Full text of DOI:10.1007/BF02665872

HgCdTe Growth on (552) Oriented CdZnTe
Journal of ELECTRONIC MATERIALS, Vol. 30, No. 6, 2001
Special Issue P7a7pe9r
by Metalorganic Vapor Phase Epitaxy
HgCdTe Growth on (552) Oriented CdZnTe
by Metalorganic Vapor Phase Epitaxy
P. MITRA,1 F.C. CASE,1 H.L. GLASS,2 V.M. SPEZIALE,2 J.P. FLINT,2
S.P. TOBIN,3 and P.W. NORTON3
1.—Lockheed Martin Missiles and Fire Control–Dallas, P.O. Box 650003, Dallas, TX 75265-0003.
2.—Honeywell Electronic Materials, 15128 E. Euclid Avenue, Spokane, WA 99216. 3.—BAE
SYSTEMS, 2 Forbes Road, Lexington, MA 02421-7306
WereportthegrowthofHgCdTeon(552)BCdZnTebymetalorganicvaporphase
epitaxy (MOVPE). The (552) plane is obtained by 180∞ rotation of the (211) plane
about the [111] twist axis. Both are 19.47 degrees from (111), but in opposite
directions. HgCdTe grown on the (552)B-oriented CdZnTe has a growth rate
similar to the (211)B, but the surface morphology is very different. The (552)B
films exhibit no void defects, but do exhibit ~40 mm size hillocks at densities of
10–50 cm–2. The hillocks, however, are significantly flatter and shorter than
those observed on (100) metalorganic vapor phase epitaxy (MOVPE) HgCdTe
films. For a 12–14 mm thick film the height of the highest point on the hillock is
less than 0.75 mm. No twinning was observed by back-reflection Laue x-ray
diffraction for (552)B HgCdTe films and the x-ray double crystal rocking curve
widths are comparable to those obtained on (211)B films grown side-by-side and
withsimilaralloycomposition. Etchpitdensity(EPD)measurementsshowEPD
values in the range of (0.6–5) ¥ 105 cm–2, again very similar to those currently
observed in (211)B MOVPE HgCdTe. The transport properties and ease of
dopant incorporation and activation are all comparable to those obtained in
(211)B HgCdTe. Mid-wave infrared (MWIR) photodiode detector arrays were
fabricated on (552)B HgCdTe films grown in the P-n-N device configuration
(upper case denotes layers with wider bandgaps). Radiometric characterization
at T = 120–160 K show that the detectors have classical spectral response with
a cutoff wavelength of 5.22 mm at 120 K, quantum efficiency ~78%, and diffusion
currentisthedominantdarkcurrentmechanismnearzerobiasvoltage. Overall,
the results suggest that (552)B may be the preferred orientation for MOVPE
growth of HgCdTe on CdZnTe to achieve improved operability in focal plane
arrays.
Key words: Mercury cadmium telluride, metalorganic vapor phase epitaxy
(MOVPE), cadmium zinc telluride, infrared detector
the demonstration of a variety of in-situ grown p-n
junction photodiode detectors as well as emitters.1–7
The surfaces of (100)-oriented films are normally
found to exhibit pyramidal hillock defects, while the
(211)B-oriented films exhibit void defects.5,8–10 Both
types of macrodefects cause the photodiodes overlap-
ping those regions to be leaky or shorted.
There have been numerous studies on the nature
and causes for macro-defects on the surface of
HgCdTe grown by MOVPE and their orientation
dependence.7–11 These defects nucleate on the sub-
strate-film interface through particulates, precipi-
tates, or due to the chemical modification of the II/VI
INTRODUCTION
Theorientationofthelattice-matchedCdZnTesub-
strates has a profound effect on the surface morphol-
ogy and the defect structure of the HgCdTe films
grown by vapor phase epitaxy. In metalorganic vapor
phase epitaxy (MOVPE) growth of HgCdTe, the pre-
ferred CdZnTe orientations are (100) misoriented 2–
8∞ towards (110) or (111)B planes, and (211)B. These
substrate orientations have been successfully used
for the growth of high quality HgCdTe films and for
(Received November 25, 2000; accepted February 19, 2001)
779

780
Mitra, Case, Glass, Speziale, Flint, Tobin, and Norton
ratio of the substrate surface induced by an etchant.
On (100)-oriented GaAs substrates pre-treatment by
alkalinemetalcontainingsolutionshasbeenreported
tosignificantlyreducethedensityofhillockdefectson
theHgCdTeepilayers.12–14 Alternateorientations,such
as the high index planes, have surfaces that can be less
susceptible to macro-defect formation or the defect
geometry itself may be more forgiving to the devices.
InthispaperwereportMOVPEinterdiffusedgrowth
of HgCdTe on the (552)-oriented CdZnTe. The (552)
plane is obtained by 180∞ rotation of the (211) plane
about the [111] twist axis. The (211) and (552) planes
both lie 19.47 degrees from (111), but in opposite
directions. HgCdTe films grown on the (552)B-ori-
entedCdZnTehavegrowthratessimilartothe(211)B
but the surface morphologies are very different. The
(552)B films exhibit no void defects but do exhibit
hillocks at densities of 10–50 cm–2. The hillocks,
however, are significantly flatter and smaller than
those observed on (100) films of similar thicknesses.
The (552)B HgCdTe film surfaces were characterized
by Nomarski optical and atomic force microscopy.
Epitaxial quality was assessed by x-ray diffraction
and etch pit density counts. The transport properties
of the n-type films were characterized by Hall effect
and lifetime measurements. Mid-wave infrared
(MWIR)photodiodefilmsgrownintheP-n-Nconfigura-
tion6 were fabricated into back-illuminated variable-
areadetectorarrays(heretheuppercasedenoteslayers
of wider bandgaps). The detectors were tested for
spectral response, quantum efficiency, current-volt-
age curves, and dynamic resistance-area products.
a
SUBSTRATES AND HgCdTe GROWTH
The(211)planesarelocated19.47degreesfrom(111)
toward the (100). At an equal angle, but in the opposite
direction, are the (552) planes. Since twins in CdZnTe
can be described as having an orientation derived from
the parent or matrix by a 180 degree rotation around a
<111> direction, the (552) can be observed as a twinned
region on a (211) wafer. In fact, our initial observation
of MOVPE growth of HgCdTe on (552) was on such a
twinned region that had previously gone unnoticed.
The (552) planes, like the (211), exhibit polarity
associated with the nearest (111) plane. This can be
revealed by the standard polarity etch15
(HF:HNO3:lactic acid in ratios 1:1:1). Etching also is
commonly used to reveal defects in CdZnTe sub-
strates. Attempts to use the defect etch developed for
the (211) orientation16 gave encouraging results. Dis-
locations were revealed on the B-faces, as previously
identifiedusingthepolarityetch.On(552)substrates,
however, dislocation etch pits were not as sharply
defined as on (211). After the defect etch, the polarity
of (211) substrate surfaces can be identified with the
unaided eye, but microscopic examination was re-
quired to identify polarity in the (552) substrates.
MOVPE growth of HgCdTe was performed using
the interdiffused process that has been previously
described in detail.5 Most of the films were grown on
the(552)BfaceoftheCdZnTesubstrates.Inanumber
b
Fig. 1. Nomarski optical micrographs at two different magnifications of
HgCdTegrownon(552)BorientedCdZnTe.Thehillock(a)isthesame
one shown in (b) at greater magnification.
of cases the films were grown side-by-side with the
(211)B substrates for comparison. The growth rates
on the two orientations are closely comparable, al-
though not identical. Thus completely new calibra-
tion was not necessary to achieve comparable x-
valuesandthicknesses.Thefilmsweretypicallygrown
tothicknessesof9–14mmwithHg1–xCdxTex-valuesin
the range of 0.21–0.40. For n-type doping the films
were doped with iodine using the precursor ethyl
iodide.17 Iodine incorporation rates in the (552)B-
oriented films were found to be within a factor of two
of the incorporation rate in (211)B-oriented HgCdTe.
This result is in marked contrast to the strong orien-
tation dependence of the iodine incorporation ob-
served18 with the tilt angles between the (100) and
(111) planes, but not unexpected, due to the similar
structures of the (552)B and (211)B surfaces. No
orientation dependence was observed for arsenic dop-
ing, which was performed with the precursor tris-
dimethylaminoarsenic.19

HgCdTe Growth on (552) Oriented CdZnTe
by Metalorganic Vapor Phase Epitaxy
781
surfacewasimagedbyatomicforcemicroscopy(AFM).
Figure 2 shows a 100 mm by 100 mm AFM scan of a
hillock from another 12.5 mm thick (552)B film. A line
scan of the tallest region of the hil-lock (region A)
shows it to be ~7500 Å high and ~15 mm wide. The
flatter region (region B) of the hillock is 1200 Å high
and~25mmwide.Inahillock-freeregionthelinescan
(region C) shows the surface roughness to be ~400 Å.
Back reflection Laue x-ray diffraction patterns
showed the HgCdTe epitaxial layer to be single crys-
tal with (552) orientation. The Laue pattern showed
no evidence of twinning. The Laue pattern of a (552)-
oriented CdZnTe substrate matched the geometry of
the HgCdTe pattern, confirming the film orientation.
A Bede 200 double-crystal diffractometer was used
to investigate the x-ray (CuKa1) rocking curves of a
number of (552)B HgCdTe films. The (331) reflection
and a 1 mm spot size was used to measure the double
crystal rocking curve (DCRC) full width at half maxi-
mum (FWHM). For films grown at 370–380∞C the
FWHM was found to be in the range of 30–40 arcsec.
The presence of twinning on the (552) surface was
studied by first locating the (331) reflection, then
moving the detector and sample appropriately for a
(422)reflectionandrotatingaboutthesamplesurface
normal to locate any (422) reflection present. In most
cases, there was no evidence of twinning. For films
grown at 340–350∞C a 1–3% twin volume was de-
tected. These films also exhibit a relatively wider
FWHM of 90–100 arcsec. Figure 3 shows a plot of the
DCRC FWHM for (552)B films grown in the tempera-
turerangeof340–380∞C. ThenarrowestFWHMarefor
films grown at 375–380∞C with values of 30–35 arcsec.
Thedislocationsinthe(552)Bfilmswerestudiedby
etch pit density (EPD) counts using the Hahnert and
Schenk dislocation revealing etch20 (HF, HNO3, H2O,
CrO3).Thisetchwasfoundtobeeffectiveinproducing
triangular pits on both A and B faces of the (552)
HgCdTefilms.TheEPDcountsforanumberof(552)B
and (211)B films are plotted in Fig. 4. The plot shows
the run-to-run variation during the growth optimiza-
tion process so all the films were not grown under
identical conditions. The growth temperature in all
cases was 380∞C. The median EPD values for the
23 (552)B films and the 15 (211)B films are 5.9 ¥ 105
and 2.6 ¥ 105 cm–2, respectively. The higher median
EPD for the (552)B-oriented films is more due to the
inclusion of the optimization runs rather than due to a
fundamental limitation in obtaining EPD values com-
parable to the (211)B films. In fact, in six of the (552)B
films the EPD is below 105 cm–2, values that are compa-
rable to the lowest EPD values in the (211)B films.
These values are also comparable to state-of-the-art
MBE and LPE grown single layer HgCdTe films.
EPDandthex-valuedepthprofileofamultilayerP-
n-NHgCdTe(552)BfilmisshowninFig.5.Thep-type
layer was doped with arsenic at 5 ¥ 1016 cm–3. The n-
type buffer and absorber layers were doped with
iodine at 4 ¥ 1015 and 1 ¥ 1015 cm–3, respectively. The
x-value depth profile was obtained by secondary ion
massspectrometry(SIMS)performedatCharlesEvans
Fig. 2. AFM Image of a 100 mm ¥ 100 mm area of (552)B HgCdTe.
Fig. 3. X-ray double crystal rocking curve data on HgCdTe films grown
on (552)B CdZnTe in the temperature range of 330–380∞C.
Fig. 4. Etch pit density data comparison between (552)B and (211)B
oriented MOVPE HgCdTe grown in a number of runs.
SURFACE MORPHOLOGY AND DEFECTS
The surface of the HgCdTe grown on the (552)B-
oriented CdZnTe exhibits hillock defects at a density
of 10–50 cm–2. Void defects, commonly observed on
(211)B HgCdTe, however, are not observed. Figure 1
shows the Nomarski optical micrographs of a (552)B
HgCdTe film that is 14 mm thick, at two different
magnifications. The surface is relatively smooth and
a bell-shaped hillock defect is visible. The hillock is
~55 mm long, ~30 mm wide, and has a higher elevation
ononeend.Toobtainaquantitativeheightprofile,the

782
Mitra, Case, Glass, Speziale, Flint, Tobin, and Norton
Fig. 6. Mobility of iodine doped MOVPE (552)B and (211)B HgCdTe
andthetrendlineforLPEHgCdTefilmsofthesamecompositiongrown
recently at BAE SYSTEMS, Lexington MA.
Fig. 5. EPD depth profile in a P-n-N HgCdTe film grown on a (552)B
substrate.
& Associates. The EPD depth profile was measured on
a differentially etched wedge from a 15 mm long strip
ofthefilmwiththicknessvaryingfrom1mmtothefull
film thickness of 13 mm. The wedge-film was then
immersed in the etchant solution to reveal the dislo-
cation pits with varying depths.21 Figure 5 shows the
EPD count to be uniform throughout the depth of the
film with an average value of ~6 ¥ 105 cm–2. Thus for
the range of x-values and doping levels of this multi-
layer film, increased EPD over single layer films are
not observed.
TRANSPORT PROPERTIES
Fig.7.LifetimesofMOVPELWIR(552)BorientedHgCdTefilmsgrown
with two different donor concentrations. The solid and dotted lines are
model calculations based on the Auger-1 and radiative recombination
processes for ND-NA of 2 ¥ 1015 and 1 ¥ 1015 cm–3, respectively.
The Hall parameters of a number of n-type single
layer (552)B HgCdTe films with x-values in the range
of 0.20 to 0.43 were measured at 77 K and a magnetic
field of 3.2 kgauss. For undoped films the background
carrier concentrations are in the range of (1–3) ¥
1014 cm–3. The 77 K mobility data for iodine-doped
films, with carrier concentrations in the range of
(0.5–2) ¥ 1015 cm–3, are plotted in Fig. 6. Also plotted
for comparison are data from similar MOVPE (211)B-
orientedfilms.Figure6showsthatthemobilitydatafor
the (552)B and (211)B HgCdTe films are comparable
and closely follow the trend line for recent Te-rich LPE
grown HgCdTe at BAE SYSTEMS, Lexington, MA.
The iodine-doped (552)B n-type films were charac-
terized for temperature dependent lifetimes by the
non-contact transient mm-wave reflectance tech-
nique.22 Inverse temperature dependent lifetimes for
two long-wave infrared (LWIR) composition films
(x ~ 0.23) with Hall carrier concentrations of 1.3¥ 1015
and 1.8 ¥ 1015 cm–3 are plotted in Fig. 7. Also plotted
for comparison are model calculations for the tem-
perature dependent lifetimes based on Auger-1 and
radiative recombination at donor concentrations of
1 ¥ 1015 and 2 ¥ 1015 cm–3. For the HgCdTe layer doped
at 1.3 ¥ 1015 cm–3 the lifetime is 0.80 msec at 77 K and
peaks at 1.85 msec at 140 K. Figure 7 shows that the
measured lifetimes fall within the calculated ranges
for the two carrier concentrations.
DETECTOR ARRAY RESULTS
For characterizing the quality of detector perfor-
manceanMWIRP-n-Nheterojunction(552)BHgCdTe
filmwasfabricatedintoback-illuminated,mesa-etched
variable-areadiagnosticphotodiodearrays.23 Thetest
array consisted of 43 widely separated circular diodes
withjunctiondiametersrangingfrom40mmto350mm.
It was indium bump-mounted to a circuit board and
illuminated through the CdZnTe substrate. Spectral
response was classical in shape with a cutoff wave-
length of 5.22 mm at 120 K (Fig. 8). Quantum effi-
ciencywasmeasuredusingacalibrated1000Kblack-
body and a narrow-band filter at 3.0 mm. From the
area dependence of detector responsivity, a 1-dimen-
sional quantum efficiency of 78% and a lateral collec-
tion length of 38 mm were determined. The quantum
efficiency is within experimental accuracy of the 79%

HgCdTe Growth on (552) Oriented CdZnTe
by Metalorganic Vapor Phase Epitaxy
783
Fig. 8. Normalized quantum efficiency as a function of wavelength.
Fig. 10. Resistance-area product as a function of bias and temperature.
Fig. 11. Resistance-areaproduct, usingtheopticallyactivearea, forall
elements of the test array. RoA is measured at zero bias and R-40A is
measured at –40 mV.
of 120 K are shown for all detectors in the array in
Fig. 11. The optical area, Aopt, is defined as the area
plusanannulusofwidthequaltothelateralcollection
length. High R0A values, up to 1 ¥ 105 ohm-cm2, and
large impedance improvements at reverse bias are
seen for nearly all detectors. The three detectors with
R-40A value less than 104 ohm-cm2 all had large
hillock defects in the HgCdTe material. These defects
were localized at one end of the array. The near
independence of RAopt with detector size is another
indication that diffusion current from the n-type
absorber layer determines R0A in these devices.
Fig. 9. Current-voltage characteristics at three temperatures.
limitimposedbyreflectionfromtheuncoatedCdZnTe
substrate surface, implying nearly 100% internal
quantum efficiency. Figures 9 and 10 show the cur-
rent-voltage and resistance-voltage characteristics of
one of the diodes at temperatures ranging from 120 K
to 160 K. The exponential voltage dependence with a
junction ideality factor of 1 and the temperature
dependence of R0A both indicate diffusion current is
the dominant mechanism near zero bias. The imped-
ance increases by more than a factor of 30 at reverse
bias, important for minimizing noise when the detec-
tors are used with a silicon multiplexer in a focal
plane array (FPA). Tunneling current, apparent at
large reverse bias at lower temperatures, is not sig-
nificant at the reverse bias of –20 mV typical of FPA
operation. The dynamic resistance-area RAopt prod-
ucts at 0 and –40 mV bias voltage for a temperature
DISCUSSION AND CONCLUSIONS
MOVPE HgCdTe films on (552)B-oriented CdZnTe
exhibit smooth surface morphology with short bell-
shaped hillocks and no void defects. For films as thick
as 14 mm the hillocks at their tallest point are no
higher than 7500 Å. In its flatter region the hillock is
only 1200 Å high. The surface morphology is similar

784
Mitra, Case, Glass, Speziale, Flint, Tobin, and Norton
to that of the off-axis (100) MOVPE HgCdTe films,
whichexhibitmuchtallerpyramidal-shapedhillocks,
and also no void defects. An off-axis (100) HgCdTe film
that is ~14 mm thick exhibits hillocks that are signifi-
cantlylarger(60–80mmlong)andtaller(4–7mm).The
hillock density in a (552)B film is typically in the
range of 10–50 cm–2. Assuming that a hillock would
cause outages on all photodiode pixels delineated in
that area, the reduction in FPA operability is not very
significant. For a 256 ¥ 256 FPA with 40 mm pitch
pixels, a hillock density of 50 cm–2 would cause only
0.23% of the pixels to be out, if each hillock causes 3
pixels to be defective. In off-axis (100) HgCdTe films
the same density of hillocks as in (552)B HgCdTe
would cause a larger number of pixel outages. During
patterning of the (100) HgCdTe films by photolithog-
raphy the larger and taller hillocks cause the photo-
resisttobeadaroundthemaffectingawiderarea, and
thus a larger number of pixels. Thus, purely from the
standpoint of outages caused by macrodefects, higher
operability is expected for FPAs fabricated on (552)B
HgCdTe films.
range of 0.10–0.25%. These estimates of defective
pixel counts would result in significant improvement
in FPA operability than the current standard based
on (211)B films grown by vapor phase epitaxy.
ACKNOWLEDGEMENTS
This work was supported by the U.S. Army Space
and Missile Defense Command Contract, monitored
by Dr. Latika Becker; and by Lockheed Martin inter-
nal R&D funds. We thank L.C. Johnson and A.X.
QuinonezfortheircontributionstotheMOVPEgrowth
experiments, P. Bodie for EPD measurements on the
CZT substrates, and Dr. M.B. Reine for his valuable
comments on the manuscript.
REFERENCES
1. S.J.C. Irvine, Narrow-Gap II-VI Compounds for Optoelec-
tronicandElectromagneticApplications,ed. P. Capper(Lon-
don: Chapman & Hall, 1997), pp. 71–96.
2. P. Mitra, S.L. Barnes, F.C. Case, M.B. Reine, P. O’Dette, R.
Starr,A.Hairston,K.Kuhler,M.H.Weiler,andB.L.Musicant,
J. Electron. Mater. 26, 482 (1997).
3. C.D. Maxey, C.L. Jones, N.E. Metcalfe, R.A. Catchpole, N.T.
Gordon, A.M. White, and C.T. Elliot, SPIE Proc. 3122, 453
(1997).
The x-ray DCRC FWHM of the (552) HgCdTe with
the (331) reflection is in the range of 30–40 arcsec,
values similar to those obtained in high quality epi-
taxial HgCdTe films. EPD counts on the films grown
with optimized growth parameters are found to be in
the range of high 104 to mid 105 cm–2, which are
comparable to the counts obtained on (211)B MOVPE
HgCdTe, and are also similar to state-of-the-art epi-
taxial HgCdTe grown by MBE and LPE. In (552)B
multilayer structures grown in the P-n-N configura-
tion, the dislocation density is found to be relatively
constant, independent of film depth and doping type.
The electrical transport properties of the (552)B
HgCdTefilmsaresimilartothosemeasuredon(211)B
MOVPE HgCdTe. The background doping level, mo-
bilitiesofiodine-dopedfilms, andcarrierlifetimesare
all found to be very similar for the (552)B and (211)B
orientations, and closely follow state-of-the-art data
for LPE HgCdTe. Backside-illuminated MWIR detec-
torsfabricatedfrom(552)BP-n-Nheterojunctionfilms
exhibit characteristics that closely match the perfor-
mance of similar (211)B films. The R0A values at 140 K
are within a factor of two of the one-dimensional
diffusion current model calculation,6 and the quan-
tum efficiency value of 78% is at the limit imposed by
reflection losses at the non-antireflection-coated
CdZnTe substrate surface. The dark current near
zero-bias voltage is limited by diffusion current, and
the spectral response is classical. Those detectors
whose junction areas overlap the bell-shaped hillocks
on the films have significantly lower R0A values (by
two or three orders of magnitude) along with lower
quantumefficiency(QE)andhigherreverse-biasdark
current. The low density (10–50 cm–2) of these
macrodefects, however, will result in relatively few
outagesintwo-dimensionaldetectorarrays. Basedon
the hillock densities and their sizes, FPAs with 30–40
mmpitchdetectorsareexpectedtohaveoutagesinthe
4. M.B. Reine, A. Hairston, P. O’Dette, S.P. Tobin, F.T.J.
Smith, B.L. Musicant, P. Mitra, and F. C. Case, SPIE Proc.
3379, 200 (1998).
5. P. Mitra, F.C. Case, and M.B. Reine, J. Electron. Mater. 27,
510 (1998).
6. P. Mitra, F.C. Case, M.B. Reine, T. Parodos, S.P. Tobin, and
P.W. Norton, J. Electron. Mater. 28, 589 (1999).
7. C.D. Maxey, M.U. Ahmed, C.L. Jones, R. A. Catchpole, P.
Capper, N.T. Gordon, M. Houlton and T. Ashley (Paper
presented at the 2000 U.S. Workshop on the Physics and
Chemistry of II-VI Materials, Albuquerque, NM Oct. 30–
Nov. 1, 2000).
8. P. Capper, C.D. Maxey, P.A.C. Whiffin and B.C. Easton, J.
Cryst. Growth 96, 519 (1989).
9. G. Cinader, A. Raizman, and A. Sher, J. Vac. Sci. Technol.
B9, 1634 (1991).
10. M.J. Bevan, N.J. Doyle, and T.A. Temofonte, J. Appl. Phys.
71, 204 (1992).
11. R. Triboulet, A. Tromson-Carli, D. Lorans, and T. Nguyen
Duy, J. Electron. Mater. 22, 827 (1993).
12. J.Geiss,J.E.Hails,A.Graham,G.Blackmore,M.R.Houlton,
J. Newey, M.L. Young, M.G. Astles, W. Bell, and D.J. Cole-
Hamiliton, J. Electron. Mater. 24 1149 (1995).
13. S.-H. Suh, J.-H. Song, and S.-W. Moon, J. Cryst. Growth 159,
1132 (1996).
14. J.E. Hails, D.J. Cole-Hamilton, and J. Geiss, J. Electron.
Mater. 27, 624 (1998).
15. T.H. Meyers, J.F. Schetzina, T.J. Magee, and R.D. Ormond,
J. Vac. Sci. Technol. A3, 1598 (1983).
16. W.J. Everson, C.K. Ard, J.L. Sepich, B.E. Dean, and H.F.
Schaake, J. Electron. Mater. 24, 505 (1995).
17. P. Mitra, Y.L. Tyan, T.R. Schimert, and F.C. Case, Appl.
Phys. Lett. 65, 195 (1994).
18. P. Mitra, F. C. Case, M. B. Reine, R. Starr and M. H. Weiler,
J. Cryst. Growth 170, 542 (1997).
19. P. Mitra, Y.L. Tyan, F.C. Case, R. Starr, and M.B. Reine, J.
Electron. Mater. 25, 1328 (1996).
20. I. Hahnert and M. Schenk, J. Cryst. Growth 101, 251 (1990).
21. P. Mitra, F.C. Case, S.L. Barnes, M.B. Reine, P. O’Dette, and
S.P. Tobin, Mater. Res. Soc. Symp. Proc. 484, 233 (1998).
22. A.J. Brouns, T.R. Schimert, P. Mitra, F.C. Case, S.L. Barnes,
and Y.L. Tyan, Semicond. Sci. Technol. 8, 928 (1993).
23. M.B. Reine, K.R. Maschhoff, S.P. Tobin, P.W. Norton, J.A.
Mroczkowski, and E.E. Krueger, Semicond. Sci. Technol. 8,
788 (1993).