APPLIED PHYSICS LETTERS 99, 052505 (2011)

Ferromagnetism in C-doped SnO2 thin films

Nguyen Hoa Hong,1,a) J.-H. Song,1 A. T. Raghavender,1 T. Asaeda,2 and M. Kurisu2
1
Nanomagnetism Laboratory, Department of Physics and Astronomy, Seoul National University, Seoul
151-747, South Korea
2
Department of Physics, Graduate School of Science and Engineering Ehime University, Matsuyama
790-8577, Japan
(Received 31 March 2011; accepted 8 July 2011; published online 3 August 2011)
Room temperature ferromagnetism (FM) was observed in laser ablated C-doped SnO2 thin films
grown on LaAlO3 substrates. The greatest value of the saturated magnetization (at 1 T) is found in
SnO2 films doped with 1% of C (about 16 emu/cm3) and reduces lightly as the C concentration
increases. By subtracting the base of pure SnO2 that is also magnetic, one can find that C-doping
actually enhances ferromagnetism in the host SnO2. It suggests that carbon really causes
defect-induced magnetism of about 3.91 lB/C into SnO2. Measurements on C-doped SnO2 bulks
show that the observed properties are unique for films (lower dimensions, having surface/interface
effect). The investigation on thickness dependence shows a change in magnetization when going
from thin to thick films, implying that somehow if magnetism is due to defects then those must be
located more on or near the surface than in deeper layers. Anisotropy is also observed, enforcing
the assumption for the origin of magnetism due to defects in C-doped SnO2. V C 2011 American

Institute of Physics. [doi:10.1063/1.3617439]

Following the theoretical prediction,1 over several years, moment. Both oxygen vacancies and confinement effects are
experimental groups have doped transition-metals (TM) into supposed to be equally important to induce FM in SnO2.13
many semiconducting oxides such as ZnO, TiO2, SnO2, and Certainly, it has confirmed that SnO2 could be classified
In2O3 in order to find new potential candidates for spintronic into a group where magnetism is possible without magnetic
applications. Actually, room temperature ferromagnetism impurities.
(FM) was found in those compounds under the thin film In 2010, a theoretical group investigated the case of car-
form.2 However, more recently, the observation of FM in bon doping, to see if defects can induce FM in SnO2.14 It
TiO2, HfO2, In2O3, and ZnO thin films confirmed that mag- was predicted that C can create surface-magnetism in SnO2,
netism is certainly possible in many types of pristine semi- and the magnetic moment was calculated to be 2 lB/C. Since
conducting and insulating oxide thin films, without any C is known as an element that can create obvious defects, it
special need of magnetic dopants that contain 3d electrons is interesting to elucidate this specific case. In this letter, we
that were usually thought to induce FM via the Ruderman- will report on structural and magnetic properties of C-doped
Kittel-Kasua-Yoshida (RKKY) interaction.3–7 Many data SnO2 thin films, thus give a direct experimental feedback to
have suggested that the origin of such an observed FM such an important theoretical statement.
should be oxygen vacancies and/or defects.4–8 The 150 nm-thick Sn1xCxO2 films (where x ¼ 0; 0.01;
As for Cr-Ni-V-doped SnO2 films,9 firstly, the observed 0.05; and 0.1) and a 45 nm-thick Sn0.95C0.05O2 film were
FM was supposed to be introduced due to doping, similarly grown on (001) LaAlO3 (LAO) substrates by using the
to what the other groups also reported.10,11 However, there pulsed laser deposition (PLD) technique (KrF laser with
are two facts that are important to clarify better here. First, k ¼ 248 nm) from corresponding ceramic targets made by a
the reported magnetic moments for low concentrations were solid state reaction method. Note also that the targets are
too big to be blamed on the spin-only values or secondary pure with no impurity peaks that could be detected from x-
phase, so that more likely the observed magnetism must ray diffractions (XRD) (impurities are of less than 102 wt.
come from electronic or lattice defects.12 Additionally, we %). The repetition rate of film deposition was 10 Hz and the
also investigated the case of pure SnO2 thin films.13 SnO2 is energy density was 1.8 J/cm2. The substrate temperature was
somewhat different to other oxides mentioned earlier, since 700  C. During deposition, the oxygen partial pressure (PO2)
there is no 3d electron in it to suggest that RKKY interaction was kept as 104 Torr, and after deposition, films were
should play a main role in introducing FM. Room tempera- cooled down to room temperature under the same PO2 as
ture FM was also observed in SnO2 films suggesting that during deposition, with a rate of 20  C/min. The structural
pristine semiconducting oxides could be ferromagnetic if ox- study was done by XRD with Cu-radiation and energy dis-
ygen vacancies and confinement effects could be formed.13 persive spectroscopy (EDS). The magnetic moment (M)
Data of Mn-doped SnO2 films showed that Mn doping does measurements were performed by a quantum design super-
not play any key role in introducing FM into SnO2, but just conducting quantum interference device (SQUID) system
degrades its structure, and therefore, reduces its magnetic under magnetic field (H) from 0 up to 0.5 T under a range of
temperatures (T) from 350 K down to 5 K.
a)
Author to whom the correspondences should be addressed. Electronic From Fig. 1, XRD data as typical example for SnO2
mail: nguyenhong@[Link]. samples doping with 1% of carbon have shown that a good

0003-6951/2011/99(5)/052505/3/$30.00 99, 052505-1 C 2011 American Institute of Physics

V

Downloaded 03 Aug 2011 to [Link]. Redistribution subject to AIP license or copyright; see [Link]
 052505-2 Nguyen et al. Appl. Phys. Lett. 99, 052505 (2011)

TABLE I. EDS data for the Sn0.90C0.10O2 target.

Element Weight % Atom %

C 2.14 7.4
O 25.87 67.34
Sn 71.99 25.26

room temperature ferromagnetic. This is confirmed also

from the magnetization versus temperature curve taken at
0.5 T for a SnO2 doped with 1% of C shown in the inset of
Fig. 3. One can see that the sample is ferromagnetic for the
whole range of temperature below 350 K. As we have al-
ready known, the laser ablated undoped SnO2 thin film could
be ferromagnetic.13 But as seen from these new data for C-
doped SnO2 films, the magnetic moment found previously in
SnO2 is obviously enhanced by doping with carbon. The
largest saturated magnetic moment is found in SnO2 films
doped with 1% of C. Increasing C concentration does not
seem to cause much change. The coercivity also does not
FIG. 1. X-ray diffraction patterns for (a) the Sn0.99C0.01O2 target and (b) a vary in any remarkable way (HC  0.013 T). Subtracting the
Sn0.99C0.01O2 film. magnetism that may come from oxygen vacancies in the
undoped SnO2 host, and then, we can estimate the magnetic
SnO2 phase is obtained even from the target (Fig. 1(a)) and moment that should come from carbon doping as about 3.91
our C-doped SnO2 films are single phase (Fig. 1(b)) with lB/C. In the modelling paper of Rahman and Garcı́a-
only well oriented peaks of SnO2 appear in the spectra. Lat- Suárez,14 a simulated value of 2 lB/C was attributed to sur-
tice parameters do not vary much as the C concentration face-magnetism, emphasizing that carbon are located only at
change (c/a ¼ 0.674; 0.679; 0.676; and 0.67 for x ¼ 0%, 1%, the surface layers. However, our larger value of magnetiza-
5%, and 10%, respectively). It seems that C is well incorpo- tion implies that in reality, carbon may also locate in deeper
rated into the SnO2 host lattice. This is also confirmed by layers. We will verify this point later when investigating
EDS data. Even though EDS is not a good method to give a thickness dependence of magnetic moment of C-doped SnO2
precise value of compositions, especially for films, one still films.
can verify if the dopant exists in the sample or not. EDS Usually for many oxide films, the surface/interface
measurements performed on C-doped SnO2 targets (that effect is present.4,6,7 In order to verify this, we investigate
from which the films were made) give us this proof. One can the magnetic properties of films with different thickness. In
see from either the EDS spectra in Fig. 2 or from the corre- general, thinner films are expected to have a poorer crystal-
sponding element concentrations (for a typical Sn0.9C0.1O2) linity (i.e., more defects) than thicker films, in which, the
listed in Table I that carbon is certainly present in the upper layers essentially grow on a buffer layer formed by the
sample. lower levels of the film at the substrate interface. Thus, an
Magnetization versus magnetic field data taken at 300 K increase in defect density in the thinner films over the thicker
for 150 nm-thick Sn1xCxO2 films are shown in Fig. 3. All films can contribute more defect-induced magnetism in com-
films of C-doped SnO2 films (C ¼ 0%; 1%; 5% and 10%) are parison to thicker ones. M-H curves taken at 300 K for films
of Sn0.95C0.05O2 with different thickness are shown in Fig. 4.

FIG. 3. (Color online) Magnetization versus magnetic field at 300 K as

magnetic field is applied parallel to the film plane for the 150 nm-thick
Sn1xCxO2 films (where x ¼ 0; 0.01; 0.05; and 0.1). The inset shows the M
FIG. 2. EDS data for the Sn0.90C0.10O2 target. (T) curve taken at 0.5 T for the Sn0.99C0.01O2 film.

Downloaded 03 Aug 2011 to [Link]. Redistribution subject to AIP license or copyright; see [Link]
 052505-3 Nguyen et al. Appl. Phys. Lett. 99, 052505 (2011)

FIG. 4. Magnetization versus magnetic field taken at 300 K as magnetic FIG. 5. Magnetization versus magnetic field taken at 300 K as magnetic
field is applied parallel to the film’s plane for the 45 nm-thick and 150 nm- field is applied parallel and perpendicular to the film’s plane for the
thick Sn0.95C0.05O2 films. The inset shows M(H) curve taken at 300 K for Sn0.99C0.01O2 film.
the Sn0.95C0.05O2 bulk (a piece cut from the corresponding target).

One can see that even there is some difference in the satu- defect-induced magnetism in the compound with magnetic
rated magnetization values of the 150 nm-thick and the 45 moment of about 3.91 lB/C. Measurements on C-doped SnO2
nm-thick films, this discrepancy is not very large. In other bulks show that the observed properties are unique for films
words, the thinner film (about 3 times thinner) has only mag- (2D). The strong anisotropy is also observed, enforcing the
netic moment of 2 times larger than that of the thick one. It assumption for the origin of magnetism due to defects in C-
is much less, in comparison to the known case of HfO2 or doped SnO2. The last, but not the less important point to men-
TiO2 films, that the difference in a similar case can be up to tion here is that, our data on thickness dependence shows that
20-40 times (more than 1 order) leading to a conclusion that C-defects should exist also in deeper layers but not only at/
in those films,4 if magnetism is due to oxygen vacancies or near the surface as theory suggested. In general, this experi-
defects, then those defects must be located mostly near to the mental feedback should be useful for fabricating nanometer
surface/interface, but not in the bulk. On the contrary, in the sized materials of SnO2 for applications.
case of C-doped SnO2 films as we see here, defects are The authors would like to thank R&D SNU Foundation
located more at/near the surface, but also are present for the financial support for our project 3348-20100016.
throughout the bulk. Again, as déjà vu in other oxides, one
must see that this defect-magnetism anyhow is not a property 1
T. Dietl, [Link], F. Matsukura, J. Cibert, and D. Ferrand, Science 287,
that we can find in ceramics, but only in thin films.4,13 The 12019 (2000).
2
M-H data of a piece of the Sn0.95C0.05O2 target is shown in 3
N. H. Hong, J. Magn. Magn. Mater. 303, 338 (2006).
the inset of Fig. 4. It is seen that the Sn0.95C0.05O2 bulk is M. Venkatesan, C. B. Fitzgerald, and J. M. D. Coey, Nature 430, 630
(2004).
diamagnetic. It means that besides the effect of oxygen 4
N. H. Hong, J. Sakai, N. Poirot, and V. Brizé, Phys. Rev. B 73, 132404
vacancies (that might be caused by film deposition) and (2006).
5
defects that were caused by C-doping, one should also take S. D. Yoon, Y. Chen, A. Yang, T. L. Goodrich, X. Zuo, D. A. Arena, K.
Ziemer, C. Vittoria, and V. G. Harris, J. Phys.: Condens. Matter 18, L355
into account the confinement effect of thin film form.4,7,8 As
(2006).
for SnO2, it appears that both factors are equally important. 6
N. H. Hong, J. Sakai, and V. Brizé, J. Phys.: Condens. Matter 19, 036219
In order to check if there is a strong anisotropy in this (2007).
7
case or not, we have perform the M-H measurement at room A. Sundaresan, R. Bhagravi, N. Rangarajan, U. Siddesh, and C. N. R. Rao,
Phys. Rev. B 74, 161306(R) (2006).
temperature for the Sn0.99C0.01O2 film in both configurations, 8
A. Barla, N. H. Hong, E. Beaurepaire, P. Imperia, J. P. Kappler, J. C.
when magnetic field is applied parallel and perpendicular to Cezar, N. B. Brookes, and J. Sakai, XMCD data, ESRF 2007
the film’s plane. One can see from Fig. 5 that while in paral- (unpublished).
9
lel configuration, the sample is purely ferromagnetic, it N. H. Hong, J. Sakai, W. Prellier, and A. Hassini, J. Phys.: Condens.
Matter 17, 1697 (2005); N. H. Hong, A. Ruyter, W. Prellier, J. Sakai, and
behaves a bit differently in the perpendicular case. The curve N. T. Huong, J. Phys.: Condens. Matter 17, 6533 (2005); N. H. Hong and
measured in H perpendicular shows a mixture of ferromag- J. Sakai, Physica B 358, 265 (2005).
10
netic and paramagnetic components. This observed strong S. B. Ogale, R. J. Choudhary, J. P. Buban, S. E. Lofland, S. R. Shinde, S.
anisotropy reinforces our conclusions for defect-induced N. Kale, V. N. Kulkarni, J. Higgins, C. Lanci, J. R. Simpson, N. D. Brown-
ing, S. D. Sarma, H. D. Drew, R. L. Greene, and T. Venkatesan, Phys.
magnetism in C-doped SnO2 thin films.4,8 Rev. Lett. 91, 77205 (2003).
In conclusion, ferromagnetism at room temperature was 11
J. M. D. Coey, A. P. Douvalis, C. B. Fitzgerald, and M. Venkatesan, Appl.
found in C-doped SnO2 thin films grown on LaAlO3 sub- 12
Phys. Lett. 84, 1332 (2004).
strates by pulsed laser deposition technique. The largest value C. B. Fitgerald, M. Venkatesan, L. S. Dorneles, R. Gunning, P. Stamenov,
and J. M. D. Coey, Phys. Rev. B 74, 115307 (2006).
of the saturated magnetization is found in Sn0.99C0.01O2 films 13
N. H. Hong, N. Poirot, and J. Sakai, Phys. Rev. B 77, 33205 (2008).
(about 16 emu/cm3), and it does not change much when the C 14
G. Rahman and V. M. Garcı́a-Suárez, Appl. Phys. Lett. 96, 052508
concentration increases. It reveals that C doping creates a (2010).

Downloaded 03 Aug 2011 to [Link]. Redistribution subject to AIP license or copyright; see [Link]