Conduction mechanism in vanadium pentoxide xerogel films

HAL Id: jpa-00210458

https://hal.archives-ouvertes.fr/jpa-00210458

Submitted on 1 Jan 1987

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Conduction mechanism in vanadium pentoxide xerogel films

K. Bali, L.B. Kiss, T. Szörényi, M.I. Török, I. Hevesi

To cite this version:

K. Bali, L.B. Kiss, T. Szörényi, M.I. Török, I. Hevesi. Conduction mechanism in vanadium pentoxide xerogel films. Journal de Physique, 1987, 48 (3), pp.431-434. �10.1051/jphys:01987004803043100�.

�jpa-00210458�

Conduction mechanism in vanadium pentoxide xerogel ^films

K. Bali, ^{L. B.} Kiss, ^T. Szörényi, ^{M. I.} Török and I. Hevesi

Research Group

Luminescence and Semiconductors of the Hungarian Academy ^{of Sciences,}

H-6720 Szeged, ^Dóm tér 9., Hungary

(Requ le 27 octobre 1986, accepté le 12 novembre 1986)

Résumé.

La dépendance temporelle de la conductivité

courant continu et les caractéristiques courant-tension des films de pentoxide de vanadium de type xerogel d’épaisseur ^{variant entre 5 et 1 000 nm, montrent} que la conduction est purement électronique ^dans

système. ^A partir ^des

^{de bruit}

1/f

estime que la borne

supérieure de la concentration des porteurs ^est de l’ordre de 107 à 6 1018 cm-3,

qui ^montre

^faible dépendance

fonction de l’épaisseur. Puisque

^{chiffres sont} approximativement de deux ordres de grandeur supérieurs ^à

obtenus dans le V2O5 monocristallin, mais aussi deux ordres de grandeur plus faibles que la concentration de V4+ dans les films

les monocristaux,

^résultats suggèrent que la plus grande conductivité des

xerogels doit être associée à l’hydratation ^{des films} qui augmente le nombre de sites actifs V4+ dans le processus de conduction.

Abstract.

The time dependence of the d.c. conductivity ^{and the} current-voltage characteristics of vanadium

pentoxide xerogel films of thicknesses ranging between 5 and 1 000

have revealed that the conduction in this system ^is purely electronic. From 1/f-noise measurements

upper limit of 1017 ^to 6 1018 cm-3 for the charge

carrier concentration could be estimated which showed weak thickness dependence. ^These figures

approx. ^two orders of magnitude higher than those of single crystal V2O5, ^{and also two orders of} magnitude lower than the

V4+ concentration in either the films

the single crystals. ^This suggests ^{that the} higher conductivity ^of xerogels ^is

associated with film hydration which increases the number of V4+ sites active in the conduction process.

Classification

Physics ^Abstracts

73.60F

1. Introduction.

Since the pioneering ^{work of} Livage and his co-workers in the late 70 s [1, 2] ^the physical-chemistry of transition metal oxide gels ^{has been} extensively ^studied mainly by

French groups [3-21]. ^{From the} viewpoint of semicon- ductor physics ^{one of the most} challenging ^{features of}

thin films deposited from vanadium pentoxide gels ^is

their surprisingly high electrical conductivity [3, ^{4, 9,} 10, 17, 19, 20].

Bullot et al. measured d.c. conductivity (u) ^values ranging ^{from 2 to 100} S/m ^for layers deposited ^from V205 gels ^{of c}

V4 ’ / (V4 + + ^{V5 + )=} 0.013-0.097 at room temperature [3, ^4, 20]. ^The gel ^{was obtained} by

fast quenching the molten oxide in water. Bullot and his co-workers observed only ^2-3-fold changes ^{in u as}

the water vapour pressure ^was changed. These authors concluded that (i) ^the conductivity ^was essentially

determined by ^{the V4}

^content of the material and (ii)

the semiconducting properties of the films were consis- tent with the small polaron model. Sanchez and his

coworkers reported very similar a-values (15-60 S/m)

for layers deposited ^from gels ^{made of} polycondensed

decavanadic acid (typically ^c

^{0.01 ;} [10, 17]). ^From

the temperature dependence ^{of the d.c. and a.c.}

conductivity ^{at various} humidity, ^{Barboux et al. con-}

cluded that polyvanadic ^acid gels ^of V205 . n H20

nominal composition (0.5 -- n -- 1.7) ^{were mixed con-} ductors, in which the ionic contribution to the conduc-

tivity preferentially depended ^{on the amount of water}

in the material [19]. ^The analysis ^{of the dielectric} spectra ^of V205 .1.6 H20 xerogels ^formed by polycon-

densation of vanadic acid led Badot et al. to the conclusion that the mechanism of electrical conduction

was hopping ^{of H+} protons ^{in this} system [22].

Our recent experiments ^on xerogel films made of CVD vanadium pentoxide [23] lent further support ^to the notion that the conductivity in these materials is

essentially governed by ^{the amount} of intercalated water [24, 25]. ^{However, the} question ^{arose how}

intercalated water can influence the conduction _process.

In the present paper ^we report ^{novel results on the}

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:01987004803043100

432

time and thickness dependence of the d.c. current-

voltage characteristics and, ^from 1 / f ^noise

measurements, give ^an estimate for the upper limit of the free charge carrier concentration. We suggest ^that in xerogel films made of CVD vanadium pentoxide ^the

conduction takes place via electron transport ^{with no} significant ionic contribution. The relatively high ^con- ductivity ^{is due to} solvation which increases the number of active V4 + sites.

2. Experimental ^methods.

Vanadium pentoxide gel ^was prepared by dissolving amorphous V205 ^{in water. The} amorphous V205 ^was prepared by ^chemical vapour deposition (CVD) ^as previously ^described [23]. The colloidal stock solution contained 0.01-0.05 M vanadium with a V4 + content of

c

0.005 ± 0.001 as determined by titrimetric method

[26]. ^Diluted aliquots ^were deposited ^onto microscope

slides which were coated with vacuum evaporated platinum electrodes in coplanar ^geometry (electrode

gap 0.1 mm). ^The samples ^{were dried at room} tempera-

ture in an exsiccator over silica gel. ^Film thicknesses

(typically between 5 and 1 000 nm) ^{were measured} by ^a Talystep mechanical stylus instrument (Rank Taylor

Hobson Ltd.). Electrical conductivity ^and 1/ f ^noise

measurements, ^{in the} frequency range of 20 Hz and 10 kHz, ^were performed in air of controlled water vapour pressure.

3. Results and discussion.

3.1 CURRENT-VOLTAGE CHARACTERISTICS.

While the samples ^were transferred from the exsiccator to the measuring ^{chamber a} gradually increasing ^con- ductivity ^was observed, i.e. under constant d.c. voltage (-- 1.0 V) ^{there was a} steady increase in the current

flowing through ^the film. This reversible gradual change continued until the sample conductivity ap-

proached ^the steady-state ^value corresponding ^to

8.8 torr water vapour pressure ^at 293 K. After the

steady-state ^{had been} achieved (a between 1 and 5 S/m depending ^{on the film thickness} [24]) ^the current-voltage characteristics were recorded between - 10 and + 10 V. At low bias, ^{when the current} flowing through ^{the film was less than 250-300 f.LA, a linear} dependence ^was always ^obtained (curve ^{a in} Fig. 1). ^As

the current exceeded this critical value the current-

voltage ^curves displayed ^a nonlinear behaviour (curves

b and c in Fig. 1). ^For samples ^{with a} thickness of less than 100 nm linear I-U curves were obtained in the entire measuring voltage range (curve ^{a in} Fig. 2). ^This

linear _response confined to a narrowing voltage ^interval

with increasing film thicknesses (curves ^{b and c in} Fig. 2). ^This phenomenon ^{could be} anticipated recalling

the switching characteristics of these films. Bullot et al.

[27] reported ^threshold voltages ^{of 17.5 and 23-24.5 V}

at room temperature ^for films of active areas of 0.2 and

Fig.l.2013d.c. current-voltage characteristics of vanadium

pentoxide xerogel ^films

function of layer ^thickness.

Fig. ^2.

M /AU plots for vanadium pentoxide xerogel ^films

of different thicknesses. The symbols correspond ^{to those of} figure ^1.

0.5 mm2 respectively. ^Our samples ^{had an active area}

of 1 mm2. For films of 0.3-0.6 _fJ.m thickness (which

were comparable ^with those of Bullot et al.) ^{the critical}

current densities varied between 2 x 105 and 5 x 105 A/m2. These values were consistent with the current density ^{in the} preswitching region ^calculated

from figure ² of reference [27] : - ^{1.7 x 105} A/M2. ^On

the basis of these results the deviation from linearity

could be attributed to Joule-heat effect.

After completing ^the current-voltage characteristics the sample ^{was biased} by ^{a constant d.c.} voltage ^chosen

to be close to the limit of the linear response interval.

The long-term ^{time course of the current} changes ^were

followed for several days. ^{No current} drop ^{could be}

observed _{in any case ;} in fact, ^a slight (a ^few percent)

increase was obtained in the case of thin samples (d

²⁰⁰ nm) ^{which was a} consequence ^of spontaneous reduction taking place ^at the film surface [24]. ^Neither polarization ^{effects nor colour} changes ^at the electrodes could be detected during ^the experiments. These results

provide evidence that the conduction of xerogel ^films

made of CVD vanadium pentoxide is electronic in nature and exclude any significant ionic contribution.

3.2 1/ f-NOISE MEASUREMENTS. - In spite ^{of the} varying conductivity during hydration/dehydration cy- cles the thermal activation energies ^remain practically

unaffected [25] suggesting that it is the charge ^carrier concentration, ^{but not} the conduction mechanism which is affected by hydration. ^For clarification of this essential point 1 / f -noise experiments ^were performed.

The low frequency part ^{of the} conductivity ^fluctu-

ation in solids is dominated by ^{the bulk} 1/ f -noise.

Although ^an indisputable theory ^of 1 / f type mobility

fluctuation does not exist [28] ^the Hooge-Vandamme

relation [29] has been found to hold in a large ^number

of experiments [28, 30] :

where the following notation is used : R is the resistance of the sample, ^AR the measured resistance fluctuation,

JL the mobility, n ^{the free} charge ^carrier concentration, Veff the effective volume, i.e. the real volume of the

sample ^{if the current} density ^is homogeneous, f ^the frequency ^and

^1. aH and JL L ^are the noise factor and the mobility ^{in the case} of pure lattice scattering.

Recalling that the maximum value of the noise factor,

a H is of the order of 10-3 [30], ^the Hooge-Vandamme

relation permits ^{a direct} estimation of the upper limit of the free charge ^carrier concentration (for ^{details see}

Ref. [31]).

In figure 3, the upper limits of the free charge ^carrier concentration, n calculated from equation (1) ^with

a H = 2 x 10- 3 ^are plotted ^{for a} typical ^{series of} samples. The n values determined (between ^{1017 and}

6 x 1018 cm- 3) ^are approximatively ^{two orders} of mag- nitude higher ^{than that of} single crystal V205 ^derived

from 1/ f -noise measurements of the same type

(1.4 ^{x 1016 cm- 3 at room} temperature [32]). ^{The room}

temperature conductivity of the films deposited ^from gels ^{was also two orders of} magnitude higher ^{than the} single crystal value, while the thermal activation ener-

gies ^were practically ^identical [25]. ^{Thus, the} 1/ f-noise

measurements have provided independent ^evidence

Fig. ^3. - Upper limits of the charge carrier concentration in vanadium pentoxide xerogel ^films derived from 1/ f-noise

measurements

function of layer ^thickness.

that the conductivity increase of xerogel films is exclu-

sively ^{due to an} increased free charge ^{carrier concen-}

tration without significant changes ^{in the conduction}

mechanism. On the other hand, it should be noted that the upper limits shown in figure ^{3 account for} only approximatively 1 % of the reduced vanadium (y4 + )

concentrations both in xerogels ^and single crystals (-2 ^{x lOZo} cm- 3). Consequently the number of free electrons must be far below the total number of

V4 + _{in any} cases.

4. Conclusions.

The ohmic character of the current-voltage ^character-

istics in the broad voltage range studied and the absence of any decrease in the conductivity with time is consistent with the pure electronic ^nature of the conduction in xerogel ^films (made ^{of CVD vanadium} pentoxide) ^{in an} atmosphere ^{of 8.8 torr water} vapour

partial pressure ^at 293 K, ^{with no} ionic contribution to the conduction process. 1/ f -noise measurements have

provided additional direct evidence that the free charge

carrier concentration in these xerogel ^films is approx-

imatively ^{two orders of} magnitude higher ^{than in} single crystal V205 ; ^this charge carrier concentration accounts

only ^for approximatively ^{1 %} of the total V4 + content of the sample. ^{It is} suggested that all these results on

xerogel ^{films versus} single crystal ^and amorphous ^films

can be reconciled by assuming ^that hydration ^increases

the number of V4 + ions active in the conduction process, i.e. the number of charge carriers. This

interpretation is further substantiated by ^{recent EPR} experiments of Araki et al. [33] ^who demonstrated that the V4 + sites are solvated by strongly ^{bonded water}

molecules.

434

References

[1] MICHAUD, M., LEROY, M. C. and LIVAGE, J., Mater. Res. Bull. 11 (1976) ^1425.

[2] LIVAGE, ^J., GHARBI, N., LEROY, ^{M. C. and} MICHAUD, M., Mater. Res. Bull. 13 (1978) ^1117.

[3] BULLOT, J., GALLAIS, O., GAUTHIER, ^{M. and} LIVAGE, J., Appl. Phys. ^{Lett. 36} (1980) ^986.

[4] BULLOT, J., CORDIER, P., GALLAIS, O., GAUTHIER,

M. and LIVAGE, J., Phys. ^{Status Solidi A 68} (1981) ^357.

[5] LIVAGE, J., ^J. Physique Colloq. ⁴² (1981) ^C4-981.

[6] ALDEBERT, P., BAFFIER, N., GHARBI, ^{N. and} LIVAGE, J., Mater. Res. Bull. 16 (1981) ^669.

[7] ALDEBERT, P., BAFFIER, N., GHARBI, ^{N. and} LIVAGE, J., Mater. Res. Bull. 16 (1981) ^949.

[8] ALDEBERT, P., BAFFIER, N., LEGENDRE, ^{J. J. and} LIVAGE, J., ^Rev. Chim. Miner. 19 (1982) 485.

[9] LIVAGE, ^{J. and} LEMERLE, J., Ann. Rev. Mater. Sci.

12 (1982) ^103.

[10] GHARBI, N., SANCHEZ, C., LIVAGE, J., LEMERLE, J., NEJEM, ^{L. and} LEFEBVRE, J., Inorg. ^Chem.

21 (1982) ^2758.

[11] ^{LEGENDRE, J.-J. and LIVAGE, J., J. Colloid} Interface

Sci. 94 (1983) ^75.

[12] LEGENDRE, J.-J., ALDEBERT, P., BAFFIER, ^{N. and} LIVAGE, J., ^{J. Colloid} Interface ^{Sci. 94} (1983)

84.

[13] ILYASOVA, ^{A. K.,} GESKINA, R. A., TARABRIN, G. K., OMAROVA, N. D., EKSHETELIS, ^{A. V.}

and BELKIN, ^A. S., Zh. Neorg. ^{Khim. 28} (1983)

279.

[14] ^{ALDEBERT, P. and} PAUL-BONCOUR, V., Mater. Res.

Bull. 18 (1983) ^1263.

[15] BOUHAOUSS, A. and ALDEBERT, P., Mater. Res.

Bull. 18 (1983) ^1247.

[16] ARAKI, B., MAILHÉ, C., BAFFIER, N., LIVAGE, ^J.

and VEDEL, J., Solid State Ionics 9-10 (1983)

439.

[17] SANCHEZ, C., BABONNEAU, F., MORINEAU, R., LIVAGE, ^{J. and} BULLOT, J., ^Philos. Mag. ^{B 47} (1983) ^279.

[18] ^{ABELLO, L. and} POMMIER, C., ^{J. Chim.} Phys. Phys.

Chim. Biol. 80 (1983) ^373.

[19] ^BARBOUX, Ph., BAFFIER, N., MORINEAU, ^{R. and} LIVAGE, J., Solid State Ionics 9-10 (1983) ^1073.

[20] BULLOT, J., CORDIER, P., GALLAIS, O., GAUTHIER,

M. and LIVAGE, J., J. Non-Crystalline ^{Solids 68} (1984) ^123.

[21] BULLOT, J., CORDIER, P., GALLAIS, O., GAUTHIER,

M. and BABONNEAU, F., ^J. Non-Crystalline

Solids 68 (1984) ^135.

[22] BADOT, ^J. C., FOURRIER-LAMER, ^{A. and} BAFFIER, N., J. Physique ⁴⁶ (1985) ^2107.

[23] SZÖRÉNYI, T., BALI, ^{K. and} HEVESI, I., ^{J. Non-} Crystalline Solids 35-36 (1980) ^1245.

[24] SZÖRÉNYI, T., BALI, K., TÖRÖK, M. I. and HEVESI, I., Thin Solid Films 121 (1984) ^29.

[25] SZÖRÉNYI, T., BALI, ^{K. and} HEVESI, I., ^J. Physique

46 (1985) ^473.

[26] DIKSHITULU, L. S. A. and GOPALA RAO, G., ^Talanta

9 (1962) ^857.

[27] BULLOT, J., GALLAIS, M., GAUTHIER, ^{M. and} LIVAGE, J., Phys. Status Solidi A 71 (1982) ^K1.

[28] ^{HOOGE, F.} N., KLEINPENNING, T. G. M. and VAN- DAMME, L. K. J., Rep. Prog. Phys. ⁴⁴ (1981)

479.

[29] HOOGE, ^{F. N. and} VANDAMME, ^{L. K.} J., Phys. ^Lett.

66A (1978) ^315.

[30] CLEVERS, ^{R. H.} M., ^{Noise in} Physical Systems ^and 1/f Noise, Eds. : A. d’Amico and P. Mazzetti

(North-Holland, Amsterdam) ^1986.

[31] ^{KISS, L. B., BALI,} K., SZÖRÉNYI, ^{T. and} HEVESI, I., Solid State Commun. 58 (1986) ^609.

[32] HEVESI, I., KISS, ^L. B., TÖRÖK, M. I. and NANAI, L., Phys. Status Solidi A 81 (1984) ^K67.

[33] ARAKI, B., AUDIERES, J. P., MICHAUD, ^{M. and} LIVAGE, J., Bull. Soc. Chim. Fr. 9-10 (1981)

366.