In this study, we investigated the basic electrical properties of Si-doped wurtzite GaN films prepared using a low-temperature pulsed sputtering deposition (PSD) process. We found that the electron concentration can be controlled in the range between 1.5 × 1016 and 2.0 × 1020 cm−3. For lightly Si-doped GaN ([Si] = 2.1 × 1016 cm−3), the room temperature (RT) electron mobility was as high as 1008 cm2 V−1 s−1, which was dominantly limited by polar optical phonon scattering. Moreover, we found that heavily Si-doped GaN prepared using PSD exhibited an RT mobility as high as 110 cm2 V−1 s−1 at an electron concentration of 2 × 1020 cm−3, which indicated that the resistivity of this film was almost as small as those of typical transparent conductive oxides such as indium tin oxide. At lower temperatures, the electron mobility increased to 1920 cm2 V−1 s−1 at 136 K, and the temperature dependence was well explained by conventional scattering models. These results indicate that Si-doped GaN prepared using PSD is promising not only for the fabrication of GaN-based power devices but also for use as epitaxial transparent electrode materials for nitride based optical devices.

GaN has excellent material characteristics for use in power devices, including a high breakdown voltage, high saturation velocity, and high thermal stability. Recent advances in bulk GaN growth technology have facilitated the development of vertical power devices such as Schottky barrier diodes,1,2 p–n junction diodes,3,4 and trench metal–oxide–semiconductor field-effect transistors.5 One of the key methods used to fabricate these devices is a light n-type doping of GaN with a low residual impurity concentration of the order of 1015 cm−3 or less. Despite intensive research efforts, the performance of GaN-based power devices has remained insufficient because of an immature epitaxial growth process. The most common method, metal organic chemical vapor deposition (MOCVD), inherently results in GaN that is contaminated by carbon, oxygen, and silicon atoms originating from the metal-organic precursors, susceptors, and reactor walls. The extent of contamination complexly depends on the growth conditions, including the growth temperature, III/V ratio, gas flow rate, and reactor pressure.6 In particular, several studies on MOCVD have reported an undesirable increase in carbon concentration in GaN with increasing TMGa flow rate7 or growth rate.8 By contrast, a new growth technique called pulsed sputtering deposition (PSD) has recently attracted much attention. Recent progress in PSD has enabled the growth of device-quality group-III nitrides at much lower temperatures than those used in a conventional MOCVD process.9–12 PSD is also a suitable method for growing high-purity GaN because its growth system does not contain carbon or hydrogen atoms in the source material. In fact, we demonstrated the growth of high-purity p-type GaN with a high room temperature (RT) hole mobility of 34 cm2 V−1 s−1. This is attributed to a reduction in the amount of residual impurities such as oxygen, hydrogen, and carbon, which should also be useful for the growth of n-type GaN.13 

In this study, we investigated how the electrical properties of GaN prepared by PSD depend on the Si doping concentrations using temperature-dependent Hall-effect measurements.

Growth of 1-μm-thick Si-doped GaN films was performed by PSD with pulsed magnetron sputtering sources in an N2/Ar atmosphere. The Si doping concentration in GaN was controlled from 2 × 1016 to 2 × 1020 cm−3 by varying Si vapor flux from a solid state single crystalline Si source. Heavily Si-doped samples ([Si] > 1 × 1017 cm−3) were grown on Fe-doped GaN on sapphire prepared by MOCVD with a latter-half threading dislocation density of 108 cm−2. Lightly Si-doped samples ([Si] < 1 × 1017 cm−3) were grown on Fe-doped bulk GaN substrates prepared by hydride vapor phase epitaxy with threading dislocation densities on the order of 106 cm−2. The concentration of Si dopants in GaN was evaluated by secondary ion mass spectroscopy (SIMS). Details of the growth method have been reported elsewhere.9–14 To fabricate van der Pauw structures, we processed all of the samples into cloverleaf patterns by photolithography and inductively coupled plasma-reactive-ion etching. Ohmic contacts were then formed with Ti/Al/Ti/Au (20/60/20/50 nm) as electrodes. Temperature-dependent Hall-effect measurements were performed using a ResiTest 8400 (Toyo Corp.) with a liquid-N2-cooled temperature-variable sample holder. The analytical protocol was based on American Society for Testing and Materials standard F76;15 the Hall scattering factor was unity.

Figure 1 shows the RT electron mobilities μe of Si-doped PSD GaN as a function of the electron concentration Ne. The RT electron concentration was controlled in the concentration range between 1.5 × 1016 and 2.0 × 1020 cm−3 by Si doping. As the Si doping concentration decreased, the RT electron mobility reached 1008 cm2 V−1 s−1 at an electron concentration of 1.5 × 1016 cm−3. On the other hand, the most heavily doped sample with a Si concentration of 2 × 1020 cm−3 exhibited an electron mobility as high as 110 cm2 V−1 s−1. These results indicate that the resistivity of this film was almost as small as those for typical transparent conductive oxides such as indium tin oxide (ITO). It can be noted that the mobility of 110 cm2 V−1 s−1 is much higher than that for ITO, which should lead to smaller free-carrier light absorption. These data indicate that the pulse-sputtered Si-doped GaN is a potential transparent electrode material for nitride based optical devices. We also note that the fabrication of such a highly conductive Si-doped GaN material by MOCVD has not been previously reported.

An S-shaped increase in the electron mobility with decreasing electron concentration is clearly observed. This trend can be empirically fit using the Caughey–Thomas equation16 

μ ( N ) = μ min + μ max μ min 1 + ( N N g ) γ ,
(1)

where μmax, μmin, and Ng are the fitting parameters, γ is unity for an n-type semiconductor, and N is the doping concentration with an assumption of N = Ne for simplicity. As evident in Fig. 1, the experimentally obtained data were well fitted by this model with the parameters μmax = 1035 cm2 V−1 s−1, μmin= 130 cm2 V−1 s−1, and Ng = 3.0 × 1017 cm−3. Mnatsakanov et al. also summarized the relationship between electron mobility and electron concentration in MOCVD-grown Si-doped GaN using the same model and extracted fitting parameters as μmax = 1000 cm2 V−1 s−1, μmin = 55 cm2 V−1 s−1, and Ng = 2.0 × 1017 cm−3. The obtained μmax for PSD-grown samples was greater than or equal to the values for the state-of-the-art MOCVD-grown samples.17 

Figures 2(a) and 2(b) show the temperature dependence of electron concentration and mobility, respectively, for Si-doped GaN between 77 and 300 K. For highly Si-doped samples ([Si] > 8 × 1018 cm−3), almost no temperature dependence in the electron concentration exists, as evident in Fig. 2(a). In this doping range, the electron mobility is also independent of the temperature at Si doping concentrations as high as 2 × 1020 cm−3, as shown in Fig. 2(b). We also observed that almost all Si atoms were active as n-type dopants in GaN, even at a Si concentration as high as 2 × 1020 cm−3. This result indicates that any self-compensation effect18 and/or the undesirable formation of silicon nitride,19 which are undesirable for heavy Si doping by MOCVD, were negligible in the case of GaN grown using PSD.

For samples doped with lower concentrations of Si ([Si] < 1 × 1017 cm−3), the experimental data related to the temperature dependence of electron concentrations were well fitted under a charge neutrality condition as a non-degenerate n-type semiconductor. The Si doping concentration obtained from the fitting well agreed with the SIMS results. The estimated activation energy of the Si dopant varied from 21 to 24 meV with decreasing Si doping concentration; this behavior is explained by the Coulomb interaction of ionized donors. The details of the fitting results will be summarized elsewhere. With respect to the temperature dependence of the electron mobilities of the samples doped with lower Si contents ([Si] < 1 × 1017 cm − 3) shown in Fig. 2(b), the highest peak mobility was 1920 cm2 V−1 s−1 at 136 K for the lightly Si-doped sample ([Si] = 2 × 1016 cm−3). We also observed that the peaks of the electron mobility were shifted toward lower temperatures. This phenomenon is explained by the reduction in the ionized impurity scattering rate with decreasing Si doping concentration. The detailed scattering mechanism is discussed later.

Next, we focus on the electron transport properties of Si-doped GaN with the highest electron mobility in Fig. 2(b). Figures 3(a) and 3(b) show a reflection high-energy electron diffraction (RHEED) pattern and a 5 × 5 μm2 atomic force microscopy (AFM) surface image of lightly Si-doped GaN ([Si] = 2 × 1016 cm−3) grown on the bulk GaN substrate. A clear streaky RHEED pattern is observed in Fig. 3(a), which indicates that the growth of Si-doped GaN using PSD proceeded in the two-dimensional growth mode. Figure 3(b) shows an atomically flat step-and-terrace structure with a terrace width of approximately 1 μm. The root mean square of the surface roughness on the terraces was as low as 0.5 nm. These AFM data correspond well with the streaky RHEED pattern.

Figure 3(c) shows the detailed temperature dependence of electron mobility between 78 and 400 K and the calculated electron mobility limited by interactions of electrons with impurities, lattice defects, and lattice vibrations. According to Matthiessen's rule, the electron mobility is expressed as 1 / μ t o t a l = i 1 / μ i , where μtotal and μi are the total electron mobility and the electron mobility limited by each scattering process, respectively. For the calculation, expressions for the scattering-limited electron mobilities are taken from the following references: ionized impurity,20 neutral impurity,21 dislocation,22 polar optical phonon,20 acoustic deformation potential,20 and piezoelectric scattering.21 We assumed that the effects of other scatterings, such as non-polar optical phonon scattering, are negligible on electron mobility. All parameters are defined in Table I. Figure 3(c) qualitatively clarifies the contribution of each scattering mechanism for temperature dependence of the experimental electron mobility. This calculation result reveals that the RT electron mobility of PSD-grown Si-doped GaN is mainly limited by polar optical phonon scattering. By contrast, at temperatures below 200 K, the ionized impurity scattering became dominant. Reducing the concentration of compensating acceptors is important for further improving the electron mobility in PSD-grown Si-doped GaN. Using high-quality bulk GaN substrates with a low threading dislocation density is also important because the highest low-temperature electron mobility was attained for the sample grown on a substrate with a very low threading dislocation density of 5 × 105 cm−2.25 

In summary, we have grown n-type GaN films with Si dopant concentrations on the order ranging from 1016 to 1020 cm−3 by PSD. The relationship between the RT electron concentration and mobility, as analyzed by the Caughey–Thomas model, revealed μmax and μmin extracted electron mobilities of 1035 and 130 cm2 V−1 s−1, respectively. For lightly Si-doped GaN ([Si] = 2 × 1016 cm−3), the RT electron mobility was 1008 cm2 V−1 s−1, which was dominantly limited by polar optical phonon scattering. With a reduction in temperature, electron mobility increases; the peak mobility was 1920 cm2 V−1 s−1 at 136 K, which is mainly limited by ionized impurity scattering. We also observed that heavily Si-doped GaN prepared by PSD exhibited an RT mobility as high as 110 cm2 V−1 s−1 at an electron concentration of 2 × 1020 cm−3, which indicates that the resistivity of this film was almost as small as those of typical transparent conductive oxides such as ITO. Because the electron mobility of the GaN films is higher than those of the transparent conductive oxides, Si-doped GaN prepared by PSD is a promising epitaxial transparent electrode material for nitride based optical devices.

This work was partially supported by the JST ACCEL project and JSPS KAKENHI Grant No. JP16H06414.

1.
Y.
Saitoh
,
K.
Sumiyoshi
,
M.
Okada
,
T.
Horii
,
T.
Miyazaki
,
H.
Shiomi
,
M.
Ueno
,
K.
Katayama
,
M.
Kiyama
, and
T.
Nakamura
,
Appl. Phys. Express
3
,
081001
(
2010
).
2.
N.
Tanaka
,
K.
Hasegawa
,
K.
Yasunishi
,
N.
Murakami
, and
T.
Oka
,
Appl. Phys. Express
8
,
071001
(
2015
).
3.
Z.
Hu
,
K.
Nomoto
,
B.
Song
,
M.
Zhu
,
M.
Qi
,
M.
Pan
,
X.
Gao
,
V.
Protasenko
,
D.
Jena
, and
H. G.
Xing
,
Appl. Phys. Lett.
107
,
243501
(
2015
).
4.
I. C.
Kizilyalli
,
A. P.
Edwards
,
O.
Aktas
,
T.
Prunty
, and
D.
Bour
,
IEEE Trans. Electron Devices
62
,
414
(
2015
).
5.
T.
Oka
,
T.
Ina
,
Y.
Ueno
,
T.
Ina
, and
K.
Hasegawa
,
Appl. Phys. Express
7
,
021002
(
2014
).
6.
D. D.
Koleske
,
A. E.
Wickenden
,
R. L.
Henry
, and
M. E.
Twigg
,
J. Crystal Growth
242
,
55
(
2002
).
7.
J. T.
Chen
,
U.
Forsberg
, and
E.
Janzen
,
Appl. Phys. Lett.
102
,
193506
(
2013
).
8.
Y.
Cao
,
R.
Chu
,
R.
Li
,
M.
Chen
,
R.
Chang
, and
B.
Hughes
,
Appl. Phys. Lett.
108
,
062103
(
2016
).
9.
K.
Ueno
,
E.
Kishikawa
,
S.
Inoue
,
J.
Ohta
,
H.
Fujioka
,
M.
Oshima
, and
H.
Fukuyama
,
Phys. Status Solidi RRL
8
,
256
(
2014
).
10.
E.
Nakamura
,
K.
Ueno
,
J.
Ohta
,
H.
Fujioka
, and
M.
Oshima
,
Appl. Phys. Lett.
104
,
051121
(
2014
).
11.
J. W.
Shon
,
J.
Ohta
,
K.
Ueno
,
A.
Kobayashi
, and
H.
Fujioka
,
Sci. Rep.
4
,
5325
(
2014
).
12.
T.
Watanabe
,
J.
Ohta
,
T.
Kondo
,
M.
Ohashi
,
K.
Ueno
,
A.
Kobayashi
, and
H.
Fujioka
,
Appl. Phys. Lett.
104
,
182111
(
2014
).
13.
Y.
Arakawa
,
K.
Ueno
,
A.
Kobayashi
,
J.
Ohta
, and
H.
Fujioka
,
APL Mater.
4
,
086103
(
2016
).
14.
K.
Sato
,
J.
Ohta
,
S.
Inoue
,
A.
Kobayashi
, and
H.
Fujioka
,
Appl. Phys. Express
2
,
011003
(
2009
).
15.
ASTM Standard F76, Test Methods for Measuring Resistivity and Hall Coefficient and Determining Hall Mobility in Single-Crystal Semiconductors (ASTM International, West Conshohocken, PA,
2011
).
16.
D. M.
Caughey
and
R. E.
Thomas
,
Proc. IEEE
55
,
2192
(
1967
).
17.
T. T.
Mnatsakanov
,
M. E.
Levinshtein
,
L. I.
Pomortseva
,
S. N.
Yurkov
,
G. S.
Simin
, and
M. A.
Khan
,
Solid-State Electron.
47
,
111
(
2003
).
18.
I.
Halidou
,
Z.
Benzarti
,
Z.
Chine
,
T.
Boufaden
, and
B. E.
Jani
,
Microelectron. J.
32
,
137
(
2001
).
19.
A.
Dadgar
,
J.
Bläsing
,
A.
Diez
, and
A.
Krost
,
Appl. Phys. Express
4
,
011001
(
2011
).
20.
K.
Seeger
,
Semiconductor Physics: An introduction
(
Springer
,
Vienna
,
2004
).
21.
E. C. H.
Kyle
,
S. W.
Kaun
,
P. G.
Burke
,
F.
Wu
,
Y. R.
Wu
, and
J. S.
Speck
,
J. Appl. Phys.
115
,
193702
(
2014
).
22.
H. M.
Ng
,
D.
Doppalapudi
,
T. D.
Moustakas
,
N. G.
Weimann
, and
L. F.
Eastman
,
Appl. Phys. Lett.
73
,
821
(
1998
).
23.
D. C.
Look
and
J. R.
Sizelove
,
Appl. Phys. Lett.
79
,
1133
(
2001
).
24.
D.
Huang
,
F.
Yun
,
M. A.
Reshchikov
,
D.
Wang
,
H.
Morkoç
,
D. L.
Rode
,
L. A.
Farina
,
Ç.
Kurdak
,
K. T.
Tsen
,
S. S.
Park
, and
K. Y.
Lee
,
Solid-State Electron.
45
,
711
(
2001
).
25.
D. C.
Look
,
C. E.
Stutz
,
R. J.
Molnar
,
K.
Saarinen
, and
Z.
Liliental-Weber
,
Solid State Commun.
117
,
571
(
2001
).