4H SiC Epitaxial Growth with Chlorine Addition

Abstract

The growth rate of a 4H-SiC epitaxial layer has been increased by a factor of 19 (up to 112 μm h–1) with respect to the standard process, with the introduction of HCl in the deposition chamber. The epitaxial layers grown with the addition of HCl has been characterized by electrical, optical, and structural characterization methods. The effects of various deposition parameters on the epitaxial growth process have been described, and an explanation of this behavior in terms of the diffusion coefficient on the surface, Ds, and the ratio between the characteristic times, τD:τG, has been provided. The diodes, manufactured on the epitaxial layer grown with the addition of HCl at 1600 °C, have electrical characteristics comparable with the standard epitaxial process. This process is very promising for high-power devices with a breakdown voltage of 10 kV.

Full text

Page 1

DOI: 10.1002/cvde.200506465
Full Paper
4H SiC Epitaxial Growth with Chlorine Addition
By Francesco La Via,* Giuseppa Galvagno, Gaetano Foti, Marco Mauceri, Stefano Leone, Giuseppe Pistone,
Giuseppe Abbondanza, Alessandro Veneroni, Maurizio Masi, Gian Luca Valente, and Danilo Crippa
The growth rate of a 4H-SiC epitaxial layer has been increased by a factor of 19 (up to 112 lm h–1) with respect to the stan-
dard process, with the introduction of HCl in the deposition chamber. The epitaxial layers grown with the addition of HCl
has been characterized by electrical, optical, and structural characterization methods. The effects of various deposition pa-
rameters on the epitaxial growth process have been described, and an explanation of this behavior in terms of the diffusion
coefficient on the surface, Ds, and the ratio between the characteristic times, sD:sG, has been provided. The diodes, manufac-
tured on the epitaxial layer grown with the addition of HCl at 1600°C, have electrical characteristics comparable with the
standard epitaxial process. This process is very promising for high-power devices with a breakdown voltage of 10 kV.
Keywords: Epitaxial growth, 4 H SiC, HCl, Schottky diodes, high-power devices
1. Introduction
Recent developments in SiC substrate production by the
sublimation technique, and epitaxial growth of high-quality
films by CVD, enable the production of suitable materials
for high-power devices.[1]
The homoepitaxial growth of a-SiC has been performed
by liquid phase epitaxy (LPE) and CVD methods. Al-
though CVD has the advantages of precise control and
uniformity of epilayer thickness and impurity doping, the
quality of the epilayers can be affected by polytype mixing.
In 1986 Matsunami et al.[2]found that single crystalline
6H-SiC can be grown homoepitaxially on off-oriented
6H-SiC (0001) at low temperatures (1400–1500°C). This
technique was named “step-controlled epitaxy”, since the
polytype can be controlled by surface steps existing on
off-oriented substrates. This technique was a really break-
through in two senses; a) the growth temperature can be re-
duced more than 300°C, and b) the quality of the resulting
epilayers is very high and suitable for device applications.
–
[*]Dr. F. La Via, Dr. G. Galvagno
CNR-IMM sezione di Catania
Stradale Primosole 50, 95121 Catania (Italy)
E-mail: francesco.lavia@imm.cnr.it
Prof. G. Foti
Physics Department, Catania University
Via S. Sofia 64, 95123 Catania (Italy)
Dr. M. Mauceri, Dr. S. Leone, Dr. G. Pistone, Dr. G. Abbondanza
Epitaxial Technology Center, c/o BIC Sicilia
Pantano d’Arci, 95030 Catania (Italy)
Dr. A. Veneroni, Prof. M. Masi
Chemistry, Materials and Chemical Engineering Department,
Politecnico di Milano
Milano (Italy)
Dr. G. L. Valente, Dr. D. Crippa
LPE
Via Falzarego 8, 20021 Bollate (Mi) (Italy)
Mirror surfaces are obtained, using this process, for C/Si
ratios between 1.4 and 2.5. The growth rate is almost con-
stant with these parameters, and increases proportionally
to the SiH4flow rate. For high flow rates Si droplets are
formed. A remarkable decrease in the growth rate is ob-
served instead for C/Si < 1.4. These results and the analysis
of the gas-phase kinetics in the growth system show that
the growth proceeds through the adsorption of Si in atomic
steps and its carbonization by hydrocarbon molecules. The
main limitation of this process is the low growth rate
(6–7 lm h–1) that is correlated to the slow silicon species
diffusion through the stagnant layer, and to the low Si/H2
ratio that cannot go over 0.05%; otherwise homogeneous
nucleation of silicon droplets in the gas phase occurs. This
last phenomenon causes the depletion of the gas-phase pre-
cursors available for the deposition and the worsening of
the surface quality.
In recent years, there have been an increasing number of
publications on high-voltage devices with a breakdown
voltage around 10 kV. This interest is due to the fact that in
this range the silicon carbide devices do not suffer competi-
tion with silicon power devices. Also, several interesting
papers have been published on power DMOSFETs,[3]im-
planted VJFETs,[4]PiN diodes,[5]and Schottky diodes.[6]In
all these devices, an epitaxial thickness around 80–100 lm
is needed to obtain a breakdown voltage between 10 and
11 kV. To obtain this large thickness with a standard epi-
taxial growth rate of 6–8 lm h–1it is necessary to use a pro-
cess time of more than ten hours with a consequent high
cost of the process.
A new epitaxial process that overcomes this limitation
and will produce a second breakthrough in the epitaxy pro-
cess has recently been developed.[7,8]The growth rate has
been increased with respect to the standard process by in-
creasing the silane flow with the introduction of HCl in the
Chem. Vap. Deposition 2006, 12, 509–515 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
509

Page 2

Full Paper
deposition chamber. With this process a high growth rate
with a good surface morphology has been obtained. The
Schottky diodes realized on these epitaxial layers show
average electrical characteristics very close to the diodes
realized on the epitaxial layers grown with the standard
process without HCl.
In this work, the influence of several deposition parame-
ters on the surface morphology, the optical and electrical
characteristics of the epitaxial layers grown with the HCl
addition are discussed in detail.
2. Results and Discussion
2.1. Electrical Characterization
The current–voltage relationship of a Schottky contact
can be described by the thermoionic emission theory shown
in Equation 1.
I ? AA?T2e?qUB?kTeq?V?IRs??nkT? 1
??
?1?
A is the diode active area, A* is the Richardson constant,
T is the absolute temperature, q is the electron charge, UB
is the Schottky barrier height, k is the Boltzmann constant,
n is the ideality factor, and Rsis the series resistance of the
diode. The forward voltage drop, VF, can be extracted from
Equation 1 to give Equation 2.
VF? IRs?nkT
q
ln
I
AA?exp ?qUB
kT
??
?
?
?
?
?2?
It is clear that, for a fixed ideality factor, n, and a fixed
Schottky barrier height, UB, the forward voltage drop, VF,
depends essentially on the series resistance, Rs, of the
diode. This value is the sum of four different contributions,
as shown in Equation 3.
Rs? Repi? Rsub? Rc? Rsystem
?3?
Repiis the series resistance due to the epitaxial layer, Rsub
is the series resistance relative to the substrate, Rcis the
contact resistance, and Rsystemis the resistance introduced
by the experimental apparatus. Maintaining the same ex-
perimental conditions (substrate characteristics, back con-
tact metallization process, and the measurement system)
different epitaxy quality can be compared. Furthermore,
the series resistance of the epitaxial layer, Repi, can be writ-
ten as Equation 4.
Repi?
Wepi
qNlepiA
?4?
Wepiis the thickness of the epitaxial layer, q the electron-
ic charge, N the doping concentration, lepithe drift mobil-
ity of the epitaxial layer, and A the area of the Schottky
diode. Then, fixing the Schottky diode area A, the epitaxial
layer thickness, Wepi, and the doping concentration of the
epitaxial layer, N, the difference between different VFdis-
tributions can be related to the mobility distribution on the
wafer.
The leakage current, Ileak, of the diodes is related to
the sum of different contributions that are shown in Equa-
tion 5.
Ileak? Itherm? Ifet? Iedge? Igen
?5?
First of all there is the contribution due to the thermoio-
nic mechanism that gives the current Itherm, which can be
obtained from the Equation 1, and shown in Equation 6.
Itherm? ?AA?T2e?q?UB?DUB??kT
?6?
The second contribution is the field-effect leakage cur-
rent, Ifet, which is particularly relevant for high reverse bias
and high doping concentration of the epitaxial layer. An-
other component of the total leakage current is related to
the leakage of the diode edge structure, Iedge. The last com-
ponent is related to the generation current Igendue to the
defects that produce levels close to the mid-gap of 4H-SiC.
Then, fixing the edge structure and considering a low re-
verse voltage with respect to the diode breakdown voltage,
VBD (VR/VBD < 0.2), the comparison between different
leakage distributions is related essentially to a different dis-
tribution of generation defects.
2.2. Si/H2Ratio
In Figure 1, four different optical microscopy images of
the SiC epitaxial surface are reported. In the processes con-
sidered, the Si/H2dilution ratio ranged from 0.01 to 0.6%.
From these images, it can be observed that, while at a low
(0 < Si/H2< 0.05%) dilution ratio (Fig. 1A) the surface
morphology is specular; by increasing the Si/H2 ratio
(0.1%) several silicon droplets appear on the surface
(Fig. 1B) and almost 30% of the surface is covered by
silicon droplets. The introduction of an HCl flux during the
reaction avoids silicon precipitation for the same dilution
ratio of 0.1% (Fig. 1C), and even for a very high silane
concentration (Si/H2=0.6%) (Fig. 1D).
These experimental data reveal that HCl addition allows
an increase in the silane concentration, without the homo-
geneous nucleation in the gas phase that is generally ob-
served in the deposition process without HCl. Without this
limitation, very high growth rates can be obtained. In Fig-
ure 2, the growth rate is reported as a function of the dilu-
tion ratio Si/H2, showing that the growth rate increases
linearly with the silane flux. In particular, a growth rate of
112 lm h–1has been reached with a dilution ratio of 0.6%.
No homogeneous gas-phase nucleation has been observed
510
www.cvd-journal.de© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Vap. Deposition 2006, 12, 509–515

Page 3

Full Paper
even at this very high silane concentration, so we suppose
that higher growth rates can be reached in the future.
The growth rate increase with HCl addition can be corre-
lated to the change in the gas-phase deposition mechanism.
In fact, under standard process conditions, the SiH4decom-
position leads to a substantial formation of gaseous Si with
limited reactivity (i.e., sticking coefficients) with respect to
film growth.[9]On the contrary, the addiction of chlorinated
species in the gas phase leads to the formation of the very
stable and high sticking coefficient radical species of gener-
al formula SiHxCly.[10]That species is unlikely to polymer-
ize at temperatures of interest here, and consequently no
accumulation of poorly reactive mass is observed. More-
over, as pointed out in most of the literature addressing Si
deposition from chlorosilanes,[11]SiHxClyhas very high re-
activity towards the growing film, indicating that those are
almost surely the most important deposition precursors.
This point indicates also that the introduction of other
chlorine-containing species, such as SiCl4, SiHCl3, SiH2Cl2,
and SiCl3CH3, will produce almost the same effect as the
simple HCl addiction, providing that the same Si/Cl/C/H
molar ratios were reproduced, because deposition condi-
tionsapproaching thermodynamics
achieved in epi-SiC processes, such as those here examined.
Accordingly, as a corollary, chlorine could be introduced
also through carbon-containing species such as CCl4,
CHCl3, C2H3Cl, and so on, because the greater affinity of
halogens to silicon instead of to carbon will reproduce the
same species distribution given above.
This growth of epitaxial layers with the new process has
been characterized also by atomic force microscopy
(AFM) to measure the surface roughness, and by KOH
etch at 500°C to measure the number of dislocations.
The average surface roughness, measured in several re-
gions of the wafers, does not depend on the growth rate
and is about 0.3 nm for all the samples. The same values
are also measured for the standard process without HCl.
Comparing the dislocation density of the process with
HCl addition with respect to the standard process without
hydrochloric acid by KOH etch at 500°C, no large differ-
ence can be observed between the two different processes.
In particular the screw dislocation densities are in the range
between 4×102cm–2and 3×103cm–2. The edge dislocation
densities are in the range between 3×104cm–2and 7×104
cm–2. Finally the plane dislocations are between 2×103cm–2
and 6×103cm–2.
equilibrium are
Chem. Vap. Deposition 2006, 12, 509–515© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimwww.cvd-journal.de
511
20 µm
(A)
(B)
(C)
(D)
Fig. 1. Optical microscopy imagesof the SiCsurfaceafter an epitaxial process
A) with a Si/H2ratio between 0 and 0.05%, or B) with a Si/H2ratio equal to
0.1%withoutHCl,andC)withHCl.In(D)theSi/H2ratiohasbeenincreased
to0.6% withHCl,andnosiliconprecipitatecanbeobserved onthesurface.
0.00.10.2 0.3
Si/H2 %
0.40.50.60.7
20
40
60
80
100
120
HCl
No HCl
Silicon precipitates
Growth rate (µm/hr)
Fig. 2. Growth rate versus the Si/H2ratio for both the process with HCl
(squares) and the process without HCl addition. The limit for the homoge-
neous nucleation of silicon precipitates for the process without HCl is re-
ported as well.

Page 4

Full Paper
2.3. Cl/Si and C/Si Ratios
In this process we have a new parameter; the hydrochlo-
ric acid flux that influences the SiC epilayer growth. The
surface roughness and the amount of defects of the epitax-
ial layer are correlated with the Cl/Si ratio. In Figure 3,
three different images of the epitaxial surfaces obtained
with various Cl/Si ratios are shown. At the lowest value
(Cl/Si=0.05) the surface of the epilayer (Fig. 3a) is very
rough on the entire wafer. Increasing the hydrochloric acid
flux by a factor of ten (Cl/Si=0.5), the surface (Fig. 3b) ap-
pears with rough and specularly flat regions. Increasing
even more the amount of chlorine in the reactor (Cl/Si=2),
the surface (Fig. 3c) is flat over the entire wafer.
The Cl/Si ratio has a great relevance also for the defects
present in the epitaxial layers. In Figure 4, three different
photoluminescence spectra measured at room temperature
are reported. It must be pointed out that, increasing the
Cl/Si ratio from 0.05 to 0.5, the large peak between 2000
and 2800 meV decreases close to the background level. In-
creasing this ratio even more, no further differences in the
photoluminescence spectrum can be observed.
Four different wafers were grown with various Cl/Si ra-
tios and with the same Si/H2ratio (0.1%) and deposition
temperature (1600°C). The epitaxial thickness was fixed at
9.5 lm and the doping in the range 2–3×1016cm—3. These
parameters give a theoretical breakdown voltage close to
1000 V. On these wafers more than 500 diodes with an area
of 1 mm2have been realized and tested both in forward
and reverse bias. The average forward drop (VF) and leak-
age current at –200 V (IR) values of the good diodes for
the different wafers are reported in Figure 5.
The forward bias drop has a maximum value of 0.2 A at
a Cl/Si ratio of 0.5, and a decrease of this value for both
Cl/Si=0.05 and Cl/Si=2. The decrease in the voltage drop,
VF, at high current (0.2 A) is related to an increase in the
drift mobility of the epitaxial layer.[12]In particular, using
Equations 2–4, it can be estimated that there is an increase
in the mobility from 28 cm2V–1s–1(Vf=1.8 V) to 700 cm2
V–1s–1(Vf=1.5 V). The reduction in this parameter should
be related to a reduction of point defects (interstitials,
vacancies, anti-site) that have a large impact on the drift
mobility.
512
www.cvd-journal.de© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Vap. Deposition 2006, 12, 509–515
100 µm
(A)
(B)
(C)
Fig. 3. Optical microscopy images of epitaxial layers grown with various Cl/
Si ratios. A) Cl/Si=0.05, B) Cl/Si=0.5, C) Cl/Si=2.
20002400
Energy (meV)
28003200
0.0000
0.0002
0.0004
0.0006
0.0008
0.0010
Cl/Si=0.05
Cl/Si=0.5
Cl/Si=2
Intensity (A.U.)
Fig. 4. Room temperature PL of three epitaxial layers grown with various
Cl/Si ratios.
0.00.5 1.01.5 2.0 2.5
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4

10
Vf @ If=0.2 A (V)
Cl/Si
10
-7
-6
10
-5
Ileakage @ -200 V (A)
Fig. 5. Average Vf(squares) and Ileakage(triangles) versus the Cl/Si ratio.
The better electrical results were obtained for a Cl/Si ratio of 2.

Page 5

Full Paper
The leakage current at –200 V (right axes) decreases
when increasing the Cl/Si ratio. This trend is related to the
reduction of defects that introduce levels close to the
4H-SiC mid-gap. In fact, from deep-level transient spec-
troscopy (DLTS) measurements (not reported here) it is
clear that, increasing the Cl/Si ratio, the main effect on the
deep levels is the reduction of the EH6/7trap concentra-
tion. In fact this level reduce its concentration from
3.5×1013cm–3for Cl/Si=0.5 to 7×1012cm–3for Cl/Si=2.
This level has been related in the literature to the presence
of carbon vacancies or carbon–silicon di-vacancies,[13,14]
and produces a large effect on both carrier mobility and
leakage current. Increasing the Si/H2ratio, i.e., increasing
the growth rate (see Fig. 2), both the Cl/Si ratio and the
C/Si ratio should be changed to obtain a perfect specular
surface. In fact, the Cl/Si ratio should be decreased from a
value of 2 to 1 going from a dilution ratio of 0.1% to 0.6%
(Fig. 6). Even the C/Si ratio should be decreased from 1.5
to 1 and finally to 0.8 for the highest dilution ratios.
The understanding of the evolution of surface morphol-
ogy during a CVD process can be obtained by the simple
comparison of two characteristics times that are inherent;
the adatoms surface diffusion (sD=L2/16Ds) and the mat-
ter supply to the surface (sG=Ns/qG) phenomena. In the
above definitions L, Ds, q, Ns, and G are the terrace length,
surface diffusivity, SiC molar density, surface site density,
and film growth rate, respectively. To obtain a good quality
epitaxial film the ratio between the two characteristic times
has to be much lower than unity.[10,13]Under these condi-
tions, an adatom can be inserted in a kink before the supply
of a new one from the deposition reactions. Thus, at the
high temperatures typical of SiC growth, the morphology is
substantially controlled by the slower moving atom on the
surface, while the matter supply to the surface is substan-
tially controlled by the transport of precursors from the gas
phase (i.e., G ≈ kC/q; k and C being the mass transport
coefficient and the gas-phase main precursor concentra-
tion, respectively).
Although the growth process is at high temperature
(T=1550–1650°C), the surface diffusivity of silicon and
carbon atoms is very different due to the difference in the
bond energies. In fact, a rough estimation of the surface
diffusivity lead to Ds=ma0
lattice vibration frequency, the lattice parameter, and the
diffusion activation energy, respectively. Roughly, this last
valueapproacheshalfof
Ed=0.5 Eb). For SiC, the corresponding values are 83, 54,
and 72 kcal mol–1for C-C, Si-Si, and Si-C, respectively.
Thus it is evident that carbon adatoms are the least mobile
species on the surface, and the ratio between the surface
diffusivities of carbon and silicon approaches 0.02 at
1650°C, the two diffusion characteristic times for carbon
and silicon being 670 ns and 15 ns, respectively. Accord-
ingly, in the following, the carbon adatom diffusivity will be
considered as the reference value. These values are almost
insensitive to the precursor inlet concentrations and pres-
sure, whilst being sensitive to temperature.
When chlorine-containing precursors are added to the
inlet mixture (e.g., HCl, SiHCl3,???) the first result is the
conversion of vapor Si to SiHxClystable species, thereby
avoiding homogeneous nucleation. Thus, higher Si/H2inlet
ratios can be safely reached with the immediate result of a
significant increase in the growth rate. However, in terms
of surface morphology, if the C/Si ratio is kept constant, it
correspond to a crystal-quality decrement because sD:sGin-
creases. To assure again mirror-like surfaces, it is necessary
to reduce the amount of carbon on the surface by a reduc-
tion in both the C/Si and the Cl/Si ratios. In fact, a higher
silicon removal from the surface is induced by the increase
in the Cl/Si ratio, thus shifting again the surface C/Si ratio
in favor of carbon. In practice, the process design can be
performed by searching for as high a growth rate as possi-
ble while maintaining the sD:sGratio well below unity (i.e.,
about 0.01).
2e–Ed/RT; m, a0, and Edbeing the
thebondstrength (i.e.,
2.4. Deposition Temperature
The deposition temperature also has a great influence on
the quality of the epitaxial layer deposited with the addi-
tion of HCl. In fact, in Figure 7 the room temperature
photoluminescence spectra of three different epitaxial
layers are reported. The spectrum of the epitaxial layer de-
posited at 1550°C shows a large peak in the region between
2100 and 2800 meV. This peak has been related to the for-
mation of stacking faults in the epitaxial layer (see trans-
mission electron microscope (TEM) cross-section in the in-
set). Increasing the deposition temperature to 1600°C, this
peak disappears and the photoluminescence spectrum is
very similar to the spectrum obtained on an epitaxial layer
grown with an optimized standard process without HCl ad-
dition. Increasing the deposition temperature to 1650°C,
no large difference can be observed in the photolumines-
cence spectrum.
A considerable difference between the deposition pro-
cess at 1600°C and the epitaxial growth at 1650°C can be
Chem. Vap. Deposition 2006, 12, 509–515© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimwww.cvd-journal.de
513
0.00.10.20.30.40.50.60.7
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
Cl/Si
Si/H2 (%)
0.6
0.8
1.0
1.2
1.4
1.6
Tdep=1650 ºC
Specular surface
C/Si
Fig. 6. Cl/Si (circles) and C/Si (squares) ratios versus the Si/H2ratio to ob-
tain a specular surface at a deposition temperature of 1650°C.

Page 6

Full Paper
observed, instead, from the electrical characteristics of
Schottky diodes both in forward (Fig. 8) and in reverse bias
(Fig. 9). In fact, the statistical distributions of the forward
voltage drop at a forward current of 0.2 A show that the
process with a deposition temperature of 1650°C has a low-
er voltage drop and a lower dispersion of the electrical
characteristics with respect to the epitaxial layer grown at
1600°C. This behavior suggests that the increase in the de-
position temperature decreases the amount of defects that
reduce the mobility of the epitaxial layer.
The increase of the deposition temperature to 1650°C
also produces a reduction of the average leakage current,
as reported in Figure 9. In fact, the average leakage current
decreases from 3.08 × 10–6A at a deposition temperature
of 1600°C, to 2.0 × 10–7A at a deposition temperature of
1650°C. These leakage current results can be related to the
energy levels introduced in the energy gap determined by
DLTS (not reported here). In fact, these measurements
show that increasing the deposition temperature from
1600°C to 1650°C, the main deep level (EH6/7) present in
the epitaxial layer reduces its concentration by an order of
magnitude from 3.5×1013cm–3to 3.2×1012cm–3. Then the
reduction in the leakage current is related to the reduction
in the generation current due to the level EH6/7that have a
distance from the conduction band of 1.6 eV. This defect
(carbon vacancy or carbon-silicon di-vacancy) should also
have a large scattering cross-section for the carriers and
then strongly influence the drift mobility.
All these effects should be related to the increase in the
diffusion coefficient on the surface, Ds, that reduce the
crystallographic defects present in the epitaxial layer and
improve the electrical characteristics of the Schottky di-
odes.
3. Conclusions
In this paper, the epitaxial process with chlorine addition
has been described and the effects of various deposition pa-
rameters has been reported in detail. With this process, a
very high growth rate (up to 112 lm h–1) can be reached
with good morphological and electrical characteristics. The
process window of this process at high growth rate is nar-
rower with respect to the standard process because the dif-
fusion coefficient on the surface, Ds, starts to be the limit-
ing factor, and the ratio between the time to diffuse on the
surface, sD, and the time to supply the matter to the sur-
face, sG, increases. This process can give thick epitaxial
514
www.cvd-journal.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Vap. Deposition 2006, 12, 509–515
2200 2400 2600 2800 3000 3200
Energy (meV) Energy (meV)
0.00.0
2.0×10 2.
-3-3
4.0×104.
-3-3
6.0×106.
-3 -3
8.0×10 8.
-3 -3
1.0×101.
-2-2
without HCl
HCl 1550
HCl 1600 HCl 1600 ºC
Intensity (A.U.)
2200 2400 2600 2800 3000 3200
10
10
10
10
10
without HCl
HCl 1550 ºC
Intensity (A.U.)
Fig. 7. Room temperature PL of three epitaxial layers grown with various
experimental conditions; the process without HCl addition, the process with
HCl and a deposition temperature of 1550°C, and the process with HCl and
a deposition temperature of 1600°C. In the inset, a TEM cross section show-
ing stacking faults is displayed.
0.0
0.6
0.1
0.2
0.3
0.4
0.5
0.6
1650 ºC
1.5
Vf @ If=0.2 A (V)
1.82.12.4
0.0
0.1
0.2
0.3
0.4
0.5
1600 ºC
Normalized yield (%)
(A)
(B)
Fig. 8. Statistical distributions of Vf@If=0.2 A for diodes with an epitaxial
layer grown A) at 1650°C, and B) at 1600°C.
-9-9-8-8 -7-7 -6-6 -5-5-4-4
0.0
101010
0.20.2
0.40.4
0.60.6
(B)(B)
Normalized Yield (%)
0.00.0
0.20.2
0.40.4
0.60.6
(A)(A)
1650 ºC
10 10
Ir@ –200 V (A)
10 10 1010 10 1010 10 10
0.0
1600 ºC
Normalized Yield (%)
10 1010 10
Fig. 9. Statistical distributions of Ir@ –200 V for diodes with an epitaxial
layer grown A) at 1650°C and B) at 1600°C.

Page 7

Full Paper
layers (80–100 lm) of good crystal quality and with a low
cost of the process for high-power devices with a break-
down voltage of 10 kV.
4. Experimental
The epitaxial films were grown in a hot-wall reactor (built by LPE
Epitaxial Technology) that can grow up to six 2 inch wafers and three 3 inch
wafers at the same time. The chamber was optimized to have a superior tem-
perature uniformity, to reduce the temperature ramp-up and ramp-down,
and particulate formation. The substrates were 4H-SiC (0001), Si face,
n-type (≅1018cm–3) off-axis (≅8° off towards the ?11?20? direction). These wa-
fers were loaded into the reactor and the system was pumped down until a
pressure of ≅10–5Torr was achieved. The growth begins with a hydrogen
etch sequence. At the deposition temperature the precursors (SiH4, HCl,
and C2H4) are introduced into the hydrogen carrier gas and the growth
starts. The epitaxial layers were grown with various Si/H2and Si/Cl ratios,
and the obtained layers were analyzed by Fourier transform infrared (FTIR)
reflectance for the thickness determination, mercury-probe C–V measure-
ments for the doping concentration, AFM for surface roughness analysis,
photoluminescence (PL), X-ray diffraction (XRD) rocking curve, and chem-
ical etch in molten KOH for defects quantification and distribution. Further-
more, PL was also used to detect the presence of 6H and 3C inclusions in
the deposited layers.
Several Schottky diodes with different contact areas were manufactured
on these wafers with a simple process using Ni2Si as the Schottky barrier
and an implanted edge termination. The thickness of the epitaxial layers
was fixed at 9.5 lm, and the dopant concentration at 2×1016cm–3. With
these parameters, a breakdown voltage of about 1000 V can be reached.
These diodes were characterized by I–V and C–V maps on the entire wafer
to obtain statistical information and spatial distribution of defects.
On selected diodes, DLTSwas usedto detect the levels introducedinto the
energygapand tosee thecorrelationswith thereverseI–Vcharacteristics.
Received: December 27, 2005
Final version: June 8, 2006
–
[1] A. Elasser, T. Paul Chow, Proc. IEEE, 2002, 90, 969.
[2] T. Ueda, H. Nishino, H. Matsunami, J. Cryst. Growth 1990, 104, 695.
[3] S. H. Ryu, S. Krishnaswami, M. O’Loughlin, J. Richmond, A. Agarwal,
J. Palmour, A. R. Hefner, IEEE Electron Device Lett. 2004, 25, 556.
[4] J. H. Zhao, P. Alexandrov, J. H. Zhang, X. Q. Li, IEEE Electron De-
vice Lett. 2004, 25, 474.
[5] M. K. Das, J. J. Sumakeris, B. A. Hull, J. Richmond, S. Krishnaswami,
A. R. Powell, Mater. Sci. Forum 2005, 483–485, 965.
[6] J. H. Zhao, P. Alexandrov, X. Li, IEEE Electron Device Lett. 2003, 24,
402.
[7] D. Crippa, G. L. Valente, A. Ruggiero, L. Neri, R. Reitano, L. Calcag-
no, G. Foti, M. Mauceri, S. Leone, G. Pistone, G. Abbondanza, G. Ab-
bagnale, A. Veneroni, F. Omarini, L. Zamolo, M. Masi, F. Roccaforte,
F. Giannazzo, S. Di Franco, F. La Via, Mater. Sci. Forum 2005, 483–
485, 67.
[8] R. L. Myers, O. Kordina, S. Rao, R. Everly, S. E. Saddow, Mater. Sci.
Forum 2005, 483–485, 73.
[9] A. Veneroni, F. Omarini, D. Moscatelli, M. Masi, S. Leone, M. Mau-
ceri, G. Pistone, G. Abbondanza, J. Cryst. Growth 2005, 275, e295.
[10] G. Valente, C. Cavallotti, M. Masi, S. Carrà, J. Cryst. Growth 2001,
230, 247.
[11] C. Cavalotti, M. Masi, in Semiconductors and Semimetals (D. Crippa,
D. L. Rode, M. Masi), Academic Press, San Diego 2001, Chap. 2, p.51.
[12] F. La Via, G. Galvagno, F. Roccaforte, A. Ruggiero, L. Calcagno,
Appl. Phys. Lett. 2005 87, 142 105.
[13] A. Pimpinelli, I. Villain, Physics of Crystal Growth, Cambridge Uni-
versity Press, Cambridge, UK 1998.
[14] K. Danno, T. Kimoto, H. Matsunami, Appl. Phys. Lett. 2005, 86, 122
104.
______________________
Chem. Vap. Deposition 2006, 12, 509–515 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimwww.cvd-journal.de
515