Journal of Physics and Chemistry of Solids 144 (2020) 109505

Contents lists available at ScienceDirect

Journal of Physics and Chemistry of Solids

journal homepage: http://www.elsevier.com/locate/jpcs

Structural properties and electrical conductivity mechanisms of

semiconducting quaternary nanocomposites: Effect of two transition
metal oxides
Dipankar Biswas a, b, Anindya Sundar Das c, *, Rittwick Mondal d, Anindita Banerjee e,
Abhigyan Dutta e, Soumyajyoti Kabi f, Debasish Roy g, Loitongbam Surajkumar Singh b
a
Department of Electronics & Communication Engineering, Regent Education and Research Foundation, Barrackpore, Kolkata, 700121, West Bengal, India
b
Department of Electronics & Communication Engineering, National Institute of Technology Manipur, Langol, Imphal, 795004, India
c
Department of Electronics & Communication Engineering, Swami Vivekananda Institute of Science & Technology, Dakshin Gobindapur, Kolkata, 700145, West Bengal,
India
d
Department of Basic Science, Edrakpur High School, Birbhum, West Bengal, 731219, India
e
Department of Physics, The University of Burdwan, Burdwan, West Bengal, 713104, India
f
Department of Physics, Hijli College, Kharagpur, 721306, West Bengal, India
g
Department of Mechanical Engineering, Jadavpur University, Jadavpur, Kolkata, 700032, India

A R T I C L E I N F O A B S T R A C T

Keywords: Several quaternary nanocomposites, doped with two transition metal oxides, of the generalised composition
Quaternary nanocomposites formula 0.4ZnO–0.1P2O5–0.5[xV2O5–(1-x) MoO3] were prepared via the melt quenching process. We investi­
X-ray diffraction gated the microstructural and conductivity properties of two ternary systems and four quaternary systems to
Optical bandgap
determine the better semiconductor nanocomposite system for wider application purposes. The X-ray diffraction
Dc and ac conductivity
CBH and NSPT model
results showed the presence of superposed nanocrystallites within amorphous networks or matrices. Analysis of
ultraviolet–visible absorption spectra showed that optical bandgap energy (Eopt) values varied with V2O5 con­
centration, and there was an inverse relationship between Eopt and average nanocrystallite size. The responsible
mechanisms of ac conductivity were examined using the model of Jonscher’s universal power-law and
Almond–West formalism, and it was found that ac conductivity increased as temperature rose, exhibiting sem­
iconducting features. All the ternary (two examples) and quaternary (four examples) nanocomposite systems
established non-linearity in dc conductivity, caused by the small polaron hopping process with dissimilar acti­
vation energies at different temperature regions. The power-law exponent (s) of Jonscher’s universal power-law
revealed that the ac conductivity mechanism could be described by correlated barrier-hopping (CBH) or non-
overlapping small polaron tunnelling (NSPT) models for different samples due to major structural alterations.
We applied modified CBH and NSPT models (as s > 0.8) to get realistic values of different fitting parameters, and
in this process, we also obtained the values of ideal or hypothetical glass transition temperature from the
theoretical models. The different activation energies associated with ac conductivity and for the small polaron
migration process diminished with an increment in conductivity. Scaled ac conductivity spectra demonstrated
that the conductivity relaxation process relied upon the composite structural features and did not depend on
temperature. These materials can be used for such applications as gas sensors, even at higher temperatures due to
their semiconducting nature and the different valence states of transition metal ions.

1. Introduction glass structure by incorporating MoO4 tetrahedral or MoO6 octahedral

structural units with some other glass-forming oxides, like phosphorus
Molybdenum trioxide (MoO3), a transition metal oxide (TMO), pentoxide (P2O5) and vanadium pentoxide (V2O5). Quite a few ternary
normally does not function as a glass-forming oxide, but transforms the glassy systems with the combination of P2O5 and MoO3 have been

* Corresponding author.
E-mail address: anindyasundardas03@gmail.com (A.S. Das).

https://doi.org/10.1016/j.jpcs.2020.109505
Received 13 November 2019; Received in revised form 10 April 2020; Accepted 11 April 2020
Available online 21 April 2020
0022-3697/© 2020 Elsevier Ltd. All rights reserved.
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

reported [1–3]. The MoO3-doped phosphate glasses have electro-optical compare variations in the structural and conductivity properties among
applications caused by electrochromism properties and higher ionic all the as-quenched ternary and quaternary glass compositions to show
conductivity [4,5]. The glass formation and structural features of which system has better conductivity properties for diverse types of
ZnO–P2O5–MoO3 glasses have been reported [6]. Furthermore, the in­ applications.
teractions of molybdenum (Mo) in phosphate glasses are advantageous
to manage radioactive wastes [6]. The electrical transport mechanisms 2. Experiment section
in ZnO–MoO3–P2O5 glasses have been described by the small polaron
hopping (SPH) process [7]. Three bridging oxygen atoms and one 2.1. Materials and sample preparation
double-bonded oxygen atom make tetrahedral units of PO4, which is the
building block P2O5-based glasses [8]. In the absence of cations, the Four quaternary and two ternary glass nanocomposite samples of the
development of different phosphate anions takes place due to the linking common formula 0.4ZnO–0.1P2O5–0.5 [xV2O5–(1-x) MoO3], with the
of covalent bridging oxygen ions with PO4 tetrahedral units. The pres­ variation of x in steps of 0.2, were synthesised via melt quench tech­
ence of different valence states and also different co-ordinations, which nique. Suitable volumes of high-purity chemicals ZnO (Loba Chemie:
form corner-sharing VO4/VO5 polyhedral units, makes V2O5 a good 99% purity), V2O5 (Loba Chemie: 98% purity), MoO3 (Loba Chemie:
glass network former, and V2O5 also performs as a better glass network 99% purity), and P2O5 (Merck Specialities Private Limited: 98% purity)
modifier when combined with the phosphate-based glass nano­ were carefully mixed in ratios according to the composition formula of
composites. Transition metal ions (TMIs) are widely used as glass the individual system. Then the mixtures were heated to the melting
composite materials since their outer d-electron orbital functions have a point of the compositions in an electric furnace in atmospheric air
wide radial distribution and owing to their profoundly sensitive condition within a temperature range of 790� C-945� C depending on the
response to the variations in neighbouring cations [9–11]. As the longer molar ratio of the composition. Ternary bulk glassy systems, i.e. samples
covalent chains shorten and local structure reforms, the phosphate with x ¼ 0.0 and 1.0, melted at 945� C and 810� C, respectively. Whereas,
network depolymerises, evolving efficiently to a refined vanadium the quaternary bulk glass compositions (x ¼ 0.2 to 0.8) melted within
phosphate network with isolated PO4 tetrahedral units [11–13]. The the temperature range from 885� C to 785� C with an increase in V2O5
literature shows that the inclusion of TMIs enhances the chemical content. We continuously monitored the melting process in every step
durability and stability of glassy systems by substituting bridging P–O–P and, after the melting of the oxide materials, we checked the fluidity.
bonds in phosphate glasses with additional moistness resistant P–O–TM Then we kept the melt at the melting temperature for 20 min to
bonds [11–13]. homogenise and equilibrate the melt and to achieve better fluidity for
The electrical transport properties of TMI-doped phosphate glasses pouring. Then the melt was transferred onto highly polished aluminium
can be enhanced as mobile charge carriers or small polarons start the plate, and quickly quenched by pressing the melt in a technique using
hopping process from lower to higher valence states. Glasses doped with another highly polished aluminium plate at room temperature. Conse­
two TMIs may attain super-exchange of electrons to a great extent be­ quently, small glassy flakes of thickness ~0.2–0.5 mm were obtained.
tween the redox ion pairs, known as hopping, leading to the non-linear Then as-quenched glassy flakes were ground into a fine pure powder for
properties of activation energy [12,13]. Such non-linearity occurs owing XRD and UV–Vis absorption measurements.
to the mixed transition metal-ion effect (MTIE) [14–16]. In glassy sys­
tems with mixed TMIs, the electrons or small polarons also hop between 2.2. Volumetric mass density measurement
diverse TMI sites, for example, Mn2þ to Fe3þ and Fe2þ to Mn3þ in
Fe2O3–MnO–TeO2 glasses [12]. After doping with two TMOs, the con­ Using the renowned Archimedes principle, volumetric mass density
ductivity of the Fe2O3–P2O5–V2O5 glassy system reduces owing to the (ρ) values were measured by immersing samples in acetone because
Anderson localisation effect [13]. After adding MoO3 to the glassy sys­ acetone is non-reactive with the samples being examined. The ρ is
tem, Mo ions exist in the stable Mo5þ (donor level) and Mo6þ (acceptor considered a significant parameter, and an influential tool to study the
level) valence states. The glass matrix becomes more compact as Mo6þ structural modifications occurring within the polycrystalline or amor­
ions reside in the network-forming positions with MoO4 tetrahedral phous material. The value of ρ of all bulk glass samples was measured
units. Contrariwise, in Mo5þO3 complexes, Mo5þ ions are positioned in using the following relationship [21]:
place-modifying positions [17]. In general, the coexistence of V4þ � �
(donor level) and V5þ (acceptor level) states of V ions are observed in ρ¼
Wair
� ρacetone (1)
vanadate-based glasses. The octahedral V4þO6 and VO4 tetrahedral sites Wair Wacetone
are formed due to the presence of V4þ ions, while V5þ ions create VO5
where ρacetone is the density of acetone and Wacetone and Wair are the
pyramidal units [18,19]. The distribution of charge carrier jump dis­
weights of the bulk glassy sample immersed in acetone and suspended in
tances or the energy differences of electrons positioned at the unrelated
air, respectively. Average ρ was determined after measuring ρ of each
potential wells due to the presence of charged impurity atoms, in­
sample quite a few times. The value of ρ provides some basic structural
fluences the electrical conduction process [20]. Additionally, the vari­
features such as molar volume (VM) and the ionic packing ratio. The VM
ation in the electrical transport mechanism and the activation energy
of each composite was computed from the values of total molecular
necessary for the conductivity mechanism is instigated by the
weight (M) of the composition and ρ according to the following rela­
thermally-activated structural vibrations or due to differences in oxygen
tionship [22]:
coordination [20].
The purpose of this work is initially to synthesise four quaternary and X xi � MWi
VM ¼ (2)
two ternary glassy systems of composition 0.4ZnO–0.1P2O5–0.5 ρ
[xV2O5–(1-x) MoO3] by using the melt quench technique. Here, we
study the electrical transport mechanism of the quaternary systems where MWi is the ith oxide molecular weight, xi is the molar fraction of
using the property of MTIE due to the presence of two TMOs: V2O5 and the ith oxide, and ρ is the determined value of the average density of each
MoO3. Furthermore, one objective of this work is to explore the micro­ glass composition.
structural and optical absorption properties of all nanocomposite sys­
tems using X-ray diffraction (XRD) and ultraviolet–visible (UV–Vis) 2.3. Characterisation techniques
absorption techniques. A second objective is to investigate the probable
dc and ac electrical conduction mechanisms and to determine the For structural characterisation, an X-ray diffractometer (Rigaku
different activation energies required for the processes. Finally, we TTRAX-III) was utilised, and the chosen wavelength of CuKα X-ray

2
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

radiation was 1.5418 Å. The scanning speed was 4� per minute in steps 3. Results and discussion
of 0.02� , while Bragg’s angle (2θ) ranged within 10–80� . According to
the Gaussian distributions [23], the microstructural properties of the 3.1. Analysis of X-ray diffractograms
samples were inspected using the line profile analysis (LPA) method.
A PerkinElmer Lambda-750 UV–Vis spectrophotometer was used in Fig. 1a presents the X-ray diffractograms of ternary (two examples)
absorbance mode to obtain the optical absorption spectra of the nano­ and quaternary (four examples) finely powdered nanocomposite sam­
composites at 300 K in a wide range of wavelengths (190–700 nm). The ples. All samples clearly exhibit a definite amount of amorphous phase,
powdered samples were dissolved in ethyl alcohol to form colloidal which is the feature of short-range order. The amount of amorphous
suspensions for the absorption procedure and homogenised using an phase present within the samples can be calculated from the relative
ultrasonic sonicator (Takashi SK-500F). percent crystallinity of each nanophase. Relative percent crystallinity
can be determined by deconvolution of the amorphous state and crys­
2.4. Electrical conductivity measurement tallinity or by the ratio of the crystalline peak intensity to the sum of the
amorphous and crystalline peak intensities, mathematically expressed
For electrical measurements, highly conductive and high- by the following relationship [25]:
temperature resistant silver paste was placed on either side of the bulk
glassy flakes to function as the electrodes. The sample temperature Relative percent crystallinity ¼ Icrystalline/(Icrystalline þ Iamorphous) (4)
within the furnace was recorded by a thermocouple, located nearby the Furthermore, according to Ruland [26], the ratio of the total area
sample surface. The thermocouple was connected to a temperature under any crystalline peak to the total area under the total diffraction
control unit to measure the operating temperature of samples with �5% pattern (crystalline plus amorphous peaks) provides the value of relative
error. Alternating current (ac) electrical measurements were performed percent crystallinity as expressed below [27,28]:
using an automatic LCR Hi-Tester (Model No: 3532–50; Hioki) in the
frequency window of 42 to 5 MHz. The values of parallel capacitance Relative percent crystallinity ð%Þ ¼
Area ​ under ​ the ​ crystalline ​ peak
(CP), dielectric loss tangent (tan δ), and conductance (G) were recorded Total area under the diffraction curve
and the experimental value of total electrical conductivity (σtotal) was ðcrystalline þ amorphousÞ
calculated using the following relationship [24]: � 100
(5)
Conductance ​ ðG ​ in ​ siemensÞ � ​ thickness ​ ðt ​ in ​ cmÞ
σtotal ¼ � (3)
​ Area ​ A ​ in ​ cm2 The values of relative percent crystallinity of each nanophase are
shown in Table 1, illustrating the development of a few nanocrystallites
The accuracy of the temperature measurements was �3 K, and within the amorphous structure of the glass network or matrices.
conductance measurements had a maximum error of �2 μS. The relative percent crystallinity values are in agreement with
Fig. 1a, which also confirms the superposition of a few nanocrystallites
over the amorphous glass network irrespective of the composition
molarity, which is characteristic of long-range order. The nanophases of
each distinct diffraction angle peaks have been identified and indexed
through laborious evaluation from existing literature data. The

Fig. 1. (a) X-ray diffractograms of all bulk nanocomposites with variations in x and (b) the composition-dependent average crystallite size of all bulk samples with
error bars, and the dotted lines are guidelines.

3
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

Table 1
Values of diffraction peak angle (2θ), Corrected Full width at half maxima (FWHM), nanocrystallite size using Scherrer equation (dC), nanocrystallite phases and
corresponding Miller indices (h, k, and l), and percentage of crystallinity of 0.4ZnO–0.1P2O5–0.5 [xV2O5–(1-x) MoO3] glass nanocomposites. Ranges of errors are
indicated within the parentheses.
x (mol %) 2θ (degree) FWHM (βcorrect) dC (nm) Nanocrystallite phase h k l Percentage
(�0.03) (�0.0035) (�1.05) Crystallinity (%) (�0.25)

0.0 19.67 0.3839 21.00 ZnP4O11 1 2 1 1.41

22.81 0.1558 52.02 Zn [MoO4] 1 1 2 0.72
24.1 0.1532 53.02 Zn [MoO4] 0 1 2 2.72
26.35 0.1667 48.94 ZnP4O11 1 2 1 1.49
28.47 0.2603 31.49 MoP3O9 3 3 0 1.29
29.79 0.1633 50.34 Zn2[P2O7] 5 1 2 1.12
33.31 0.2151 38.55 Zn [MoO4] 0 1 3 0.82
33.85 0.3389 24.50 Zn [MoO4] 2 2 1 1.00

39.98 (Average) 10.57 (Total)

24.11 0.2201 36.91 Zn [MoO4] 0 1 2 0.38

25.39 0.2461 33.09 MoP3O9 3 3 2 0.40
0.2 26.54 0.3762 21.69 Zn2.5MoVO8 1 2 2 0.99
27.61 0.7308 11.19 MoV2O8 6 0 0 0.98
28.34 0.9946 8.23 MoV2O8 4 0 1 1.80

22.23 (Average) 4.55 (Total)

26.53 0.3715 21.97 Zn2.5MoVO8 1 2 2 0.18

0.4 27.61 1.0601 7.72 MoV2O8 6 0 0 1.20
28.82 0.3178 25.81 Zn3MoV2O11 0 0 2 0.17
18.50 (Average) 1.55 (Total)

28.13 0.6053 13.53 MoV2O8 3 1 0 0.79

0.6 28.81 0.4873 16.83 Zn3MoV2O11 0 0 2 0.26

15.18 (Total) 1.05 (Total)

0.8 28.46 0.3465 23.65 Zn3MoV2O11 2 1 1 2.69

28.81 0.3596 22.81 Zn3MoV2O11 0 0 2 1.19
34.86 0.5989 13.90 Zn3MoV2O11 3 1 0 0.96
37.89 0.3846 21.84 ZnMoO4 2 0 0 0.40

20.55 (Average) 5.24 (Total)

1.0 16.61 0.2564 31.31 V4[P2O7]3 4 0 0 0.96

19.86 0.2145 37.60 ZnP4O11 1 2 1 0.64
26.33 0.4118 19.81 ZnP4O11 1 2 1 1.55
28.8 0.2323 35.31 VP3O9 3 1 0 2.40
33.78 0.1796 46.23 Zn2[V2O7] 2 2 2 0.82
42.27 0.2271 37.50 [VO][PO4] 2 2 0 0.55
56.57 0.3608 25.00 V2O5 2 1 1 0.69

33.25 (Average) 7.61 (Total)

development of superimposed ZnP4O11 [29], ZnMoO4 [30], MoP3O9 nanocomposites (x ¼ 0.2 to 0.8) are seen in Fig. 1a and Table 1.
[31], Zn2P2O7 [32], Zn2.5MoVO8 (ICDD-50-1894), MoV2O8 [33], Although samples of x ¼ 0.0 (ZnO–P2O5–MoO3 ternary system) and x ¼
Zn3MoV2O11 [34], V4[P2O7]3 [35], VP3O9 [36], Zn2V2O7 [37], VOPO4 1.0 (ZnO–P2O5–V2O5 ternary system) reveal the presence of a definite
[38], and V2O5 [39] nanophases within the amorphous glassy networks amount of amorphous phase (Table 1), the number of developed nano­
or matrices have been predicted and confirmed. The LPA method is used crystallites exceeds those in the quaternary systems except for x ¼ 0.8,
to analyse the microstructural properties following Gaussian distribu­ and these samples are considered polycrystalline due to the absence of a
tions [40]. The broadening effect of the instrument is corrected by broad hump. For quaternary systems with values of x ¼ 0.2, 0.4, and 0.6,
considering the βinstrumental value of standard crystalline silicon. The a broad hump is visible around 2θ ¼ 27� , which confirms the total
corrected full width at half maximum (FWHM) (βcorrect) values of the amorphous structure of the samples with very few nanocrystallites. The
diffraction peaks are the result of subtraction of the observed broad­ evolution of total amorphous structure with the presence of a broad
ening, βobserved, and instrumental broadening, βinstrumental, effects. Thus, hump appears for x ¼ 0.2 and then increases progressively with rising
the value of the Gaussian distribution function, βcorrect, used in the cal­ V2O5 concentration up to x ¼ 0.6; for x ¼ 0.8, the broad hump disap­
culations can be expressed as follows [41]: pears due to major structural variations, thus representing a poly­
�0:5 crystalline sample. The presence of superposed nanocrystallites inside
βcorrect ¼ β2observed β2instrumental (6) the amorphous glass networks or matrices alters the structural features
Using the well-known Scherrer equation, the size of each nano­ and also electrical conductivity.
crystallite (dC) is determined by the following equation [41,42]: Fig. 1b illustrates average size of nanocrystallites with variation in
V2O5 concentration (x). For x ¼ 0.0, the average size of nanocrystallites
dC ¼
0:89 λ
(7) is highest and followed by a decrement in the average size of nano­
βcorrect cos θ crystallites with inclusion of V2O5 up to x ¼ 0.6 due to formation of the
total amorphous nature of the corresponding composites. Average dC
where λ is Cu-Kα X-ray radiation wavelength (1.54184 Å). The super­ then increases as samples (x ¼ 0.8 and 1.0) develop more crystallinity
posed nanophases with their peak θ, βcorrect, values of the equivalent than the pure amorphous state.
Miller index, and each dC are presented in Table 1. The distinct varia­
tions of the XRD patterns of ternary (x ¼ 0.0 and 1.0) and quaternary

4
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

3.2. Analysis of ρ and VM separation <dMo–Mo> [22,45]. For that, the volume of 1 mol of V (VVM)
and Mo (VMoM ) within the given structure can be computed using the
Fig. 2 presents the differences in ρ (Eq. (1)) and VM (Eq. (2)) values following expressions [22,45]:
with variations in V2O5 concentration (x) and shows the usual inverse
VM VM
relation of VM and ρ [22]. According to the composition formula, the VVM ¼ ​ and VMo
M ¼ (8)
2ð1 XV Þ ð1 X Mo Þ
molecular weight of each sample increases with increasing V2O5 con­
ion ion

centration (x) (Table 2) because of the higher molecular weight of V2O5 where XV-ion and XMo-ion are the molar fractions of V2O5 and MoO3,
(181.88 g mol 1) compared to MoO3 (143.94 g mol 1). Thus, incorpo­ respectively. The <dV–V> and <dMo–Mo> (Table 2) are evaluated using
ration of V2O5 in the glass matrix acts as a modifier, causing an incre­ the following relationships [22,45]:
ment in non-bridging oxygen ions (NBOs), which escalates the breaking
� �1=3 � �1=3
of covalent bonds within the network. For x ¼ 0.0 (ternary system of VVM VMo
composition ZnO–P2O5–MoO3), the value of ρ is higher than x ¼ 1.0 < dV V > ¼ and < dMo Mo > ¼ M
(9)
NA NA
(ternary system of composition ZnO–P2O5–V2O5) due to MoO6 poly­
hedra group formation [43]. where NA is Avogadro’s number (6.0228 � 1023 mol 1). The <dV–V>
The higher concentration of MoO3 typically reorganises the structure value increases and <dMo–Mo> value reduces progressively with incre­
of the glass network or matrix by creating additional oxygen ions, ment in V2O5 concentration (x) (Table 2). Therefore, due to the cumu­
resulting in the changeover of MoO4 to Mo2O7 or MoO6 structural units lative effect of inclusion of V2O5 strongly expanding the quaternary glass
[43]. During the development of quaternary glassy systems with the network, this directly supports the decline in ρ and the increasing trend
inclusion of V2O5, the ρ value clearly declines up to x ¼ 0.8 (Fig. 2), of VM for as-quenched nanocomposites. Such nature of ρ and VM may
revealing a loosely packed amorphous glass network, and ρ increases arise from the formation of excess NBO bonds, thus expanding the glass
again for x ¼ 1.0 (ZnO–P2O5–V2O5, ternary system). The decrement of ρ network structure [22,45]. Using the concept of chemical bond
value with V2O5 concentration rise is caused by a number of factors, approach in explaining VM and glass ρ behaviour, the oxygen molar
firstly, introduced V ions reside in interstitial sites, thus reducing the volume (VO) and oxygen-packing density (OPD) of each nanocomposite
vacancies inside the glass matrix [44]. Furthermore, the conversion of are computed using the following ratios [45]:
six- to four-fold coordination states of V and Mo ions also reduces the ρ � ρ �
VM
value [44]. For x ¼ 1.0 (ZnO–P2O5–V2O5, ternary system), the ρ value VO ¼ P and OPD ¼ 1000 C (10)
x ​ O M
increases considerably due to a slight decrease in volume due to the n n n

formation of different vanadate structural groups, and the nonexistence

where On is the number of oxygen atoms existing in each oxide, C is the
of molybdate structural groups in the composition. The change in ρ
number of oxygen atoms per formula unit, and M is molecular weight.
value with increment in V2O5 content supports the transformation of
The calculated values of VO and OPD of all glass composites are recorded
V2O5 from the glass network modifier to network former for x ¼ 0.2–1.0.
in Table 2. The VO reduces and OPD increases with V2O5 concentration
The increase in VM results from the development of NBOs with addition
rise, thus confirming NBO bond formations with the replacement of
of V ions, which indicates that V ions either act as modifiers in glass
MoO3 by V2O5 within the glass structure [45]. With V2O5 addition, the
matrix or cause field strength variation [44].
progressive transformation of V ions from V4þ to V5þ state occurs and
Conversely, VM is directly related to the internal spatial structure of
V5þ, in this case, modifies the glass structure by collapsing local sym­
materials [14]. Fig. 2 displays the change of estimated VM values as a
metry and producing coordinated bond defects, which also enhance
function of the V2O5 molar ratio (x). The increasing nature of VM also
electrical conductivity. The total concentration of V and Mo ions can be
assists in the increment of bond length, causing less compaction of the
determined using the following expression [22]:
glass structures. The lower compaction of quaternary glass structures
due to V2O5 inclusion or MoO3 exclusion can be confirmed by deter­ mol ​ % ​ of ​ oxide � NA � ρ
N ​ V=Mo ¼ (11)
mining the average values of V–V separation <dV–V> and Mo–Mo
ions
Molecular ​ weight ​ of ​ composition ðMÞ
It is possible to predict the inter-nuclear spacing between V4þ and
V (RV) and Mo5þ and Mo6þ (RMo) separately by utilising the estimated
5þ

value of NV/Mo-ions (Eq. (11)) as follows [22]:

� �1=3
1
R ​ V=Mo ¼ (12)
NV=Mo ions

The estimated values of NV-ions and NMo-ions (Table 2) reveal the

anticipated results that V ion concentration rises and Mo ion concen­
tration declines with inclusion of V2O5. Variations in the inter-nuclear
spacing of RV and RMo with the variation in V2O5 concentration are
shown in Fig. 3.
There is an increase in RMo as MoO3 concentration (NMo-ions) de­
creases, whereas RV declines with rising V2O5 concentration, initiating
the small polaron formation, thus enhancing conductivity.

3.3. Determination of optical bandgap energy

The absorption spectrum in pure crystalline semiconductors usually

ceases suddenly at the energy gap. Conversely, in materials consisting of
crystalline and amorphous phases, a band tail state intrudes into the
bandgap region in the absorption spectrum, as a consequence of the
Fig. 2. Variation of density (ρ) and molar volume (VM) of all the nano­ presence of disorder or defect and localised states [46]. The absorption
composite samples. The vertical lines are error bars, and dotted lines coefficient (α) of UV–Vis electromagnetic waves due to the inter-band
are guidelines. transition near the bandgap of the glassy systems has been noted and

5
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

Table 2
Molecular weight (M) of each composition, vanadium volume (VVM), average vanadium–vanadium separation <dV–V>, molybdenum volume (VMo M ), average molyb­
denum–molybdenum separation <dMo–Mo>, molar volume of oxygen (VO), oxygen-packing density (OPD), V ion concentration (NV-ions), and Mo ion concentration
(NMo-ions) of all the glass samples for different values of x. For the calculation, we considered the valence number of V as five and Mo as six. Ranges of errors are
indicated within the parentheses.
x (mol Molecular Weight VVM (g <dV–V> VMo
M (g <dMo–Mo> VO (cm3 OPD (mol NV-ions ( � 1022 ions/ NMo-ions ( � 1022 ions/
%) (M) cm–3) (Å) cm–3) (Å) mol 1) l 1) cm3) cm3)
(g mol 1) (�0.20) (�0.04) (�0.20) (�0.04) (�0.20) (�0.05) (�0.25) (�0.25)
(�0.02)

0.0 127.53 – – 68.52 4.85 14.28 70.05 – 1.56

0.2 131.32 19.70 3.20 59.10 4.61 13.64 73.32 0.25 1.24
0.4 135.11 23.34 3.38 53.35 4.46 13.34 74.98 0.48 0.91
0.6 138.91 28.56 3.62 49.98 4.36 13.33 75.03 0.69 0.58
0.8 142.70 34.80 3.87 46.40 4.26 13.05 76.62 0.91 0.29
1.0 146.50 41.16 4.09 – – 12.11 82.60 1.18 –

Fig. 3. Variation of inter-nuclear spacing of V (RV) and Mo (RMo) ions of all

nanocomposite samples with different values of x. The vertical lines are error
bars and dotted lines are guidelines.

carefully analysed to obtain the optical bandgap energy (Eopt) by using

Tauc’s plot technique [47,48]. In general, the Eopt value can be predicted
via the following relationship [47,48]:
� ��p
αhν ¼ ​ C ​ hν Eopt (13)

where C denotes the band-tailing parameter and hν represents the en­

ergy of the UV–Vis photon. The exponent (p) value depends on the
electronic nature of the bandgap, i.e. directly allowed, indirect allowed,
direct forbidden, and indirect forbidden electronic transitions are rep­
resented by p values of 1/2, 2, 3/2, and 3, respectively. Fig. 4a illustrates Fig. 4. (a) Tauc’s plots of all samples, dotted lines are the extrapolation of the
Tauc’s plots of bulk composites assuming indirect allowed transition (p linear region and (b) the plot of optical bandgap energy (Eopt) against the
¼ 2), i.e. the plot of (αhν)0.5 against hν. To compute the values of Eopt, the concentration of V2O5 (x). The vertical lines are error bars and dotted lines
linear section of the Tauc’s plot is elongated to reach (αhν)0.5 ¼ 0 on the are guidelines.
X-axis (Fig. 4a). For x ¼ 0.8 and 1.0, another absorption peak is observed
due to formation of nanocrystallites, superposing over the amorphous (Fig. 1 and Table 1). Due to formation of more amorphous phase up to x
phase (Fig. 4a). For a complicated system such as the present one, we ¼ 0.6, the average dC declines, causing development of localised or
have to consider the linear region of the first absorption peak because extended states in which electrons are confined within the surface or
extrapolation of the linear region of the second peak does not intersect at potential well and this confinement increases Eopt. For x ¼ 0.8 and 1.0,
(αhν)0.5 ¼ 0 on the X-axis. The non-linear nature of composition- the addition of V2O5 opens up the glass network by breaking oxygen
dependent Eopt, i.e. the variation of Eopt with different V2O5 molar ra­ bonds, therefore, the amount of NBO grows, resulting in changes in the
tios (x), is shown in Fig. 4b. The error of Eopt value is assumed to be 10% absorption characteristics. The alteration in NBOs and the growth of
of the estimated Eopt value. The Eopt value increases with addition of new nanocrystallites reduces the Eopt as the number of localised or
V2O5 (x) up to x ¼ 0.6, after which it decreases (Fig. 4b). extended states reduces.
Comparison of Figs. 1b and 4b shows the inverse relationship be­ Davis and Mott described that in low crystalline or amorphous solids,
tween Eopt and average dC. Thus, Eopt values become higher as average the Eopt usually relates to the density of defects or localised states [49].
dC decreases or for samples with a higher degree of amorphous phase They also proposed, within the amorphous glass structure, the existence

6
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

of a number of defect states and disorder forms localised or extended 0.0, 0.4, 0.8, and 1.0 are shown in Fig. 6a–d, respectively. Qualitatively
states inside the optical bandgap or mobility gap. In the non-linearity all the other samples also exhibit similar σtotal behaviour. The σtotal of all
plot of the absorption coefficient (ln α), near the optical bandgap bulk nanocomposites increases non-linearly with equal increment in
edge, an exponentially decaying band tail is detected that is expressed as temperature, confirming semiconducting behaviour. Individually iso­
Urbach energy (EU) or band tail energy. The EU value can be estimated therms illustrate two typical features, i.e. a plateau or a flat dc con­
using the following Urbach empirical rule [22,50]: ductivity (σdc) region, which is nearly frequency independent owing to
� � the effect of space charge polarisation and the conduction process is the
hν
α ¼ α0 ​ exp (14) consequence of small polarons or charge carrier hopping mechanism.
EU
The second feature is the dispersion that varies almost linearly at the
higher frequencies. The range of frequency-independent conductivity
where α0 is a constant term when (hν/EU) → 0. Generally, EU refers to the
increases to a higher frequency with temperature rise and dispersion
amount of defect, extended, or localised states present in the vicinity of
regions nearly coincide with each other at higher frequencies due to the
Eopt [51]. The value of the reciprocal of the slope of the linear region of
mixed electronic and ionic conductivity nature of the systems (Fig. 6a
the plot of ln (α) versus hν provides the EU value, as demonstrated by
and b). In contrast, for x ¼ 0.8 and 1.0, the range of frequency inde­
solid lines in Fig. 5a. The composition-dependent values of EU with the
pendent and dispersive regions is almost the same for the entire tem­
variation of V2O5 molar ratio (x) are displayed in Fig. 5b. The EU value
perature range (Fig. 6c and d), i.e. the crossover frequency does not alter
drops up to x ¼ 0.6 and then rises again for x ¼ 0.8 and 1.0. Thus, it can
much, which confirms that conductivity is electronic or polaronic in
be concluded that for the present nanocomposite systems, the EU values
nature. The dispersal behaviour of frequency-dependent conductivity
decline as the degree of amorphous state rises. Comparative study of Eopt
originates either due to tunnelling or hopping mechanisms of small
and EU, as shown in the inset of Fig. 5b, reveals the inverse relationship
polarons or mobile charge carriers through or among the defect or
between Eopt and EU, i.e. as Eopt increases, EU falls, and dC also declines
localised states existing within the range of Eopt or owing to the ionic
(Fig. 1a). The XRD data show the existence of low crystallinity or
motion of mobile ions inside the glassy matrices [52,53]. The unified
amorphous nature of all the nanocomposites (Fig. 1a), which is
forward and backward motion of charge carriers or small polarons
confirmed by the optical absorbance spectra.
during hopping or tunnelling mechanisms causes the dispersive nature
of σtotal spectra. During the hopping process, small polarons or charge
3.4. Study of electrical conductivity carriers generally hop or jump from one localised or defect state to
another, which requires less energy than Eopt, thus resulting in the
Characteristic plots of total ac conductivity (σtotal) isotherms for x ¼ reduction of activation energy with a rise in frequency. All σtotal plots
(Fig. 6) present characteristics of the electronic or small polaronic
transport mechanism between sites with randomly separated space and
energy. The σtotal plots disclose that σtotal is the sum of both ac con­
ductivity (σac (ω)), which is frequency-dependent but weakly
temperature-dependent and the temperature-dependent and
frequency-independent σdc. The σtotal can be examined by the following
Almond–West formalism [52,53]:
� �
ω n
σtotal ¼ σdc 1 þ (15)
ωH

where ωH specifies the hopping or crossover frequency of small polarons

or charge carriers, i.e. the representative of the crossover from dc to
dispersive conductivity at ω ¼ ωH. In addition, n represents the fractional
exponent, which decides the charge transport type and indicates the
ionic/non-ionic boundary and the dimension of the conduction process
[54]. The value of n depends on material and temperature [52–54]. For
n < 1, the charge carriers or small polarons are assumed to take a
translational motion with sudden hopping, whereas n > 1 suggests a
localised hopping of the charge carriers or small polarons without
leaving the neighbourhood. Solid black lines in Fig. 6 disclose the
non-linear best-fit curves using Almond–West formalism (Eq. (15)), and
after fitting with the investigational measured data, the values of fitting
parameters, i.e. σdc, n, and ωH are obtained. The as-obtained average
values of frequency or fractional exponent (n) of all the nanocomposite
samples are given in Table 3.
The relative study of the σtotal spectra of all bulk glass composites at a
definite temperature (653 K) clarifies that the inclusion of V2O5 (x) in
quaternary systems increases the σtotal value (Fig. 7). Comparative study
of σtotal suggests that x ¼ 0.0 (ternary system of composition
ZnO–P2O5–MoO3) has lower values of σtotal than x ¼ 1.0 (ternary system
of composition ZnO–P2O5–V2O5). Thus, V2O5 acts more effectively as a
glass network modifier compared to MoO3. Alternatively, the σtotal
values of the quaternary system of x ¼ 0.8 show the highest electrical
conductivity, which is also higher than for x ¼ 1.0 (Fig. 7). Therefore, it
Fig. 5. (a) Dependence of ln (α) of all samples on incident photon energy (hν), can be concluded that the MTIE with a proper stoichiometry signifi­
with solid black lines the best-fitted linear fit data and (b) variation of Urbach cantly enhances total electrical conductivity (σtotal). The present results
energy (EU) with different values of x; vertical lines are error bars and dotted also disclose that conductivity does not increase immediately with an
lines are guidelines.

7
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

Fig. 6. (a) Temperature- and frequency-dependent total ac conductivity (σtotal) spectra of the glass compositions for (a) x ¼ 0.0, (b) x ¼ 0.4, (c) x ¼ 0.8, and (d) x ¼
1.0. Black solid lines indicate best non-linear curve fitting data using Almond–West formalism.

Table 3
The average values of frequency exponent (n), composition-dependent values of ac conductivity activation energy (Eac), dc conductivity activation energy (Edc), the
radius of small polarons and density of defect states at the high-temperature region with errors within the parentheses.
x (mol %) Average value of n Eac (eV) Dc conductivity activation energy (Edc) (eV) Small polaron radius Density of defect states at high temperature
(�0.05) N(EF) ( � 1021 eV 1cm 3)

Elow (eV) Ehigh (eV) RP-V (Å) RP-Mo (Å) N(EF) of V ions N(EF) of Mo ions
(�0.02) (�0.02) (�0.02) (�0.02)

0.0 1.612 0.84 0.95 1.91 – 1.61 – 3.96

0.2 1.125 0.70 0.69 2.69 2.99 1.74 0.85 4.29
0.4 1.519 0.42 0.67 1.88 2.39 1.93 1.73 3.29
0.6 0.961 0.38 0.55 1.62 2.12 2.24 3.05 2.57
0.8 1.112 0.23 0.39 0.96 1.93 2.84 5.54 1.75
1.0 0.951 0.36 0.52 1.44 1.77 – 5.41 –

increment in V2O5 concentration, but a drop in the rate of crossover or independent of frequency.
hopping frequency is also detected as observed for x ¼ 0.2 (Fig. 7). Hence, it can be asserted that the reciprocal temperature response of
Using the fitting of Almond–West formalism with the experimental σac directs that the conduction process is due to space charge conduc­
data, we obtain σdc and σac is calculated as follows: σac ¼ σtotal – σdc. The tion. The calculation of ac conductivity activation energy (Eac) for each
σac dependency with the variation of frequency of reciprocal tempera­ nanocomposite at various frequencies is carried out using the following
ture for x ¼ 0.4 reveals that σac increases consistently with an increment empirical relationship [55,56]:
in frequency and temperature (Fig. 8a). The σac usually depends on the � �
Eac
structural symmetry, grain and crystallite size, and distribution of grains σac ¼ σ0 exp (16)
KB ​ T
within the glassy network [55]. In bulk polycrystalline or amorphous
materials, the σac mechanism is either due to the long-range hopping
where KB is Boltzmann constant (¼ 8.617 � 10 5 eV K 1) and σ0 is the
mechanism of charge carriers or small polarons or to the localised
value of ac conductivity at very high temperature when (Eac/KBT) → 0.
transportation resulting from oxygen vacancies [55]. The slopes of the
The Eac can be determined by the slope of the linear relationship, as
Arrhenius plot reduce with increasing frequency, and σac increases with
revealed by solid lines in Fig. 8a and b and Eq. (16). The estimated Eac
a rise in frequency, indicating the dispersive response of σac (Fig. 8a).
values are shown in Fig. 8a for different frequencies and in Table 3 for
This is the result of the possible release of space charges at higher fre­
different composites of varying values of molar ratio (x). The Eac value
quencies, even at lower temperatures. At a higher temperature for
reduces as frequency rises (Fig. 8a), i.e. less energy is required to counter
different frequencies, σac does not differ much because at a higher
the oscillations of excited charge carriers at high frequencies due to
temperature the space charges release and recombine almost at the same
improved jumping or hopping of electrons between neighbouring
rate, therefore, the σac of the composite samples becomes almost
localised states [56]. Frequency-dependent Eac values show that for

8
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

(0.3 eV) [57]. A comparison of Eac for different compositions with the
variation in the molar ratio of V2O5 concentration (x) at a specific fre­
quency (3 kHz) is presented in Fig. 8b. The Eac values of different
composites are presented in Table 3, indicating the inverse relationship
of conductivity and Eac. The Eac value for x ¼ 0.0 and 0.2 is nearby the
activation energy of the second ionisation of oxygen vacancies (0.7 eV;
Table 3) [57]. Relatively, Eac values for x ¼ 0.4–1.0 are in the range of
the first ionisation of oxygen vacancies (0.3 eV) [57]. With increasing
V2O5 concentration, Eac value falls as the inclusion of V2O5 induces an
increment in the number of glass network defects as the amorphous
phase of the samples increases (Fig. 1a) [55]. The diminishing nature of
Eac can also be ascribed to electrons trapped on the defects [55].
The variations of crossover or hopping frequency (ωH) of charge
carriers or small polarons with the reciprocal of temperature, which
obeys the Arrhenius equation, are shown in Fig. 9. The values of the
small polaron migration or crossover frequency activation energy (EH)
can be realised using the least-squares straight-line fit of the data as
shown by the solid lines in Fig. 9 using the following Arrhenius equation
[58]:
� �
Fig. 7. Total ac conductivity (σtotal) spectra of all nanocomposite samples with ωH ¼ ωe exp
EH
(17)
the variation of V2O5 molar ratio (x) at 653 K. KB ​ T

where ωe is the effective attempt hopping frequency and T is the absolute

temperature in Kelvin. The EH is the minimum energy essential for small
polarons hopping or tunnelling among or through localised or defect
sites alienated by a potential barrier from each other. With an increment
in σtotal, the EH value falls (Fig. 9). Thus, it is emphasised that due to rise
in temperature, the small polarons or charge carriers require lesser en­
ergy as the high temperature starts oscillation of small polarons or
charge carriers, which prompts small polarons or charge carriers to hop
or tunnel among or through the defect or localised states to another. For
x ¼ 0.8, the variation of ωH is lower (Fig. 9) compared to other samples
or ωH is almost independent of temperature due to the MTIE, owing to
the formation of Zn3MoV2O11 nanocrystallite. In amorphous or poly­
crystalline solids like the present case, the random lattice potential in­
volves Anderson-like localisation [13]. The randomisation occurs
substantially owing to the dissimilar atomic potential of V4þ/V5þ and
Mo5þ/Mo6þ ions in as-formed Zn3MoV2O11 and Zn2.5MoVO8 nano­
crystallites, which provides the MTIE. In such a random atomic potential

Fig. 8. Variations of ac conductivity (σac) with reciprocal of temperature (a) at

different frequencies for x ¼ 0.4 and (b) at a fixed frequency (3 kHz) of all
samples. Solid red lines are the linear fit data, and vertical lines are error bars.
(For interpretation of the references to colour in this figure legend, the reader is
referred to the Web version of this article.)

frequency up to 1 MHz, Eac > 1 (Fig. 8a), which describes the σac
mechanism on account of long-range hopping of electrons or small po­ Fig. 9. Plot of the reciprocal temperature dependence of hopping or crossover
larons. Whereas, in the high frequency (�3 MHz) region, Eac value is frequency of all nanocomposites. Solid lines represent Arrhenius linear fits with
nearby the activation energy of the first ionisation of oxygen vacancies calculated activation energy values.

9
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

well, electrons are trapped, forming localised sites, and electrons or governed by the hopping process between similar TMI sites (between
small polarons can hop between different valence states of similar or V4þ and V5þ and between Mo5þ and Mo6þ) as well as unlike TMI sites,
dissimilar TMIs. such as between V5þ and Mo6þ and between V4þ and Mo5þ [14–16]. The
The σdc data, obtained from Almond–West formalism, strongly maximum conductivity for x ¼ 0.8 is due to the diminishing effect of
depend on the temperature; and the non-linearity of σdc is due to the Anderson localisation instigated by an increment in disorder states when
involvement of dissimilar conduction mechanisms [59,60]. The two TMIs are mixed, promoting electronic or polaronic motion, thus
temperature-dependent σdc plot is demonstrated in Fig. 10, and σdc can causing an enhanced σdc. Additionally, the formation of Zn3MoV2O11
be expressed by the Arrhenius equation [59,60]: nanocrystallites greatly raises σdc owing to a higher degree of MTIE.
� � � � Significant lattice distortion is a criterion for SPH to occur, and this
E ​ high E ​ low
σdc ¼ σ ​ high exp þ σ ​ low exp (18) process is usually a multi-phonon hopping process. In TMO-doped bulk
KB T KB T
glass, the SPH model describes the σdc using the following relationship
Two terms appear in Eq. (18) due to the two different σdc mecha­ [61,62]:
nisms in dissimilar temperature regions [60]. In Eq. (18), the first term � � � �
ν e2 W
defines σdc in the high-temperature region due to the band conduction σdc ¼ ​ C ​ ð1 CÞexpð 2αRÞexp (19)
KB T R KB ​ T
process through the extended states, developed inside the mobility gap
as a result of the amorphous nature of the nanocomposites. The second
where W is activation energy, ν is optical phonon frequency, T is abso­
term of Eq. (18) refers to the conduction process arising due to the
lute temperature (in Kelvin), and c is the redox ratio of TMIs [(reduced
hopping of small polarons or electrons via localised or extended or
TMI)/(total TMI)]. Here R implies the hopping distance concerning the
defect states at a lower temperature range or due to the tunnelling
available localised centres, and α represents the wave function of small
mechanism between unoccupied sites of the nearest neighbouring sites.
polarons at a localised centre decay as exp ( αR). The small polaron
The pre-exponential factors σhigh and σlow are strongly composition
radius (RP) can be predicted, knowing the value of the inter-nuclear
dependent [59,60].
distance of V and Mo ions using the following formula [22]:
However, values of activation energies associated with dc conduc­ � �
tivity at high (Ehigh) and low (Elow) temperature regions are obtained RV = ​ RMo �π�1=3
RPðV=MoÞ ¼ (20)
from the gradient of the fitted lines indicated by the solid lines in Fig. 10. 2 6
The estimated values of Ehigh and Elow are given in Table 3. The Elow
values decrease with rising V2O5 concentration, and the lowest value of with a rise in V2O5 concentration, the estimated value of the small
Elow is for x ¼ 0.8, which reveals the highest conductivity (Fig. 10). The polaron radius of V ions (RP-V) reduces and the small polaron radius of
Ehigh value reduces as the atomic bonding disorder among the adjacent Mo ions (RP-Mo) increases (Table 3). According to the concept of small
sites increases, which enhances the density of tail and extended states polaron radius, σdc increases as the radii of small polarons reduces,
within the region of Eopt [59,60]. With the addition of V2O5 in the therefore, it can be concluded that V ions or V2O5 act as glass network
ternary system of the composition ZnO–P2O5–MoO3, the Ehigh value modifiers, and assist in enhancing conductivity for our present systems.
increases for x ¼ 0.2 due to substantial structural changes or configu­ The Fermi level defect states density [N(EF)] can be anticipated by
ration disorder, materialising because of the mixing of two TMIs [11] applying the assessed values of RMo and RV using the following expres­
and the development of new crystallites over the amorphous matrix. sion [22]:
Initially, for x ¼ 0.2, the σdc is not much altered at the low temperature 3
region, but the slope changes noticeably for higher temperatures NðEF Þ ¼ (21)
4 π ðRMo =RV Þ3 ​ Ehigh
(Fig. 10) owing to the interference introduced by Mo and V ions to
electronic motion. With further increasing V2O5 content, σdc increases The approximated and reasonable N(EF) values for localised states in
due to the MTIE, where the small polaron transport mechanism is the high-temperature region are noted in Table 3. With rising V2O5
concentration (x), the N(EF) of V ions advances more than for Mo ions,
which affirms the glass network modifier function of V2O5. The N(EF)
produces higher quantities of localised states inside the optical bandgap
or mobility gap, supporting the σdc hopping or diffusion process.
The frequency-dependent ac conductivity dispersion region is usu­
ally examined using the Jonscher universal power-law equation as fol­
lows [54]:
σðωÞ ¼ A ​ ωs (22)

where A is a temperature-independent or -dependent constant, which

can determine the polarisability strength [56], and s is the power-law
exponent, describing the degree of interactions of small polarons or
electrons with the neighbouring nanocrystallites or amorphous
matrices. The value of s typically determines the nature of the charge
transport mechanism. For s < 1, the charge carriers usually have a
translational motion or hopping conduction over the potential barrier
concerning two sites [63,64]. According to Gilroy et al. [65], s > 1 is
generally ascribed to the charge carriers’ motion from a defect or
localised site to another with a quantum mechanical tunnelling process
among asymmetric double-well potentials. In Fig. 6, for each isotherm
curve over the frequency range, where linear behaviour is observed,
linear least-square lines are fitted to the investigational data. The slopes
of the linear least-square fit lines directly provide s values. Fig. 11 il­
lustrates the variations of s as functions of varying temperatures for both
Fig. 10. The dc conductivity (σdc) variations against the reciprocal temperature ternary and quaternary bulk nanocomposites. The value of s depends on
of all composites; solid lines are linear fit data.

10
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

the critical energy required for the transportation of localised small

polarons trapped in potential wells of nanocrystallites to the conduction
band [22,67]. For our present complex nanocomposite system, as
highest s values exceed 0.8 (Fig. 11a), thus we use the modified rather
than the existing CBH model. The modified CBH model is defined as
below [22]:
6 ​ KB ðT T0 Þ
s¼ 1 (24)
WM þ ​ KB ðT T0 Þ ln ðωτ0 Þ
In Fig. 11a, solid lines illustrate the non-linear fitting of investiga­
tional data with the modified CBH model (Eq. (24)). The judicious fitting
parameter values, WM, τ0, and T0, are introduced in Table 4, showing
that values of each fitting parameter fall with rising V2O5 concentration
(x). Therefore, ac conductivity rises according to the CBH model as an
outcome of the diminishing values of τ0 and WM.
The NSPT mechanism describes the relationship between s and
characteristic relaxation time (τH) mathematically as expressed in the
following equation [64,68]:
4
s¼1 � � � � (25)
1 WH
ln ω τH KB T

where WH is the activation energy required for transportation of small

polarons. Eq. (25) implies that with an increment in temperature, the s
value rises. The s value is greater than unity (s > 1; Fig. 11b), hence we
apply the modified NSPT model to realise reasonable parameter values
using the experimental data with the following s–T relationship [58]:
4
s¼1 � � � � (26)
1 WH
ln ω τH KB ðT T0 Þ

The modified NSPT model parameters, WH, τH, and T0, are obtained
Fig. 11. Variation of power-law exponent (s) as a function of temperature (T) by non-linear curve fit of the s–T plot with the modified NSPT model (Eq.
of all nanocomposites with different values of x. Solid lines are the non-linear (26)) at a constant f ¼ 1 kHz; the as-obtained values of WH, τH, and T0
best-fit data of (a) modified CBH model and (b) modified NSPT model. drop with an increase in V2O5 concentration (x) (Table 4). Therefore, the
increment in ac conductivity according to the NSPT model is the effect of
temperature (Fig. 11), which efficiently shows the most appropriate the reduction in values of WH and τH.
model for the ac conduction mechanism. Scaling property is used to identify the dependence of the conduc­
The values of s decrease with rising temperature for x ¼ 0.2, 0.4, 0.6, tivity relaxation process on temperature and frequency as the scaled ac
and 1.0 (Fig. 11a), indicative that a correlated barrier-hopping (CBH) conductivity spectra merge into a solitary master curve [69,70]. The
mechanism is relevant for ac conductivity [66,67]. The s values increase frequency (X-axis) and ac conductivity (Y-axis) of each ac conductivity
with temperature increment (Fig. 11b), suggesting that a spectra are scaled or normalised by dividing by the values of ωH and σdc
non-overlapping small polaron tunnelling (NSPT) process is applicable individually [69,70]. The scaled spectra of a particular composition at
for ac conductivity for x ¼ 0.0 and 0.8 [63,68]. different temperatures are called temperature scaling spectra, as
In the CBH process, it is presumed that small polarons or charge depicted in Fig. 12a for x ¼ 0.8. For the present complicated
carriers do not interact with each other and are located in the crystallite
potential well, permitting nearest-neighbour jumps only [66,67]. The
small polaron jump or hopping probabilities with time can be presumed Table 4
as symmetric or asymmetric, i.e. back and forth hopping or jumping All values of fitting parameters of modified CBH model (for x ¼ 0.2, 0.4, 0.6, and
processes among the localised or defect sites or among localised states to 1.0) and the modified NSPT model (for x ¼ 0.0 and 0.8). Errors are indicated in
defect states. Due to the effect of thermal energy, the vibrational fre­ parentheses.
quency of electrons or small polarons at different well-defined sites Modified CBH model Modified NSPT model
enhances, leading to escalation of the SPH process. The localised or x (mol
WM (eV) τ0 (s) T0 WH (eV) τH (s) T0
%)
defect states also account for the ac conductivity process in response to (�0.004) (�0.05 � (K) (�0.014) (�0.05 � (K)
their closer proximity, in that way providing the distribution of different 10 –3) (�2) 10 –3) (�2)
heights of energy barrier separating them. It is predicted that separation 0.0 – – – 0.185 2.04 � 632
between two centres of a localised or defect state is higher than the 10 6
radius of the small polaron wave function of each site. According to the 0.2 0.089 1.38 � 604 – – –
10 3
CBH model, the relation of different parameters such as s, Debye
0.4 0.053 3.21 � 564 – – –
relaxation time (τ0), and temperature (T) is described by the following 10 3
expression [66,67]: 0.6 0.044 9.11 � 532 – – –
10 4
6 KB T 0.8 0.113 3.68 � 484
s¼ 1 (23) – – –
WM þ KB T ln ​ ðωτ0 Þ 10 9
1.0 0.034 2.01 � 440 – – –
With a rise in temperature, the s value drops in Eq. (23), where WM is 10 5

11
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

bandgap caused EU values to drop drastically, thus instigating the SPH

mechanism. The analysis of total ac conductivity spectra disclosed
frequency-independent (σdc) and -dependent (σac) regions and in both
cases, conductivity escalated with an increment in temperature owing to
the variations of the bond structures of the glassy systems and the for­
mation of defect, localised, or extended states, thus exhibiting semi­
conductor behaviour. For x ¼ 0.8, the conductivity was the highest
because of the cumulative effect of two TMIs, where small polarons hop
among similar (from V4þ to V5þ or Mo5þ to Mo6þ) and dissimilar TMIs
(from V5þ to Mo6þ or V4þ to Mo5þ), especially within Zn3MoV2O11
nanocrystallites. The σac increased due to reduction of the Eac and EH
values. The CBH method was the most suited σac mechanism for x ¼ 0.2,
0.4, 0.6, and 1.0, because the s value fell as temperature rose. The NSPT
mechanism was held responsible for the σac mechanism for x ¼ 0.0 and
0.8 since the s values rose with incremental temperature. Modified CBH
and NSPT models were adopted to compute different fitting parameters
and ideal thermodynamic glass transition temperature values for all
samples. Scaled σac spectra established that the conductivity relaxation
process depended on the structural features of nanocomposites, but was
temperature independent. Overall, it can be concluded that the con­
ductivity of the ternary system 0.4ZnO–0.1P2O5–0.5V2O5 was better
than that of the 0.4ZnO–0.1P2O5–0.5MoO3 ternary system. However,
the intensifying effect of two TMIs contributed to the highest conduc­
tivity of the quaternary 0.4ZnO–0.1P2O5–0.5(xV2O5–(1-x) MoO3) sys­
tem for x ¼ 0.8. Therefore, it can also be concluded that V2O5 had a role
as a more improved glass network modifier than MoO3, and the present
semiconducting system can be used for different types of sensors even at
higher temperatures.

Data availability

The raw/processed data required to reproduce these findings cannot

be shared at this time as they also form part of an ongoing study.

Fig. 12. Normalised plot of (a) temperature scaling spectra of ac conductivity Declaration of competing interest
for x ¼ 0.8 and (b) composition scaling spectra of ac conductivity at 693 K for
all nanocomposite samples with the variation in V2O5 concentration (x). The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
(quaternary) glassy systems, temperature scaling spectra comply with the work reported in this paper.
the time-temperature superposition (TTS) principle because all the
scaled spectra at different temperatures overlap almost perfectly [69, CRediT authorship contribution statement
71]. A wide distribution of electron trapping sites forms due to the ex­
istence of structural disorder, with unlike energy barrier heights, Dipankar Biswas: Investigation, Methodology, Software, Valida­
creating a wide-ranging distribution of back and forth hopping or tion. Anindya Sundar Das: Conceptualization, Visualization, Supervi­
tunnelling rates that maintain the TTS principle [71]. The scaling pro­ sion, Writing - original draft, Writing - review & editing. Rittwick
cess for all the composites at a precise temperature is designated as the Mondal: Validation, Formal analysis. Anindita Banerjee: Software,
composition scaling spectra (Fig. 12b), which indicates that composition Data curation. Abhigyan Dutta: Resources, Supervision. Soumyajyoti
scaling spectra are not suitably merged into a single unique curve. Kabi: Resources, Funding acquisition. Debasish Roy: Resources. Loi­
Conductivity scaling phenomena (i.e. temperature scaling and compo­ tongbam Surajkumar Singh: Project administration.
sition scaling) reveal that the conductivity relaxation process of mobile
charge carriers or small polarons does not change with variations of Acknowledgements
temperature but relies on the structural features of each composition, i.
e. on the molar ratio of V2O5 concentration (x). The authors wish to acknowledge Mr. Anish Karmahapatra and Ms.
Soma Roy of SINP, Kolkata, for their assistance in producing the XRD
4. Summary and conclusions patterns and UV–Vis absorption spectra, respectively. Dr. Abhigyan
Dutta thankfully acknowledges the financial support of SERB (Govt. of
Two ternary and four quaternary nanocomposite samples of India) through Grant no. EMR/2017/000325.
0.4ZnO–0.1P2O5–0.5[xV2O5–(1-x) MoO3] composition, where x varies
in steps of 0.2, were synthesised via the melt quenching process. An Appendix A. Supplementary data
inverse proportional relationship between ρ and VM was recognised. The
XRD analysis confirmed the development of ZnP4O11, ZnMoO4, MoP3O9, Supplementary data to this article can be found online at https://doi.
Zn2P2O7, Zn2.5MoVO8, MoV2O8, Zn3MoV2O11, V4[P2O7]3, VP3O9, org/10.1016/j.jpcs.2020.109505.
Zn2V2O7, VOPO4, and V2O5 nanocrystallites, superposing within the
amorphous glass networks. The dC and Eopt were in an inverse rela­
tionship, whereas EU values were consistent with nanocrystallite size.
The formation of additional defect or localised states within the optical

12
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

References [33] P. Pailleret, J. Borensztajn, W. Freundlich, A. Rimsky, Structure crystalline de la

phase MoV2O8, C.R. Seances Acad. Sci. Ser. C. 263 (1966) 1131–1133.
[34] M. Kurzawa, M. Bosacka, Novel compound Zn3V2MoO11 and the reactivity of
[1] L. Abbas, L. Bih, A. Nadiri, Z. El Amraoui, H. Khemakhem, D. Mazzane, Chemical
ZnMoO4 towards Zn2V2O7 in the solid state, Solid State Phenom. 90–91 (2003)
durability of MoO3-P2O5 and K2O-MoO3-P2O5 glasses, J. Therm. Anal. Calorim. 90
341–346.
(2007) 453–458.
[35] K.K. Palkina, S.I. Maksimova, N.T. Chibiskova, K. Schlesinger, G. Ladwig, Double
[2] L. Bih, A. Nadiri, Y. El Amraoui, Investigation of the physico-chemical properties of
octahedron cluster [V2O9] in the crystal structure of vanadium (III) diphosphate,
NaPO3-MoO3 glasses, J. Phys. IV France 123 (2005) 165–169.
V4(P2O7)3, Z. Anorg. Allg. Chem. 529 (1985) 89–96.
[3] S.H. Santagneli, C.C. de Araujo, W. Strojek, H. Eckert, G. Poirier, S.J.L. Ribeiro,
[36] S.A. Linde, Y.E. Gorbunova, A.V. Lavrov, X-ray diffraction study of VP3O9 systems,
Y. Messadeq, Structural studies of NaPO3 MoO3 glasses by solid-state nuclear
Russ. J. Inorg. Chem. 28 (1983) 16–18.
magnetic resonance and Raman spectroscopy, J. Phys. Chem. B 111 (2007)
[37] V.D. Zhuravlev, Y.A. Velikodnyi, L.L. Surat, A powder X-ray diffraction study of
10109–10117.
Mn2V2O7-M2V2O7 systems, (M ¼ Ba, Sr, Ca, Co, Zn, Cu, or Ni), Russ. J. Inorg.
[4] I.R. Zakis, A.R. Lusis, Y.L. Lagzdons, Color centers in tungsten phosphate glasses
Chem. 38 (1993) 1133–1136.
and amorphous tungsten trioxide films, J. Non-Cryst. Solids 47 (1982) 267–269.
[38] B.D. Jordan, C. Calvo, α-V1.08P0.92O5 at 22� C, Acta Crystallogr. B 32 (1976)
[5] F. Studer, A. Lebail, B. Raveau, Local environment of tungsten in mixed valence
2899–2900.
tungsten phosphate glasses: an EXAFS study, J. Solid State Chem. 63 (1986)
[39] J. Galy, Vanadium pentoxide and vanadium oxide bronzes-structural chemistry of
414–423.
single (S) and double (D) layer MxV2O5 phases, J. Solid State Chem. 100 (1992)
[6] J. Sub�
� cík, L. Koudelka, P. Mo�sner, L. Montagne, G. Tricot, L. Delevoye, I. Gregora,
229–245.
Glass-forming ability and structure of ZnO–MoO3–P2O5 glasses, J. Non-Cryst.
[40] A. El Kasmi, Z.Y. Tian, H. Vieker, A. Beyer, T. Chafik, Innovative CVD synthesis of
Solids 356 (2010) 2509–2516.
Cu2O catalysts for CO oxidation, Appl. Catal., B 186 (2016) 10–18.
[7] M.A. Ghauria, S.A. Siddiqi, M.G.B. Ashiq, Band gap measurement of ZnO-MoO3-
[41] A.S. Das, M. Roy, D. Roy, S. Bhattacharya, P.M.G. Nambissan, Defects
P2O5 glasses by photoconductivity, Glass Phys. Chem. 40 (2) (2014) 151–156.
characterization and study of amorphous phase formation in xV2O5-(1-x)Nd2O3
[8] A.K. Hassan, Properties of oxychloride glass system in relation to fast ion
binary glass nanocomposites using positron annihilation and correlated
conduction, J. Phys. Condens. Matter 11 (1999) 7995–8004.
experimental techniques, J. Alloys Compd. 753 (2018) 748–760.
[9] G. Tricot, L. Montagne, L. Delevoye, G. Palavit, V. Kostoj, Redox and structure of
[42] B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley, Reading MA, 1956.
sodium-vanadophosphate glasses, J. Non-Cryst. Solids 345–346 (2004) 56–60.
[43] H. Tian, C.A. Roberts, I.E. Wachs, Molecular structural determination of
[10] N. Vedeanu, O. Cozar, I. Ardelean, S. Filip, Spectroscopic investigation on some
molybdena in different environments: aqueous solutions, bulk mixed oxides, and
calcium-phosphate glasses, J. Optoelectron. Adv. Mater. 8 (2006) 1135–1139.
supported MoO3 catalysts, J. Phys. Chem. C 114 (2010) 14110–14120.
[11] N. Vedeanu, O. Cozar, I. Ardelean, B. Lendl, D.A. Magdas, Raman spectroscopic
[44] I. Kashif, S.A. Rahman, A.A. Soliman, E.M. Ibrahim, E.K. Abdel-Khalek, A.
study of CuO–V2O5–P2O5–CaO glass system, Vib. Spectrosc. 48 (2008) 259–262.
G. Mostafa, A.M. Sanad, Effect of alkali content on AC conductivity of borate
[12] S. Annamalai, R.P. Bhatta, I.L. Pegg, B. Dutta, Mixed transition-ion effect in the
glasses containing two transition metals, Physica B 404 (2009) 3842–3849.
glass system: Fe2O3-MnO-TeO2, J. Non-Cryst. Solids 358 (2012) 1380–1386.
[45] D.P. Singh, G.P. Singh, Conversion of covalent to ionic behavior of
[13] B. Dutta, N.A. Fahmy, I.L. Pegg, Effect of mixed transition-metal ions in glasses. I.
Fe2O3–CeO2–PbO–B2O3 glasses for ionic and photonic application, J. Alloy.
The P2O5–V2O5–Fe2O3 system, J. Non-Cryst. Solids 351 (2005) 1958–1966.
Compd. 546 (2013) 224–228.
[14] I. Kashif, S.A. Rahman, A.A. Soliman, E.M. Ibrahim, E.K. Abdel-Khalek, A.
[46] A.R. Zanatta, Revisiting the optical bandgap of semiconductors and the proposal of
G. Mostafa, A.M. Sanada, Effect of alkali content on AC conductivity of borate
a unified methodology to its determination, Sci. Rep. 9 (2019) 11225.
glasses containing two transition metals, Physica B 404 (2009) 3842–3849.
[47] J. Tauc, Optical properties and electronic structure of amorphous Ge and Si, Mater.
[15] L. Srinivasa Rao, M. Srinivasa Reddy, D. Krishna Rao, N. Veeraiah, Influence of
Res. Bull. 3 (1968) 37–46.
redox behavior of copper ions on dielectric and spectroscopic properties of
[48] J. Tauc, Amorphous and Liquid Semiconductors, Plenum, New York, 1974,
Li2O–MoO3–B2O3: CuO glass system, Solid State Sci. 11 (2009) 578–587.
pp. 159–220.
[16] A. Yadav, S. Khasa, A. Hooda, M.S. Dahiya, A. Agarwal, P. Chand, EPR and
[49] E.A. Davis, N.F. Mott, Conduction in non-crystalline systems V. Conductivity,
impedance spectroscopic investigations on lithium bismuth borate glasses
optical absorption and photoconductivity in amorphous semiconductors, Phil.
containing nickel and vanadium ions, Spectrochim. Acta 157 (2016) 129–137.
Mag. 22 (1970) 903–922.
[17] L. Bih, M. Elomari, J.M. Reau, M. Hadded, D. Boudlich, A.M. Yocaubi, A. Nadiri,
[50] F. Urbach, The long-wavelength edge of photographic sensitivity and of the
Electronic and ionic conductivity of glasses inside the Li2O–MoO3–P2O5 system,
electronic absorption of solids, Phys. Rev. 92 (1953) 1324.
Solid State Ionics 132 (2000) 71–85.
[51] A.S. Hassanien, A.A. Akl, Effect of Se addition on optical and electrical properties
[18] L.S. Rao, T.V. Rao, DSC, ESR and IR spectral studies on Li2O–WO3–B2O3 glass
of chalcogenide CdSSe thin films, Superlattice. Microst. 89 (2016) 153–169.
system doped with vanadium ions, Funct. Mater. Lett. 2 (2009) 127–130.
[52] D.P. Almond, A.R. West, Anomalous conductivity prefactors in fast ion conductors,
[19] L.S. Rao, N. Veeraiah, T.V. Rao, AC conduction phenomenon of Li2O–WO3–B2O3
Nature 306 (1983) 456–467.
glasses doped with V2O5, Funct. Mater. Lett. 6 (1–5) (2013) 1350032.
[53] D.P. Almond, G.K. Duncan, A.R. West, The determination of hopping rates and
[20] M.G. El-Shaarawy, Physical studies on ternary vanadium–phosphate glasses,
carrier concentrations in ionic conductors by a new analysis of ac conductivity,
J. Phys. Soc. Jpn. 71 (4) (2002) 1118–1125.
Solid State Ionics 8 (1983) 159–164.
[21] A.S. Hassanien, A.A. Akl, Estimation of some physical characteristics of
[54] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Di Electrics, London, 1983,
chalcogenide bulk Cd50S50 xSex glassy systems, J. Non-Cryst. Solids 428 (2015)
pp. 10–152.
112–120.
[55] A. Cizman, E. Rysiakiewicz-Pasek, M. Krupi� nski, M. Konon, T. Antropova,
[22] D. Biswas, R.K.N. Ningthemcha, A.S. Das, L.S. Singh, Structural characterization
M. Marszałek, The effect of Fe on the structure and electrical conductivity of
and electrical conductivity analysis of MoO3–SeO2–ZnO semiconducting glass
sodium borosilicate glasses, Phys. Chem. Chem. Phys. 19 (2017) 23318–23324.
nanocomposites, J. Non-Cryst. Solids 515 (2019) 21–33.
[56] A. Dhahri, E. Dhahri, E.K. Hlil, Electrical conductivity and dielectric behaviour of
[23] A.A. Akl, S.A. Mahmoud, S.M. AL-Shomar, A.S. Hassanien, Improving
nanocrystalline La0.6Gd0.1Sr0.3Mn0.75Si0.25O3, RSC Adv. 8 (2018) 9103–9111.
microstructural properties and minimizing crystal imperfections of nanocrystalline
[57] Z. Zhao, X. Song, T. Zhang, K. Hu, X. Liang, S. Li, Y. Zhang, I. Baturin, V. Shur,
Cu2O thin films of different solution molarities for solar cell applications, Mater.
Influence of lanthanum substitution on microstructure and impedance behavior of
Sci. Semicond. Process. 74 (2018) 183–192.
barium strontium titanate glass-ceramics, J. Appl. Phys. 126 (2019), 074101.
[24] S.B. Kolavekar, N.H. Ayachit, Conductivity and dielectric relaxations in Bi2O3-
[58] A.S. Das, D. Biswas, Investigation of Ac conductivity mechanism and dielectric
doped phospho-vanadate glasses, J. Mater. Sci. Mater. Electron. 30 (1) (2019)
relaxation of semiconducting neodymium-vanadate nanocomposites: temperature
432–449.
and frequency dependency, Mater. Res. Express 6 (2019), 075206.
[25] S. Park, John O. Baker, M.E. Himmel, P.A. Parilla, D.K. Johnson, Cellulose
[59] N.F. Mott, E.A. Davis, Electronic Processes in Non-crystalline Materials, Clarendon
crystallinity index: measurement techniques and their impact on interpreting
Press, Oxford, 1978, pp. 3–152.
cellulase performance, Biotechnology for Biofuels, Biotechnol. Biofuels 3 (2010)
[60] A.S. Hassanien, A.A. Akl, Electrical transport properties and Mott’s parameters of
10.
chalcogenide cadmium sulphoselenide bulk glasses, J. Non-Cryst. Solids 432
[26] W. Ruland, X-ray determination of crystallinity and diffuse disorder scattering,
(2016) 471–479.
Acta Crystallogr. 14 (1961) 1180–1185.
[61] N.F. Mott, Conduction in glasses containing transition metal ions, J. Non-Cryst.
[27] B. Papajani, E. Qoku, P. Malkaj, T. Dilo, The study of phase compound and the
Solids 1 (1968) 1–17.
degree of crystallinity of recycled LDPE by X-ray diffractometer and optical
[62] N.F. Mott, Conduction in non-crystalline materials III. Localized states in a
microscope, Int. J. Sci. Res. 4 (2) (2015) 2228–2232.
pseudogap and near extremities of conduction and valence bands, Phil. Mag. 19
[28] S. Sharma, A.A. Singh, A. Majumdar, B.S. Butola, Tailoring the mechanical and
(1969) 835–852.
thermal properties of polylactic acid-based bio-nanocomposite films using
[63] S.R. Elliott, Ac conduction in amorphous chalcogenide and pnictide
halloysite nanotubes and polyethylene glycol by solvent casting process, J. Mater.
semiconductors, Adv. Phys. 36 (1987) 135–217.
Sci. 54 (2019) 8971–8983.
[64] S.R. Elliott, Frequency-dependent conductivity in ionically and electronically
[29] C.B. Doelle, D. Stachel, I. Svoboda, H. Fuess, Crystal structure of zinc ultra-
conducting amorphous solids, Solid State Ionics 70–71 (1994) 27–40.
phosphate, ZnP4O11, Z. Kristallogr. 203 (1993) 282–283.
[65] K.S. Gilroy, W.A. Phillips, An asymmetric double-well potential model for
[30] J. Meullemeestre, E. Penigault, Les molybdates neutres de zinc, Bull. Soc. Chim. Fr.
structural relaxation processes in amorphous materials, Philos. Mag. B. 43 (1981)
10 (1972) 3669–3674.
735–746.
[31] I.M. Watson, M.M. Borel, J. Chardon, A. Leclaire, Structure of the trivalent
[66] G.E. Pike, Ac conductivity of scandium oxide and a new hopping model for
molybdenum metaphosphate Mo(PO3)3, J. Solid State Chem. 111 (1994) 253–256.
conductivity, Phys. Rev. B 6 (1972) 1572–1580.
[32] C. Calvo, The crystal structure and phase transitions of β-Zn2P207, Can. J. Chem. 43
[67] S.R. Elliott, Temperature dependence of a.c. conductivity of chalcogenide glasses,
(5) (1965) 1147–1153.
Philos. Mag. B. 37 (1978) 553–560.

13
D. Biswas et al. Journal of Physics and Chemistry of Solids 144 (2020) 109505

[68] S.R. Elliott, A theory of ac conduction in chalcogenide glasses, Philos. Mag. 36 [70] R. Mondal, D. Biswas, D. Mandal, A.S. Das, S. Kabi, Consequence of the heat-
(1978) 1291–1304. treatment on the ionic conductivity of silver-tellurite glass nanocomposites, J. Non-
[69] A. Ghosh, A. Pan, Scaling of the conductivity spectra in ionic glasses: dependence Cryst. Solids 506 (2019) 51–57.
on the structure, Phys. Rev. Lett. 84 (2000) 2188–2190. [71] B. Roling, A. Happe, K. Funke, M.D. Ingram, Carrier concentrations and relaxation
spectroscopy: new information from scaling properties of conductivity spectra in
ionically conducting glasses, Phys. Rev. Lett. 78 (1997) 2160–2163.

14