Polarization Effects of GaN and AlGaN:
Polarization Bound Charge, Band Bending,
and Electronic Surface States
Brianna S. Eller, Jialing Yang & Robert
J. Nemanich

Journal of Electronic Materials
ISSN 0361-5235
Journal of Elec Materi
DOI 10.1007/s11664-014-3383-z

1 23

Your article is protected by copyright and all
rights are held exclusively by The Minerals,
Metals & Materials Society. This e-offprint is
for personal use only and shall not be selfarchived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23

Author's personal copy
Journal of ELECTRONIC MATERIALS

DOI: 10.1007/s11664-014-3383-z
Ó 2014 The Minerals, Metals & Materials Society

Polarization Effects of GaN and AlGaN: Polarization Bound
Charge, Band Bending, and Electronic Surface States
BRIANNA S. ELLER,1 JIALING YANG,1 and ROBERT J. NEMANICH1,2
1.—Department of Physics, Arizona State University, Tempe, AZ 85287-1504, USA. 2.—e-mail:
robert.nemanich@asu.edu

GaN-based devices are currently limited by reliability issues such as gate
leakage and current collapse, where the mechanisms responsible for degradation are closely related to the electronic surface state configuration.
Therefore, understanding the electronic surface state configuration of GaNbased materials will help improve device performance. Since GaN has an
inherent polarization, these materials are also subject to a bound polarization
charge, which influences the electronic state configuration. In this study, the
surface band bending of N-face GaN, Ga-face GaN, and Ga-face AlGaN was
measured with x-ray photoemission spectroscopy after various cleaning steps
to investigate the effects of the polarization. Despite the different surface
bound charge on these materials, similar band bending was observed
regardless of the magnitude or direction of the charge. Specifically, the band
bending varied from 0.1 eV to 0.9 eV on these samples, which supported the
models of a Fermi level pinning state at 0.4 eV to 0.8 eV below the conduction band. Based on available literature, we suggest this pinning state is
indirectly evident of a nitrogen vacancy or gallium-dangling bond.
Key words: GaN, AlGaN, polarization, surface states, band bending

INTRODUCTION
Mitigating multi-lateral ecological and environmental concerns will define next-generation technology. In particular, improvement in power
electronic technologies ensures progress towards
this goal. GaN-based semiconductors thus remain
promising candidates, where GaN has several
advantages over competing semiconductors in power applications—e.g., Si, SiC, and GaAs—due to
superlative material properties.1 Specifically, GaN
is characterized by high power per unit width,
which allows for smaller devices, easier manufacturing, and higher impedance. This characteristic
also facilitates system matching that may be difficult with other materials such as GaAs. Moreover,
the high breakdown field of GaN supports higher
operating voltages, reduced voltage conversion,
decreased power requirements, and simpler cooling.

(Received June 9, 2014; accepted August 16, 2014)

Consequently, many GaN-based devices have
demonstrated superior performance.2 However,
despite the success of GaN-based devices, there are
three issues that require resolution before GaN
technology can replace existing Si technology: (1)
the development of high-quality growth methods of
single-crystal epitaxial GaN, (2) the selective formation of regions, and (3) the minimization of electronic states at the gate dielectric and GaN or
AlGaN interface.3 In this article, we address the last
one. More specifically, this work investigates the
effects of the polarization bound charge associated
with III–V nitrides on surface states in an attempt
to reveal relevant pinning states. In particular, this
work focuses on oxygen-terminated surfaces, which
are more representative of the conditions at the
device interface. Results of these oxygen-terminated
surfaces show band bending is independent of the
magnitude or direction of the polarization; these
similarities indicate a likely pinning surface state
located 0.4 eV to 0.8 eV below the conduction band
that can accommodate both positive and negative

Author's personal copy
Eller, Yang, and Nemanich

The spontaneous polarization, on the other hand, is
large for GaN and AlN; calculations using the Berryphase approach and local density14–16 or generalized
gradient approximations15,17 determine the spontaneous polarization is 0.029 C/m2 and 0.081 C/m2
for wurtzite GaN and AlN, respectively. This calculation assumes the respective (0001) Ga- and Al-face,
suggesting the spontaneous polarization is directed
towards the N-face. In addition, the magnitude of
the polarization increases with aluminum content,
as the spontaneous polarization is sensitive to
structural parameters. Therefore, the longer anioncation bond length along the (0001) axis of AlN corresponds to an increase in magnitude along the
c-axis of the wurtzite structure.18 The polarization of
AlxGa(1x)N can thus be determined by linear interpolation as displayed in Table I.
This polarization charge gives rise to a bound
surface charge,
*

rb ¼ P  c^:

Fig. 1. (Color online) Crystal structure, spontaneous polarization
fields (PSP), and piezoelectric polarization fields (PPE) for GaN (top)
and AlxGa(1x)N (bottom). Reprinted from Yu et al.18 Ó1999,
American Vacuum Society.

charge. Based on available literature, we suggest
the relevant pinning state is likely related to the
nitrogen vacancy or gallium-dangling bond.
Unlike Si, GaN and other wurtzite, III–V nitrides
*
are characterized by a macroscopic polarization, P.
This polarization arises
from the material properties
*
of the nitrides, where P is the sum of the spontaneous
*
polarization inherent to the equilibrium lattice, PSP ,
and
the piezoelectric polarization created by strain,
*
PPE . (See Fig. 1.) Using ab initio calculations and
material constants, the polarization along the c-axis
as induced by the piezoelectric effect is


*
a  ao
C13
e31 
e33 c^;
PPE ¼ 2
(1)
ao
C33
where C13 and C33 are elastic constants, e31 and e33
are piezoelectric coefficients, and ao and a are lattice
constants.4–12 For relaxed GaN and AlN, the piezoelectric polarization is negligible.13 (However, this
component of the polarization depends on the strain
of the crystal and may vary with growth method,
substrate material, or temperature).

(2)

There is thus a negative bound charge of
1.81 9 1013 charges/cm2 and 5.06 9 1013 charges/
cm2 for the Ga- and Al-face of GaN and AlN crystals,
respectively. Consequently, an equivalent positive
bound polarization charge exists on the N-face of GaN
and AlN. Since the internal electric field of a widebandgap semiconductor is zero or near zero, the system adjusts to satisfy surface conditions of near
charge neutrality. Therefore, the intrinsic material
properties give rise to a distribution of inherent
electronic states. The nature and distribution of
the compensation charge affect the internal electric field of the materials and ultimately device
performance.
This phenomenon is better understood in terms of
surface band bending, which is directly related to
the space charge region. More specifically, compensation charge in semiconductors, can take two
forms: (1) the formation of an internal space-charge
layer that consists of ionized donors and defects
near the surface, or (2) external charged surface or
interface states. These internal and external
screening mechanisms are inversely related as
shown in Fig. 2, where the larger the compensation
from the internal space-charge layer (and thus the
smaller the net concentration of surface states), the
larger the band bending. The band bending is thus
calculated from the density of internal screening
charge:
2
qNss
;
(3)
Us ¼ 
2eeo Nd
where q is the charge of an electron, e is the relative
permittivity, eo is the permittivity of free space, Nd
is the doping density, and Nss is the net surface
charge in charges/cm2. Assuming a doping density
of 1017 charges/cm3 and a net polarization charge of

Author's personal copy
Polarization Effects of GaN and AlGaN: Polarization Bound Charge, Band Bending, and Electronic
Surface States
Table I. Band gap, polarization, and corresponding polarization bound charge for GaN, AlN, and AlxGa(12x)N

Band gap (eV)
Spontaneous polarization (C/m2)
Polarization bound charge (1013 charges/cm2)

GaN

AlN

AlxGa(12x)N

3.4
0.029
1.81

6.2
0.081
5.06

3.4 (1  x) + 6.2x
0.029 (1  x) 0.081x
1.81 (1  x) + 5.06x

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2Us eeo Nd
:
Nss ¼ 
q

(4)

This equation suggests a 0.1 eV change in band
bending corresponds to a 3.2 9 1011 charges/cm2
change in the concentration of surface states.
EXPERIMENT

Fig. 2. (Color online) Theoretical (a) and experimental (b) band
bending schematic for Ga- and N-face GaN. Both surfaces are
screened by 1013 charges/cm2. (the position of the ionized donors
and electrons in the material corresponds to their physical position
rather than their energy level within the band gap.) Reprinted from
Eller et al.2 Ó2013, American Vacuum Society.

1.81 9 1013 charges/cm2 for GaN, this calculation
suggests a surface potential of 420 V, which corresponds to 420 eV of upwards band bending and an
average electric field of 200 MV/m at the surface of
GaN. In equilibrium, this large internal field results
in inversion or accumulation, and the band bending
is thus limited to approximately the band gap of the
material, 3.4 eV. (See Fig. 2a.)19,20 Therefore, ionized donors cannot be solely responsible for the
compensation of the polarization bound charge.
Experimental band bending measurements indicate the band bending is well below the band gap.
(See Fig. 2b.). In fact, most experimental band
bending experiments for n-type Ga-face GaN typically report measurements between 0.3 eV and
1.5 eV.21–23 These measurements can then determine the concentration of charged surface states
from the band bending, where

In this study, we determine the concentration of
surface states from the experimentally measured
band bending of several different sample surfaces,
including the Ga-face of GaN and Al0.25Ga0.75N
films as well as the Ga- and N-faces of freestanding GaN. These various sample surfaces enable us to examine different surfaces associated
with several polarization bound charge conditions.
Freestanding wafers were 450-lm-thick, n-type,
as-grown via hydride vapor phase epitaxy (HVPE)
purchased from READE Advanced Materials with
a Si doping density of 8 9 1017 cm3; this doping
density determined the position of the Fermi level
to be 0.1 eV lower than the conduction band
minimum. Additional n-type Ga-face epitaxial
GaN wafers were also used. The samples were
5 ± 1 lm thick, as grown by HVPE on sapphire
substrates purchased from READE Advanced
Materials. The doping density was still
1017 cm3, which established a similar Fermi
level position. To investigate the effects of a larger
concentration of surface bound charge, Ga-face
AlGaN was also used with 25% aluminum content
as purchased from NTT Advanced Technology.
Al0.25Ga0.75N samples were 50 nm thick, as
deposited on Si substrates with a doping density
of 1017 cm3. This doping density ensured a
similar Fermi level position as the other samples,
0.1 eV below the conduction band.
As-received wafers were cleaned ex situ via sonication in acetone, methanol, and NH4OH for 10 min
each. Samples were then rinsed in DI water for
1 min and dried with nitrogen. After the chemical
cleaning, samples were loaded into an ultra high
vacuum system with base pressure of 4 9 1010
Torr. The inclusive UHV system allowed for in-situ
cleaning, which reduced the oxygen coverage using
NH3 plasma and additional NH3 gas annealing at
680°C for 15 min each. The plasma was operated at
100 W with a constant gas flow of 90 sccm and

Author's personal copy
Eller, Yang, and Nemanich

pressure of 60 mtorr. Characterization was subsequently conducted using in-situ XPS.
More specifically, XPS spectra were used to
determine the stoichiometric ratios and surface
band bending. These spectra were obtained at a
base pressure of 8 9 1010 Torr. Mg Ka
(=1253.6 eV) x-ray radiation was used as a radiation
source, except when scanning the C 1s peak; the Ga
LMM Auger lines and C 1s peak overlap, and,
therefore, Al Ka (=1486.6 eV) x-rays were used. The
non-monochromatic x-ray source used a 4.4 A
filament current, 16 mA emission current, and
13 kV accelerating voltage. Survey scans were
repeated 30–80 times with a pass energy of 20 eV.
The spectra were dispersed with a Fisons Clam II
hemispherical analyzer at a resolution of 1.0 eV.
Through curve fitting of the core level peaks, the
peak positions could be resolved to ±0.1 eV. These
measurements included adjustments according to a
calibration using a gold foil; typical corrections were
0.1 and 0.2 eV for respective Mg and Al radiation
sources. XPS measurements were also used to
determine the concentration and the atomic ratio of
constituents near the surface; C 1s and O 1s spectra
indicated the effectiveness of the cleaning process,
while Ga 3d and N 1s spectra indicated the stoichiometric ratio of the GaN at the surface.
RESULTS
The oxygen coverage was defined as the number
of absorbed oxygen atoms per Ga (Al) or N atoms at
the c plane surface, where one oxygen atom per
surface lattice site referred to a single monolayer
(ML) of coverage. This ratio was calculated by the
following24:

 1


IO SGa X
ndGaN
Ho ¼
exp
;
(5a)
SO IGa n¼0
kGa Cos½/
where IGa and IO were the integrated intensities of
the respective Ga 3d and O 1s peaks, SO and SGa
were the atomic sensitivity factors for respective O
1s and Ga 3d (0.66 and 0.31),19 kGa was the inelastic
mean free path of Ga 3d electrons with kinetic
energies 1200 eV (24 Å),25 / was the angle
between the normal direction and the analyzer
(20°), and d was the distance between two Ga planes
(2.6 Å). For Al0.25Ga0.25N, this calculation was
modified, where one in every four Ga atoms was
replaced with an Al atom, giving

IO
Ho ¼
SO
2
6
6
4

SGa

,

P1

n¼0

3
IGa
exp

h

ndGAlaN
kGa Cos½/

IAl

iþ
SAl

P1

n¼0

exp



ndGAlaN
kAl Cos½/

7
7
5;

(5b)
where IAl was the integrated intensity for the Al 2p
peak, SAl was the atomic sensitivity factor for Al 2p
(0.185), kAl was the approximate average inelastic
mean free path of Al 2p electrons with kinetic
energies 1400 eV (24 Å),26 and dAlGaN was the
distance between two Ga/Al planes (2.6 Å). The
oxygen coverage on each sample after the different
cleaning states is summarized in Table II, and the
XPS core levels are shown in Fig. 3. In general, the
AlGaN surfaces are more resistant to oxygen
reduction;27–29 this is expected given the difficulty of
breaking Al–O bonds during the cleaning process.
Moreover, the NH3 cleaning reduced carbon below
the XPS detection limit.
The atomic concentration ratios of the samples
were also determined after the different processing
steps from relative XPS intensities, as shown in
Fig. 4 and summarized in Table III.
Band bending (BB) of oxygen-terminated GaN
and AlGaN was also calculated from the position of
the Ga 3d core level and inherent material properties, as shown in Fig. 5:
BB ¼ ðEV  ECL ÞGaN þEg ECL;XPS þ EC ;

(6)

where EC was the position of the conduction band
with respect to the Fermi level as determined by
the doping density (0.1 eV), Eg was the
band gap of the material (Eg,GaN = 3.4 eV and
Eg,AlGaN = 4.0 eV30), and ECl, XPS was the position of
the Ga 3d core level for GaN and Al 2p core level for
AlGaN. It is worth noting here that this analysis
focused on the Ga 3d peak, where the core
level intensity and position were less sensitive to the
overlayer, thereby providing more reliable band bending measurements. Additionally, (ECL  EV)GaN represented the binding energy difference of the core
level of Ga 3d in GaN with respect to the VBM.
According to electronic state studies of GaN,31–33
the difference between the Ga 3d core level and the
valence band maximum was 17.7 eV–17.8 eV. In

Table II. Oxygen coverage (in ML) on N-face GaN, Ga-face GaN, and Ga-face AlGaN as determined after the
various cleaning steps as given by XPS
Oxygen coverage
As received
Ex situ cleaning
In situ cleaning

N-face GaN

Ga-face GaN

Ga-face Al0.25Ga0.75N

4.5
2.6
1.1

5.1
3.1
1.1

2.6
2.2
1.8

Author's personal copy
Polarization Effects of GaN and AlGaN: Polarization Bound Charge, Band Bending, and Electronic
Surface States

(a)

(b)

inversely related to oxygen coverage as shown in
Fig. 6a and positively correlated to nitrogen content
as shown in Fig. 6b. In addition, while there was a
small disparity in the band bending on the
N- and Ga-face—which may be the result of different interface dipoles36—all three surfaces were
characterized by similar band bending after cleaning, ranging from 0.1 eV to 0.9 eV. The magnitude
of this band bending suggests the presence of an
electronic surface state responsible for Fermi level
pinning 0.1 eV to 0.9 eV below the conduction
band edge. Moreover, the similarity of the band
bending regardless of the polarization bound charge
suggests this pinning state can accommodate charge
transfer to compensate both positive and negative
charge. This pinning state is thus as modeled in
Fig. 7.
DISCUSSION

(c)

Fig. 3. (Color online) O 1s peak for N-face GaN (a), Ga-face GaN
(b), and Ga-face Al0.25Ga0.75N (c) as received (i), after ex-situ
cleaning (ii), and after in-situ cleaning (iii). Note that core levels were
shifted to the corresponding flat band position, allowing direct comparison of the oxygen states.

this study, 17.8 eV was assumed for (ECL  EV)GaN.
Similar electronic-state studies of Al0.25Ga0.75N
indicated 17.5 eV34 and 71.5 eV35 were the respective differences between the Ga 3d and Al 2p core
levels and the valence band maximum, which were
used to determine the band bending at the AlGaN
surface. Please note that, while the Ga and Al core
level peaks include a component due to Ga–O and
Al–O bonding, this component does not affect the
peak position. The experimental band bending of
each sample after various stages of cleaning is summarized in Table IV, and the corresponding external
compensation charge is summarized in Table V.
These results demonstrated there was a relationship between film content and band bending. In
particular, it was revealed that band bending was

In order to identify the microscopic nature of this
pinning state, we have considered surface reconstructions, defect states, and adsorbates; however,
direct comparison between experiment and theory is
often not a straightforward means of identifying the
microscopic nature of a surface.
To date, most related theoretical research has
focused on the nature of pinning states on Ga-face
GaN as dependent on surface reconstructions. For
example, at the Ga-face surface, several (2 9 2)
surface reconstructions are thought to be the most
stable, including the gallium adatom, nitrogen
adatom, and gallium vacancy structures.37,38 Nevertheless, most experimental studies do not explicitly agree with these results, where (1 9 1), (2 9 2),
(2 9 4), (5 9 5), and (6 9 4) reconstructions have all
been observed.39–41 In other words, although the
existence of Fermi level pinning surface states has
been well documented in GaN-based materials,42–44
a microscopic understanding of these states and
their dependence on polarity is still unclear. The
discrepancy between experimental and theoretical
values is likely the result of two factors. First, a
detailed understanding of the theoretical and
experimental band structure and related density of
states of experimentally relevant GaN surface
reconstructions is lacking. Second, theoretical
studies generally assume clean, ordered surface
conditions when such surfaces rarely exist; actual
GaN surfaces are commonly oxidized or metal rich.
Recent studies have attempted to rectify this disparity, where Himmerlich et al.45 described experimental and theoretical surface studies to determine
the microscopic surface conditions.
However, these perfectly ordered surface reconstructions are not sufficient to explain observed
Fermi level pinning; it is thus more likely these
results were caused by a defect state. Given the
magnitude of the experimentally observed band
bending and the proportional relationship between
the band bending and nitrogen concentrations, it

Author's personal copy
Eller, Yang, and Nemanich

(a)

(b)

(c)

Fig. 4. (Color online) XPS results for N-face GaN (a), Ga-face GaN (b), and Ga-face Al0.25Ga0.75N (c) as received (i), after ex situ cleaning (ii),
and after in-situ cleaning (iii).

Table III. Atomic ratio of N/Ga(Al) as determined from Al 2p, Ga 3d, and N 1s core level intensities, respective
atomic sensitivity factors of 0.19, 0.31, and 0.42, and effective attenuation lengths as determined from the
NIST database58
Atomic ratios

N-face GaN

Ga-face GaN

Ga-face Al0.25Ga0.75N

As received
Ex situ cleaning
In situ cleaning

0.92
1.02
2.03

0.74
0.78
1.42

1.17
0.84
1.28

has been proposed that the observed pinning state
behavior is related to a nitrogen vacancy or gallium
dangling bond. These states have been theoretically

and experimentally determined to be located
0.37 eV46 or 0.5 eV–0.7 eV47 below the conduction
band edge, which agrees with the experimental

Author's personal copy
Polarization Effects of GaN and AlGaN: Polarization Bound Charge, Band Bending, and Electronic
Surface States

findings in this research as well. Furthermore,
these defect states have been linked back to device
behavior, where Hashizume and Hasegawa46

Fig. 5. (Color online) Surface bend bending measurements (bottom)
were determined from the position the Ga 3d core level as determined from XPS (top) by the given equation (middle).

demonstrated that passivating GaN-based samples
with nitrogen improves device reliability.
While this evidence is suggestive, there are several other factors that may affect the band bending
on these samples as well. In particular, we cannot
overlook the effects of oxygen coverage. To date, a
number of studies have investigated the effects of
various ex-situ and in-situ treatments, including
HF, NH4OH, annealing, as well as N2 and H2
plasma.2,48–54 These studies demonstrate the difficulty in removing native oxygen, likely GaO2 or
Ga2O3,51 from GaN surfaces without damaging the
surface reconstruction, as it leaves nitrogen vacancies and/or gallium dangling bonds. Consequently,
the cleaning method used should passivate these
states. This was the case in this experiment and
may explain the reciprocal nature between the
oxygen coverage and nitrogen content as shown in
Fig. 6. On the other hand, it may be possible that
the oxygen coverage introduces additional charge
states. For example, there is evidence that dissociation of H2O and O2 on GaN surfaces introduces O
and OH groups on the surface.53,54 Removal of this
negative charge would thus describe the increase in
band bending on the Ga-face as well. It is thus unclear whether the oxygen coverage contributes to
the band bending as well, especially on the N face.
Moreover, the plasma used in the cleaning may also
influence the band bending. While NH3 plasma
annealing has been shown to reduce oxygen effectively, the plasma may also introduce interstitial
nitrogen. Therefore, an additional component of the
band bending may not necessarily be attributable to
surface reconstructions or point defects.
In reality, these explanations are not mutually
exclusive. It is likely that nitrogen interstitials,
oxygen adsorbates, and vacancies influence the
observed band bending at different cleaning stages.
Furthermore, additional research has provided evidence for a fixed charge at the interface,55–57 which
may affect the Fermi level pinning position;

Table IV. Band bending (in eV) for N-face GaN, Ga-face GaN, and Ga-face AlGaN as determined after the
various cleaning steps as given by XPS
Band bending

N-face GaN

Ga-face GaN

Ga-face Al0.25Ga0.75N

As received
Ex situ cleaning
In situ cleaning

0.1
0.1
0.6

0.2
0.4
0.9

0.2
0.2
0.4

Table V. Concentration of external compensation charge (1013 charges/cm2) on N-face GaN, Ga-face GaN, and
Ga-face AlGaN as determined after the various cleaning steps as given by XPS
Net external compensation
As received
Ex situ cleaning
In situ cleaning

N-face GaN

Ga-face GaN

Ga-face Al0.25Ga0.75N

1.8
1.8
2.0

+1.8
+1.7
+1.5

+2.6
+2.6
+2.5

Author's personal copy
Eller, Yang, and Nemanich

(a)

(b)

1.0

N-face GaN
Ga-face GaN
Ga-face AlGaN

Band Bending (eV)

0.8
0.6
0.4
0.2
0.0
-0.2

1

2

3

4

5

Oxygen Coverage (ML)

6 0.0

0.5

1.0

1.5

2.0

N:Ga Concentration Ratio

Fig. 6. (Color online) Correlation between the surface band bending and film content, i.e. oxygen coverage (a) and relative nitrogen surface
concentration (b) for N-face GaN, Ga-face GaN, and Ga-face Al0.25Ga0.75N.

face states and affected the band bending. Future
research will continue to investigate the effects of
polarity with more emphasis on surface bonding to
better understand the microscopic nature of these
states and their effects on the band bending at the
surface of GaN-based materials as well as correlate
CV and photoemission measurements to resolve
potential inconsistencies.
ACKNOWLEDGEMENT
Fig. 7. Suggested charge transfer model for pinning state on N- and
Ga-face GaN and AlGaN.

This research was supported by the Office of
Naval Research through the DEFINE MURI program, N00014-10-1-0937.

however, without additional CV measurements, it is
difficult to resolve the discrepancy between the
electrical measurements and photoemission measurements.

REFERENCES

CONCLUSION
In summary, we have used XPS to determine the
net concentration of surface states on N-face GaN,
Ga-face GaN, and Ga-face Al0.25Ga0.75N. The results
demonstrated a similar band bending regardless of
the magnitude or direction of the polarization bound
charge. (There is a disparity between the N- and
Ga-face; however, this difference is likely the result
of an interface dipole, which will be explored further.) Specifically, the band bending varies from
0.1 eV to 0.9 eV on these samples, which indicated
that there was likely a Fermi pinning state 0.4 eV
to 0.8 eV below the conduction band minimum. We
suggested the observed pinning state was related to
the nitrogen-related defect or gallium dangling
bond, as supported by experimental and literary
results—though the state was not directly observed.
This proposal is consistent with other experimental
and theoretical research; however, there were
likely additional mechanisms that influence surface
states. In particular, it is still unclear how oxygen
adsorbates impact the electronic states configuration. Moreover, the plasma cleaning process likely
introduced additional nitrogen interstitial or sur-

1. U.K. Mishra, P. Parikh, and Y.F. Wu, Proc. IEEE 90, 1022
(2002).
2. B.S. Eller, J. Yang, and R.J. Nemanich, J. Vac. Sci. Technol.
A 31, 050807 (2013).
3. Y. Niiyama, S. Ootomo, J. Li, T. Nomura, S. Kato, and T.P.
Chow, Semicond. Sci. Technol. 25, 125006 (2010).
4. V.A. Savastenko and A.U. Sheleg, Phys. Status Solidi A 48,
K135 (1978).
5. Y. Takagi, M. Ahart, T. Azuhata, T. Sota, K. Suzuki, and S.
Nakamura, Physica B 219–220, 547 (1996).
6. A. Polian, M. Grimsditch, and I. Grezegory, J. Appl. Phys.
79, 3343 (1996).
7. R.B. Schwartz, K. Khachaturyan, and E.R. Webber, Appl.
Phys. Lett. 70, 1122 (1997).
8. C. Deger, E. Born, H. Angerer, O. Ambacher, M. Stutzmann,
J. Hornsteiner, E. Riha, and G. Fischerauer, Appl. Phys.
Lett. 72, 2400 (1998).
9. K. Tsubouchi and N. Mikoshiba, IEEE Trans. Sonics
Ultrason. 32, 634 (1985).
10. L.E. McNeil, M. Grimsditch, and R.H. French, J. Am. Ceram. Soc. 76, 1132 (1993).
11. K. Kim, W.R.L. Lambrecht, and B. Segall, Phys. Rev. B 53,
16310 (1996).
12. A.F. Wright, J. Appl. Phys. 82, 2833 (1997).
13. O. Ambacher, J. Smart, J.R. Shealy, N.G. Weimann, K. Chu,
M. Murphy, W.J. Schaff, L.F. Eastman, R. Dimitrov, L.
Wittmer, M. Stutzmann, W. Rieger, and J. Hilsenbeck,
J. Appl. Phys. 85, 3222 (1999).
14. F. Bernardini, V. Fiorentini, and D. Vanderbilt, Phys. Rev. B
56, R10024 (1997).
15. F. Bernardini, V. Fiorentini, and D. Vanderbilt, Phys. Rev. B
63, 193201 (2001).
16. Y. Duan, J. Li, S.-S. Li, and J.-B. Xia, J. Appl. Phys. 103,
023705 (2008).

Author's personal copy
Polarization Effects of GaN and AlGaN: Polarization Bound Charge, Band Bending, and Electronic
Surface States
17. K. Shimada, A. Zenpuku, K. Fukiwara, K. Hazu, S.F.
Chichibu, M. Hata, H. Sazawa, T. Takada, and T. Sota,
J. Appl. Phys. 110, 074114 (2011).
18. E.T. Yu, X.Z. Dang, P.M. Asbeck, S.S. Lau, and G.J.
Sullivan, J. Vac. Sci. Technol. B 17, 1742 (1999).
19. J. Yang, B.S. Eller, C. Zhu, C. England, and R.J. Nemanich,
J. Appl. Phys. 112, 053710 (2012).
20. H.W. Jang, J.-H. Lee, and J.-L. Lee, Appl. Phys. Lett. 80,
3955 (2002).
21. M. Hong, K.A. Anselm, J. Kwo, H.M. Ng, J.N. Baillargeon,
A.R. Kortan, J.P. Mannaerts, A.Y. Cho, C.M. Lee, J.I. Chyi,
and T.S. Lay, J. Vac. Sci. Technol. B 18, 1453 (2000).
22. Y.Q. Wu, T. Shen, P.D. Ye, and G.D. Wilk, Appl. Phys. Lett.
90, 143504 (2007).
23. W.J. Mecouch, B.P. Wagner, Z.J. Reitmeier, R.F. Davis, C.
Pandarinath, B.J. Rodriguez, and R.J. Nemanich, J. Vac.
Sci. Technol. A 23, 72 (2005).
24. V.M. Bermudez, J. Appl. Phys. 80, 1190 (1996).
25. M. Krawczyk, L. Zommer, A. Jablonski, I. Grzegory, and M.
Bockowski, Surf. Sci. 566–568, 1234 (2004).
26. H.-H. Wu, Ph.D. Thesis. Ohio State University, 2011.
27. B. Brennan, X. Qin, H. Dong, J. Kim, and R.M. Wallace,
Appl. Phys. Lett. 101, 211604 (2012).
28. X. Qin, H. Dong, B. Brennan, A. Azacatl, J. Kim, and R.M.
Wallace, Appl. Phys. Lett. 103, 221604 (2013).
29. X. Qin, B. Brennan, H. Dong, J. Kim, C.L. Hinkle, and R.M.
Wallace, J. Appl. Phys. 113, 244102 (2013).
30. R.R. Pela, C. Caetano, M. Marques, L.G. Ferreira, J.
Furthmuller, and L.K. Teles, Appl. Phys. Lett. 98, 151907
(2011).
31. T.E. Cook Jr, C.C. Fulton, W.J. Mecouch, R.F. Davis, G.
Lucovsky, and R.J. Nemanich, J. Appl. Phys. 94, 7155
(2003).
32. J.R. Waldrop and R.W. Grant, Appl. Phys. Lett. 68, 2879
(1996).
33. J. Hedman and N. Mårtensson, Phys. Scr. 22, 176 (1980).
34. Y.L. Chiou and C.T. Lee, IEEE Trans. Electron Devices 58,
3869 (2011).
35. M.H.S. Owen, M.A. Bhuiyan, Q. Zhou, Z. Zhang, J.S. Pan,
and Y.-C. Yeo, Appl. Phys. Lett. 104, 091605 (2014).
36. J. Yang, B.S. Eller, and R.J. Nemanich, J. Appl. Phys.
(submitted).
37. J.E. Northrup, J. Neugebauer, R.M. Feenstra, and A.R.
Smith, Phys. Rev. B 61, 9932 (2000).
38. D. Segev and C.G. Van de Walle, Surf. Sci. 601, L15 (2007).

39. A.R. Smith, R.M. Feenstra, D.W. Greve, M.-S. Shin, M.
Skowronski, J. Neugebauer, and J.E. Northrup, Surf. Sci.
423, 70 (1999).
40. S.W. King, C. Ronning, R.F. Davis, M.C. Benjamin, and R.J.
Nemanich, J. Appl. Phys. 84, 2086 (1998).
41. S. Vézian, F. Semond, J. Massies, D.W. Bullock, Z. Ding,
and P.M. Thibado, Surf. Sci. 541, 242 (2003).
42. A.R. Smith, R.M. Feenstra, D.W. Greve, M.S. Shin, M.
Skowronski, J. Neugebauer, and J.E. Northrup, J. Vac. Sci.
Technol. B 16, 2242 (1998).
43. C.G. Van de Walle and D. Segev, J. Appl. Phys. 101, 081704
(2007).
44. L. Ivanova, S. Borisova, H. Eisele, M. Dahne, A. Laubsch,
and Ph Eben, Appl. Phys. Lett. 93, 192110 (2008).
45. M. Himmerlich, L. Lymperakis, R. Gutt, P. Lorenz, J.
Neugebauer, and S. Krischok, Phys. Rev. B 88, 125304
(2013).
46. T. Hashizume and H. Hasegawa, Appl. Surf. Sci. 234, 387
(2004).
47. D. Segev and C.G. Van de Walle, Europhys. Lett. 76, 305
(2006).
48. S.W. King, J.P. Barnak, M.D. Bremser, K.M. Tracy, C.
Ronning, R.F. Davis, and R.J. Nemanich, J. Appl. Phys. 84,
5248 (1998).
49. T. Hashizume, S. Ootomo, T. Inagaki, and H. Hasegawa,
J. Vac. Sci. Technol. B 21, 1828 (2003).
50. R.D. Long and P.C. McIntyre, Mater 5, 1297 (2012).
51. C.L. Hu, J.Q. Li, Y.F. Zhang, X.L. Hu, N.X. Lu, and Y. Chen,
Chem. Phys. Lett. 424, 273 (2006).
52. R. Meunier, A. Torres, and M. Charles, ECS Trans. 50, 451
(2013).
53. J. Elsner, R. Gutierrez, B. Hourahine, R. Jones, M.
Haugk, and Th Frauenheim, Solid State Comm. 108, 953
(1998).
54. V.M. Bermudez and J.P. Long, Surf. Sci. 450, 98 (2000).
55. M. Esposto, S. Krishnamoorthy, D.N. Nath, S. Bajaj, T.H. Hung, and S. Rajan, Appl. Phys. Lett. 99, 1233503
(2011).
56. S. Ganguly, J. Verma, G. Li, T. Zimmermann, H. Xing, and
D. Jena, Appl. Phys. Lett. 99, 193504 (2011).
57. J. Son, V. Chobpattana, B.M. McSkimming, and
S. Stemmer, Appl. Phys. Lett. 101, 102905 (2012).
58. NIST, Electron Effective-Attenuation-Length Database,
SRD-82, Version 1.1 (Gaitherburg: National Institute of
Standards and Technology, 2003).