We report on the influence of surface reconstruction on silicon dopant incorporation and transport properties during molecular-beam epitaxy of GaAs(Bi) alloys. GaAs(Bi) growth with an (n × 3) reconstruction leads to n-type conductivity, while growth with a (2 × 1) reconstruction leads to p-type conductivity. We hypothesize that the presence or absence of surface arsenic dimers prevents or enables dopant incorporation into arsenic lattice sites. We consider the influence of bismuth anions on arsenic-dimer mediated dopant incorporation and the resulting electronic transport properties, demonstrating the applicability of this mechanism to mixed anion semiconductor alloys.

Elements from column IV of the periodic table, including C, Ge, and Si, may act as either donors or acceptors in III–V semiconductors, depending on their incorporation into group III or group V lattice sites. During molecular-beam epitaxy (MBE) of GaAs, C most often occupies an As site, leading to p-type conductivity.1 On the other hand, Ge may occupy a Ga lattice site, GeGa, or an As lattice site, GeAs, resulting in n- or p-type conductivity, respectively.2 Typically, Si occupies a Ga lattice site, SiGa, resulting in n-type conductivity. However, when Si is able to outcompete arsenic for group V sites, p-type conductivity is observed. For example, due to the high density of Ga dangling bonds on (n11)A (n ≤ 3) GaAs surfaces, Si-doping typically leads to p-type conductivity, with a transition to n-type conductivity occurring as the group V/III flux ratio is increased.3–5 However, for (001) GaAs surfaces, Si-doping typically leads to n-type conductivity, with a transition to p-type conductivity using a shuttering sequence or group V/III flux ratios near the transition from anion-rich to cation-rich conditions, i.e., the stoichiometric threshold.6–8 For growth at 600 °C, Zhang et al. observed a transition from n-type to p-type conductivity at the so-called stoichiometric threshold, corresponding to (2 × 4) and (4 × 2) GaAs surface reconstructions, respectively.8 For lower substrate temperatures (520 °C), Quivy et al. obtained p-type GaAs:Si by alternate surface exposure to As, then Ga plus Si, leading to (2 × 4) and (4 × 2) surface reconstructions, respectively.6 For the lowest substrate temperatures (270 °C), a transition from n- to p-type conductivity has been reported for the GaAs growth near the stoichiometric threshold.7 To date, the atomistic mechanisms for this transition have not yet been considered. We hypothesize that the presence or absence of surface arsenic dimers prevents or enables dopant incorporation into arsenic lattice sites. We consider the influence of bismuth anions on arsenic-dimer mediated dopant incorporation and the resulting electronic transport properties, demonstrating the applicability of this mechanism to mixed anion semiconductor alloys.

GaAs(Bi):Si alloy films were grown on semi-insulating (001) GaAs substrates by MBE, using solid Ga, As4, Bi, and Si sources. The surface reconstruction was monitored in-situ with reflection high-energy electron diffraction (RHEED), and the substrate temperature was measured by a thermocouple in contact with the back of each molybdenum block, calibrated by setting the oxide desorption temperature to 600 °C. Following the growth of a 500-nm thick GaAs buffer at 600 °C, the substrate temperature was lowered to 280 °C during the growth of 250- to 550-nm thick GaAs. Finally, 400-nm GaAs(Bi):Si layers, with or without Si and Bi, were grown at various growth rates (0.25–2.0 μm/h), group V/III beam equivalent pressure (BEP) ratios (6–20), and Bi:As BEP ratios (0–0.03). The Si cell temperature was chosen to target n = 1 × 1018 or 7 × 1018 cm−3 in GaAs:Si control films.

For the GaAs(Bi):Si alloys, the surface morphology was examined ex-situ with atomic force microscopy (AFM) and scanning electron microscopy (SEM). The in-plane and out-of-plane strain and lattice parameters were determined from an analysis of (004) and (224) high resolution x-ray rocking curves (HRXRCs),9 in both φ = 0° and φ = 180° configurations. To determine the Bi fraction, x, we use a linear interpolation of the measured GaAs (5.6533 Å) and calculated GaBi (6.324 Å) lattice parameters10 along with a Poisson's ratio of 0.33. In many cases, x was also determined using Rutherford backscattering spectrometry (RBS) measurements in conjunction with simulation of the nuclear reaction analysis (SIMNRA) code; the HRXRC and RBS determinations of x agree to within ±0.005 for x < 0.04 and ±0.01 for x > 0.04. For all the GaAs(Bi):Si films, room temperature resistivity and Hall measurements were performed in the van der Pauw configuration,11 using an external magnetic field of 0.124 T and a DC current between 5 μA and 6 mA. In0.52Sn0.48 Ohmic contacts were applied to the corners of 4 mm × 4 mm cleaved squares and annealed at 420 °C for 2 min. The composition profiles across the surface droplets were determined using energy dispersive x-ray spectroscopy (EDS) and/or Auger electron spectroscopy. For select films, surface Ga droplets were removed via surface etching with HCl:H2O (1:3),12,13 followed by subsequent AFM, SEM, EDS, and resistivity and Hall measurements.

To monitor the surface reconstruction, RHEED patterns were collected along the [110] and [11¯0] directions, as shown in Fig. 1. During GaAs growth at 600 °C, a streaky (2 × 4) pattern is observed, indicating group V termination of the surface [Figs. 1(a) and 1(b)]. As the substrate temperature is lowered to 280 °C, the RHEED pattern transforms to a streaky (2 × 3) pattern [Figs. 1(c) and 1(d)], independent of the presence of Si. In some cases, spotty (2 × 1) RHEED patterns with chevrons [Figs. 1(e) and 1(f)] emerge during the growth of GaAs(Bi):Si films; these films exhibit p-type conductivity. In other cases, within 10 s of GaAs(Bi):Si growth, streaky (n × 3) patterns [Figs. 1(g) and 1(h)] appear, with (2 × chevron) patterns emerging following at least 45 s of growth; these films exhibit n-type conductivity.

FIG. 1.

Reflection high-energy electron diffraction patterns collected along [110] and [11¯0] axes during the GaAs(Bi):(Si) film growth. [(a) and (b)] a (2 × 4) pattern during GaAs growth at 600 °C; [(c) and (d)] a streaky (2 × 3) pattern during GaAs(Si) growth at 280 °C; [(e) and (f)] a dim (2 × chevron) pattern during p-type GaAs(Bi):Si growth at 280 °C, with arrows indicating the chevrons; [(g) and (h)] a streaky (1 × 3) pattern during n-type GaAs(Bi):Si growth at 280 °C.

FIG. 1.

Reflection high-energy electron diffraction patterns collected along [110] and [11¯0] axes during the GaAs(Bi):(Si) film growth. [(a) and (b)] a (2 × 4) pattern during GaAs growth at 600 °C; [(c) and (d)] a streaky (2 × 3) pattern during GaAs(Si) growth at 280 °C; [(e) and (f)] a dim (2 × chevron) pattern during p-type GaAs(Bi):Si growth at 280 °C, with arrows indicating the chevrons; [(g) and (h)] a streaky (1 × 3) pattern during n-type GaAs(Bi):Si growth at 280 °C.

Close modal

We now discuss the structure and composition of the GaAs(Bi):(Si) films. A representative series of HRXRCs about the (004) GaAs diffraction peak are shown in Fig. 2 for GaAs1−xBix films with x ranging from 0 to 0.048. For all films, diffractions peaks associated with GaAsBi and GaAs are observed, with the peak separations increasing with increasing x. For low x, Pendellosung fringes are apparent, indicating smooth epilayer surfaces and interfaces. As x increases, the Pendellosung fringes disappear, and the GaAsBi diffraction peak broadens, presumably due to roughening of the surfaces and interfaces due to misfit strain relaxation.14 In Fig. 3(a), we plot the in-plane (a) and perpendicular (a) lattice parameters as a function of x for GaAs(Bi) films, with and without Si. For x < 0.04, a remains constant while a increases linearly with x, consistent with a pseudomorphically strained film. For x > 0.04, a begins to increase while a remains nearly constant, suggesting that the film lattice has begun to relax to its intrinsic lattice parameter. We note that both a and a are independent of the free carrier type and Si concentration.

FIG. 2.

Representative high-resolution x-ray rocking curves for GaAs1-xBix films, with x ranging from 0 to 0.048. For x < 0.03, Pendellosung fringes indicate a high crystalline quality and coherent interfaces. As x increases, the Pendellosung fringes disappear, and the GaAsBi diffraction peak broadens.

FIG. 2.

Representative high-resolution x-ray rocking curves for GaAs1-xBix films, with x ranging from 0 to 0.048. For x < 0.03, Pendellosung fringes indicate a high crystalline quality and coherent interfaces. As x increases, the Pendellosung fringes disappear, and the GaAsBi diffraction peak broadens.

Close modal
FIG. 3.

(a) In-plane, a, and perpendicular, a, lattice parameters and (b) FWHM of the (004) GaAsBi diffraction peak and upper bound for excess arsenic concentration, all as a function of Bi fraction, x. For x < 0.04, a increases with increasing x, independent of doping type and concentration. For x > 0.04, a begins to increase while a is nearly constant, suggesting that the film lattice has begun to relax to its intrinsic lattice parameter. For undoped GaAsBi and n-type GaAsBi:Si with x < 0.04, the FWHM of the GaAsBi peak is nearly constant; for x > 0.04, the FWHM increases with x. For p-type GaAsBi:Si, the FWHM of the GaAsBi peak increases monotonically with x.

FIG. 3.

(a) In-plane, a, and perpendicular, a, lattice parameters and (b) FWHM of the (004) GaAsBi diffraction peak and upper bound for excess arsenic concentration, all as a function of Bi fraction, x. For x < 0.04, a increases with increasing x, independent of doping type and concentration. For x > 0.04, a begins to increase while a is nearly constant, suggesting that the film lattice has begun to relax to its intrinsic lattice parameter. For undoped GaAsBi and n-type GaAsBi:Si with x < 0.04, the FWHM of the GaAsBi peak is nearly constant; for x > 0.04, the FWHM increases with x. For p-type GaAsBi:Si, the FWHM of the GaAsBi peak increases monotonically with x.

Close modal

Since the GaAs(Bi) layers are grown at low temperatures, the dopant incorporation and transport properties may be influenced by As antisites, [AsGa]. [AsGa] is often determined from HRXRC measurements of an increase in the lattice parameter of the low temperature GaAs (LT-GaAs) layer according to Δa/a = 1.24 × 10−23 [AsGa], where a is the lattice parameter of the GaAs substrate, Δa is the difference between a and the LT-GaAs lattice parameter, and [AsGa] is in cm−3.15 Due to the absence of a distinct LT-GaAs diffraction peak, we estimate an upper bound for [AsGa] using the FWHM of the GaAs peak. For all the GaAs layers, [AsGa] ≤ 3 × 1019 cm−3. Similarly, for the GaAsBi layers, we consider the FWHM of the GaAsBi film peak as an upper bound for [AsGa]. In Fig. 3(b), the FWHM of the GaAsBi diffraction peak and corresponding [AsGa] upper bound are shown as a function of x. For undoped GaAsBi and n-type GaAsBi:Si with x < 0.04, the FWHM of the GaAsBi peak is typically 48 ± 7 arc sec, corresponding to an [AsGa] upper bound of ∼(3.0 ± 0.5) × 1019 cm−3. For x ≥ 0.04, the FWHM increases to 500 ± 100 arc sec, corresponding to an [AsGa] upper bound of ∼(3.0 ± 0.6) × 1020 cm−3. For p-type GaAsBi films, the FWHM increases from 150 to 500 arc sec, corresponding to the [AsGa] upper bound increasing from 1 to 3 × 1020 cm−3. In undoped GaAs with [AsGa] up to 1020 cm−3, free electron concentrations up to 3 × 1016 cm−3 have been reported.16,17 Since the carrier concentrations of our n-type and p-type GaAs(Bi):Si films exceed 1017 cm−3, it is expected that compensation by excess-As-induced free carriers is insignificant for the GaAs(Bi) films.

The conductivity types for GaAs(Bi):Si films are summarized in Fig. 4, a plot of arsenic BEP vs. gallium BEP, with growth rate on the upper x-axis for the As-rich growth regime. A dashed line indicates where the group V/III BEP ratio is equal to 10, “BEPratio10.” All GaAsBi:Si films above [below] the BEPratio10 line are grown on (n × 3) [(2 × 1)] reconstructed surfaces and are n-type [p-type]. A similar transition from (n × 3) to (2 × 1) with decreasing group V/III ratio has been reported for GaAsBi growth with As2.18 In addition, along the BEPratio10 line, a conversion from n- to p-type conductivity occurs as the growth rate increases (from 0.25 to 2.0 μm/h), and the surface reconstruction changes from (n × 3) to (2 × 1). This growth-rate-dependence of the transition to a (2 × 1) surface reconstruction is consistent with the growth-rate-dependence of the stoichiometric threshold reported for GaAsBi growth using As2.19 

FIG. 4.

Conductivity type for GaAs(Bi):Si films: arsenic beam-equivalent pressure (BEP) vs. Ga BEP, with growth rate on the upper x-axis for the As-rich growth regime. A dashed line indicates the regime where the group V/III BEP ratio is equal to 10, “BEPratio10.” All GaAsBi:Si films above (below) the BEPratio10 line are n-type (p-type). In addition, along the BEPratio10 line, a conductivity type conversion from n- to p-type occurs as the growth rate increases (from 0.25 to 2.0 μm/h), consistent with the growth rate dependence of the stoichiometric threshold reported in Ref. 19.

FIG. 4.

Conductivity type for GaAs(Bi):Si films: arsenic beam-equivalent pressure (BEP) vs. Ga BEP, with growth rate on the upper x-axis for the As-rich growth regime. A dashed line indicates the regime where the group V/III BEP ratio is equal to 10, “BEPratio10.” All GaAsBi:Si films above (below) the BEPratio10 line are n-type (p-type). In addition, along the BEPratio10 line, a conductivity type conversion from n- to p-type occurs as the growth rate increases (from 0.25 to 2.0 μm/h), consistent with the growth rate dependence of the stoichiometric threshold reported in Ref. 19.

Close modal

We next discuss a mechanism for the transition from n- to p-type conductivity. During the growth on a (1 × 3) or (2 × 3) reconstructed surface, Si must break an As dimer in order to occupy a group V site;20 thus, Si more easily incorporates into group III sites, resulting in n-type conductivity. However, for a (2 × 1) reconstruction without As dimers,21–23 Si incorporation into group V sites is enabled, leading to p-type conductivity. This arsenic-dimer mediated dopant incorporation mechanism also explains previous reports of conductivity type conversion in GaAs. Zhang8 and Quivy6 observed an n- to p-type conductivity conversion following the transition from a (2 × 4) to a (4 × 2) reconstruction, associated with the presence24 and absence25 of surface As dimers. In the case of Fukushima,7 the conductivity type conversion occurred at an As4/Ga BEP ratio of 12, with a substrate temperature of 270 °C, which typically correspond to a transition from an (n × 3) to a (2 × 1) surface reconstruction, associated with the presence and absence of surface As dimers.

This surface-reconstruction-induced transition in conductivity type provides the opportunity to examine Si as both an n- and p-type dopant in III–V semiconductor compounds, especially for alloys such as GaAsBi, which are typically grown at low temperatures near the stoichiometric threshold.26 In Fig. 5, we plot conductivity, σ, as a function of carrier concentration for our GaAs(Bi):Si films, in comparison with those of literature reports for GaAsBi films with various n- and p-type dopants.27–31 In the plot of σ vs. n, dashed lines form an envelope corresponding to mobilities of 2300 and 45 cm2/(V-s). The n-type film mobilities are typically near the upper end of the range, ∼2300 cm2/(V-s), whereas the p-type film mobilities are typically near the lower end, ∼45 cm2/(V-s), all independent of x. For the n-type films, the mobilities are consistent with those of conventionally grown GaAs.32 Indeed, as shown in the SEM and AFM images in Figs. 6(a) and 6(c), n-type GaAsBi films are smooth and droplet-free.

FIG. 5.

Conductivity vs. carrier concentration for bulk GaAs1-xBix:Si films, in comparison with literature reports for GaAsBi films doped with C or Be. The dotted lines form an envelope corresponding to mobilities of 2300 and 45 cm2/(V-s). The n-type film mobilities are typically near the upper end of the range (∼2300 cm2/(V-s)) whereas the p-type film mobilities are typically near the lower end (∼45 cm2/(V-s)), all independent of Bi fraction. For our n-type films grown near the transition from n- to p-type conductivity, mobilities lie near the mid-range of the envelope. The uncertainties in the conductivities and carrier concentrations are comparable to the size of the data points.27–31 

FIG. 5.

Conductivity vs. carrier concentration for bulk GaAs1-xBix:Si films, in comparison with literature reports for GaAsBi films doped with C or Be. The dotted lines form an envelope corresponding to mobilities of 2300 and 45 cm2/(V-s). The n-type film mobilities are typically near the upper end of the range (∼2300 cm2/(V-s)) whereas the p-type film mobilities are typically near the lower end (∼45 cm2/(V-s)), all independent of Bi fraction. For our n-type films grown near the transition from n- to p-type conductivity, mobilities lie near the mid-range of the envelope. The uncertainties in the conductivities and carrier concentrations are comparable to the size of the data points.27–31 

Close modal
FIG. 6.

Scanning electron microscopy, corresponding energy dispersive x-ray spectroscopy (EDS), and atomic force microscopy (AFM) images for [(a)–(c)] n-type GaAs0.992Bi0.008:Si, as well as p-type GaAs0.991Bi0.009:Si [(d)–(f)] before and [(g)–(i)] after etching with HCl:H20 (1:3). In the EDS images, red, green, and yellow correspond to Ga, As, and Bi, respectively. The gray-scale ranges displayed in the AFM images are (c) 10 nm, (f) 160 nm, and (i) 100 nm. Line-cuts of the tip height across the AFM images are shown beneath the images, with height axis from 0 to 250 nm.

FIG. 6.

Scanning electron microscopy, corresponding energy dispersive x-ray spectroscopy (EDS), and atomic force microscopy (AFM) images for [(a)–(c)] n-type GaAs0.992Bi0.008:Si, as well as p-type GaAs0.991Bi0.009:Si [(d)–(f)] before and [(g)–(i)] after etching with HCl:H20 (1:3). In the EDS images, red, green, and yellow correspond to Ga, As, and Bi, respectively. The gray-scale ranges displayed in the AFM images are (c) 10 nm, (f) 160 nm, and (i) 100 nm. Line-cuts of the tip height across the AFM images are shown beneath the images, with height axis from 0 to 250 nm.

Close modal

In Figs. 6(d) and 6(f), representative SEM and AFM images are shown for p-type GaAs(Bi):Si films, which contain sub-micron diameter surface droplets. A corresponding EDS image, with red, green, and yellow representing Ga, As, and Bi, shown in Fig. 6(e). The surface droplets are composed primarily of Ga, often with segregated patches of Bi, consistent with similar 2D EDS mapping of GaAsBi films.19 Droplets presumably begin to form at the start of the GaAs(Bi):Si growth and gradually develop, leading to both a dimming of the RHEED pattern and the appearance of chevrons, likely associated with crater formation, similar to earlier reports of curved crystalline surface formation during growth.33,34 AFM reveals that the droplets are displaced from 20 ± 10 nm deep craters [circled in Fig. 6(f)], consistent with reports of Ga droplets deposited at 350 °C.35 Following removal of the droplets via chemical etching, the surface craters are retained, as shown in Fig. 6(g). The corresponding EDS image in Fig. 6(h) confirms the absence of Ga-rich regions within the craters. Furthermore, for the p-type GaAsBi:Si film shown in Fig. 6, the carrier concentration is 2.1 ± 0.2 × 1018 cm−3, the hole mobility is 29.5 ± 1.5 cm2/(V-s), and the conductivity is 9.8 ± 0.6 S/cm, both before and after etching. The apparent lack of influence of the Ga surface droplets on the GaAs(Bi):Si film transport properties suggests that our 20 ± 2% droplet surface coverage is below the percolation threshold for conductivity, typically >70%.36,37

The p-type GaAs(Bi):Si films with high hole concentrations (p = 2–6 × 1018 cm−3) exhibit mobilities consistent with conventionally grown GaAs.38 For lower hole concentrations, p ∼ 2 × 1017 cm−3, GaAs(Bi):Si hole mobilities are lower than those of typical GaAs but similar to those reported for GaAsBi doped with Be or C.29–31 Since similar hole mobilities are observed in the p-type films, independent of the presence of Bi, the hole mobility reduction is likely related to point defects intrinsic to Ga-rich growth, such as As vacancies.39 The n-type GaAsBi films grown at the lowest group V/III BEP ratios are compensated, with mobilities near the mid-range of the envelope in Fig. 5.40 Since the group V/III BEP ratios of these compensated films lie very close to the transition from n- to p-type conductivity, it is likely that Si donors are compensated by Si acceptors. Work is underway to quantify and control the relative concentrations of Si donors and acceptors in both GaAs and GaAsBi films.

In summary, we report on the influence of surface reconstruction on Si doping type and transport properties of GaAs(Bi) alloys. A transition from n- to p-type conductivity is observed as the surface reconstruction transitions from (n × 3) to (2 × 1). We hypothesize that the presence (absence) of surface arsenic dimers prevents (enables) dopant incorporation into arsenic lattice sites, resulting in n-type (p-type) conductivity. We consider the influence of bismuth anions on arsenic-dimer mediated dopant incorporation and the resulting electronic transport properties, demonstrating the applicability of this mechanism to mixed anion semiconductor alloys.

This work was supported by the National Science Foundation (Grant No. DMR 1410282).

1.
H.
Ito
and
T.
Ishibashi
,
Jpn. J. Appl. Phys.
30
,
L944
(
1991
).
2.
A. Y.
Cho
and
I.
Hayashi
,
J. Appl. Phys.
42
,
4422
(
1971
).
3.
N.
Sakamoto
,
K.
Hirakawa
, and
T.
Ikoma
,
Appl. Phys. Lett.
67
,
1444
(
1995
).
4.
K.
Sato
,
M. R.
Fahy
,
M. J.
Ashwin
, and
B. A.
Joyce
,
J. Cryst. Growth
165
,
345
(
1996
).
5.
S.
Subbanna
,
H.
Kroemer
, and
J. L.
Merz
,
J. Appl. Phys.
59
,
488
(
1986
).
6.
A. A.
Quivy
,
A. L.
Sperandio
,
E. C. F.
da Silva
, and
J. R.
Leite
,
J. Cryst. Growth
206
,
171
(
1999
).
7.
S.
Fukushima
,
T.
Obata
, and
N.
Otsuka
,
J. Appl. Phys.
89
,
380
(
2001
).
8.
D.
Zhang
,
K.
Radhakrishnan
, and
S.
Yoon
,
J. Cryst. Growth
135
,
441
(
1994
).
9.
Bede D1 System.
10.
A.
Janotti
,
S.-H.
Wei
, and
S.
Zhang
,
Phys. Rev. B
65
,
115203
(
2002
).
11.
ASTM F76-08(2016)
,
Standard Test Methods for Measuring Resistivity and Hall Coefficient and Determining Hall Mobility in Single-Crystal Semiconductors
(
ASTM International
,
West Conshohocken, PA
,
2016
).
12.
C. X.
Zheng
,
W. X.
Tang
, and
D. E.
Jesson
,
Appl. Phys. Lett.
100
,
071903
(
2012
).
13.
C.
Somaschini
,
S.
Bietti
,
N.
Koguchi
, and
S.
Sanguinetti
,
Nano Lett.
9
,
3419
(
2009
).
14.
D. F.
Reyes
,
F.
Bastiman
,
C. J.
Hunter
,
D. L.
Sales
,
A. M.
Sanchez
,
J. P. R.
David
, and
D.
Gonzalez
,
Nanoscale Res. Lett.
9
,
23
(
2014
).
15.
X.
Liu
,
A.
Prasad
,
J.
Nishio
,
E. R.
Weber
,
Z.
Liliental-Weber
, and
W.
Walukiewicz
,
Appl. Phys. Lett.
67
,
279
(
1995
).
16.
J.
Betko
,
P.
Kordos
,
S.
Kuklovsky
,
A.
Förster
,
D.
Gregusová
, and
H.
Lüth
,
Mater. Sci. Eng. B
28
,
147
(
1994
).
17.
S.
O'Hagan
and
M.
Missous
,
J. Appl. Phys.
75
,
7835
(
1994
).
18.
M.
Masnadi-Shirazi
,
D. A.
Beaton
,
R. B.
Lewis
,
X.
Lu
, and
T.
Tiedje
,
J. Cryst. Growth
338
,
80
(
2012
).
19.
G.
Vardar
,
S. W.
Paleg
,
M. V.
Warren
,
M.
Kang
,
S.
Jeon
, and
R. S.
Goldman
,
Appl. Phys. Lett.
102
,
42106
(
2013
).
20.
T.
Kita
,
O.
Wada
,
T.
Nakayama
, and
M.
Murayama
,
Phys. Rev. B
66
,
195312
(
2002
).
21.
D. J.
Chadi
,
J. Vac. Sci. Technol., A
5
,
834
(
1987
).
22.
P.
Laukkanen
,
M. P. J.
Punkkinen
,
H.-P.
Komsa
,
M.
Ahola-Tuomi
,
K.
Kokko
,
M.
Kuzmin
,
J.
Adell
,
J.
Sadowski
,
R. E.
Perälä
,
M.
Ropo
,
T. T.
Rantala
,
I. J.
Väyrynen
,
M.
Pessa
,
L.
Vitos
,
J.
Kollár
,
S.
Mirbt
, and
B.
Johansson
,
Phys. Rev. Lett.
100
,
086101
(
2008
).
23.
F.
Bastiman
,
A. G.
Cullis
,
J. P. R.
David
, and
S. J.
Sweeney
,
J. Cryst. Growth
341
,
19
(
2012
).
24.
25.
I.
Kamiya
,
D. E.
Aspnes
,
L. T.
Florez
, and
J. P.
Harbison
,
Phys. Rev. B
46
,
15894
(
1992
).
26.
S.
Tixier
,
M.
Adamcyk
,
T.
Tiedje
,
S.
Francoeur
,
A.
Mascarenhas
,
P.
Wei
, and
F.
Schiettekatte
,
Appl. Phys. Lett.
82
,
2245
(
2003
).
27.
R. N.
Kini
,
L.
Bhusal
,
A. J.
Ptak
,
R.
France
, and
A.
Mascarenhas
,
J. Appl. Phys.
106
,
043705
(
2009
).
28.
B.
Fluegel
,
R. N.
Kini
,
A. J.
Ptak
,
D.
Beaton
,
K.
Alberi
, and
A.
Mascarenhas
,
Appl. Phys. Lett.
99
,
162108
(
2011
).
29.
D. A.
Beaton
,
R. B.
Lewis
,
M.
Masnadi-Shirazi
, and
T.
Tiedje
,
J. Appl. Phys.
108
,
083708
(
2010
).
30.
R. N.
Kini
,
A. J.
Ptak
,
B.
Fluegel
,
R.
France
,
R. C.
Reedy
, and
A.
Mascarenhas
,
Phys. Rev. B
83
,
075307
(
2011
).
31.
K.
Kado
,
T.
Fuyuki
,
K.
Yamada
,
K.
Oe
, and
M.
Yoshimoto
,
Jpn. J. Appl. Phys.
51
,
040204
(
2012
).
32.
J. R.
Meyer
and
F. J.
Bartoli
,
Phys. Rev. B
36
,
5989
(
1987
).
33.
T.
Hanada
,
B. H.
Koo
,
H.
Totsuka
, and
T.
Yao
,
Phys. Rev. B: Condens. Matter Mater. Phys.
64
,
1653071
(
2001
).
34.
D. W.
Pashley
,
J. H.
Neave
, and
B. A.
Joyce
,
Surf. Sci.
476
,
35
(
2001
).
35.
C.
Somaschini
,
S.
Bietti
,
S.
Sanguinetti
,
A.
Fedorov
,
N.
Koguchi
, and
A.
Fedorov
,
Nanotechnology
21
,
125601
(
2010
).
36.
A.
Okazaki
,
K.
Horibe
,
K.
Maruyama
, and
S.
Miyazima
,
Phys. Rev. E
61
,
6215
(
2000
).
37.
S.
Wagner
and
A.
Pundt
,
Phys. Rev. B
78
,
155131
(
2008
).
38.
W.
Walukiewicz
,
Phys. Rev. B
41
,
10218
(
1990
).
39.
J. E.
Northrup
and
S. B.
Zhang
,
Phys. Rev. B
50
,
4962
(
1994
).
40.
W.
Walukiewicz
,
L.
Lagowski
,
L.
Jastrzebski
,
M.
Lichtensteiger
, and
H. C.
Gatos
,
J. Appl. Phys.
50
,
899
(
1979
).