Quartz-crystal microbalance study of the growth of Zn(Se,O) thin-films in a chemical bath. A sequential electroless-chemical process

Article abstract of DOI:10.1016/S0013-4686(01)00815-5

The chemical bath deposition of Zn(Se,O) thin films is studied with a quartz crystal microbalance (QCM). The deposition velocity is measured with QCM under different experimental conditions, including substrate properties, bath temperature and bath composition. The plots of the growth velocity vs. time reflect the sequence of processes for the deposition of the film. At the beginning, an induction period takes place, which ends up with the formation of first ZnSe particles. These particles activate an electroless reaction induced by a reducer, hydrazine or hydroxide, which gives rise to deposition of a film with ZnO and ZnSe composition. This step is necessary to attain compact and thicker films with appropriate conditions for solar cell application. The electroless mechanism is decelerated after deposition of some nanometers of film, hence, a second growth mechanism starts to predominate consisting in the chemical reaction of soluble selenide with hydroxide and Zn²⁺ cations on the substrate surface. The layers resulting from this second mechanism have higher proportion of ZnSe than the previously electroless generated layers. At longer times, the predominant growth mechanism is the deposition of ZnSe clusters formed in the solution, which produces less compact and poorly adherent top layers. This sequence of mechanisms gives rise to films with heterogeneous morphology and composition in depth.

Full text of DOI:10.1016/S0013-4686(01)00815-5

Electrochimica Acta 47 (2001) 977–986
www.elsevier.com/locate/electacta
Quartz-crystal microbalance study of the growth of Zn(Se,O)
thin-films in a chemical bath. A sequential electroless-chemical
process
A.M. Chaparro *, M.T. Gutie´rrez, J. Herrero
Departamento de Energ´ıas Reno6ables, CIEMAT, A6da. Complutense 22, 28040 Madrid, Spain
Received 27 June 2001; received in revised form 30 August 2001
Abstract
The chemical bath deposition of Zn(Se,O) thin films is studied with a quartz crystal microbalance (QCM). The deposition
velocity is measured with QCM under different experimental conditions, including substrate properties, bath temperature and bath
composition. The plots of the growth velocity vs. time reflect the sequence of processes for the deposition of the film. At the
beginning, an induction period takes place, which ends up with the formation of first ZnSe particles. These particles activate an
electroless reaction induced by a reducer, hydrazine or hydroxide, which gives rise to deposition of a film with ZnO and ZnSe
composition. This step is necessary to attain compact and thicker films with appropriate conditions for solar cell application. The
electroless mechanism is decelerated after deposition of some nanometers of film, hence, a second growth mechanism starts to
predominate consisting in the chemical reaction of soluble selenide with hydroxide and Zn2+ cations on the substrate surface. The
layers resulting from this second mechanism have higher proportion of ZnSe than the previously electroless generated layers. At
longer times, the predominant growth mechanism is the deposition of ZnSe clusters formed in the solution, which produces less
compact and poorly adherent top layers. This sequence of mechanisms gives rise to films with heterogeneous morphology and
composition in depth. © 2001 Elsevier Science Ltd. All rights reserved.
Keywords: ZnSe; ZnO; Chemical bath deposition; Electroless; Photovoltaics
photon energy absorption, indicating that the CIS/ZnSe
interface formed is less appropriate for solar energy
1. Introduction
conversion than the CIS/CdS. As soon as the interfacial
properties are determined during the process of deposi-
tion of the buffer layer, notwithstanding possible effects
of post-deposition processing, a closer view to the CBD
kinetics may help in understanding the buffer/absorber
interface properties, and to orient efforts towards fu-
ture improvement.
The general mechanisms scheme accepted for the
CBD process consists in two consecutive steps, first the
reaction of dissolved precursors of the halogen and the
metal at the substrate surface (heterogeneous chemical),
and then the deposition by diffusion of precipitates and
clusters formed in solution [1,4,5]. For the case of
CBD–ZnSe films, this scheme may in principle explain
the important presence of oxide in the films [6], and
heterogeneous composition and morphology [7]. Layers
deposited at the beginning of the process are more
The chemical bath deposition (CBD) is a low temper-
ature, soft method for the growth of thin films. It is
very appropriate when a conforming film is needed with
preservation of the properties of the substrate under
few monolayers. For this reason, CBD is applied to the
deposition of the buffer layer onto the semiconductor
absorber of thin-film solar cells [1]. One good candidate
for buffer layer of next generation Cd-free thin film
solar cells is CBD deposited ZnSe. Results on solar cells
with CuInSe2 absorber have given efficiencies of almost
14% [2], and with CuInS2 absorber almost 10% effi-
ciency is obtained [3]. These efficiencies are still below
those of cells using CdS buffer layer, in spite of the
higher band gap of ZnSe which allows for higher
* Corresponding author. Fax: +34-91-3466037.
E-mail address: antonio.mchaparro@ciemat.es (A.M. Chaparro).
0013-4686/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0013-4686(01)00815-5

978
A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
compact and richer in zinc atoms proportion due to
parallel zinc oxide deposition during the heterogeneous
stage. These films are better designated as of Zn(Se,O)
composition, indicating that they are formed by a close
mixture of ZnSe and ZnO nanocrystals in variable
proportion (no evidence for solid solution is found and
no intermediate compound is known). On the other
hand, it is known that the growth of films with high
photovoltaic quality requires the addition of hydrazine,
a reducer and complexing agent, which improves com-
pactness and adherence of the films. In spite of its
important effect, the participation of hydrazine in the
deposition mechanism has not been clarified. It is also
found that the proportion of ZnO in the film is very
dependent on the conductivity of the substrate, increas-
ing with the conductivity of the substrates [8]. It will be
shown here that hydrazine and substrate conductivity
are closely related factors, responsible for a transient
electroless mechanism which results in the deposition of
ZnSe and ZnO. This mechanism is more important at
the beginning of the deposition process, and is deter-
mining for the properties of the films.
The technique used for the study of the growth of
Zn(Se,O) in a chemical bath is the quartz crystal mi-
crobalance (QCM). QCM is an in situ technique which
monitors in real time the thickness of the film, giving
access to the velocity of deposition [9]. QCM technique
has been applied to the study of CBD growth of CdS
[10] and CdSe [11] thin films. The experiments are
carried out under different substrate, temperature and
bath composition conditions in order to characterise
the reaction mechanisms that contribute to the growth
of Zn(Se,O) films.
A quartz crystal microbalance system (Maxtek,
INC.) was used, run by a PC computer and a program
(quickbasic) for real time acquisition of the resonance
frequency and mathematical conversion to film thick-
ness. The corrected expression for loaded crystal was
used for the conversion of frequency to film thickness
[13], using for the acoustic impedance and density of
ZnSe the values v=12.23×1010 kg m−2 s−1 and
z=5.26 g cm−3, respectively. This expression gives for
low loads (DfB0.05f0) a quasi-constant conversion fac-
tor of 30 Hz nm−1. The derivative of the film-thickness
vs. time gives the growth velocity vs. time plots, which
have the meaningful chemical information [11]. The
commercial quartz crystal sensors used have f0=5
MHz fundamental frequency, 3.14 cm2 area, with a
central disk of 1.37 cm2 coated with Au or Pt. Polished
and unpolished crystals were used as purchased, with 5
and 250 nm RMS roughness, respectively. The sensor
head was immersed vertically in the chemical bath. Due
to the sensitiveness of the QCM to small changes of
temperature, viscosity and/or hydrodynamic conditions,
special care was taken before the start of the measure-
ments. After all the reactants were added to the chemi-
cal bath, except selenourea, a settle time of some
minutes was allowed until attaining a constant reso-
nance frequency within 10–20 Hz, which indicated
stabilised conditions inside the bath (viscosity, tempera-
ture, agitation). Then addition of solid selenourea was
carried out and at the same time the count of reaction
time was initiated. Resonance frequency measurements
(film thickness) were taken each 30 s during the first
10–15 min, and at longer intervals afterwards. XPS
measurements for determination of the composition of
films were carried out as described elsewhere [7].
2. Experimental
3. Results
The chemical bath composition included selenourea
(SU) (NH2ꢀCSeꢀNH2) and zinc sulphate (ZnSO4) as
selenium and zinc sources, respectively, aqueous ammo-
nia (NH3) as the complexing agent of Zn2+, sodium
sulphite (Na2SO3) and hydrazine hydrate (N2H4·H2O)
[12]. The base composition in the experiments was 30
mM ZnSO4, 7.5 mM NH2CSeNH2, 1.4 M NH3, 30 mM
Na2SO3, and 1.6 M N2H4. Changes from this composi-
tion are indicated in each case. Reagent grade chemicals
and Millipore-Q water were used for the preparation of
solutions. The pH of the departing chemical bath solu-
tion was 11.5 (25 °C), and was found to diminish at a
rate of 0.3 units per hour. This diminution must be
attributed almost entirely to ammonia evaporation, as
pure ammonia solutions showed the same drift, which
means that the chemical deposition reaction does not
cause an important change of pH. The bath was water
thermostatised, stirred with a magnet and left in open
air during the deposition.
The effect of different bath parameters on QCM
growth curves is studied.
3.1. Substrate
Fig. 1 shows the growth rate of Zn(Se,O) thin films
on three different substrates, Au-polished, Au-unpol-
ished and Au covered with isolating SiO2 (Rd=1 kV).
Similar trends can be observed on the three substrates,
consisting in an induction period, an initial peak, a
second broader peak, and a final decay. The induction
time shows little dependence on substrate type, whereas
rate and time of appearance of the peaks are substrate
dependent. It is observed that the peaks are more
intense on the unpolished Au surface, the first one
appearing at the same time on both polished and
unpolished surfaces, whereas the second one is some-
what delayed on the polished surface. On Au/SiO2

A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
979
3.2. Temperature
Growth rate curves of CBD–Zn(Se,O) thin films on
Au-unpolished substrate at four bath temperatures
(30, 40, 50 and 60° C) are shown in Fig. 2. The two
rate peaks already mentioned are observable between
60 and 40° C (curves 1–3), whereas at 30° C (curve 4)
the rate curve shows a first rise and fairly constant rate
afterwards. Arrhenius plots can be constructed for the
velocity of the peaks and for the inverse of the induc-
tion time (t−in 1), as shown in the inset of Fig. 2. The
activation energies (Eact) can be determined for the
corresponding processes:
Fig. 1. Growth rate curves corresponding to chemical deposition of
ZnSe thin films on three different substrates, Au-unpolished (1),
Au-polished (2), Au/SiO2 (3). The base composition was used at
50 °C. The vertical line indicates the time at which a precipitate is
distinguishable in solution.
Log 6=log A−Eact/kT,
(1)
where 6 is the velocity (i.e. peaks rate and t−in 1) and the
other factors have their usual meaning. The obtained
Eact are tabulated in Table 1 together with reaction
orders obtained for the three processes (see below). It
must be remarked that this analysis is less accurate for
the second peak, as the initial conditions may be altered
by the previous deposition peak. The curves of Fig. 2
show that temperature dependence of the growth rate
after 70–80 min is reversed, reflecting a growth regime
by deposition of particles formed in solution. In this
case, heavier clusters are formed at higher temperatures
which deposit preferentially on the bottom of the vessel
and not on the QCM substrate.
3.3. Zinc sulphate and selenourea concentrations
Fig. 2. Growth rate curves corresponding to deposition of ZnSe thin
films at 60 °C (1), 50 °C(2), 40 °C (3) and 30 °C (4), on Au-unpol-
ished using base composition. Lines indicate the time at which
turbidity starts in solution. Inset: Arrhenius plot for the rates of the
induction process (t−in 1), peak 1 and peak 2 (Fig. 1).
Growth curves of CBD–Zn(Se,O) thin films on Au-
unpolished substrate, at different selenourea and zinc
sulphate concentrations are shown in Figs. 3 and 4,
respectively. The range of concentration for each com-
pound covers the values for which good film quality is
obtained. It is observed for some of the curves an initial
period of apparent negative growth rate, which seems
more pronounced with low SU and high ZnSO4 con-
centrations, probably attributed to stress generated ini-
tially by the film, as occurs in the electrodeposition of
metallic layers [14]. In other cases, oscillations may be
observed during the induction time, reflecting deposi-
tion–dissolution and/or adsorption–desorption pro-
cesses before the start of film growth.
surface only one broad peak appears. A vertical line in
the figure indicates the appearance of precipitate in
solution, which is simultaneous with the initial peak for
the Au substrates, and some minutes before the broad
maximum of the Au/SiO2 surface. After 70 min, the
growth rate is identical for the uncovered Au sub-
strates, and lower on the Au/SiO2.
Table 1
Experimental activation energies (Eact) and reaction orders (n), for the induction process and the two growth peaks
Eact (kJ mol−1
)
n
ZnSO4
NH2CSeNH2
Na2SO3
N2H4
NH3
Induction
Peak 1
(Peak 2)
61
50
(6)
−0.3
0.3
(0.5)
1.3
0.24
(1.1)
0
0.3
(0.4)
0
1.8
–
−1.7
0.7
(B0)

980
A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
The insets of the Figs. 3 and 4 show logarithmic plots
of 6 vs. reactant concentration for the induction and
growth peaks, from which the experimental reaction
orders (n) can be obtained using the relation:
Log 6=log k+n log C,
(2)
where k is the experimental rate constant and C the
concentration. The results are given in Table 1 for the
three processes. The observed negative order respect to
ZnSO4 for the induction process is related with the
acidification produced by Zn2+ cations (see Section 4).
The values of pH experimentally measured in the CBD
solution, and inverse induction time (t−in 1) are given as
a function of zinc sulphate concentration in Fig. 5.
Fig. 3. Growth rate curves corresponding to deposition of ZnSe thin
films on Au-unpolished at 50 °C, using [SU]=15 mM (1), 7.5 mM
(2), 5 mM (3), 1.5 mM (4), and the other components as in base
composition. Inset: reaction order determination respect to [SU] for
induction, peak 1 and 2.
3.4. Sulphite
Growth curves under different concentrations of sul-
phite are given in Fig. 6. In the absence of sulphite
(curve 1), the initial growth peak is enhanced to 4.5 nm
min−1, followed by a fast decay at the same time that
red Se0 precipitate is observed in the solution. The
resulting film is very thin and of poor quality. The
addition of sulphite avoids the precipitation of Se0
(curves 2 and 3). Under the highest sulphite concentra-
tion some increase in the growth rate is observed with-
out change of the shape of the curve. The obtained
reaction orders from the plot in the inset of Fig. 6 are
in Table 1.
3.5. Hydrazine
Growth curves at different hydrazine concentrations
are shown in Fig. 7. The main effect of hydrazine is to
enhance the velocity of the first peak, whereas the
induction time is not affected by changes in its concen-
tration. However, in the absence of hydrazine (curve 4),
the induction time is longer (tin=9 min), and the rate
Fig. 4. Growth rate curves corresponding to deposition of ZnSe thin
films on Au-unpolished at 50 °C, using [ZnSO4]=300 mM (1), 60
mM (2), 30 mM (3), 8 mM (4), and the other as in base bath
composition. Inset: reaction order determination respect to [ZnSO4]
for induction, peak 1 and 2.
Fig. 6. Growth rate curves corresponding to deposition of ZnSe thin
films on Au-unpolished at 50 °C, using [Na2SO3]=0 M (1), 30 mM
(2), 60 mM (3), and other components as in base. Inset: reaction
order determination respect to [Na2SO3] for peak 1 and 2.
Fig. 5. Plot of the experimental pH and (t−in 1) as a function of
[ZnSO4]. Other bath conditions as in Fig. 4.

A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
981
Two sets of growth curves under high ammonia
concentration (2.8 M) with and without hydrazine in
the chemical bath and under variable selenourea con-
centration are given in Fig. 9. One peak is observed,
which is more intense in the presence of hydrazine (Fig.
9a). The induction time follows the same trend as
observed in Figs. 2 and 7, being longer at lower sele-
nourea concentration, and in the absence of hydrazine.
The obtained reaction order respect to selenourea, for
induction (n=1.0) and the first peak (n=0.24) are in
agreement with those obtained in Fig. 3 with Au
substrate.
Fig. 7. Growth rate curves corresponding to deposition of ZnSe thin
films on Au-unpolished at 50 °C, and [N2H4]=6.4 M (1), 3.2 M (2),
1.6 M (3), 0 M (4), and the other components as in base conditions.
Inset: reaction order determination respect to [N2H4] for peak 1.
4. Discussion
In view of the experimental results of the previous
section, the growth of Zn(Se,O) films in a chemical bath
curve shows one unique peak of lower velocity, fol-
lowed by a plateau, and negative growth velocity after
150 min which indicates film dissolution. The reaction
order for the first growth peak is obtained from the plot
in the inset (Table 1).
3.6. Ammonia concentration
Fig. 8 shows growth rate plots with different ammo-
nia concentrations in the solution bath. Ammonia in-
creases the induction time and the intensity of the
initial peak. Reaction order for induction and first peak
are determined in the inset (Table 1). The second peak
seems absent (or overlapped by the first) at high con-
centrations, so it is not included in the analysis. Pt
substrate was used in these experiments, instead of Au
as in previous figures.
Fig. 9. (a) Growth rate curve corresponding to deposition of ZnSe
thin films on Pt-unpolished, at 50 °C, 2.8 M NH3 and 1.6 M N2H4
Fig. 8. Growth rate curves corresponding to deposition of ZnSe thin
films on Pt-unpolished, at 50 °C and [NH3]=2.8 M (1), 2.1 M (2),
1.4 M (3), 0.7 M (4), and other as in base composition. Marked the
time at which turbidity starts in solution. Inset: reaction order
determination respect to [NH3] for induction and peak 1.
(a) and
0 M N2H4 (b), at different selenourea concentrations:
[NH2CSeNH2]=30 mM (1), 15 mM (2), 7.5 mM (3), 3.8 mM (4) and
1.9 mM (5). Other bath conditions are in base. Insets: reaction order
determination respect to selenourea (SU) for induction and peak 1.

982
A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
(n= −0.3). On the other hand, little or no dependence
on the substrate surface is observed (Fig. 1). During the
induction, the reaction of selenourea with a nucleophile
agent in solution, like hydroxide, ammonia or hydra-
zine, must occur to generate the selenide ions. The main
nucleophile is hydroxide, as can be inferred from the
plot in Fig. 10 where the concentrations of OH−, free
NH3 and free N2H4 are given as a function of ZnSO4
concentration, together with the inverse induction time
(t−in 1). In this figure, the concentration of OH− was
obtained from Fig. 5, whereas NH3 and N2H4 were
calculated considering equilibrium reactions with Zn2+
and water. The parallel trend of (t−in 1)and OH− concen-
tration evidences that hydroxide is the main reactant
for the decomposition of selenourea:
−1
in
Fig. 10. Plots of [OH−], [NH3], [N2H4] and t
[ZnSO4]. Bath conditions as in Fig. 4.
as a function of
NH2CSeNH2+OH−“NH2C((O)NH2+HSe−
(3)
where decomposition of urea continues upon reaction
with more hydroxide, until the final formation of car-
bonate and ammonia. Fig. 11 shows that the experi-
mental reaction order for induction process is close to
one at high [OH−], and may be higher at lower concen-
tration. The negative order respect to zinc sulphate for
the induction process must be attributed to the con-
sumption of OH− by Zn2+ ions (Fig. 5). Zn2+ cations
participate in two main equilibria, with ammonia and
hydroxide:
can be divided into different stages. There is first an
induction process independent on substrate surface,
which is followed by one or two growth processes,
characterised by substrate dependent peaks in the
growth curves, and a final decay (Fig. 1). Two rate
peaks have also been observed in QCM experiments on
CBD-CdSe thin films [11], for which independent and
consecutive steps in the reaction mechanisms were pos-
tulated, consisting first in a cylindrical nucleation with
two-dimensional growth, and then a three-dimensional
growth. That analysis allowed for the application of a
model based on the geometry of the nucleation [15].
Our analysis here will concentrate on the chemical
nature of the processes, using the experimentally de-
termined reaction orders and activation energies of
Table 1.
Zn2+ +4NH3lZn(NH3)42+
K=108.2
K=1018
(4)
(5)
Zn2+ +4OH−lZnOH22− +2H2O
Being Zn(NH3)24+ predominant at the ammonia con-
centrations used in the bath. Once biselenide species are
formed after reaction 3, they may react with SO32−
(reactions 6 and 7) and with Zn(NH3)24+ (reaction 8):
4.1. Induction process
HSe−+1/2O2“Se+OH−
(6)
(7)
The determining species for this process are sele-
nourea (n=1.3), ammonia (n= −1.7), and ZnSO4
Se+SO23− lSeSO23−
Zn(NH3)24+ +HSe−+OH−“ZnSe+4NH3+H2O
(8)
where in bold are the solid compounds that contribute
to the growth of the film. The results of Fig. 6 show
that reactions 6 and 7 are the stabilising mechanism of
selenide in solution (a stronger reducer like hydrazine is
not effective in avoiding HSe− oxidation). SeSO32− may
be used as departing Se source for the deposition of
ZnSe thin films [16]. Reaction 8 is favoured by the high
insolubility of ZnSe (ks=3×10−17), and is responsible
for the appearance of the first solid ZnSe particles. The
negative order with respect to ammonia in the induc-
tion process is explained by the stabilisation of the
ammonia complex (reaction 4), and implies that reac-
tion 8 also determines the induction time. Hence, the
Fig. 11. Log (t−in 1) vs. log [OH−]. Bath conditions as in Fig. 4.

A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
983
induction process is governed by a net reaction of the
type:
which in the presence of the zinc complex can be
written as:
Zn(NH3)24+ +4H2O+4e−
NH2CSeNH2+Zn(NH3)24+ +2OH−
“Zn(OH)2+2H2+2OH−+4NH3
E= −0.72 V
“ZnSe+4NH3+H2O+NH2C(O)NH2
(9)
(12)
which includes the decomposition of selenourea (reac-
tion 3) and precipitation of ZnSe (reaction 8). The
small negative order respect to zinc sulphate indicates
that the participation of the zinc–ammonia complex as
a reactant in reaction 9 is more than counterbalanced
by the acidification effect (reactions 4 and 5). This
reaction gives rise to appearance of first ZnSe particles
in the bath, but it does not contribute significantly to
the growth of the film, as shown by the differences
between this process and the first growth peak reflected
by the kinetic data in Table 1.
This reaction gives rise to zinc hydroxide deposition,
which further dehydrate to ZnO at bath temperatures
above 50° C [23]:
Zn(OH)2“ZnO+H2O
(13)
The reduction of water in the presence of biselenide,
formed after reaction 3, and the zinc complex, yields
the electrochemical deposition of ZnSe:
Zn(NH3)24+ +3H2O+HSe−+4e−
“ZnSe+2H2+3OH−+4NH3
E= −0.55 V
(14)
4.2. Electroless mechanisms
Reactions 11–14 are favoured by the high negative
free energy of formation of the compounds (−555,
−318 and −145 kJ mol−1 for Zn(OH)2, ZnO and
ZnSe, respectively [21]), an effect that favours the elec-
trochemical deposition of compounds and alloys [24–
26]. It must be noticed that the addition of ammonia
diminishes the thermodynamic force of zinc compound
deposition through stabilisation of the aqueous ammo-
nia complex (reaction 4), whereas QCM experiments
show positive reaction order of ammonia for the first
maximum (Fig. 9 and Table 1). This result indicates
that the main effect of ammonia during the electroless
deposition mechanism is the acceleration of the oxida-
tion of hydrazine (reaction 10) by increment of OH−.
The resulting overall reactions for film growth in-
duced by oxidation of hydrazine are:
The first growth peak is characterised by small influ-
ence of selenourea (n=0.24) and ZnSO4 (n=0.2), but
high positive reaction orders of hydrazine (n=1.8) and
NH3 (n=0.7). It is also observed that roughness and
conductivity of the substrate have a positive effect, as
shown in Fig. 1. These facts indicate that the initial
growth of the film is due to the oxidation of hydrazine
on the surface of the substrate:
N2H4+4OH−“N2+4H2O+4e−
E°= −1.16 V
(10)
where E° is the standard reduction potential with re-
spect to NHE. This reaction occurs as a four electrons
oxidation process on Pt and Au electrode [17,18], with
greater catalytic activity on Pt, and important electrode
structure sensitiveness due to strong interactions [19].
The value of Eact=50 kJ mol−1 is comparable to that
measured for hydrazine induced electroless deposition
of metals [20]. Reaction 10 may give rise to the reduc-
tion of three species present in solution, ie. dissolved
oxygen, water, or Zn(NH3)24+. Most favoured thermo-
dynamically is the reduction of oxygen, which in the
presence of the zinc complex gives rise to ZnO
deposition:
Zn(NH3)24+ +1/2N2H4+2OH−+1/2O2
“1/2N2+4NH3+2H2O+ZnO
DG= −340 kJ mol−1
(15)
(16)
(17)
Zn(NH3)24+ +N2H4+2OH−
“N2+4NH3+H2O+2H2+ZnO
DG= −168 kJ mol−1
Zn(NH3)24+ +N2H4+OH−+HSe−
“N2+4NH3+H2O+2H2+ZnSe
DG= −236 kJ mol−1
Zn(NH3)24+ +1/2O2+2e−“ZnO+4NH3
E=0.60 V
(11)
where in reaction 16 has also been included reaction 13.
Deposition reactions 15–17 give rise to a film with a
mixed ZnSe–ZnO composition [7].
Further evidence of the electroless deposition mecha-
nism is given by XPS results of Zn(Se,O) layers de-
posited on different substrates, given in Table 2. It is
observed that the ZnO proportion in the layers is much
(E are reduction potentials calculated from the data in
Refs. [21,22] at 298 K). However, the concentration of
dissolved oxygen is probably very low (B10−3 M)
with hydrazine and sulphite in solution, so this reaction
may have limited importance for the formation of the
film. A second possibility implies the reduction of water

984
A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
Table 2
Composition of CBD–ZnSe thin films grown on conductive SnO2 substrate, semiconducting CuInS2 and glass substrates, using base bath
composition
Substrate
Zn/Se
ZnSe (%)
SeO2 (%)
Se (%)
Zn(O,OH),Zn (%)
SnO2
CuInS2
Glass
1.81
1.19
0.95
39
79
77
–
–
4
16
5
14
45
16
5
Films were grown at 50 °C bath temperature, during 15 min, on SnO2 and CuInS2, and 60 °C and 20 min on glass.
higher on the films deposited on conductive SnO2 (45%)
and semiconducting CuInS2 (16%), whereas the film on
isolating glass has the lowest oxide proportion (5%).
These results show that the electroless deposition of
oxide is favoured by the conductivity of the substrate.
Hydrazine may also reduce the Zn(NH3)24+ complex,
giving rise to an alternative route for film formation.
Reduction of the zinc–ammonia complex is experimen-
tally supported by detection of metallic Zn0 in CBD–
Zn(Se,O). The Zn LMM Auger signals reproduced in
Fig. 12 correspond to the three CBD–ZnSe films de-
posited under identical conditions on different sub-
strates, as in Table 2. The signal is composed of two
peaks, one due to Zn2+ located at 988–989 eV, and the
other assigned to Zn0 at 992 eV [27]. The intensity of
the Zn0 contribution follows the same trend as the
conductivity of the substrate, explained by the electro-
less formation. Zn0 will most probably be an intermedi-
ate for zinc compound deposition, being the reduction
of the zinc tetraamino-complex the first step:
Zn0+2OH− “ZnO+H2O+2e−
E= −2.04 V
(19)
Zn0+OH− +HSe− “ZnSe+H2O+2e−
E= −1.60 V
(20)
Reactions 19 and 20 liberate more electrons which
may further accelerate the reduction of the zinc com-
plex (reaction 18), giving rise to an autocatalytic effect.
The possibility of remaining Zn0 in the film may be due
to a passivation effect.
In the absence of hydrazine an electroless deposition
route is still possible from the oxidation of OH−:
4OH− “O2+2H2O+4e−
E°=0.40 V
(21)
Semireaction reaction 21 is kinetically difficult and
not probable to occur close to the standard potential
(already the single electron process OH−“[OH]S+e−
on Pt takes place at E= −0.05 V in alkaline solutions
[28]). However, considering final formation of ZnO
(reaction 11) may give rise to film growth:
Zn(NH3)24+ +2e−“4NH3+Zn0
E= −1.04 V
(18)
Zn(NH3)24+ +2OH−“4NH3+H2O+ZnO
DG= −78 kJ mol−1
(22)
This reaction alone is weakly favourable with a re-
ducer like hydrazine (reaction 10). But it is favoured
when followed by formation of ZnO and ZnSe:
This reaction has the same terms as a pure chemical
precipitation, but in this case catalysed by electroless
mechanism on a conductive substrate. This route seems
important under high ammonia concentration, as in
Fig. 9b, and explains the growth of zinc rich films also
in the absence of hydrazine [6].
Fig. 13a shows schematically the electroless deposi-
tion mechanism, exemplified by net reaction 16, with
oxidation of hydrazine, reduction of water and final
formation of ZnO (analogous schemes may be given for
reactions 15 and 17). The mechanism through the auto-
catalytic formation of Zn0 intermediate is schematised
in Fig. 13b (reactions 10, 18–20). The electroless depo-
sition proceeds while a positive driving force exits,
reactants are available and charge exchange on the
substrate is possible. It is observed that in time addition
of hydrazine does not recovers the initial growth rate,
which indicates that the electroless route may become
kinetically impeded by the growing film. As a result
Fig. 12. Zn LMM Auger signal corresponding to three CBD–ZnSe
thin films deposited under identical conditions on SnO2:F (1), CuInS2
(2) and glass (3) substrates.
a
second heterogeneous mechanism starts to
predominate.

A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
985
Fig. 13. Reaction schemes for the deposition of Zn(Se,O) thin films in a chemical bath. (a) Electroless deposition via reaction 10 (hydrazine
oxidation), reaction 12 (H2O reduction and Zn(OH)2 deposition) and reaction 13 (Zn(OH)2 dehydration), ie. net reaction 16. (b) Electroless
deposition with intermediate formation of Zn0 (reactions 10, 18–20). (c) Chemical deposition via reaction 24.
4.3. Heterogeneous chemical growth
DG= −137 kJ mol−1
(25)
Now, the negative order respect to ammonia is due
to the stabilisation of the zinc–ammonia complex. This
process gives rise to growth velocities smaller than
previous electroless mechanisms, and predominates af-
ter about 20 min of deposition on metallic substrates
(Fig. 1). It must be predominant on non-conducting
substrate (SiO2 covered Au substrate (Fig. 1)) and in
the absence of hydrazine and low ammonia concentra-
tion (Fig. 7, curve 3). The resulting film is less compact
than previous electroless deposited layers, and will have
larger proportion of ZnSe [7]. In Fig. 13c the chemical
deposition of ZnSe is schematised.
This mechanism is responsible for the second rate
peak observed after 20–25 min deposition. Its heteroge-
neous character is evidenced by the influence of the
substrate surface on the second peak (Fig. 1). The film
growth is now characterised by low activation energy,
about 6 kJ mol−1, and high reaction order for sele-
nourea (n=1.1, Table 1). These facts point to a reac-
tion controlled by the diffusion of soluble selenium
species. The predominant selenide precursor in solution
is selenosulphate, formed after reactions 6 and 7, which
hydrolysis liberates the biselenide anions:
SeSO32− +[OH−]S“SO42− +[HSe−]S
Zn(NH3)24+ +[HSe−]S+[OH−]S
“ZnSe+4NH3+H2O
(23)
4.4. Growth by ZnSe cluster deposition
(24)
In the end, the deposition of ZnSe clusters predomi-
nates, formed after reaction 25 in the bulk solution.
Diffusion of the clusters to the substrate is rate deter-
mining, competing with the deposition on the bottom
of the vessel, more important at higher temperature
when heavier clusters are formed. The layers will be less
compact, more porous, and mainly composed of ZnSe.
The adsorbed species are tentatively introduced to
account for the heterogeneous character of this mecha-
nism. The overall reaction is:
Zn(NH3)24+ +SeSO23− +2[OH−]S
“4NH3+H2O+SO24− +ZnSe

986
A.M. Chaparro et al. / Electrochimica Acta 47 (2001) 977–986
Kolb (Eds.), Advances in Electrochemical Science and Engi-
5. Conclusions
neering, vol. 6, Wiley-VCH, Weinheim, 1999 Chapter 3.
[2] A. Ennaoui, U. Blieske, M.Ch. Lux-Steiner, Prog. Photovolt.
Res. Appl. 6 (1998) 447.
The deposition of ZnSe thin film in a chemical bath
has been studied with a quartz crystal microbalance. The
presence in the chemical bath of a reducer, hydrazine,
necessary to adequate the properties of the film to
photovoltaic application, gives rise to film formation by
an electroless mechanism. The different stages of film
deposition have been identified. After an induction
period governed by formation of first ZnSe particles
(reaction 9), three consecutive mechanisms are found to
be responsible for the growth of the film. First, the
electroless mechanisms in the presence of hydrazine, with
reduction of water, gives rise to deposition of a mixed
ZnSe–ZnO film (reactions 15–17). Hydrazine induced
deposition may also occur via intermediate formation of
metallic Zn0 (reactions 18–20). An electroless route is
also possible by oxidation of hydroxide (reaction 21),
which is more important under high ammonia concentra-
tion. A second mechanism is the chemical reaction of the
soluble selenide precursor (SeSO23−) with hydroxide and
Zn(NH3)24+ resulting in a film with higher proportion of
ZnSe (reactions 23 and 24). Finally, the precipitation of
ZnSe clusters formed in the bulk gives rise to porous top
layers. This sequence of mechanisms explains the inho-
mogeneous composition observed in these films and the
presence of ZnO [7]. The ZnO content will modify the
transparency, conductivity and gap energy of the films,
which are determinant properties for their buffer layer
behaviour in thin film solar cells.
[3] A.M. Chaparro, R. Bayo´n, M.T. Gutie´rez, J. Herrero, J.
Klaer, K. Siemer, D. Bra¨unig, Proceedings of the 16th Eu-
ropean Photovoltaic Solar Energy Conference and Exhibition,
Glasgow, UK, 2000.
[4] K.L. Chopra, R.C. Kainthla, D.K. Pandya, A.P. Thakoor, in:
G. Hass, M. Francombe, J. Vossen (Eds.), Physics of Thin
Films, vol. 12, Academic Press, London, 1982.
[5] O. Savadogo, Sol. Energy Mater. Sol. Cells 52 (1998) 361.
[6] C.D. Lokhande, P.S. Patil, H. Tributsch, A. Ennaoui, Sol.
Energy Mater. Sol. Cells 55 (1998) 379.
[7] A.M. Chaparro, C. Maffiotte, M.T. Gutie´rrez, J. Herrero,
Thin Solid Films 358 (2000) 22.
[8] A.M. Chaparro, C. Maffiotte, J. Herrero, M.T. Gutie´rrez,
Surf. Interf. Anal. 30 (2000) 522.
[9] K.K. Kanazawa, J.G. Gordon II, Analytica Chimica Acta 175
(1985) 99.
[10] R. Ortega-Borges, D. Lincot, J. Electrochem. Soc. 140 (1993)
3464.
[11] H. Cachet, H. Essaaidi, M. Froment, G. Maurin, J. Elec-
troanal. Chem. 396 (1995) 175.
[12] J.M. Don˜a, J. Herrero, J. Electrochem. Soc. 142 (1995) 764.
[13] K.K. Kanazawa, in: H.-M. Tong, L.T. Nguyen (Eds.), New
Characterisation Techniques for Thin-Film Polymers, Wiley,
New York, 1990.
[14] E. Gileadi, V. Tsionsky, J. Electrochem. Soc. 147 (2000) 567.
[15] M.Y. Abyaneh, M. Fleishmann, J. Electrochem. Soc. 138
(1991) 2491.
[16] L. Grozdanov, Semic. Sci. Technol. 9 (1994) 1234.
[17] J.A. Harrison, Z.A. Khan, J. Electroanal. Chem. 26 (1970) 1.
[18] K. Koc´ınek, J. Koryta, M. Musilova, J. Electroanal. Chem. 21
(1969) 319.
[19] R. Go´mez, J.M. Orts, A. Rodes, J.M. Feliu´, A. Aldaz, J.
Electroanal. Chem. 358 (1993) 287.
[20] C.D. Iacovangelo, K.P. Zamoch, J. Electrochem. Soc. 138
(1991) 983.
[21] W.M. Latimer, Oxidation Potentials, 2nd ed., Prentice Hall,
New Jersey, 1952.
Acknowledgements
[22] D. Lide, H. Frederiksen (Eds.), Handbook of Chemistry and
Physics, 77th ed., CRC, Boca Raton, 1996.
This work was partially financed by the MARISOL
project (Comunidad de Madrid). A.M.C. thanks to
Ministerio de Ciencia y Tecnolog´ıa for a subvention for
his contract (Programa de Reincorporacio´n de Doc-
tores).
[23] T. Pauporte´, D. Lincot, Electrochim. Acta 45 (2000) 3345.
[24] S. Peulon, D. Lincot, J. Electrochem. Soc. 145 (1998) 864.
[25] F.A. Kro¨ger, J. Electrochem. Soc. 125 (1978) 2028.
[26] C. Ko¨nigstein, M. Neumann-Spallart, M. Sharon, Proceedings
of the 14th European Photovoltaics Solar Energy Conference,
Barcelona, 1997.
[27] T.F. McGee III, H.J. Comelissen, Appl. Surf. Sci. 35 (1988–
89) 371.
References
[28] K. Igarashi, A. Aramata, S. Taguchi, Electrochim. Acta 46
(2001) 1773.
[1] D. Lincot, M. Froment, H. Cachet, in: R.C. Alkire, D.M.