Phase differences in RHEED oscillations of various beams during epitaxial growth of GaAs(100) by MBE

Vacuum~volume 41/numbers 4-6/pages 1052 to 1067/1990 Printed in Great Britain

0042-207X/90S3.00 + .00 Pergamon Press plc

Extended abstracts Phase differences in RHEED oscillations of various beams during epitaxial growth of GaAs(100) by MBE

2.0-

(a)

-r

/

J S Resh, K D Jamison, J Strozier*, A Bensaoula and A Ignatiev, Space Vacuum Epitaxy Center, University of Houston,

Houston, TX 77204-5507, USA Intensity oscillations are routinely used to determine growth rates, I I I - V flux ratios, and composition in ternary compounds during MBE growth ]. The oscillations are usually measured in the specular beam, even though the nonspecular beams sometimes have oscillations of greater amplitude than the specular beam 2,3. Intensity oscillations in the 00, 01 and 0i beams diffracted from the GaAs(100) surface during MBE growth have been systematically studied as a function of diffraction geometry. Although the oscillations from the various beams all have the same period, they exhibit a phase relationship with respect to one another that is very sensitive to the incident angle 0 of the electron beam and the azimuthal angle ~b of the substrate. A video RHEED system was used to simultaneously monitor the intensity of several beams 4. Following the convention set by Zhang et al 5, the phase of a particular beam was determined by calculating the ratio (t3/2/T), where t3/2 is the time from shutter opening to the second minimum of the oscillation and T is the steady state oscillation period. A (t3/2/T) ratio of 1.5 indicates a 'normal phase', i.e. the intensity begins at a maximum at zero layer coverage and falls to a minimum at half-layer coverage. A (t3/2 T) ratio of I or 2 corresponds to oscillations which begin at a minimum, i.e. 180° out of phase with the preceding description. Figures l(a and b) show plots of t3/2 VS t~ for two different incident angles. An azimuthal angle of 0.0 ° indicates exact alignment of the electron beam with the [110] direction. Due to the glancing angle of incidence, the 01 and 0T beams are both present for only a very small azimuthal angular range. Note that the data is symmetric about a symmetry axis in the shape,

I I /// 1.01-- H~" f

/

~ e I O0 l~om - - - - I - - - 01 beam ---.-o---- O~ beam

I

-0.3

~"-.. I-~'~I"~-~ "J-" t - ~ '

I

-0.1

0.1

"%% 0.3

Azimuthal angle (degrees)

(b)

2.0 I

- - - = - - - 01beam ~=~ O0 beam -- --o---- O~'blmm

I-

1.5

1.0

-0.7

I

0

I

0.7

Azimuthal angle (degrees)

Figure 1. t3/2/T ratio vs azimuthal angle u tot the 00, 01 and the 0T beams (a) 0 = 2.1°, (b) 0 = 1.9°. Lines were added to aid the eye only.

q~=0 °. Kinematic scattering theory applied to a simple growth model predicts oscillations with identical phase in all diffracted beams 6. Within that analysis, the RHEED beam intensity minima always occur at half-layer coverage, with the intensity always dropping at the onset of growth. Figure 1 clearly shows, however, that significant phase differences between oscillations in different diffracted beams occur and are sensitive to the experimental geometry. The possible role of inelastic processes in the phase change between various beams has been examined. A number of authors claim that interaction of the diffracted beam with forward scattered diffuse electrons that then undergo a Kikuchi type scattering are responsible for the change of phase away from the expected t3/2/T= 1.5 for the specular beam 5,7,s. In this model they state that the number of diffuse electrons

*Permanent address: Empire College, State University of New York, Stony Brook, NY, USA. 1052

available for Kikuchi processes is largest when the surface is in the highest degree of disorder. Kikuchi line intensity should, therefore, oscillate 180° out of phase with the 'normal phase' of elastically scattered beams. Phase changes in the diffracted RHEED beams may then occur through a superposition of inelastically scattered and elastically scattered electrons. The data presented here, however, shows phase shifts in the 00, 01, and 01" beams over a continous range of incident and azimuthal angle, regardless of whether or not there are Kikuchi lines near the particular beam(s). Therefore, it is doubtful that Kituchi line scattering is entirely responsible for the phase shifts observed. In summary, significant phase differences have been observed between the RHEED intensity oscillations measured for different diffracted beams which cannot be explained by kinematic theory. Whether an elastic, multiple-scattering, dynamical approach can explain these phase differences needs to be determined. It is clear, however, that diffraction conditions (as opposed to growth conditions) can drastically affect the phase

Extended abstracts

SnO2-Si(100) interface

of the RHEED oscillations. As a result, it should be realized that the maximum intensity of a diffracted RHEED beam does not necessarily correspond to a completed layer during growth.

S P da Cunha, M Schreiner and C I Z Mammana, lnst Microelectronics/Centro Tecnologico para Inform~tica, CP 6162, 13081 Campinas, SP, Brazil

Acknowledgements This work has been partially supported by NASA through grant NAGW-977 and by the R A Welch Foundation.

and

References

A P Mammana and R Landers, FEE and I F G W / U N I C A M P , CP 6162, 13081 Campinas, SP, Brazil

J M van Hove, P R Pukite and P I Cohen, J Vac Sci Technol, B1, 741 (1983). 2j H Neave and B A Joyce, Appl Phys, A31, 1 (1983). 3 p j Dobson, B A Joyce, J H Neave and J Zhang, J Crystal Growth, 81, 1 (1987). 4 j S Resh, J Strozier, K D Jamison and A Ignatiev, Rev Scient Instrurn, 61, 771 (1990). 5j Zhang, J H Neave, P J Dobson and B A Joyce, Appl Phys, A42, 317 (1987). 6j Resh, K D Jamison, J Strozier, A Bensaoula and A Ignafiev, Phys Rev B15, 40, 779 (1989). 7G E Crook, E G Eyink, A C Campbell, D R Hinson and B G Streetman, J Vac Sci Technol, A7, 2549 (1989). 8 p K Larsen, G Meyer-Ehmsen, B Bolger and A J Hoeven, J Vac Sci Technol, A5, 611 (1987).

Thin films of semiconductor oxides are very promising in applications where a high transmittance can improve the response to blue light as well as the quantum efficiency. Further, tin dioxide presents physical and chemical properties that make it compatible with the materials and processes normally employed in integrated circuits, mainly temperatures of deposition, melting point and selectivity to wet etching. These aspects motivated us to investigate the use of tin dioxide as the gate material in MIS structures in place of aluminum or polysilicon envisaging future applications in opto-electronic devices where a high transmittance is desirable. We have studied the interface of thin films of SnO2 and single crystalline Si, using n and p Si wafers of 0.5 to 20 ~ cm over which we have deposited SnO2 films by a simple CVD process, using tin tetrachloride and methanol vapors carried by nitrogen, at fluxes of 100 to 700 cm 3 rain- i. The reaction is carried

A2 / ~

C(pF)

t6 ¢1114

- ~ ~^1o F-,zj

350 ~

300-

250-

20

L,oSOO -- o,O:°

200-

N B = 7 . 6 x t O I 4 c.~ 3 -t8 -t6

150-14 ,t2

'10

IOO"

.8 "6

50-

4 .2 I -12

I

I -10

:

; -8

Sn02(-)

:

_~

:

_~,

:

_~,

I

0

2

4

6

8 VOLTAGE (Volts)

10

12 sn02(+)

Figure l(a) Caption on next page. 1053