Journal of Crystal Growth 223 (2001) 466–483

GaN decomposition in H

2

and N

2

at MOVPE temperatures

and pressures

D.D. Koleske*, A.E. Wickenden, R.L. Henry, J.C. Culbertson, M.E. Twigg

Laboratory for Advanced Material Synthesis, Code 6800, Electronics Science and Technology Division, Naval Research Laboratory,

Washington, DC 20375, USA

Received 31 July 2000; accepted 12 December 2000

Communicated by C.R. Abernathy

Abstract

GaN decomposition rates were measured in H

2

, N

2

, and mixed H

2

and N

2

flows for pressures and temperatures

typically encountered in metalorganic vapor phase epitaxy. The rates for GaN decomposition, Ga desorption, and Ga
droplet accumulation, were obtained from weight measurements before and after annealing the GaN films in a close-
spaced showerhead reactor. In H

2

at constant temperature, the GaN decomposition rate is enhanced when the reactor

pressure is greater than 100 Torr. Unlike H

2

, the decomposition rate in N

2

did not change as a function of pressure. The

enhanced GaN decomposition rate in H

2

is not due to an increase in the Ga desorption rate, which is constant vs.

pressure, but instead is due to H

2

dissociation on the surface followed by NH

3

formation and desorption. NH

3

formation is suggested by the cubic decrease in the GaN decomposition rate as N

2

is substituted for H

2

. The measured

activation energies, E

A

, for the GaN decomposition range from 0.34 to 3.62 eV and depend strongly on the annealing

conditions. By comparing measured and literature values of the E

A

, four distinct groupings of the E

A

are observed. The

four distinct groupings of the E

A

imply that there are possibly four different reactions which limit the GaN

decomposition rate. Connections between the GaN decomposition and improved GaN growth are discussed. This
includes a discussion of changes that occur in the nucleation layer during the ramp from low to high temperature, as
well as increases in GaN grain size as the growth pressure is increased. Published by Elsevier Science B.V.

PACS:

81.05.Ea; 81.15.Gh; 82.60.Cx; 81.05.D; 68.55; 61.16.B; 61.72

Keywords:

A1. Decomposition; A1. Desorption; A1. Grain size; A1. Growth; A3. Metalorganic vapor phase epitaxy; B1. Gallium nitride

1. Introduction

Despite the lack of a lattice matched substrate,

GaN of sufficient quality has been grown to
produce high brightness blue LEDs [1], lasers [2],
and microwave power devices [3–5]. Because of

these device successes, detailed understanding of
GaN growth is of great interest for further
material and device improvements. Studies of
GaN growth using molecular beam epitaxy
(MBE) have provided far greater insight into the
growth details because the growth species are
simpler and in situ diagnostics are available [6–11].
However, a similar level of understanding has not
been attained in metalorganic vapor phase epitaxy

*Corresponding author. Fax: [phone].

E-mail address:

[e-mail] (D.D. Koleske).

0022-0248/01/$ - see front matter Published by Elsevier Science B.V.
PII: S 0 0 2 2 - 0 2 4 8 ( 0 1 ) 0 0 6 1 7 - 0

(MOVPE) [12–15],

1

because of the more complex

chemical reactions and changes at the surface are
difficult to measure during growth.

2

Optimal

MOVPE growth conditions vary from reactor to
reactor, primarily reflecting the influence of
reactor design on the individual chemical reaction
rates [16]. In addition, some of the chemical
reactions, specifically NH

3

decomposition, have

been suggested to be catalytic, relying on specific
surface sites for full activation [17–19].

It is well established that GaN material quality

depends heavily on the nucleation layer [20] and
high temperature GaN layer [16] growth condi-
tions. Recently, we have reported the influence of
pressure, P, on GaN film quality [21,22]. Using
cross sectional TEM, the GaN grain size was
shown to increase as the growth P increased
[21,22]. Prior to this work we observed that GaN
decomposition also increases as the reactor P
increases [23–25], implying a correlation between
enhanced decomposition at high P and the larger
grained GaN films [21]. These studies [21–25]
illustrate a close link exists between GaN film
quality and the initial growth stages, specifically
the competition between GaN growth and decom-
position.

Although GaN decomposition has been exten-

sively studied, there are substantial differences in
the reported kinetic parameters and decomposi-
tion mechanisms. This is partly due to the
differences in P, temperature, T, and gas flows
used to study GaN decomposition. For example,
the onset of GaN decomposition has been
reported for T as low as 4008C [26] and as high
as 10708C [27]. Additionally, activation energies,
E

A

, as low as 0.4 eV [28] and as high as 3.93 eV [29]

have been reported. Literature values for E

A

and

the pre-exponential factor, A

0

, are listed in Table 1

[6,19,24,26,28,30–40]. The E

A

and A

0

are sepa-

rated into groups labeled A–D to reflect their
dependence on annealing conditions.

The principal aim of this paper is to demon-

strate how P; T , and gas flow (i.e. H

2

or N

2

)

change the GaN decomposition rates and mechan-
isms. The paper also unifies previous literature on
GaN decomposition kinetics, since many of the
previously measured kinetic values are corrobo-
rated in the present study by varying the annealing
conditions.

The paper is structured as follows: in Section 2,

results

from

previous

GaN

decomposition

studies are discussed along with proposed decom-
position mechanisms. In Section 3, the experi-
mental details of the weight loss measurements
and annealing conditions are presented. In Section
4, the experimental results on GaN decompo-
sition, Ga desorption, and Ga surface droplet
accumulation are presented. Details of the Ga
droplet coalescence are presented along with a
comparison to a theoretical droplet coalescence
model. Also in Section 4, the GaN decomposition
in mixed H

2

and N

2

flows will be presented,

along with an Arrhenius analysis of the GaN
decomposition kinetic parameters. In Section 5,
mechanisms for GaN decomposition under the
various experimental conditions are discussed.
This will include a discussion of the four different
groups of kinetic parameters listed in Table 1.
Also in Section 5, the implications of this
decomposition study to GaN growth are consid-
ered, including a discussion of how the GaN
nucleation layer changes during the high tempera-
ture anneal and how the growth conditions
affect GaN grain size. Section 6 contains conclu-
sions from this study on decomposition and
general implications for other GaN reactor con-
figurations.

2. Previous studies of GaN decomposition

Many early reports of GaN growth also

included observations and concern over GaN
decomposition, mostly because decomposition
competed with growth [41,42]. In 1932, GaN
synthesis and decomposition were reported by
Johnson, Parsons, and Crew, who noticed that
GaN decomposes when the temperature, T , was
greater than 8008C [41]. In later papers, GaN

1

For recent reviews of GaN growth, see Ref. [12].

2

While in situ reflectometry and optical pyrometery are

useful for determining surface roughness and growth rate, they
do not measure atomic level changes in the surface structure
like RHEED does.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

467

decomposition was found to initiate at lower T
(400–8008C) in H

2

[26,27,41–44] compared inert

gases (N

2

or Ar) or vacuum (>9008C) [26,27,30–

32,43–48].

Early on, several different mechanisms were

proposed to explain GaN decomposition. These
included decomposition into gaseous Ga and
nitrogen (Eq. (1)) [31], liquid Ga and nitrogen
(Eq. (2)) [26,31,32], and sublimation of GaN as a
diatomic or polymeric product (Eq. (3)) [31].

2GaN

ðsÞ ! 2GaðgÞ þ N

2

ðgÞ:

ð1Þ

2GaN

ðsÞ ! 2GaðlÞ þ N

2

ðgÞ ! 2GaðgÞ:

ð2Þ

GaN

ðsÞ ! GaNðgÞ or ½GaN

x

ðgÞ:

ð3Þ

Evidence for all three reactions has been observed
experimentally. For example, N

2

evolution has

been observed using mass spectroscopy [27,32]
along with GaN

+

and Ga

2

N

2

+

species [47,48].

Table 1
Measured kinetic parameters for GaN decomposition (groupings A–D) and Ga desorption (grouping E) gathered from the literature
and the present work. For each grouping label (i.e. A–E) the conditions for pressure and temperature are listed. The first column lists
the measurement technique used for measuring the kinetic parameters. The second column lists the pre-exponential factor, A

0

, in units

of cm

2

s

1

, while the third column lists the base 10 logarithm of A

0

. The fourth column lists the activation energy, E

A

, for each

measurement in eV. The last column lists the reference from the literature and the year the work was conducted in square brackets

Event and experimental conditions

A

0

(cm

2

s

1

)

Log

10

ðA

0

Þ (cm

2

s

1

)

E

A

(eV)

Ref. [Year]

(A) GaN decomposition, P476 Torr, T> 9258C,

N

2

formation and desorption limited

(1) Thermogravimetric

}

}

2.7

30 [1956]

(2) Thermogravimetric

4

 10

29

29.60

3.1

31 [1965]

(3) Mass spectroscopy in vacuum

5

 10

28

28.70

3.1

32 [1974]

(4) Mass spectroscopy in vacuum

1.2

 10

31

31.08

3.93

33 [1996]

(5) Ga flux in vacuum

5.0

 10

29

29.70

3.45

34 [1998]

(6) Reflectivity data in vacuum

5.1

 10

31

31.7

1.2

3.7

0.3

35 [1999]

(7) Microbalance at 760 Torr

1.4

 10

29

29.1

3.2

36 [2000]

(8) Weight loss in H

2

at 40 and 76 Torr

2.8

 10

29

29.4

1.8

3.4

0.2

This work

(9) Weight loss in N

2

at 76 and 150 Torr

1.2

 10

29

29.1

1.6

3.62

0.14

This work

(B) GaN decomposition, P476 Torr, T 59258C,

N

2

H

25x54

formation and desorption limited

(10) Surface photoadsorption N loss in H

2

}

}

0.91

37 [1998]

(11) Weight loss in H

2

at 40 and 76 Torr

8.7

 10

19

19.94

0.40

0.98

0.07

This work

(C) GaN decomposition, P5150 Torr, T 59008C,

NH

3

formation and desorption limited

(12) Redhead analysis of NH

3

TPD peak

}

}

1.6–2.0

19 [1999]

(13) Gravimetric in H

2

at 760 Torr

}

}

1.8

26 [1974]

(14) Reflectometry at 760 Torr

1

 10

25

25

1.87

28 [1999]

(15) Weight loss in H

2

at 150 and 250 Torr,

7.9

 10

24

24.9

1.3

1.7

0.2

This work

(D) GaN decomposition P5150 Torr, T > 9008C,

Ga diffusion limited

(16) Ga diffusion E

A

}

theory calculation

}

}

0.4

38 [1999]

(17) Reflectometry at 760 Torr

3

 10

17

17.5

0.38

28 [1999]

(18) Microbalance at 760 Torr

1

 10

15

15

0.44

36 [2000]

(19) Weight loss in H

2

at 150 and 250 Torr

3.5

 10

17

17.54

0.53

0.34

0.1

This work

(E) Ga desorption, in vacuum and at all pressures in N

2

, and H

2

(20) Desorption from liquid Ga

}

}

2.8

39 [1969]

(21) RHEED study of hexagonal GaN

}

}

2.76

40 [1996]

(22) RHEED study, cubic GaN

1.0

 10

28

28.0

0.26

2.69

0.05

6 [1996]

(23) Weight loss in H

2

at 40–250 Torr

6.5

 10

26

26.81

0.34

2.74

0.08

24 [1999]

(24) Weight loss in N

2

at 76 and 150 Torr

5.4

 10

25

25.73

1.5

2.69

0.4

24 [1999]

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

468

Occasionally, Ga droplets 20–30 mm in size have
been observed on the GaN surface [27,30–32,43],
indicating that the GaN decomposition rate, k

GaN

,

can exceed the Ga desorption rate, k

Ga

. Liquid Ga

and In droplets have been shown to catalyze GaN
decomposition [49–51].

As mentioned above, the onset T for GaN

decomposition is lower in H

2

compared to inert

environments (i.e. N

2

, Ar, and vacuum). Hydrogen

could assist decomposition by the reverse GaN
synthesis reaction, i.e. the reformation of NH

3

via

GaN

ðsÞ ! 3=2H

2

! GaðlÞ þ NH

3

ðgÞ

ð4Þ

To measure the NH

3

formation rate, Thurmond

and Logan cleverly titrated the furnace exhaust
during GaN decomposition in H

2

into a column of

HCl solution containing methyl red indicator [42].
The NH

3

reformation reaction is significant for

GaN growth because it removes N from the
surface and provides a reversible pathway for
bringing GaN synthesis closer to equilibrium.

Previously measured values of the activation

energy, E

A

, and the pre-exponential factor, A

0

, for

GaN decomposition are listed in Table 1. For H

2

pressure, P, less than 76 Torr, in N

2

, or in vacuum

(Table 1A) the E

A

is 2.7–3.93 eV. Surprisingly, this

is lower than the E

A

of 4.7 eV for GaAs decom-

position [52,53], despite the stronger bond strength
in GaN (4.1 eV) compared to GaAs (2.0 eV) [54].
When GaN films are annealed in H

2

at higher P

and lower T , the E

A

decreases (Table 1B–D). For

example, Morimoto [26] and Rebey et al. [28] have
measured E

A

of 1.8 eV and 1.87 eV, respectively.

In H

2

at higher P and higher T , two groups have

measured lower E

A

of 0.38 eV [28] and 0.44 eV

[36]. These lower E

A

for decomposition in H

2

suggests hydrogen aids in N removal from the
surface, possibly by NH

3

formation (i.e. Eq. (4)).

Unlike the GaN decomposition kinetics, the

kinetic parameters for Ga desorption have only a
weak dependence on pressure or gas flow [23,24].
The measured kinetic parameters listed in Table
1E are similar to the parameters for Ga desorption
from GaAs surfaces [55] and liquid Ga [39]. Good
agreement has been observed in the measured A

0

(ranging from 1

 10

28

to 6

 10

29

cm

2

s

1

) and

E

A

(ranging from 2.69 to 2.76 eV) [6,23,24,39,40].

The close agreement between the heat of forma-

tion for Ga (i.e. 2.83 eV) [56] and the measured E

A

suggests a simple Ga–Ga bond cleavage for the
rate-limiting step.

The N desorption kinetic parameters have been

measured in a RHEED study on cubic GaN
[6]. Both the measured E

A

(6.1 eV) and A

0

(2

 10

44

cm

2

s

1

) are large for N desorption [6].

Extrapolating the N desorption rate, k

N

, to

10208C gives a k

N

that is 1000 times k

Ga

[16].

3. Experimental details

The experimental procedure is described in Ref.

[23]. The GaN films used in these studies were
grown in a close-spaced showerhead (CSS) reactor
[3] and in a custom-designed vertical reactor [16].
Growth in both reactors have produced GaN films
with specular morphology and Si doped electron
mobilities>600 cm

2

V

1

s

1

[22,23].

The decomposition study was conducted in the

CSS reactor.

3

In the CSS reactor, the temperature,

T

, measurement and control consists of a W/Re

thermocouple in direct contact with the backside
of the susceptor. The T was calibrated by
observing the melting of 0.005

00

diameter Au wire

placed on the sapphire and correlating it with the
T

of the W/Re thermocouple. The T

was

reproducible to within 108C after 2 years of use.

For the decomposition study, pieces of the GaN

grown on a-plane

ð1 1 2 0Þ sapphire were cleaved

and weighed to within 0.1 mg using an analytical
balance [23]. The pieces were loaded into the
reactor and the process flows established. To
minimize the effects of decomposition during the
ramp from low to high T , the pieces were ramped
at a controlled rate (258C/min) from 8008C to the
final annealing T . After annealing for a set time,
ranging from 3 to 180 min

1

, the heater was shut

off. Each piece was re-weighed to determine the
mass loss.

In many instances, Ga droplets were observed

on the surface and were removed in dilute HNO

3

acid. After removal of the Ga droplets, each piece

3

Manufactured by Thomas Swan & Co., Ltd. The use of a

commercial product does not imply endorsement by the U.S.
Government.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

469

was weighed again to determine the weight of the
liquid Ga. The weight of the decomposed GaN
was calculated by subtracting the weight after
removal of the Ga droplets from initial weight.
The weight of desorbed Ga was calculated from
the mass balance difference between the decom-
posed GaN and the liquid Ga. Finally, the piece
was annealed at 10808C until all traces of GaN
were removed from the sapphire piece. The area,
A

, of the irregularly shaped pieces was determined

from the bare sapphire weight using the formula,
A

¼ m=td, where m is the sapphire weight, t is the

sapphire thickness (typically 0.033 cm), and d is
the sapphire density (3.98 g cm

3

). After convert-

ing the weights to molar quantities, the kinetic
rates (atoms cm

2

s

1

) were determined by divid-

ing by the sapphire area and anneal time. During
the ramp from 8008C to the final T , some GaN
decomposition occurred. Direct comparison of
258C/min and 508C/min ramp rates produced
similar kinetic rates, therefore decomposition
during the ramp was not corrected for in the
kinetic rates.

The surface morphology after annealing was

viewed using a Nomarski microscopy in the phase
contrast and transmission mode. Transmission
images were analyzed using NIH public domain
software

4

. The atomic-force-microscope (AFM)

measurements were performed under atmospheric
conditions using a Park Scientific Instruments
system in the contact mode.

5

4. Results

4.1. Ga droplet accumulation and growth

After annealing in H

2

, the most notable change

in the GaN surface morphology is the appearance
of Ga droplets as shown in Fig. 1. In Fig. 1, the

GaN pieces were annealed at 9928C for 10 min in
6 SLM of H

2

at pressures of 76 (Fig. 1(a)) and

150 Torr (Fig. 1(b)). Note that the Ga droplets are
larger and more numerous in Fig. 1(b) (150 Torr)
compared to Fig. 1(a) (76 Torr). Generally, as the
P

increased, the size of the Ga droplets increased.

In pure N

2

, Ga droplets were not observed for

T

510008C i n N

2

. However, Ga droplets were

observed in N

2

for T > 10008C. For similar T and

P

, the Ga droplet size was

 10 times smaller in N

2

compared in H

2

.

The Ga droplet size also increases as a function

of time at fixed T and P. In Fig. 2, images of the
GaN surface after (a) 3, (b) 10, (c) 20, and (d)
80 min of annealing in 6 SLM H

2

at 150 Torr and

8118C are shown. In Fig. 3, Ga droplet size
distributions, N

S

, vs. size are shown for annealing

times of (a) 10, (b) 20, (c) 30, and (d) 45 min. The
N

S

were obtained from Nomarski transmission

images using an analysis program to measure the
droplet area [52,53]. The droplet size was then
calculated as the square root of the area. The N

S

for 20, 30, and 45 min are shifted vertically in
Fig. 3. The solid line is a fit to N

S

based on power

law and bell shaped curve distributions [55]. Note
that the mean droplet size (i.e. peak of bell shaped
distribution) in Fig. 3 is

 8–9 mm after 10 min and

increases to

 20 mm after 45 min. Comparison of

N

S

shown in Fig. 3 to the growth model of Family

and Meakin [57] is discussed in Section 5.1.

4.2. Kinetic rates vs. pressure in H

2

The total P had a strong influence on the GaN

decomposition rate, k

GaN

, in flowing H

2

. The k

GaN

is plotted in Fig. 4 vs. pressure for T of 811 (filled
diamonds), 902 (open squares), and 9928C (filled
circles). It is clear from Fig. 4 that the k

GaN

increases as the H

2

pressure increases at constant

T

. A more dramatic increase in the k

GaN

occurs at

lower T. For example, at 9028C the k

GaN

at

400 Torr is 9 times the k

GaN

at 40 Torr. Also, the P

where the k

GaN

is enhanced shifts to higher P as

the T decreases. This can be seen in Fig. 4 by
comparing the onset for enhanced decomposition
at 9928C, which occurs at a P of 76 Torr to the
onset at 9028C and 8118C, which occur at 120 and
200 Torr, respectively. It is also clear from Fig. 4

4

Images were analyzed using NIH public domain software

(NIH Image program developed at the U.S. National Institutes
of Health and available on the Internet at http://rsb.info.nih.-
gov/nih-image/).

5

Manufactured by Park Scientific Instruments. The use of a

commercial product does not imply endorsement by the U.S.
Government.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

470

Fig. 1. Nomarski phase contrast pictures of the GaN surface after heating GaN for 10 min at 9928C i n H

2

pressures of (a) 76 Torr and

(b) 150 Torr. The droplets are liquid Ga. The bar on (a) indicates a length of 100 mm.

Fig. 2. Nomarski phase contrast pictures of the GaN surface after annealing GaN films at 8118C for (a) 3, (b) 10, (c) 20, and (d) 80 min
i n H

2

at a pressure of 150 Torr. The bar on (d) indicates a length of 50 mm.

Fig. 3. The natural logarithm of the GaN droplet size
distribution is plotted for GaN surfaces annealed at 8118C for
(a) 10, (b) 20, (c) 30, and (d) 45 min in H

2

at a pressure of

150 Torr. The solid line fit is combination of a power law
dependence (at small sizes) and a bell shaped curve (at large
sizes).

Fig. 4. GaN decomposition rate measured is 6 SLM of H

2

plotted vs. pressure at temperatures of 9928C (filled circles),
9028C (open squares), and 8118C (filled diamonds). The solid
and dashed lines are guides to the eye.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

471

that the k

GaN

at higher P (i.e. >200 Torr) and

lower T (i.e. 9028C) can exceed the k

GaN

at lower P

(i.e. 5 40 Torr) and higher T (i.e. 9928C). The
strong influence of P on the k

GaN

suggests a

change in the GaN decomposition mechanism as P
increases.

For the same T and P shown in Fig. 4, the liquid

Ga accumulation rate, k

S;Ga

, on the surface is

plotted in Fig. 5. This rate was obtained from the
weight difference between the annealed GaN piece
(with Ga droplets) and the same piece after
removal of the Ga in dilute HNO

3

. The k

S;Ga

closely coincide with the enhanced k

GaN

shown in

Fig. 4. Note that for P greater than 100 Torr, the
k

S;Ga

at 9028C exceeds the k

S;Ga

at 9928C. This is

due to the increased k

GaN

at higher P and the

reduced Ga desorption rate at lower T.

The Ga desorption rate, k

Ga

, is plotted in Fig. 6

for the same T and P shown in Figs. 4 and 5. Note
that in Fig. 6, the k

Ga

are plotted on an

exponential scale. Unlike the k

GaN

and k

S;Ga

, the

k

Ga

is relatively constant as a function of P at each

T

, only increasing at P of 80 Torr at 9928C,

110 Torr at 9028C, and 120 Torr at 8118C. The
increase in the k

Ga

at these P may be due to the

increased Ga surface area that occurs at onset of
droplet coalescence as described in Appendix A.

To achieve a measurable weight loss for the

rates shown in Figs. 4–6, the annealing time was
varied from 10 to 180 min. For Figs. 4–6, it is
assumed that the rates were constant in time (i.e.

steady state). This assumption was checked by
measuring the three rates as a function of time. In
Fig. 7, the k

GaN

, k

S;Ga

, and k

Ga

in 150 Torr of H

2

at

8118C are plotted as a function of total annealing
time. Note that after 20 min all three rates become
constant and that the k

Ga

is constant after 10 min.

At shorter times, the k

GaN

is larger and the k

GaN

and the k

S;Ga

are equal for times 510 min. At

higher T the incubation time to achieve constant
rates (in time) will be shorter because the rates
depend exponentially on T .

Fig. 5. Same as Fig. 4, except the Ga surface accumulation rate
is plotted.

Fig. 6. Same as Fig. 4, except the Ga desorption rate is plotted.

Fig. 7. GaN decomposition (filled circles), Ga surface accumu-
lation (open squares), and Ga desorption (filled diamonds) rates
are measured in 6 SLM H

2

for various annealing times. For this

plot the temperature was 8118C and the pressure was 150 Torr.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

472

The observation that the enhanced k

GaN

coin-

cides with the formation of liquid Ga on the
surface suggests that liquid Ga metal catalyzes the
decomposition of GaN [47,50,51]. To confirm this
observation, the GaN surface was predosed with
trimethylgallium for 10 min at 6008C to deposit
liquid Ga on the surface. On the Ga predosed
surface, the k

GaN

increased by 30% for annealing

at 9928C in 6 SLM of H

2

at 76 Torr. This increase

in k

GaN

it is not as large as the 300% increase in

the k

GaN

as P increases at T

¼ 9928C as shown in

Fig. 4.

Images of the surface morphology after etching

the Ga droplets in HNO

3

are shown in Fig. 8. For

this sample the GaN film was annealed at 8118C
for 60 min in 40 Torr of H

2

. The length scales are

(a) 40

 40 mm

2

, (b) 10

 10 mm

2

, and (c) 1

 1 mm

2

.

The raised features in Fig. 8(a) correlate well with
previous locations of Ga droplets when the images
of Ga droplets (see Fig. 1(a) of Ref. [22]) are
compared. The sunken features shown in Fig. 8(a)
extend

 1000 A˚ into the GaN film. In Fig. 8(a)

and (b) the measured step heights ranged from 50
to 200 A˚ and the maximum height (dark to light)
relief in Fig. 8(a) was approximately 2000 A˚. In
Fig. 8(c), atomic step heights of

 2.5 A˚ are

measured. The AFM images indicate that the
terraces are atomically smooth, suggesting that
decomposition occurs in a step flow fashion.

4.3. Kinetic rates vs. P in N

2

and mixed H

2

and N

2

flows

As N

2

flow is substituted for H

2

flow, the

k

GaN

, k

S;Ga

, and k

Ga

all decrease. This is shown in

Fig. 9, where the kinetic rates are plotted vs. the

N

2

fractional flow (i.e. [N

2

]/[N

2

] +[H

2

]) at a T of

9928C and a P of 76 Torr. In pure N

2

, the k

GaN

and k

Ga

drop by a factor of

 10 compared to the

rates in pure H

2

. The k

S;Ga

decreases exponentially

as N

2

is added, until in pure N

2

, Ga droplets are

not observed.

Unlike the k

GaN

measured in H

2

, no strong P

effect was observed for the k

GaN

i n N

2

. This is

shown in Fig. 10, where the k

GaN

i n N

2

are similar

at P of 76 and 150 Torr. Note in Fig. 10 that the
k

GaN

at 76 (open squares) and 150 (solid circles)

torr have the same curvature as N

2

is substituted

for H

2

. The solid and dashed lines are cubic fits to

k

GaN

vs. N

2

fractional flow, with only the k

GaN

at

100% N

2

not included in the fit. For T > 10008C,

the k

GaN

i n N

2

was less uniform, typically

Fig. 8. Atomic force microscopy images of the GaN surface after annealing at 8118C for 60 min in 40 Torr of H

2

. The images are

(a) 40

 40 mm

2

, (b) 10

 10 mm

2

, and (c) 1

 1 mm

2

.

Fig. 9. GaN decomposition (filled circles), Ga surface accumu-
lation (filled squares), and Ga desorption (open diamonds) rates
are measured at 9928C and 76 Torr as a function of N

2

fraction

of the total N

2

and H

2

flow.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

473

decomposing from the edges of the sapphire pieces
before the center.

4.4. Arrhenius parameters for GaN decomposition

The k

GaN

in pure H

2

and pure N

2

are plotted vs.

reciprocal T in Fig. 11. The k

GaN

was measured in

H

2

at P of 40 Torr (filled circles) and 76 Torr (open

circles), and in N

2

at P of 76 Torr (filled squares)

and 150 Torr (open squares). In H

2

a pre-

exponential, A

0

, of 2.8

 10

29

cm

2

s

1

and an

activation energy, E

A

, of 3.4 eV are measured,

while in N

2

a A

0

of 1.2

 10

29

cm

2

s

1

and an E

A

of 3.62 eV are measured. These values of A

0

and

E

A

along with Log

10

ðA

0

Þ are listed on lines 8 (H

2

)

and 9 (N

2

) of Table 1.

As shown in Fig. 4, the k

GaN

i n H

2

shows a

strong dependence on P. For this reason the k

GaN

were measured vs. T at P ranging from 40 to
250 Torr. An Arrhenius plot of the k

GaN

i n H

2

is

shown in Fig. 12 for P of 40 (filled circles), 76
(open circles), 150 (filled diamonds) and 250 Torr
(open diamonds). Plotting the k

GaN

this way,

shows that the two lower and two higher P have
nearly the same k

GaN

vs. 1=T . In addition, both

the low and high P data sets have a break in slope.
As a result, the k

GaN

in Fig. 12 were fit with four

exponentials: (A) the P476 Torr and T > 9008C
(line 8 of Table 1), (B) the P476 Torr and T

59008C (line 11), (C) the P5150 Torr and T
59008C (line 15), and (D) the P5150 Torr and
T >

9008C (line 19).

4.5. Decomposition rate vs. H

2

flow rate

The k

GaN

is plotted vs. H

2

flow rate in Fig. 13 at

a P of 76 Torr and a T of 9928C. As shown in

Fig. 10. GaN decomposition rate at 76 Torr (open squares) and
150 Torr (filled circles) at 9928C as a function of N

2

fraction of

the total N

2

and H

2

flow.

Fig. 11. Arrhenius plot of the GaN decomposition rate in H

2

at

40 (filled circles) and 76 Torr (open circles) and in N

2

at 76 Torr

(filled squares) and 150 Torr (open squares). Parameters for the
exponential fits (lines 8 and 9) are listed in Table 1.

Fig. 12. Arrhenius plot of the GaN decomposition rate in H

2

at

40 (filled circles), 76 Torr (open circles), 150 Torr (filled
diamonds), and 250 Torr (open diamonds). Parameters for the
exponential fits (lines 8, 11, 15, 19) are listed in Table 1.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

474

Fig. 13, the k

GaN

increases by 25% as the H

2

push

flow increases from 1 to 10 SLM. A much larger
increase in the k

S;Ga

is observed (4.5 times) as the

H

2

push flow is increased. A similar increase in the

k

GaN

vs. the H

2

push flow was previously reported

[28].

4.6. GaN decomposition measured vs. time

In Fig. 14, the k

GaN

is measured over 18 months

for 10 min anneals at 9928C in 6 SLM of H

2

at

76 Torr. As shown in Fig. 14, the initial k

GaN

measured in the Fall of 1997 [23] was larger than
measurements under identical annealing condi-
tions in March through August of 1999 after
cleaning the showerhead. In November of 1999 a
new showerhead was installed on the CSS reactor.
After the showerhead change, the k

GaN

was

measured four times (see Fig. 14). The k

GaN

after

the showerhead change are similar to those
measured in March through August of 1999.

4.7. Growth rate vs. growth pressure

The GaN growth rate at 10308C is plotted vs. P

in Fig. 15(c). For the growth, 2 SLM NH

3

, 4 SLM

H

2

, and 32 mmol of TMGa were used. In Fig. 15(b)

the k

GaN

is plotted for the same conditions used in

Fig. 15(c) except that no TMGa was used. Note
that as the P increases the k

GaN

increases and the

GaN growth rate decreases. For comparison the
k

GaN

in pure H

2

at 9928C vs. P (from Fig. 4) are

plotted in Fig. 15(a). The k

GaN

in both H

2

Fig. 13. GaN decomposition (filled circles), Ga surface accu-
mulation (open squares), and Ga desorption (filled diamonds)
rates at 9928C and 76 Torr as a function of H

2

flow rate.

Fig. 14. GaN decomposition rate measured over several
months time. The GaN pieces were annealed at a temperature
of 9928C and a pressure of 76 Torr in 6 SLM H

2

for 10 min. The

filled circle is the measured decomposition rate reported in Ref.
[19] the open squares were measured after cleaning the
showerhead, and the filled diamonds were measured after the
installation of a new showerhead.

Fig. 15. Plots of (a) the GaN decomposition rate in 6 SLM of
H

2

(open circles) at T

¼ 9928C, (b) the GaN decomposition rate

in 2 SLM of NH

3

and 4 SLM of H

2

(filled circles) at

T

¼ 10308C, (c) the GaN growth rate in 2 SLM of NH

3

,

4 SLM of H

2

, and 32 mmol/min of TMGa (open squares) at

T

¼ 10308C, and (d) the calculated incorporation rate obtained

by adding curve b to curve c (filled diamonds) as a function of
reactor pressure.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

475

(Fig. 15(a)) and mixed H

2

and NH

3

(Fig 15(b))

have a similar shape for P from 40 to 300 Torr. In
fact, if the k

GaN

in Fig. 15(b) are multiplied by

 30, the k

GaN

i n H

2

(Fig. 15(a)) and in mixed H

2

and NH

3

(Fig. 15(b)) overlap, suggesting a similar

mechanism for GaN decomposition in both gas
environments.

5. Discussion

In the previous section, changes in the GaN

decomposition rate

ðk

GaN

Þ, liquid Ga accumula-

tion rate

ðk

S;Ga

Þ, and Ga desorption rate ðk

Ga

Þ

were presented for a range of gas mixtures, P, and
T

. In this section we compare the measurements

presented in Section 4 to the measurements of
GaN decomposition reported in the literature and
offer explanations as to how the GaN decomposi-
tion mechanism changes as the annealing condi-
tions change.

5.1. Ga droplet accumulation and growth

Previous studies of GaAs decomposition by

Carlow et al. showed that Ga droplets form in the
absence of an As overpressure when the substrate
is heated to 6608C [58]. Further heating of the
GaAs surface causes the Ga droplets to grow in
size and coalesce. Carlow et al. suggested that the
Ga droplet coalescence mechanism was consistent
with the growth model of Family and Meakin [57].
To verify the Family and Meakin model, Carlow
et al. used the combination of a power law and bell
shaped curve to fit the Ga droplet distribution.
Likewise, this combination of fits was used in
Fig. 3.

The coalescence growth model of Family and

Meakin assumes that when two spherical droplets
coalesce they form a single droplet on the center of
mass of the original droplets [57]. The droplets
initiate from the Ga accumulation on the surface,
because the k

GaN

>

the k

Ga

. When droplets begin

to coalesce, the droplet size increases [57]. Droplet
formation is shown in Figs. 2(a) and (b), where the
droplets are

 1 mm. Over time these small

droplets coalesce and the droplet size increases as
shown in Fig. 3. Evidence of droplet coalescence

can be seen in bottom left corner of Fig. 2(b). The
close fit by the combination power law and bell
shaped curve suggests good agreement between the
droplet size distributions and the theoretical model
[57]. Further comparison between the droplets and
the model [57] is outside this papers scope.

6

Clearly, as the H

2

pressure increases, both the

k

GaN

(Fig. 4) and the k

S;Ga

(Fig. 5) increase.

Despite this coincidence, the presence of droplets
are most likely the result of the increased k

GaN

at

higher P. Liquid Ga enhances GaN decomposition
locally on the surface [51], because H

2

is cracked

on liquid Ga surfaces [59,60]. However, the
primarily reason for the increases in k

GaN

and

k

S;Ga

vs. H

2

pressure is because the N removal rate

exceeds k

Ga

as is discussed in Section 5.4.

5.2. Kinetics of Ga desorption

Except for the increase near 80–120 Torr, the

k

Ga

does not depend on H

2

pressure. As shown in

Table 1E, the E

A

for Ga desorption are very

similar. The E

A

measured for Ga desorption are

nearly identical to the E

A

of 2.8 eV measured for

Ga desorption from liquid Ga (line 20, Table 1)
[39] and the heat of vaporization for Ga which is
2.8 eV [56]. This close agreement of E

A

suggests

that the barrier to Ga desorption is the breaking of
a single Ga–Ga bond. Despite the similarity in the
E

A

, the value of A

0

is 12 times larger in H

2

compared to N

2

(compare lines 23 and 24 in Table

1). Possible origins for the difference in A

0

are

discussed in Appendix B. An explanation for the
increase in the k

Ga

between 80 and 120 Torr in

Fig. 6 is suggested in Appendix A.

5.3. Kinetic rates vs. time

When measuring kinetic rates it is important

that the rates approach a constant value in time,

6

According to the coalescence model of Ref. [54], the change

in droplet size from 30 to 45 min should be larger than the
experimental Ga droplet distributions, where no significant
increase in the droplet size is observed between Figs. 2(c) and
2(d). The discrepancy between the model and the experimental
data may be due to a change in k

Ga

, which depends on the

surface area as described in Section 5.2 and is not included in
the coalescence model.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

476

especially if time is varied to measure slow rates. In
Fig. 7, the kinetic rates at a T of 8118C and P of
150 Torr are shown vs. annealing time. Despite a
short induction time, which is

 10 min, all three

rates become constant at longer times. Because the
rates become constant, the slower kinetic rates
could be measured by increasing the anneal time.
Several factors may contribute to the observed
induction time, including an increase in the k

GaN

at

the surface due to the larger concentration of
subsurface H [61]. The larger surface H concentra-
tion may enhance decomposition mechanisms with
lower E

A

such as NH

3

formation. Because the

k

GaN

is initially larger, the k

S;Ga

observed at 10 min

in Fig. 7 is larger. Note that at higher T the
induction time will be shorter than 10 min, due to
the increased rates at higher T.

5.4. Rate of atomic H production in CSS reactor

For NH

3

formation, H

2

must be cracked on the

GaN surface to form adsorbed H. Ga metal is
known to dissociate H

2

at high temperatures to

form Ga hydrides [59,60]. Recently, Bartram and
Creighton have shown that H

2

can be dissocia-

tively chemisorbed at elevated temperatures on the
GaN surface [19]. Under an atomic H flux of

 10

18

cm

2

s

1

, King et al. have calculated that at

least 50% of the surface N bonds should be H
terminated at 8008C, while at higher atomic H
fluxes (

 10

20

cm

2

s

1

), 50% H termination

occurs at 10008C [62].

Although

the

degree

of

atomic

H,

H

at

,

production was not measured, an estimate can be
made from the NH

3

production rate. In Fig. 10,

the k

GaN

is

 1.6  10

16

cm

2

s

1

at 150 Torr at

9928C. In order to produce NH

3

at this rate, the

steady state H

at

flux must be greater than

4.8

 10

16

cm

2

s

1

or 42 ML s

1

. However, NH

3

formation from surface N and H requires the
formation of three N–H bonds in three sequential
reversible chemical reactions. This suggests that
the H

at

flux generated at the GaN surface

is

1–4

orders

of

magnitude

larger

than

4.8

 10

16

cm

2

s

1

. Certainly, H

at

is produced at

a substantial rate in the MOVPE reactor during
growth and decomposition.

5.5. Reaction order for NH

3

production during

GaN decomposition in H

2

Dissociated H on the GaN surface can recom-

bine with surface N to form NH

3

. This reaction

channel is shown in Fig. 10 by the decrease in the
k

GaN

as N

2

is substituted for H

2

. The solid

(150 Torr) and dashed (76 Torr) lines are cubic fits
to the data, showing that k

GaN

is proportional to

[H

2

]

3

. This is the expected number of H

2

that need

to be cracked on the surface to form NH

3

, if only

one H per H

2

molecule attaches to a surface N

atom. Recently, Mayumiet al. measured the k

GaN

to be proportional to partial P of H

2

to the 2/3

power [36]. The difference in power dependence
between the data in Fig. 10 (third power) and in
Ref. [36] (2/3 power) is not currently understood.
This difference may be due to the lower T used in
Ref. [36], which would reduce the H

2

desorption

rate from Ga atoms on the surface. Further work
is needed to understand the reactor influence on
the H

2

dissociation rate, because it plays a large

role in GaN decomposition and growth.

5.6. Grouping of GaN decomposition kinetic rates

The dependence of the GaN decomposition

kinetic parameters on process conditions is sum-
marized in Table 1. Unexpectedly, the measured
kinetic parameters fall into one of four different
groups, denoted as A–D. The four distinct slopes
in Fig. 12 imply that four different decomposition
reactions become rate limiting. Here, we propose
how each of these reactions limit GaN decom-
position.

In Table 1A, the Arrhenius parameters are listed

for T > 9008C, in H

2

at P of 40 and 76 Torr (line 8)

and in N

2

at P of 76 and 150 Torr (line 9). Under

these conditions, an E

A

of 3.4

0.2 eV is measured

i n H

2

, while a slightly larger E

A

of 3.62

0.14 eV is

measured in N

2

. The values for the E

A

compare

well to previous E

A

measured in vacuum (lines 1–

7, Table 1), suggesting that under these conditions
the GaN decomposition mechanism is similar. The
proposed rate limiting step for GaN decomposi-
tion is the formation and desorption of N

2

[26,31,32]. Forming N

2

on the surface would

involve the diffusion of at least one N atom to

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

477

form N

2

, however N surface diffusion is unlikely

because of a large kinetic barrier [6,16]. Alterna-
tively, a subsurface N could combine with a
surface N to form N

2

. This mechanism has

recently been observed by Bartram and Creighton,
where

15

N

14

N was shown to desorb after dosing

the surface with

15

NH

3

[19]. Nitrogen formation

from surface and subsurface N atoms would
follow first order desorption kinetics as previously
observed [6]. The E

A

(Table 1A) for forming N

2

would then equal to the energy required to remove
a subsurface N atom, i.e. the energy required to
create a N-vacancy. Estimates of the N-vacancy
formation energy range from 2.7 eV for a Ga-rich
surface to 4.6 eV for a N-rich surface [63]. Note
that the N-vacancy formation energy is within the
range of measured E

A

in Table 1A. The increased

A

0

i n H

2

vs. N

2

is most likely related to the

increased A

0

also observed for Ga desorption in

H

2

vs. N

2

.

7

For T 59258C at P476 Torr, the measured E

A

and A

0

decrease for the k

GaN

(see Table 1B). The

decrease in E

A

and A

0

is seen in Fig. 12, where for

T

59258C the k

GaN

is larger than the expected

k

GaN

based on extrapolation of line 8. A fit to k

GaN

for T59258C gives an E

A

of 0.98 eV (line 11 of

Table 1). This E

A

is close to the E

A

(0.91 eV)

measured Kobayashiand Kobayashiin 76 Torr of
H

2

(line 10 of Table 1) [37]. The decrease in E

A

from

 3.5 to  1 eV implies a change in the rate

limiting step for decomposition. At lower T , the H
coverage will increase due to the decreased H
desorption rate from N sites [37], resulting in
increased stability of –NH and –NH

2

species.

8

Because N–H bonds are strong (i.e. 4.5 eV [56]),
the –NH and –NH

2

species will be more weakly

bound to the surface. We propose that two

partially hydrogenated N species bond to form a
hydrogenated dinitrogen molecule

ðN

2

H

x

Þ, which

then desorbs from the surface. The formation of
this molecule would have a lower E

A

compared to

the E

A

for N

2

formation and desorption. While the

H surface coverage at lower P is larger at lower T,
the coverage is not large enough to form NH

3

.

For T59008C and P5150 Torr the k

GaN

increases compared to the k

GaN

at lower P (Table

1C). This can be seen in Fig. 12 by comparing line
11 (lower P) and line 15 (higher P). Under these
annealing conditions an E

A

of 1.7 eV is calculated

(line 15 of Table 1). Similar E

A

have been observed

by Morimoto [26] (line 13, Table 1) and Rebey
et al. [28] (line 14, Table 1) at similar T and P. The
rate limiting step for GaN decomposition under
these conditions (i.e. T59008C and P 5150 Torr)
i s NH

3

desorption from the surface. Recently,

Bartram and Creighton found that the NH

3

desorption rate from the GaN surface peaks at a
T

of 4478C [19]. Using T

¼ 4478C, a Redhead

analysis gives an E

A

of 1.6–2.0 eV if an A

0

of 10

13

–

10

20

cm

2

s

1

is assumed [64]. The range of 1.6–

2.0 eV agrees well with the E

A

listed on lines 13–15

of Table 1, suggesting that NH

3

desorption limits

GaN decomposition at lower T and higher P.

Finally, as the T is increased at higher P, the

k

GaN

is relatively constant as shown by line 19 in

Fig. 12 (see Table 1D). An E

A

of 0.34 eV is

calculated from line 19. This value of E

A

agrees

well with the E

A

of 0.38 eV measured by Rebey

et al. (line 17, Table 1) [28] and the E

A

of 0.44 eV

measured by Mayumiet al. (line 18, Table 1) [36].
At these T and P, Ga droplets are prevalent
because k

GaN

exceeds k

Ga

. The change in slope

near 9008C suggests that NH

3

desorption no

longer limits the k

GaN

. Instead the lower E

A

is

suggestive of a kinetic barrier for metallic surface
diffusion [65]. Recently, the Ga surface diffusion
barrier on a Ga terminated surface was calculated
to be

 0.4 eV (line 16 in Table 1) [38]. This value

for the Ga diffusion barrier agrees well with the
experimentally measured E

A

listed in Table 1D. As

a result, the limiting factor for GaN decomposi-
tion is the rate of Ga diffusion so that the next
layer of N atoms are exposed for removal via the
formation of NH

3

. Note that the k

S;Ga

plotted in

Fig. 5 drops dramatically for T 59008C and

7

The increase in the pre-exponential factor, A

0

, for GaN

decomposition, k

GaN

, i n H

2

vs. N

2

is similar to the increased A

0

for Ga desorption in H

2

vs. N

2

described in Ref. [23]. Note that

in both cases the k

GaN

and k

Ga

are larger in H

2

than in N

2

. In

Appendix B, it was proposed that the increase A

0

for Ga

desorption is due to an increased entropy of more volatile Ga–
H species. Increased desorption of Ga would increase the
number of N exposed at the surface, which in turn would
increase the number of N

2

desorbing per second.

8

An increased surface and subsurface H coverage were

measured in MOVPE grown GaN films in Ref. [61].

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

478

P >

100 Torr, in agreement with the change in the

rate limiting step from Ga diffusion limited (Table
1D) to NH

3

desorption limited (Table 1C).

It is apparent from the A–D grouping in Table 1

that GaN decomposition strongly depends on
P

, T , and gas type (i.e. H

2

or N

2

). One key to

understanding the changes in the k

GaN

is the rate

of H

at

production. While the production rate of

H

at

on the GaN surface may be reactor dependent,

in general it should increase as P increases. In fact,
one of the major distinguishing characteristics
between atmospheric and low P GaN growth may
be the H

at

production rate and its subsequent

influence on GaN growth and decomposition.

5.7. GaN nucleation layer decomposition and
evolution

Several groups have shown that the nucleation

layer, NL, evolves during the ramp from low to
high T used for the main layer growth [21,29,66–
70]. As shown by X-ray diffraction, annealing of
the GaN NL at high T increases its crystallinity
[66]. Recently Sugiura et al. have shown that a T
of 850–9008C (P

¼ 760 Torr) produces an optimal

NL for high quality films, while above 9508C
reevaporation of the NL can occur [67]. Annealing
the NL at higher P increases the NL roughness
and the GaN electron mobility when compared to
annealing the NL at lower P [29]. Similar results
have been reported by Wickenden et al., where
higher m and larger GaN grain sizes were produced
on GaN NL annealed at 150 Torr compared to
76 Torr [21]. The grain size in the NL also
increases during annealing at 10808C with the
tops of the GaN NL grains becoming flatter [68–
70]. The flattening occurs because the NL grains
decomposition during the high T anneal and the
Ga atoms reform GaN by reaction with NH

3

. If

the N loss from the surface is not compensated by
enough NH

3

during NL annealing, the entire GaN

NL decomposes. This has been observed by
Kobayashiet al., who noted the loss of the GaN
NL when heated to 10208C i n NH

3

(0.25 SLM)

and H

2

[71]. Because the k

GaN

is enhanced at

higher P, the NL may evolve at a faster rate when
annealed at high P. Therefore, conditions that

increase GaN decomposition should accelerate the
rate of the GaN NL evolution.

5.8. Relationship of GaN decomposition to the
quality of GaN growth

While there is not an obvious link between the

high T GaN growth and GaN decomposition, we
have grown better quality GaN growth in the CSS
reactor at P > 100 Torr [21,22]. When the GaN
epitaxial layer is grown at 150 Torr, we find a near
doubling of the electronic mobility (m > 500 cm

2

/

V s) compared to films growth at 76 Torr on
identically grown NL [21]. For films grown above
100 Torr, the GaN grain size increased from 51 to
2–5 mm, which may be directly responsible for the
increased mobility [21,22]. Other groups using CSS
or high speed rotating disk reactors have also
reported improved electric properties when GaN is
grown at P > 100 Torr [29,72]. Watanabe and
coworkers have shown that the dislocation density
decreases, as the growth P is increases [73].
Previously, we speculated that having some degree
of GaN decomposition during GaN growth is
important for removing more weakly bound Ga
and N atoms, which would increase ordering
during growth [16]. From a thermodynamic
perspective, lower activation barriers for decom-
position bring the GaN growth closer to equili-
brium.

Theoretically, growth rate equals the incorpora-

tion rate minus the decomposition rate [16]. At a
constant growth flux, the incorporation rate
should not change; therefore any increase in the
decomposition rate will decrease the growth rate.
This is shown in Fig. 15(d), where the incorpora-
tion rates (solid diamonds) are calculated vs. P by
adding the k

GaN

(Fig. 15(b)) to the GaN growth

rates (Fig. 15(c)). Below 150 Torr, the incorpora-
tion rate is constant at

 1.7  10

15

cm

2

s

1

,

however at 300 Torr the incorporation rate de-
creases. This decrease in the incorporation rate
may be due to gas phase pre-reaction [74] or site
blocking at the GaN surface [75]. At equilibrium,
the decomposition and incorporation rates are
equal and no growth occurs. Note that at
300 Torr, the k

GaN

and the incorporation rate are

the closest for all P, suggesting that at this P the

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

479

GaN growth is closest to equilibrium. The k

GaN

in

pure H

2

(Fig. 15(a) and in mixed H

2

and NH

3

(Fig. 15(b)) increase similarly as P increases
suggesting that the same decomposition mechan-
ism, i.e. H

2

dissociation and removal of surface N

as NH

3

, occurs in both cases as P increases.

Other groups have also observed decreases in

growth rate when the growth P is increased
[28,33,76,77]. For example, Khan et al. observed
a factor of 2 decrease in the GaN growth rate as P
was increased from 40–100 Torr [76]. Reductions
in growth rate have also been observed when H

2

is

substituted for N

2

as the carrier gas [33]. The use

of higher H

2

push flows during growth has

previously been shown to lead to discontinuous
growth followed by no GaN film growth [77].
Sasaki also showed that the transition between
continuous and discontinuous growth occurred at
lower H

2

push flows when the growth P was

increased from 70 to 780 Torr [77]. The observa-
tions of Sasaki are consistent with the increase in
k

GaN

shown in Fig. 14 and in Ref. [28]. For each

case the likely cause for the decreased GaN growth
rate is an increase in the k

GaN

.

Besides affecting the GaN growth rate, the GaN

crystal quality is also improved when pure H

2

is

used instead of pure N

2

. In a comparison study of

GaN growth in H

2

vs. N

2

, Kistenmacher et al.

showed that the FWHM of the GaN films grown
i n H

2

had narrower X-ray rocking curve line-

widths and were better aligned compared (i.e.
smaller mosaic dispersion) to GaN films grown in
only N

2

[78]. Scho¨n and coworkers find smoother

morphologies and better electrical properties when
growth is conducted in H

2

compared to N

2

[79].

Tadatomo and coworkers find that the lateral
growth rate in N

2

increases compared to growth in

H

2

, however, there is a greater degree of mis-

orientation in the overgrown region and a smooth
surface cannot be attained [80]. These studies
indirectly suggest higher quality growth in H

2

compared to growth in N

2

. We speculate that an

increase in the k

GaN

when H

2

is used as the carrier

gas plays a role in the initial grain formation and
hence the resulting GaN quality.

It is not clear to what extent the surface H atoms

etch the GaN surface or block sites necessary for
GaN growth. Site blocking may account for the

decreased incorporation rate

at 300 Torr in

Fig. 15(d). Recently, Briot and coworkers pro-
posed that the GaN growth rate decreases at high
NH

3

flow because the excess NH

3

blocks surface

sites necessary for growth [75]. When large
amounts of H

at

are generated on the GaN surface,

the H coverage may become large enough to block
surface sites, thereby reducing the incorporation
and subsequent growth rates. Further study is
necessary to determine the relative contributions
of site blocking and k

GaN

to limiting the GaN

growth.

Finally, a benefit of large surface H coverages

might be an increase in the Ga diffusion length [16]
through the formation of more volatile and mobile
Ga–H species. For example, Morishita et al.
measured an increased Ga diffusion length in H

2

and H

at

. The increased diffusion length was

speculated to be due to the formation of Ga–H
species [81]. In addition, Okamoto and coworkers
have recently showed a suppression of 3D growth
morphology when H

at

is used during MBE growth

of GaN [82], implying increased surface mobility
of the Ga–H species. Note that the A

0

for the k

Ga

are larger in H

2

(line 23) than in N

2

(line 24) as

shown in Table 1. As proposed in Appendix B,
Ga–H species that desorb may redeposit at a
different position along the GaN surface, conse-
quently increasing the Ga diffusion length. High
levels of surface H may also scavenge and remove
graphitic C from the surface. Growth parameters
that enhance k

GaN

should also decrease the surface

and bulk C concentrations.

6. Conclusions

This paper has shown how the kinetics rates of

GaN decomposition, Ga surface accumulation,
and Ga desorption change under different anneal-
ing conditions. We have shown that the GaN
decomposition kinetic parameters can be arranged
into four distinct groups. We have also suggested
possible chemical reactions, which ultimately limit
the decomposition. The reformation of ammonia
is suggested when GaN is decomposed in H

2

at

higher P. The observation of the NH

3

reformation

reaction is important for growth because it

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

480

suggests that at higher P GaN growth is closer to
equilibrium [16]. Unlike GaN decomposition, the
Ga desorption kinetics are similar in H

2

and N

2

environments with only a difference in the pre-
exponential factor.

The occurrence of GaN decomposition during

GaN growth has several implications. Clearly,
GaN decomposition during growth affects the
quality of the GaN films. Annealing the NL at
higher temperatures and pressures allows the NL
to evolve to a greater extent for the formation of
large grained GaN [21]. Limiting GaN grain
formation at higher P (because of the increased
GaN decomposition) also contributes to the
increased GaN grain size observed in the full
GaN films [21,22]. From Fig. 14 it appears that the
decomposition rate is nearly constant; this even
after cleaning the showerhead and replacement
with a new showerhead.

The extent of GaN decomposition that occurs

during growth is condition (see Fig. 15(b)) as well
as reactor dependent [16]. Here, we have shown
that the H

2

dissociation rate plays a critical role in

the decomposition rates. The kinetic rates for
many of the GaN growth reactions such as the
cracking rate of H

2

may explain, at least partly, the

reason why GaN growth conditions are not
directly transferable from reactor to reactor.

Finally, because of the differences in reactor de-

signs (horizontal, vertical, two flow, etc.) and flow
dynamics (gas composition, preheating, residence
time, etc.), the measured GaN decomposition rates
in other types of reactors may differ from those
presented here. However, the general trends and
rate limiting decomposition mechanisms presented
here are likely present in all GaN MOVPE
reactors. Clearly, more research is necessary to
measure kinetic factors that control GaN growth.

Acknowledgements

We thank V.A. Shamamian, V.M. Bermudez,

R.J. Gorman, J.A. Freitas, Jr., and M. Fatemi for
their assistance in this work. We thank J.E. Butler
for the use of a phase contrast microscope. This
work was supported by the Office of Naval
Research.

Appendix A

In this section, we explain the peak in the Ga

desorption rate between 80 and 120 Torr in Fig. 6.
This peak is probably due to an increase in the
surface area, S, of the Ga droplets that occurs in
this pressure regime. An increase in S is expected
from the model of Family and Meakin as the
droplets begin to coalesce [57]. If the droplet size,
R

, has a combination power law, i.e. R

x

, and bell-

shaped distribution, i.e. exp

ðdðR  R

0

Þ

2

Þ, the

droplet size distribution, N

ðRÞ, is given by

N

0

R

x

exp

ð  dðR  R

0

Þ

2

Þ, where N

0

is the num-

ber density as R

! 0; x is the power law depen-

dence, d is a constant proportional to the width of
the bell shaped distribution, and R

0

is the peak

droplet size of the bell shaped distribution. The S
of a spherical droplet on a flat surface, is given by
pR

2

ð1 þ tan

2

ðy=2ÞÞ, where y is the contact angle

[83]. To estimate the total S of the Ga droplets
requires summing S of each droplet over the
droplet distribution, N

ðRÞ. Assuming that the

contact angle between the Ga droplets and
the GaN surface is constant, the surface area
distribution,

S

ðRÞ ¼ const ðNðRÞÞ

2

¼ const ½N

0

R

x

exp

ðdðR  R

0

Þ

2

Þ

2

. Integration of S

ðRÞ over

R

would provide the total S. Therefore, the change

in S

ðRÞ can be estimated from the change in NðRÞ

as the droplets initially form, begin to coalesce,
and grow in size.

Based on Family and Meakin’s theory, the

power law term, R

x

, is nearly constant and only

the bell shaped term, exp

ðdðR  R

0

Þ

2

Þ, changes

as the droplets coalesce [57]. This is observed in the
fits in Figs. 3 and 4 of Ref. [57]. Initially, when the
Ga droplets first appear, the exp

ðdðR  R

0

Þ

2

Þ

term is negligible, S

ðRÞ / ðN

0

R

x

Þ

2

, and integra-

tion of S

ðRÞ over R will be a constant. As the

droplets begin to coalesce, there will be a large
intensity of droplets at a small value of R

0

,

increasing the exp

ðdðR  R

0

Þ

2

Þ term, which will

increase the total S. As the droplets further
coalesce, the exp

ðdðR  R

0

Þ

2

Þ term will decrease

as R

0

increases and as a result the total S will

decrease. Eventually, as the droplets become
larger, the contribution to the total S from the
bell shaped distribution will again become negli-
gible, S

ðRÞ / ðN

0

R

x

Þ

2

, and S will again be a

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

481

constant. Therefore, the Ga droplet surface area is
initially constant before coalescence, increases at
the onset of coalescence, and then decreases back
to the initial value as the Ga droplets grow in size.
This behavior is qualitatively observed in Fig. 6, as
the H

2

pressure increases at 9928C and the Ga

droplets initially form (10–40 Torr), begin to
coalesce (76–150 Torr), and increase in size (250–
700 Torr).

Appendix B

The larger pre-exponential factor, A

0

, for Ga

desorption in H

2

compared to N

2

may be due to an

increase in the frequency factor, v, or the reaction
entropy, S, if a standard thermodynamic inter-
pretation of A

0

is used, i.e. A

0

¼ ve

ðS=kÞ

, where k is

the Boltzman constant [84]. Previously, we pro-
posed that a reduced v for Ga desorption in N

2

vs.

H

2

might be possible because of the more efficient

heat transfer between N

2

and the hot surface [24].

Additionally, increases in v and S might be
possible if the Ga adatoms become hydrogenated
i n H

2

. Increases have been observed in the A

0

for

Ga surface diffusion when the GaAs surface is
exposed to H

2

and atomic H [81]. In this study, the

increase in A

0

for Ga surface diffusion was

attributed to the formation of GaH species, which
were more mobile on the surface [81]. Recent
thermodynamic calculations show that when GaN
is heated in H

2

near 10008C GaH species are more

abundant in the gas phase than Ga atoms [50].
Another calculation shows that Ga desorption in
mixed H

2

and NH

3

environments is in the form of

Ga hydrides rather than Ga atoms [85]. Finally,
the gas phase entropy for GaH is larger than for
Ga atoms [56]. Therefore, the larger A

0

measured

i n H

2

vs. N

2

could be the result of forming Ga–H

species, which have larger entropy and/or the
reduced heat transfer between H

2

and the surface,

discussed in Ref. [24].

References

[1] S. Nakamura, M. Senoh, T. Mukai, Appl. Phys. Lett. 64

(1994) 1687.

[2] S. Nakamura, M. Senoh, S. Nagahama, N. Iwasa, T.

Matushita, T. Mukai, MRS Internet J. Nitride Semicond.
Res. 4S1 (1999) G1.1.

[3] S.N. Mohammad, A.A. Salvador, H. Morkoc, Proc. IEEE

83 (1995) 1306.

[4] S. Strite, H. Morkoc, J. Vac. Sci. Technol. B 10 (1992)

1237.

[5] R.F. Davis, Proc. IEEE 79 (1991) 702.
[6] O. Brandt, H. Yang, K.H. Ploog, Phys. Rev. B 54 (1996)

4432.

[7] R. Held, B.E. Ishaug, A. Parkhomovsky, A.M. Dabiran,

P.I. Cohen, J. Appl. Phys. 87 (2000) 1219.

[8] D. Crawford, R. Held, A.M. Johnston, A.M. Dabiran, P.I.

Cohen, MRS Internet J. Nitride Semicond. Res. 1 (1996)
12.

[9] E.J. Tarsa, B. Heying, X.H. Wu, P. Fini, S.P. DenBaars,

J.S. Speck, J. Appl. Phys. 82 (1997) 5472.

[10] C.R. Jones, T. Lei, R. Kaspi, K.R. Evans, Mater. Res.

Soc. Symp. Proc. 395 (1996) 141.

[11] R. Held, D.E. Crawford, A.M. Johnston, A.M. Dabiran,

P.I. Cohen, Surf. Rev. Lett. 5 (1998) 913.

[12] S.P. DenBaars, S. Keller, Semicond. Semimet 50 (1998) 11.
[13] O. Ambacher, J. Phys. D 31 (1998) 2653–2710.
[14] O. Briot, J.P. Alexis, M. Tchounkeu, R.L. Aulombard,

Mater. Sci. Eng. B 43 (1997) 147.

[15] R.D. Dupuis, J. Crystal Growth 178 (1997) 56.
[16] D.D. Koleske, A.E. Wickenden, R.L. Henry, W.J.

DeSisto, R.J. Gorman, J. Appl. Phys. 84 (1998) 1998.

[17] S.S. Liu, D.A. Stevenson, J. Electrochem. Soc. 125 (1978)

1161.

[18] M. Kamp, M. Mayer, A. Pelzmann, K.J. Ebeling, Mater.

Res. Soc. Symp. Proc. 449 (1997) 161.

[19] M.E. Bartram, J. R. Creighton, MRS Internet J. Nitride

Semicond. Res. 4S1 (1999) G3.68.

[20] H. Amano, N. Sawaki, I. Akasaki, Y. Toyoda, Appl. Phys.

Lett. 48 (1986) 353.

[21] A.E. Wickenden, D.D. Koleske, R.L. Henry, R.J. Gor-

man, J.C. Culbertson, M.E. Twigg, J. Electron. Mater. 28
(1999) 299.

[22] A.E. Wickenden, D.D. Koleske, R.L. Henry, R.J. Gor-

man, M.E. Twigg, M. Fatemi, J.A. Freitas Jr., W.J.
Moore, J. Electron. Mater. 29 (2000) 21.

[23] D.D. Koleske, A.E. Wickenden, R.L. Henry, M.E. Twigg,

J.C. Culbertson, R.J. Gorman, Appl. Phys. Lett. 73 (1998)
2018; Erratum Appl. Phys. Lett. 75 (1999) 1646.

[24] D.D. Koleske, A.E. Wickenden, R.L. Henry, M.E. Twigg,

J.C. Culbertson, R.J. Gorman, MRS Internet J. Nitride
Semicond. Res. 4S1 (1999) G3.70.

[25] D.D. Koleske, A.E. Wickenden, R.L. Henry, MRS

Internet J. Nitride Semicond. Res. 5S1 (2000) W3.64.

[26] Y. Morimoto, J. Electrochem. Soc. 121 (1974) 1383.
[27] M. Furado, G. Jacob, J. Crystal Growth 64 (1983) 257.
[28] A. Rebey, T. Boufaden, B. El Jani, J. Crystal Growth 203

(1999) 12.

[29] J. Han, T.-B. Ng, R.M. Biefeld, M.H. Crawford, D.M.

Follstaedt, Appl. Phys. Lett. 71 (1997) 3114.

[30] R.J. Sime, J.L. Margrave, J. Phys. Chem. 60 (1956) 810.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

482

[31] Z.A. Munir, A.W. Searcy, J. Chem. Phys. 42 (1965) 4223.
[32] R. Groh, G. Gerey, L. Bartha, J.I. Pankove, Phys. Status

Solidi A 26 (1974) 353.

[33] O. Ambacher, M.S. Brandt, R. Dimitrov, T. Metzger, M.

Stutzmann, R.A. Fischer, A. Miehr, A. Bergmaier, G.
Dollinger, J. Vac. Sci. Technol. B 14 (1996) 3532.

[34] R. Held, D.E. Crawford, A.M. Johnston, A.M. Dabiran,

P.I. Cohen, Surf. Rev. Lett. 5 (1998) 913.

[35] Grandjean, J. Massies, F. Semond, S.Yu. Karpov, R.A.

Talalaev, Appl. Phys. Lett. 74 (1999) 1854.

[36] M. Mayumi, F. Satoh, Y. Kumagai, K. Takemoto, A.

Koukitu, Jpn. J. Appl. Phys. 39 (2000) L707.

[37] Y. Kobayashi, N. Kobayashi, J. Crystal Growth 189/190

(1998) 301.

[38] T. Zywietz, J. Neugebauer, M. Scheffler, Appl. Phys. Lett.

73 (1998) 487.

[39] R.E. Honig, D.A. Kramer, RCA Rev. 30 (1969) 285.
[40] P. Hacke, G. Feuillet, H. Okumura, S. Yoshida, Appl.

Phys. Lett. 69 (1996) 2507.

[41] W.C. Johnson, J.B. Parsons, M.C. Crew, J. Phys. Chem.

36 (1932) 2651.

[42] C.D. Thurmond, R.A. Logan, J. Electrochem. Soc. 119

(1972) 622; R.A. Logan, C.D. Thurmond, J. Electrochem.
Soc. 119 (1972) 1727.

[43] G. Jacob, R. Madar, J. Hallais, Mater. Res. Bull. 11 (1975)

445.

[44] M.R. Lorenz, B.B. Binkowski, J. Electochem. Soc. 109

(1962) 24.

[45] M. Kuball, F. Demangeot, J. Frandon, M.A. Renucci, J.

Massies, N. Grandjean, R.L. Aulombard, O. Briot, Appl.
Phys. Lett. 73 (1998) 960.

[46] C.J. Sun, P. Kung, A. Saxler, H. Ohsato, E. Bigan, M.

Razeghi, D.K. Gaskill, J. Appl. Phys. 76 (1994) 236.

[47] A.S. Bolgar, S.P. Gordienko, E.A. Ryklis, V.V. Fesenko,

in: G.V. Samsonov (Ed.), Chemistry and Physics of the
Nitrides, Naukova, Dumka, Kiev, 1968, p. 151 (in
Russian).

[48] I.G. Pichugin, D.A. Yas’kov, Izv. Akad. Nauk SSSR,

Neorganicheskie Mater. 6 (1970) 1973.

[49] R.C. Schoonmaker, A. Buhl, J. Lemley, J. Phys. Chem. 69

(1965) 3455.

[50] A. Davydov, private communication, 1998.
[51] A. Pisch, R. Schmid-Fetzer, J. Crystal Growth 187 (1998)

329.

[52] T. Kojima, N. Kawai, T. Nakagawa, K. Ohta, T.

Sakamoto, M. Sawashima, Appl. Phys. Lett. 47 (1985)
286.

[53] M. Van Hove, P.I. Cohen, Appl. Phys. Lett. 47 (1985) 726.
[54] Y. Seki, H. Watanabe, J. Matsui, J. Appl. Phys. 49 (1978)

822.

[55] J.R. Arthur Jr., J. Appl. Phys. 39 (1968) 4032.
[56] R.C. Weast (Ed.), CRC Handbook of Chemistry and

Physics, 55th Edition, CRC Press, Cleveland, 1974.

[57] F. Family, P. Meakin, Phys. Rev. Lett. 61 (1988) 428.
[58] G.R. Carlow, R.J. Barel, M. Zinke-Allmang, Phys. Rev. B

56 (1997) 12519.

[59] H. Remy, Treatise on Inorganic Chemistry, Elsevier, New

York, 1960, p. 18.

[60] W.R.S. Garton, Proc. Phys. Soc. A 64 (1951) 509.
[61] M.M. Sung, J. Ahn, V. Bykov, J.W. Rabalais, D.D.

Koleske, A.E. Wickenden, Phys. Rev. B 54 (1996) 14652.

[62] S.W. King, E.P. Carlson, R.J. Therrien, J.A. Christman,

R.J. Nemanich, R.F. Davis, J. Appl. Phys. 86 (1999) 5584.

[63] I. Gorczyca, A. Svane, N.E. Christensen, Phys. Rev. B 60

(1999) 8147.

[64] P.A. Redhead, Vacuum 12 (1962) 203.
[65] E.G. Seebauer, C.E. Allen, Prog. Surf. Sci. 49 (1995) 265.
[66] A.E. Wickenden, D.K. Wickenden, T.J. Kistenmacher, J.

Appl. Phys. 75 (1994) 5367.

[67] L. Sugiura, K. Itaya, J. Nishio, H. Fujimoto, Y. Kokubun,

J. Appl. Phys. 82 (1997) 4877.

[68] X.H. Wu, D. Kapolnek, E.J. Tarsa, B. Heying, S. Keller,

B.P. Keller, U.K. Mishra, S.P. DenBaars, J.S. Speck,
Appl. Phys. Lett. 68 (1996) 1371.

[69] S.D. Hersee, J. Ramer, K. Zheng, C. Kranenberg, K.

Malloy, M. Banas, M. Goorsky, J. Electron. Mater. 24
(1995) 1519.

[70] K. Lorenz, M. Gonsalves, Wook Kim, V. Narayanan, S.

Mahajan, Appl. Phys. Lett. 77 (2000) 3391.

[71] Y. Kobayashi, T. Akasaka, N. Kobayashi, Jpn. J. Appl.

Phys. 37 (1998) L1208.

[72] B.T. McDermott, R. Pittman, E.R. Gertner, J. Krueger, C.

Kisielowski, Z. Lilienthal-Weber, and E. Weber, Talk D2.2
at Fall Material Research Society Meeting, Boston, 1997.

[73] A. Watanabe, H. Takahashi, T. Tanaka, H. Ota, K.

Chikuma, H. Amano, T. Kashima, R. Nakamura, I.
Akasaki, Jpn. J. Appl. Phys. 38 (1999) L1159.

[74] A. Thon, T.F. Kuech, Appl. Phys. Lett. 69 (1996) 55.
[75] O. Briot, S. Clur, R.L. Aulombard, Appl. Phys. Lett. 71

(1997) 1990.

[76] M.A. Khan, R.A. Skogman, R.G. Schulze, and M.

Gershenzon, Appl. Phys. Lett. (1983) 43492.

[77] T. Sasaki, J. Crystal Growth 129 (1993) 81.
[78] T.J. Kistenmacher, D.K. Wickenden, M.E. Hawley, R.P.

Leavitt, Mater. Res. Soc. Symp. Proc. 395 (1996) 261.

[79] O. Scho¨n, B. Schineller, M. Heuken, R. Beccard, J. Crystal

Growth 189/190 (1998) 335.

[80] K. Tadatomo, Y. Ohuchi, H. Okagawa, H. Itoh, H.

Miyake, K. Hiramatsu, MRS Internet J. Nitride Semi-
cond. Res. 4S1 (1999) G3.1.

[81] Y. Morishita, Y. Nomura, S. Goto, Y. Katayama, Appl.

Phys. Lett. 67 (1995) 2500.

[82] Y. Okamoto, S. Hashiguchi, Y. Okada, M. Kawabe, Jpn.

J. Appl. Phys. 37 (1998) L1109.

[83] J. Israelachvili, Intermolecular and Surface Forces, 2nd

Edition, Academic Press, London, 1991, p. 149.

[84] S.W. Benson, The Foundations of Chemical Kinetics,

McGraw-Hill, New York, 1960, p. 72.

[85] Beaumont, P. Gibard, J.P. Faurie, J. Crystal Growth 156

(1995) 140.

D.D. Koleske et al. / Journal of Crystal Growth 223(2001) 466–483

483