We present measurements of the temperature dependence of P,Q,R ltnninescence and P,Q,R lifetimes that support this conclusion. This is the first investigation of the photoluminescence properties of Si-rich Si-Ge alloy semiconductors.
CHAPTER 1
INTRODUCTION
In Chapter 2, an investigation into the effect of increasing impurity concentrations on the photoluminescence spectrum is presented. This is the first investigation into the photoinnate properties of Si-rich Si-Ge alloy semiconductors.
TI-IE PHOTOLUMINESCENCE TECHNIQUE
1he effect of the Boltzmann distribution of kinetic energies in free exciton center of mass is included in EFE(l), 1his effect results in a predicted line shape for free exciton luminescence, which is C3). As a result, the strength of the Auger process depends on the k-space spread of the carrier functions.
APPLICATIONS OF CURRENT INTEREST
As a result, there is considerable interest in using the photolliIDlinescence technique to ~xamine irradiated semiconductors. Measuring the dependence of the luminescence intensity on the characteristics of the sample can provide important information about.
ENERGY (meV)
CHAPTER 2
INFLUENCE OF IMPURITY CONCENTRATION ON THE PHOTOLUMINESCENCE SPECTRUM - A CASE STUDY OF THE SYSTEM Si : (B, In). At impurity concentrations well below those of technological interest, a damping of the shallow impurity luminescence is observed. Physically, si is the ratio of the FE capture rate, yi', to the total BE.
Naturally, the application of this method requires that the absence of saturation effects be attributed. Consider the case of a single impurity~ impurity i~ and consider variation of the parameter £i. Also in this case it is important to note that Eq, Ci.10) simply indicates the behavior of the ratio I./I. and I., we must again return to Eq.
As a result of this coupling of N. the rate theory of the system predicts that it may be possible to quench BE completely. Ltuninescence intensity simply by increasing. We wish to determine at what point the simple velocity theory presented in Eq. 2.1) and (2, 2) become inapplicable as the impurity concentrations increase. Total quenching of low tuninescence will be observed in many samples, but due to the coupling effect between I. vation alone it is not sufficient to conclude that the rate theory has become inapplicable. Si:(B,In) was chosen as a model system for the study because highly doped Si:In is a very important material for detector applications and because the low concentration photollililuminescence properties of the Si:B and Si:In subsystems are relatively well understood.
TO) INTEGRATED INTENSITY
- DISCUSSION OF 1HE EXPERIMENJ:AL RESULTS .1 The Low Concentration Result
- SUMMARY AND CONCLUSION
- INTRODUCTION
- MJDEL FOR P,Q,R LUMINESCENCE
Ptunp power dependence of the NP BE1n luminescence intensity in sample 2104011. a) The low pump power regime where the luminescence intensity is proportional to the pump power. The results of the straight line fit with low pump power and the resulting intensity ratios are listed in Table 2.2. The experimental results for the Si:(B,In) system presented in Section 2.3 show that the expected behavior based on the theory in Section 2.2 is achieved in the low concentration regime.
Based on the experimental results presented in Section 2.3, it is clear that the simple theory obtained in the low concentration limit becomes inapplicable at even moderately elevated concentrations. The variation in the calculated experimental result for each sample is mainly a consequence of the tmcertainty in Nln'. This splitting of the BEin luminescence was attributed to the presence of a low-lying excited state of the BEin complex.
Finally, certain properties of BEin have been derived as a result of detailed investigations of the BEin luminescence. The energy and assignment of some of the lines observed in this study of the photoluminescence spectrum of Si:In. From the spectra shown in Figure 3.3, measurements of the P,Q,R line intensity ratios were made as a function of temperature.
Finally, line Q is the result of a transition from an excited state of the bound exciton to an excited state of the isoelectronic complex. The qualitative features of the model described in the previous section can be incorporated into a series of rate comparisons.
ENERGY LEVEL SCHEME FOR P,Q,R LUMINESCENCE ll.E+l.7meV l PLINE (1117.6meV)
ENHANCEMENT OF P,Q,R LUMINESCENCE .1 Origin of P,Q,R Luminescence
However, we have discovered an unusual property of the P,Q,R luminescence that makes it impossible to continue on such a course. Of course, the absolute 11.ID1insence intensities will depend sensitively on the optical alignments and instrl.ID1ental. responses inherent to the experimental measurement, as well as to the condition of the sample surface. In Fig. 3.12 we present several measurements of the photo- 11.ID1incentration spectrum.ID1 of sample Cll7A. The experimental conditions were identical for each measurement. The only difference between the spectra is that the sample was removed from and placed back into the dewar between each spectrum.
Saturation results in a spatial profile of bound exciton densities, which leads to a dependence of the luminescence intensity on the optical alignment. Of course, as already discussed, spectra from the same region were reproducible if the sample remained in the dewar at low temperatures between measurements. We are therefore forced to consider our sample handling during luminescence measurements as the source of this lack of reproducibility. In particular, it leads us to consider the effect of such an operation on the surface characteristics of the sample.
In the first step of the procedure, the sample is subjected to a 2-hour annealing at 1000 C in a dry He atmosphere. P-line 11. Enhancement of fluorescence in response to room temperature treatment of the surface of Si:In sample Cl17A. The second model involves the suggestion that P-line luminescence is the result of recombination occurring at the surface of the semiconductor and which, again, is associated with an impurity complex involving In.
SUMMARY AND CONCLUSION
Quantitative experimentation is certainly required under conditions under which controlled surface treatments can be obtained. iii) Temperature dependent measurements of P,Q,R luminescence were obtained. 1his result suggests that the Q line may be the result of a transition from an excited state. An Arrhenius plot of the P,Q,R line intensity ratios reveals that the P,Q,R luminescence is not due to recombination of different BE states of the same center.
This observation supports the interpretation that line R is a local phonon mode replica of line P. This is of course also consistent with the interpretation that lines P and Rare are the result of BE recombination at independent centers, but the temperature dependence measurements effectively rule out this possibility. It has been proposed that lines l P,Q and R originate from the same isoelectronic complex, Line Q is then a transition from an excited state of the complex. Line R is a local phonon mode replica of line P. Rate equations based on this model have been solved and applied in the high temperature regime where thermalization is the dominant decay mechanism.
The thermal binding energies obtained in this way support the interpretation that the P,Q,R lines are the result of isoelectronic spherical exciton recombination. Moreover, they also support the interpretation that line R is a local mode replica of line P. viii). It was found that the intensity of P,Q,R luminescence is sensitively dependent on the surface treatment of the sample. This is the first observation of surface-dependent photoluminescence.
CHAPTER 4
Properties of alloy semiconductors have been of general interest for some time, such compounds provide a convenient system with which to study experimentally and theoretically the effects of . disorder, which can be changed with the alloy composition, the 1be band gap can also be changed with the alloy composition, and therefore considerable effort has been made towards the development of intrinsic and extrinsic alloy photodetectors, 1be properties study of alloy semiconductor luminescence. can provide useful and relatively easily interpretable information about the properties of bonds and the consequences of their disordered nature. The greatest attention has been directed towards the ternary alloys III.,_v, where the luminescence processes have been extensively studied C-1-31, Lt The .nnines-cence of the ternary alloys II-VI~ mostly Hg1 -x x Cd Te C4) where IR detector applications are particularly important. Si1_xGex are not particularly well known, although free and bound exciton (S), donor-acceptor (6) and electron-hole dot (?) recombination have been observed in Ge-rich alloys. In this chapter we report the first detailed measurements. of luminescence from Si. Alloy compositions are obtained by a variety of techniques; the results of electron microprobe, density, and X-ray diffraction measurements were in excellent agreement.
In this section, we consider the effect of increasing sample temperature on the photoluminescence spectrum of sample C077. At high temperatures, the luminescence of free excitons (FE) takes a form characteristic of luminescence of free excitons in Si. Note that at these temperatures the FE emission takes a form characteristic of the free exciton emission in Si.
In all cases the adaptation temperature obtained in this way was within 1 K of the measured bath temperature. Moreover, the threshold energy remained constant within 0.05 meV. On the basis of this analysis, the line labeled FE is identified as a result of no-phonon (NP) FE recombination. Note that this intrinsic NP luminescence is greatly enhanced in the alloy as the Ge atoms can act as momentum-conserving scattering centers, C5). 1 the identification of the FE line obtained in the previous section, division between the FE threshold and the BEP line peak positions, and the thermal behavior shown in fig.
ENERGY
It therefore seems reasonable to conclude that the line marked BE in the luminescence of sample C077 is mainly due to NP BEP recombination. We see that, for the impurity concentrations specified in the figure, the BEp line is about an order of magnitude more intense than the ~ line. To begin with, we see that the BE line is significantly broadened in the alloy.
In this model, the broadening of the BE line is accounted for by relatively large-scale fluctuations in the alloy composition. There is a considerable amount of evidence supporting the nearest-neighbor configuration model for BE line extension in . alloy These results indicate that the closest configuration of alloy atoms plays a crucial role in determining the ionization energies of impurities in the alloy.
However, the shift in the peak position of the BEp line with temperature is expected based on the nearest neighbor configuration model, as the higher energy lines will thermalize at lower temperatures. If vp(k) < (vFE+yp), then the long-term decay rate is dominated by vp(k) and we would expect some variation in the BEP line shape during the decay. In the limited capture case, the decay rate of all components is (vFE + Yp), which is independent of k.
