Full text of DOI:10.1063/1.114542

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Thermal stability of the midgap acceptor rhodium in indium phosphide
A. Näser, A. Dadgar, M. Kuttler, R. Heitz, D. Bimberg, J. Y. Hyeon, and H. Schumann
Citation: Applied Physics Letters 67, 479 (1995); doi: 10.1063/1.114542
View online: http://dx.doi.org/10.1063/1.114542
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/67/4?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Neutron Capture and Transmission Measurements on 103Rh Down to Thermal Energies
AIP Conf. Proc. 769, 953 (2005); 10.1063/1.1945162
NO adsorption on Rh(100). II. Stability of the adlayers
J. Chem. Phys. 119, 12534 (2003); 10.1063/1.1627757
Performance, Thermoelectric Homogeneity and Stability of Au/Pt and Pt10 %Rh/Pt Thermocouples
AIP Conf. Proc. 684, 475 (2003); 10.1063/1.1627171
Thermal reduction of activation energy of tricyclohexylphosphine on a rhodium crystal surface
J. Chem. Phys. 115, 9513 (2001); 10.1063/1.1389277
Acceptor implantation in Fedoped, semiinsulating indium phosphide
J. Appl. Phys. 52, 6623 (1981); 10.1063/1.328430
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sat, 20 Dec 2014 00:37:26

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Thermal stability of the midgap acceptor rhodium in indium phosphide
¨
A. Naser, A. Dadgar, M. Kuttler, R. Heitz, and D. Bimberg
¨
¨
¨
Institut fur Festkorperphysik, Technische Universitat Berlin, Hardenbergstrasse 36, D-10623 Berlin,
Germany
J. Y. Hyeon and H. Schumann
¨
¨
Institut fur Anorganische und Analytische Chemie, Technische Universitat Berlin,
Strasse des 17. Juni 135, D-10623 Berlin, Germany
͑Received 24 March 1995; accepted for publication 24 May 1995͒
We investigated the thermally induced redistribution of Rh in low pressure MOCVD grown InP
structures by means of secondary-ion-mass-spectroscopy. Analogous measurements for InP:Fe
structures serve as reference. On alternately Rh-doped/undoped InP structures an upper limit for the
diffusion coefficient of DRh(800 °C)р1ϫ10Ϫ14 cm2/s is established much smaller than
DFe(750 °C)ϭ1ϫ10Ϫ11 cm2/s. No exchange reactions are observed at the interface of p-InP/
InP:Rh structures. Only Rh implanted into InP shows defect induced redistribution into amorphous
areas. Rh is superior to Fe as far as thermal stability is concerned. © 1995 American Institute of
Physics.
Semi-insulating ͑SI͒ InP is a key material for the fabri-
cation and monolithic integration of InP-based high speed
electronic and optoelectronic devices.1–3 Today SI InP sub-
strates and epitaxial layers are grown by doping with the 3d
transition metal ͑TM͒ Fe.4 The deep acceptor iron compen-
sates n-type material, but has a low thermal stability5,6 and
shows a pronounced interdiffusion ͑of e.g., Zn and Fe͒ in
contact with p-type material,7 leading to conductivity type
inversion. p-Fe-p layers show no blocking behavior,8 and
parasitic currents occur in ridge lasers overgrown with
InP:Fe.9 The application of Fe-doped layers in device struc-
tures is restricted by these properties. The aim of this letter is
to present results of the search for alternative compensating
transition metals with improved thermal stability.
but undoped material ͑face-to-face-method͒ in a H2 atmos-
phere. Depth profiles are recorded by SIMS with an
ATOMIKA 6500 using an oxygen primary ion beam while
monitoring 103Rhϩ or a CAMECA IMS-4f measuring the
͑CsZn͒ and ͑CsRh͒ cluster.
Figure 1͑a͒ compares Rh depth profiles measured for an
alternating Rh-doped/undoped InP layer structure as grown
and after annealing at 800 °C for 20 min. Independent of the
197
ϩ
237
ϩ
4d and 5d TMs are expected to form deep donor and
acceptor levels similar to 3d TMs, but should be less diffu-
sive due to their larger atomic and ionic radii. First investi-
gations on ion-implanted III–V materials indicated a supe-
rior thermal stability of Ru and Os.10 However, the electrical
properties of 4d and 5d metals are almost unknown yet. Only
recently deep level transient spectroscopy revealed two deep
hole traps in Rh-doped InP with activation energies of
ERhAϭ710 meV and ERhBϭ620 meV.11,12 These near
midgap acceptor levels are in principle suited to compensate
n-type InP. In this letter we investigate the diffusive proper-
ties of Rh in InP doped either in situ during MOCVD growth
or by ion implantation.
InP growth is performed by low pressure ͑20 mbar͒
metalorganic chemical vapor phase deposition ͑MOCVD͒
using TMIn and PH3. The temperature of the substrate is
680 °C. All layers are deposited on ͑100͒ oriented p- or
n-type InP substrates with a growth rate of 2 ␮m/h and a
V/III ratio of 150. For Rh-doping cyclopentadienyl͑cyclooc-
tadienyl͒rhodium ͓Rh͑cp͒cod͔ or cyclopentadienyl͑norborna-
dien͒rhodium ͓Rh͑cp͒͑nor͔͒ are used as metalorganic
sources. The second one, named CYNOR, has a higher vapor
pressure than ͓Rh͑cp͒͑cod͔͒ allowing Rh concentrations at
1019 cmϪ3 without degradation of the surface. Annealing is
performed for all samples placed upside down on the same
FIG. 1. Comparison of SIMS depth profiles of ͑a͒ Rh and ͑b͒ Fe in InP:Rh
and InP:Fe, respectively, before and after thermal annealing.
Appl. Phys. Lett. 67 (4), 24 July 1995
0003-6951/95/67(4)/479/3/$6.00
© 1995 American Institute of Physics
479
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sat, 20 Dec 2014 00:37:26

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
FIG. 3. SIMS depth profiles of 103Rh in InP implanted at 200 keV/1
ϫ1014 cmϪ2 beofre and after face-to-face annealing at 750 °C for 20 min.
material. In the interface region ͑dϷ0.3 ␮m͒ Zn atoms popu-
late In vacancies generated by the formation of Fe intersti-
tials leading to a total depletion of Fe atoms now homoge-
neously distributed in the p-InP:Zn toplayer.17
For Rh in InP we observe no influence of the position of
the Fermi level on diffusion. Figure 2͑a͒ demonstrates a
complete lack of interdiffusion at the p-InP:Zn/InP:Rh inter-
face. InP:Rh was deposited on a SI InP:Fe substrate and
covered with 2 ␮m InP:Zn ͑ϳ3ϫ1018 cmϪ3). Memory
effects in the MOCVD reactor causes the small but distinct
Rh concentration in the Zn-doped toplayer. Like in Fig.1͑a͒
Rh shows no thermally induced redistribution, even after an-
nealing at 700 °C for 20 min.
FIG. 2. SIMS depth profiles of ͑a͒ Fe and Zn in a p-InP:Zn layer deposited
on InP:Fe, and ͑b͒ of Rh and Zn in a p-InP/InP:Rh structure before and after
thermal annealing.
In the defect-rich surface region of MOCVD grown
InP:Rh layers a weak accumulation of Rh can be observed
even without annealing, as shown in Fig. 2͑b͒. To acquire
more information about defect related diffusion of Rh atoms
we implanted 103Rh with 200 keV at room temperature ͑dose
1ϫ1014 cmϪ2͒ into Zn-doped InP substrates. Figure 3
compares the Rh depth profiles of as-implanted and annealed
͑750 °C, 20 min͒ samples. TRIM ͑transport of ions in matter͒
simulations18 are carried out in order to evaluate the distri-
bution of the implanted Rh ions as well as of the implanta-
tion induced damage. The as-implanted profile is in very
good agreement with the expected TRIM profile. The pro-
jected range (Rp) of the as-implanted profile is 87 nm and
the standard deviation (⌬Rp) is 39 nm. To explain the mea-
sured profile observed after annealing, two different regions
have to be distinguished. In the amorphous zone up to 130
nm below the surface Rh atoms redistribute, due to the com-
bination of a nonFickian uphill diffusion caused by the im-
plantation induced damage and an outdiffusion to the surface
by indiffusion of vacancies ͑Frank–Turnbull mechanism͒.
Both effects act attractive to Rh atoms in the amorphous
region up to Rpϩ⌬Rp . The tail shows, as expected, no
diffusion into the bulk. Recently, similar defect related dif-
fusion properties were observed for Zr and Hf.19
Rh concentration, all seven doped layers show no diffusion
into undoped regions. Numerical simulations adjusted to
the measured profiles give an upper limit for the diffusion
coefficient of DRh(800 °C)р1ϫ10Ϫ14 cm2/s.
For Fe an analogous structure in undoped InP is grown at
640 °C ͓Fig. 1͑b͔͒. The structure is already disordered after a
20 min thermal annealing at 780 °C. At 750 °C the Fe boxes
only are not totally washed out ͑not shown here͒ giving a
diffusion coefficient of DFe (750 °C)р1ϫ10Ϫ11 cm2/s.
This value is in good agreement with previously published
data.13,14 The attempt to grow a similarly modulated Fe
structure in homogeneously p-doped InP failed. SIMS depth
profiles ͑not shown here͒ show an almost homogeneous Fe
distribution in the whole p-InP layer, indicating a consider-
ably higher diffusion coefficient in p-type than in n-type ma-
terial. The diffusion of Fe is governed by the equilibrium
between substitutional and interstitial Fe atoms depending on
the position of the Fermi level. Fe on interstitial positions
can rapidly diffuse making the amount of interstitial Fe
atoms decisive for the diffusion coefficient. A low Fermi
level like in p-type material favors the creation of interstitial
Fe atoms.15
More severe for device fabrication is the above-
mentioned breakdown of the SI behavior in InP:Fe between
p-type InP contact layers16 due to Fermi level dependent ion
exchange via the interface. A low level indiffusion of Zn far
into the nominally Fe-doped InP ͓Fig. 2͑b͔͒ produces p-type
In summary, we find that Rh is a nondiffusion dopand in
InP. The thermal stability is only affected by defect induced
redistribution. No ion exchange reactions with shallow ac-
ceptors from adjacent p-type layers are observed presenting a
decisive advantage as compared to SI InP:Fe. The dramatic
¨
Naser et al.
480
Appl. Phys. Lett., Vol. 67, No. 4, 24 July 1995
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sat, 20 Dec 2014 00:37:26

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
difference between the thermal stability of Rh and Fe is most
probably caused by the larger ionic radius of Rh. The very
effective substitutional-interstitial diffusion process dominat-
ing for Fe is replaced for Rh by the much less effective
kick-out process. Rh induces a deep acceptor in InP ͑Ref. 20͒
and, thus, is an excellent candidate for thermally stable SI
InP layers.
9 S. Asada, S. Sugou, K.-I. Kasahara, and S. Kumashiro, IEEE J. Quantum
Electron. QE-25, 1362 ͑1989͒.
M. Kuttler, A. Knecht, D. Bimberg, and H. Krautle, Mater. Res. Soc.
10
¨
Symp. Proc. 316, 179 ͑1994͒.
11 H. Scheffler, B. Srocka, A. Dadgar, M. Kuttler, A. Knecht, R. Heitz, D.
Bimberg, J. Y. Heyon, and H. Schumann, Mater. Res. Soc. Symp. Proc.
325, 329 ͑1994͒.
12 A. Dadgar et al. ͑unpublished͒.
13 M Gauneau, H. L. Haridon, A. Rupert, and M. Salvi, J. Appl. Phys. 53,
6823 ͑1982͒.
This work was partly supported by the Deutsche For-
schungsgemeinschaft ͑Bi 284/10͒.
14 H. Kamada, S. Shinoyama, and A. Katsui, J. Appl. Phys. 55, 2881 ͑1984͒.
15 T. Wolf, T. Zinke, A. Krost, H. Scheffler, H. Ullrich, D. Bimberg, and P.
Harde, J. Appl. Phys. 75, 3870 ͑1994͒.
1
¨
¨
E. H. Bottcher, D. Kuhl, F. Hieronimy, E. Droge, T. Wolf, and D. Bimberg,
16 E. W. A. Young and G. M. Fontijn, Appl. Phys. Lett. 56, 146 ͑1990͒.
17 C. Blaauw, B. Emmerstorfer, R. A. Bruce, and M. Benzaguen, in Semiin-
sulating III-V-Materials, Toronto, 1990, edited by A. Milnes and C. J.
Miner ͑Adam Hilger, Bristol, 1990͒, p. 137.
IEEE J. Quantum Electron. QE-28, 2343 ͑1992͒.
2
¨
M. Putz,, J. Cryst. Growth 107, 806 ͑1991͒.
3 H. Wada, H. Horikawa, Y. Ogawa, and Y. Kawai, Appl. Phys. Lett. 55, 723
͑1989͒.
4 J. A. Long, V. G. Riggs, and W. D. Johnston, J. Cryst. Growth 69, 10
͑1984͒.
18 J. P. Biersack and L. G. Haggmark, Nucl. Instrum. Methods 174, 257
5 M. Gauneau, R. Chaplin, and A. Rupert, J. Appl. Phys. 57, 1029 ͑1985͒.
6 D. E. Holmes and R. G. Wilson, J. Appl. Phys. 52, 3396 ͑1981͒.
7 C. Kazmierski, D. Robein, and Y. Gao, J. Cryst. Growth 116, 75 ͑1992͒.
8 W.-H. Cheng, S. Y. Huang, A. Appelbaum, J. Pooladdej, K. D. Buehring,
D. Wolf, D. S. Renner, K. L. Hess, and S. W. Zehr, IEEE J. Quantum
Electron. QE-25, 1353 ͑1989͒.
͑1980͒.
19
¨
A. Knecht, M. Kuttler, H. Scheffler, T. Wolf, D. Bimberg, and H. Krautle,
Nucl. Instrum. Methods 80/81, 683 ͑1993͒.
20 D. Bimberg, A. Dadgar, R. Heitz, A. Knecht, A. Krost, M. Kuttler, H.
¨
Scheffler, A. Naser, B. Srocka, T. Wolf, T. Zinke, J. Y. Heyon, S. Wernik,
and H. Schumann, J. Cryst. Growth 145, 435 ͑1994͒.
¨
Appl. Phys. Lett., Vol. 67, No. 4, 24 July 1995
Naser et al.
481
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sat, 20 Dec 2014 00:37:26