In situ spectroscopic ellipsometry in molecular beam epitaxy for photonic devices

Applied Surface Science 63 (1993) 1-8 North-Holland

applied surface science

In situ spectroscopic ellipsometry in molecular beam epitaxy for photonic devices G.N. Maracas, J.L. Edwards, D.S. Gerber

and R. Droopad

Arizona State University, EE Department/CSSER, Tempe, A Z 85287, USA

Received 2 June 1992; accepted for publication 31 July 1992

In situ spectroscopic ellipsometry (SE) has been shown to be a versatile technique for monitoring growth in ultrahigh vacuum epitaxial growth systems. For instance, typical MBE parameters of substrate temperature, growth rate, alloy composition and thickness of growing layers have been measured during the growth of heterostructures in solid-source and gas-source MBE. The growth of AIAs/GaAs quantum wells has also been investigated in studies where the growth was monitored in real time with and without growth interruption. The difference in interfacial abruptness of the heterojunction was then determined. This paper first discusses some practical considerations of implementing an SE onto a gas-source MBE. Examples of monitoring substrate temperature, oxide desorption, surface smoothing and heterojunction growth of a test structure will then be presented followed by a demonstration of quantum well growth and growth interruption. The first use of high-temperature GaAs and AlAs optical constants for thickness and alloy composition determination at the MBE growth temperature is presented which enabled the growth and calibration of distributed Bragg reflectors for use in vertical cavity lasers and modulators.

1. Introduction S e m i c o n d u c t o r p h o t o n i c devices such as vertical cavity surface e m i t t i n g lasers ( V C S E L ) [1,2] a n d e l e c t r o - o p t i c spatial light m o d u l a t o r s ( S L M ) [3] have t h e p o t e n t i a l to b e c o m e t h e basic c o m p o n e n t s for c h i p - t o - c h i p fiber optic b a s e d optical i n t e r c o n n e c t s a n d f r e e - s p a c e optical signal p r o cessors a n d c o m p u t e r s . T h e c o m p l i c a t e d device s t r u c t u r e s have strict m a t e r i a l s r e q u i r e m e n t s that pose c e r t a i n p r o b l e m s for crystal growers. F o r instance, in a vertical cavity laser the gain m e d i u m is a q u a n t u m well i m b e d d e d into a d i s t r i b u t e d B r a g g r e f l e c t o r ( D B R ) cavity. T h e D B R m i r r o r p e a k r e f l e c t a n c e b a n d a n d t h e q u a n t u m well emission n e e d to b e " t u n e d " to the s a m e e n e r g y for efficient device o p e r a t i o n . T h e q u a n t u m well r e q u i r e s m o n o l a y e r c o n t r o l over a p p r o x i m a t e l y 100 ~, to o b t a i n an a c c u r a t e emission w a v e l e n g t h while the q u a r t e r - w a v e l e n g t h D B R layers (several h u n d r e d ~mgstr6ms thick) n e e d a high d e g r e e o f

l a y e r - t o - l a y e r thickness a n d alloy c o m p o s i t i o n u n i f o r m i t y to e n s u r e o p t i m u m reflectance. This p a p e r a d d r e s s e s the issues in realizing r e p r o d u c i b l e I I I - V s e m i c o n d u c t o r p h o t o n i c devices grown by m o l e c u l a r b e a m epitaxy ( M B E ) a n d gas-source M B E ( G S M B E ) . T h e use o f in situ s p e c t r o s c o p i c e l l i p s o m e t r y for r e a l - t i m e m o n itoring a n d c o n t r o l o f M B E growth is discussed as a p p l i e d to the growth of q u a n t u m wells a n d D B R s t h a t a r e the basic c o m p o n e n t s o f e l e c t r o - o p t i c m o d u l a t o r s a n d V C S E L s . U s i n g the h i g h - t e m p e r a t u r e A l A s optical c o n s t a n t s t h a t we have m e a s u r e d has e n a b l e d t h e c a l i b r a t i o n a n d m o n i t o r i n g of thickness a n d alloy c o m p o s i t i o n at t h e growth t e m p e r a t u r e for the first time.

2. Ellipsometry review S p e c t r o s c o p i c e l l i p s o m e t r y (SE) is b a s e d on m e a s u r i n g t h e c h a n g e in p o l a r i z a t i o n state o f the

0169-4332/93/$06.00 © 1993 - Elsevier Science Publishers B.V. All rights reserved

2

(;.N. Mara<'as et al. / hi ~itu SE in MBE t~r photoni<' det'icc~

reflected light from a surface at multiple wavelengths. A complex reflection coefficient ratio p is measured from the reflected light which is defined as the quotient of the complex reflection coefficients for light polarized parallel (Rp) and perpendicular (R~) to the plane of incidence. In the simplest case of the two-phase model that consists of ambient (air or vacuum) / substrate with no additional layers, Rp and R~ represent the Fresne[ reflection coefficients of the system. Usually, p is transformed into the "ellipsometric parameters" t// and ,.1, which characterize the polarization state of the reflected light according to the relation:

R. P

[ Rp I

--

e i~'~,' 's,~= tan(~/J) e ia.

(1

where tan(d+) is the amplitude change and A is the phase difference between the p and s components of the electric field. The SE experimental measurements are expressed as ~l~(hu,, q)/) and :3(h~,,, 4)), where hu, is the photon energy and 4>, is the external angle of incidence. Measurements in an MBE system [4-7] are most often performed at a fixed angle because the optical port geometry is fixed. Varying the angle of incidence provides more information on a particular structure enabling a more accurate structure model to be achieved. This technique called variable-angle spectroscopic ellipsometry (VASE) is usually performed outside the growth chamber on a go_ niometer stage. The measured ellipsometric parameters, ~k and A, are sensitive to the structure (i.e., layer thicknesscs, alloy compositions, microstructure, etc.) and the optical constants of the sample. A model for the MBE structure is first constructed with initial estimates, for instance, of layer thickness and alloy composition and the parameters varied to generate ellipsometric curves that fit the measured data. Data fitting is achieved by minimizing the mean square error (MSE) between the measured data and the generated eIlipsometric fit. The number of angles needed increases with the complexity of the structure being analyzed. We find that five angles are usually sufficient for

most MBE structures analyzed. The MSE is given by

MSE

'~i~l + ( a ? ''~ - a;~ ...... )-~],

(_~)

where the sum is taken over all measured wavelengths.

3. S E / M B E system overview The ellipsometer system is of a rotating analyzer design and can measure ~ and ~1 over a wavelength range of 250-1000 nm. Optical and electronic filters allow measurements in ambient lighting from the effusion cells. The reader is referred to ref. [4] which describes the in situ e[lipsometric system attached to the MBE chamber in more detail and epitaxial reactor design considerations for implementing in situ optical growth measurement systems. The e]lipsometer ports were designed to bc at an angle of 75 ° to the substrate normal. This angle was chosen because it is the optimum angle for studying GaAs and related compounds. Wavelength tunabi[ity is achieved by using a xenon arc light source and monochromator whose chopped output passes through an optical fiber, to a tilt stage and polarizer that are attached to strain-free windows on the MBE system. A tilt stage and stainless steel bellows assembly provides optical alignment so that all optical elements can be made normal to the light beam. The reflected beam is analyzed by a silicon photodetector housed in an enclosure with a rotating analyzer. Strain-free windows (Bomco Co.) are used to minimize birefringence effects. In the gas-source MBE system, additional gate valves are used to reduce the window coating effects of As~ and P.. The externally pumpable valves also allow removal of the windows and R H E E D screen for cleaning without requiring system venting to atmosphere. Whereas the in situ measurements can only be carried out using a single angle of incidence, the ex situ VASE has the additional

G.N. Maracas et al. / In situ SE in MBE for photonic det,ices

capability of varying the angle of incidence to analyze complex structures. The ellipsometer hardware is modular in design so that the input and output optical assemblies can easily be removed from the MBE system and mounted onto the ex situ VASE goniometer. The ellipsometer, goniometer and data acquisition are fully automated and controlled by an 80386 microcomputer, which is also used for data analysis. Alignment can typically be performed in 15 rain. This system was designed and constructed in conjunction with the J.A. Woollam Co. in Lincoln, Nebraska. All layers were grown in a VG V80H dualchamber M B E system; one chamber equipped with conventional solid sources and the other with elemental Group III and gaseous Group V sources.

4. Substrate temperature It has been shown that the temperature of a semiconductor surface can be determined by spectroscopic ellipsometry [4,8] if the optical constants as a function of temperature are known a priori. The procedure is to measure a known structure and fit to the temperature-dependent optical constants with temperature as the variable. The ultimate accuracy of the temperature determination depends on the accuracy of the optical constant information and presence of surface roughness. Using the optical constants for GaAs in ref. [8] we have obtained agreement between optical pyrometer and SE measurements of GaAs substrate temperature within 10°C. Thermocouple readings are notoriously inaccurate, differing by as much as 180°C [4] from both optical measurements in growth systems where the thermocouple is not in direct contact with the substrate.

5. Optical constants of AlAs versus temperature Ultimately the success of an in situ diagnostic tool is its applicability to monitor processes occurring under actual operating conditions (tern-

3

perature, pressure, appropriate time scales, etc.). For I I I - V semiconductor epitaxial growth, this requirement is that alloy composition and thickness be measurable at the growth temperature (400-700°C). There is a noticeable lack of hightemperature optical-constant data in the literature which has compelled us to obtain such constants for GaAs and AlAs. Previous optical-constant temperature ranges [10,11] have been limited to below temperatures where G r o u p V surface desorption roughens the surface and makes ellipsometric analysis ambiguous. Because MBE has the capability of producing high Group V surface overpressures, a stoichiometrically intact (smooth) surface can be maintained to high temperatures. We have used this advantage to measure the optical constants of AlAs up to approximately 700°C. The temperature-dependent optical constants will be published elsewhere [9]. The structure used to measure the optical constants was 10 000 A of AlAs grown on a semi-insulating GaAs substrate without a GaAs or As cap layer. The temperature of the GaAs substrate was calibrated with the ellipsometer and an optical pyrometer as discussed previously. The angle of incidence (75.2 °) was also determined by SE to reduce errors due to angular uncertainties. Ellipsometric scans were fit to the structure using optical constants, n and k as independent fitting parameters. Optical constants were extracted at room temperature and 400-700°C in 25°C increments after allowing sufficient time for the substrate temperature to stabilize. Knowledge of the high-temperature AlAs optical constants has enabled the thickness and alloy composition extraction of structures presented in this paper at the growth temperature of 623°C.

6. Dynamic growth rate modeling One important p a r a m e t e r in epitaxial layer growth is the growth rate. Reflection high-energy electron diffraction ( R H E E D ) is typically used because one cycle of intensity oscillations corresponds to the growth of one atomic layer of material. Three problems with this surface-sensi-

4

(;.N. Maracas et al. / ht ~'im SE in M B E / o r photonic decices

tive technique arc that it cannot be used with substrate rotation, the oscillation intensity decays after the growth of tens of monolayers and oscillations occur only in layer-by-layer growth mode. SE, in comparison, can be performed under substrate rotation. Depending on the incident optical energy uscd in SE (typically 1-4 eV), the light can sample both the surface and the entire w)lume of the epitaxial layer as opposed to R H E E D (with energies of 5-20 keV) which has a sampling depth (electron penetration depth) of a few ~'mgstr6ms. For this same reason SE data can bc taken in either layer-by-layer or island growth mode. If the optical constants at the growth temperature are known, then measurement of ellipsometric parameters can be made and thickness and alloy composition can be extracted as a function of time (dynamically) during the growth. We find that after an initial spectroscopic scan, only three wavelengths are required to obtain a reasonablc accuracy in growth rate determination. Typically data is taken at a rate of 3.3 seconds per wavelength. In an AlAs test structure the SE determined growth rate of (t.77 M L / s agrees within 10% with the 0.83 ML/s measured using the R H E E D intensity oscillation technique. This discrepancy occurs because the growth rate on thick structures can decrease with effusion cell cooling while the shutter is open. Since SE measures thc entire structure, it measures a growth rate avcraged over any effusion cell flux temporal variations and thus can yield lower growth rates than R H E E D . This is illustrated more completely in the DBR growth section.

7. Quantum well growth Quantum devices requirc thickness and heterojunction abruptness control on the order of one monolayer (ML) to produce narrow energy optical transitions that occur at the designed energy. We were able to observe the QW thickness and interface properties using SE during the growth cycle. Two single quantum well structures were grown consecutively under identical growth conditions

I

..... 16

interrupt on "inverted" I no interrupt on "inverted" I ~ ....

14

_2

X

3550/~. line

;.i;,s" ' 'i'n; ' '~' , "~ ' ]nt[ '~ ; ; ; ; .... ! 's'o'lidt (I) \

~

. . . . .

., GaA/ ",

V Vc~rv~

~

GaAs

"-

10 8

~ .. [ GaAs ~ ' ~ ' / cap 6 ~ ~ t ~ AA s ~ ~ dc:t~:e d (NI) 4 . . . . i . . . . . . ~'.. ~. . . . . , . . . . . . . . . . . . . . . . . 5 6 7 8 9 10 11 12 13 t(min) Fig. 1. T r a c k i n g of two q u a n l u m well g r o w t h runs. O n e q u a n t u m well h a d a 60 s i n t e r r u p t i o n at the " i n v e r l e d ' " G a A s - o n - A I A s s u r f a c e while l h c o t h e r h a d n o n e .

except that one had a 60 s growth interruption a! the bottom GaAs-on-A1As "inverted" interface. Thcsc will be called sample I (with interruption) and N) (no interruption) for convenience. The "inverted" interface has been found to produce poor electrical properties in high electron mobility transistors because of impurity incorporation and roughness caused by alloy intermixing at the interface. This effect is also responsible in some cases for broad luminescence transition linewidths. Two test samples were grown consisting of a 100(} ,~ GaAs buffer layer, folk)wed by a 100 ,~ AIAs barrier, a 70 A GaAs well, a 100 ,~ AIAs barrier and finally a 20 ,~ GaAs cap layer. Both structures were undoped, grown at 600°C with a 60 s growth interruption at the top A 1 A s / G a A s QW interface. One QW structure (NI) had no growth interruption at the bottom GaAs-on-AIAs interface and the other layer (1) had a 60 s interruption. Fig. 1 shows 4J (at A = 355(I A) as a function of time during the growth of the two quantum wells with and without a 60 s interruption at the "'inverted" interface. The wavelength of 3550 A, was used for display purposes because it was more sensitive to near-surface phenomena due to its small absorption depth at this temperature. The sloped characteristics of the individual AIAs lay-

G.N. Maracas et al. / In situ SE in MBE for photonic decices

ers arise from the layers being optically thin at this wavelength. During the growth interruption time (labeled "int."), the slopes of the curves at this wavelength are approximately zero. Fits to the ellipsometric spectra yielded a quantum well width of 70 A and barriers of 100 * indicating that the growth of the wells was successful. Independent photoluminescence measurements confirmed the well thickness in both samples and showed a difference in the n = 1 heavy-hole transitions of less than 0.5 meV indicating high reproducibility of the growth process. Shoulders of 4 m e V on either side of the main photoluminescence transition indicated that the quantum well grown without interruption (NI) had interface roughness on the order of _+1 monolayers. The measured shoulder energies agreed very well with solutions to Schr6dinger's equation for quantum wells having + 1 ML differences in width. SE analysis [8] extracted an interfacial layer thickness in the quantum well of 1.4 ML which agrees with the photoluminescence. Thus it is demonstrated that SE can have thickness sensitivity on the order of a monolayer and can be used to monitor the width of a quantum well during the growth process. This information allows the grower to control the thickness by adjusting the growth rate a n d / o r time for each layer without requiring removal from the M B E for ex situ material structure evaluation.

8. Bragg reflectors

Dielectric mirrors require layer thickness uniformity among successive periods to produce high reflectance in a given wavelength range. This uniformity is a result of interface flatness and a random or systematic variation thickness in successive layers. Interface roughness degrades the peak reflectance [12] while systematic and random thickness variations skew and increase the side lobe reflectance, respectively. Our calculations show that the thickness deviation required to maintain near-peak reflectance in a D B R stack is less than 8% for a 20-period DBR. The design of a mirror presented here with stop band cen-

5

tered at 867 nm consists of AIAs/AI0.2GaAs quarter-wavelength layers of thicknesses 728 and 628 A, respectively. Thus a maximum 8% variation imposes the requirement that the optical thickness ( = layer thickness/index of refraction) variation among the layers be less than 58 and 50 A, respectively. This tolerance must be maintained over a relatively thick (2.7/xm) 20-period mirror with control over both epitaxial layer thickness and alloy composition. The problem is increased in a vertical cavity structure where two Bragg reflectors must be stacked with an intermediate quantum well gain region. An in situ monitoring and control technique such as SE is thus desirable for such complicated structures. A growth temperature of 623°C measured by optical pyrometer or 616°C by SE (750°C as measured by the MBE thermocouple) was used to grow the 20-period D B R with thicknesses mentioned previously centered at 867 nm. To reduce interface roughness, the Al0.2Ga0.sAs quarterwavelength layers were approximated by a GaAs and A10.3Ga0.7As superlattice. The mirror was grown on a 3300 A n+-GaAs buffer on an n + substrate. The AlAs optical constants [9] for the growth temperature were used to fit the ellipsometric data for thickness and alloy composition. Before growth, a full spectroscopic scan (200-770 nm) was performed on the GaAs substrate to measure the temperature. @ and A were then measured at three wavelengths: 3500, 5500 and 6500 A at a rate of one point (three wavelengths) every three seconds. Full spectroscopic scans were performed after the growth of periods 1, 2-5, 6-10, 11-15 and 16-20 to illustrate the thickness variation among successive layers and to monitor the evolution of the mirror reflectance. The growth rate was calibrated by SE after the growth of period 1. Fig. 2 is a plot of the 0(t, 3500 ,~) data with the a f o r e m e n t i o n e d intervals superimposed. While 4, and A for all three wavelengths were used in the structural calculations, only ~ is plotted for illustrative purposes. The time axes on the successive mirror portions were shifted such that each portion began at t = 0. The numbers in fig. 2 indicate the times where the data for each portion ended and the arrows denote the AlAs o

(;.N. Maracas et al. / In situ SE in M BE for photonic deHces

6

a n d Al0.2Ga0.~As superlattice (SL) layers. The reproducibility of ~ in s u b s e q u e n t l y grown layers shows that there is a high degree of optical thickness uniformity a m o n g all the periods of the D B R i n d i c a t i n g a high effusion cell flux stability. T h e thickness a n d growth rate were calculated dynamically d u r i n g the growth a n d c o m p a r e d well with the R H E E D oscillations. R Ho E E D m e a s u r e d growth rates of 97.6 a n d 154.4 A / m i n for AlAs a n d the Al0.2Ga0.sAs SL layer;s, respectively, while SE m e a s u r e d 95.3 a n d 156.8 A / r a i n for AlAs and the Al0.2Ga0.sAs SL layers, respectively. An eff e c t i v e - m e d i u m a p p r o x i m a t i o n was used for analyzing the superlattice regions. This a s s u m p t i o n of a u n i f o r m average alloy composition for the superlattice proved to be sufficiently accurate fl)r this application. Fig. 3 shows SE scans after the growth of thc various m i r r o r portions with the substrate scan i n c l u d e d for reference, to is plotted since it is the ratio of the a m p l i t u d e c h a n g e u p o n reflection (eq. (1)) and thus is more closely related to the n o r m a l i n c i d e n c e reflectance t h a n is A. It is i n t e r e s t i n g to note that the m a x i m u m value of to saturates with increasing n u m b e r of periods at

AlAs AIGaAs 742/~, 605A i ....

18 16

i ....

~

T=623 C f ....

/, 3500A i .... i ....

~

eriods

~ -5

j

11.15 6-10

14

6

,

0

,

,

i

10

,

,

,

i

20

•

,

30 40 Time in minutes

50

60

Fig. 2. Dynamicgrowthmonitoring of a 20-periodAIAs/AIGaAs distributed Bragg reflector in the MBE at the growth temperature of 623°C. The periodicity of the mirror is evident in ~ as a function of time. Superimposed are the curves of periods 1, 2 5, 6-10, 11 15 and 16-20 identified when the periods end. The time axis has been shifted such that successive period scans begin at t = 0. A high degree of thickness and alloy compositkm reproducibility is evident from the superimposed curves.

4 0

35

AIAs/AIGaAs BDR measured in situ at 6 2 3 C

Z

20

~

20 periods 15 periods

)~:/.---~-~.~.~ . ~ .... "~#

10

J

eriods

I"~

,

15

~ 0

''

5000

substrate

. . . .

6000

7000

8000

9000

Wavelength (A)

Fig. 3. Mcasurement of ~// versus wavelength of a distributed Bragg reflector at the growth temperature of 623°(" in the MBff. The spectra were taken belt)re the epitaxial structure was grown and after the growth of 1, 5, 1(), 15 and 20 periods of the mirror. The tingle of incidence of the incident light was 75L

approximately 15, as does the n o r m a l incidence reflectance. Evolution of the side lobes is also evident in the g/(A) curves for periods greater than 5. It should be recalled that to and ,3 are measured at oblique incidence (75 ° ) a n d c a n n o t be converted to an absolute, normal incidence reflectance using just a two-layer model. T h e normal incidence reflectance can, however, be obt a i n e d from tO and A if the structural i n f o r m a t i o n (thicknesses a n d optical constants) is used and an analysis similar to the scattering matrix approach [13] performed. Shown in fig. 4 is a c o m p a r i s o n of the norma[ incidence reflectance and to which was m e a s u r e d at 75 ° . This data was m e a s u r e d at room t e m p e r a ture outside of the MBE. T h e most n o t a b l e feature is the shift in the stop b a n d . This arises partly from the difference in the angle of incid e n c e b u t also from n o n - u n i f o r m i t y in the G r o u p III flux m a k i n g the position of the in situ meas u r e m e n t on the substrate different than that of the ex situ m e a s u r e m e n t . N o r m a l incidence reflectance was m e a s u r e d against a s t a n d a r d reflector using a m e a s u r e m e n t system of our own design which a l t e r n a t e s the reference and reflected b e a m s onto one detector.

G.N. Maracas et al. / In situ SE in MBE for photonic decices 1

0.9 0.8

.. /

~0.7

40

:

0.6

"1~

~- 0.5

30 -6

f"

.

/

'\

20 "V"\ ..',~

0.4

"/

"J

...... "Ro'~m iemp 0.2 ~ - - ' - - ' - - ' J - - ' - - ~ 5000 6000 7000 8000 9000 Wavelength (A)

10

0.3

0 10000

Fig. 4. Room-temperature normal incidence reflectance of the 20-period D B R and corresponding oblique angle (75 °) ~ versus wavelength. The peak normal incidence reflectivity was measured to be 0.995 _+0.004 at A = 8630 .~.

Plotted in fig. 5 is the calculation for the nominal design structure centered at 8670 A, the calculation for a mirror using the ellipsometrically determined thicknesses and compositions (at the growth temperature of 623°C) and the measured reflectance of the DBR discussed in this section. Thicknesses of the AIGaAs (superlattice) and AlAs were 606.2 and 742.5 ,~, respectively, and x = 0. 19 and 1.0, respectively, as measured by SE at the growth temperature. The peak reflectance was measured to be 0.995_+ 0.004 at

1

,

,

,

i

,,,~.': 0.8

'

.....

"-'~'=~'i

-,

measured (8638A)

it"

17

ore0.6

-

", '\i ';

',

calculated using SE data (8630A)

"5 m

"~ 0.4 rr"

design

(8670A)

0,2

~

0

8000

,

,

,

I

8200

,

i

i

i

.

.

.

i

;

8400 8600 Wavelength (A)

.

.

i

8800

,

.

.

9000

Fig. 5. Normal incidence reflectance of a 20-period AIAs/A1GaAs distributed Bragg reflector. The mirror was designed for A = 867 nm. The as-grown structure parameters were measured by SE at 623°C. Measured and calculated reflectance curves for the mirror are shown.

A =8638 ,~ (32 ,~ from the design value) as compared to the calculated 0.997 at 8630 ,~. This reflectivity is higher than any previously reported for a 20-period DBR. In summary, it was shown that SE is a viable in situ monitoring technique for growth of I I I - V photonic devices. SE can be used to measure substrate temperature, growth rate, alloy composition, thickness and heterojunction interface properties. It can also be used to monitor the evolution of reflectivity and the stop band in distributed Bragg reflectors. Combination of all these aspects of SE with MBE can thus provide the material grower with information for control of complicated structures required to realize reproducible photonic device operating characteristics.

Acknowledgements We wish to acknowledge Chuck Wheeler for assistance with the reflectance calculations. This work was supported by D A R P A contract #N00014-89-J-3120 and A F O S R c o n t r a c t AFOSR-90-0118-DEF.

References [1] J.L. Jewel, A. Scherer, S.L. McCall, Y.H. Lee, S. Walker, J.P. Harbison and L.T. Florez, Electron. Lett. 25 (1989) 1123. [2] F. Koyama, S. Kinoshita and K. Iga, Appl. Phys. Lett. 55 (1989) 221. [3] D.A.B. Miller, D.S. Chemla, T.C. Damen, A.C. Gossard, W. Wiegman, T.H. Wood and C.A. Burrus, Appl. Phys. Lett. 45 (1984) 13. [4] G.N. Maracas, J.L. Edwards, K. Shiralagi, K.Y. Choi, R. Droopad, B. Johs and J.A. Woollam, J. Vac. Sci. Technol. A 10 (1992) 1832. [5] D.E. Aspnes, Thin Solid Films 89 (1982) 249. [6] D.E. Aspnes, W.E. Quinn and S. Gregory, Appl. Phys. Len. 56 (1990) 2569. [7] B. Johs, J.L. Edwards, K.T. Shiralagi, R. Droopad, K.Y. Choi, G.N. Maracas, D. Meyer, G. Cooney and J.A. Woollam, Materials Research Society, Spring 1991 Meeting, Anaheim, CA. [8] H. Yao, P.G. Synder and J.A. Woollam, J. Appl. Phys. 70 (1991) 3261. [9] G.N. Maracas and J.L. Edwards, to be published.

8

G.N. Maracas et aL / In situ SE in MBE for photonic derices

[1(1] M. Garriga, P. Lautenschlagcr, M. Cardona and K. Ploog, Solid State C o m m u n . 61 (1987) 157. [1 l] E.D. Palik, in: Handbook of Optical Constants of Solids 11 (Academic Press, San Diego, CA, 1991) pp. 489 499. [12] J.D. Walker, K. Malloy, S. Wang and J.G. Smith, Appl. Phys. Lelt. 56 (1990) 2943.

[13] R.M.A. Azzam and N.M. Bashara, Ellipsometry and Po larized Light (North-Holland, Amsterdam, Iq77) pp. 288 315.