clusters in low-temperature-grown GaAs
W. N. Lee, Y. F. Chen, J. H. Huang, X. J. Guo, and C. T. Kuo
Citation: Journal of Vacuum Science & Technology B 23, 2514 (2005); doi: 10.1116/1.2131872
View online: http://dx.doi.org/10.1116/1.2131872
View Table of Contents: http://scitation.aip.org/content/avs/journal/jvstb/23/6?ver=pdfcov
Published by the AVS: Science & Technology of Materials, Interfaces, and Processing
Articles you may be interested in
Dual accumulation and depletion behaviors of the arsenic precipitation in low-temperature-grown Be delta-doped GaAs
Appl. Phys. Lett. 82, 305 (2003); 10.1063/1.1536248
Role of As precipitates on ultrafast electron trapping in low-temperature-grown GaAs and AlGaAs alloys
J. Appl. Phys. 91, 9863 (2002); 10.1063/1.1477614
Enhanced precipitation of excess As on antimony delta layers in low-temperature-grown GaAs
Appl. Phys. Lett. 74, 1588 (1999); 10.1063/1.123625
Scanning tunneling microscopy and spectroscopy of arsenic antisites in low temperature grown InGaAs
Appl. Phys. Lett. 74, 1439 (1999); 10.1063/1.123575
Two-dimensional arsenic precipitation in superlattice structures of alternately undoped and heavily Be-doped GaAs grown by low-temperature molecular beam epitaxy
arsenic clusters in low-temperature-grown GaAs
W. N. Leea兲
Department of Materials Science and Engineering, National Chiao Tung University, Hsinchu 300, Taiwan
Y. F. Chen and J. H. Huangb兲
Department of Materials Science and Engineering, Materials Science Center, National Tsing Hua University, Hsinchu 300, Taiwan
X. J. Guo
Institute of Physics, Academia Sinica, Taipei 11529, Taiwan
C. T. Kuo
Department of Materials Science and Engineering, National Chiao Tung University, Hsinchu 300, Taiwan
共Received 27 April 2005; accepted 10 October 2005; published 11 November 2005兲
In this study, the effects of doping type and concentration on arsenic precipitation in low-temperature-grown GaAs upon postgrowth annealing at 600, 700, and 800 °C were investigated. Three undoped/Si-doped/undoped 共i-n-i兲 regions and three undoped/Be-doped/ undoped共i-p-i兲 regions were grown by low-temperature molecular beam epitaxy. The results show that arsenic precipitation is dependent on doping type and doping concentration. Arsenic depletion was observed in all Be-doped layers for all annealing temperatures. However, a “dual” arsenic precipitation behavior was observed in Si-doped layers: As accumulates in 关Si兴=2⫻1018cm−3 doped layers, while it depletes in关Si兴=2⫻1016 and 2⫻1017 cm−3 doped layers. We attribute this “dual” As precipitation phenomenon in Si-doped layers to the different depletion depths. © 2005
American Vacuum Society. 关DOI: 10.1116/1.2131872兴
I. INTRODUCTION
GaAs and related compounds grown by low-temperature molecular-beam epitaxy 共LT-MBE兲 have attracted much at-tention due to their unique electronic and optical properties.1–3When grown at 200–300 °C, the LT-GaAs lay-ers are very nonstoichiometric, containing about 1 at. % ex-cess As and a high concentration of As related defects.4Upon postgrowth annealing above 500 °C, the excess As precipi-tate into clusters5and the LT layer’s resistance changes to an extremely high-resistivity state共⬎106⍀-cm兲.6
This property offers the benefits of excellent device isolation in integrated circuits for the elimination of sidegating or backgating effects.3 The reduced recombination time of about 400 fs7,8 makes LT GaAs very suitable for ultrafast optoelectronic applications.
Although the annealed LT GaAs has demonstrated suc-cessful device applications, the out diffusion of As precipi-tates during annealing remains a major constraint in device design. It is, therefore, very important to control the density and distribution of As precipitates for potential device appli-cations. Previous studies showed that As precipitation in an-nealed LT materials can be controlled by doping effects.5,8–15 It has been found that As precipitates preferentially form in Si-doped GaAs then intrinsic and least favorably in Be-doped GaAs for moderately Be-doped GaAs. However, when the Si and Be doping concentrations reach certain levels
共艌5⫻1018 cm−3兲, an opposite precipitation behavior
occurs.9,16Chang et al.13examined the As precipitation in an
n-p1-n-p2-n-p3 or an i-n1-i-n2-i-n3-i structure, where “i,” “n,” and “p” denotes undoped, Si-doped, and Be-doped GaAs, respectively, with doping levels ranging from 2 ⫻1018 共n
1 or p1兲 to 2⫻1016cm−3 共n3 or p3兲. Upon
post-growth annealing at 800 °C, distinctive depletion zones could be observed along the i-n1 interface, as generally ex-pected. However, the depletion zone was less well defined or even unrecognizable at the i-n2 and i-n3 interfaces, and
depletion zones along the p1-n interface were narrower than
those at the p2-n and p3-n interfaces; these precipitation
be-haviors can not be fully explained by the so-called “doping” or “Fermi-level” effect. Both the inter-diffusion of As related defects and doping effects during annealing among the LT layers make this a more complicated system for understand-ing the As precipitation behavior in lightly-doped LT GaAs. Thereby, in this study we designed a LT-GaAs structure con-sisting of three i-n-i and three i-p-i active regions, with dop-ing levels rangdop-ing from 1016 to 1018cm−3, in which each active region is separated with a 10 nm AlAs marker layer. Owing to the high activation energy but small diffusion con-stant for Al and Ga inter-diffusion,17the thin AlAs layer also acts as an inter-diffusion barrier between adjacent active re-gions. By this way, the dependence of doping type and con-centration on As precipitation behavior can be clearly distin-guished. The As precipitation behavior was carefully characterized by transmission electron microscopy 共TEM兲 and discussed.
a兲Also with: Materials Science Center, National Tsing Hua University,
Hsin-chu 300, Taiwan; electronic mail: [email protected]
II. EXPERIMENT
The samples used in this study were grown by a Varian Modular GEN-II MBE system. The growth rates of 0.8m / h for GaAs and 0.2m / h for AlAs and the V/III beam equivalent pressure共BEP兲 ratio of 25 were used. Fol-lowing native oxide desorption, a 100 nm GaAs buffer layer was first grown at 580 °C to smooth the surface, followed by a 10 nm AlAs diffusion barrier at the same temperature. Growth was then interrupted by closing the Al effusion fur-nace shutter, the substrate temperature was ramped down to 250 °C and the As shutter was closed when the substrate temperature was below 400 °C to maintain a clear 2⫻4 sur-face reconstruction, as observed by reflection high-energy electron diffraction共RHEED兲. It took about 15 min to stabi-lize the substrate temperature. Subsequently, the low-temperature active layers and a 35 nm GaAs cap layer were grown. As shown in Fig. 1, the LT-GaAs sample consisted of six active regions, i.e., three regions of undoped/Be-doped/ undoped 共i-p-i兲 multilayers and another three regions of undoped/Si-doped/undoped共i-n-i兲 multilayers. Each i-n-i or
i-p-i region was separated by a thin AlAs layer, which acted
as both a diffusion barrier and a marker layer. The thickness of each doped or undoped GaAs layer was 35 nm. The doping levels in the p1, p2, and p3 layers 共and n1, n2, and
n3 layers兲 were 2⫻1016, 2⫻1017, and 2⫻1018cm−3,
respectively.
Postgrowth annealing was carried out in a rapid thermal annealing共RTA兲 system in forming gas ambient with a GaAs proximity cap at 600, 700 and 800 °C for 30 s. Arsenic pre-cipitation in annealed samples was examined using a JEOL JEM-2010 transmission electron microscope. Cross-sectional
samples parallel to 共110兲 planes were prepared convention-ally by mechanical thinning and Ar-ion milling. To further characterize the concentration of excess As incorporated into the LT-GaAs layers, three 1m thick LT GaAs control samples关Si兴=1018cm−3 doped,关Be兴=1018cm−3 doped, and
undoped were also grown under the same growth conditions as aforementioned. Double-crystal x-ray diffraction共DXRD兲 rocking curves were examined using a Philip DCD-3 double-crystal diffractometer.
III. RESULTS AND DISCUSSION
Figures 2共a兲–2共c兲 show the bright-field TEM images of six active regions after 30 s anneals at 600, 700, and 800 °C, respectively. The bright stripes seen in each picture are the AlAs layers. It is noted that AlAs layers provide good mark-ers among the 共i-n-i兲 or 共i-p-i兲 regions. As precipitation depletion was observed in all Be-doped layers for all anneal-ing temperatures; moreover, the As clusters accumulate to-ward the undoped layers. The DXRD rocking curves of GaAs 共400兲 reflection for the three control samples are shown in Fig. 3. Clearly, peak separations of 108 and 72 arc s, caused by excess As incorporated in LT GaAs, were observed in Be-doped and undoped LT-GaAs epilayers, re-spectively; however, no obvious peak separation was ob-served in the Si-doped LT GaAs. This suggests that the Be doping enhanced the incorporation of excess As into the lat-tice of LT GaAs while the Si doping suppressed this effect. According to a previous report,11 the concentration of As antisites,关AsGa兴, is directly proportional to the lattice
expan-sion of the LT-GaAs epilayer. Therefore, we can conclude that关AsGa兴p3⬎关AsGa兴i⬎关AsGa兴n3, where关AsGa兴xdenotes the
As antisite concentration in the x-doped control sample. Moreover, during postgrowth annealing, the reduction of strain caused by the AsGadefects is also a driving force for
precipitation. For the i-p3-i region, the AsGadefects diffused
from the p3to i layers, and for the i-n3-i region from i to n3
layer, which is opposite to the VGadefects. Our experimental
result about i-n3-i region and all i-p-i regions is therefore consistent with the vacancy-assisted As antisite diffusion mechanism.8,10,12,13
Interestingly, we observed a “dual” As precipitation phenomenon9 in Si-doped regions: As precipitates accumu-late toward the center of the关Si兴=2⫻1018cm−3doped layer
for all annealed temperatures; however, As precipitates are depleted away the other two lightly Si-doped regions 共n=2 ⫻1016 and 2⫻1017cm−3兲 toward the undoped regions. The
latter depletion behavior contrasts with the earlier observa-tion by Chang et al.13In their work, the i-n1-i-n2-i-n3-i
mul-tiplayer structure did not consist of any isolation layer to prevent the inter-diffusion of defects and doping effect from
n1 to n2 or n3. As a result, As precipitation behaviors were
disturbed and the correlation of doping concentration to As precipitation could not be clearly identified.
The “dual” As precipitation phenomenon in Si-doped GaAs layers indicates that other effects must be considered in addition to the vacancy-assisted As antisite diffusion mechanism. According to the analysis of O’Hagan et al.,16 FIG. 1. Schematic of the LT-GaAs structure containing six active regions
共i-n1-i , i-n2-i , i-n3-i , i-p1-i , i-p2-i, and i-p3-i multilayers兲. The “i” denotes 35
nm thick undoped GaAs layer, “n1,” “n2,” and “n3”共“p1,” “p2,” and “p3”兲
denote the Si-doped共Be-doped兲 GaAs layers with doping concentration of 1016, 1017, and 1018cm−3, respectively. Each active region was separated by
a 10 nm AlAs layer.
As precipitation at the i / n interface can be influenced by the Debye length,18 D=兵0kT /关q2共ND− NA兲兴其1/2. Following
this argument, band bending14,18,19and depletion of the “n” regions to a depth of up to a few Debye lengths would occur at each i / n interface. Moreover, according to the model pro-posed by Chandra et al.,18 the Debye lengths at the current annealing temperatures in our structure are estimated to be
around 560, 180, and 50 Å for n1, n2, and n3 layers,
respec-tively. Therefore, the total depletion depth should be ⬃1120, ⬃360, and ⬃100 Å for n1, n2, and n3layers,
respec-tively, due to the double n / i interfaces in each active region. Therefore, since each Si-doped layer is 350 Å thick, it is likely that complete depletion occurred in the n1and n2 lay-ers but an incomplete depletion in the n3layer, the latter was also noted as an “As accumulation” layer.
Figure 4 shows the average density of As clusters in each
i-x-i region after annealing at 700 and 800 °C, respectively.
The density of As clusters was determined by assuming that the sample thickness under TEM investigation was 100 nm. The analysis reveals that the cluster densities of the i-p-i
FIG. 2. TEM bright field images showing arsenic precipitates in different active regions after annealing at 共a兲 600, 共b兲 700, and 共c兲 800 °C, respectively.
FIG. 3. DXRD rocking curves of GaAs 共400兲 for 1m thick undoped,
关Si兴=1018cm−3 doped, and 关Be兴=1018cm−3 doped LT GaAs control
samples.
FIG. 4. As cluster density in each i-x-i region after annealing at 700 and 800 °C, where x = nior pi.
regions are higher than those of the i-n-i regions under the same doping concentration and annealing condition, which is consistent with the DXRD results.
IV. CONCLUSION
The effect of doping type and concentration on As pre-cipitation in LT-GaAs upon post-growth annealing at 600, 700, and 800 °C was investigated. The results show that the As precipitation is dependent on doping type and doping concentration. As depletion were observed in all Be-doped regions. However, a “dual” arsenic precipitation phenom-enon was observed in Si-doped regions: As precipitation ac-cumulation was observed in关Si兴=2⫻1018cm−3doped layer for all annealing temperatures, while As precipitation deple-tion was observed in关Si兴=2⫻1016and 2⫻1017cm−3doped layers. The “dual” arsenic precipitation phenomenon in Si-doped layers can be attributed to the different depletion depths caused by various doping concentrations.
ACKNOWLEDGMENTS
This work was supported by the National Science Coun-cil, Republic of China, under Contract Nos. NSC 92-2112-M-007-032 and NSC 92-2120-M-007-006.
1D. C. Look and D. C. Walters, Phys. Rev. B 42, 3578共1990兲. 2D. D. Nolte, M. R. Melloch, J. M. Woodall, and S. J. Ralph, Appl. Phys.
Lett. 62, 1356共1993兲.
3F. W. Smith, A. R. Calawa, C.-L. Chen, M. J. Manfra, and L. J. Mahoney,
IEEE Electron Device Lett. 9, 77共1988兲.
4T. J. Rogers, C. Lei, B. G. Streetman, and D. G. Deppe, J. Vac. Sci.
Technol. B 11, 926共1993兲.
5M. R. Melloch, N. Otsuka, K. Mahalingam, P. D. Kirchner, J. M.
Wood-all, and A. C. Warren, Appl. Phys. Lett. 61, 177共1992兲.
6A. C. Warren, J. M. Woodwall, J. L. Freeouf, D. Grischkowsky, D. T.
McInturff, M. R. Melloch, and N. Otsuka, Appl. Phys. Lett. 57, 1331 共1990兲.
7S. Gupta, M. Y. Frankel, J. A. Valdmanis, J. F. Whitaker, G. A. Mourou,
F. W. Smith, and A. R. Calawa, Appl. Phys. Lett. 59, 3276共1991兲.
8D. E. Bliss, W. Walukiewicz, J. W. Ager, III, E. E. Haller, K. T. Chan, and
S. Tanigawa, J. Appl. Phys. 71, 1699共1992兲.
9J. H. Huang, L. Z. Hsieh, X. J. Guo, and Y. O. Su, Appl. Phys. Lett. 82,
305共2003兲.
10M. R. Melloch, N. Otsuka, K. Mahalingam, C. L. Chang, J. M Woodall,
G. D. Pettit, P. D. Kirchner, F. Cardone, A. C. Warren, and D. D. Nolte, J. Appl. Phys. 72, 3509共1992兲.
11X. Liu, A. Prasad, J. Nishio, E. R. Weber, Z. Liliental-Weber, and W.
Walukiewicz, Appl. Phys. Lett. 67, 279共1995兲.
12M. N. Chang, J.-W. Pan, J.-I. Chyi, K. C. Hsieh, and T.-E. Nee, Appl.
Phys. Lett. 72, 587共1998兲.
13M. N. Chang, K. C. Hsieh, T.-E. Nee, and J.-I. Chyi, J. Appl. Phys. 86,
2442共1999兲.
14M. Missous and S. O’Hagan, J. Appl. Phys. 75, 3396共1994兲.
15T. M. Cheng, C. Y. Chang, and J. H. Huang, Jpn. J. Appl. Phys., Part 1
34, 1185共1995兲.
16S. P. O’Hagan, M. Missous, A. Mottram, and A. C. Wright, J. Appl. Phys.
79, 8384共1996兲.
17L. L. Chang and A. Koma, Appl. Phys. Lett. 29, 138共1976兲.
18A. Chandra, C. E. C. Wood, D. W. Woodard, and L. F. Eastman,
Solid-State Electron. 22, 645共1979兲.
19D. C. Look, D. C. Walters, M. Mier, C. E. Stutz, and S. K. Brierley, Appl.
Phys. Lett. 60, 2900共1992兲.
