Physica B 307 (2001) 105–110

Ageing effect of Sb2Te3 thin films

P. Arun*, Pankaj Tyagi, A.G. Vedeshwar, Vinod Kumar Paliwal

Department of Physics and Astrophysics, University of Delhi, Delhi 110 007, India

Received 8 March 2001; received in revised form 6 August 2001

Abstract

Post deposition variation in film resistance of Sb2Te3 films deposited on glass substrates at room temperature and an

elevated temperature (1101C) are investigated. The resistance of films grown at room temperature shows a non-lineardecrease with time which is thickness dependent as opposed to the increasing resistance of film grown at elevatedtemperature. The decreasing resistance with time can be attributed to the transformation of an amorphous phase of the

as-grown film to a micro-crystalline phase as revealed by X-ray diffraction. The increasing resistance was found to bedue to the surface oxidation (Sb2O3) and a diffusion as a function of time. However, the underneath layer of Sb2Te3

below the top Sb2O3 layer remains amorphous even after 2 years from the date of fabrication. r 2001 Published by

Elsevier Science B.V.

PACS: 73.61; 73.61.G; 81.40.C

Keywords: Thin films; Chalcogenides; Ageing materials

1. Introduction

Post deposition variation in the physical proper-ties of thin films as a function of time is calledageing. Such post deposition variation has beenreported in various films, like silver, copper [1] andCdSexTe1�x [2,3] etc. In these studies the postdeposition film resistance was found to increasewith time and then saturate. Many theories havebeen put forward to explain this increase in filmresistance. Agglomeration of the islands withinthe film, where coalescence is the force driving theagglomeration process, has been considered thecause of increase in film resistance. The increase in

resistance was shown to be directly proportional totime (lnRðtÞ=Rð0Þpln t). The conduction mecha-nism across the voids between the islands/grains isconsidered to be either due to quantum tunnellingor electron emission [4]. Fehlner [5], Erhlich [6]and Despande [7] assumed the oxidation of themetal islands’ surface as the cause for theincreasing resistance. This causes increase in theaverage inter-island spacing and a change in workfunction which leads to a decrease in tunnellingprobability, thus leading to an increase in the filmsresistance with time. Morris [8] however explainedan increase in resistance due to a reduced electronemission due to a decrease in film temperature. Inshort, the literature is full of examples where theresistance of the film increases with time. Theproperties of Sb2Te3 films also show a variationwith time, immediately after deposition. However,

*Corresponding author. Tel.: +91-11-7667-793; fax: +91-

11-7667-061.

E-mail address: agni@physics.du.ac.in (P. Arun).

0921-4526/01/$ - see front matter r 2001 Published by Elsevier Science B.V.

PII: S 0 9 2 1 - 4 5 2 6 ( 0 1 ) 0 1 0 3 2 - 8

it is seen that in films grown on glass substrates atroom temperature, the resistance falls with time.Films of different materials seem to have uniqueprocess leading to ageing. Various peculiaritiesseen during the ageing of Sb2Te3 films havebeen reported in this manuscript. Therefore, itprompted us to study ageing effect of Sb2Te3 filmsin detail as reported here.

2. Experimental

Thin films of Sb2Te3 were grown on glasssubstrates kept at room temperature, using ther-mal evaporation method. Sb2Te3 ingot of highpurity (99.99%) supplied by Aldrich (USA) wereused as the starting material. The crushed ingotwere evaporated from molybdenum boat at avacuum better than 10�6 Torr. The film thicknesswas measured using Dektek IIA surface profilerwhich uses the method of a mechanical stylusmovement on the surface. The movement of thestylus across the edge of the film determines thestep height or the film thickness. The film thicknesswas found to be uniform over the area 5 cm� 5 cmwith an error of 3% at the edges.

Before the glass substrates were placed in thechamber, indium contacts were grown on them. Astrip of Sb2Te3 film of dimensions 2.3 cm�

1.65 mm could be fabricated on the pre-grownindium contacts using a mask. The I–V character-istics of the films were measured by four probemethod. It was found to be linear between 25 mV–24 V, showing the ohmic nature of indiumcontacts. The variation in films’ resistance withtime were measured by a digital multimeter. Thevariation in resistance of films grown at roomtemperature were done both in vacuum and atatmospheric pressure. However, variation in re-sistances of films grown on heated substrates wasstudied only at atmospheric pressure. The struc-tural and compositional analysis of these filmswere done using Phillips PW1840 X-ray diffracto-meter and Shimadzu ESCA750 (Electron Spectro-scopy for Chemical Analysis). The films werefound to be stoichiometrically uniform over thearea 5 cm� 5 cm as determined by ESCA carriedout in various regions of the film.

3. Results and discussion

We have observed a quite different nature ofageing effect for films grown at room temperatureand elevated temperatures. The resistance of as-grown films decreases with time non-linearly andshows film thickness dependence as shown inFig. 1. However, the resistance of films grown at

Fig. 1. The variation in resistance of Sb2Te3 thin film with passage of time immediately after deposition shown for three different film

thickness (a) 500 nm, (b) 250 nm and (c) 160 nm.

P. Arun et al. / Physica B 307 (2001) 105–110106

elevated temperature (1101C) shows an initiallinear increase and then a parabolic increase withtime. We will discuss it separately in contrast withthat of at room temperature. Coming to the caseof films grown at room temperature, the decreasein resistance with time is irrespective of whetherthe film was kept in vacuum or in air at ambient.The decrease in resistance with time immediatelyrules out the possibility of any oxidation evenwhen film is exposed to air which has beenconfirmed by ESCA. Almost all earlier studies[4–7] invoke the idea of oxidation in the film toexplain the increasing resistance with time. This isquite probable in case of metal films and somecompound’s films containing few unreacted che-mical species. However, the decreasing resistancewith time is hardly reported in literature.

The thickness dependent behaviour of thedecreasing film resistance with time is illustratedin Fig. 1. We can notice few things clearly. First ofall, the sharp decrease in resistance commencesafter some time of initial slow decrease which isthickness dependent. Thinner the film, earlier is thecommencement of the sharp decrease. Secondly,the duration of sharp decrease also shows thick-ness dependence. This duration increases with filmthickness. Thirdly, after the sharp decrease thefilms resume again a slowly decreasing behaviourand saturate after a long time. The saturation timeshows a linear dependence with time as displayedin Fig. 2.

The film resistance or in other words resistivitymeasured either immediately after deposition or

after saturation shows similar systematic beha-viour with film thickness as shown in Figs. 3 and 4,respectively. We have carried out the structuralanalysis of the films as a function of time by X-raydiffraction to get any clue regarding the decreasingresistance of the films. A film of thickness 380 nmwas selected, since the larger the thickness, sloweris the variation in resistance with time. We haveshown the results of X-ray diffraction as a functionof time in Fig. 5.

All the as-grown films were amorphous withoutexception as revealed by diffractogram (a) inFig. 5. Till the diffractogram (d) there is no muchdirect revelation of structural changes. The micro-crystalline nature of the film can clearly be realisedby the appearance of a broad peak in diffracto-gram (d) of Fig. 5, which was taken after the filmresistance saturated. However, a close look at the

Fig. 2. The time taken for the resistance of Sb2Te3 thin film to

saturate as a function of film thickness.

Fig. 3. The resistivity of as-grown film as a function of film

thickness.

Fig. 4. The saturated resistivity of as-grown Sb2Te3 film as a

function of film thickness.

P. Arun et al. / Physica B 307 (2001) 105–110 107

diffractograms would indicate the evolution ofmicro-crystalline phase from an amorphous phase.Therefore, we believe, the decreasing film resis-tance with time is mainly due to the evolution ofmicro-crystallinity from an amorphous phase.However, the onset and end point of the sharpdecrease in the film resistance of Fig. 1 could notbe identified explicitly in the structural analysis.

The evolution of micro-crystalline phase startsfrom the relatively unstable amorphous phasethrough the atomic re-arrangements correctingthe ‘wrong bonds’, normally existing even instoichiometric amorphous phase, at a nucleationcentre. Considering the various bond strengths, i.e.Sb–Sb (3.11 eV), Te–Te (2.67 eV) and Sb–Te(2.88 eV) in Sb2Te3 [9], we can see that thedifference between the average bond strengths of(Sb–Sb, Te–Te) and 2(Sb–Te) is nearly 20 meV,comparable to the thermal energy at roomtemperature. Therefore, the atomic arrangementcan take place by the thermal vibration of atomsresulting in ‘right bonds’ leading to a final micro-crystalline phase. The initial time lag for thecommencement of sharp decrease in RðtÞ showinga linear thickness dependence could be assumed tobe the initialisation time for the re-arrangementprocess. Further, the time duration for either thelinear sharp fall of RðtÞ or the saturation alsoshows a linear thickness dependence. This could beexplained if we assume the amount of ‘wrongbonds’ to be proportional to film thickness whichis quite reasonable. This kind of atomic re-

arrangement cannot be followed by a simple X-ray diffraction analysis. However, the change infilm resistance could easily be thought as due to abetter connectivity for electron conduction result-ing from atomic re-arrangements. Returning toFigs. 3 and 4, it should be noted that the nature ofresistivity’s variation with thickness is identical. Itis as if there is a significant background initially,which diminishes with time. The resistivity isexplained by Matthiessen’s additive rule [10],which states the contributions due to thermalresistivity, from point defects, dislocations etc. addup linearly

rðdÞ ¼ rTðdÞ þ rpðdÞ þ rdðdÞ þ?: ð1Þ

Since the change in resistance is attributed toincreasing crystallinity, it may be assumed that thedecrease in resistivity is due to the decrease incontributions from those of point defects anddislocations. However, since these terms, them-selves are thickness dependent, a quantitativeestimate is difficult. The variation in resistivitywith film thickness is inversely proportional to thefilm thickness and is in accordance to Mayadasequation [11].

The resistance was found to be increasing withtime linearly initially and then parabolically forfilms grown at an elevated temperature of 1101Cand exposed to air immediately after deposition asshown in Fig. 6. The X-ray diffractograms takenimmediately after fabrication, showed these sam-ples were amorphous in nature.

The increasing film resistance of a quite thickfilm can be suspected as due to oxidation anddiffusion in the film. Therefore, we have carriedout ESCA analysis, Fig. 7, on these films whichshows the surface oxidation. The oxidation can bedetected by the intensity ratio of antimony’s 3d3/2

and 3d5/2 peaks which is equal to 0.66 for pureantimony. Since Sb 3d5/2 and O 1s peaks mergedue to identical binding energies, the intensity ofthe undeconvoluted Sb 3d5/2 peak increases so thatthe ratio will decrease from 0.66, as explained byMorgan et al. [12]. We have observed the ratio todue to be 0.44 indicating oxidation. The absence ofTe peaks in the survey scan indicates the completeoxidation at the surface. Since ESCA is a surfacetechnique limited to about 5 nm depth [13], the

Fig. 5. X-ray diffractograms of as-grown Sb2Te3 film taken

after (a) 40 min, (b) 21 h, (c) 167 h and (d) 600 h.

P. Arun et al. / Physica B 307 (2001) 105–110108

estimate of the extent of oxidation and diffusion isnot possible by this technique. The oxidation anddiffusion of oxide in the film has been explained bySze [14]. The thickness of oxide layer at the surfaceincreases linearly with time initially. The chang-ing film resistance due to the initial oxidation isgiven by

RðtÞ ¼rl=wd

1 � Bt=Ad; ð2Þ

where B=A is called the linear rate constant, ameasure of the rate of oxidation. We get the linearrate constant as 50.16� 10�3 mm/h by fittingEq. (2) to the initial linear increase of film

resistance with time of Fig. 6. After the initialoxidation, the surface oxide layer acts as aprotective layer preventing further oxidation.However, the film resistance continues to increasedue to the diffusion of oxide into the depth of thefilm. The diffused layer thickness increases para-bolically due to which the film resistance variesaccording to

RðtÞ ¼rl=wd

1 �ffiffiffiffiffiBt

p=d

; ð3Þ

where B is called parabolic rate constant. Eq. (3)fits very well to the experimental data as shown inFig. 6 by the continuous curve yieldingB ¼ 10:1 � 10�6 mm2/h. We estimate the initialoxide layer thickness of about 25 nm and a totaldiffused layer thickness of about 61 nm in a 380 nmthick film using the above analysis.

The resistivity of films grown on heatedsubstrates were found to be only 10–55% of thatgrown at room temperature, with thinner filmsshowing larger variation as compared to theircounterparts which were grown at room tempera-ture. For example, the as-grown film of thickness380 nm had a resistivity of 400� 10�4Om, whichon ageing reduced to 4� 10�4Om, while film ofsame thickness grown on a heated substrate had aresistivity of 168� 10�4Om. However, the resis-tivity of films grown on heated substrates does notshow any systematic variation with thickness. The

Fig. 7. ESCA spectra of antimony’s 3d peaks.

Fig. 6. The variation in film resistance with time of Sb2Te3 film grown at substrate temperature of 1101C and exposed to air

immediately after deposition.

P. Arun et al. / Physica B 307 (2001) 105–110 109

resistivity of films grown on heated substratesbeing less than that of as-grown films furtherindicates that ordering contributes in the loweringof resistivity.

The XRD carried out on this film continues toshow amorphous nature of the film even after 2years since deposition. This means both Sb2O3 (asdetermined by ESCA) and Sb2Te3 at the bottomare in amorphous phase as X-ray penetrationdepth is quite high in diffraction experiments athigher angles. This fact we have shown in Fig. 8.Therefore, a protective oxide layer prevents Sb2Te3

to change over to micro-crystalline phase from anamorphous phase. This could well be due to therestriction of degree of freedom for the atomicvibration and hence the atomic re-arrangementleading to micro-crystalline phase as discussedearlier. Even the diffusion of oxide species intoSb2Te3 will enable the restriction of Sb2Te3 re-arrangement.

4. Conclusion

The ageing of as-grown Sb2Te3 films at roomtemperature shows a non-linear decrease inresistance due to the transformation of initialamorphous phase to micro-crystalline phase asrevealed by X-ray diffraction. This smooth trans-

formation of phases takes place with the atomic re-arrangements healing the wrong bonds. In con-trast, the ageing of the film grown at an elevatedtemperature showed an initial linear and a laterparabolic increase in film resistance due to surfaceoxidation resulting from the immediate exposureto air after deposition. The buried Sb2Te3 layerunder the surface Sb2O3 layer does not show anytransformation of amorphous to micro-crystallinephase even after 2 years since deposition. Thiscould be well due to the restriction imposed on theatomic re-arrangements of Sb2Te3 due to the inter-diffusion of Sb2O3.

Acknowledgements

We are grateful to Mr. Padmakshan, Depart-ment of Geology, University of Delhi, for carryingout the X-ray diffraction studies.

References

[1] V.D. Das, M.S.M. Sastry, M. Pattabi, Phys. Stat. Sol. A 96

(1986) 677.

[2] P.J. Sebastian, V. Sivaramakrishnan, Phys. Rev. B 65

(1989) 237.

[3] P.J. Sebastian, V. Sivaramakrishnan, Phys. Stat. Sol. A

124 (1991) 505.

[4] C.M. Hurd, J. Phys. C: Solid State Phys. 18 (1985) 6487.

[5] F.P. Fehlner, J. Appl. Phys. 38 (1967) 2223.

[6] G. Erhlich, J. Chem. Phys. 35 (1961) 2165.

[7] S.M. Despande, Thin Solid Films 5 (1970) 125.

[8] J.E. Morris, Vacuum 22 (1972) 153.

[9] David R. Lide (Ed-in-Chief), CRC Handbook of Chem-

istry and Physics, 81st Edition, CRC Press, Boca Raton,

FL, 2000.

[10] T. Schober, J.B. Condon, H. Trinkaus, J. Appl. Phys. 69

(1991) 2981.

[11] C.R. Tellier, A.J. Tosser, C. Boutrit, Thin Solid Films 44

(1977) 201.

[12] W.E. Morgan, W.J. Stec, J.R. Von Wazer, Inorg. Chem.

12 (1973) 953.

[13] D. Briggs, Handbook of X-ray and Ultra-violet Photon-

electron Spectroscopy, Heyden, London, 1977.

[14] S.M. Sze, Semiconductor Devices, Physics and Technol-

ogy, Wiley, New York, 1993.

Fig. 8. X-ray diffractogram of an Sb2Te3 film of Fig. 6 taken

after 2 years from the date of fabrication.

P. Arun et al. / Physica B 307 (2001) 105–110110