R&D Review of Toyota CRDL, Vol.43 No.

1 (2012) 53-59 53

Research Report
Effect of Defects on the Electrical Properties of 4H-SiC Schottky Barrier
Diodes
Takashi Katsuno, Yukihiko Watanabe and Tsuyoshi Ishikawa
Report received on Jan. 25, 2012

The effect of defects on the electrical properties of 4H-SiC Schottky barrier diodes
(SBDs) was investigated. Nanoscaled circular cone-shaped pits (nanopits) were observed at the leakage
current sources of 4H-SiC SBD using atomic force microscopy (AFM). The leakage currents were generated
due to the concentration of electric fields at the nanopits during measurement of the reverse bias
characteristics. The positions of the nanopits correspond to the positions of threading dislocations (TDs),
which were identified from molten potassium hydroxide (KOH) etching. The most important factor of
leakage current generation was determined to be the surface morphologies of the TDs rather than their
presence. Furthermore, there is no difference in the sensitivities on the leakage currents between threading
screw dislocations (TSD) and threading edge dislocations (TED), because both nanopit shapes are almost
the same. As a result, the leakage currents of 4H-SiC SBDs were proportional to the etch pit density (EPD)
obtained from both TSDs and TEDs.

SiC, Schottky Barrier Diodes (SBDs), Defect, Nanopit, Surface Morphology,

Etch Pit Density, Molten KOH Etching

1. Introduction of 4H-SiC SBDs.(8,9) Different sensitivities to other

surface defects, such as triangular defects in 4H-SiC
(13,14)
Silicon carbide (SiC) is a suitable material for use in epilayers, have been reported by many groups, but
next-generation power devices, because it has superior these defects mainly cause an increase of the leakage
thermal, mechanical, and electrical properties compared current in the reverse bias characteristics of 4H-SiC
to silicon (Si).(1-3) The high breakdown field of SiC SBDs.
material enables high power densities to be realized in Furthermore, various sensitivities of the reverse
SiC devices. Consequently, SiC power devices are characteristics to threading dislocations (TDs) have
(13-19)
highly suitable for use in a variety of applications, such also been reported. Some reports have noted the
(13-16)
as power conversion in electric vehicles (both fully influence of TDs on the leakage current, and
cell-powered and hybrid vehicles). SiC power devices contrasting reports have not noted such an influence
(18,19)
require several-hundred-ampere diodes with low by TDs. We suggest that the surface morphologies
specific on-resistance. In addition, high reliability and at the leakage current sources of 4H-SiC diodes are
high reproducibility of the electrical properties are changed, and this is a possible reason for the different
required. The advantages of 4H-SiC Schottky barrier sensitivities. Step bunching on SiC surfaces is known
(20,21)
diodes (SBDs) are high switching speeds and low to cause an increase of the leakage current;
switching losses due to low surge currents during high therefore, it is suspected that unusual surface
frequency operation, and many groups have investigated morphology may be observed at the leakage current
4H-SiC SBDs with high electrical performance.(4-7) sources.
However, dispersion of the electrical properties in 4H- To confirm the change of surface morphology at
SiC SBD diodes, ranging from low blocking voltage leakage current sources, the leakage current sources of
to high leakage current, is observed and is mainly 4H-SiC SBDs were determined using emission
caused by the influences of surface and crystalline microscopy, and their surface morphologies were
defects.(8-12) investigated using atomic force microscopy (AFM).
Micropipes and downfalls are generally the cause of Molten KOH etching was then performed to assist
low blocking voltage in the reverse bias characteristics identification of crystalline defects and to determine

© Toyota Central R&D Labs., Inc. 2012 [Link]

54 R&D Review of Toyota CRDL, Vol.43 No.1 (2012) 53-59

the presence of TDs at the leakage current sources. The the leakage current sources was determined.
etch pit density (EPD) of 4H-SiC diodes on crystalline Crystalline defects were transformed into etch pits
defects by molten KOH etching were investigated to by molten KOH etching at 500 °C for 5 min, which
clarify the relationship between the dispersed reverse were then observed using laser microscopy (Keyence
characteristics of 4H-SiC SBDs and the crystalline VK-9710).
defects.
3. Results and Discussion
2. Experimental Methods
Typical reverse characteristics of 4H-SiC SBDs are
A 4° off-axis n-type 4H-SiC repeated a-face (RAF) shown in Fig. 1. The blocking voltage of the 4H-SiC
substrate with a 13 mm thick epitaxial layer doped with SBD was kept over 1200 V and the leakage level was
15 –3 –3 2
5.0×10 cm of nitrogen was used to fabricate 4H- less than 10 A/cm at 1200 V. It was evident that the
SiC SBDs. A field-limiting ring (FLR) was used as a reverse properties were dispersed. The leakage current
(22)
termination structure, and the breakdown voltage was seems to follow a thermionic field emission model.
designed to be approximately 1400 V. The SiC surface In this device, the Schottky barrier height FB is 1.1 eV
was chemically cleaned by immersion in H2SO4/H2O2 calculated from the forward characteristic, and the
solution for 10 min, and Al was implanted at a dose of electric field strength at 1200 V is 1.5 MV/cm.
19 –3
1×10 cm to complete the FLR structure. The FLR Figure 2 shows an emission microscope image of
structure was then annealed at 1600 °C for 30 min to the 4H-SiC SBD, of which the sample is the middle
remove damage due to the creation of defects by Al leakage current level shown in Fig. 1. Six emission
implantation. Ti/Ni layers were deposited on the current points with various intensities (determined by
backside and annealed at 900 °C for 30 min to ensure
the formation of an ohmic contact. A 100 nm thick, 400
mm diameter Mo layer was deposited on the surface as
a Schottky electrode and an Al electrode (4000 nm
thick) was deposited on the Mo electrode. The reverse
bias characteristics of the 4H-SiC SBD were measured
to evaluate the electrical properties. An emission
microscope (Hamamatsu Photonics PHEMOS-1000)
was used to determine the leakage current sources of
the 4H-SiC SBD. A cooled CCD image sensor with a
sensitivity range of 300 to 1100 nm was equipped in
the emission microscope. The leakage current
(emission current) of 4H-SiC SBD was observed
through the partially-polished backside electrode using Fig. 1 Reverse bias characteristics of 4H-SiC SBDs
the CCD image sensor. A reverse voltage of 1200 V measured at room temperature.
was applied between the surface and backside. Visual
markings with points approximately 10 mm apart from
the sources of the emission currents were applied using
a focused ion beam (FIB) from the surface side to
assist location of the leakage current sources. After
removal of the surface electrodes by H2SO4/H2O2
solution for 10 min, the surface morphologies of the
leakage current sources around the visual markings
were measured by AFM (Veeco Nanoscope V); the
observation areas for leakage current sources were
2 mm squares. After measuring the surface morphologies
of the leakage current sources, the presence of Fig. 2 Emission microscope images of 4H-SiC SBD at
1200 V. (a) View of the entire electrode.
crystalline defects (e.g., TSDs, TEDs, and basal plane
(b) Enlarged view of emission points (i), (ii) and
dislocations (BPDs)) corresponding to the positions of (iii).

© Toyota Central R&D Labs., Inc. 2012 [Link]

 R&D Review of Toyota CRDL, Vol.43 No.1 (2012) 53-59 55

the contrast), as indicated by the circles in Fig. 2(a), pits caused by molten KOH etching. The shapes of the
were detected in the 400 mm diameter device. Strong nanopits at positions (i) and (ii) are very similar. The
intensity (contrast) indicates a high leakage current. positions of the leakage current sources correspond to
Figure 2(b) shows an enlarged view of 3 emission the positions of the nanopits. The maximum field
current points, where the intensities of positions (i) and strength was applied at the interface of the SiC surface
(ii) are almost the same, and the intensity of position and Mo (Schottky metal) during the reverse bias
(iii) is weaker than those of (i) and (ii). The visual measurement due to the presences of the pits, and the
markings produced by FIB are visible at the upper electric fields concentrate at the nanopits; therefore,
edge of Fig. 2(b). leakage currents are generated at the nanopits.
AFM was used to investigate the surface morphologies Figure 4(a) shows simulation results of nanopits (dpit
at the leakage current sources. Figures 3(a) and (b) = 50 nm, αpit = 120°) on the SiC surface, where the pits
show the surface morphologies of the leakage current cause an increase of the reverse leakage current in the
sources at positions (i) and (ii) shown
in Fig. 2, respectively. The different
color contrast indicates that the
surface morphologies of leakage
current sources (i) and (ii) are not flat.
To confirm the shapes of the leakage
current sources, line profiles measured
between the arrows shown in Figs.
3(a) and (b) are presented in Figs. 3(c)
and (d), respectively. Very small
circular cone-shaped pits are evident
at the leakage current sources. The pit
at position (i) is 240 nm wide and 48
nm deep with a base angle of 136°, as
shown in Fig. 3(c). The pit at position
(ii) is 220 nm wide and 44 nm deep
with a base angle of 136°, as shown in
Fig. 3(d). In contrast, the pit at
position (iii) is 230 nm wide and 20
nm deep with a base angle of 160°.
These pits are nanosized and are Fig. 3 AFM images of leakage current sources (i) and (ii). Surface morphology
therefore referred to as nanopits. It of (a) position (i), and (b) position (ii). Line profiles of (c) the nanopit in
should be noted that these are not etch (a) and (d) the nanopit in (b).

–
Fig. 4 (a) Comparison of reverse I-V curves with (αpit = 120°, dpit = 50 nm, n-EPD: 35400 cm 2) and without etch
pits (flat surface). (b) Contour plot of electric field at VR = 1200 V around the etch pit.

© Toyota Central R&D Labs., Inc. 2012 [Link]

56 R&D Review of Toyota CRDL, Vol.43 No.1 (2012) 53-59

reverse voltage.(23) This leakage current is due to positions (i) and (ii) in Fig. 2(b) and the depth of the
electron tunneling current at the bottom of the nanopit. nanopit at position (iii) is only 20 nm. Accordingly, the
This tunneling current is caused by the electric field leakage current at position (iii) is lower than that at
enhancement at the bottom of nanopit. The electric positions (i) and (ii), due to the lower electric field than
field at the metal-semiconductor interface (Schottky that at positions (i) and (ii).
junction) near the bottom of pits is four times greater To confirm the surface morphologies of the leakage
than that of a flat surface area with application of current sources, different leakage current sources were
reverse voltage VR = 1200 V, as shown in Fig. 4(b). also measured by AFM. Surprisingly, nanopits were
Figure 5 shows the simulation result for the impact present at all emission points (total: n=24) in some 4H-
of the nanopit shape (depth dpit, and opening angle αpit) SiC SBDs. The distributions of the nanopit depth,
(23)
on the leakage current. Although the leakage current width and base angle are summarized in Figs. 6(a)-(c),
(24)
is sensitive to both dpit and αpit, no significant increase respectively. TEDs (n=22/24) and TSDs (n=2/24)
in the leakage current is observed in the depth range are represented by the hatched and gray fills,
of less than 20 nm. There is a good relationship respectively. There is no difference in the nanopit
between the emission intensities and the nanopit depths shapes of the TSDs and TEDs. Thus, the sensitivities
at positions (i) and (ii); both the depths of the nanopits of TSD and TED to the leakage current are almost the
and the emission intensities are also the same. The same. As proposed, the presence of nanopits is an
intensity of position (iii) is weaker than that at important factor in the increase of the leakage current.
The relationship between the positions of the
nanopits and crystalline defects was confirmed by laser
microscopy observation of the etch pits formed by
molten KOH etching. Six etch pits denoted by circles
are evident in Fig. 7(a), the positions of which
correspond to the emission points shown in Fig. 2(a).
Figure 7(b) shows an enlarged view of the 3 etch pits
with the same field of view as that in Fig. 2(b). The
positions of the nanopits (equal to the leakage current
sources) also correspond to the positions of TDs. From
the shapes of the etch pits, the crystalline defects at
positions (i) and (iii) are identified as TEDs and that
at position (ii) appears to be a TSD. TSDs are generally
more sensitive than TEDs with respect to the leakage
(12,13,15)
Fig. 5 Dependence of leakage current IR (at VR = 1200 V) current; however, this is not in agreement with
on the opening angle αpit and depth dpit of etch pits. the results, because the sensitivities of TSDs and TEDs
–
(n-EPD: 1.27×106 cm 2). to the leakage current are almost the same. Figure 4(b)

Fig. 6 Distribution of nanopit (a) depth, (b) width and (c) base angle in 4H-SiC SBDs.

© Toyota Central R&D Labs., Inc. 2012 [Link]

 R&D Review of Toyota CRDL, Vol.43 No.1 (2012) 53-59 57

shows that the TDs at positions (i) and (ii) are different, The 4H-SiC SBDs samples investigated by the former
although the emission intensities are almost the same, groups had nanopits (corresponding to the position of
as shown in Fig. 2(b). Therefore, the leakage current TDs), which generate the leakage current. In contrast,
is not dependent on the type of crystalline defect (TSD the samples of the latter groups did not have nanopits
or TED), but is dependent on the presence of nanopits (or the surfaces were extremely flat), and therefore no
and their respective depths. Therefore, the factor of leakage current was generated, and no dependence on
leakage current generation is the surface morphology TDs was evident.
of the TD rather than the presence and type of TD. In this study, Mo was used as a Schottky barrier
The relationship between leakage current and EPD metal for the 4H-SiC SBD. The leakage currents
calculated from the medium and small etch pits formed caused at the nanopits would be influenced by different
from TSDs and TEDs, respectively, were investigated. Schottky barrier metals such as Ti, Ni, and Au.
BPDs, which produce shell pits by molten KOH However, the nanopits of the 4H-SiC PN diodes would
etching, were not observed in the samples under not influence the leakage currents, because the position
analysis; thus, BPDs are not discussed in this paper. of maximum electric field strength is the PN junction
The numbers of etch pits (both TSDs and TEDs) in the formed below the SiC surface and not the nanopits on
area of interest were counted. Figure 8 shows the the SiC surface. Therefore, the increase of leakage
relationship between the EPD and the leakage current currents caused at nanopits would be dependent on the
at 1200 V. The leakage current increases at an almost structures (SBD, junction barrier Schottky (JBS) and
3 4 –2
constant rate with the EPD from 10 to 10 cm . Based PN diodes).
on this trend, the etch pits from TDs are considered to
–8
increase the leakage current by approximately 10 4. Conclusions
2
A/cm . Due to the content of nanopits in these samples,
the leakage current increases with increasing EPD. In summary, the leakage current sources of 4H-SiC
The creation of nanopits is influenced by the SBDs were analyzed using AFM. The leakage current
presence of TDs; however, it is unclear why the sources have nanoscaled circular cone-shaped pits
nanopits are created. TDs are crystalline defects, and (nanopits). The electric field at the nanopits is
the top of TDs at SiC surfaces can be easily damaged concentrated during measurement of the reverse bias
by thermal conditions due to the weak bonds at TDs. characteristics and leakage currents are generated. The
Therefore, it is possible that the nanopits are present positions of the nanopits correspond to the positions
in a bare epi-wafer or are produced during device of TD etch pits, which were determined by molten
processing. It has been reported that TDs do influence KOH etching. The important factor in the generation
(13-16)
the leakage current, yet other reports suggest that of the leakage current is the surface morphology of the
there is no influence of TDs on the leakage TD rather than the presence and type of TD. Therefore,
(18,19)
current. We propose that both results are correct. the sensitivities of TSD and TED on the leakage

Fig. 7 Laser microscope image of 4H-SiC SBD after molten

KOH etching. (a) View of the entire electrode.
(b) Enlarged view of emission points (i), (ii) and (iii) in Fig. 8 Relation between EPD and the leakage
the same field of view as that shown in Fig. 2(b). current of 4H-SiC SBDs at 1200 V.

© Toyota Central R&D Labs., Inc. 2012 [Link]

58 R&D Review of Toyota CRDL, Vol.43 No.1 (2012) 53-59

current are almost the same. Microelectron. Reliab., Vol.48 (2008), pp.1664-1668.
(13) Wahab, Q., et al., “Influence of Epitaxial Growth and
Substrate-induced Defects on the Breakdown of
Acknowledgements 4H-SiC Schottky Diodes”, Appl. Phys. Lett., Vol.76
(2000), 2725.
The authors would like to thank H. Fujiwara, (14) Berechman, R. A., et al.,“Electrical and Structural
M. Konishi, H. Naruoka, J. Morimoto, A. Adachi, and Investigation of Triangular Defects in 4H-SiC
T. Ohnishi of Toyota Motor Corporation and T. Junction Barrier Schottky Devices”, J. Appl. Phys.,
Morino, T. Endo, T. Yamamoto, and T. Yamamoto of Vol.105 (2009), 074513.
DENSO CORPORATION for helpful discussions. (15) Izumi, S., et al., “Analysis for Structural Defects in
the 4H-SiC Epilayers and Their Influence on
This study received the corporation and technical Electrical Properties”, Mater. Sci. Forum, Vol.457-
support of Y. Kimoto and K. Kataoka at the Toyota 460 (2004), pp.1085-1088.
Central R&D Labs. (16) Katsuno, T., et al., “Effects of Surface and Crystalline
Defects on Reverse Characteristics of 4H-SiC
Junction Barrier Schottky Diodes”, Jpn. J. Appl. Phys.,
References Vol.50 (2011), 04DP04.
(1) Ruff, M., et al., “SiC Devices – Physics and (17) Katsuno, T., et al., “Analysis of Surface Morphology
Numerical-simulation”, IEEE Trans. Electron at Leakage Current Sources of 4H-SiC Schottky
Devices, Vol.41 (1994), pp.1040-1054. Barrier Diodes”, Appl. Phys. Lett., Vol.98 (2011),
(2) Kimoto, T., et al., “Performance Limiting Surface 222111.
Defects in SiC Epitaxial p-n Junction Diodes”, IEEE (18) Morisette, D. T. and Cooper, J. A., “Impact of
Trans. Electron Devices, Vol.46 (1999), pp.471-477. Material Defects on SiC Schottky Barrier Diodes”,
(3) Casady, J. B. and Johnson R. W., “Status of Silicon Mater. Sci. Forum, Vol.389-393 (2002), pp.1133-
Carbide (SiC) as a Wide-bandgap Semiconductor for 1136.
High-temperature Applications: A Review”, Solid- (19) Saitoh, H., et al., “Origin of Leakage Current in SiC
State Electron., Vol.39 (1996), pp.1409-1422. Schottky Barrier Diodes at High Temperature”,
(4) Singh, R., et al., “SiC Power Schottky and PiN Mater. Sci. Forum, Vol.457-460 (2004), pp.997-1000.
Diodes”, IEEE Trans. Electron Devices, Vol.49 (20) Syvajarvi, M., et al., “Step-bunching in SiC Epitaxy:
(2002), pp.665-672. Anisotropy and Influence of Growth Temperature”,
(5) Tumakha, S., et al., “Defect-driven Inhomogeneities J. Cryst. Growth, Vol.236 (2002), pp.297-304.
in Ni/4H-SiC Schottky Barriers”, Appl. Phys. Lett., (21) Guy, O. J., et al., “Improved Schottky Contacts to
Vol.87 (2005), 242106. Annealed 4H-SiC Using a Protective Carbon Cap:
(6) Ewing, D. J., et al., “Inhomogeneities in Ni/4H-SiC Investigated Using Current Voltage Measurements
Schottky Barriers: Localized Fermi-level Pinning by and Atomic Force Microscopy”, Diamond Relat.
Defect States”, J. Appl. Phys., Vol.101 (2007), Mater., Vol.15 (2006), pp.1472-1477.
114514. (22) Hatakeyama, T., et al., “Optimum Design of a SiC
(7) Skromme, B. J., et al., “Electrical Characteristics of Schottky Barrier Diode Considering Reverse Leakage
Schottky Barriers on 4H-SiC: The Effects of Barrier Current due to a Tunneling Process”, Mater. Sci.
Height Nonuniformity”, J. Electro. Mater., Vol.29 Forum, Vol.433-436 (2002), pp.831-834.
(2000), pp.376-383. (23) Ishikawa, T., et al., “Critical Density of Nanoscale
(8) Neudeck, P. G., “Electrical Impact of SiC Structural Pits for Suppressing Variability in Leakage Current of
Crystal Defects on High Electric Field Devices”, a SiC Schottky Barrier Diode”, Mater. Sci. Forum,
Mater. Sci. Forum, Vol. 338-342 (2000), pp.1161- (2012), in print.
1166. (24) Katsuno, T., et al., “Surface Morphology of Leakage
(9) Kamata, I., et al., “Electroluminescence Analysis of Current Source of 4H-SiC Schottky Barrier Diode by
Al+ and B+ Implanted pn Diodes”, Mater. Sci. Atomic Force Microscopy”, Mater. Sci. Forum,
Forum, Vol.389-393 (2002), pp.1137-1140. (2012), in print.
(10) Fujiwara, H., et al., “Characterization of In-grown
Stacking Faults in 4H-SiC (0001) Epitaxial Layers Figs. 2, 3 and 7
and Its Impacts on High-voltage Schottky Barrier Reprinted from Appl. Phys. Lett., Vol.98 (2011), 222111,
Diodes”, Appl. Phys. Lett., Vol.87 (2005), 051942. Katsuno, T., Watanabe, Y., Fujiwara, H., Konishi, M.,
(11) Lee, K.-Y. and Capano, M. A., “The Correlation of Naruoka, H., Morimoto, J., Morino, T. and Endo, T.,
Surface Defects and Reverse Breakdown of 4H-SiC Analysis of Surface Morphology at Leakage Current
Schottky Barrier Diodes”, J. Electron. Mater., Vol.36 Sources of 4H-SiC Schottky Barrier Diodes, © 2011 AIP,
(2007), pp.272-276. with permission from American Institute of Physics.
(12) Grekov, A., et al., “Effect of Crystallographic Defects
on the Reverse Performance of 4H-SiC JBS Diodes”,

© Toyota Central R&D Labs., Inc. 2012 [Link]

 R&D Review of Toyota CRDL, Vol.43 No.1 (2012) 53-59 59

Takashi Katsuno
Research Field:
- Development of Compound
Semiconductor Power Devices
Academic Degree: Ph. D
Academic Society:
- The Japan Society of Applied Physics

Yukihiko Watanabe
Research Field:
- Development of SiC Power Devices
Academic Degree: Dr. Eng.
Academic Societies:
- The Japan Society of Applied Physics
- IEEE Electron Device Society

Tsuyoshi Ishikawa
Research Field:
- Development of Semiconductor Power
Devices
Academic Degree: Dr. Eng.
Academic Societies:
- The Japan Society of Applied Physics
- The Japanese Society of Microscopy

© Toyota Central R&D Labs., Inc. 2012 [Link]