Thermochromic effect in NdNiO3?? thin films annealed in ambient air

Abstract

The synthesis of NdNiO3 perovskite structure was achieved by soft post deposition annealing of initially amorphous thin films reactively sputter deposited on silicon substrates. The physical measurements were fully consistent with the properties expected for the thermochromic NdNiO3 phase. Upon heating, the optical transmission that was correlated with the electrical properties decreased in the infrared domain showing a thermochromic effect in this optical region. The metal–insulator transition temperature was found to be −68 °C for the specimen tested and the jump in resistance at the transition was 1.4 orders of magnitude. The state-of-the-art methods that normally involve an annealing at a high oxygen pressure (200 × 105 Pa) or epitaxial stabilization were considerably improved in this work by the use of dc sputtering and the optimization of the deposition conditions. Therefore, the novel soft process proposed here opens up numerous research possibilities.

Full text

Thermochromic effect in NdNiO3−δ thin films annealed in ambient air
This article has been downloaded from IOPscience. Please scroll down to see the full text article.
2009 J. Phys. D: Appl. Phys. 42 182006
(http://iopscience.iop.org/0022-3727/42/18/182006)
Download details:
IP Address: 193.50.29.182
The article was downloaded on 10/05/2011 at 15:35
Please note that terms and conditions apply.
View the table of contents for this issue, or go to the journal homepage for more
Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS
J. Phys. D: Appl. Phys. 42 (2009) 182006 (4pp) doi:10.1088/0022-3727/42/18/182006
FAST TRACK COMMUNICATION
Thermochromic effect in NdNiO3−δthin
films annealed in ambient air
F Capon1, D Horwat1, J F Pierson1, M Zaghrioui2and P Laffez2
1Institut Jean Lamour, D´
epartement CP2S, (UMR CNRS 7198), Ecole des Mines, Parc de Saurupt,
CS 14234, 54042 Nancy Cedex, France
2Universit´
e de Tours, IUT de Blois, Laboratoire d’Electrodynamique des Mat´
eriaux Avanc´
es, UMR
CNRS CEA 6187, Place Jean Jaur`
es, 41029 Blois cedex, France
E-mail: fabien.capon@mines.inpl-nancy.fr
Received 9 July 2009, in final form 10 August 2009
Published 28 August 2009
Online at stacks.iop.org/JPhysD/42/182006
Abstract
The synthesis of NdNiO3perovskite structure was achieved by soft post deposition annealing
of initially amorphous thin films reactively sputter deposited on silicon substrates. The
physical measurements were fully consistent with the properties expected for the
thermochromic NdNiO3phase. Upon heating, the optical transmission that was correlated
with the electrical properties decreased in the infrared domain showing a thermochromic effect
in this optical region. The metal–insulator transition temperature was found to be −68 ◦C for
the specimen tested and the jump in resistance at the transition was 1.4 orders of magnitude.
The state-of-the-art methods that normally involve an annealing at a high oxygen pressure
(200 ×105Pa) or epitaxial stabilization were considerably improved in this work by the use of
dc sputtering and the optimization of the deposition conditions. Therefore, the novel soft
process proposed here opens up numerous research possibilities.
Transition metal oxide-based materials find many applications
in electronic and optical devices due to their wide range of
electrical, magnetic and optical properties. Among these,
the optimization of thermochromic or electrical switching is
challenging. The RENiO3materials (RE =Pr, Nd, Sm or Gd)
are of interest for their sharp metal–insulator (MI) transition
easily tunable around room temperature [1]. Therefore the
light transmission, reflection or emissivity in the IR region
may be passively controlled, placing this material as a potential
candidate for sensors, modulated switches or for furtivity
applications [2] in the field of smart material coatings.
Compounds of this family crystallize in the orthorhombic
perovskite structure (space group Pbnm) and exhibit a metal–
insulator transition temperature (TMI)of−140 ◦C, −70 ◦C,
130 ◦C and 290 ◦C for RE =Pr, Nd, Sm and Gd,
respectively [3]. Since the electronic properties of oxides
show a significant dependence on oxygen content, additional
oxidation treatments after synthesis are required. In the case of
RENiO3(nickelates), the nickel must adopt the Ni3+ oxidation
state to form the perovskite structure, which is difficult to
stabilize. Powder or ceramics synthesis usually requires
annealing in an oxidation atmosphere with pressure varying
from 2 ×106up to 6 ×109Pa, depending on the synthesis
method [3–7]. Other variations in the process route have
been described. However, nickelates prepared following these
routes tend to be oxygen deficient, and show only marginal
transformation, characterized by a change in the slope of
the resistance–temperature curve. So far, only specimens
produced using high temperature and high pressure conditions
have shown the characteristic first-order MI phase transition.
In the field of thin films, NdNiO3has also been synthesized by
RF sputtering [8–11] with high oxygen pressure post-annealing
treatments to obtain optimum electrical switching properties.
The high oxygen pressures currently required to produce good
quality nickelates limit the access for wider research in these
materials.
The experimental conditions need to be simplified for
NdNiO3-based systems and applications. It has been shown
[12,13] that, owing to the production of Ni3+ ions in the
plume during deposition, pulsed-laser deposition (PLD) can be
0022-3727/09/182006+04$30.00 1© 2009 IOP Publishing Ltd Printed in the UK

J. Phys. D: Appl. Phys. 42 (2009) 182006 Fast Track Communication
used for preparing NdNiO3thin films under relatively mild
conditions without the need for extreme conditions of post-
deposition oxidant annealing. Concurrently, epitaxial RENiO3
thin films with RE =Pr–Gd were synthesized under reduced
oxygen pressures (∼2×103Pa) by metal organic chemical
vapour deposition (MOCVD) [14,15]. Nevertheless, the
thermochromic nickelates can only be reached by MOCVD
or PLD on cubic or pseudo-cubic single crystalline substrates
presenting a lattice parameter close to that of the phases of
interest: the stress imposed by the substrate on the film leads
to thermodynamic stabilization. A resistance jump of about
two orders of magnitude was observed. These considerations
show that it is therefore important to search for alternative
methods for thin film synthesis. In this paper we report on the
feasibility of NdNiO3coatings deposited on silicon substrates
by pulsed direct current (dc) reactive magnetron co-sputtering
followed by a simple annealing treatment at low temperatures
in ambient air. The use of dc sputtering, starting with metallic
targets, instead of RF sputtering with one ceramic target, allows
a better oxidation of the film during the synthesis.
The films were synthesized in a 40 L sputtering chamber
pumped down via an oil diffusion suction system allowing
a base vacuum of about 10−4Pa. The Nd–Ni–O films
(500 nm thick) were sputter deposited on silicon (1 0 0) by
co-sputtering of Nd and Ni metallic targets in the presence
of a reactive argon–oxygen gas mixture. The substrates were
positioned on a rotating substrate holder at 70 mm away from
the targets. No external heating was provided during the
film deposition and the temperature deposition was close to
70 ◦C. The Ni and Nd targets, 50 mm in diameter and 3 mm
in thickness, were powered by Advanced Energy generators
dc supply and Pinnacle+ (pulsed dc supply), respectively. The
main deposition parameters were determined by considering
the behaviour of Nd and Ni targets sputtered together with
a discharge current of 1 A and 0.22 A, respectively. The
current dissipated on the Nd target was fixed while the current
dissipated on the Ni target was tuned to match the atomic ratio
Nd/Ni of 1 in the deposited films.
This value was measured via a Philips XL30.S field
effect gun scanning electron microscope (SEM), also equipped
with an energy dispersive spectroscope (EDS). The EDS
spectra were collected for 100 s under an acceleration voltage
of 15 kV. The chemical composition was confirmed by
Rutherford backscattering spectrometry (RBS). Measurements
were carried out in the Laboratoire Pierre S¨
ue (LPS) at CEA-
Saclay in France. The experimental conditions at the LPS were
a proton beam of 3.2 MeV with a beam spot of 5 µm. RBS
measurements were done by an annular PIPS detector. The
reactive mode was achieved with an Ar flow rate of 21 sccm
and an O2flow rate of 7 sccm at a total pressure (PT) of 0.5Pa.
The argon flow and the oxygen flow were controlled using
MKS flowmeters and PTwas measured using an MKS Baratron
gauge. To optimize the deposition conditions for operating in
the reactive sputtering mode, the PTevolution was monitored
while varying the oxygen flow rate (q) (figure 1).
Without discharge, PTin the sputtering chamber follows
the dashed line when the oxygen flow rate is varied. In the
presence of a discharge, a hysteresis on PTis evidenced on
Figure 1. Evolution of the total pressure as a function of the oxygen
flow rate. The current applied to the Nd and the Ni targets were 1 A
and 0.22 A, respectively. The dashed line represents the evolution of
PTversus qwithout discharge.
the increasing and the decreasing qcurves, indicating the
occurrence of an instability of the sputtering conditions for
the Nd + Ni system. This hysteresis results from the coupling
of three major process phenomena: target poisoning, gettering
efficiency of the receiving surfaces and suction capacity of the
reactor. qbeing used as the only control parameter, we observe
a huge PTincrease from A to B for increasing values of qand
PTavalanches from C to D for decreasing values of q. Figure 1
indicates that a minimum of 7 sccm of oxygen is required to
reach an equilibrated synthesis condition in the chamber when
qis increased. In practice, the discharge currents are applied
to the targets at qinitial =10 sccm for 10 min and the value of
qis decreased to qwork =7 sccm to optimize the deposition
rate. The films were deposited for 1 h leading to a thickness of
approximately 500 nm as measured by the step method with
a Talysurf profilometer allowing an accuracy of about 20–
30 nm. The structural features of the coatings were obtained
by glancing angle x-ray diffraction (λ=0.178 897 nm, Co Kα
radiation) using an INEL diffractometer with an incident angle
of 2◦.
The as-deposited Nd–Ni–O films were amorphous.
To induce film crystallization, annealing treatment was
performed. The lowest crystallization temperature, in ambient
air, was found to be 640 ◦C. Figure 2shows the x-ray
diffractogram of a film annealed at 640 ◦C for 3 days
that is characteristic of an orthorhombic NdNiO3perovskite
structure.
Note that polycrystalline randomly oriented thin films
are obtained. Although NdNiO3is known to crystallize in
the Pbnm space group, we assumed a pseudocubic structure
for the films, and the lattice constant is determined to be
0.380 32±0.000 03 nm. This value is slightly lower than those
usually observed in bulk samples (a=0.3810 nm) [16].
The coating is dense either as-deposited or after annealing.
A SEM micrograph performed on a brittle-fracture cross
section after annealing is presented in figure 3. The
micrographs are obtained via the SEM with the secondary
electron detector under an acceleration voltage of 5 kV and
the weight distance between the detector and the sample is
around 5 mm.
2

J. Phys. D: Appl. Phys. 42 (2009) 182006 Fast Track Communication
Figure 2. X-ray diffractogram for the NdNiO3thin film deposited
on a Si (1 00) substrate after air annealing. Indexation refers to
orthorhombic cell Pbnm with average a=√2 ap, b=√2 ap, and
c=2×ap. Average pseudocubic parameter is 0.380 32nm.
Figure 3. Brittle-fracture cross section SEM micrographs of
NdNiO3coatings deposited on silicon substrates.
Figure 4. Evolution of the electrical resistance versus the
temperature for an annealed film. TMI is defined when dR/dT
changes its sign (see the inset).
The evolution of electrical resistance (R) of the annealed
film was characterized on heating between −170 and 25 ◦C
using the four point probe method with a silver paste
contact. Figure 4clearly exhibits a phase transition from
a high temperature metallic state to a low temperature
semiconducting one.
Figure 5. Infrared transmittance spectra at different temperatures
for an annealed NdNiO3film. (∗) represents the absorption bands of
SiO2and NiO6octahedra. Inset shows the average intensity, at each
temperature, in the atmospheric transparency (band III).
The determination of the TMI was achieved when the
derivative dR/dTchanges sign (see the inset in figure 4).
A value of −68 ◦C has been obtained using this method.
The presence of this transition is one of the most valuable
facts of the process. Indeed, conditions for sputtered oxide
film growth do not necessarily need to be consistent with
the thermodynamic phase stability of the compound. Thus,
the reactive sputtering process can stabilize some metastable
phases, leading here to the oxidation of nickel through a
soft process. The electrical behaviour is similar to the
measurements obtained by Laffez et al [17] on polycrystalline
NdNiO3thin film annealed at 200 bar of oxygen, especially
the transition of 1.4 orders of magnitude. However, we
do not assume that we entirely reached the oxidation state
Ni3+. Indeed, considering the work of Nikulin et al [18]on
NdNiO3−δ, the shape of figure 4fits with an oxygen deficient
rare-earth nickelate. In their work, they show that a phase
transition is still observed in NdNiO3−δas long as δ<0.2.
In the semiconductor state the lack of oxygen was found to
induce donor or acceptor sites in the band gap decreasing the
resistance. In the metallic state the electrons are scattered
by the lack of oxygen and thus increase the resistance. As a
consequence, the material exhibits an unsharpened transition.
From the results of Nikulin et al [18], the amount of oxygen
in our nickelate films may be close to NdNiO2.9.
The electrical measurements versus temperature are
complemented by infrared (IR) transmission measurements
in the same temperature range using a Perkin Elmer FTIR
spectrometer with a DTGS detector in the 400–7000 cm−1
wavenumber range (i.e. 1.42–25 µm). The thin film is
positioned in a variable temperature cell (Specac P/N
21500), which permits transmission mode measurement. The
background is the IR light reaching the detector with the empty
cell placed in the path of the IR light beam. The general shape
of the different spectra (figure 5) is fully consistent with those
previously reported for textured NdNiO3and Eu0.7Nd0.3NiO3
thin films [8,19] annealed at high pressures.
3

J. Phys. D: Appl. Phys. 42 (2009) 182006 Fast Track Communication
At −170 ◦C, the thin film shows an IR transmittance
characteristic of a semiconductor state. The absorption bands
at 9.1, 12.3, 13.4 and 16.4 µm are attributed to a layer of
SiO2[20] that grows during the annealing process, at the
interface of the film/substrate and on the uncovered surfaces of
the silicon substrate. For higher wavelengths, the film presents
two absorption bands at 17.5 µm and 22.5 µm dominated by
the antisymmetric T1u stretching and deformation modes of the
NiO6octahedra, respectively [21]. In agreement with previous
studies performed on thin films the temperature dependence
of transmittance appears only above 5 µm. Inspection of the
semiconductor state reveals that the transmittance gradually
increases in the 7–15 µm range owing to the gap absorption
edge. These wavelengths are higher than those measured
in [19]. We attribute this effect to microstructure driven
perturbations in the 5–10 µm range, as a NdNiO3−δthin film
exhibits a smaller resistance jump leading to a deterioration
of the thermochromic effect. Nevertheless, upon heating, the
electrical behaviour is associated with a clear change in the
transmittance. The screening effect which corresponds to
the increase in density of charge carriers in the low energy
spectrum is well evidenced. Indeed, when the temperature
increases, the maximum at 15 µm in the semiconductor state
is progressively masked, and the material becomes opaque at
TMI. The decrease in transmittance ranges from about 19%
at −170 ◦Cto7%at−68 ◦C. For potential applications it
is useful to report the transmittance versus temperature in
the different transparency bands of atmosphere. In order
to quantify this property, we have selected the atmospheric
transparency band III (8–12 µm) and computed the average
intensity (1/λ) λ2
λ1(I (λ)/I0)dλfor each temperature. The
inset of figure 5shows the evolution of this average intensity
with temperature.
In conclusion, NdNiO3−δthin films were reactively
deposited by co-sputtering from Nd and Ni metallic targets
on Si (1 0 0) substrates. We have adjusted the oxygen flow
rate to grow the thin films in the reactive sputtering mode.
The deposition process was followed by an annealing step at
640 ◦C for 3 days in ambient air. The characterizations, such
as x-ray diffraction, electrical and optical transmittance versus
temperature, demonstrate that the as-deposited Nd–Ni–O thin
films are amorphous while the annealed ones crystallize in
the perovskite structure. The resistance versus temperature
presents characteristics similar to polycrystalline NdNiO3
thin film post-annealed at high oxygen pressures. Thus,
while avoiding a drastic annealing condition or epitaxial
stabilization, a thermochromic thin film is synthesized. These
results represent a great improvement in the simplification
of this thermochromic material and would be expected to
stimulate extensive studies on RENiO3materials. We suggest
that this work opens a new field of applications for metal–
semiconductive nickelate compositions. We will now focus the
future studies on the optimization of the transition temperature
and the sharpness of the transition.
Acknowledgment
The authors thank A Gordon for his help with the English
language.
References
[1] Frand G, Bohnke O, Lacorre P, Fourquet J L, Carre A,
Eid B, Th´
eobald J G and Gire A 1995 J. Solid State Chem.
120 157
[2] Napierala C, Edely M, Laffez P and Sauques L 2009 Opt.
Mater. 31 1498
[3] Lacorre P, Torrance J B, Pannetier J, Nazzal A I, Wang P W
and Huang T C 1991 J. Solid State Chem. 91 225
[4] Demazeau G, Marbeuf A, Pilchard M and Hagenmuller P
1971 J. Solid State Chem. 3582
[5] Garcia J, Gonzalez A, Sanchis M J, Marcos M D, Martinez E,
Sapi˜
na F, Beltran D and Beltran A 1993 Solid State Ion.
63–65 52
[6] Vassiliou J K, Hornbostel M, Ziebarth R and Disalvo J F 1989
J. Solid State Chem. 81 208
[7] Tiwari A and Rajeev K P 1999 Solid State Commun. 109 119
[8] DeNatale J F and Kobrin P H 1995 J. Mater. Res. 10 2992
[9] Laffez P, Zaghrioui M, Monot I, Brousse T and Lacorre P 1999
Thin Solid Films 354 50
[10] Laffez P, Retoux R, Boullay P, Zaghrioui M, Lacorre P,
Van Tendeloo G 2000 Eur. Phys. J. Appl. Phys. 12 55
[11] Laffez P, Ruello P, Edely M 2008 Leading Edge in Materials
Science Research vol 277 ed P W Lamont ISBN
1-60021-798-2 (Hauppauge, NY: Nova Science) p 310
[12] Catalan G, Bowman R M and Gregg J M 2000 J. Appl. Phys.
87 606
[13] Catalan G, Bowman R M and Gregg J M 2000 Phys. Rev. B
62 7892
[14] Novojilov M A, Gorbenko O Yu, Graboy I E, Kaul A R,
Zandbergen H W, Babushkina N A and Belova L M 2000
Appl. Phys. Lett. 76 2041
[15] Conchon F, Boulle A, Girardot C, Pignard S, Guinebretiere R,
Dooryhee E, Hodeau J L, Weiss F, Kreisel J and Berar J F
2007 J. Phys. D: Appl. Phys. 40 4872
[16] Medarde M L 1997 J. Phys.: Condens. Matter 91679
[17] Laffez P, Zaghrioui M, Retoux R and Lacorre P 2000 J. Magn.
Magn. Mater. 211 111
[18] Nikulin I V, Novojilov M A, Kaul A R, MudretsovaSNand
Kondrashov S V 2004 Mater. Res. Bull. 39 775–91
[19] Capon F, Laffez P, Bardeau J F, Simon P, Lacorre P and
Zaghrioui M 2002 Appl. Phys. Lett. 81 619
[20] Grecea M et al 1999 J. Mol. Struct. 480 607
[21] Baran E J 1990 Catal. Today 8133
4