Eur. Phys. J. B 48, 47–53 (2005) DOI: 10.1140/epjb/e2005-00379-8 T HE EUROPEAN P HYSICAL JOURNAL B Stress in polycrystalline GaN films prepared by r.f sputtering M. Pal Chowdhury, R.K. Roy, S.R. Bhattacharyya, and A.K. Pal a Department of Materials Science, Indian Association for the Cultivation of Science, Calcutta-700 032, India Received 25 June 2005 Published online 9 December 2005 – c  EDP Sciences, Societ`a Italiana di Fisica, Springer-Verlag 2005 Abstract. Undoped, Be-doped and Si-doped polycrystalline GaN films were deposited by R.F. sputtering onto fused silica substrates. The films were deposited at various deposition temperatures ranging from 300 K to 623 K and characterized by optical measurements while the microstructural information was obtained from SEM and XRD studies. The compositional study for the GaN film was carried out using SIMS. Residual stresses in these films were evaluated from the band tail of the absorption spectra as well as from direct measurements of hardness by commercially available depth sensing indentometer. It was observed that undoped GaN films had the highest hardness followed by that for Be-doped and Si-doped films. The values of hardness obtained form the above optical measurement tallied quite well with those obtained from direct indentation measurement. PACS. 81.05.Ea III-V semiconductors – 78.20.-e Optical properties of bulk materials and thin films – 62.20.-x Mechanical properties of solids 1 Introduction The use of polycrystalline semiconductors in general had attracted much interest in an expanding variety of applica- tions in electronic and opto-electronic devices [1–3]. The main technological interest in the polycrystalline based devices lie in its very low-cost production and possibility of using low-cost substrates. Survey of literature indicates that not very many studies are reported so far on GaN in polycrystalline form. Only in recent years, some groups explored the feasibility of obtaining polycrystalline GaN layers [4-6]. The devices based on GaN epilayers suffered a setback due to the presence of biaxial strain compo- nent originating from the growth on lattice mismatched substrates with different thermal expansion coefficients and hydrostatic strains originating from incorporation of point defects altering the material’s lattice constant [7,8]. Polycrystalline semiconductor films are known to contain residual stress of the order of 10 9 –10 11 Nm -2 . The na- ture of the stress (tensile or compressive) depends on the deposition temperature and the sign of the misfit fac- tor between the substrate and the film [9–12]. The stress field results in higher concentration of dislocations or de- fects at the grain boundary. The presence of these dis- order or defect states would cause a random spatial po- tential fluctuation (together with band gap fluctuation) which in turn would provide an excess optical absorption in the below band gap region. Therefore, if the amount of defect states present in the grain boundary region of a e-mail: msakp@iacs.res.in a ploycrystalline film could be obtained from the below- band-gap optical absorption traces, one may estimate the corresponding stress or strain and hence the microhard- ness of the film [13–16]. In this communication, we present the synthesis of polycrystalline GaN films (doped and undoped) deposited onto fused silica (quartz) substrates kept at different tem- peratures by r.f. sputtering technique and evaluate the stress present in these films from the modification of the band edge absorption due to the presence of inherent elec- tric field and mechanical stress in the grain boundary re- gions. 2 Experimental details Polycrystalline GaN in thin film form was deposited onto fused silica substrates by r.f. sputtering of a GaN target (99.999%) target in argon plasma at a system pressure of ∼0.5 Pa. p- and n-doped GaN films were also deposited by using targets containing 1 at % Be and 1 at % Si in GaN respectively. The films were deposited at four differ- ent substrate temperature (T s ∼ 300, 423, 523 and 623 K) and for a fixed deposition time of ∼2 h. The substrates were placed on a heavy circular copper block that could be heated by appropriate heating coils passed through holes laterally drilled through the copper block. The tempera- ture of the substrates could be monitored and controlled by a copper-constant thermocouple by an on/off electronic temperature controller. Before starting the actual deposi- tion, the target was pre-sputtered with a shutter located in