scispace - formally typeset

Journal ArticleDOI

4H-SiC(0001) Surface Faceting during Interaction with Liquid Si

01 May 2016-Materials Science Forum (Trans Tech Publications)-Vol. 858, pp 163-166

AbstractThe aim of this study was to find conditions allowing the "natural" formation of a regular and controllable step bunched morphology on a 4H-SiC seed without the need of any SiC deposition. This was performed by melting a bulk piece of Si on a 4°off 4H seed in the temperature range of 1500 - 1600°C, for 15 min. After etching the remaining Si, the 4H surface was found to be successfully highly step bunched with steps very parallel and regular. A mechanism of dissolution-precipitation was proposed, which could occur both on a short (step to step) and long (centre to periphery) range. This process is kinetically limited at low temperature (1500-1550°C) and considered to be close to the equilibrium at 1600°C.

Topics: Faceting (52%), Atmospheric temperature range (50%)

Summary (1 min read)

Introduction

  • The formation of liquid Si droplets during 4H-SiC epitaxial growth is a common feature when using Si-rich chemical vapor deposition (CVD) conditions.
  • Such growth conditions are usually to be avoided because the growth kinetics under the droplet is altered, resulting in a highly step bunched surface at the places where the Si droplets formed.
  • On the other hand, it was shown recently that a step bunched surface can lead to threading dislocation conversion during epitaxial growth either by CVD [1] or liquid phase epitaxy (LPE) [2].
  • In ref [1] the step bunched morphology was artificially created by photolithography patterning while in ref [2] it is simply the consequence of growth from a liquid phase.
  • The aim of this preliminary study is to find conditions allowing the "natural" formation of a regular and controllable step bunched morphology on a 4H-SiC seed without the need of any SiC deposition.

Experimental

  • Considering the occurrence of a step bunched morphology after CVD growth below Si droplets, the authors investigated the intentional interaction between liquid Si and a SiC seed on a larger scale than what can be made using micrometer size droplets.
  • Towards this end, a bulk piece of n type (phosphorus doped).
  • The experiments were performed in a homemade vertical cold wall CVD reactor working at atmospheric pressure.
  • After cooling, the solidified Si was acid etched and the.

Results and discussion

  • The obtained morphologies after 15 min treatment are shown in Fig. 1.
  • One can see that, for all these experiments, a step bunched morphology was succe sfully obtained, which evidences the effective interfacial reconstruction happening when SiC is put in contact with liquid Si. Furthermore, on all the samples, the steps are found parallel, very straight (with only slight long range undulation at 1600°C) and with very few disturbances.
  • For 1500°C, despite the step width distribution being centred around 2-3 µm, steps width up to 37 µm could be found in the studied area (bigger ones could even be found in specific places, see Fig. 1a and 1b).
  • It suggests that some matter redistribution is occurring at long distance from the centre to the edges of the drop.
  • The latter factor is obviously predominant at low a) b) .

Summary

  • A parallel and highly step bunched surface was successfully obtained on a large scale by putting in contact a big Si drop to the 4H-SiC(0001) 4°off seed.
  • The proposed mechanism of reaction involves a dissolution/precipitation process acting both in short and long range.

Did you find this useful? Give us your feedback

...read more

Content maybe subject to copyright    Report

ETH Library
4H-SiC(0001) surface faceting
during interaction with liquid Si
Conference Paper
Author(s):
Soulière, Véronique; Davy, Carole; Camarda, Massimo; Woerle, Judith ; Grossner, Ulrike ; Dezellus, Olivier; Ferro, Gabriel
Publication date:
2016-05
Permanent link:
https://doi.org/10.3929/ethz-b-000117357
Rights / license:
In Copyright - Non-Commercial Use Permitted
Originally published in:
Materials Science Forum 858, https://doi.org/10.4028/www.scientific.net/MSF.858.163
This page was generated automatically upon download from the ETH Zurich Research Collection.
For more information, please consult the Terms of use.

4H-SiC(0001) surface faceting during interaction with liquid Si
Véronique Soulière
1,a
, Davy Carole
1,b
, Massimo Camarda
2,c
, Judith Wörle
2,3,d
,
Ulrike Grossner
3,e
, Olivier Dezellus
1,f
, Gabriel Ferro
1,g*
1
Laboratoire des Multimateriaux et Interfaces, Université de Lyon, 43 Bd du 11 nov. 1918, 69622
Villeurbanne, France
2
Laboratory for Micro- and Nanotechnology, Paul Scherrer Institute, 5232 Villigen PSI, Switzerland
3
Advanced Power Semiconductor Laboratory, ETH Zürich, Physikstrasse 3, 8092 Zürich,
Switzerland
a
Veronique.Souliere@univ-lyon1.fr,
b
davy.carole@univ-lyon1.fr,
c
massimo.camarda@psi.ch,
d
judith.woerle@psi.ch,
e
ulrike.grossner@ethz.ch,
f
olivier.dezellus@univ-lyon1.fr,
g
gabriel.ferro@univ-lyon1.fr
Keywords: Liquid silicon, CVD, faceting, steps
Abstract. The aim of this study was to find conditions allowing the "natural" formation of a regular
and controllable step bunched morphology on a 4H-SiC seed without the need of any SiC
deposition. This was performed by melting a bulk piece of Si on a 4°off 4H seed in the temperature
range of 1500 - 1600°C, for 15 min. After etching the remaining Si, the 4H surface was found to be
successfully highly step bunched with steps very parallel and regular. A mechanism of dissolution-
precipitation was proposed, which could occur both on a short (step to step) and long (centre to
periphery) range. This process is kinetically limited at low temperature (1500-1550°C) and
considered to be close to the equilibrium at 1600°C.
Introduction
The formation of liquid Si droplets during 4H-SiC epitaxial growth is a common feature when
using Si-rich chemical vapor deposition (CVD) conditions. Such growth conditions are usually to be
avoided because the growth kinetics under the droplet is altered, resulting in a highly step bunched
surface at the places where the Si droplets formed. On the other hand, it was shown recently that a
step bunched surface can lead to threading dislocation conversion during epitaxial growth either by
CVD [1] or liquid phase epitaxy (LPE) [2]. In ref [1] the step bunched morphology was artificially
created by photolithography patterning while in ref [2] it is simply the consequence of growth from
a liquid phase. The formation and control of a highly step bunched morphology is the study of
recent works for the case of LPE [2-4] while this is almost a blank topic for CVD.
The aim of this preliminary study is to find conditions allowing the "natural" formation of a
regular and controllable step bunched morphology on a 4H-SiC seed without the need of any SiC
deposition.
Experimental
Considering the occurrence of a step bunched morphology after CVD growth below Si droplets,
we investigated the intentional interaction between liquid Si and a SiC seed on a larger scale than
what can be made using micrometer size droplets. Towards this end, a bulk piece of n type
(phosphorus doped) Si substrate of 5x5 mm was molten on top of a n type (5x10
15
cm
-3
) epilayer
deposited on Si face 4H-SiC(0001), 4°off seed of ~1 cm
2
. This created a circular liquid Si with an
average diameter of 5-6 mm (see fig. 1). The experiments were performed in a homemade vertical
cold wall CVD reactor working at atmospheric pressure. The interactions were done at 1500, 1550
and 1600°C for 15 min under high purity H
2
. After cooling, the solidified Si was acid etched and the

surface morphology was then studied using optical and scanning electron microscopies. Mechanical
profilometry was used to evaluate the macroscopic shape of the surface below the Si drop.
Results and discussion
The obtained morphologies after 15 min treatment are shown in Fig. 1. One can see that, for all
these experiments, a step bunched morphology was successfully obtained, which evidences the
effective interfacial reconstruction happening when SiC is put in contact with liquid Si.
Furthermore, on all the samples, the steps are found parallel, very straight (with only slight long
range undulation at 1600°C) and with very few disturbances. Obviously, the main difference
between these samples is the step width: at 1500°C there is a high dispersion of step width while for
higher temperatures the width is more regular. From a first glance, one can notice a general trend,
i.e. a decrease of the mean step width with increasing temperature. In order to confirm this point,
both SEM and optical images were quantitatively analysed for step width evaluation. The statistics
obtained on a large number of steps (between 200 to 350) are shown in Fig. 2.
200 µm
1 mm
1500°C
200 µm
1550°C
200 µm
1600°C
Figure 1. Optical (up) and SEM (down) images of the 4H-SiC seeds surface morphology after liquid Si
interaction at different temperatures: a), b) 1500°C, c), d) 1550°C and e), f) 1600°C. The insert a) is a low
magnification observation showing the average shape and size of the original Si droplet.
For 1500°C, despite the step width distribution being centred around 2-3 µm, steps width up to 37
µm could be found in the studied area (bigger ones could even be found in specific places, see Fig.
1a and 1b). It is important to mention that these very large steps observed for 1500°C are most
probably not due to liquid Si induced step bunching. Indeed, when looking to Fig. 1b, if the large
flat area on the upper part of the image was corresponding to a step bunching, then the associated
step height at one extremity would be 7 µm. Such high step height would have been easily noticed
both by SEM or even optical microscopy. We believe that these large flat areas are in fact unreacted
zones, i.e. areas where mechanism leading to step bunching formation at the liquid Si/SiC has not
occurred yet. Under this assumption and considering that the most kinetically advanced (or closer to
equilibrium) sample is the one at 1600°C, one can propose that the evolution of step width
distribution with increasing temperature reflects the progress of the interfacial reaction: at 1500°C
some large areas are left unreacted; at 1550°C the large areas have reacted lately which gives steps
100 µm
100 µm
100 µm
a)
b)
c)
d)
e)
f)

of intermediate width; while at 1600°C the reaction kinetics is so fast that all the intermediate width
steps were converted into smaller ones.
0 5 10 15 20 25 30 35 40
0
10
20
30
40
Frequency (%)
Step widthm)
500 1000 1500 2000 2500
0
Step height (nm)
T = 1500°C
Not a step
0 5 10 15 20 25 30 35 40
0
5
10
15
20
Frequency (%)
Step width (µm)
500 1000 1500 2000 2500
0
Step height (nm)
T = 1550°C
500 1000 1500 2000 2500
0
Step height (nm)
T = 1600°C
0 5 10 15 20 25 30 35 40
0
10
20
30
40
Frequency (%)
Step widthm)
Figure 2. Step width statistics of the studied
samples. a) 1500°C, b) 1550°C and c) 1600°C.
The upper scale giving the step height was
calculated from the step width and the 4°off
angle of the 4H-SiC seeds.
Another interesting feature to be noticed on sample treated at 1500°C is the remarkable separation
by a defect line between a step bunched area and an unreacted one (Fig. 1b). This defect line seems
to act either as a blocking or a generating agent. According to the higher density of steps reaching
from this defect line compared to the surrounding, we believe that the line is generating the step
bunching. Also, from the same sample, many steps seem to be originating from the Si drop edges.
In order to have complementary information on the overall mechanism occurring below the drop,
mechanical surface profiling was performed on these samples. A typical result is given in Fig. 3.
One can see that, compared to the mean altitude of the seed outside the drop, the surface at the
periphery of the drop is several µm above it while it is ~1 µm below it at the centre of the drop. It
suggests that some matter redistribution is occurring at long distance from the centre to the edges of
the drop. Note also that at the centre the seed surface presents long range undulations.
Though it is most probable that the driving force leading to such reconstruction is the minimization
of the interfacial energy, interpretation of our results in terms of interfacial energy is rather difficult.
Considering all the observations made so far and the fact that the only reaction that may occur
between liquid Si and SiC is SiC dissolution in the liquid, we can propose two distinct mechanisms
based on the same process of dissolution-precipitation.
The first mechanism would be a local one between neighbouring steps (transfer of atoms
from one step to another via the liquid) and would result in the faceting of the steps. This
mechanism could be activated by two factors: 1) kinetically activated by temperature and/or
2) catalysed by a surface imperfection. The latter factor is obviously predominant at low
a)
b)
c)

temperature (1500°C). Note that the drop edge could be also seen as a local imperfection
(solid/liquid/vapour triple line) since it was found to be also a source of step generation.
The second mechanism would be a long range dissolution-precipitation process from the
centre to the edge of the drop, with matter transport through the liquid.
Both mechanisms would happen simultaneously. The effect of treatment duration at 1500 and
1550°C is currently under investigation in order to have a clearer insight on the dissolution
mechanism.
Finally, it is worth considering that Si evaporation from the drop was found be too small to have any
significant influence on the mechanism discussed. Indeed, though some Si evaporation was
effectively detected by the formation of a network of µm size droplets surrounding the big one, this
did not drastically change the main Si droplet size, even at the highest temperature of 1600°C. Some
slight receding (~50-100 µm) of the drop was detected from the step bunched print left by the
droplet at the periphery, but this is clearly small compared to the drop diameter of 5-6 mm.
Summary
A parallel and highly step bunched surface was successfully obtained on a large scale by putting in
contact a big Si drop to the 4H-SiC(0001) 4°off seed. The proposed mechanism of reaction involves
a dissolution/precipitation process acting both in short and long range.
References
[1] H. Tsuchida, R. Takanashi, I. Kamata, N. Hoshino, E Makino, J. Kojima, J. Crystal Growth
402 (2014) 260–266
[2] S. Harada, Y. Yamamoto, K. Seki, A. Horio, M. Tagawa, T. Ujihara, Acta Materialia 81 (2014)
284–290
[3] T. Mitani, N. Komatsu, T. Takahashi, T. Kato, K. Fujii, T. Ujihara, Y. Matsumoto, K.
Kurashige, H. Okumura, J. Crystal Growth 401 (2014) 681–685
[4] S. Endo, K. Kamei, Y. Kishida, K. Moriguchi, Mater. Sci. Forum 821-823 (2015) 26
Horizontal profile (µm)
1000
2000
3000
4000
5000
6000
7000
8000
0
-1
0
1
2
3
4
Vertical profile (µm)
5
6
Horizontal profile (µm)
1000
2000
3000
4000
5000
6000
7000
8000
0
-1
0
1
2
3
4
Vertical profile (µm)
45°
5
6
Horizontal profile (µm)
1000
2000
3000
4000
5000
6000
7000
8000
0
-1
0
1
2
3
4
Vertical profile (µm)
90°
Figure 3: Mechanical profiles of the surface of 1550°C treated sample a) parallel to the steps (0°), b) 45°
rotated and c) perpendicular to the steps (90°). These profiles were found very similar to the ones obtained
after treatment at different temperatures. The dashed line corresponds to the surface mean altitude outside
the droplet.
a)
b)
c)
Citations
More filters

Journal ArticleDOI
Abstract: Current generations of 4H-SiC metal-oxide-semiconductor field-effect transistors are still challenged by the high number of defects at the SiO2/SiC interface that limit both the performance and gate reliability of these devices. One potential source of the high density of interface defect states (D-it) is the stepped morphology on commonly used off-axially grown epitaxial surfaces, favoring incomplete oxidation and the formation of defective transition layers. Here we report measurements on intentionally modified 4H-SiC surfaces exhibiting both atomically flat and stepped regions where the generation of interface defects can be directly linked to differences in surface roughness. By combining spatially resolving structural, chemical, optical, and electrical analysis techniques, a strong increase of D-it for stepped surfaces was revealed while regions with an atomically flat SiC surface exhibited close-to-ideal interface properties.

12 citations


Journal ArticleDOI
Abstract: In this study, we analyze the electrical properties of MOS capacitors fabricated on different surface morphologies. Comparing a standard, low-roughness ( ox = 3.6 nm to d ox = 32 nm. The extracted interface state traps (D it ) shows a small, but systematic, decrease/improvement of ∼10–15 % for the samples with macrosteps.

2 citations


Cites methods or result from "4H-SiC(0001) Surface Faceting durin..."

  • ...In this study, we firstly produced 4H-SiC macrostepped surfaces using a bulk Si melting process (see ref.[9] for details)....

    [...]

  • ...The authors speculated that the five times increase of the mobility was due to a lower density of interface traps rather than due to a different scattering mechanisms of the carriers in the channel, but reported no measurements for channels along different crystallographic orientations (as compared to the used one, the [1-100]) nor measurements of the Dit....

    [...]


Proceedings ArticleDOI
01 Oct 2019
Abstract: SiO 2 /SiC interfaces have been under intensive research, because the quality of the interfaces can significantly affect the performance and reliability of SiC power devices. For offering microscopic insights on the generation of interface defects, here we describe the application of our novel scanning probe microscopy method called local deep level transient spectroscopy (local DLTS) to a thermally oxidized Si-face 4H-SiC epitaxial layer. Our sample is an off-axially grown epitaxial layer which is intentionally modified to have wide flat terraces and macro-stepped features instead of typical terraces separated by atomic steps. A spatially resolved map of the interface defect density shows that the macro-steps have a significantly higher interface defect density than the wide flat terraces. The result suggests that high interface defect densities on typical off-axis SiO 2 /4H-SiC stacks might be related to the steps created during the step-controlled epitaxial growth. Local DLTS is useful for the microscopic evaluation of the interface quality.

Frequently Asked Questions (1)
Q1. What are the contributions in this paper?

The aim of this study was to find conditions allowing the `` natural '' formation of a regular and controllable step bunched morphology on a 4H-SiC seed without the need of any SiC deposition. The formation and control of a highly step bunched morphology is the study of recent works for the case of LPE [ 2-4 ] while this is almost a blank topic for CVD. The aim of this preliminary study is to find conditions allowing the `` natural '' formation of a regular and controllable step bunched morphology on a 4H-SiC seed without the need of any SiC deposition. Experimental Considering the occurrence of a step bunched morphology after CVD growth below Si droplets, the authors investigated the intentional interaction between liquid Si and a SiC seed on a larger scale than what can be made using micrometer size droplets. After cooling, the solidified Si was acid etched and the surface morphology was then studied using optical and scanning electron microscopies. One can see that, for all these experiments, a step bunched morphology was successfully obtained, which evidences the effective interfacial reconstruction happening when SiC is put in contact with liquid Si. Furthermore, on all the samples, the steps are found parallel, very straight ( with only slight long range undulation at 1600°C ) and with very few disturbances.