Impedance analysis and protonic conduction mechanism in RbH 2PO4/SiO2 composite systems

Article abstract of DOI:10.1016/j.electacta.2010.07.006

The protonic conductivity of (1 - x)RbH₂PO₄-xSiO ₂ composites prepared by mechanical milling has been investigated in the humid atmosphere. In comparison with pure RbH₂PO₄, the conductivity enhancemen

Full text of DOI:10.1016/j.electacta.2010.07.006

Electrochimica Acta 55 (2010) 7298–7304
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Impedance analysis and protonic conduction mechanism in RbH2PO4/SiO2
composite systems
Zikun Li∗
Department of Physics, Hong Kong Baptist University, 224 Waterloo Road, Kowloon, Hong Kong, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 April 2010
Received in revised form 28 June 2010
Accepted 3 July 2010
Available online 3 August 2010
The protonic conductivity of (1 − x)RbH2PO4–xSiO2 composites prepared by mechanical milling has been
investigated in the humid atmosphere. In comparison with pure RbH2PO4, the conductivity enhance-
ment was observed and it was strongly dependent on the amount of SiO2 additives. At low-temperature
(LT) phase, the conductivity of composites increased with the addition of SiO2 till x = 0.6 at which the
maximum conductivity reached. At high-temperature (HT) phase, it was approximately equal to that
of pure RbH2PO4 and almost kept unchanged till x = 0.5. Therefore, the jump in conductivity at ∼280 ◦C
(superprotonic phase transition temperature of RbH2PO4) tended to decrease as x increased and finally
disappeared. Furthermore, the conductivity of both LT and HT phases went down at large x values. The
conductivity enhancement has been confirmed not to be influenced by absorbed water on the surface of
SiO2 particles. In fact, it is due to the formation of structurally disordered or amorphous interface phase.
Also, we summarized the different theoretical models of the conductivity enhancement in composites.
© 2010 Elsevier Ltd. All rights reserved.
Keywords:
RbH2PO4
Composite
Impedance analysis
Conductivity
Amorphous phase
1. Introduction
Heterogeneous doping by the dispersion of inert oxide particles
in ionic salts has been shown to be an effective method not only
Fuel cells with solid acid electrolytes offer a considerable advan-
tage over the traditional ones based on solid oxides and polymers
due to the anhydrous proton transport, high electrocatalytic activ-
ity, a better tolerance of catalysts to CO, relatively low operating
temperatures and so on [1–7]. However, they have not been
developed sufficiently because of the absence of good protonic
conductors with high thermal stability and high conductivity at
intermediate temperatures (IT). CsHSO4 and CsH2PO4 are well
known as the proton-conductive solid acids for IT applications, their
conductivity increases drastically at 141 ◦C and 231 ◦C, respectively,
because of structural phase transition. Recently, fuel cells based on
CsHSO4 or CsH2PO4 electrolytes were demonstrated successfully
[1,2], but the conductivity of electrolytes is somewhat lower than
that requested for practical use, also, the operating temperature
range is very narrow as a result of the superprotonic phase tran-
sition, especially, the chemical instability of superprotonic phase
is an important limiting factor. Therefore, extensive investigation
of the solid acid electrolytes should be done in order to improve
the protonic conductivity, extend the operating temperature by
suppressing the phase transition, and stabilize the high conductive
phase.
to improve ionic conductivity, but also to obtain robustness. This
method was first reported by Liang in 1973 that the composites
of LiI and Al2O3 were observed to have a lithium ion conductiv-
ity up to 12 times higher than that of LiI alone [8]. Subsequently,
it stimulated a number of other studies and the ionic conductivity
enhancement indeed was observed in many similar heterogeneous
systems, e.g., AgI/X (X = SiO2, Al2O3, TiO2), AgCl/Al2O3, CuCl/Al2O3,
CsCl/Al2O3, etc. [9–17]. Generally, it was believed that the inter-
facial layers between both components play an important role
in the ionic conductivity enhancement and act as the additional
source of point defects [18], although the specific mechanism has
not yet been clarified. Considering the feasibility of solid acids
as fuel-cell electrolytes, the enhancement of their LT conductiv-
ity has been also investigated by heterogeneous doping approach.
Ponomareva et al. reported in a series of papers that the con-
ductivity of the LT phase in CsHSO4-doped system was improved
remarkably, however, affected by the specific surface area, pore
sizes, and molar ratio of inert oxides [19–24]. Quite recently, the
composite electrolyte based on CsH2PO4/SiO2 was reported, the
conductivity enhancement at LT was also found [25,26]. In the case
of CsH2PO4/SiO2 composite, which is different from CsHSO4/SiO2 in
the measurement process, the humidification of atmosphere was
necessary to prevent the salt dehydration. However, whether or
not the conductivity was influenced by the water absorbed on the
surface of SiO2 particles should be considered. As one member of
CsH2PO4 family, RbH2PO4 shows the same property as CsH2PO4:
∗
Corresponding author. Tel.: +852 3411 5873; fax: +852 3411 5813.
E-mail addresses: 08466173@hkbu.edu.hk, lizikun2004@163.com.
0013-4686/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2010.07.006

Z. Li / Electrochimica Acta 55 (2010) 7298–7304
7299
superprotonic phase transition from monoclinic to an unidentified
phase at ∼276 ◦C, accompanying which is a jump in conductivity.
This has been confirmed in our previous report [27]. Therefore, it
is of considerable scientific and practical interest to find heteroge-
neous additives capable of increasing the protonic conductivity of
RbH2PO4 at its LT phase.
In this paper, our main objective is to investigate the influ-
ence of heterogeneous doping of SiO2 particles in RbH2PO4 on the
protonic conductivity, and elucidate the relationship between the
structure of composites and conductivity enhancement through
ac impedance analysis. Generally, the water absorbed on the sur-
faces of highly dispersed SiO2 was considered as a possible factor
to influence the conductivity. In order to find out the real rea-
son for conductivity enhancement, we also investigated the effects
of atmospheric humidity on its conductivity. In the last part, we
summarized the different theoretical models of the conductivity
enhancement and discussed the protonic conduction mechanism
in CsH2PO4/SiO2 composite systems.
2. Experimental
Single crystals of RbH2PO4 were grown by slow evaporation of
water at room temperature from an aqueous solution of rubidium
carbonate (Rb2CO3) and phosphorous acid (H3PO4) in which the
Rb:PO4 molar ratio was set at 1:1. The polycrystalline samples were
prepared by milling single crystals in an agate mortar. SiO2 particles
with specific surface area at Ssp ≈ 150 m2/g determined by the BET
method were used as dispersing matrices of the composites. They
were calcined in air at ∼500 ◦C for 3 h before use. The molar frac-
tion of SiO2 additives in the (1 − x)RbH2PO4–xSiO2 composites was
varied from 0 to 0.7, corresponding to 0–55 vol.% SiO2. The compos-
ites were prepared by using a high-energy ball mill, mechanically
milled for 24 h, and then pressed at 7 t for 5 min to form pellets. The
resultant pellets were heated at ∼250 ◦C, close to the melting point
of RbH2PO4, for 20 min under highly humidified N2 flow condition
with 56 mol% H2O/N2. The pellet’s relative density decreased as x
increased from 0 to 0.7 and was about 75–85%. Silver electrodes
were attached to both sides of the pellets by silver paint following
to silver vapor deposition for impedance analysis.
X-ray diffraction (XRD) measurement of the product was per-
formed in a Rigaku Rotaflex RU-200B diffractometer using Cu-K␣
radiation (ꢀ = 1.5418 Å), to confirm its identity. Differential scan-
ning calorimetry (DSC) was performed in a Perkin-Elmer Diamond
DSC thermal analyzer. The sample was sat in a platinum pan, to
be heated at a fixed rate of 10 ◦C/min, to 475 ◦C, in a dynamic
atmosphere of dried N2 flowing at 50 ml/min. Impedance mea-
surements were carried out with an HP 4284A Precision LCR Meter
over a frequency range of 20 Hz to 1 MHz. The pellet stayed inside
a tube furnace. The atmosphere with water partial pressure at
∼0.56 atm was maintained by flowing N2 that had bubbled through
water heated to around 90 ◦C; the relative humidity at the gas
outlet was read from a sensor and held near 80%. The samples
were kept for 30 min at each temperature so they could stabilize
before the impedance measurement. The conductivity, ꢁ, of pure
RbH2PO4 and its composites was calculated from resistance and
pellet dimensions, i.e., ꢁ = L/(AR), where R is resistance and repre-
sents R1 + R2 for composites, L and A are thickness of the pellet and
electrode surface area, respectively.
Fig. 1. (a) Temperature dependence of the conductivity of (1 − x)RbH2PO4–xSiO2
composites; (b) conductivity isotherms as a function of molar fraction x at different
temperatures.
was adopted. It was shown that RbH2PO4 undergoes superpro-
tonic phase transition at ∼276 ◦C and the abrupt increase in
conductivity is not related to the presence of water. Additionally,
after the heating/cooling cycles reproducible values of conduc-
tivity were also observed. The same result has been obtained
in other solid acid like CsH2PO4 [25]. This confirmed that the
protonic conductivity of RbH2PO4 is intrinsic, not caused by the
possible absorbed water. Therefore, the conductivity measurement
of RbH2PO4-based composites was also carried out under this
humidity condition. Temperature dependence of the conductiv-
ity for (1 − x)RbH2PO4–xSiO2 composites (x = 0–0.7) is presented
in Fig. 1(a). As can be seen clearly, the conductivity of both pure
RbH2PO4 and composites abruptly changed at around 280 ◦C. For
pure RbH2PO4, the conductivity increase by about 3 orders of
magnitude is due to a structural transformation [27]. For the
composites, however, their properties are very different at high
(HT > 280 ◦C) and low (LT < 280 ◦C) temperatures. The conductiv-
ity depends on x, at LT increasing gradually with addition of SiO2,
whereas at HT approximately equal to that of pure RbH2PO4 till
x = 0.5. The jump in conductivity tended to decrease as x increased
and finally disappeared at x = 0.6. With increasing molar fraction
of SiO2 additives above x = 0.6, the conductivity at both LT and HT
began to decrease. These results are approximately consistent with
those previously reported by Ponomareva et al. in terms of CsHSO4
[19–22], although the ionic salts under study were different. The
phase transition temperature of RbH2PO4 in the composites almost
kept unchanged at ∼280 ◦C from x = 0 to 0.6. This point is somewhat
different from previous results in other composite systems such as
CsHSO4/SiO2 [19–22], CsH5(PO4)2/SiO2 [29], XNO3/SiO2 (X = Cs, Rb)
3. Results
The thermal properties of pure RbH2PO4 have been studied in
detail in our previous report [28], its initial dehydration takes place
at ∼250 ◦C. In order to suppress the dehydration and keep com-
plete structure of RbH2PO4 humidified N2 flow (PH ≈ 0.56 atm)
O
2

7300
Z. Li / Electrochimica Acta 55 (2010) 7298–7304
4.2. Impedance analysis of protonic conduction
Our experimental results suggest that the ionic conduction
is strongly related with the structure of heterogeneous doping
systems. Here the relationship between conductivity and struc-
ture of RbH2PO4-based composites is discussed with reference
to impedance analysis. Typical impedance spectra obtained from
pure RbH2PO4 and RbH2PO4/SiO2 composites, below and above the
superprotonic phase transition, are shown in Fig. 3 in Nyquist (Zꢀ
vs Zꢀꢀ as parametric functions of frequency). For pure RbH2PO4 in
LT region, a semicircle is seen clearly (Fig. 3(a)). In this case, the
effective dc resistance of the sample was determined by fitting the
data to an equivalent circuit model composed of a resistor, R, and a
constant phase element, Q, in parallel with each other. Its complex
impedance is defined as:
Fig. 2. Humidity dependence of the conductivity of (1 − x)RbH2PO4–xSiO2 compos-
1
ZQ
=
(1)
ites at different temperatures and molar fractions.
[Y(iω)n]
where Y and n are frequency-independent parameters, i and ω
denote imaginary unit and angular frequency, respectively. For
n = 1, the constant phase element reduces to an ideal capacitor with
the capacitance Y, and for n = 0 to a simple resistor with the admit-
tance Y. However, this semicircle no longer exists in HT region
because the protonic conduction occurs too fast to be tested with
the present impedance analyzer. On the contrary, a slow protonic
diffusion process at the interface between electrode and elec-
trolyte, and capacitance-like behavior was observed as a straight
line (Fig. 3(b)). Its dc resistance can be determined by extrapolat-
ing the impedance spectra to the real axis. For the composite of
30 mol% RbH2PO4 in LT region, its impedance showed two overlap-
ping semicircles (Fig. 3(c)). The equivalent circuit model consists
in series of two parallel combinations of a resistor and a constant
phase element. The higher-frequency semicircle is identified as
semicircle 1 and the lower-frequency semicircle as semicircle 2 for
convenience. In HT region, however, only one electrode response
is apparent within the measured frequency range as a straight line
just as the pure RbH2PO4 (Fig. 3(d)).
Each parameter value of the equivalent circuit can be obtained
by using the least-squares refinement program equivalent circuit.
As for the overlapping semicircles, we could first determine R1 by
the fitting program, and R2 was obtained by subtracting R1 from
the overall resistance. According to the fitting data and semicir-
cles in Nyquist plot, n value in Eq. (1) was very close to unity and
the constant phase element Q could be regarded as a capacitor. Its
capacitance C could be calculated from the resistance R and peak
angular frequency ωp using the relationship:
[30,31]. On the other hand, the activation energy for proton trans-
port seems not to depend on x and is equal to 0.75–0.80 eV at LT
and 0.52 eV at HT.
Conductivity isotherms as a function of molar fraction are
presented in Fig. 1(b). The LT conductivity goes through a maxi-
mum at x = 0.45–0.65, corresponding to 30–50% volume fraction,
which is about 2 orders of magnitude higher than that of pure
RbH2PO4. The reason for conductivity enhancement may be the
formation of pathways, which lead to highly proton-conductive
interfaces. We will discuss it in detail in later section. Moreover, the
HT conductivity of composites with 0.1 < x < 0.5 is slightly higher
than that of pure RbH2PO4, and then slowly decreases at x ≥ 0.6.
This conductivity enhancement is probably the result of molten
RbH2PO4 stabilized on the surface of highly dispersed SiO2 par-
ticles. Both HT and LT conductivity decrease at large x values
is due to the percolation effect of the conductor–insulator type,
where SiO2 acts as insulator phase embedded into the proton-
conductive medium and inhibit the protonic conduction pathways
when the SiO2 concentration surpassed the percolation thresh-
old.
4. Discussion
4.1. Water-effect on the conductivity
In principle, the conductivity enhancement of RbH2PO4-based
composites in comparison with pure RbH2PO4 at LT could be caused
by water absorbed on the surfaces of dispersed SiO2 particles. In
order to eliminate the possible contribution of water, conductivity
measurement of pure RbH2PO4 and its composites under differ-
ent humidity conditions were carried out, respectively. Here, we
only repeated the experiment below 250 ◦C at which the thermal
decomposition of RbH2PO4 occurs, so as to avoid the interference of
dehydration. As we can see in Fig. 2, the different relative humidity
of atmosphere (PH = 0.007 − 0.56 atm) did not influence their
1
(RC)
ωp
=
(2)
As shown in Fig. 4, the value of Q1 in composites, C1, is almost in
the same order (∼10−10) as that in pure RbH2PO4, C. Therefore,
the protonic conduction process for semicircle 1 in the composites
is primarily associated with the intrinsic proton transport of pure
RbH2PO4. Taking into account the fact that semicircle 2 appeared
with addition of SiO2 particles into RbH2PO4, it was considered to
be attributable to the dispersion of SiO2.
The resistances within the error of 5%, R of pure RbH2PO4, R1
and R2 of composites with increasing temperature to 260 ◦C, are
presented in Fig. 5. Clearly, R1 is significantly smaller than that
of pure RbH2PO4 at each temperature. Furthermore, comparing
R1 and R2 in the composites with different SiO2 molar fractions,
R2 is always far more than R1. This suggests that R2 is dominant
for the conductivity of composites. The significant decrease in R2
thus could result in the increase in conductivity of composite. In
addition, R1 for the case of x = 0.3 is approximately equal to that
of x = 0.5, however, the difference between R2 values for these two
O
2
LT conductivity, which confirms that the conductivity enhance-
ment of RbH2PO4/SiO2 composites occurs via structural protons
rather than hydronium ions. Though some authors reported that
the composite conductivity can be affected to some extent by the
adsorbed water on SiO2 particles, we claim that the water adsorp-
tion cannot explain the significant increase of their conductivity,
in particular, at the temperature below the superprotonic phase
transition. Therefore, the adsorption of water is not the dominant
factor for the conductivity enhancement, other reasons should be
considered.

Z. Li / Electrochimica Acta 55 (2010) 7298–7304
7301
Fig. 3. Typical Nyquist plots of (1 − x)RbH2PO4–xSiO2 composites at: (a) x = 0.0 and T = 200 ◦C; (b) x = 0.0 and T = 285 ◦C; (c) x = 0.3 and T = 200 ◦C; (d) x = 0.3 and T = 285 ◦C.
cases becomes very apparent. It is claimed that R2 strongly depends
on the amount of SiO2 particles. As we know, a more dispersive
distribution of SiO2 particles occurs in the composites with higher
SiO2 content. The agglomeration of SiO2 particles and the formation
of voids or cracks also take place subsequently, which can inhibit
the protonic conduction pathways. Therefore, it is considered that
semicircle 2 is associated with slow protonic conduction at the
grain boundary generated by the agglomeration of SiO2 particles
and the formation of voids. However, the conductivity can be sig-
nificantly improved by more adequately controlling the dispersed
structure of SiO2 particles. Prolonging the time limit of mechan-
ical milling of RbH2PO4/SiO2 composites in the agate mortar is a
not only simple but also effective method in terms of the structure
control. As shown in Fig. 5(a) and (b), after increasing the milling
time from 24 to 48 h, R2 values decrease significantly because the
mechanical milling promotes the dispersion of SiO2 particles and
eliminates their agglomeration and voids to some extent.
On the other hand, the corresponding R1 values exhibit the sim-
ilar property: a slight shift down with increasing the milling time.
What is the reason for such a behavior? Generally, it has been
accepted that an interface phase forms between ionic salt and inert
oxide particles owing to their interface interaction. If this interface
interaction is strong enough then it can result in the stabilization of
partially or completely disordered state of ionic component at the
interface, e.g., or amorphous phase, resulting in the ionic conduc-
tivity enhancement in many composite systems [11,32–34]. The
experimental results in Fig. 1 have shown that the superprotonic
phase transition temperature of RbH2PO4 in composites almost
keeps unchanged at ∼280 ◦C with the addition of SiO2. Therefore,
no superprotonic phase of RbH2PO4 exists at the interfacial layer
below 280 ◦C. The interface phase is thought to be structurally dis-
ordered or amorphous rather than superprotonic which will be
further confirmed in next section. The second advantage of prolong-
ing the mechanical milling time is to increase the contact interface
between RbH2PO4 and SiO2, and promote the formation of amor-
phous phase of RbH2PO4 on the surface of SiO2 particles, resulting
in the slight decrease in R1. As XRD patterns in Fig. 5(c) show,
the intensity of peaks decrease and the peaks become broadened
with increasing the milling time, suggesting either a decrease of
the grain size or a partial amorphization of RbH2PO4. Therefore,
RbH2PO4 in the composites has two main phases, i.e., a bulky phase
and an amorphous interface phase. We believe that semicircle 1 is
Fig. 4. Capacitance values of pure RbH2PO4 and its composite in LT phase plotted
against temperature.

7302
Z. Li / Electrochimica Acta 55 (2010) 7298–7304
host ionic salt (phase A) as a result of the interface interac-
tion. An increase in defect or charge carrier concentration in
the space–charge region enhances the ionic conductivity of
the composites. However, the assumption of such a model has
a limit that the volume fraction of phase B should be much
smaller than that of phase A. It cannot account for the case of
high volume fraction of phase B.
(II) Percolation model or random-resistor model (RRM)
This type of model [37–40] regards the composite as three
components due to their distinct interfacial properties, i.e.,
normally conducting, non-conducting and highly conducting
bonds. The bonds are distributed in space according to a ran-
dom distribution of phase B in the normally conducting phase
A and the highly conducting interface in between. It shows
the existence of two percolation thresholds, namely, inter-
face percolation at concentration equal to that of phase B,
and bulk percolation at concentration equal to that of phase
A, where a conductor–insulator transition takes place. How-
ever, this model can only give a numerical solution without a
quantitative analytical relationship between conductivity and
microstructure.
(III) Defect-induced order–disorder phase transition
This type of model focuses on more microscopic studies
including the interaction between the defects at two-
phase boundary in the composites. It was claimed that an
order–disorder phase transition is promoted due to plenty of
defects existing in the interface region, giving rise to the max-
imum ionic conductivity at temperature far below the bulk
superionic transition [41]. In other words, the superionic phase
of ionic salt in the composites can be realized even at lower
temperature than its bulk superionic phase transition tem-
perature. Experimentally, this superionic phase transition, i.e.,
sublattice melting phenomenon has been observed at ionic salt
thin film interface or thin film itself [42,43].
(IV) Formation of a disordered or amorphous phase
By the investigation of AgI/Al2O3 nanocomposites, Uvarov et
al. [32–34] pointed out that the change of bulk characteristics
of ionic salts in the nanocomposites takes place due to the for-
mation of a new interface phase between ionic salt and inert
oxides. The main reason for the stabilization of interface phase
is the adhesion energy. The adhesion is the result of the inter-
face interaction, leading to the formation of the double layer
formed by the point defects in the interface region. In the case
of the strong adhesion, the structural reconstruction or the for-
mation of the metastable interface phase takes place. Their
structure may be either epitaxial crystalline, or amorphous.
If the surface structures are very different or when compos-
ites are obtained in strongly non-equilibrium conditions, the
interface phase may be amorphous. The amorphous phase,
being disordered structurally, facilitates ionic conductivity of
composites.
Fig. 5. Temperature dependence of the resistances: (a) R1 and (b) R2 of
(1 − x)RbH2PO4–xSiO2 composites; (c) XRD patterns of the composite at x = 0.3 with
different mechanical milling times.
associated with protonic conduction through the bulky part of
RbH2PO4 and the interface phase which cannot be separated indi-
vidually by the impedance measurement.
4.3. Mechanism discussion of protonic conduction at interface
phase
The atomic mechanism of the interface interaction is not yet
clear. Generally, it can be regarded as a chemical adsorption of
ionic salt on the surface of inert oxides, interionic, dipole–dipole
and Van der Waals interactions between surface atoms [32]. In
order to explain the conductivity enhancement in the compos-
ites in more detail, a comprehensive model of ionic conduction
should be constructed. Combining with our experimental results
and the impedance analysis of protonic conduction, we agree well
with model (IV) stated above: an amorphous interface phase on
the surface of SiO2 particles is the source of protonic conductivity
enhancement. DSC and XRD data for pure RbH2PO4 and compos-
ites with different SiO2 content are presented in Fig. 6. The DSC
curve of pure RbH2PO4 exhibited two main endothermic peaks at
∼250 and ∼280 ◦C, corresponding to the initial dehydration and
Heterogeneous doping by the dispersion of inert oxide particles
in ionic salts has been extensively studied as an effective approach
to enhance ionic conductivity. Many theoretical studies have been
devoted to explain the mechanism of such kind of composites, and
several ionic conduction models have been proposed as follows:
(I) Space–charge model
Jow and Wagner [35,36] are the first who attempted to
explain the enhancement of ionic conductivity in composite
systems. They proposed that ultrafine dispersoids (phase B)
form well-defined space–charge regions in contact with the

Z. Li / Electrochimica Acta 55 (2010) 7298–7304
7303
Fig. 7. Schematic representation of protonic conduction pathway in the
RbH2PO4/SiO2 composites.
Therefore, the formation of amorphous ionic salt on the surface of
dopants is the main reason for the conductivity enhancement at
LT.
5. Conclusions
The conclusions obtained through the conductivity measure-
ment of RbH2PO4/SiO2 composites and the identification of
protonic conduction process by impedance analysis are summa-
rized below:
Fig. 6. (a) DSC curves and (b) XRD patterns of (1 − x)RbH2PO4–xSiO2 composites at
different molar fractions.
(1) The conductivity of (1 − x)RbH2PO4–xSiO2 composites is
strongly related with the SiO2 content. In comparison with pure
RbH2PO4, its LT conductivity increases with the addition of SiO2
at x ≤ 0.6, then decreases when x > 0.6. However, the HT conduc-
tivity almost keeps unchanged at x ≤ 0.5, then decreases when
x ≥ 0.6. Therefore, the jump in conductivity at ∼280 ◦C gradually
decreases as x increases and finally disappears.
(2) Water-effect on the conductivity has been also investigated.
The absorbed water on the surface of SiO2 particles indeed
has no any contribution to the conductivity enhancement of
composites.
(3) Different theoretical models of the conductivity enhancement
in composites have been summarized. The impedance analysis,
DSC and XRD data support the position that the conductivity
enhancement of RbH2PO4/SiO2 composites should be due to the
formation of interface phase on the surfaces of SiO2 particles,
which is structurally disordered or amorphous.
superprotonic phase transition. As x increases, the enthalpy of
superprotonic transition of RbH2PO4 in the composites signifi-
cantly decreases and this endothermic peak is almost indiscernible
at x = 0.5. This proves that thermodynamic parameters of RbH2PO4
change markedly due to its heterogeneous doping. In the case
of XRD patterns, the diffractograms of composites before heating
treatment are identical to that of pure RbH2PO4. However, the
intensity of all reflections decreases and the main signal peaks
slightly broaden with increasing SiO2 molar fraction after HT treat-
ment, which cannot be associated with amorphous SiO2. This
phenomenon suggests that mixing RbH2PO4 and SiO2, and sinter-
ing this mixture induce structural disorder of RbH2PO4. Therefore,
the line shapes of DSC and XRD confirm the formation of a new
phase of RbH2PO4 in composites, which is amorphous on the
surfaces of SiO2 particles. On the other hand, the mechanical
milling using a high-energy ball mill was adopted and the amor-
phous RbH2PO4 is easier to form to improve the conductivity of
RbH2PO4-based composites. The protonic conduction process is
represented schematically in Fig. 7. Generally, different pathways
involving bulk pathway, transport across boundary as well as along
boundary exist in a heterogeneous system. However, a fast pro-
tonic conduction pathway via the interfacial layer appeared in
RbH2PO4/SiO2 composites after treatment. We have known that
the RbH2PO4/SiO2 composites consist of three parts, i.e., RbH2PO4
bulk, SiO2 bulk, and the interface phase between them which
is disordered or amorphous structure of RbH2PO4 [44]. Through
impedance analysis, the protonic conduction can be classified as
two processes: the first one through RbH2PO4 bulk and the inter-
face phase which corresponds to high-frequency semicircle 1,
and the second one through the gain boundary generated by the
agglomeration of SiO2 particles and the formation of voids which
corresponds to low-frequency semicircle 2. The former process is
a positive effect on conductivity, whereas the latter is negative.
Acknowledgment
This work is supported financially by the Research Grant Council
of Hong Kong (HKBU 210907).
References
[1] S.M. Haile, D.A. Boysen, C.R.I. Chisholm, R.B. Merly, Nature 410 (2001) 910.
[2] D.A. Boysen, T. Uda, C.R.I. Chisholm, S.M. Haile, Science 303 (2004) 68.
[3] S.M. Haile, C.R.I. Chisholm, K. Sasaki, D.A. Boysen, T. Uda, Faraday Discuss. 134
(2007) 17.
[4] T. Matsui, S. Takeshita, Y. Iriyama, T. Abe, Z. Ogumi, J. Electrochem. Soc. 152
(2005) A167.
[5] T. Matsui, T. Kukino, R. Kikuchi, K. Eguchi, J. Electrochem. Soc. 153 (2006) A339.
[6] G.V. Lavrova, M.V. Russkikh, V.G. Ponomareva, N.F. Uvarov, Russ. J. Electrochem.
41 (2005) 485.
[7] G.V. Lavrova, M.V. Russkikh, V.G. Ponomareva, N.F. Uvarov, Solid State Ionics
177 (2006) 2129.
[8] C.C. Liang, J. Electrochem. Soc. 120 (1973) 1289.

7304
Z. Li / Electrochimica Acta 55 (2010) 7298–7304
[9] P. Chowdhary, V.B. Tare, J.B. Wagner Jr., J. Electrochem. Soc. 132 (1985) 123.
[10] Z.Y. Zhao, C.Y. Wang, S.Y. Dai, L.Q. Chen, Solid State Ionics 9–10 (1983) 1175.
[11] N.F. Uvarov, L.I. Brezhneva, E.F. Hairetdinov, Solid State Ionics 136 (2000)
1273.
[12] X. Glipa, J.M. Leloup, D.J. Jones, J. Roziere, Solid State Ionics 97 (1997) 227.
[13] K. Shahi, J.B. Wagner Jr., J. Electrochem. Soc. 128 (1981) 6.
[14] N.J. Dudney, J. Am. Ceram. Soc. 68 (1985) 538.
[27] Z.K. Li, T.B. Tang, Mater. Res. Bull., submitted for publication.
[28] Z.K. Li, T.B. Tang, Thermochim. Acta 501 (2010) 59.
[29] G.V. Lavrova, V.G. Ponomareva, Solid State Ionics 179 (2008) 1170.
[30] V.G. Ponomareva, G.V. Lavrova, L.G. Simonova, Solid State Ionics 136–137
(1999) 1279.
[31] G.V. Lavrova, V.G. Ponomareva, N.F. Uvarov, Solid State Ionics 136–137 (2000)
1285.
[15] M.C.R. Shastry, K.J. Rao, Solid State Ionics 51 (1992) 311.
[16] J. Maier, J. Phys. Chem. Solids 46 (1985) 309.
[17] J. Maier, Mater. Res. Bull. 20 (1985) 383.
[32] N.F. Uvarov, P. Vanek, J. Mater. Synth. Process. 8 (2000) 319.
[33] N.F. Uvarov, B.B. Bokhonov, A.A. Politov, P. Vanek, J. Petzelt, J. Mater. Synth.
Process. 8 (2000) 327.
[18] J. Maier, Solid State Ionics 75 (1995) 139.
[19] V.G. Ponomareva, N.F. Uvarov, G.V. Lavrova, E.F. Hairethinov, Solid State Ionics
90 (1996) 161.
[20] V.G. Ponomareva, G.V. Lavrova, L.G. Simonova, Solid State Ionics 118 (1999)
317.
[21] V.G. Ponomareva, G.V. Lavrova, L.G. Simonova, Solid State Ionics 119 (1999)
295.
[22] V.G. Ponomareva, G.V. Lavrova, Solid State Ionics 145 (2001) 197.
[23] V.G. Ponomareva, G.V. Lavrova, L.G. Simonova, Inorg. Mater. 34 (1998) 1136.
[24] V.G. Ponomareva, E.S. Shutova, Solid State Ionics 178 (2007) 729.
[25] J. Otomo, N. Minagawa, C.J. Wen, K. Eguchi, H. Takahashi, Solid State Ionics 156
(2003) 357.
[34] N.F. Uvarov, P. Vanek, M. Savinov, V. Zelezny, V. Studnicka, J. Petzelt, Solid State
Ionics 127 (2000) 253.
[35] T. Jow, J.B. Wagner Jr., J. Electrochem. Soc. 126 (1979) 1963.
[36] S. Jiang, J.B. Wagner Jr., J. Phys. Chem. Solids 56 (1995) 1101.
[37] A. Bunde, W. Dieterich, H.E. Roman, Phys. Rev. Lett. 55 (1985) 5.
[38] H.E. Roman, A. Bunde, W. Dieterich, Phys. Rev. B 34 (1986) 3439.
[39] H.E. Roman, M. Yusouff, Phys. Rev. B 36 (1987) 7285.
[40] A.G. Rojo, H.E. Roman, Phys. Rev. B 37 (1988) 3696.
[41] T. Ishii, J. Kawamura, J. Phys. Soc. Jpn. 67 (1998) 3517.
[42] N. Hainovsky, J. Maier, Solid State Ionics 76 (1995) 199.
[43] N. Hainovsky, J. Maier, Phys. Rev. B 51 (1995) 15789.
[44] H. Shigeoka, J. Otomo, C.J. Wen, M. Ogura, H. Takahashi, J. Electrochem. Soc. 151
(2004) J76.
[26] V.G. Ponomareva, E.S. Shutova, Russ. J. Electrochem. 43 (2007) 521.