- Email: [email protected]
Photoluminescence studies of surface damage states in InP
L470
Surface Science 108 (1981) L470- L476 North-Holland Publishing Company
SURFACE SCIENCE LETTERS PHOTOLUMINESCENCE
STUDIES OF SURFACE DAMAGE STATES IN InP
J.D. OBERSTAR and B.C. STREETMAN Departttlent of Electrical Ellgineering and Coordinated Sciettce Laboratory, nois at C’,barla~Cltatttpai~tt, Urbana, Ilhois 61801, USA Keceivcd
L’ttil’ersit_v of Illi-
10 March 1981
WC have observed a radiative transition (1.306 eV) in the photoluminesccnce which WC show is due to near surface states created by surface damage.
spectra
of InP
Recent publications have drawn attention to the unique surface properties of InP in comparison with other III-V compounds. Capacitance-voltage studies on metal-insulator-semiconductor (MIS) n-type InP capacitors report (almost independently of the dielectric employed) interface state densities from (2-5) X 10” cnie2 eV_’ [l-4]. This density is lower than that for n-type GaAs (-2 X 1012 cme2 eV -‘) an d m ak es n-type InP an attractive material for integrated logic circuits employing normally-off devices [5-71. Additionally, several articles have focused interest on the surface-related optical properties of InP. Of particular interest is the report of a low surface recombination velocity for n-type InP of -lo3 cm s-l (versus -10’ cm s-l in GaAs) and a photoluminescence (PL) intensity -100 times greater than that obtained from GaAs of similar free electron concentration [8]. We have observed another unusual optical feature of the InP surface, suggesting the presence of radiative states in damaged surfaces. Photoluminescence spectra were obtained from samples mounted in a gas exchange liquid helium cryostat and held at 5 K. Excitation was provided by the 5 145 .8 line of an argon ion laser with an incident power density of lo3 W/cm* and a beam diameter of 40 tn. The front surface luminescence signal was focused onto the slit of a : m Spex 1302 double-grating monochromator and was detected by a photomultiplier (S-l response) using lock-in amplifier techniques. The resolution used for this work was 6 A. The emission spectra are uncorrected for system response. All single crystal (lOO)InP samples examined here originated from the same wafer slice of an InP boule grown by the liquid-encapsulation Czochralski (LEC) method at the Naval Research Laboratory [9]. As a result of intentional Fe-doping, these InP samples are semi-insulating with a resistivity of 10’ .Q cm. Three types of surfaces have been investigated and these are shown in fig. 1. Fig. lc shows the surface of a sample prior to any polishing. This surface is the 003%6028/8
1/OOOO-OOOO/SO2.50 0 North-fhlland
J. D. Oberstar, B. G. Streetman / PL studies of surface damage states in InP
L471
Fig. 1. Optical micrographs of the three surfaces discussed: (a) “polished”; (b) “semi-polished”; (c) “as-sawn”.
L412
J.D. Oberstar,
B.C. Streetrtaarl / PL studies of surface
dar?laKc states in Irrl’
result of the sawing operation used to obtain (100) oriented wafers from the single crystal boule. Such surfaces will be referred to here as “as-sawn”. Fig. la shows the results of chemically polishing an “as-sawn” surface with a 1% bromine-methanol solution and a rotating polishing pad. The results of this operation are monitored with an optical microscope and the polishing is continued until scratches, digs. or pits are no longer visible on the surface. The mirror-like surface of fig. la is referred to here as “polished”. The surface displayed in fig. lb results from chemical polishing similar to that employed for “polished” surfaces, but in this case the polishing is terminated before a perfect mirror surface is obtained. The scratches visible in fig. lb are the result of incomplete polishing of “as-sawn” surfaces. Surfaces similar to that in fig. lb are referred to here as “semi-polished”. Fig. 2 displays the photolummescence spectra obtained from the three surfaces described above. Referring to fig. 2a the spectral features characteristic of InP are seen. The dominant peak at 1.41 eV is band-edge luminescence due to recombination of excitons bound to neutral donors (D’X) and acceptors (AOX) [ 10,l l] . The peak at 1.38 eV results from donor-acceptor (D-A) pair transitions [ 12,131. Longitudinal optic phonon (LO) replicas of the (D-A) pair transition occur at intervals of 43 meV, of which only the I-LO peak at 1.34 eV is shown in fig. 2 [14]. An expected result of this work is the observation of a decrease in luminescence intensity when the surface is less than perfectly polished. The effects of polishing damage on the PL intensity of GaAs have been reported in the past [ 15-171. These studies indicate that surface damage can decrease the radiative efficiency by at least an order of magnitude from that obtained from the best surface. We observe a similar result here for InP surfaces. Comparing the integrated intensities of the “semipolished” and “as-sawn” spectra to that of the “polished” spectra we obtain 20% and 40% respectively. Presumbaly such decreases in luminosity can be attributed to surface damage on a macroscopic scale (e.g., scratches) and on a microscopic scale (e.g., dislocations, vacancy-complexes) which introduce nonradiative recombination channels [ 181. The difference in luminosity between ‘semi-polished” and “as-sawn” surfaces may appear contradictory; however, we feel the majority of this effect can be explained in terms of the latter surface having a much larger absorption of the incident radiation. We have made crude measurements of the reflected power from the surfaces of fig. 1. “Polished” and “semi-polished” surfaces yield comparable reflection while the reflected power from “as-sawn” surfaces is -20-40 times weaker. These measurements, together with the equation for luminescent intensity given in ref. [B], suggest that absorption effects are the predominant cause of the observed luminosity differences seen in figs. 2b and 2c. Recent publications prove that the PL intensity from 1nP is strongly affected by surface band-bending effects leading to a space charge layer and an optically dead layer [ 19-2 11. Differences in surface band-bending between “as-sawn” and “semi-polished” sur-faces may also play a role in the observed luminosity differences. An unusual result of our work is the appearance of a new spectral feature at 1.30(6) eV. As can be seen from fig. 2a “polished” surfaces exhibit no PL peaks in
J.D. Oberstar,
B. G. Streetman
I
128 ,
/ PL studies of surface
Energy 1.32 I I
I
(eV) 1.36 I
damage
1.40 I
1
states in InP
1
L413
1.44 I
InP (100) Substrate (a) Polished
4
x180
(b) Semi-Polished
x10
(c) As- Sawn
x4
t 1
I
9600
1
I
,
9600
Fig. 2. Typical photoluminescence temperature for these measurements
,
,
,
9400 9200 Wavelength
spectra from was 5 K.
,
/
,
9000 (A)
the three
,
surfaces
,
(
8800
8600
shown
in fig. 1. The sample
L474
J.D. Oberstar,
B.G. Streetman
/ PL studies of surface damage
states in InP
this energy range, whereas spectra obtained from “semi-polished” surfaces exhibit a weak luminescence in this region. Greater damage to the surface in the “as-sawn” case results in the strongest 1.30(6) eV feature. Fig. 2 suggests that the evolution of this transition is due to surface damage states in the band gap as a result of mechanical damage. We have also observed this transition under other circumstances which further support this view of its origin. “Polished” surfaces degraded by exposure to concentrated HCl exhibit weak spectral features in the 1.30(6) eV region which were not present in the spectra prior to the HCl treatment. This I .30(6) eV feature has also been observed when the excitation beam is incident on an isolated pit of an otherwise perfect “polished” surface. Our results support the surface damage origin of this transition proposed by Street et al., who were to our knowledge the first to observe this feature [21]. They observed this PL band from samples which cleaved poorly under vacuum and resulted in surfaces with a very large density of surface steps. We wish to point out that our material and processing techniques differ in several respects: (1) Street et al. performed experiments on (llO)InP while ours were on (lOO)InP; (2) their crystals were lightly doped n and p-type while ours are semi-insulating; (3) their samples were cleaved and examined in ultrahigh vacuum while ours have been exposed to air. Despite these differences we observe the identical transition reported by Street et al. Surface studies performed in the last few years indicate that no intrinsic surface states exist in the fundamental gap of vacuum cleaved GaAs or Inp(ll0) surfaces [22,23]. It is now believed that previous reports of Fermi level pinning were due to extrinsic states. In light of these findings and considering the extent of damage to our surfaces (together with their exposure to air), we feel we can rule out transitions due to intrinsic surface states. Using a value of 1.423 eV vor the bandgap energy at 5 K would place the observed transition at 117 meV from the bulk conduction band edge; however, the exact location of this level with respect to the surface conduction band is unknown due to probable band bending at the surface [24]. In this regard the 1.30(6) eV feature may be a transition to an effectively shallow level (bent band edge to surface damage state level) or a 117 meV transition to an optically deep level. The detailed nature of the defects giving rise to the observed transition is not presently known. In this regard the influence of the anion should not be overlooked, since it appears to affect Schottky barrier heights and the surface recombination velocity in other multicomponent semiconductors [25271. Experimental evidence suggests that chemical adsorption of oxygen on InP directly affects the P surface atoms but not the In surface atoms [22]. Williams et al. have presented a model for ‘Schottky barriers which incorporates the role of phosphorus vacancies on the I.nP surface [28]. The chemical volatility of the phosphorus component of InP would appear to make such models based on phosphorusrelated defects likely candidates for understanding the detailed nature of the 1.30(6) eV transition. Recently Street and Williams have published further results of their work on the 1.3 eV transition [29]. Our data in agreement with their observations, except for
J.D. Oberstar, B. G. Streetman 1 PL studies of surface damage states in InP
LA75
the long-term stability of defects reported here. They observe a decay in the radiative luminescence of the 1.3 eV transition as a result of room temperature annealing over a period of days. It is stated that this may explain why this feature is not routinely observed in the spectra of InP. Although we have made no effort to examine annealing effects, we have observed the transition in “as-sawn” surfaces months after the sawing operation. This may be the result of the much greater damage of these surfaces in comparison to those examined by Street and Williams. We wish to thank R.L. Henry and E.M. Swiggard of the Naval Research Laboratory for providing us with the InP substrates. Discussions with K. Hess and Nick Holonyak, Jr. were greatly appreciated. This work was supported in part by the Office of Naval Research under Contract NOOO14-76-C-0806 and in part by the Joint Services Electronics Program under Contract NOOOl4-79-C-0424.
References [l] [2] [3] [4] [5] [6] [7] (81 [9]
L. Messick, J. Appl. Phys. 47 (1976) 4949. D.L. Lile and D.A. CoIlins, Appl. Phys. Letters 28 (1976) 554. C.W. Wilmsen, Critical Rev. Solid State Sci. 5 (1975) 313. G.G. Roberts, K.P. Pande and W.A. Barlow, Electron. Letters 13 (1977) 581. D.L. Lile and D.A. CoIlins, Thin Solid Films 56 (1979) 225. D.L. Like, A.R. CIawson and D.A. Collins, Appl. Phys. Letters 29 (1976) 207. L.J. Messick, IEEE Trans. Electron Devices 28 (1981) 218. H.C. Casey, Jr. and E. Buehler, Appl. Phys. Letters 30 (1977) 247. R.L. Henry and E.M. Swiggard, in: Proc. 6th Intern. Symp on GaAs and Related Compounds, Edinburgh, 1976, Inst. Phys. Conf. Ser. 33b (1977) 28. [lo] J.U. Fischbach, G. Benz, N. Stath, M.H. Pilkuhn and K.W. Benz, Solid State Commun. 11 (1972) 721. [ 111 E.W. Williams, W. Elder, M.G. Astles, M. Webb, J.B. Mulhn, B. Straughan and P.J. Tufton, J. Electrochem. Sot. 120 (1973) 1741. [ 121 U. Heim, Solid State Commun. 7 (1969) 445. [13] R.C.C. Leite, Phys. Rev. 157 (1967) 672. [14] 0. Roder, U. Heim and M.H. Pilkuhn, J. Phys. Chem. Solids 31 (1970) 2625. [15] B. Tuck, Phys. Status Solidi 36 (1969) 285. [ 161 V.A. Zuev, V.G. Litovchenko, G.A. Sukach and D.V. Korbutyak, Phys. Status Solidi 17 (1973) 353. [17] A. Karpol and B. Pratt, Solid State Commun. 12 (1973) 325. [ 181 J.I. Pankove, Optical Processes in Semiconductors (Dover Publications, New York, 1971) p. 164. [19] K. Ando, A. Yamamoto and M. Yamaguchi, J. Appl. Phys. 51 (1981) 6432. [20] D.B. Witty and D.F. Kyser, J. Appl. Phys. 38 (1967) 375. [21] R.A. Street, R.H. Williams and R.S. Bauer, J. Vacuum Sci. Technol. 17 (1980) 1001. (221 W.E. Spicer, I. Lindau, P.E. Gregory, C.M. Garner, P. Pianetta and P.W. Chye, J. Vacuum Sci. Technol. 13 (1976) 780. [23] A. McKinley, C.P. Srivastava and R.H. Williams, J. Phys. C (Solid State Phys.) 13 (1980) 1581. [24] S.B. Nam, DC. Reybolds, C.W. Litton and T.C. Collins, Phys. Rev. B13 (1976) 1643.
L476
J.D. Oberstar,
B.G. Streetman
[25] J.O. McCaldin,
[29]
of surfacedamage
states in InP
T.C. McGill and CA. Mead, Phys. Rev. Letters 36 (1976) 56. Jr., R.D. Burnham, H.R. Zwicker, D.L. Keune, W.O. Groves, M.G. Crawford and J.W. Burd, Solid-State Electron. 14 (1971) 949. M.S. Daw and D.L. Smith, Appl. Phys. Letters 36 (1980) 690. R.H. WiJliams, R.R. Varma and V. Montgomery, J. Vacuum Sci. Technol. 16 (1979) 1418. R.A. Street and R.H. Williams, J. Appl. Phys. 52 (1981) 402.
[ 261 D.R. Scifres, N. Holonyak, [27] [28]
/ PL studies
Surface Science 108 (1981) L470- L476 North-Holland Publishing Company
SURFACE SCIENCE LETTERS PHOTOLUMINESCENCE
STUDIES OF SURFACE DAMAGE STATES IN InP
J.D. OBERSTAR and B.C. STREETMAN Departttlent of Electrical Ellgineering and Coordinated Sciettce Laboratory, nois at C’,barla~Cltatttpai~tt, Urbana, Ilhois 61801, USA Keceivcd
L’ttil’ersit_v of Illi-
10 March 1981
WC have observed a radiative transition (1.306 eV) in the photoluminesccnce which WC show is due to near surface states created by surface damage.
spectra
of InP
Recent publications have drawn attention to the unique surface properties of InP in comparison with other III-V compounds. Capacitance-voltage studies on metal-insulator-semiconductor (MIS) n-type InP capacitors report (almost independently of the dielectric employed) interface state densities from (2-5) X 10” cnie2 eV_’ [l-4]. This density is lower than that for n-type GaAs (-2 X 1012 cme2 eV -‘) an d m ak es n-type InP an attractive material for integrated logic circuits employing normally-off devices [5-71. Additionally, several articles have focused interest on the surface-related optical properties of InP. Of particular interest is the report of a low surface recombination velocity for n-type InP of -lo3 cm s-l (versus -10’ cm s-l in GaAs) and a photoluminescence (PL) intensity -100 times greater than that obtained from GaAs of similar free electron concentration [8]. We have observed another unusual optical feature of the InP surface, suggesting the presence of radiative states in damaged surfaces. Photoluminescence spectra were obtained from samples mounted in a gas exchange liquid helium cryostat and held at 5 K. Excitation was provided by the 5 145 .8 line of an argon ion laser with an incident power density of lo3 W/cm* and a beam diameter of 40 tn. The front surface luminescence signal was focused onto the slit of a : m Spex 1302 double-grating monochromator and was detected by a photomultiplier (S-l response) using lock-in amplifier techniques. The resolution used for this work was 6 A. The emission spectra are uncorrected for system response. All single crystal (lOO)InP samples examined here originated from the same wafer slice of an InP boule grown by the liquid-encapsulation Czochralski (LEC) method at the Naval Research Laboratory [9]. As a result of intentional Fe-doping, these InP samples are semi-insulating with a resistivity of 10’ .Q cm. Three types of surfaces have been investigated and these are shown in fig. 1. Fig. lc shows the surface of a sample prior to any polishing. This surface is the 003%6028/8
1/OOOO-OOOO/SO2.50 0 North-fhlland
J. D. Oberstar, B. G. Streetman / PL studies of surface damage states in InP
L471
Fig. 1. Optical micrographs of the three surfaces discussed: (a) “polished”; (b) “semi-polished”; (c) “as-sawn”.
L412
J.D. Oberstar,
B.C. Streetrtaarl / PL studies of surface
dar?laKc states in Irrl’
result of the sawing operation used to obtain (100) oriented wafers from the single crystal boule. Such surfaces will be referred to here as “as-sawn”. Fig. la shows the results of chemically polishing an “as-sawn” surface with a 1% bromine-methanol solution and a rotating polishing pad. The results of this operation are monitored with an optical microscope and the polishing is continued until scratches, digs. or pits are no longer visible on the surface. The mirror-like surface of fig. la is referred to here as “polished”. The surface displayed in fig. lb results from chemical polishing similar to that employed for “polished” surfaces, but in this case the polishing is terminated before a perfect mirror surface is obtained. The scratches visible in fig. lb are the result of incomplete polishing of “as-sawn” surfaces. Surfaces similar to that in fig. lb are referred to here as “semi-polished”. Fig. 2 displays the photolummescence spectra obtained from the three surfaces described above. Referring to fig. 2a the spectral features characteristic of InP are seen. The dominant peak at 1.41 eV is band-edge luminescence due to recombination of excitons bound to neutral donors (D’X) and acceptors (AOX) [ 10,l l] . The peak at 1.38 eV results from donor-acceptor (D-A) pair transitions [ 12,131. Longitudinal optic phonon (LO) replicas of the (D-A) pair transition occur at intervals of 43 meV, of which only the I-LO peak at 1.34 eV is shown in fig. 2 [14]. An expected result of this work is the observation of a decrease in luminescence intensity when the surface is less than perfectly polished. The effects of polishing damage on the PL intensity of GaAs have been reported in the past [ 15-171. These studies indicate that surface damage can decrease the radiative efficiency by at least an order of magnitude from that obtained from the best surface. We observe a similar result here for InP surfaces. Comparing the integrated intensities of the “semipolished” and “as-sawn” spectra to that of the “polished” spectra we obtain 20% and 40% respectively. Presumbaly such decreases in luminosity can be attributed to surface damage on a macroscopic scale (e.g., scratches) and on a microscopic scale (e.g., dislocations, vacancy-complexes) which introduce nonradiative recombination channels [ 181. The difference in luminosity between ‘semi-polished” and “as-sawn” surfaces may appear contradictory; however, we feel the majority of this effect can be explained in terms of the latter surface having a much larger absorption of the incident radiation. We have made crude measurements of the reflected power from the surfaces of fig. 1. “Polished” and “semi-polished” surfaces yield comparable reflection while the reflected power from “as-sawn” surfaces is -20-40 times weaker. These measurements, together with the equation for luminescent intensity given in ref. [B], suggest that absorption effects are the predominant cause of the observed luminosity differences seen in figs. 2b and 2c. Recent publications prove that the PL intensity from 1nP is strongly affected by surface band-bending effects leading to a space charge layer and an optically dead layer [ 19-2 11. Differences in surface band-bending between “as-sawn” and “semi-polished” sur-faces may also play a role in the observed luminosity differences. An unusual result of our work is the appearance of a new spectral feature at 1.30(6) eV. As can be seen from fig. 2a “polished” surfaces exhibit no PL peaks in
J.D. Oberstar,
B. G. Streetman
I
128 ,
/ PL studies of surface
Energy 1.32 I I
I
(eV) 1.36 I
damage
1.40 I
1
states in InP
1
L413
1.44 I
InP (100) Substrate (a) Polished
4
x180
(b) Semi-Polished
x10
(c) As- Sawn
x4
t 1
I
9600
1
I
,
9600
Fig. 2. Typical photoluminescence temperature for these measurements
,
,
,
9400 9200 Wavelength
spectra from was 5 K.
,
/
,
9000 (A)
the three
,
surfaces
,
(
8800
8600
shown
in fig. 1. The sample
L474
J.D. Oberstar,
B.G. Streetman
/ PL studies of surface damage
states in InP
this energy range, whereas spectra obtained from “semi-polished” surfaces exhibit a weak luminescence in this region. Greater damage to the surface in the “as-sawn” case results in the strongest 1.30(6) eV feature. Fig. 2 suggests that the evolution of this transition is due to surface damage states in the band gap as a result of mechanical damage. We have also observed this transition under other circumstances which further support this view of its origin. “Polished” surfaces degraded by exposure to concentrated HCl exhibit weak spectral features in the 1.30(6) eV region which were not present in the spectra prior to the HCl treatment. This I .30(6) eV feature has also been observed when the excitation beam is incident on an isolated pit of an otherwise perfect “polished” surface. Our results support the surface damage origin of this transition proposed by Street et al., who were to our knowledge the first to observe this feature [21]. They observed this PL band from samples which cleaved poorly under vacuum and resulted in surfaces with a very large density of surface steps. We wish to point out that our material and processing techniques differ in several respects: (1) Street et al. performed experiments on (llO)InP while ours were on (lOO)InP; (2) their crystals were lightly doped n and p-type while ours are semi-insulating; (3) their samples were cleaved and examined in ultrahigh vacuum while ours have been exposed to air. Despite these differences we observe the identical transition reported by Street et al. Surface studies performed in the last few years indicate that no intrinsic surface states exist in the fundamental gap of vacuum cleaved GaAs or Inp(ll0) surfaces [22,23]. It is now believed that previous reports of Fermi level pinning were due to extrinsic states. In light of these findings and considering the extent of damage to our surfaces (together with their exposure to air), we feel we can rule out transitions due to intrinsic surface states. Using a value of 1.423 eV vor the bandgap energy at 5 K would place the observed transition at 117 meV from the bulk conduction band edge; however, the exact location of this level with respect to the surface conduction band is unknown due to probable band bending at the surface [24]. In this regard the 1.30(6) eV feature may be a transition to an effectively shallow level (bent band edge to surface damage state level) or a 117 meV transition to an optically deep level. The detailed nature of the defects giving rise to the observed transition is not presently known. In this regard the influence of the anion should not be overlooked, since it appears to affect Schottky barrier heights and the surface recombination velocity in other multicomponent semiconductors [25271. Experimental evidence suggests that chemical adsorption of oxygen on InP directly affects the P surface atoms but not the In surface atoms [22]. Williams et al. have presented a model for ‘Schottky barriers which incorporates the role of phosphorus vacancies on the I.nP surface [28]. The chemical volatility of the phosphorus component of InP would appear to make such models based on phosphorusrelated defects likely candidates for understanding the detailed nature of the 1.30(6) eV transition. Recently Street and Williams have published further results of their work on the 1.3 eV transition [29]. Our data in agreement with their observations, except for
J.D. Oberstar, B. G. Streetman 1 PL studies of surface damage states in InP
LA75
the long-term stability of defects reported here. They observe a decay in the radiative luminescence of the 1.3 eV transition as a result of room temperature annealing over a period of days. It is stated that this may explain why this feature is not routinely observed in the spectra of InP. Although we have made no effort to examine annealing effects, we have observed the transition in “as-sawn” surfaces months after the sawing operation. This may be the result of the much greater damage of these surfaces in comparison to those examined by Street and Williams. We wish to thank R.L. Henry and E.M. Swiggard of the Naval Research Laboratory for providing us with the InP substrates. Discussions with K. Hess and Nick Holonyak, Jr. were greatly appreciated. This work was supported in part by the Office of Naval Research under Contract NOOO14-76-C-0806 and in part by the Joint Services Electronics Program under Contract NOOOl4-79-C-0424.
References [l] [2] [3] [4] [5] [6] [7] (81 [9]
L. Messick, J. Appl. Phys. 47 (1976) 4949. D.L. Lile and D.A. CoIlins, Appl. Phys. Letters 28 (1976) 554. C.W. Wilmsen, Critical Rev. Solid State Sci. 5 (1975) 313. G.G. Roberts, K.P. Pande and W.A. Barlow, Electron. Letters 13 (1977) 581. D.L. Lile and D.A. CoIlins, Thin Solid Films 56 (1979) 225. D.L. Like, A.R. CIawson and D.A. Collins, Appl. Phys. Letters 29 (1976) 207. L.J. Messick, IEEE Trans. Electron Devices 28 (1981) 218. H.C. Casey, Jr. and E. Buehler, Appl. Phys. Letters 30 (1977) 247. R.L. Henry and E.M. Swiggard, in: Proc. 6th Intern. Symp on GaAs and Related Compounds, Edinburgh, 1976, Inst. Phys. Conf. Ser. 33b (1977) 28. [lo] J.U. Fischbach, G. Benz, N. Stath, M.H. Pilkuhn and K.W. Benz, Solid State Commun. 11 (1972) 721. [ 111 E.W. Williams, W. Elder, M.G. Astles, M. Webb, J.B. Mulhn, B. Straughan and P.J. Tufton, J. Electrochem. Sot. 120 (1973) 1741. [ 121 U. Heim, Solid State Commun. 7 (1969) 445. [13] R.C.C. Leite, Phys. Rev. 157 (1967) 672. [14] 0. Roder, U. Heim and M.H. Pilkuhn, J. Phys. Chem. Solids 31 (1970) 2625. [15] B. Tuck, Phys. Status Solidi 36 (1969) 285. [ 161 V.A. Zuev, V.G. Litovchenko, G.A. Sukach and D.V. Korbutyak, Phys. Status Solidi 17 (1973) 353. [17] A. Karpol and B. Pratt, Solid State Commun. 12 (1973) 325. [ 181 J.I. Pankove, Optical Processes in Semiconductors (Dover Publications, New York, 1971) p. 164. [19] K. Ando, A. Yamamoto and M. Yamaguchi, J. Appl. Phys. 51 (1981) 6432. [20] D.B. Witty and D.F. Kyser, J. Appl. Phys. 38 (1967) 375. [21] R.A. Street, R.H. Williams and R.S. Bauer, J. Vacuum Sci. Technol. 17 (1980) 1001. (221 W.E. Spicer, I. Lindau, P.E. Gregory, C.M. Garner, P. Pianetta and P.W. Chye, J. Vacuum Sci. Technol. 13 (1976) 780. [23] A. McKinley, C.P. Srivastava and R.H. Williams, J. Phys. C (Solid State Phys.) 13 (1980) 1581. [24] S.B. Nam, DC. Reybolds, C.W. Litton and T.C. Collins, Phys. Rev. B13 (1976) 1643.
L476
J.D. Oberstar,
B.G. Streetman
[25] J.O. McCaldin,
[29]
of surfacedamage
states in InP
T.C. McGill and CA. Mead, Phys. Rev. Letters 36 (1976) 56. Jr., R.D. Burnham, H.R. Zwicker, D.L. Keune, W.O. Groves, M.G. Crawford and J.W. Burd, Solid-State Electron. 14 (1971) 949. M.S. Daw and D.L. Smith, Appl. Phys. Letters 36 (1980) 690. R.H. WiJliams, R.R. Varma and V. Montgomery, J. Vacuum Sci. Technol. 16 (1979) 1418. R.A. Street and R.H. Williams, J. Appl. Phys. 52 (1981) 402.
[ 261 D.R. Scifres, N. Holonyak, [27] [28]
/ PL studies