Materials Chemistry and Physics 113 (2009) 756–763

Contents lists available at ScienceDirect

Materials Chemistry and Physics

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

(Cu, Fe, Co, or Ni)-doped tin dioxide films deposited by spray pyrolysis:
Doping influence on thermal stability of the film structure
G. Korotcenkov a,b,∗ , Sang Do Han a,∗
a
Future Fundamental Technology Research Department, Korea Institute of Energy Research, 71-2 Jang-dong, Yuseong-gu, Daejeon 305-343, Republic of Korea
b
Technical University of Moldova, Bld. Stefan cel Mare 168, Chisinau 2004, Republic of Moldova

a r t i c l e i n f o a b s t r a c t

Article history: The results of doping influence on thermal stability of the SnO2 film morphology are presented in this
Received 9 May 2008 article. The SnO2 films doped by Fe, Cu, Ni, Co (16 at.%) were deposited by spray pyrolysis from 0.2 M
Received in revised form 6 August 2008 SnCl4 –water solution at Tpyr 350–450 ◦ C. The annealing at 850–1030 ◦ C was carried out in the atmosphere
Accepted 11 August 2008
of the air. The change of such parameters as film morphology, the grain size, texture and the intensity
of X-ray diffraction (XRD) peaks have been controlled. For structural analysis of tested films we have
Keywords:
been using X-ray diffraction, Scanning Electron Microscopy (SEM), and Atomic Force Microscopy (AFM)
Oxides
techniques. It was established that the doping does not improve thermal stability of both film morphology
Thin films
Annealing
and the grain size. It was made a conclusion that the increased contents of the fine dispersion phase of
Microstructure tin oxide in the doped metal oxide films, and the coalescence of this phase during thermal treatment are
the main factors, responsible for observed changes in the morphology of the doped SnO2 films.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction including SnO2 , In2 O3 , and TiO2 [3,10–14]. It is necessary to note

that the prevention of the grain growth during any thermal treat-
High thermal stability of structure is indispensable requirement ments is also important for achievement the high sensitivity of gas
to any material designed for practical application, including metal sensors [15].
oxides [1]. It is known that the sintering rate of ceramic materials For the thin films doped by various additives we do not have
is usually controlled by the grain boundary transport kinetics. It such a reliable experimental basis for conducting a similar anal-
was also established that the appearance of additional intergrain ysis. As it is known, the structure and the composition of doped
boundaries between the adjacent grains of metal oxides inhibits metal oxides, formed under the conditions of the powder synthe-
the cation inter-diffusion and the crystallite growth. It means that sis and the film deposition, differ from each other. It means that
the incorporation of the second phase through modification of the peculiarity of the change of both the morphology and the crys-
grain boundaries’ properties could provide controlled influence on tallite size in doped metal oxide films during thermal treatment
the mechanism and kinetics of the sintering and the final proper- could be essentially different from the regularities established for
ties of the ceramic materials [2]. So, there is an assumption that the metal oxide ceramics. So, the enlarging research in the field of
the doping, accompanied by the appearance of the second phase, doped metal oxide thin film’s study is required.
can promote the improvement of thermal stability of the metal The results presented in this article are a continuation of
oxide structure. Numerous experiments carried out with metal research, started in Refs. [16,17]. In indicated papers we have con-
oxide powders, synthesized by using standard methods, such as sidered the influence of doping by Cu, Fe, Co and Ni (0–16 at.%)
sol–gel method, have confirmed this assumption [3–9]. Metal oxide on the cathodoluminescence and the structure of the SnO2 films,
powders doped by Cr, Nb, Ce, Er, Ni and Al had considerably bet- deposited by spray pyrolysis method. We have established that the
ter stability of the grain size during thermal treatments than the SnO2 doping was accompanied by the sharp quenching of cathodo-
undoped powders. Such effect was observed for many metal oxides, luminescence in comparison with undoped samples [17]. We have
also found out that the increase in the contents of the fine dispersed
phase in as-deposited films is one of the most important conse-
quences of the SnO2 doping by mentioned above impurities [16].
∗ Corresponding authors at: Future Fundamental Technology Research Depart-
In present article we are trying to answer the questions regard-
ment, Korea Institute of Energy Research, 71-2 Jang-dong, Yuseong-gu, Daejeon
ing our ability to control the morphology of doped SnO2 films
305-343, Republic of Korea. Tel.: +82 42 860 3509; fax: +82 42 860 3307.
E-mail addresses: [email protected] (G. Korotcenkov), [email protected] during thermal treatments, and regarding the importance of the
(S.D. Han). appearance of mentioned above fine dispersed phase for thermal

0254-0584/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.matchemphys.2008.08.031
 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763 757

Fig. 1. SEM images of undoped SnO2 films and films doped by Co, Ni, Fe or Cu before and after annealing at T = 850 ◦ C. Films were deposited at Tpyr = 450 ◦ C and had thickness
∼120 nm.
758 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763

stability of those films’ structural parameters. The wide use of the films. Even more, at used concentrations (0–16 at.%) those additives
SnO2 doping by Cu, Fe, Co and Ni for optimization of the gas sensors do not control that changes. Noticeable distinction in the morphol-
parameters is an additional motivation of present research [1,13,18]. ogy of as-deposited films and similarity of their structure after
As it is known, working temperatures of gas sensors could be in the annealing allows concluding that the doping influence on the film
range of 100–600 ◦ C; and the determination of consequences of morphology at the stage of deposition is much stronger than dur-
doping for the structure stability of gas sensing materials gained a ing the process of the following annealing. It means that the grain
special significance for the improvement of working characteristics shape after annealing is being determined first of all by the tem-
of those devices. perature of thermal treatment and the grain size, but not by the
doping element.
2. Experimental details The comparison of SEM images presented in Figs. 1 and 2 shows
that the shape of the grains, formed as a result of annealing at
During present research we have been using the SnO2 films similar to the ones
described in Refs. [16,17]. Films, which thickness varied from 50–70 nm to 400 nm, Tan ∼ 850 ◦ C does not depend on Tpyr of annealed SnO2 films. After
were deposited by spray pyrolysis from 0.2 M SnCl4 –water solution on oxidized annealing at 850 ◦ C the SnO2 films deposited at Tpyr = 350 ◦ C and
Si substrate at Tpyr = 350 ◦ C and 450 ◦ C [19,20]. Such wide range of the thicknesses 450 ◦ C had very similar SEM images. Further thermal treatment at
was conditioned by the fact that SnO2 films find an application in devices, where
Tan = 950 ◦ C did not influence the film morphology and the grain
requirements to the thickness are being varied in very wide range.
The annealing was carried out in a flow type reactor in the atmosphere of air. The
size, confirming our assumption that the main changes in SnO2
annealing temperature was varied from 850 ◦ C to 1030 ◦ C. The choice of indicated film morphology were connected with coalescence of the fine dis-
temperatures was necessary for obtaining effect, applicable for analysis. As it was persion phase and the small grains, which has already taken place at
found in Refs. [21,22], noticeable changes in the morphology and the grain size of
the SnO2 films deposited by spray pyrolysis take place only after thermal treatments
at Tan higher 600–700 ◦ C. The duration of annealing was usually equaled to 1 h.
Fe, Cu, Co, and Ni have been used as doping additives, embedded into sprayed
solution in the form of chlorides. The concentration of doping elements in sprayed
solution was equaled 16 at.%. At lower concentration of additives (1–8 at.%) the
doping influence was not so clear shown as at the concentration equaled 16 at.%.
Therefore it was decided that consideration of the behavior of the film structure
with maximum concentration of doping additives would be enough for establishing
a general peculiarities of the influence of the dopants’ nature on the transformation
of the films morphology during thermal treatment.
X-ray diffraction (XRD), Scanning Electron Microscopy (SEM), and Atomic Force
Microscopy (AFM) have been used for structural characterization of analyzed films.
XRD measurements were conducted by a Rigaku Rotaflex X-ray diffractometer with a
rotating anode source, working with the K␣ of the Cu. For structural characterization
we have been using the /2 mode of measurements. The size of the crystallites,
forming the film, was evaluated by Scherrer’s formula.
For SEM measurements we have been using scanning electronic microscopes
Philips XL30, and Stereoscan JS360 Cambridge Instruments with structural resolu-
tion ∼0.3–0.4 nm. AFM images were obtained by using MultiMode Scanning Probe
Microscope with Nanoscope IIIa Controller of Digital Instruments. These measure-
ments have been carried out in contact scanning mode using AS-0.5 scanner (scan
rate 2 Hz; resolution 0.4 nm).

3. Experimental results and discussions

3.1. Annealing influence on morphology of doped SnO2 films

In Fig. 1 it is shown the influence of thermal treatment on the

morphology of the SnO2 films doped by various additives. As it is
seen from SEM images, the annealing sufficiently influences the
morphology of deposited films. At that all films, regardless of dop-
ing additives are exposed to transformation. If before annealing at
SEM images for the majority of the SnO2 films we have been observ-
ing the diamond like formations; after thermal treatment they were
transformed into formations without clear cutting. The variety of
aggregate specimens’ configurations in as-deposited SnO2 films
we connect with multiple twinning of the SnO2 crystallites. As it
is known, SnO2 is inclined to twinning during the growth of the
films [23,24]; and, as it was shown in Ref. [16], the doping had an
influence on this process.
We do not know the reasons of such strong change of the film
morphology and the shape of the grains (crystallites). However one
can suppose that existing changes in the crystallite shape is a con-
sequence of the grain size’s change connected with a coalescence
of grains. We believe that after coalescence of the grains, forming
agglomerates, the new grains acquire a shape, peculiar to their big-
ger size. As it is seen in the Fig. 1, after annealing the SnO2 films’
morphology obtains common type for all studied films, regardless
of initial morphology and doping additive. Thus, observed changes Fig. 2. SEM image of undoped SnO2 film (Tpyr = 350 ◦ C; d ∼ 250 nm) before and after
testify that doping does not inhibit structural changes in the SnO2 annealing at Tan = 850 ◦ C and Tan = 1030 ◦ C.
 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763 759

lower temperature. The following considerable change of the SnO2

films’ morphology took place only at Tan > 1000 ◦ C, when the notice-
able increase of the crystallites size was observed. The typical image
of the SnO2 films after annealing at T > 1000 ◦ C is shown in Fig. 2.
The same processes take place for doped films. We have to note,
that observed regularities are very much similar to those, we have
been observing while making thermal stability study of undoped
In2 O3 films’ structure [21].

3.2. Transformation of XRD patterns of doped films during

thermal annealing

Typical XRD patterns of studied films are presented in Fig. 3. As

an example there are shown XRD patterns of undoped SnO2 and
doped SnO2 :Cu films, for which the strongest influence of the dop-
ing on the film morphology is observed (see Fig. 1). As it follows
from the analysis of XRD patterns, the X-ray diffraction peak’s posi-
tion corresponds to cassiterite structure, characteristic for SnO2 . At
that the doping in used range of dopants concentrations (0–16 at.%)
does not exert an essential influence on the XRD patterns’ shape.
XRD patterns of doped SnO2 are only slightly different from the
ones of undoped SnO2 . As it is seen from the results obtained for
films deposited at different temperatures and presented in Fig. 3,
the shape of XRD patterns depends more on the temperature of
deposition then on the doping. It means that in spite of existing
changes in the morphology of doped SnO2 films (see Fig. 1), the
films retain crystallographic structure, peculiar to undoped films.
The same result was also described in other papers [12,13]. For
example, in Ref. [12] XRD measurements of Fe- or Ni-doped SnO2 ,
synthesized by the polymeric precursor method, have shown the
presence of only the SnO2 tetragonal phase, even at 30 mol.% of both
additives.
Present research has shown that the thermal treatment in the
range of 800–1000 ◦ C has also weak influence on the shape of
XRD patterns of doped SnO2 . XRD patterns after annealing had the
shape, similar to the one of XRD patterns of undoped SnO2 films
[22,25]. We just observe some change in the correlation of inten-
sities of individual XRD peaks. In particular, for the doped SnO2
films deposited at Tpyr = 450 ◦ C, we have observed the increase of
the ratio of the main XRD peaks intensities I(1 1 0)/I(2 0 0) in 1.5–1.7
folds. However as it was established, these changes were almost
similar to the ones observed for undoped SnO2 films. It means that
observed changes in the shape of XRD patterns are not connected
with the doping process.
The absence of any additional peaks, related to the second phase
at XRD patterns of doped SnO2 , in spite of the fact that, according
to EDAX measurements, this phase exists [16], allows confirming
the conclusion made in Ref. [16], where it was established that the
concentration of doping element in grown SnO2 films was close
to concentration of this element in sprayed solution. In Ref. [16] it Fig. 3. XRD patterns of undoped and Cu (16 at.%) doped SnO2 films: (a) films
was concluded that in doped SnO2 films very low concentration of deposited at Tpyr = 450 ◦ C, d ∼ 70 nm; (b) films deposited at Tpyr = 350 ◦ C, d ∼ 250 nm.

dopants is found in the state of bulk solid solution; while the most
part of the dopants is segregated on the surface of the SnO2 crys- of SnO2 . It is necessary to note that the small limiting solubility of
tallites in the form of small grains with size smaller than 2–4 nm. various metal oxides is typical for SnO2 [12,14].
It is known that X-ray diffraction on the crystallographic planes of
the grains with such small size does not give any noticeable peaks. 3.3. Doping influence on grain size stability
The absence of significant shifts in the position of X-ray
diffraction peaks in comparison with undoped SnO2 is additional Regarding the influence of the nature of doping additives on the
confirmation of the made-above conclusion. Even after annealing stability of the SnO2 grain size, obtained XRD data allows to make
at 1000 ◦ C the shift in XRD peak position in doped films in compar- the following conclusions.
ison with undoped ones was in the limits of experimental error. It Additives actually affect the thermal stability of the SnO2 grain
means that the doping is not accompanied by the change of lattice size. However it is necessary to note, that optimizing influence, i.e.
parameter of SnO2 . Such situation at used levels of doping is pos- the decrease of the grain size change during thermal treatment,
sible only in case when the solubility of doping elements is small, in contrast to the results obtained for the SnO2 powders, is not
i.e. only small part of doping additives incorporates in the lattice being observed for all studied additives. For example, the addition
760 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763

Table 1
Doping influence on the intensity of X-ray diffraction at (1 1 0) plane of SnO2 grains

Tpyr (◦ C) d (nm) Doping element XRD intensity (1 1 0)

As-deposited Tan = 850 ◦ C

350 250 ± 20 – 80 244

Cu 52 487

450 120 ± 10 – 356 586

Cu 158 804
Fe 189 510
Co 308 736
Ni 365 538

450 400 ± 20 – 1660 2500

Cu 1620 3000
Fe 1670 2520
Co 1400 2800

a rule, the bigger the drop of XRD peaks’ intensity in as-deposited

doped films is, the greater is the growth of the XRD peaks’ intensity,
Fig. 4. Influence of annealing on average grain size in undoped and Cu-doped and the bigger is the growth of the grain size after annealing (see
SnO2 films estimated by XRD method. Studied films were deposited at Tpyr = 350 ◦ C Fig. 4 and Table 1). It means that the changes, taking place during
(d ∼ 250 nm). the annealing, are mainly a reflection of the changes, existing during
the film deposition.
of nickel practically does not influence the stability of the tin diox- In Ref. [16] we have concluded that the doping of SnO2 during
ide grain size during thermal treatment at 850 ◦ C. The changes of spray pyrolysis deposition is accompanied by the increase of con-
average grain sizes estimated on the base of XRD data in SnO2 and tents of the fine dispersed amorphous-like SnO2 phase. Under fine
SnO2 :Ni films with the same thickness were similar. While in the dispersed phase we meant the SnO2 grains with size smaller than
presence of Fe, Co and especially cooper additives, thermal stability 2–3 nm. As it is known, X-ray diffraction at crystallites of this size
of structural parameters of doped SnO2 films is being considerably (t < 2–3 nm) does not show well-defined peaks at the XRD patterns.
deteriorated. We believe that discussed in this paper effect of strong increase
At that indicated influence was in full accordance with regu- of X-ray diffraction peaks’ intensity for doped SnO2 films after their
larities established for undoped SnO2 films. The higher is pyrolysis annealing is a consequence of the presence of this fine dispersed
temperature and the less is film thickness, the smaller is the change SnO2 phase. Due to coalescence with already existing bigger crys-
of the crystallites size during the annealing [21]. For example, tallites during thermal treatment it promotes both the increase of
if in doped SnO2 :Cu films with thickness ∼120 nm deposited at the grain size and the improvement of the film crystallinity.
Tpyr = 450 ◦ C, after thermal treatment at Tan = 850 ◦ C the average However the second oxide phase (Fe2 O3 , NiO, Cu2 O or CoO), in
grain size changed on average for 3–5 nm; while for the films spite of the fact that it is segregated on the surface of the grains
deposited at Tpyr ∼ 350 ◦ C with the same thickness the grain size of basic oxide [12,14] is not able to block this process. Apparently
change exceeds 15 nm. For thicker films (d ∼ 250 nm; Tpyr = 350 ◦ C) the fine dispersed phase of the basic oxide (SnO2 ), filling inter-
those changes are even bigger and exceed 20 nm (see Fig. 4). For crystallite space, even at the presence of the second oxide phase
information, in as-deposited undoped SnO2 films with thickness is able to form the channels for intensive mass transfer between
∼120 nm deposited at Tpyr = 450 ◦ C the average grain size is equaled crystallites (from the smaller crystallite to the bigger one), neces-
∼27 nm, and in as-deposited SnO2 :Cu films the average grain size sary for their growth (see Fig. 5). Either theoretical calculations, and
is equaled ∼18 nm. Films with thickness 250 nm (Tpyr = 350 ◦ C) had experimental results have shown that already at the crystallites size
initial average crystallite size equaled 13–14 nm (see Fig. 4). smaller than 5 nm, the temperature of melting could decrease in
more than 2–3 times [21,26,27]. As a result, already at sufficiently

3.4. Mechanism of the grain size instability in doped SnO2 films

Analyzing obtained results we have concluded that the doping

exerts the most evident influence on the intensity of XRD patterns.
It was established that after thermal treatment the intensity of char-
acteristic X-ray diffraction peaks has been sharply increasing (see
Table 1). At that the maximum increase of XRD peaks’ intensity
appears mainly for the films, doped by Cu and Fe, for thin films, and
for the films, deposited at low pyrolysis temperatures. As it is seen
from data presented in Table 1, during the annealing at T = 850 ◦ C
the intensity of X-ray diffraction peaks of doped SnO2 films, espe-
cially of SnO2 :Cu, may increase up to 8 times. For undoped films for
the same conditions the X-ray diffraction peaks’ intensity does not
increase in more than 3 times. As it is known, the increase of XRD
peak’s intensity usually is being considered as an evidence of the
improvement of the tested films’ crystallinity.
At that the changes of the intensity after the annealing were in
Fig. 5. Diagram illustrating the influence of the fine dispersed phase on the forming
direct correlation with both the doping influence on XRD patterns of channels for mass transfer between the grains of SnO2 in the presence of the
of as-deposited films, and with observed change of the grain size. As second oxide phase between them.
 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763 761

low temperatures the necks with essentially big area (diameter) ence of the fine dispersed phase. After annealing at Tan = 850 ◦ C
could be formed between crystallites. The presence of the necks all AFM images even for SnO2 films doped by cooper (see Fig. 6b)
between grains can be considered as factor, promoting the coa- had clearly defined contours of grains, confirming the absence in
lescence of closely located grains. This fact could be an important the SnO2 matrix of the grains with sizes, smaller than the limit of
reason of degradation of the sensors parameters after high tem- resolution of used measuring instruments.
perature annealing. In this context, a considerable improvement of It is important to note that the morphology of the SnO2 films
the sensors parameters after mechanical milling of sintered metal after thermal treatments replicates the macrostructure of the films
oxide powders could be considered as a result of those necks’ break- formed by agglomerates (see Fig. 6). Such behavior one can con-
ing [28]. sider as the evidence of the fact that the transformation of the
The mechanism of the appearance of the fine dispersed phase we film morphology first of all takes place through the coalescence
have discussed earlier in Ref. [16]. According to suggested model, of crystallites forming the agglomerates.
the growth of contents of the fine dispersed phase in doped films is Comparing the influence of annealing on the XRD patterns of
conditioned by the appearance of additional nucleation centers for both thick and thin SnO2 films, deposited at the same tempera-
the SnO2 growth, formed by the second oxide phase. As we indi- ture, one can see that the change of the X-ray diffraction peaks’
cated earlier the second oxide phase, appearing during the process intensity after thermal treatments is higher for the thin films (see
of SnO2 growth as a rule is being segregated on the surface of basic Table 1). For the SnO2 films with thickness of 400 nm (Tpyr = 450 ◦ C),
oxide grains [4,12,29], which stimulates the appearance on the sur- the change of the X-ray diffraction peaks’ intensity after annealing
face of big crystallites of the layer of the fine dispersed phase, filling at T = 850 ◦ C was equaled 1.5–2 times, while for the films with thick-
inter-crystallite space. It is necessary to note that we are not the ness of 100 nm this change could exceed 3–5 times. Basing on our
only ones, using such approach in analyzing structural properties assumptions this effect can be easily explained, since in the thin
of the SnO2 films. In Ref. [30] it was also mentioned the form- films the contents of the fine dispersion phase is bigger. This is
ing of the fine dispersed amorphous-like phase during the growth understandable, because with the thickness’ growth the grain size
of the SnO2 films. The authors of mentioned work, as well as in is being increased as well [20,25]. Besides that, as we have shown
our case, have been observing both the increase of contents of the earlier in Ref. [15], the SnO2 films at the interface with substrate
fine dispersed phase at the lowering of deposition temperature and always had transient layer with thickness of ∼3–10 nm, consist-
the sharp reduction of this phase’s concentration at the following ing from randomly oriented fine dispersed crystallites. Therefore,
annealing at Tan > 500 ◦ C. the contribution of this transition layer crystallization and struc-
On the base of the assumption that the main changes in the ture ordering in XRD patterns of the thin films should be obviously
grain size of doped SnO2 films during the annealing at Tan ∼ 850 ◦ C higher, than in the thick films.
are connected with the presence of the fine dispersed phase, and The comparison of data presented in Table 1 also indicates that
basing on the data of the annealing influence on the intensity of the films deposited at low temperatures (Tpyr ∼ 350 ◦ C) subjected
X-ray diffraction peaks, we receive the confirmation of the con- to much more structural changes than the films, grown at higher
clusion made in Ref. [16] that the maximum contents of the fine temperatures (Tpyr > 400–450 ◦ C). This result is quite appropriate,
dispersion phase in the SnO2 films during spray pyrolysis deposi- because the coalescence of the fine dispersed grains formed dur-
tion are being formed after doping by Cu. With regard to another ing the process of the growth with big crystallites could take place
doped SnO2 films, we can concluded that the contents of the fine already at the stage of the film growth. It means that devices, man-
dispersed phase have been decreasing in the line SnO2 :Cu, SnO2 :Fe, ufactured of the base of films, grown at high temperatures would
SnO2 :Co, SnO2 :Ni, SnO2 . Given line reflects the degree in the change be more stable during exploitation in comparison with the ones,
of the XRD peaks intensity as a result of thermal treatment. fabricated on the base of low temperature films.
The same conclusion one can make analyzing AFM images of From our point of view the increase of contents of the fine dis-
doped SnO2 films (see Fig. 6). As we have shown in Ref. [16], the persed phase in doped SnO2 films has great importance for doped
sharpness (image detail) of the grain contours, observed at AFM films’ application in real devices, since the appearance of the fine
images for doped SnO2 films is being worsening due to the pres- dispersed phase can change such metal oxides’ parameters as real

Fig. 6. AFM images of SnO2 :Cu films (a) before and (b) after thermal treatments at 850 ◦ C. SnO2 films were deposited at Tpyr = 350 ◦ C and had thickness equaled d ∼ 250 nm.
762 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763

area of intergrain contacts, the active surface area, the surface activ- size after thermal treatment. At that the contents of this phase
ity, gas penetrability of gas sensing matrix, and so on. For example, are being increased in the line SnO2 , SnO2 :Ni, SnO2 :Fe, SnO2 :Co,
the presence of the fine dispersed phase, inclined to coalescence and SnO2 :Cu. It was assumed that the fine dispersed phase, fill-
already at low temperatures could be a reason, responsible for ing inter-crystalline space, besides coalescence with the big grain
observed temporal drift of the sensors parameters [1,15,31–33]. As creates conditions for mass transition from one grain to another
it is known, the instability of the gas sensing material structure is one, necessary for coalescence of bigger crystallites, forming film.
the main reason of instability of the gas sensors parameters [15,32]. After annealing the morphology of doped and undoped SnO2 films
Therefore, indicated peculiarity of doped films should be taken into becomes similar. The grains lose clearly defined faceting.
account while analyzing the doping influence on electro-physical Such unexpected influence of additives on the SnO2 films struc-
properties of deposited films. In particularly, for avoidance of pos- ture, i.e. the decrease of the grain size in as-deposited SnO2 films in
sible temporal instability it is desirable to age as-deposited doped comparison with undoped films and the increase of the grain size
films through their additional thermal treatment. after annealing, forces researchers to be careful while elaborating
It is necessary to note that this statement receives good exper- the technologies of gas sensing layers’ forming on the base of doped
imental confirmation. Numerous researches have shown that the and multi-component materials. For example, the presence of the
decrease of the grain size is inevitably accompanied by the degrada- fine dispersed phase could be a reason of instability of the metal
tion of thermal stability of the film structure even for metal oxides, oxide films’ structure and generate temporal drift of the parameters
having maximum thermodynamic stability among well-known of gas sensors, manufactured on their base.
materials [21,34,35]. It was shown that the fine dispersed phase Therefore, while choosing additives for either the gas sensor
with the grain size smaller than 1–3 nm was inclined to coalescence optimization or for the film structure stabilization one should take
with bigger crystallites already at low temperatures, which did not into account that after changing of the conditions of synthesis,
exceeded 300–400 ◦ C, promoting their growth [21,26,36,37]. exploitation, or parameters of post-deposition thermal treatments,
Structural disorder and the growth of concentration of the point the consequences of the doping could be substantially different
defects in doped SnO2 films is another important factor, stimulating from the ones, observed during our initial experiments with as-
structure’s instability of doped films during the process of anneal- deposited films.
ing. As we have established earlier in Ref. [17], basing on the analysis
of cathodoluminescence and optical transmission, the doping of Acknowledgements
the SnO2 films is accompanied by structure disordering, appearing
through the growth of the concentration of centers of non-radiating This work was supported by the Supreme Council of Science
recombination and the appearance of the tails of electronic states and Advanced Technology of the Republic of Moldova, as well as by
inside the band gap. At that maximal effect from the doping was the Korean Brain Pool Program. The authors are thankful to Prof.
observed exactly for the films, doped by Cu, for which the maxi- J. Schawnk from University of Michigan, USA, and Dr. V. Brinzari
mum instability of the film’s structural parameters was found. The from Technical University of Moldova for their help in structural
authors of Refs. [13,38] have also indicated an essential growth of characterization of studied samples.
structural defects in doped SnO2 films. For example in Ref. [13] the
increase in the stacking fault density of the SnO2 films has been References
observed already after doping by Fe with concentration 2%. In Ref.
[38] the authors believed that Co creates oxygen vacancies in the [1] G. Sberveglieri (Ed.), Gas Sensors—Principles Operation and Developments,
Kluwer Academic Publishers, The Netherlands, 1992.
SnO2 lattice. According to Ref. [14], the growth of structural defects [2] L.C. Dufour, J. Nowotny (Eds.), Surface and Near-Surface Chemistry of Oxide
concentration stimulates surface diffusion, responsible for mass Materials, Elsevier, Amsterdam, 1988.
transfer during the annealing process. [3] N.L.V. Carreno, A.P. Maciel, E.R. Leite, P.N. Lisboa-Filho, E. Longo, A. Valentino,
L.E.D. Probst, C.O. Paiva-Santos, W.H. Schreiner, Sens. Actuators B: Chem. 86
High degree of densification, which could be attained for doped
(2002) 185.
metal oxides under conventional sintering at high temperatures, [4] D. Szezuko, J. Werner, S. Oswald, G. Behr, K. Wetzing, Appl. Surf. Sci. 179 (2001)
is a direct consequence of those processes [11,14,38–40]. As it is 301.
[5] C.-M. Liu, L.-M. Fang, X.-T. Zu, W.-L. Zhou, Chin. Phys. Soc. (IOP) 16 (1) (2007)
known, the undoped tin dioxide is characterized by a relatively low
95.
densification even at high temperatures of sintering [14]. It is neces- [6] K.N. Yu, Y. Xiong, Y. Liu, C. Xiong, Phys. Rev. B 55 (4) (1997) 2666.
sary to note that recorded effect of densification evidently is not an [7] S.H. Shek, J.K.L. Lai, G.M. Lin, Nanostruct. Mater. 11 (7) (1999) 887.
optimizing effect for gas sensing material and cannot promote the [8] G. Yang, Z. Haibo, Z. Biying, J. Mater. Sci. 35 (2000) 917.
[9] P. Qi, J.-F. Wang, W.-B. Su, H.-C. Chen, G.-Z. Zang, C.-M. Wang, B.-Q. Ming, Mater.
improvement of operating characteristics of SnO2 -based gas sen- Chem. Phys. 92 (2–3) (2005) 578.
sors. As it is known, high porosity and big active surface area are [10] S.D. Choi, D.D. Lee, Sens. Actuators B: Chem. 77 (2001) 335.
the metal oxides’ parameters, important for gas sensor applications [11] M. Ferroni, M.C. Carotta, V. Guidi, G. Martinelli, F. Ronconi, M. Sacerdoti, E.
Traversa, Sens. Actuators B: Chem. 77 (2001) 163.
[1,35,41]. [12] R.H.R. Castro, P. Hidalgo, R. Muccillo, D. Gouve, Appl. Surf Sci. 214 (2003) 172.
[13] S. Rani, S.C. Roy, M.C. Bhatnagar, Sens. Actuators B: Chem. 122 (2007) 204.
[14] R.H.R. Castro, G.J. Pereira, D. Gouve, Appl. Surf. Sci. 253 (2007) 4581.
[15] G. Korotcenkov, Sens. Actuators B: Chem. 107 (1) (2005) 209.
4. Conclusion [16] G. Korotcenkov, V. Brinzari, I. Boris, J. Mater. Sci. 43 (2008) 2761.
[17] G. Korotcenkov, Sang-Do Han, E.V. Kolesnikova, M. Nazarov, Do Young Noh,
It was established, that SnO2 doping by Cu, Co, Ni, and Fe Mater. Chem. Phys., submitted for publication.
[18] J. Tamaki, K. Shimanoe, Y. Yamada, Y. Yamamoto, N. Miura, N. Yamazoe, Sens.
during spray pyrolysis deposition does not improve thermal sta-
Actuators B: Chem. 49 (1998) 121.
bility of the SnO2 film’s structure in studied range of annealing [19] G. Korotcenkov, V. Brinzari, J. Schwank, A. Cerneavschi, Mater. Sci. Eng. C 19
temperatures 600–1000 ◦ C. Moreover, such additives as cooper (1–2) (2001) 73.
[20] G. Korotcenkov, V. Brinzari, M. DiBattista, J. Schwank, A. Vasiliev, Sens. Actuators
stimulate noticeable grain’s growth during the process of high tem-
B: Chem. 77 (1–2) (2001) 244.
perature annealing in comparison with the grain size change in [21] G. Korotcenkov, V. Brinzari, M. Ivanov, A. Cerneavschi, J. Rodriguez, A. Cirera, A.
undoped SnO2 films. It was concluded that observed effect was Cornet, J. Morante, Thin Solid Films 479 (1–2) (2005) 38.
connected with increased contents of the fine dispersed phased in [22] G. Korotcenkov, M. DiBattista, J. Schwank, V. Brinzari, Mater. Sci. Eng. B 77 (2000)
33.
doped SnO2 films. The bigger contents of the fine dispersed phase [23] J. Lu, J. Sundqvist, M. Ottosson, A. Tarre, A. Rosental, J. Aarik, A. Harsta, J. Cryst.
are in as-deposited films, the greater is the change of the grain Growth 260 (2004) 191.
 G. Korotcenkov, S.D. Han / Materials Chemistry and Physics 113 (2009) 756–763 763

[24] X. Pan, J. G. Zheng, in: J. Im, S. Yalisove, B. Adams, Y. Zhu, F.R. Chen (Eds), Polycrys- [32] H. Meixner, U. Lampe, Sens. Actuators B: Chem. 37 (1996) 198.
talline thin films-structure, texture, properties and applications III, Proceedings [33] P.K. Sharma, P.C.H. Chan, Z. Tang, G. Yan, I.-M. Hsing, J.K.O. Sin, Sens. Actuators
of Materials Research Society Symposium (MRS ‘97, Boston, USA, December B: Chem. 81 (1) (2001) 9.
1997), vol. 472 (1997) 87. [34] C.V. Thompson, Annu. Rev. Mater. Sci. 30 (2000) 159.
[25] G. Korotcenkov, A. Cornet, E. Rossinyol, J. Arbiol, V. Brinzari, Y. Blinov, Thin Solid [35] N. Barsan, M. Schweizer-Berberich, W. Gopel, F. Fresenius, Anal. Chem. 365
Films 471 (1–2) (2005) 310. (1999) 287.
[26] A.M. Mazzone, J. Phys.: Condens. Mater. 14 (2002) 12819. [36] C.H. Shek, J.K.L. Lai, G.M. Lin, Y.F. Zheng, W.H. Liu, J. Phys. Chem. Solids 58 (1)
[27] G. Guisbiers, S. Pereira, Nanotechnology 18 (2007) 435710. (1997) 13.
[28] S. Abe, K. U-Sung Choi, N. Shimanoe, Yamazoe, Sens. Actuators B: Chem. 107 [37] C.H. Shek, J.K.L. Lai, G.M. Lin, J. Phys. Chem. Sol. 60 (1999) 189.
(2005) 516. [38] B.-C. Kim, J.-I. Jung, J.-H. Lee, J.-J. Kim, Solid State Ionics 144 (2001) 321.
[29] T. Kobayashi, Y. Kimura, H. Suzuki, T. Sato, T. Tanigaki, Y. Sato, C. Kito, J. Cryst. [39] B. Panchapakesan, D.L. De Voe, M.R. Widmaier, R. Cavicchi, S. Semancik, Nan-
Growth 243 (2002) 143. otechnology 12 (2001) 336.
[30] V.M. Jimenez, J.P. Espinos, A. Caballero, L. Contreras, A. Fernandez, A. Justo, A.R. [40] J.P. Bonnet, N. Dolet, J.M. Heintz, J. Eur. Ceram. Soc. 16 (1996) 1163.
Gonzalez-Elipe, Thin Solid Films 353 (1999) 113. [41] G. Zhang, M. Liu, Sens. Actuators B: Chem. 69 (2000) 144.
[31] W. Gopel, K.D. Scherbaum, Sens. Actuators B: Chem. 26-27 (1995) 1.