presented in Fig. 4. By XPS quantitative analysis, the relative acceptors in ZnO [22]. The other N 1s peak located at higher
element composition of the film is Zn 48.93%, O 47.02%, Mn binding energy of 399.5 eV is usually ascribed to several defects,
4.12% and N 0.93%. The binding energy of Zn 2p3/2 is located at such as (NC)O, N–O and/or (N2)O complexes in ZnO [23–25].
1021.7 eV, suggesting a single component of Zn2 þ ions in the These defects are known as shallow donors and could cause
ZnO lattice. The O 1s peak at 530.7eV is ascribed to O2  ions in server compensation for the acceptors (NO) in ZnO. Therefore,
Zn–O and Mn O bonds. The binding energy of Mn 2p3/2 state is even if a high concentration of N was incorporated into ZnO
located at 641.0 eV, no signals from metallic Mn (637.7 eV) and lattice, the p-type ZnO:Mn–N film we obtained only has such
Mn4 þ ions (642.4 eV) are detected, indicating the doped Mn ions low hole concentration of 2.1  1016 cm  3.
are mainly in divalent states. From the Fig. 4(d), one can see The magnetic field dependence of magnetization (M versus H)
clearly that the N 1s spectra can be well fitted by two Gaussian of the ZnO:Mn, as-implanted and annealed ZnO:Mn–N films were
curves. The N 1s peak around 397 eV is generally attributed to measured at 300 K, as shown in Fig. 5. Since the used quartz glass
the signal of the substitution of N atom for O sublattice (NO), substrate is diamagnetic, to eliminate the influence of substrates,
1628 H.B. Ruan et al. / Solid State Communications 152 (2012) 1625–1629

1.5 introduces carriers. Exchange interactions between one localized

ZnO:Mn(a) c hole and the surrounding Mn2 þ ions align all the Mn2 þ spins
1.2
as-implanted ZnO:Mn-N(b) around the hole localization center, forming a BMP. As the (No)
0.9 annealed ZnO:Mn-N(c) b defect density is increased, the overlapping of neighboring BMPs
can result in the long range Mn2 þ –Mn2 þ ferromagnetic coupling
Magnetization (µB/Mn)

0.6
in ZnO:Mn–N films and thus a macroscopic spin polarization
0.3 a becomes favorable. Certainly, if more effective N acceptors are
activated upon thermal annealing, as N and Mn atoms prefer to
0.0
0.52
exist as the nearest neighbors [13,15–17], thus more BMPs are
H = 1000 Oe
-0.3 formed corresponding to the larger magnetic moment. So it is

M (µB/Mn)
0.48 easy to understand that the stronger ferromagnetism can be
-0.6
0.44 FC
obtained in the annealed p-type ZnO:Mn–N films. Recently,
-0.9 whether ferromagnetism occurs only in a p-type ZnO:Mn or not
0.40 is still debatable. Even for the (Mn, N) co-doped ZnO, the
-1.2 0 100 200 300
Temperature (K)
ferromagnetic behavior was observed not only in p-type [14],
-1.5 but also in n-type [11,12] conducting or nearly insulating [10,13]
-6000 -4000 -2000 0 2000 4000 6000 samples. In our study, considering RT ferromagnetism has been
H (Oe) observed for both n-type and weak p-type ZnO:Mn–N films, we
propose that the existence of ferromagnetism in ZnO:Mn system
Fig. 5. (Color online) Room temperature M–H curves of ZnO:Mn, as-implanted
has no direct correlation with the conductivity type for the
and annealed ZnO:Mn–N films. The inset shows the M–T curve of the annealed
p-type film measured in a magnetic field of 1 kOe.
specific samples.

M–H curves of substrates were also measured under an identical 4. Conclusions

procedure after treatment. Then the diamagnetic contribution of
the quartz glass substrates was subtracted in the raw data. From ZnO:Mn–N thin films with c-axis preferred orientation were
the M–H curve of the ZnO:Mn film, it can be seen that the Mn fabricated by radio-frequency (RF) magnetron sputtering in
mono-doped ZnO film only exhibits paramagnetic properties. combination with N þ -implantation. In comparison with the Mn
As for the M–H curve of the as-implanted ZnO:Mn–N film, it mono-doped ZnO film only exhibits paramagnetic properties, RT
shows a clear hysteresis loop with a saturation magnetization ferromagnetism has been observed not only in weak p-type but
of 0.65 mB/Mn and a coercive field of  200 Oe, revealing a also in n-type ZnO:Mn–N films. Additionally, we also found that
room-temperature ferromagnetic characteristic. Obviously, the the magnetization of the annealed p-type ZnO:Mn–N film
incorporation of N ions plays a crucial role in the activation (1.35 mB/Mn) is larger than that of the as-implanted sample
of RT ferromagnetic behavior in the ZnO:Mn systems, which (0.65 mB/Mn). These results clearly indicate that the codoping N
is consistent with the recent experiments [10–14] and theoretical acceptors are favorable for ferromagnetic ordering of Mn2 þ ions
calculations [15–17]. Furthermore, we also found that the mag- in ZnO, in agreement with the recent theoretical calculations. The
netic behavior of the ZnO:Mn–N films can be enhanced upon mechanism of ferromagnetic coupling in our samples is discussed
thermal annealing. The saturation magnetic moment of the based on the bound magnetic polaron model.
annealed ZnO:Mn–N film is  1.35 mB/Mn, much larger than that
(  0.65 mB/Mn) of the as-implanted one. The inset of Fig. 5 shows
the temperature dependence of magnetization (M–T) curve of the Acknowledgments
annealed ZnO:Mn–N film. It is seen that the magnetization
decreases slightly with increasing temperature. However, the This work was supported by the Fundamental Research Fund
transition from ferromagnetic to paramagnetic state does not for the Central Universities no. CDJXS10102207, the National
occur in the temperature range of 5–300 K, indicating a TC in Natural Science Foundation of China (Grant nos. 11075314 and
excess of at least 300 K. As the possible mixed phase, including 50942021) and the Natural Science Foundation of Chongqing City
Mn metal and almost all the Mn-based oxides (except Mn3O4 under Grant CSTC, 2011BA4031, the Third Stage of ‘‘211’’ Innova-
with TC 43 K), is antiferromagnetic, thus the observed ferromag- tive Talent Training Project (no. S-09109) and the Sharing Fund
netism in our samples should be most probably intrinsic. It is also of Large-scale Equipment of Chongqing University (Grant nos.
noted that the measured saturation magnetization (  1.35 2010063072 and 2010121556).
mB/Mn) in our case is still much smaller than the theoretical
value of 5 mB for a free Mn2 þ ion, but larger [12] than or in
References
agreement with [10,11]values previously reported, which usually
is ascribed to antiferromagnetic superexchange interactions
[1] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287 (2000) 1019.
between neighboring Mn2 þ ions [9,26]. [2] T.L. Phan, Solid State Commun. 151 (2011) 24.
To explain the observed ferromagnetic behavior in our [3] S. Kolesnik, B. Dabrowski, J. Appl. Phys. 96 (2004) 5379.
samples, it should be noted the Zener model [1] does not apply [4] T. Fukumura, Z. Jin, M. Kawasaki, T. Shono, T. Hasegawa, S. Koshihara,
H. Koinuma, Appl. Phys. Lett. 78 (2001) 958.
here since the hole concentration (2.1  1016 cm  3) is too small to [5] Y.M. Kim, M. Yoon, I.W. Park, Y.J. Park, J.H. Lyou, Solid State Commun. 129
induce a robust ferromagnetism. Alternatively, the ferromagnetic (2004) 175.
mechanism here could be attributed to the formation of bound [6] H.Y. Xu, Y.C. Liu, C.S. Xu, Y.X. Liu, C.L. Shao, R. Mu, J. Chem. Phys. 124 (2006)
074707.
magnetic polarons (BMPs) [10,12]. When ZnO:Mn films were
[7] G. Lawes, A.S. Risbud, A.P. Ramirez, R. Seshiadri, Phys. Rev. B 71 (2005)
implanted with N ions, though n-type conductivity has been 045201.
determined by Hall measurements, it is still possible that some [8] Z. Yang, J.L. Liu, M. Biasini, W.P. Beyermann, Appl. Phys. Lett. 92 (2008)
uncompensated or localized holes exist in the samples. Generally, 042111.
[9] M. Ivill, S.J. Pearton, Y.W. Heo, J. Kelly, A.F. Hebard, D.P. Norton, J. Appl. Phys.
the hole is tightly localized by the effective N acceptor (No). 101 (2007) 123909.
Doping Mn atoms provides the localized spins while doping N [10] K.R. Kittistved, N.S. Norberg, D.R. Gamelin, Phys. Rev. Lett. 94 (2005) 147209.
 H.B. Ruan et al. / Solid State Communications 152 (2012) 1625–1629 1629

[11] Z.B. Gu, M.H. Lu, J. Wang, D. Wu, S.T. Zhang, X.K. Meng, Y.Y. Zhu, S.N. Zhu, [19] B.T. Adekore, J.M. Pierce, R.F. Davis, D.W. Barlage, J.F. Muth, J. Appl. Phys. 102
Y.F. Chen, X.Q. Pan, Appl. Phys. Lett. 88 (2006) 082111. (2007) 024908.
[12] H.Y. Xu, Y.C. Liu, C.S. Xu, Y.X. Liu, C.L. Shao, R. Mu, Appl. Phys. Lett. 88 (2006) [20] C.Y. Kong, G.P. Qin, H.B. Ruan, M. Nan, R.J. Zhu, T.L. Dai, Chin. Phys. Lett. 25
242502. (2008) 1128.
[13] W.S. Yan, Z.H. Sun, Q.H. Liu, Z.R. Li, T.F. Shi, F. Wang, Z.M. Qi, G.B. Zhang, [21] H.B. Ruan, L. Fang, G.P. Qin, C.Y. Kong, J. Nanosci Nanotechnol 11 (2011) 11049.
S.Q. Wei, Appl. Phys. Lett. 89 (2007) 242509. [22] Y. Zhu, S.S. Lin, Y.Z. Zhang, Z.Z. Ye, Y.F. Lu, J.G. Lu, B.H. Zhao, Appl. Surf. Sci.
[14] Z.L. Lu, G.Q. Yan, S. Wang, W.Q. Zou, Z.R. Mo, L.Y. Lv, F.M. Zhang, Y.W. Du, 255 (2009) 6201.
M.X. Xu, Z.H. Xia, J. Appl. Phys. 104 (2008) 033919. [23] Craig L. Perkins, S. Lee, X. Li, Sally E. Asher, Timothy J. Coutts, J. Appl. Phys. 97
[15] Q. Wang, Q. Sun, P. Jena, Y. Kawazoe, Phys. Rev. B 70 (2004) 052408. (2005) 034907.
[16] L. Zhao, P.F. Lu, Z.Y. Yu, X.T. Guo, Y. Shen, H. Ye, G.F. Yuan, L. Zhang, J. Appl. [24] K. Tang, S. Gu, S. Zhu, J. Liu, H. Chen, J. Ye, R. Zhang, Y. Zheng, Appl. Phys. Lett.
Phys. 108 (2010) 113924. 95 (2009) 192106.
[17] M.H.N. Assadi, Y.B. Zhang, S. Li, J. Phys.:Condens. Matter 21 (2009) 185503. [25] P. Hoffmann, C. Pettenkofer, Phys. Status Solidi B 248 (2) (2011) 327–333.
[18] C.C. Lin, S.Y. Chen, S.Y. Cheng, H.Y. Lee, Appl. Phys. Lett. 84 (2004) 5040. [26] Q. Wang, Q. Sun, B.K. Raol, P. Jena, Phys. Rev. B 69 (2004) 233310.