Triisopropylindium for OMVPE growth

Journal of Crystal Growth 124 (1992) 88—92 North-Holland

~

CRYSTAL OROWT H

Triisopropylindium for OMVPE growth C.H. Chen, C.T. Chiu, G.B. Stringfellow Department of Materials Science and Engineering, University of Utah, Salt Lake City, Utah 84112 USA

and R.W. Gedridge, Jr. Chemistry Division, Research Department, Naval Air Warfare Center, China Lake, California 93555, USA

The organometallic vapor phase epitaxial (OMVPE) growth of In-containing Ill—V semiconductors typically uses trimethylindium (TMIn). However, TMIn suffers from several problems. This work reports the first decomposition and OMVPE growth studies for a newly developed indium source, triisopropylindium (TIPIn). The decomposition study shows that the temperature for 50% decomposition is 110°C for TIPIn in a He ambient, much lower than for TMIn. InAs epilayers with good surface morphologies were obtained at temperatures as low as 300°C.The necessary V/Ill ratio increases as the growth temperature is decreased, due to the incomplete decomposition of AsH 3 at low temperatures. Since TIPIn contains C3H7 radicals which are far less reactive than the CH3 radicals in TMIn, the InAs grown using TIPIn has carbon concentrations several orders of magnitude lower than when TMIn is used.

1. Introduction

2. Experimental procedure

Trimethylindium (TMIn) has been the standard In source for organometallic vapor phase epitaxy (OMVPE). However, problems exist with the use of TMIn. They include the variable effective vapor pressure [1,2], the presence of methyl ligands that cause carbon contamination [3,4], and slow decomposition at very low temperatures (<300°C) [3]. Ethyldimethylindium (EDMIn) is an effective replacement for low temperature growth [1,3,5]. However, it is not certain how ligand exchange reactions [61 affect the final molecules leaving the liquid source. Moreover, EDMIn still has two methyl ligands, which give carbon contamination at low temperatures [3]. In this work, we report the results of the first study of triisopropylindium [(C3H7)31n,TIPIn] as a possible TMIn replacement for OMVPE growth of In-containing materials. Both decomposition and OMVPE growth results are presented.

The TIPIn source was synthesized at the Naval Air Warfare Center. The vapor pressure and the synthesis process are reported in ref. [7]. The decomposition experiments were conducted in an isothermal, flow-tube, Si02 ersatz reactor at atmospheric pressure. A detailed description of the apparatus has been published previously [81.The TIPIn source was held at 23°Cand the carrier gas was He with a flow rate of 40 SCCM. Unless otherwise specified, the TIPIn source was purged with He for more than 12 h before each experiment. For OMVPE growth of InAs, an atmospheric pressure horizontal reactor was used [91. The arsenic source was 100% arsine. The carrier gas for the sources was palladium-diffused 2 with a total flow rate of about 2.5 liter/mm. Separate stainless steel tubing was used for the group III and V reactants in order to minimize possible parasitic reactions. The mixing of the

0022-0248/92/$05.00 © 1992

—

Elsevier Science Publishers B.V. All rights reserved

C.H. Chen et aL 8

/

Triisopropylindium for OMVPE growth

The Van der Pauw technique was used to obtain the room temperature carrier concentrations for layers grown on semi-insulating InP substrates. The In contacts on the four corners of the rectangular samples were annealed at 300°Cfor 1—2 mm under N2. The magnetic field was 5 kG

________

To A + •

A

m/e m/e..421 mie ~43I m/e =

89

~

.

7~j

‘

and the sample current was about 10 ~A. The carbon concentration in the epilayers was measured using a Perkin-Elmer 6300 secondary ion mass spectrometry (SIMS). The standard was

0

ionituplanted GaAs. A sample was measured twice to ensure the reliability of the measure-

2’

0

i’i’I’’’

0

50

100

150

200

250

T (°C) Fig. 1. Mass spectral intensities versus temperature at several values of m/e.

group III and V reactants occurred immediately before entering the quartz 3/min reactor.with Thethetypical bubTIPIn flowheld rate atwas 300 cm AsH bler being 22°C.The 3 flow rate was 3/min. on the order of 20 cm

For low temperature photoluminescence (PL) measurements, the excitation source was an ar2. gon ion laser operating at 488 nm. The beam was The excitation was approximately focused to a spotintensity size of approximately 0.5 mm20 2

W/cm The samples were bonded to the cold finger of a closed cycle He cryostat. The window of the cryostat was made of JR transmitting BaF 2. A off-axis paraboloidal focused thepair PL of onto the entrance slit of reflectors a half-meter Spex M500 the spectrometer. A GaAs usedthe to block scattered laser lightfilter and was to pass

40p,m

Tg=500°C V/III= 154

Tg=400°C V/III= 165

Tg=300°C —

Fig. 2. Surface morphology of InAs layers grown on InAs substrates using TIPIn and AsH 3 at 500, 400, and 300°Cwith V/Ill ratios of approximately 150.

90

C.H. Chen et aL

/

Trusopropylindium for OMVPE growth

desired radiation. The filter was carefully checked by FTIR transmittance and was found to be

is about 310°C[10].The ease of decomposition is most likely due to the weak In—C3H7 bond. It is

transparent at wavelengths as long as 16 ~.tm.The PL was detected using an InSb detector cooled to liquid N2 temperature.

known that the H—alkyl bond strength decreases in the order [11]: CH3—H> C2H5—H> i—C3H7—H. The alkyl—In bond strengths are expected to follow the same order. Thus, TIPIn is expected to decompose at lower temperatures than TMIn. The OMVPE growth experiments were carried out in a typical OMVPE growth reactor, not the one used for the pyrolysis studies. Fig. 2 shows the surface morphologies of InAs layers grown at 500, 400, and 300°C,with V/Ill ratios of about 150. The surface morphology degrades as the growth temperature is reduced. This is not related to the use of TIPIn, but to the incomplete decomposition of AsH3. Since AsH3 decomposes slowly at low growth temperatures [12],the real V/Ill ratio at the interface will be much smaller than the ratio of input molar flow rates of AsH3

3. Results and discussion As mentioned in section 2, the pyrolysis experiments were carried out in a flow tube, isothennal reactor. The products observed for T1PIn decomposition occur at m/e = 39, 42, 43, and 71. No peaks are observed for m/e = 100—300. Peak intensities are plotted versus temperature in fig. 1. The intensities increase rapidly from 100 to 125°C and saturate above 125°C.The results show that the TIPIn decomposes at low temperatures, with a value of T50 (temperature for 50% decomposition) of about 110°C.For comparison, the value of T50 obtained for TMIn using similar conditions

Temperature (°C) 600

21

10

1020

nforTtPlnsource

•

CforllPlnsource

A

n for TMIn&EDMIn source CforlMlnsource C for EDMIn source

0 ~

C 0

t~ C

400

A

o

C

E

500

________________________________________

300

r

10~

10~

0

I

o A

o

1017.

A 16

A

A ________________________________________________________________________________________________

10~ 1.0

I

1.2

I

1.4

I

1.6

1.8

2.0

1OOO/Tg(K~) Fig. 3. Room temperature electron concentration measured using the Van der Pauw technique and carbon concentration measured using SIMS for InAs grown using TIPIn, TMIn or EDMIn as a function of growth temperature.

C.H. Chen et aL

/

Triisopropylindium for OMVPE growth

over TIPIn. For the sample grown at Tg = 300°C and an input V/Ill ratio of 144, the V/Ill ratio

440

420

400

91 En.rgy (meV) 380

360

at the interface is probably less than unity. Thus,

the surface appears to be black to the naked eye, due to well-known whisker growth [13,14]. The

surface morphologies for the samples grown at 400 and 300°Ccan be improved to smooth surfaces electron Thewith as-grown and an carbon increase InAs concentrations in epilayers V/Ill ratio areare to 460 n-type. shown [7]. The in fig. 3 as a function of growth temperature. The electron concentration does not change when the growth temperature is reduced from 500 to 400°C.

Tg=400°C

This level of donor impurity is probably caused by an extrinsic residual impurity in the TIPIn source.

Tg=500 °C

This is not surprising since this is the first bottle

Moreimpurity TIPIn significant evergrade. used for isItthe increase growth: in electron Itfurther isbackand not ground of electronic can isOMVPE expected be reduced that by this purification of thelevel TIPIn. carbon concentrations for temperatures below 400°C.Since the electron and carbon concentrations are nearly equal, it is concluded that the carbon impurity is the primary donor species for the InAs grown using T1PIn. In previous studies [3,4], the carbon impurity has also been found to be a donor in InAs rather than an acceptor as for GaAs. This has been explained in terms of the relative strengths of the C—As and C—In bonds [3,4]. Recently, InAs has been grown using TMIn (or EDMIn) and AsH3. It was found that the electron and carbon concentrations also increase as the growth temperature is reduced [3,4]. The results are shown in fig. 3 for comparison. It is seen that the carbon incorporation increases for InAs grown at lower temperatures using TIPIn, TMIn, and EDMIn: The absolute carbon level is more than 10 times lower for the layers grown using TIPIn. It is possible that all three sources contain an extrinsic carbon impurity to cause the similar trend in carbon incorporation. More likely, the carbon incorporation is intrinsic: It comes from the alkyl ligands. Both TMIn and EDMIn contain methyl ligands. Upon decomposition, some of the methyl radicals would be bound to the surface, leading to carbon incorporation into

2.8

3.0

32 3.6 Waveiength (pan) Fig. 4. Low temperature (10 K) PL spectra for InAs samples grown using TIPIn and AsH 3 at several temperatures. The V/Ill ratios for the samples shown are about 150. The PL spectrum of an InAs substrate is also shown for comparison.

the solid. On the other hand, the main decomposition products for TIPIn are isopropyl radicals which are far less reactive than the methyl radicals. Thus, they are far more likely to desorb before decomposition. As a result, the carbon incorporation for InAs grown using TIPIn is more than an order of magnitude lower than that for InAs grown using TMIn or EDMIn. Fig. 4 shows the low temperature PL results for samples grown at several temperatures. The PL from an InAs substrate is also shown for comparison. For the InAs substrate, the two low energy peaks at about 3.08 and 3.25 ~.tm have been assigned to emission processes involving either impurity or defect states [15]. The two high energy peaks located near 3 ~tm are due to bandto-band and exciton recombination [16,171. For samples grown using T1PIn and AsH3, the PL consists mainly of two peaks. The higher energy

92

C.H. Chen et aL

/

Triisopropylindium for OMVPE growth

peak at about 3 ~m is apparently due to a combination of peaks from band-to-band and exciton recombination. The lower energy peak at about 3.08 p~mis apparently the same impurity/defect peak as for the substrate. The PL intensities are similar for samples grown at 500 and 400°C.The PL intensity is much weaker for the sample grown at 300°Cwith a V/Ill ratio of 150. Of course, a V/Ill ratio of 150 at 300°Cproduces rough surface morphologies, as seen in fig. 2. This may

Van der Pauw measurements and Ray Menna at the David Sarnoff Research Center for performing SIMS measurements. Financial support of the this work was provided by the Army Research Office, Office of Naval Research, and Office of Naval Technology.

References

partially explain the low PL intensity. However,

even with a V/Ill ratio of 460, the PL intensities of samples grown at 300°Care still roughly 60 times weaker than for samples grown at higher

temperatures even though the surface morphologies are smooth. The observed temperature dependence of PL intensity is similar to that observed for InAs grown using TMIn and AsH3 [4].

[1] C.P. Kuo, R.M. Fletcher, T.D. Osentowski, G.R. Trott and i.E. Fouquet, 4th Biennial Workshop on Organometallic Vapor Phase Epitaxy, Monterey, CA, October 1989. [2] AKZO Chemicals Inc., Newsletter, October, 1991.

[31K.Y.

Ma, Z.M. Fang, R.M. Cohen and G.B. Stringfellow, J. AppI. Phys. 70 (1991) 3940. [4] Z.M. Fang, K.Y. Ma, R.M. Cohen and G.B. Stringfellow, AppI. Phys. Letters 59 (1990)1446. [5] K.L. Fry, C.P. Kuo, C.A. Larsen, R.M. Cohen, G.B.

4. Conclusions [6]

TIPIn has been investigated as a possible replacement for TMIn in OMVPE growth. From the decomposition results, it is found that TIPIn decomposes with a value of T50 of about 110°C, approximately 200°C lower than the value for

TMIn under similar conditions. The growth resuits show that good surface morphology InAs can be obtained provided that the V/Ill ratio is sufficiently high. The required V/Ill ratio has to be increased as the growth temperature is low-

ered because less AsH3 is decomposed at lower temperatures. Both the background electron and carbon concentrations increase as the growth temperature is reduced below 400°C.The background carbon concentration is reduced by about two orders of magnitude when TIPIn replaces TMIn. Acknowledgements

The authors would like to thank S.H. Soh and K.T. Huang for assisting with room temperature

[7] [8] [9] [10] [11] [12]

[13] [14]

Stringfellow and A. Melas, J. Electron. Mater. 15 (1986) 91. P.D. Agnello and S.K. Ghandhi, J. Crystal Growth 94 (1989) 311. C.H. Chen, G.B. Stringfellow and R.W. Gedridge, Jr., J. Electron. Mater., submitted. N.!. Buchan, C.A. Larsen and G.B. Stringfellow, AppI. Phys. Letters 51(1987)1024. C.H. Chen, M. Kitamura, R.M. Cohen and G.B. Stringfellow, AppI. Phys. Letters 49 (1986) 963. N.I. Buchan, C.A. Larsen and GB. Stringfellow, J. Crystal Growth 92 (1988) 591. G.B. Stringfellow, Organometallic Vapor Phase Epitaxy: Theory and Practice (Academic Press, New York, 1989) ch. 2. G.B. Stringfellow, Organometallic Vapor Phase Epitaxy: Theory and Practice (Academic Press, New York, 1989) section 4.2.2.1. G.B. Stringfellow, Organometallic Vapor Phase Epitaxy, Theory and Practice (Academic Press, New York, 1989) pp. 84—86. R.S. Wagner, in: Whisker Technology, Ed. A.P. Levitt (Wiley, New York, 1970).

[15] R.D. Robert, H.D. Drew, i-I. Chyi, S. Kalem and H. Morkoc, J. AppI. Phys. 65 (1989) 4079. [16] A. Mooradian and H.Y. Fan, in: Proc. 7th Intern. Conf. on Physics of Semiconductors, Paris, 1964 (Academic Press, New York, 1965) Vol. 4, p. 39. [17] Z.M. Fang, K.Y. Ma, D.H. Jaw, R.M. Cohen and G.B. Stringfellow, J. AppI. Phys. 67 (1990) 7034.