PHOTOLYSIS OF SIH BY THE THIRD HARMONIC OF A Nd : YAG LASER AT 355 NM

The photolysis of silane (Sill4) was carried out using the third harmonic of a Nd: YAG laser at 355 nm, at a fixed Sill4 pressure of350 Torr, varying the laser energy fluence in the range of30-300 Jcm2. The emission spectra indicates that the photofragments formed are Sill2, Sill, Si, H2, and H. The (A1B1-X1A1) transitions at 552.7 nm, 525.3 nm, 505.6 nm, and 484.7 nm of Sill are due to a two photon absorption process. The (A2A-X2n) transitions of Sill at 425.9 nm, 418 nm, 414.2 nm, 412.8 nm and 395.6 nm are due to a three photon absorption process. The brownish white deposit on the cell windows indicates the presence of amorphous silicon (a:Si-H). The two atomic lines of Si(4s po 3p21D2) at 288.1 nm, and (4s3Pj 3P3Pj) at 251.6 nm are observed. The atomic Si transitions are due to a three photon absorption. We observed seven transitions due to molecular hydrogen at wavelengths 577.5 nm, 565.5 nm, 534.4 nm, 542.5 nm, 471 nm, 461.7 nm, and 455.4 nm. These bands are due to a four photon absorption proc6ss. In addition to the molecular bands we also observed hydrogen atomic lines Ha, H and H.


INTRODUCTION
The photolysis of silane can be useful in developing growth models for thin films in microelectronics industry. 1 '2 The photochemical pathways of the silane dissociation process have had a long and controversial history and finally have been resolved in favour of a three-center elimination reaction to form silylene (Sill2) and molecular hydrogen. [3][4][5][6] The primary photodissociative mechanism of Sill 4 in the formation of silylene may be described as follows: Sill, -+ SiH (1) Sill 4 Sill 2 + H 2 (2) SiH, Sill 2 + 2H (3) The other possible mechanism in the formation of atomic silicon and silyl radical may be described as follows: The formation of Sill may be described as follows: Sill4 --* Sill + H 2 + H (7) Though the breaking of Si-H bond requires about 2.5 eV which corresponds to a single photon energy at 355 nm, scission of that bond is not necessarily the step by which Sill4 dissociation is initiated. A detailed kinetic and spectroscopic study is necessary to establish a definite photochemical pathways.
The excimer laser induced fluorescence (LIF) emission from silylene has been r.eported in the photodissociation of silane. 7  In an attempt to provide further experimental information on the photodissociation of silane, we have focused the third harmonic of a Nd:YAG laser at 355 nm, creating a small interaction zone of volume (10-6cm3). The short pulse duration of the laser (8 ns) eliminates the chain reactions. The emission due to various photofragments are collected by a Photomultiplier .Tube (PMT)and recorded by a multichannel analyzer (MCA) at a fixed silane pressure of 350 Torr and varying the intensity of the laser. The experimental method used and the results obtained will be discussed below.

EXPERIMENTAL
The experimental set-up used in the present investigation is shown in Figure 1. The silane used in this experiment was semiconductor grade, obtained from Air products. The sample cell was made up of stainless steel tubing of one inch internal diameter.
A constricted tubing of 1/4" internal diameter, also made of stainless steel tubing was welded inside the main cell with clear access to Sill 4 fillings and measuring the pressure in the constricted tubing. Both ends of the cell were attached with 1/4" thick quartz windows and Viton O-rings. The total length of the cell was 40 cm. The cell was evacuated prior to filling the silane gas. The gas pressure was measured by a calibrated pressure gauge.
The photolysis was carried out using the third harmonic of a Nd:YAG laser at 355 nm (JKJ model HY-500). The laser was a multimode, linearly polarized with 8 ns pulse duration and 6.3 cmline-width. The laser was operated at a repetition rate of 10 Hz throughout the experiment. The laser was focused at the center of the cell by a lens of focal length 34 cm. The spot size of the focused beam was 100 lam. The output beam which consisted of the pump beam and emissions due to various photofragmented radicals was collimated by an another lens of focal length 50 cm, dispersed by a Norland 5600 multichannel analyzer (MCA) in the multichannel scalling (MCS) mode. Multichannel scalling is a time sweep of the channels in the MCA, with each channel being an interval oftime equal to the total sweep time/total channel swept. During each channel time interval (is) in our case, the memory content for that channel is available for input data counting in the form of serial digital pulses. Thus, the resulting display is a frequency/wavelength histogram. In order to establish the possible photochemical pathway, at a fixed Sill 4 pressure of 350 Torr we varied the energy fluence of the photolysing laser in the range of 30-300 Jcm-2, by keeping the number of pulses fixed (10 pulses per second) and varying the average energy per pulse, we recorded the emission spectrum due to various photofragments. The results will be discussed as follows:

RESULTS AND DISCUSSION
Since our interest is to investigate whether the dissociation of Sill 4 and the subsequent emission from the excited states of various fragments are due to multiphoton absorption, 248 K. SENTRAYAN et al.
we varied the energy fluence of the laser in the range of 30-300 Jcm-2, The emission spectra due to various excited states of the fragments such as SiH2, Sill, H2, H, and Si dissociated from SiH, at three different energy fluences namely, 31 Jcm-2, 188 Jcm-2, and 307 Jcm-2, at a fixed Sill 4 pressure of 350 Torr is shown in Figure 2.
The (A1 B1 X1A1) transition is a well known transition of SiH2 and often used to detect the Sill 2 radical. 12 The Based on the energy level diagram of Sill2 given (see Fig. 3.), SiHz in the 1B state can be generated from SiH, with excitation energy more than 4.6 eV. This clearly indicates that the (1B1 A) transitions are due to a two photon absorption. In order to establish a relationship between the emission intensity of (B-A) transitions and the laser energy fluence, we measured the relative intensity of (B-A) transitions at various laser energy fluences and the results are plotted in Figure 4. A nearly quadratic dependence of the emission intensity on the laser energy fluence for all (B-A) transitions was observed. This indicates that these transitions are due to a two photon absorption.
We have observed a band system shown as "c" in Figure 2  The threshold energy to produce Sill (AA) radical is in the range of 9.57-10.03 ev 5 which corresponds to three photons at 355 nm. The experimentally measured emission intensity on the laser energy fluence plotted in Figure 5 confirms that these transitions are due to a three photon absorption process. The observation of atomic lines of silicon and molecular bands of hydrogen is similar to that of Bosov et al. 6 who showed that a cw CO 2 laser dissociates SiH, at a pressure of 200 Torr into silicon and hydrogen. The threshold energy for the formation of Si 4slPo is in the range of 10.03 10.64 eV. 5 This indicates that the Si atomic transitions are due to a three photon absorption process. Although the transition probability for the line at 288.1 nm and 251.6 nm are equal, 7 we have observed that the intensity of the transition at 288.1 nm is nearly twice stronger than that at 251. Energy-level diagram for Sill:. Ln Laser Fluence in Jcm-2-pulse) Figure 4 The relative intensity of (A1B1. Ln (Laser Fluence in Jcm-2-pulse) is three times stronger that at 251.6 nm. This disagrees with electron impact experiment in which 251.6 nm is observed to be three times stronger than that at 288.1 nm. 8 The two atomic lines of Si at 288.1 nm and 251.6 nm observed in this study shown as "d" in Figure 2. We have observed seven transitions due to the molecular hydrogen wavelengths at 577.5 nm, 565.5 nm, 543.4 nm, 542.5 nm, 471 nm, 461.7 nm, and 455.4 nm and are indicated as "a" in Figure 2. The upper electronic levels involved in the molecular hydrogen emission lies 100,000 cm- ( the time resolution of our detection system, we were unable to detect the any ionic fragments. It is evident from many broad peaks, the spectra displayed in Figure 2a and 2b are emission spectra of silane plasma. Many lines are neither resolved due to the limitation in the resolution ofour detection system nor identified due to the fact that these transitions do not follow simple multiphoton processes but represent complex plasma interactions. The quantum mechanical calculation of the wve numbers of these transitions is very cumbersome and it is beyond the scope of this investigation. We have observed brownish white deposit of a:Si-H on the cell windows due to decomposition of Sill 4. Compared to unhydrogenated a:Si a:Si-H gives great improvements in the photo-conductivity and minority carrier life time 2 and the utility to dope n-and p-type 21 thin films in microelectronics industry. In a: Si-H formation, possible involvement of Sill 2 (and/or SiHa) has been suggested by knights 22 based on photolysis and electron impact dissociation studies by Lampe et al. 2a Though our result support this hypothesis, one cannot ignore the role ofionic species for the formation of a: Si-H due to plasma decomposition of Sill4.

CONCLUSION
The photolysis ofsilane has been carried out using the third harmonic of a Nd:YAG laser at 355 nm. The photofragments identified in the present work are Sill2, Sill, Si, H2, and H. The (A1B1-X1A1) transitions of Sill 2 and (A2A-X2r0 transition of Sill are due to a two and three photon absorption respectively. The Si atomic lines are due to a three photon absorption process. The molecular bands of hydrogen are due to a four photon process. We observed the hydrogen atomic lines H a, Hr, and H. We observed a brownish white deposition on the cell windows due to amorphous silicon (a:Si-H). The exact composition of Si and H are not known and mass spectrometric probing is necessary to establish the / and SiH are not detected in composition of Si and H. The ionic radicals SiH-, Sill 3 this investigation. The absence of these ions may be due to the fact that at a high silane pressure of 350 Torr, the recombination and ionic collision process are fast (few ps) in comparison with our time resolution of our detection system.