Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Journal of Thermal Analysis and Calorimetry

https://doi.org/10.1007/s10973-020-09345-z

WO3 thermodynamic properties at 80–1256 K revisited


Bing‑yuan Han1 · Andrey V. Khoroshilov2 · Alexander V. Tyurin2 · Alexander E. Baranchikov2 · Mikhail I. Razumov2 ·
Olga S. Ivanova2 · Konstantin S. Gavrichev2 · Vladimir K. Ivanov2

Received: 6 September 2019 / Accepted: 13 January 2020


© Akadémiai Kiadó, Budapest, Hungary 2020

Abstract
Thermodynamic properties (heat capacity Cp, entropy S°, enthalpy change ΔH° and reduced Gibbs energy ΔΦ°) of crystalline
­WO3 in a temperature range of 80–1256 K were refined using two complementary calorimetric methods, namely differential
scanning calorimetry and adiabatic vacuum calorimetry. Enthalpy changes during a number of successive phase transitions
in ­WO3 were calculated from differential scanning calorimetry data in both heating and cooling regimes and compared
to previously reported values. For the first time, enthalpy changes for the orthorhombic to monoclinic and monoclinic to
tetragonal high-temperature phase transitions in ­WO3 were estimated.

Keywords Tungsten oxide · WO3 · Heat capacity · Phase transitions · Thermodynamic properties · DSC · Adiabatic
calorimetry

Introduction various reducing and oxidizing gases provide it with excel-


lent gas-sensing properties [11–13].
Tungsten trioxide ­(WO3), due to the unique combination of Exact knowledge of the physical properties of a mate-
its electronic properties and diverse polymorphism, attracts rial is a touchstone for its practical applications and for the
a great deal of attention in modern materials science. ­WO3 modelling of technological processes. The thermodynamic
is an n-type semiconductor material, with a band gap of parameters (enthalpy, entropy, Gibbs energy and heat capac-
2.7–3.2 eV depending on oxygen vacancy concentration ity) are the most significant ones, requiring the highest accu-
[1–4] capable of catalysing various oxidation processes upon racy and reproducibility. Certainly, special attention should
visible light irradiation, thus possessing valuable photocata- be paid to materials which can easily change their structure
lytic properties for waste water treatment and environmental upon heating, including tungsten trioxide. W ­ O3 is one of the
remediation [5–8]. The reversibility of ­W+6 ↔ W+5 transi- well-known transition metal oxides, and its thermodynamic
tion, accompanied by colour changes of the material, enables properties are generally considered to be well established.
the use of ­WO3-based thin films and coatings in a number Surprisingly, our thorough literature survey revealed confus-
of advanced applications, e.g. smart windows, anti-dazzle ing discrepancies in its thermodynamic data. For instance,
mirrors and information displays [9, 10]. The reversible and the reported data on W ­ O3 heat capacity at 298.15 K range
quick changes in W ­ O3 conductivity upon interaction with from 73.1 to 81.5 J K−1 mol−1 [14–16]. Low-temperature
(~ 10–55 K) ­WO3 heat capacity was reported by Bevolo et al.
[17], while at higher temperatures (~ 100–300 K) specific
heat (in J K−1 g−1) was measured by Sawada [18], but his
Electronic supplementary material The online version of this
article (https​://doi.org/10.1007/s1097​3-020-09345​-z) contains data did not agree with low-temperature measurements [17].
supplementary material, which is available to authorized users. Moreover, the data presented by Sawada’s group in differ-
ent reports did not match each other (see Fig. S1) [18, 19].
* Alexander E. Baranchikov The heat capacity of ­WO3 in the range of~ 70–300 K has
a.baranchikov@yandex.ru
also been determined by Seltz et al. [14], their data being
1
Jiangsu University of Technology, Changzhou 213001, consistent with low-temperature measurements [17], but
China not matching Sawada’s data [18]. Up to now, no attempt
2
Kurnakov Institute of General and Inorganic Chemistry,
Moscow, Russia 119991

13
Vol.:(0123456789)
B. Han et al.

has been made to determine ­WO3 heat capacity in a wide


0.2
temperature range.
Such discrepancies are due to a number of different rea-
0.1
sons. Experimental investigations of thermodynamic data

DSC signal/mW mg–1


were mostly conducted in the 1940s through to the 1970s
a' b' c' d'
using ­WO3 samples prepared by different procedures. The 0.0
a bc d
difficulties with exact determination of ­WO3 thermodynamic
data can be also connected with the similarity of the crystal – 0.1

structure of various ­WO3 polymorphs.


Unlike hexagonal ­WO3 [20, 21], the structure of a stoi- – 0.2 Heating
chiometric ­WO3 belongs to the ­ReO3-type, consisting of Cooling
corner-interconnected ­[WO6] octahedra [22]. Diverse W ­ O3 – 0.3
polymorphism arises from the slight mutual rotation of the 200 400 600 800 1000 1200
octahedra or their distortions, e.g. due to the cooperative dis- T/K
placement of tungsten atoms. Even a slight disturbance in
the structure significantly affects the symmetry of the crystal Fig. 1  Thermal analysis data (DSC) of ­
WO3 sample in the 313–
lattice. Nonetheless, the substantial feature of phase transi- 1273 K temperature range upon heating and subsequent cooling in
tions in ­WO3 is the stability of the main motif of the ­WO3 argon
crystal structure, namely the framework of interconnected
­[WO6] octahedra. The structure of various W ­ O3 polymorphs
was extensively studied in some recent experiments including of thermodynamic functions in a temperature range of
in situ high-temperature diffraction measurements [23–28]. 100–3000 K are given in a 100 K step.
According to one of the most reliable sources in ther- The present work was aimed at the heat capacity meas-
modynamic data, the IVTANTERMO database [16] and urements of ­WO3 in a wide temperature range, from ~ 80
the reported structural data [23, 24, 29–31] solid tungsten to ~ 1250 K, using two complementary calorimetric meth-
trioxide undergoes six phase transitions at the following ods, namely differential scanning calorimetry and adiabatic
temperatures: 228 K (− 45.5 ± 5 °C), triclinic (α) → triclinic vacuum calorimetry. Special attention was paid to reconcil-
(β); 291 K (18 ± 2 °C), triclinic (β) → monoclinic (γ); 603 K ing the results of calorimetric measurements in low-tem-
(330 ± 20 °C), monoclinic (γ) → orthorhombic; 1013 K perature and high-temperature regions. In our experiments,
(740 ± 10 °C), orthorhombic → tetragonal (t); 1173 K we also focused on the re-determination of ΔH values for
(900 ± 10 °C), most probably, tetragonal ­(t1) → tetragonal the phase transitions in W­ O3, paying special attention to
­(t2) [28]; 1493 K (1220 ± 20 °C), tetragonal ­(t2) → cubic. the temperature range of 920–1090 K. In this range, ΔH
According to Howard et al. [28], another monoclinic struc- values for the orthorhombic to monoclinic and monoclinic
ture exists (sp. gr. P21/c) in the temperature range from to tetragonal high-temperature phase transitions have not yet
about 993 K (720 °C) to 1073 K (800 °C). Surprisingly, been determined.
NIST–JANAF Thermochemical Tables [15] mention only
one phase transition in solid ­WO 3 (at 1000…1050 K,
orthorhombic to tetragonal). Experimental
The literature survey also showed that even the heat func-
tions values for the listed phase transitions are sometimes Synthesis
missing. ΔH and ΔS values have not been determined for
the first two phase transitions, which proceed below 298 K WO3 powder used in the calorimetric experiments was
(i.e. triclinic to triclinic and triclinic to monoclinic). The ΔH synthesized by thermolysis of ammonium B-paratungstate,
value for the monoclinic to orthorhombic phase transition according to the previously reported procedure [36]. Briefly,
(1.380 ± 0.209 kJ mol−1) has been reported in Sawada’s paper 1.0 g of ­(NH4)10[H2W12O42]·2.6H2O in a wide alundum cru-
only [18]; the ΔH value for the orthorhombic to tetragonal cible was heated in air from room temperature to 1273 K
phase transition (1.882 ± 0.209 kJ mol−1) has been deter- (1000 °C) at a rate of 10 K min−1 and then held at 1273 K for
mined by several authors, namely Bokov and Myl’nikova 24 h in air. Next, the furnace was switched off and allowed
[32], Leute [33] and Sawada [18, 34, 35]; ΔH values for to cool naturally to room temperature overnight, and then,
the tetragonal to tetragonal (1.171 ± 0.125 kJ mol−1) and the sample was drawn out. The resultant yellowish powder
for the tetragonal to cubic (0.502 ± 0.125 kJ mol−1) phase was divided into parts which were analysed by powder X-ray
transitions have also been determined by Sawada [18, 34]. diffraction, differential scanning calorimetry and adiabatic
Note that, in a NIST–JANAF database [15], the values calorimetry.

13
WO3 thermodynamic properties at 80–1256 K revisited

Table 1  Parameters of the phase transitions (Ton—onset temperature; Tpeak—temperature of the peak maximum; and ΔH—enthalpy of the transi-
tion) in ­WO3 determined by DSC measurements in the temperature range of 313–1253 K
Effect Literature data This work
Chang Tpeak, K ∆H, Refs. Ton, K Tpeak, ∆Н, J∙mol–1
es in (°C) J∙mol–1 (°C) K (°C)
crystal
system a

Heati
ng
a m o 593…623 – [18,19,24,31,43] 595.75 599.55 –73.12±0.28
(320…35 1380±2 (322.6 (326.4 –79.202±0.309**
0) 09 0) 0)
±0.25 ±0.25
b o 1013.7 1020.5 – –
m’ 5 5 1035.6±26. 1222.22±30.92 b
(740.6 (747.4 2
983…103 0) 0) –

3 [18,19,24,33,34,4 ±0.35 ±0.35 1449.18±36.66 c
1882±2
c m' t (710…76 4,45] 1031.6 1052.3 –
09
0) 1 5 186.62±2.3
(758.4 (779.2 5
6) ) ±0.25
±0.35
d t t 1118…11 – [18,34,42,46,47] 1172.8 1175.4 –290.72±3.26
83 1171±1 5 5 –340.0±8.74**
(845…91 25 (899.7 (902.3
0) 0) 0)
±0.35 ±0.35
Cooli
ng
d t t 1153 – [42] 1164.5 1168.6 230.26±3.48
(880) 5 5
(891.4 (895.5
0) 0)
±0.35 ±0.35
c t m’ – – 1056.1 1047.3 387.16±9.9 1215.73±31.00 b
5 5 5
(783.0 (774.2
) ±0.3 ) ±0.3
b m' o 1003.3 997.65 828.57±21.
5 (724.5 13
(730.1 ) ±0.3
8)
±0.35
a o m – – 600.25 585.85 23.62±0.09
(327.1 (312.7
0) ) ±0.25
±0.25
A comparison with previously reported data is also given
a
m, m’—monoclinic; o—orthorhombic; t—tetragonal
b
Cumulative enthalpy corresponding to b + c or b’ + c’ heat effects
c
ΔH0 values calculated from heat capacity measurements

13
B. Han et al.

Powder X‑ray diffraction experimental data was performed using the spline method,
which is unrelated to a certain Cp (T). The method allows
X-ray diffraction analysis of the samples was carried out on approximation of Cp (T) functions with various anomalies
a Bruker D8 Advance diffractometer on ­CuKα1, 2 radiation (related to phase transitions in a material, for instance). Rela-
in the angular range 3–100° 2θ, in steps of 0.02° 2θ, with tive masses of the data were calculated using a piecewise
rotation of the cuvette (30 rpm), at a counting time of 0.5 s polynomial approximation, where a polynomial degree in
per step. For a full-profile analysis of the diffraction pat- each segment was chosen according to the Fisher crite-
tern, JANA2006 software [37] was applied. Fourth-order rion. The error for each point was calculated as a difference
Chebyshev polynomials with a reciprocal term were used to between an experimental value and approximated value.
fit the background. The overall fitting was performed using The error of thermodynamic functions was calculated from
the fundamental parameter approach. the standard deviation of the experimental points from the
approximated curve and from the error of the extrapolation
Adiabatic calorimetry to 0 K.
Thermodynamic functions, Cp0 (T), S0 (T), 𝛷0 (T), ΔH 0 (T),
WO3 heat capacity in a temperature range of 80.71–348.18 K were calculated within the extrapolation region according to
was measured using an adiabatic vacuum calorimeter BKT-3 the additive Debye–Einstein model. Values of enthalpy
(CJSC “Termis”, Russia). The heat capacity was measured change H 0 (T) − H 0 (0) and absolute entropy S0 (T) were cal-
at a step of 1–4 K in the range of 80–200 K and at a step culated according to the following equations:
of 4–7 K at temperatures above 200 K. All measurements
T
were taken in automatic mode. For the calorimetric experi- ΔH 0 (T) = H 0 (T) − H 0 (0) = ∫ Cp0 (T)dT,
ments, a special thin-walled cylindrical titanium vessel, with 0

a screw cap and a gasproof indium seal, was used. The inner
volume of the ampule was 1 cm3 and its mass was ~ 1.6 g. T Cp0 (T)
The ampule was evacuated, and measurements were taken S0 (T) = ∫ dT.
0 T
at a pressure not higher than 0.015 Pa. The temperature
of the sample was controlled by means of an iron–rho- Reduced Gibbs energy, 𝛷]0 (T) , is defined as
dium resistance thermometer (resistance at room tempera- 𝛷 (T) = S (T) − T H (T) − H 0 (0) .
0 0 1[ 0

ture ≈ 100 Ω). The sensitivity of the thermometric scheme


was 1 × 10−3 K. The absolute error of temperature measure- Differential scanning calorimetry
ments was ± 5 × 10−3 K, in accordance with the international
temperature scale (ITS-90). The reliability of measurements Differential scanning calorimetry measurements were taken
was checked by experimental determination of the heat using a Netzsch STA 449F1 Jupiter synchronous thermal ana-
capacity of metallic copper, standard synthetic corundum lyser under argon (99.999%) according to the requirements of
and benzoic acid (ACS grade). For the measurements, a the standard test method ASTM E 1269. Before the experi-
1.81356(3) g ­WO3 sample was used. For the calculations, ments, the temperature and sensitivity calibration was per-
­WO3 molar mass was taken to be equal to 231.8382 g mol−1 formed using certified metallic standards (DIN 51007, DIN
[38]. ASTM E 967/DIN EN 10 204—2.1). The standard samples
The low-temperature heat capacity data were spline- were placed in platinum–rhodium crucibles with thin-walled
approximated using software which is part of an IVTAN- alundum liners. Before the measurements, the inner volumes
TERMO database [16]. of thermobalance and furnace were evacuated and filled with
Experimental data were processed in two steps. At the argon (twice). Thermal analysis of the ­WO3 sample was per-
first step, the experimental heat capacity values were extrap- formed in a temperature range of 313–1273 K (40–1000 °C);
olated to 0 K using the data presented in the report of Bevolo the sample mass was 92.71 mg. The following heating regime
et al. [17]. For the extrapolation to 0 K, an additive scheme was used: thermal equalizing at 313 K (40 °C) for 15 min;
was used: heating to 1273 K (1000 °C) at a heating rate of 20 K min−1;
n−1
cooling down to 313 K (40 °C) at a cooling rate of 20 K min−1.
The analysis of differential scanning calorimetry data was
( ) ( )
𝛩D ∑ 𝛩E
CV = 3D +3 Ei ,
T i=1
T performed using Netzsch Proteus v.5.2.1 software. The soft-
ware was used to calculate ΔH values and the temperature
where D—Debye function, Ei—Einstein function, ΘD— values corresponding to the beginning of the thermal effects
Debye temperature (K), ΘE—Einstein temperature (K) and (Ton) and the maxima of the peaks (Tpeak) in a DSC curve.
T—absolute temperature (K). The approximation of the

13
WO3 thermodynamic properties at 80–1256 K revisited

(a) (b) 0.10


0.00

DSC signal/mW mg–1


DSC signal/mW mg–1

Heating
– 0.05
0.05

– 0.10
Cooling

0.00
– 0.15
∆DSC*1000/

∆DSC*1000/
mW mg–1

mW mg–1
2 2
0 0
–2 –2
950 1000 1050 1100 950 1000 1050 1100
T/K T/K

Fig. 2  Fitting the DSC signal in a temperature range of 920–1100 K using pairs of asymmetric Gaussians (depicted as blue and red curves): a
heating curve and b cooling curve. The difference between experimental data and the fitting curve is given at the bottom of each graph

For the specific heat capacity measurements, the following 250


0 50 100 150 200 250 300
100
heating regime was used: thermal equalizing at 303 K (30 °C) 1
for 15 min; then heating to 313 K (40 °C) at a heating rate of 2
5 K min−1; equalizing at 313 K (40 °C) for 15 min; and heat- 200 3
ing to 1276 K (1003 °C) at a heating rate of 20 K min−1. The 50 4
5
sample mass was 68.08 mg. As a reference, a pellet (6.0 mm
Cp/J mol–1 K–1

150 6
diameter, 0.5 mm height, 56.6 mg mass) of a certified (DIN
51007, DIN 53765, DIN ASTM E 968/DIN EN 10 204—2.1) 0
100
synthetic sapphire (α-Al2O3,) was used. Before the measure-
ments, the inner volumes of thermobalance and furnace were
evacuated and filled with argon (twice). 50
Specific heat capacity values were calculated using sap-
phire as a reference. For the calculations, the following ratio
was used: 0 500 1000 1500
T/K
mref DSCsample − DSCbas
cp = ⋅ ⋅ cp, ref ,
msample DSCref − DSCbas Fig. 3  WO3 heat capacity versus temperature according to (1) Seltz
et al. [14]; (2) Sawada [18]; (3) Bevolo et al. [17]; (4) IVTAN-
where cp—calculated specific heat value (stated in TERMO [16]; and (5) NIST–JANAF [15]. The black line (6) cor-
J K−1 g−1); cp, ref —specific heat of a reference substance; responds to the experimental data obtained in this work (both DSC
mref and msample—masses of a reference and a sample, and adiabatic vacuum calorimetry). Inset: Magnified low-temperature
region of the graph
respectively; DSCref , DSCsample and DSCbas—DSC values
for a reference, a sample and a baseline, respectively. The
molar heat capacity (Cp ) was calculated by multiplying the
Results and discussion
specific heat (cp ) by the molar mass.
The fitting of the temperature dependence of heat capac-
For the synthesis of the W
­ O3 sample, we chose prolonged
ity in the temperature ranges between phase transitions of
isothermal heating of ammonium B-paratungstate at 1273 K
­WO3 was performed using a Maier–Kelley equation:
in air, followed by slow cooling. Such conditions favour
Cp (T) = a + b ⋅ T + c ⋅ T −2 . the formation of a stoichiometric W­ O3 and the absence of
ammonium tungsten bronzes (see, e.g. [39, 40]) in a powder.
According to our recent data, the traces of ammonium ions
may be present in tungsten trioxide synthesized from ammo-
nium B-paratungstate upon its heating to 873 K (600 °C), or

13
B. Han et al.

Table 2  Maier–Kelley equation Range Temperature ­rangea/K Maier–Kelley equation coefficients


coefficients calculated from
fitting the dependence of a/J K−1 mol−1 b/J mol−1 c/J K−3 mol−1
­WO3 specific heat capacity on
temperature measured by DSC 1 316–599.55 87.54 0.01635 − 1.746 × 106
2 599.55–1052.35 87.59 0.01592 − 1.764 × 106
3 1052.35–1175.45 108.6 − 5.977 × 10−4 − 1.121 × 107
4 1175.45–1276 193.3 − 4.541 × 10−2 − 5.660 × 107
a
The values for the temperature ranges’ boundaries correspond to Tpeak values (see Table 1)

Table 3  WO3 heat capacity in T/K Cp/J K−1 mol−1 T/K Cp/J K−1 mol−1 T/K Cp/J K−1 mol−1 T/K Cp/J K−1 mol−1
the range of 80.71–348.18 K
calculated using adiabatic 80.71 23.23 130.98 40.65 205.43 59.12 278.37 71.32
calorimetry data
83.56 24.45 134.49 41.66 209.85 60.02 282.41 71.85
85.79 25.38 138.02 42.67 214.25 60.89 286.45 72.37
88.03 26.29 143.72 44.26 218.63 61.74 290.44 72.86
90.31 27.19 148.05 45.44 222.95 62.44 294.39 73.34
92.57 28.07 151.55 46.38 227.32 63.18 298.29 73.80
94.84 28.92 155.05 47.30 231.62 63.85 302.12 74.24
97.12 29.76 163.13 49.39 236.02 64.62 306.40 74.72
99.52 30.63 167.46 50.48 240.43 65.28 312.65 75.40
102.54 31.69 170.98 51.35 244.73 66.01 319.91 76.15
102.54 31.69 178.39 53.13 249.02 66.72 327.13 76.87
106.03 32.89 182.67 54.14 253.30 67.43 334.28 77.55
109.67 34.10 186.52 55.02 257.52 68.13 341.21 78.18
113.18 35.23 190.81 55.99 261.75 68.85 348.18 78.79
116.70 36.35 194.36 56.77 265.96 69.55
120.22 37.44 197.90 57.54 270.13 70.21
126.65 39.38 201.46 58.29 274.27 70.77

even to 1073 K (800 °C) [36]. According to X-ray diffrac- DSC curve shows exothermic heat effects corresponding to
tion analysis, the ­WO3 powder used in our work for calori- the above-listed phase transitions (d′, b′ + c′, a′), indicating
metric experiments consisted of only γ-WO3 (monoclinic that all these transitions are fully reversible. Evidently, the
system, sp. gr. P21/n). Full-profile analysis of the powder temperatures of these effects in a cooling DSC curve were
X-ray diffraction pattern (Fig. S2) resulted in the follow- shifted to lower values.
ing lattice parameters: a = 7.30093(15) Å, b = 7.53729(15)
Å, c = 7.69010(16) Å, β = 90.8749(18)° (GOF = 1.40, a and a′ thermal effects
Rp = 6.87, wRp = 9.10).
According to thermal analysis performed in argon in a In the temperature range of 566–625 K (293–352 °C), a
temperature range of 300–1250 K, the mass of the sam- small thermal effect is observed, which can unambiguously
ple did not change upon heating or subsequent cooling, be attributed to a monoclinic (sp. gr. P21/n) → orthorhom-
the change in sample mass did not exceed 0.015%, and no bic (sp. gr. Pcnb [23]) ­WO3 transformation. Note that the
mass loss due to the elimination of oxygen from ­WO3 was effect is quite weak, and in some recent works it has not
detected (Fig. S3). been detected at all, (see, for example, the results reported
Figure 1 presents the results of the DSC study upon by Sale et al. [41]). The parameters of this phase transi-
heating and subsequent cooling. The heating branch of the tion (Ton, Tpeak, ΔH) are listed in Table 1. Note that the ΔH
DSC curve shows several endothermic effects which can be values obtained in our experiments are ~ 20–50 times lower
assigned to previously reported phase transitions in ­WO3: (a) than the previously reported value (− 73.12 ± 0.28 J mol−1
monoclinic → orthorhombic; (b + c) orthorhombic → tetrag- at heating and 23.62 ± 0.09 J mol −1 at cooling versus
onal; (d) tetragonal → tetragonal. The cooling branch of the − 1.380 ± 0.209 kJ·mol−1) [16].

13
WO3 thermodynamic properties at 80–1256 K revisited

Table 4  WO3 heat capacity T/K Cp/J mol−1 K−1 T/K Cp/J mol−1 K−1 T/K Cp/J mol−1 K−1 T/K Cp/J mol−1 K−1
in the temperature range of
316–1276 K calculated using 316 75.24 561 91.17 801 97.60 1046 102.6
DSC data
321 75.86 566 91.35 806 97.71 1052.35 102.7
326 76.45 571 91.52 811 97.82 1052.35 97.80
331 77.03 576 91.70 816 97.94 1056 97.88
336 77.58 581 91.87 821 98.05 1061 97.97
341 78.11 586 92.04 826 98.16 1066 98.06
346 78.62 591 92.21 831 98.27 1071 98.00
351 79.12 596 92.37 836 98.38 1076 98.23
356 79.59 599.55 92.48 841 98.49 1081 98.32
361 80.05 599.55 92.21 846 98.60 1086 98.40
366 80.50 606 92.44 851 98.71 1091 98.49
371 80.93 611 92.60 856 98.81 1096 98.57
376 81.35 616 92.75 861 98.92 1101 98.65
381 81.75 621 92.91 866 99.03 1106 98.73
386 82.14 626 93.06 871 99.14 1111 98.81
391 82.52 631 93.21 876 99.24 1116 98.89
396 82.89 636 93.36 881 99.35 1121 98.97
401 83.25 641 93.51 886 99.45 1126 99.04
406 83.59 646 93.65 891 99.56 1131 99.12
411 83.93 651 93.80 896 99.66 1136 99.19
416 84.26 656 93.94 901 99.77 1141 99.27
421 84.58 661 94.08 906 99.87 1146 99.34
426 84.89 666 94.22 911 99.97 1151 99.41
431 85.19 671 94.36 916 100.1 1156 99.48
436 85.49 676 94.50 921 100.2 1161 99.55
441 85.78 681 94.63 926 100.3 1166 99.62
446 86.06 686 94.77 931 100.4 1171 99.68
451 86.33 691 94.90 936 100.5 1175.45 99.74
456 86.60 696 95.03 941 100.6 1175.45 98.98
461 86.87 701 95.17 946 100.7 1181 99.11
466 87.12 706 95.30 951 100.78 1186 99.23
471 87.37 711 95.43 956 100.88 1191 99.34
476 87.62 716 95.55 961 100.98 1196 99.44
481 87.86 721 95.68 966 101.08 1201 99.54
486 88.10 726 95.81 971 101.18 1206 99.64
491 88.33 731 95.93 976 101.28 1211 99.73
496 88.56 736 96.06 981 101.38 1216 99.82
501 88.78 741 96.18 986 101.48 1221 99.91
506 89.00 746 96.30 991 101.57 1226 99.99
511 89.21 751 96.42 996 101.67 1231 100.1
516 89.42 756 96.54 1001 101.77 1236 100.1
521 89.63 761 96.66 1006 101.87 1241 100.2
526 89.83 766 96.78 1011 101.96 1246 100.3
531 90.03 771 96.90 1016 102.06 1251 100.3
536 90.23 776 97.02 1021 102.16 1256 100.4
541 90.42 781 97.14 1026 102.25 1261 100.5
546 90.61 786 97.25 1031 102.35 1266 100.5
551 90.80 791 97.37 1036 102.44 1271 100.6
556 90.98 796 97.48 1041 102.54 1276 100.6

Bold italics indicate the temperatures of phase transitions (Tpeak, see Table 1) and the corresponding heat
capacity values for W
­ O3 polymorphs stable at higher and lower temperatures

13
B. Han et al.

Table 5  Calculated values of T/K Cp°(T)/J K−1 mol−1 S°(T)/J K−1 mol−1 H0(T) − H0(0)/J mol−1 Φ0(T) − Φ0(0)/J K−1 mol−1
­WO3 thermodynamic functions
(heat capacity Cp, entropy S0, 5 0.02518 0.000103 0.000476 0.0000078
enthalpy change ΔH0, and
10 0.1997 0.0356 0.3080 0.0048
reduced Gibbs energy ΔΦ0) in a
temperature range of 5–1250 K 15 0.8159 0.2238 2.734 0.0416
20 1.670 0.5691 8.836 0.127
25 2.873 1.064 20.03 0.262
30 4.441 1.721 38.18 0.448
35 6.243 2.539 64.83 0.687
40 8.139 3.496 100.8 0.977
45 10.05 4.565 146.2 1.315
50 11.96 5.723 201.3 1.698
60 15.75 8.238 339.8 2.575
70 19.57 10.95 516.4 3.576
80 23.37 13.81 731.1 4.675
90 27.09 16.78 983.5 5.855
100 30.68 19.82 1272 7.099
110 34.09 22.91 1596 8.396
120 37.33 26.02 1954 9.735
130 40.39 29.12 2342 11.11
140 43.29 32.23 2761 12.50
150 46.05 35.31 3208 13.92
160 48.67 38.36 3681 15.35
170 51.16 41.39 4181 16.80
180 53.52 44.38 4704 18.25
190 55.78 47.33 5251 19.70
200 57.92 50.25 5819 21.15
210 59.95 53.13 6409 22.61
220 61.87 55.96 7018 24.06
230 63.69 58.75 7646 25.51
240 65.42 61.50 8291 26.95
250 67.05 64.20 8954 28.39
260 68.59 66.86 9632 29.82
270 70.05 69.48 10,330 31.22
280 71.42 72.05 11,030 32.66
290 72.71 74.58 11,750 34.06
298.15 73.72 76.61 12,350 35.19
300 73.94 77.07 12,490 35.43
310 75.09 79.51 13,230 36.83
320 76.18 81.91 13,990 38.19
330 77.20 84.27 14,760 39.54
340 78.17 86.59 15,530 40.91
350 79.08 88.87 16,320 42.24
400 83.16 99.71 20,380 48.76
450 86.27 109.7 24,620 54.98
500 88.73 118.9 29,000 60.92
550 90.76 127.5 33,480 66.58
599.55 92.48 135.4 38,030 71.95
599.55 92.23 135.5 38,110 71.95
600 92.24 135.6 38,150 71.99
650 93.77 143.0 42,800 77.17
700 95.14 150.0 47,520 82.13
750 96.40 156.6 52,310 86.88

13
WO3 thermodynamic properties at 80–1256 K revisited

Table 5  (continued) T/K Cp°(T)/J K−1 mol−1 S°(T)/J K−1 mol−1 H0(T) − H0(0)/J mol−1 Φ0(T) − Φ0(0)/J K−1 mol−1

800 97.57 162.9 57,160 91.43


850 98.68 168.8 62,060 95.81
900 99.74 174.5 67,030 100.0
950 100.8 179.9 72,040 104.1
1000 101.7 185.1 77,100 108.0
1050 102.7 190.1 82,210 111.8
1052.35 102.8 190.3 82,450 112.0
1052.35 97.80 191.8 83,900 112.1
1100 98.63 196.2 88,580 115.7
1150 99.39 200.6 93,530 119.3
1175.45 99.74 202.8 96,070 121.0
1175.45 98.98 203.0 96,410 121.0
1200 99.52 205.0 98,850 122.7
1250 100.3 209.1 103,800 126.0

Bold italic indicates the temperatures of phase transitions (Tpeak, see Table 1) and the corresponding values
of thermodynamic functions for ­WO3 polymorphs stable at higher and lower temperatures

Upon heating from 625 K (352 °C) to 925 K (652 °C), no Pearson VII functions. The results of the fitting are presented
thermal effects in the ­WO3 sample were observed. in Fig. 2a and 2b and in Table 1.
As can be judged from the narrow temperature range
b + c and b′ + c′ thermal effects (1014–1032 K) of the intermediate monoclinic phase exist-
ence, as well as from the shape of the DSC peaks, we can
The most appreciable observations were made within a tem- assume that in this temperature range, under certain condi-
perature range of 925–1083 K (652–810 °C). According to tions, the coexistence of tetragonal and monoclinic phases
well-known data [23], at these temperatures, W ­ O3 under- is possible.
goes a single-step phase transition, from orthorhombic (sp.
gr. Pcnb [23]) to tetragonal (sp. gr. P4/ncc [29]). On the d and d′ thermal effects
other hand, the recent results by Howard et al. [28], obtained
using high-temperature neutron powder diffraction, indicate The next reversible phase transformation was observed in the
the existence of an intermediate low-symmetry monoclinic ­WO3 sample within the temperature range of 1118–1198 K
­WO3 polymorph (sp. gr. P21/c) in a temperature range from (845–925 °C) upon heating (effect d) and of 1148–1183 K
about 993 K (720 °C) to 1073 K (800 °C). Note that upon (875–910 °C) upon cooling (effect d′). This observation is
the temperature rise, phase transformations in a solid usually in line with recently reported data [29]. Interestingly, this
occur according to successive increase in structure symme- phase transformation proceeds without the changes in crys-
try, and the decrease in symmetry (from Pcnb to P21/c) is tal system (tetragonal to tetragonal transformation) and is
quite eye-catching. accompanied by the change in space group (P4/ncc to P4/
Our data inevitably showed two relatively well-resolved nmm) due to antiphase rotations of the ­[WO6] octahedra
thermal effects that should correspond to two distinct phase about [001] [26, 29]. The parameters of this phase transition
transitions. Surprisingly, a similar double thermal effect (Ton, Tpeak, ΔH) are given in Table 1.
upon ­WO3 heating has been demonstrated earlier (e.g. by Figure 3 and Fig. S4 present the experimental data
Kehl et al. [42]), but these observations have not been dis- for ­WO3 heat capacity versus temperature in the range of
cussed at all. Thus, our results strongly support the diffrac- 80.71–1276 K as measured using differential scanning calo-
tion data obtained by Howard et al. [28] and indicate that rimetry and adiabatic vacuum calorimetry. In Fig. 3 and Fig.
orthorhombic to tetragonal phase transformation in W ­ O3 S4, reference data on ­WO3 heat capacity are also given.
proceeds through an intermediate phase formation. The Adiabatic vacuum calorimetry data were recorded in the
endothermic character of both thermal effects observed in a temperature range of 81–349 K. In this range, the corre-
heating branch of a DSC curve is indirect evidence that all sponding heat capacity curve monotonically increases, with
the above-listed phases are thermodynamically stable. no noticeable anomalies. Thus, previously reported low-
The fitting of a DSC signal within the temperature range temperature (< 298 K) phase transitions were not detected
of 920–1100 K was performed using two asymmetrical in our experiments. Our data are in a good agreement with

13
B. Han et al.

the results of low-temperature measurements (12–54 K) in the presence of Fenton-like reagent. Appl Catal B. 2013;138–
reported by Bevolo et al. [17]. Our data also coincide with 139:311–7. https​://doi.org/10.1016/j.apcat​b.2013.03.006.
6. Gondal MA, Hameed A, Yamani ZH. Laser induced photo-
the data presented in the IVTANTERMO database [16] catalytic splitting of water over W
­ O3 catalyst. Energy Sources.
and are very close to those in the NIST–JANAF database 2005;27:1151–65. https​://doi.org/10.1080/00908​31049​04795​
[15]. The heat capacity values reported by Seltz et al. [14] 74.
in a temperature range of 80–120 K almost coincide with 7. Szilágyi IM, Fórizs B, Rosseler O, Szegedi Á, Németh P, Király
P, Tárkányi G, Vajna B, Varga-Josepovits K, László K, Tóth
our data, being somewhat higher at higher temperatures. AL, Baranyai P, Leskelä M. W ­ O3 photocatalysts: influence of
Sawada’s [18] data are lower than ours below 150 K and are structure and composition. J Catal. 2012;294:119–27. https​://
higher at T > 150 K (Fig. S5). Surprisingly, we have noticed doi.org/10.1016/j.jcat.2012.07.013.
an excellent coincidence of our data with the heat capacity 8. Nagy D, Nagy D, Szilágyi IM, Fan X. Effect of the morphol-
ogy and phases of W ­ O 3 nanocrystals on their photocatalytic
values measured by Seltz et al. for M
­ oO3 [14] (see Fig. S6). efficiency. RSC Adv. 2016;6:33743–54. https​://doi.org/10.1039/
The fitting of the experimental dependence of heat capac- C5RA2​6582G​.
ity on temperature was performed using the Maier–Kelley 9. Avellaneda C. Photochromic properties of W ­ O 3 and W ­ O 3:
equation. The fitting results are presented in Table 2 as X (X = Ti, Nb, Ta and Zr) thin films. Solid State Ion.
2003;165:117–21. https​://doi.org/10.1016/j.ssi.2003.08.023.
Maier–Kelley equation coefficients. 10. Blackman CS, Parkin IP. Atmospheric pressure chemical vapor
The treatment of all data sets obtained in this work deposition of crystalline monoclinic ­WO3 and ­WO3 − x thin films
allowed us to calculate the standard isobaric heat capacity of from reaction of W ­ Cl 6 with O-containing solvents and their
­WO3 and a set of thermodynamic functions in the 5–1250 K photochromic and electrochromic properties. Chem Mater.
2005;17:1583–90. https​://doi.org/10.1021/cm040​3816.
temperature range. For the details, see Tables 3–5. 11. Wang G, Ji Y, Huang X, Yang X, Gouma P-I, Dudley M. Fabri-
cation and characterization of polycrystalline ­WO3 nanofibers
and their application for ammonia sensing. J Phys Chem B.
Conclusions 2006;110:23777–82. https​://doi.org/10.1021/jp063​5819.
12. Meng D, Yamazaki T, Shen Y, Liu Z, Kikuta T. Preparation of
­WO3 nanoparticles and application to N
­ O2 sensor. Appl Surf Sci.
Our experimental measurements support some recent exper- 2009;256:1050–3. https​://doi.org/10.1016/j.apsus​c.2009.05.075.
imental data on the thermodynamic properties of W ­ O3 and 13. Balázsi C, Wang L, Zayim EO, Szilágyi IM, Sedlacková K, Pfeifer
provide precise information on W ­ O3 heat capacity over J, Tóth AL, Gouma P-I. Nanosize hexagonal tungsten oxide for
gas sensing applications. J Eur Ceram Soc. 2008;28:913–7. https​
a wide temperature range (80–1256 K). We have revised ://doi.org/10.1016/j.jeurc​erams​oc.2007.09.001.
the data on the temperatures of phase transitions in W ­ O3 14. Seltz H, Dunkerley FJ, DeWitt BJ. Heat capacities and entro-
and the corresponding enthalpy changes. For the first time, pies of molybdenum and tungsten trioxides. J Am Chem Soc.
we have estimated the enthalpy changes during recently 1943;65:600–2. https​://doi.org/10.1021/ja012​44a03​0.
15. Chase MW. NIST–JANAF thermochemical tables. 4th ed.
reported high-temperature phase transitions in ­WO3, namely Washington: American Chemical Society; 1998.
orthorhombic to monoclinic and monoclinic to tetragonal. 16. http://www.chem.msu.su/cgi-bin/tkv.pl?lette ​ r =&joule​
s=1&show=termo​d ata&tabno​= 56&allow​_ addit​i onal​_ eleme​
Acknowledgements The work was supported by the Russian Sci- nts=&volno​=7&pg=0&no=1202&allow​_more_atoms​=&allow​
ence Foundation (Project 18-73-10150). This research was performed _no_ions=&brutt​o=WO3.
using shared experimental facilities supported by IGIC RAS state 17. Bevolo AJ, Shanks HR, Sidles PH, Danielson GC. Heat capacity
assignment. of hexagonal tungsten bronzes. Phys Rev B. 1974;9:3220–8. https​
://doi.org/10.1103/PhysR​evB.9.3220.
18. Sawada S. Thermal and electrical properties of tungsten oxide
­(WO3). J Phys Soc Jpn. 1956;11:1237–46. https:​ //doi.org/10.1143/
References JPSJ.11.1237.
19. Sawada S, Ando R, Nomura S. Thermal expansion and spe-
1. Thummavichai K, Xia Y, Zhu Y. Recent progress in chromogenic cific heat of tungsten oxide at high temperatures. Phys Rev.
research of tungsten oxides towards energy-related applications. 1951;84:1054–5. https​://doi.org/10.1103/PhysR​ev.84.1054.2.
Prog Mater Sci. 2017;88:281–324. https:​ //doi.org/10.1016/j.pmats​ 20. Tilley RJD. The crystal chemistry of the higher tungsten oxides.
ci.2017.04.003. Int J Refract Met Hard Mater. 1995;13:93–109. https​://doi.
2. Granqvist CG. Handbook of inorganic electrochromic materials. org/10.1016/0263-4368(95)00004​-6.
Amsterdam: Elsevier; 1995. 21. Szilágyi IM, Pfeifer J, Balázsi C, Tóth AL, Varga-Josepovits K,
3. Koffyberg FP, Dwight K, Wold A. Interband transitions of Madarász J, Pokol G. Thermal stability of hexagonal tungsten
semiconducting oxides determined from photoelectrolysis trioxide in air. J Therm Anal Calorim. 2008;94:499–505. https​://
spectra. Solid State Commun. 1979;30:433–7. https ​ : //doi. doi.org/10.1007/s1097​3-007-8601-y.
org/10.1016/0038-1098(79)91182​-7. 22. Zhuiykov S. Two-dimensional semiconductor nanocrystals:
4. González-Borrero PP, Sato F, Medina AN, Baesso ML, Bento new direction in science and technology. Nanostructured semi-
AC, Baldissera G, et al. Optical band-gap determination of nano- conductor oxides for the next generation of electronics and
structured ­WO3 film. Appl Phys Lett. 2010;96:061909. https:​ //doi. functional devices. Amsterdam: Elsevier; 2014. https​: //doi.
org/10.1063/1.33139​45. org/10.1533/97817​82422​242.139.
5. Lee H, Choi J, Lee S, Yun S-T, Lee C, Lee J. Kinetic enhance-
ment in photocatalytic oxidation of organic compounds by ­WO3

13
WO3 thermodynamic properties at 80–1256 K revisited

23. Vogt T, Woodward PM, Hunter BA. The high-temperature MN, Dorofeev SG, Ivanov VK. Ultrasonic disintegration of tung-
phases of W ­ O3. J Solid State Chem. 1999;144:209–15. https​:// sten trioxide pseudomorphs after ammonium paratungstate as a
doi.org/10.1006/jssc.1999.8173. route for stable aqueous sols of nanocrystalline ­WO3. J Mater Sci.
24. Perri JA, Banks E, Post B. Study of phase transitions in ­WO3 2018;53:1758–68. https​://doi.org/10.1007/s1085​3-017-1668-3.
with a high-temperature X-ray diffractometer. J Appl Phys. 37. Petříček V, Dušek M, Palatinus L, Petrícek V, Dušek M, Palatinus
1957;28:1272–5. https​://doi.org/10.1063/1.17226​31. L. Crystallographic computing system JANA2006: general fea-
25. Thummavichai K, Wang N, Xu F, Rance G, Xia Y, Zhu Y. In situ tures. Zeitschrift für Krist Cryst Mater. 2014;229:345–52. https​
investigations of the phase change behaviour of tungsten oxide ://doi.org/10.1515/zkri-2014-1737.
nanostructures. R Soc Open Sci. 2018;5:171932. https​://doi. 38. Wieser ME, Berglund M. Atomic weights of the elements 2007
org/10.1098/rsos.17193​2. (IUPAC technical report). Pure Appl Chem. 2009;81:2131–56.
26. Locherer KR, Salje EKH. The refinement of a tetragonal phase of https​://doi.org/10.1351/PAC-REP-09-08-03.
­WO3 using a novel PSD high temperature X-ray powder diffrac- 39. Szilágyi IM, Sajó I, Király P, Tárkányi G, Tóth AL, Szabó A,
tometer. Ph Trans. 1999;69:85–93. https​://doi.org/10.1080/01411​ Varga-Josepovits K, Madarász J, Pokol G. Phase transforma-
59990​82080​10. tions of ammonium tungsten bronzes. J Therm Anal Calorim.
27. Pokhrel S, Birkenstock J, Dianat A, Zimmermann J, Schowalter 2009;98:707–16. https​://doi.org/10.1007/s1097​3-009-0287-x.
M, Rosenauer A, Ciacchi LC, Mädler L. In situ high tempera- 40. Szilágyi IM, Madarász J, Hange F, Pokol G. Partial thermal reduc-
ture X-ray diffraction, transmission electron microscopy and tion of ammonium paratungstate tetrahydrate. J Therm Anal Calo-
theoretical modeling for the formation of W ­ O 3 crystallites. rim. 2007;88:139–44. https:​ //doi.org/10.1007/s10973​ -006-8078-0.
CrystEngComm. 2015;17:6985–98. https​: //doi.org/10.1039/ 41. Sale FR. Heat capacities of the tungsten oxides W ­ O3, ­W20O58,
C5CE0​0526D​. ­W18O49 and ­WO2. Thermochim Acta. 1979;30:163–71. https​://
28. Howard CJ, Luca V, Knight KS. High-temperature phase transi- doi.org/10.1016/0040-6031(79)85051​-0.
tions in tungsten trioxide—the last word? J Phys Condens Matter. 42. Kehl WL, Hay RG, Wahl D. The structure of tetragonal
2002;14:377–87. https​://doi.org/10.1088/0953-8984/14/3/308. tungsten trioxide. J Appl Phys. 1952;23:212–5. https​: //doi.
29. Locherer KR, Swainson IP, Salje EKH. Transition to a new org/10.1063/1.17021​76.
tetragonal phase of W ­ O3: crystal structure and distortion param- 43. Tanisaki S. Phase transition in tungsten trioxide. J Phys Soc Jpn.
eters. J Phys Condens Matter. 1999;11:4143–56. https​://doi. 1959;14:680–1. https​://doi.org/10.1143/JPSJ.14.680.
org/10.1088/0953-8984/11/21/303. 44. Sawada S, Ando R, Nomura S. On the ferroelectric curie point of
30. Hamdi H, Salje EKH, Ghosez P, Bousquet E. First-principles re- tungsten oxide. Phys Rev. 1951;82:952–3. https:​ //doi.org/10.1103/
investigation of bulk W ­ O3. 2016. http://arxiv.​ org/abs/1609.04207.​ PhysR​ev.82.952.2.
31. Rosen C, Banks E, Post B. The thermal expansion and phase 45. Ueda R, Ichinokawa T. On the phase transition of tungsten tri-
transitions of ­WO3. Acta Crystallogr. 1956;9:475–6. https​://doi. oxide. Phys Rev. 1951;82:563–4. https​://doi.org/10.1103/PhysR​
org/10.1107/S0365​110X5​60013​03. ev.82.563.2.
32. Bokov VA, Myl’nikova IE. Piezoelectric properties of new com- 46. Brækken H. Two higher phase transitions of W ­ O3. J Appl Phys.
pound single crystals with perovskite structure. Fiz Tverd Tela. 1966;37:3635–6. https​://doi.org/10.1063/1.17089​22.
1960;2:2728–32. 47. Sawada S. Microscopic and X-ray studies on tungsten oxide
33. Leute V. Das Wolframtrioxid und seine Reaktion mit den Oxiden ­(WO3). J Phys Soc Jpn. 1956;11:1246–52. https:​ //doi.org/10.1143/
zweiwertiger Metalle. Zeitschrift für Phys Chemie. 1966;48:307– JPSJ.11.1246.
18. https​://doi.org/10.1524/zpch.1966.48.5_6.307.
34. Sawada S. Thermal and electrical properties and crystal structure Publisher’s Note Springer Nature remains neutral with regard to
of tungsten oxide at high temperatures. Phys Rev. 1953;91:1010– jurisdictional claims in published maps and institutional affiliations.
1. https​://doi.org/10.1103/PhysR​ev.91.1010.
35. Sawada S, Ando R. Study on the ferroelectricity of tungsten oxide.
II. Thermal properties at high temperatures. Tokyo Univ Rep Inst
Sci Technol. 1951;5:143–8.
36. Shekunova TO, Baranchikov AE, Yapryntsev AD, Rudakovskaya
PG, Ivanova OS, Karavanova YA, Kalinina MA, Rumyantseva

13

You might also like