Effect of hydrogen dilution on the deposition rate of hydrogenerated amorphous silicon films in a modified pulsed plasma discharge

C. Mukherjee, C. Anandan, Tanay Seth, P. N. Dixit, R. Bhattacharyya
1996 Applied Physics Letters  
Hydrogen dilution effects on the growth of a-Si:H films in a modified pulsed plasma discharge are studied for two different silane flow conditions with hydrogen dilution ranging from 0% to 80%. The increase of deposition rate (r d ) due to hydrogen dilution is attributed mainly to changes in electron density (n e ) and electron decay time constant ( e ). Concurrently, it appears that hydrogen dilution mitigates the deleterious effects of secondary plasma reactions. Increasing the dwell time of
more » ... he high power period under hydrogen dilution also increases r d , accompanied by an enhancement of photoconductivity ( ph ) as compared to films grown under similar conditions without dilution. The decrease in r d at dilution Ͼ25% is seen as the onset of a process of etching, presumably by atomic hydrogen, and this limits r d irrespective of the availability of the number of silane molecules at such dilutions. Time-resolved optical emission spectroscopy results for H* emission supports the above views. © 1996 American Institute of Physics. ͓S0003-6951͑96͒04206-9͔ In the continuing search for device quality a-Si:H material that could also be grown at high deposition rates, many innovations have been proposed and tested. 1, 2 In this context, pulsed plasma methods 3-6 have been investigated. The attractive features of this technique that have been cited are ͑i͒ basic equipment and system configurations remain the same as conventional rf PECVD; ͑ii͒ the technique lends itself to a variety of manipulations ͑digital processing for instance͒; and ͑iii͒ uniformity of deposition is generally improved in this technique. In our initial experiments with a modified pulsed plasma discharge to obtain higher growth rates, 7 interestingly, it was found that at very low modulation frequency, for rf 13.56 MHz excitation, deposition rates (r d ) were less than those obtained by continuous wave ͑cw͒ discharge. Taking a clue from Fleddermann et al.'s 8 work, it was thought that a lower electron density (n e ) and low electron decay time constant ( e ) in undiluted silane, due to electron attachment to silane dissociation products, may be the reason for such low r d values. It may be noted that in the above-cited reference, Fleddermann et al. 8 investigated whether dissociative attachment to hydrogen molecule, a by-product of silane discharge, can be a factor contributing to an appreciable amount of electron loss. During these experiments, they found that for a discharge in a mixture of molecular hydrogen and helium, e increases, which is in marked contrast to SiH 4 /He discharges. Again, it was found that when H 2 was added in the SiH 4 /He mixture, e increased. These authors finally concluded that dissociative attachment to hydrogen may not be a significant process, and perhaps H 2 addition increases e . Thus, anticipating a higher n e and longer e in SiH 4 /H 2 discharges compared to an undiluted silane discharge, a modified pulsed plasma discharge of the SiH 4 /H 2 mixture was carried out systematically with an aim to achieve higher growth rates. Further, in order to understand the basic kinetics involved in the deposition process, two different schemes of dilution were tested. In one scheme, the total flow of the mixture was maintained constant ͑40 sccm͒ while changing the flow of the diluting gas. In the other scheme, the silane flow was maintained constant flow ͑20 sccm͒ and different amount of dilutions were used while keeping other parameters unchanged. Time-resolved optical emission spectroscopy ͑TROES͒ studies were also carried out to see whether excited species can give some insight into the kinetics involved. The a-Si:H films were deposited at 0.5 Torr in a capacitively coupled PECVD system ͑Multizone PECVD system, GSI, USA͒. In contrast to all earlier pulsed plasma work, a nonzero low power level was maintained throughout the discharge in our modified pulsed plasma discharge. 7 Thicknesses were determined by step height measurement equipment ͑Talystep, Rank Tailor Hobson͒. Average thickness values were used for plotting the data points. TROES studies were carried out using a combination of monochromator, UV-NIR photodiode, and a boxcar averager ͑single point analysis mode͒. Figure 1 shows the variation of deposition rate ͑r d , Å/s͒ with dwell time ͑, ms͒. For this study, the pulse power levels used were 75 W high power level ͑HPL͒ and 10 W low power level ͑LPL͒. These parameters were chosen because a͒ Electronic mail: npl@sirnetd.ernet. in FIG. 1 . Deposition rate as a function of dwell time for undiluted and H 2 diluted silane discharge. 835
doi:10.1063/1.116549 fatcat:lekjbvye7jd5td3okolqv3upaa