TRANSMISSION ELECTRON MICROSCOPY AND THERMAL RESIDUAL STRESS ANALYSIS OF AlN CRYSTAL by RAC. G. LEE., B.S. A Thesis In MECHANICAL ENGINEERING Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE Approved Dr. J. Chaudhuri Dr. A. Idesman Dr. Y. Ma John Borrelli Dean of the Graduate School May, 2007
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
TRANSMISSION ELECTRON MICROSCOPY AND THERMAL RESIDUAL
STRESS ANALYSIS OF AlN CRYSTAL
by
RAC. G. LEE., B.S.
A Thesis
In
MECHANICAL ENGINEERING
Submitted to the Graduate Faculty of Texas Tech University in
Partial Fulfillment of the Requirements for
the Degree of
MASTER OF SCIENCE
Approved
Dr. J. Chaudhuri
Dr. A. Idesman
Dr. Y. Ma
John Borrelli Dean of the Graduate School
May, 2007
Texas Tech University, Rac. G. Lee, May 2007
ii
ACKNOWLEDGEMENTS
I would like to acknowledge the support and guidance of my committee chair, Dr. J.
Chaudhuri, who donated countless hours to investigate the high energy gab
semiconductor used in this work. Thanks for the help during the long afternoons in the
office, week ends, and most importantly thanks for your open-mind.
Appreciation is also given to the remaining members of my advisory committee,
Dr. A. Idesman, Dr. Y. Ma, for their time and effort.
Thanks to my wife, Sun Jin Jang, who is everything to me, and I dedicate this
work to my three months old daughter, A Hyun Lee (Katie Lee). Most of all, I would like
to give special thanks to my parents for always supporting my decisions.
Source : (Robert R. Reeber et al. 2001) for AlN and (H. B. Huntington) for W
Table 2.2. Thermal expansion coefficients at 1700 K (α, 10-6 K-1) Crystal α 11 α 22 α 33
AlN 6.6 6.6 5.8 W 5.7 5.7 5.7
Source : (H. B. Huntington. P 70)
Geometry of the AlN film grown on tungsten substrate is modeled as a single or
several thin islands placed on a large substrate. A representative sample size was
determined by simulating the stresses with different thickness ratios between the film and
the substrate and also for different planar dimensions of both the film and the substrate.
Based on the convergence of the overall stress-strain response, it was determined that the
thickness of the substrate should be 10 times the thickness of the film. In addition, planar
dimensions of the substrate in directions perpendicular to the thickness, are much larger
then the corresponding dimensions of the islands. A 2-D formulation, with the plane-
stress state and similar to the approach by W. M. Ashmawi (W. M. Ashmawi et al. 2004)
that corresponds to a thin sample, as well as a general 3-D formulation are considered.
For simplicity, rectangular islands (i. e. rectangles and parallelepipeds in the 2-D and 3-D
Texas Tech University, Rac. G. Lee, May 2007
6
cases, respectively) are studied (more complicated shapes of islands can be analyzed
without any difficulty as well).
Due to symmetry, one half of a sample in the 2-D case and one quarter of a
sample in the 3-D case are considered for problems with one island (Figure 2.1). For the
case of several islands, a symmetric sample with one plane of symmetry for the 2-D case
(Figure 2.1 (b)) and two planes of symmetry for the 3-D case (Figure 2.1 (c)), in addition
to above configuration, are considered. This geometrical symmetry along with the
boundary conditions reduces the dimension of the system for finite element applications
and does not affect the analysis of residual stresses. For the 3-D case, the Cartesian axes
notations are designated as x1 and x2 are the in-plane directions and x3 is the thickness
direction. For the 2-D case, x1 along the in-plane direction and x2 along the thickness
direction are. Normal displacements and tangential forces are zero on planes and lines of
symmetry; all other surfaces are free of stresses.
Texas Tech University, Rac. G. Lee, May 2007
χ3
χ1
χ2
dHdL1
dW
dS
• •A B
7
(a)
SLd
Hd
Sd
1Ld
χ1
χ3
(Island Ⅰ)
Substrate
(b) (c)
Figure 2.1. Sample geometry: the sample consists of a substrate and one or three islands. (a) Sketch of the entire sample with one island. Due to the symmetry, (b) one half of the sample geometry for the 2-D case and (c) one quarter of the sample for the 3-D case are used for all calculations and reporting results. An additional island is shown with the dotted lines in Figure 1 (c) which is the representative of 3 islands for the whole sample.
The introduction of the dimensionless coordinates e.g.H
X
dd 1 ,
H
X
dd 2 and
H
X
dd 3
essentially simplifies the analysis, where is the height or thickness of the island. It
can be shown that for the linear thermo elastic problems considered, strains and stresses
in the sample depend on the ratio
Hd
H
L
dd 1 in the 2-D case and on the ratios
H
W
dd
and H
L
dd 1 in the
Texas Tech University, Rac. G. Lee, May 2007
3-D case, and are independent of the absolute value of the island height, ( and
are the dimensions of the island along the coordinate axes x
Hd 1Ld Wd
1 and x2, respectively. The
dimensions of the substrate along the x1 and x2 axes are and , respectively. In the
2-D and 3-D cases stresses
SLd SWd
ijσ can be expressed in terms of dimensionless quantities as:
),,,,,,,,( 21321 θεσσ iif
mnsmn
HHH
S
H
L
H
L
H
Lijij cc
dx
dx
dd
dd
dd
dd
Δ= (4)
and
),,,,,,,,,( 321321 θεσσ iif
mnsmn
HHHH
S
H
L
H
L
H
Lijij cc
dx
dx
dx
dd
dd
dd
dd
Δ= (5)
respectively, where and are the elastic coefficients for the film and the substrate,
and the strain, , is given by Equation (3). For simplicity, and are taken as
equal to each other for the all calculations. The axes of material symmetry used in
Equation (1) are directed in the common Cartesian system as follows: the 11-axis
coincides with the x
fmnc s
mnc
θε iiΔ Wd 1Ld
1 -axis ( i. e. a-axis [1-210] for hexagonal and a-axis [100] for cubic),
the 22-axis coincides with the x2 –axis (i. e. direction [01-10] for hexagonal and b-axis
[ 010] for cubic), and the 33-axis coincides with the x3-axis( i. e. c-axis [0001] for
hexagonal and c-axis [001] for cubic). For the 2-D case 22-axis coincides with the x2-axis
(i. e. c-axis [0001] for hexagonal and c-axis [001] for cubic).
8
Finite element calculations were performed using the commercial finite element
code ABAQUS (ABAQUS. 2004) with quadrilateral quadratic elements. Due to non-
homogeneous stress distributions, non-uniform meshes with refinement near the interface
between the film and substrate are generated for the improved accuracy. Approximately
3200 quadrilateral elements for AlN and 48000 quadrilateral elements for W were used.
Texas Tech University, Rac. G. Lee, May 2007
Due to discontinuities in thermo mechanical properties across the interface
between the film and the substrate, stresses across the interface are also discontinuous
and have to be analyzed in the film and in the substrate separately along the interface.
To determine the optimum growth condition for the AlN crystal, several
problems as follows are studied; the effect of the growth of the island or film in the in-
plane direction, the effect of different thicknesses of the island and the tungsten substrate,
misorientaion of grains in AlN, and the effect of the interaction between islands. The
results of these calculations are presented in the following sections.
2.3 Effect of the meshes
For the considered problems, the stress reaches a maximum value at the corner of
the interface between the island and the substrate. In other words, the maximum stress
around the corner obtained with the finite element method (or any numerical method) is
mesh-dependent. For example, the distribution of the normal stress in the film along the
x1 direction at the interface for a 3-D problem with one island is given in Figure 2.2 for
the uniform coarse meshes (with 20 nodes along the interface), uniform fine meshes (with
80 nodes along the interface), and non-uniform fine meshes with refinement in the
vicinity of the corner (with 100 nodes along the interface). For the uniform coarse meshes,
the maximum normal stress at the corner is 1.789 GPa. For the uniform fine meshes the
stress is 2.458 GPa and for the non-uniform fine meshes, the stress becomes 2.844 GPa.
It can be seen that stresses for the coarse and fine meshes coincides except a small region
near the corner (i. e. 0.1 from the corner). Therefore, in order to obtain mesh
independent results, we will consider stresses up to the distance 0.1 from the corner
Hd
Hd
9
Texas Tech University, Rac. G. Lee, May 2007
for all analysis. Another possibility to avoid mesh-dependent results is to use non-linear
models (e.g., elastoplastic models).
0 2 4 6 80.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
S-S
11 (G
Pa)
X1/dH
Figure 2.2. The distribution of in-plane stress component, 11σ in the film along the x1 axis at the interface for different finite element meshes. Curves A, B and C correspond to the non-uniform fine mesh with mesh refinement near the corner (100 nodes along the interface), the uniform fine mesh (80 nodes along the interface), and uniform coarse mesh ( 20 nodes along the interface), respectively.
2.4 Effect of the size of the AlN film
In this section, the effect of the planar dimensions of the AlN film on residual
thermal stress is analyzed. A thick and large substrate is considered e. g. the thickness
and planar dimensions of the substrate are 10 and 24 times the thickness of the island
(H
S
dd
= 10, H
SL
dd
= 24 for the 2-D case and H
S
dd
= 10 and H
SL
dd
= H
SW
dd
= 24 for the 3-D case,
where dS, dSL and dSW are the thickness and planar dimensions of the substrate, 10
Texas Tech University, Rac. G. Lee, May 2007
respectively, and dH is the thickness of the island (Figure 2.1)). The dimensions of the
island were varied as follows: H
L
dd 1 = 4, 6 and 8 for both 2-D and 3-D cases where dL1 is
the length of the film along the x1 direction (see Figure 2.1 (b) for the 2-D case, and
Figure 2.1(c) for the 3-D case). For example, for the island thickness, = 0.1 mm, these
ratios correspond to the following planar dimensions of the island = = 0.4, 0.6
and 0.8 mm, respectively, and the substrate dimensions are = 1mm and = =2.4
mm.
Hd
1Ld Wd
Sd SLd SWd
For the 2-D formulation (plane stress case), Figures 2.3 (a), (b) and (c) show the
distribution of the normal and shear components of the stresses in the film and substrate
for H
L
dd 1 as 8 . Figure 2.4 (a) shows the comparison of variation of σ11 in the film along the
x1 axis at the interface and Table 2.3 summarizes the stress values at the center and corner
in the film and the substrate along the interface for all three cases. For the 3-D
formulation, Figures 2.5 (a), (b) and (c) show the distribution of stresses for H
L
dd 1 as 8,
Figure 2. 4 (b) shows the comparison of variation of σ11 in the film along the x1 axis at
the interface and Table 2.4 summarizes the stress values at the center and corner in the
film and the substrate along the interface for all three cases. The stress components are
almost constant along the interface except at the corner where there is a stress
concentration and the stress increases enormously. The in-plane stresses (i. e. σ11 and σ22)
in AlN are tensile and higher than the yield strength of AlN (i. e. 0.3 GPa (I. Yonenaga et
11
Texas Tech University, Rac. G. Lee, May 2007
12
al.)) while the normal component along the thickness direction (i. e. σ33) is almost zero
and shear component (i. e. σ13) is compressive.
In-plane stress in AlN increases as the islands become larger and may lead to the
formation of microscopic cracks especially at the corner. The stresses in the tungsten
substrate are compressive except at the corner, where the normal stress becomes tensile
and the shear stress is zero. The reversal of stress across the interface (compressive in the
substrate and tensile in the film) may cause delamination or separation of the film.
The values of all stress components in the 3-D case are much higher as compared
to those in the 2-D case. This indicates that stresses in 2-D calculations are
underestimated and should be used for very rough estimations. Therefore, all results that
follow are analyzed using the 3-D formulation.
Texas Tech University, Rac. G. Lee, May 2007
(a)
(b)
(c)
Figure 2.3. The distribution of stress components (a) σ11, (b) σ22, and (c) σ12 for H
L
dd 1 = 8
for the 2-D case.
13
Texas Tech University, Rac. G. Lee, May 2007
-1 0 1 2 3 4 5 6 7 8 90.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
CBAS
-S11
(GP
a)
X1/dH
A: dL1/dH = 4B: dL1/dH = 6C: dL1/dH = 8
-1 0 1 2 3 4 5 6 7 8 90.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
CB
A
S-S
11 (G
Pa)
X1/dH
A: dL1/dH = 4B: dL1/dH = 6C: dL1/dH = 8
(a) (b) Figure 2.4. The variation of stress component, σ11, in the film along x1 direction at the
interface (i. e. from the center to the edge) for H
L
dd 1 = 4, 6 and 8 for (a) the 2-D case and
(b) 3-D case.
Table 2.3. The minimum and maximum values of the stress components σ11, σ22, σ12, at the interface in AlN and W for the 2D case
Figure 2.5. (a). The distribution of stress components (a) σ11 and (b) σ22 , for H
L
dd 1 = 8 for
the 3-D case.
15
Texas Tech University, Rac. G. Lee, May 2007
(c)
(d)
Figure 2.5. (b). The distribution of stress components (c) σ33 and (d) σ12 for H
L
dd 1 = 8 for
the 3-D case.
16
Texas Tech University, Rac. G. Lee, May 2007
(e)
(f)
Figure 2.5. (c) The distribution of stress components (e) σ13 and (f) σ23 for H
L
dd 1 = 8 for the
3-D case.
17
Texas Tech University, Rac. G. Lee, May 2007
Table 2.4. The minimum and maximum values of the normal stress components, σ11 , σ22, and σ33, and shear stress component, σ13, at the interface between AlN and W for the 3D case
Figure 2.7. The variation of the stress component, σ11, in the film along the x1 direction at the interface for different thicknesses of the film. Curves A, B, C, D and E correspond to
the following ratios H
L
dd 1 = 32, 16, 8, 4 and 2, respectively.
2.6 Effect of the thickness of the substrate
The thickness of the substrate can be one of the parameters that could be varied
for the investigation of the optimization of the crystal growth. Therefore, to understand
the effect of the thickness of the substrate on thermal residual stress in the film,
calculations of thermal residual stress with various thicknesses of the substrate were
performed. Residual stresses with variable thickness of the substrate (i. e. H
S
dd
=10, 5,
and 2 with the thickness of the film, dH, as 0.1 mm, and the thickness of the substrate, ds,
as 1.0, 0.5, and 0.2 mm) were calculated. Figure 2.8 shows the variation of the σ11 stress
component in the film at the interface along the x1 axis. The decrease in the substrate
thickness from 1.0 to 0.2 mm leads to the decrease in the in- plane stress by 35 %. Thus
formation of microscopic cracks and delamination in the film might be avoided by using
Figure 2.8. The variation of the normal stress in the x1 direction (σ11) along the interface in the film for different thicknesses of the substrate. Curves A, B, and C correspond to
the following ratios, H
S
dd
=10, 5, and 2, respectively.
2.7 Effect of the interaction of the islands (i. e. the island structure of the film).
For the analysis of the influence of the island structure of the film on the stress
distribution, the film consisting of three islands has been considered (Figure 2.1 (c)). The
lengths of the islands I and II were selected as H
L
dd 1 =4 and
H
L
dd 3 =8, respectively.
Preliminary study has found that the stresses in the islands are affected by the distance
between the islands, H
L
dd 2 , and the dimensions of islands
H
L
dd 1 and
H
L
dd 3 , and are
independent of the number of islands. The distances between the islands were taken
asH
L
dd 2 = 0 (infinitely small separation), 0.2, 1 and ∞, where dH is 1.0 mm. Figure 2.9 (a)
21
Texas Tech University, Rac. G. Lee, May 2007
shows distribution of in-plane stress component, σ11, for H
L
dd 2 ≈ 0 and 2.9 (b) shows the
comparison of variation of σ11 in the film along the x1 axis at the interface for all four
cases. At dL2 =1 and ∞, the effect of the interaction between islands is very small, and the
stress in the islands is close to that for a single island. For = 0.2 and 0, the effect of
interaction increases the stress by 80% and 100% at the corner, respectively. This stress
build up could be the origin of the formation of threading dislocations and grain
* TaC and NbC are belongs to the group of RX-type crystals. RX-type crystals have the structure of NaCl, sodium chloride(Ralph W. G. Wyckoffpp 85~93). Atomic positions in the unit cube : R: (4a) 000 or ½ ½ 0; ½ 0 ½; 0 ½ ½ X: (4b) ½ ½ ½; or ½ 0 0 ; 0 ½ 0 ; 0 0 ½ Source : Elastic constants for W and Al2O3 (H. B. Huntington), AlN and SiC (Robert R. Reeber et al. 2001), NbC (H. M. Ledbetter) and Tac (L. Lopez de la Torre)
30
Texas Tech University, Rac. G. Lee, May 2007
31
Table 3.2. Thermal expansion coefficients at 2000 K (α, 10-6 K-1) AlN Al2O3 α-SiC(6H) TaC NbC W
Source : thermal expansion coefficients for AlN and α-SiC(6H)( Robert R. Reeber. et al.), Al2O3 ,NbC, TaC and W(Y. S. Touloukian et al. 1977)
3.3 Sapphire (Al2O3) substrate
Figure 3.2. illustrates the stress component, σ11, using Al2O3 substrate. Sapphire
substrate has much higher thermal expansion coefficient and lattice parameter of the C-
direction [0001] is much bigger than aluminum nitride, hence it produces much higher
defect densities and thermal stress. The thermal stress of aluminum nitride grown on
sapphire shows under horizontal compressed stress σ11 (figure 3.5) and vertical tensile
stress σ33 (figure 3.6). Table 3.3 shows the stress component at the center ( B point from
figure 3.2 (a)) and corner (C point from figure 3.2 (a)), and the stresses are maximum at
the corner because of the stress concentration. Aluminum nitride grown on sapphire
substrate has the highest thermal stress than other substrate compared. However, the
substrate is under tensile stress near the interface between aluminum nitride and sapphire
substrate. This opposite stress component at the interface are arranged from -2.496 GPa
(σ11) (at the center of aluminum nitride) to 0.763 GPa (σ11) ( at the center of sapphire
substrate), and similar result are showing for all the substrate. At the end of the corner
point (c) of the substrate, the tensile stress 0.264 GPa (σ11) changes to the compressed
stress -0.479 GPa (σ11) at the vicinity of the corner because of the thermal expansion of
the aluminum nitride, thus gradually increasing to the zero stress which shows far from
Texas Tech University, Rac. G. Lee, May 2007
the corner. The majority of the film area is under small shear stress σ12 and σ23. Shear
stress σ13 is plotted in figure 3.4 (c).
Figure 3.2. The illustration of the stress component, σ11, using Al2O3 substrate
3.4 TaC and NbC substrate
Figure 3.3 (a) and (b) illustrate the stress component, σ11 using tantalum carbide
and niobium carbide substrate. The advantage of using TaC and NbC is well presented
by B. Liu (B. Liu et al.). Compare to sapphire substrate, thermal expansion coefficients
are small, and lattice parameter is much closer to the aluminum nitride, but thermal
32
Texas Tech University, Rac. G. Lee, May 2007
expansion coefficient are higher than aluminum nitride. Therefore aluminum nitride
grown on tantalum carbide (TaC) or niobium carbide (NbC) behave the same way as
grown on sapphire substrate. However, the range of stress is smaller than using sapphire
substrate. For the aluminum nitride grown on tantalum nitride substrate, the minimum
compressed stress area at the bottom corner is not concentrated. According to the figures
3.5 ,3.6 and Table 2.3, it is clear that the aluminum nitride grown on tantalum carbide
(TaC) and niobium carbide (NbC) produces much small thermal stress and will prevent
micro crack caused by thermal residual stress.
Figure 3.3 (a). The illustration of the stress component, σ11, using tantalum carbide substrate
33
Texas Tech University, Rac. G. Lee, May 2007
Figure 3.3 (b). The illustration of the stress component, σ11, using niobium carbide substrate
3.5 α-SiC(6H) and W substrate
Figure 3.4 (a) and (b) illustrate the stress component, σ11 using silicon carbide and
tungsten substrate. Thermal expansion coefficient of the silicon carbide and tungsten are
smaller than aluminum nitride. Specially, thermal expansion coefficient of silicon carbide
shows much smaller than tungsten substrate, and lattice parameter of C-direction [0001]
is even bigger than sapphire. Using silicon carbide as a substrate increased coefficient of
thermal expansion mismatch thus higher thermal stress. The thermal stress of aluminum
nitride grown on silicon carbide or tungsten shows under horizontal tensile stress σ11
(figure 3.5) and vertical compressed stress σ33 (figure 3.6). The substrates, silicon
carbide and tungsten, are under compressive stress between aluminum nitride and
34
Texas Tech University, Rac. G. Lee, May 2007
substrates. Comparing thermal residual stress of tungsten with silicon carbide, it is clear
that using tungsten as substrate produces smaller thermal stress for aluminum nitride
single crystal. In addition to the thermal residual stress, tungsten crucibles have been
chosen to avoid carbon contamination and resistant to attack by Al vapor at the high
temperature (G. A. Slack et al. 1976, 1977).
Figure 3.4 (a). The illustration of the stress component, σ11, using silicon carbide substrate
35
Texas Tech University, Rac. G. Lee, May 2007
Figure 3.4 (b) . The illustration of the stress component, σ11, using tungsten substrate
36
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 8
-6
-4
-2
0
2
4
6
8
S-S1
1 (G
Pa)
Length, mm
W substrate NbC substrate TaC substrate SiC substrate Al2O3 substrate
(a)
0 2 4 6 8-5
-4
-3
-2
-1
0
1
2
3
4
5
6
S-S2
2 (G
Pa)
Length, mm
W substrate NbC substrate TaC substrate SiC substrate Al2O3 substrate
(b)
0 2 4 6 8
-4
-3
-2
-1
0
1
2
3
4
5
6
S-S
13 (G
Pa)
Length, mm
W substrate NbC substrate TaC substrate SiC substrate Al2O3 substrate
(c)
Figure 3.5. The variation of the stress component, (a) σ11, (b) σ22 and (d) σ13 in the film along x1 direction at the interface between island and substrate. Curves are corresponding to the following substrates, W, NbC, TaC, SiC and Al2O3, respectively.
37
Texas Tech University, Rac. G. Lee, May 2007
Table 3.3. The minimum and maximum values of the normal stress components, σ11 , σ22, and σ33, and shear stress component, σ13, in the film at the interface between AlN and substrates.
Substrates Stress (GPa) Al2O3 α-SiC(6H) TaC NbC W σ11(GPa) (center~corner)
Figure 3.6. The variation of the normal stress component σ33 in the film along x3 direction (thickness direction) at the center of island. Curves are corresponding to the following substrates, W, NbC, TaC, SiC and Al2O3, respectively.
38
Texas Tech University, Rac. G. Lee, May 2007
39
3.6 Conclusions
Suitability of several refractory materials such as tungsten (W), tantalum carbide
(TaC) and niobium carbide (NbC) for aluminum nitride (AlN) crystal growth was
discussed by B. Liu and it has been concluded that W is a durable and less expensive
crucible material. Thermal residual stress of AlN grown on sapphire, silicon carbide,
tungsten, tantalum carbide and niobium carbide were investigated to produce high purity
aluminum nitride single crystal. Thermal residual stress of AlN grown on silicon carbide
and tungsten shows compressive stress (σ 11) at the end of the island and, because of
thermal expansion coefficient, tungsten shows the least amount of thermal stress. When
sapphire, TaC and NbC were used as substrate, the thermal stress (σ 11) of AlN shows
tensile stress at the end of the island. Even though sapphire has hexagonal structure
which is same as AlN structure, it has the highest amount of thermal stress, thus it can be
the worst substrate. TaC and NbC will be the good substrate to grow AlN crystal. At the
interface, compressive and tensile stresses are conjunct, and this different stress state can
cause of the generation of dislocation and micro crack.
Texas Tech University, Rac. G. Lee, May 2007
40
CHAPTER 4
THE ANALYSIS OF BULK AlN CRYSTAL GROWN ON TUNGSTEN SUBSTRATE
4.1 Introduction
The unique properties of the group III-nitrides, such as wide direct band gap, high
thermal conductivity, and high thermal stability have made GaN and AlN the most
serious candidates for the high-power and high-frequency electronic and deep ultraviolet
(UV) optoelectronic devices (B. Monemar. et al. 1999) (S. Strite et al. 1992) . AlN is one
of the most promising substrates for group III-nitride based devices due to lower lattice
constant and thermal expansion coefficient mismatch and hence reduced defect densities
and residual thermal stresses (J. H. Edgar et al. 2002). Currently, several groups are
producing bulk AlN crystals by the sublimation-recondensation technique (J. H. Edgar et
al. 2002) (J. Carlos Rojo et al. (2002) (R. Sclesser et al. 2002)
.
4.2 Sublimation method
Sublimation (one kind of physical vapor deposition, PVD) is the most successful
method to grow AlN bulk crystals so far (G. A. Slack et al. 1977). Vapor deposition can
be separated by two methods. One is Physical Vapor Deposition (PVD) and another is
Chemical Vapor Deposition (CVD). When the vapor state of the source material changes
to solid state at the substrate, different solidification procedure occurs by PVD and CVD
respectively. In detail, for PVD, sintered or melted composite materials are used as solid
target materials and heated by laser beam or thermal method to deposit to the substrate.
Texas Tech University, Rac. G. Lee, May 2007
High Vacuum state is required to prevent vapor state target material to react with other
vapor molecules. The element attached to the substrate has the same composition with the
element arrived to the target by vapor state. Figure 4.1 shows the sketch of the AlN
crystal growth process. AlN powder source sublimes and re-condenses on a colder seed
crystal (R. Dalman, 2005).
Figure 4.1. Sketch of the AlN crystal growth process: Source; (R. Dalman, 2005)
41
Texas Tech University, Rac. G. Lee, May 2007
42
The source material for AlN crystal growth in this study was sintered AlN. The
original AlN powders had an average agglomerated particle size of 1.8 mm, and
contained 0.9 wt% O and 0.06wt% C as the major impurities, as reported by the vendor.
The sintered AlN source was easy to handle and the distance between the source material
and the growth region did not change much in the initial growth period. Sintered AlN
source was prepared by heating AlN powders at 1960 C for up to 4 h. The sublimation
growth was conducted in a resistively heated tungsten furnace. The furnace was heated
by two tungsten wire mesh heating elements with a maximum operation temperature of
2400 C at one atmospheric pressure as claimed by the vendor. The heating elements
provided an axial temperature gradient of 3–5 C/mm between the source material and the
crystal growth region, which was the driving force for the sublimation growth. The
growth chamber consisted of a tungsten crucible contained in a concentric tungsten retort,
to prevent the escape of vaporized species during the growth. The growth temperature
was measured by an optical pyrometer focused on the top lid of the outside tungsten
retort, and was controlled by the furnace output power (Z. Gu. et al. 2006). Figure 4.2 (a)
shows the aluminum nitride boul grown by sublimation method and (b) illustrate the
crystal structure of aluminum nitride. Aluminum and nitride can be occupied by arbitrary.
Texas Tech University, Rac. G. Lee, May 2007
(a)
(b)
Figure 4.2 aluminum nitride poly crystal grown on tungsten by sublimation method (a) and computed atomic structure (b).
43
Texas Tech University, Rac. G. Lee, May 2007
44
4.3 Surface morphology by SEM When aluminum nitride was grown on tungsten substrate, it has been investigated
that the crystal has smooth and rough surfaces. In order to see it closely, scanning
electron microscope (SEM) image was taken (figure 4.3). The nature of the scattering can
result in different angular distributions. Scattering can be either forward scattering or
backward scattering. The forward scattered electrons are used for transmission electron
microscopy (TEM) and backward scattered electrons can be detected by SEM. When the
electron passes through the specimen, it may interact with the electron cloud, and make
small angular deviation. However, when it pass through the electron cloud and
approaches the nucleus, the electron (negative charge) may be attracted and scattered
through a large angle. Backward scattered electron consists of Secondary electrons from
within the specimen and Incoherent elastic backscattered electrons. As the specimen gets
thicker, fewer electrons are forward scattered and more are backscattered until primary
incoherent backscattering is detectable in bulk, nontransparent specimens. Because
backscattered electrons has high angle (> ~10°), the amount of energy loss is higher than
secondary electrons, which dedicates to the high resolution surface morphology. Figure
4.3 (a) was taken by secondary electrons and (b) was taken by backscattered electrons
which show rough surface morphology clearly.
Texas Tech University, Rac. G. Lee, May 2007
(a)
(b)
Figure 4.3. Scanning Electron Microscopy (SEM) image of rough and smooth surface by (a) secondary Electrons and (b) back scattered electrons.
45
Texas Tech University, Rac. G. Lee, May 2007
46
4.4 EDAX result using SEM Energy Dispersive X-ray Analysis (EDAX) technique has been used to investigate
the compositional analysis for polarity determination. Since the surface morphology
consists of rough and smooth surface, it was assumed that the surface polarity (N- polar
or Al- polar) makes this difference. When the incident beam hit the specimen, the
electron from the incident beam emit the electrons from the electron cloud and specific
amount of energy releases when electrons form next shell occupy emitted position of the
inner shell. The amount of energy released by the transferring electron depends on which
shell it is transferring from, as well as which shell it is transferring to. Thus, by
measuring the amounts of energy present in the X-rays being released by a specimen
during electron beam bombardment, the identity of the atom from which the X-ray was
emitted can be established. If the emitted x-ray satisfies Bragg’s angle (figure 4.4), then
the detector can catch the emitted x-ray with specific angle.
Figure 4.4. The description of Bragg angle. The path difference between reflected waves is AB + BC.
B
θ
θ θ
θ
A C
d
Incident plane wave
Scattered plane wave
Texas Tech University, Rac. G. Lee, May 2007
47
n λ = AB + BC = 2d sin θ (6)
Equation (6) shows the path difference between reflected waves called Bragg’s Law in
figure 4.4. If we know the wave length (λ) for the incident electron and we can measure θ
experimentally, then we can calculate the interplanar spacing (d) in crystal. The X-rays
are generated in a region about 2 microns in depth (tear drop effect), and thus EDAX is
not a surface science technique. However, in this experiment, it has been used for the
guide line. Figure 4.5 shows the EDAX result by Hitachi S-4300 SEM. For EDAX
experiment, the region used for x-ray analysis was shown in figure 4.5 (a). From figure
4.5 (a), there are two different type of the surface such as smooth and rough surface.
Figure 4.5 (b) shows the EDAX result from smooth surface and 4.5 (c) are from the
rough surface. EDAX result shows that the smooth surface contains nitride and atomic
weight of aluminum is 26.62 % and 11.78% for nitride, respectively. Carbon and oxide
also detected and those elements are detected from the grid and epoxy.
However, no nitride or very small amount of nitride was detected in the rough surface.
Nitride peak is not showing in figure 4.5 (c), but it is probably over-wrapped by carbon
peak, because of small amount of nitride compared to the smooth surface.
Texas Tech University, Rac. G. Lee, May 2007
(a)
(b)
(c)
Figure 4.5. Pointed aluminum nitride surface morphology (a) and EDAX result of smooth (b) and rough(c) surface.
48
Texas Tech University, Rac. G. Lee, May 2007
4.5 Dislocation Study by Etching
In general, the shape and morphology of the etch pit depend on the type of defect,
the crystal orientation and polarity, the basic crystal structure and symmetry, and the
etchant and its composition. Usually, the planes exposed by etching have the slowest etch
rate. For hexagonal symmetry crystals, such as wurtzite structured GaN, AlN, and SiC
(6H–SiC or 4H–SiC), hexagonal pits are usually produced on the (0 0 0 1) planes. Such
pits can be delimited by six planes inclined to the c-axis (Z. Gu., 2006). For this study,
AlN crystal was etched in molten KOH at 405 °C for 2 minutes by Dr. Edgar’s group at
Kansas State University. Figure 4.6 (a) shows the surface of AlN crystal before etching
and (b) shows after etching
(a) (b)
Figure 4.6. Optical microscope image of the AlNsurface (a) before etching and (b) after etching in molten KOH at 405 °C for 2 minute.
Figure 4.7 shows (a) the optical microscope image with high magnification and
(b) scanning electron microscopy image in etch pit area. The etch pit density after etching
at 405 °C for 2 minute was approximately 10.4 * 1010 Cm-2, as determined by counting
49
Texas Tech University, Rac. G. Lee, May 2007
the number of etch pits in a randomly chosen area. Etch pits were found to be of
hexagonal shape with a deviation of the apex from center. These dislocations are
associated with edge and screw dislocation. The final shape of the etch pit gives
information about the type of dislocation since the etch pit is centered around the
dislocation. If the dislocation line is normal to the etched surface, the etch pit will be a
perfect hexagonal with the apex at the center, which is the case of edge dislocation with
burger’s vector, 1/3<11-20> and 1/3<-1100>, and in case of screw dislocations, if the
screw dislocations have burger’s vector, [0001] or ½[0001], then the etch pit will be a
perfect hexagonal with the apex at the center also. On the other hand, If the dislocations
have the burger’s vector, 1/3<11-23> or 1/6<-2203>, then the dislocation line is oblique
to the etched surface and the apex of the pit will be off center. The study about the etch
pit formation for NaCl crystal was done by J. J. Gilman. (J. J. Gilman, et. al., 1958).
50
(a) (b) Figure 4.7 shows (a) the optical microscope image with high magnification and (b) scanning electron microscopy image in etch pit area.
Texas Tech University, Rac. G. Lee, May 2007
4.6 HRTEM images of AlN crystal
The Transmission Electron Microscope (TEM) has been used in all areas of
biological, biomedical investigations and material science. For the crystallographer,
metallurgist or semiconductor research scientist, current high voltage/high resolution
TEMs, utilizing 200 keV to 1 MeV, have permitted the routine imaging of atoms,
allowing materials researchers to monitor and design materials with custom-tailored
properties. Figure 4.8 (a) shows the actual image of the equipment and (b) shows the
electron diffraction to project the image onto the screen.
(a) (b) Figure 4.8 (a) The actual image of the TEM and (b) shows the electron diffraction
to project the image onto the screen.
51
Texas Tech University, Rac. G. Lee, May 2007
The defect and the quality of aluminum nitride crystal were examined by high
resolution transmission electron microscopy (HRTEM) and energy dispersive x-ray
analysis using ZEOL 2010F TEM which has field emission gun at the University of New
Mexico. Figure 4.9 shows the sample made by rough and smooth separately. Samples are
made by conventional method using disk cutter, dimple grinder and ion milling machine
(PIPS) to achieve very thin area. Both of the samples were tilted to the [11-20] zone axis.
(a) (b) Figure 4.9. The optical microscope image of the cross-sectional TEM sample for (a) smooth surface (b) rough surface. In order to take High Resolution Image, the beam source should be located in the
exact zone axis where we intend to study. Therefore, tilting sample is extremely
important technique in TEM. If the sample is thick enough to produce the Kikuchi line,
the sample will be tilted to the zone axis using Kikuchi line. The reason we form Kikuchi
pattern is that, if the specimen is thick enough, we will generate a large number of
scattered electrons which travel in all directions; i.e., they have been incoherently
scattered but not necessarily in-elastically scattered. They are sometimes referred to as
diffusely scattered electrons (David. B. Williams., Transmission Electron Microscopy,
52
Texas Tech University, Rac. G. Lee, May 2007
vol 2, chapter 19). Figure 4.10 shows the Kikuchi map for Hexagonal Structure (Defect
analysis in electron microscopy, appendix 2)
(a)
(b)
Figure 4.10 shows the Kikuchi map for Hexagonal Structure for (a) two standard triangles and (b) centered on [0001] zone axis.
53
Texas Tech University, Rac. G. Lee, May 2007
54
Figure 4.11 shows the TEM image taken from rough surface and indicate stacking
fault of the hexagonal system. Stacking sequence of perfect single crystal is
ABABAB…However, when the layer A slips to the 1/3<10-10> direction, C point of
figure 4.11 (b), from the perfect crystal, the next layer should be on a A point, thus
stacking sequence needs to be changed from ABABABAB...to ABABCACA...It is called
second order intrinsic stacking fault. First order intrinsic stacking fault occurs when the A
layer is eliminated and vacancies are produced in that area. After that, upper lattice (B)
needs to slip to the 1/3<10-10> direction to reduce the stacking energy. In this case,
stacking sequence needs to be changed from ABABAB…to ABABCBCB…We call the
stacking fault produced by vacancy agglomeration “intrinsic stacking fault”
(http://www.tf.uni-kiel.de/matwis/amat/def_en/kap_5/backbone/r5_4_1.html). Another
possible stacking fault is extrinsic stacking fault produced by interstitial agglomeration.
In this case the stacking sequence needs to be changed from ABABAB….to
ABABCABAB…. Figure 4.12 and 4.13 show the TEM images taken rough and smooth
surface separately.
Texas Tech University, Rac. G. Lee, May 2007
(a)
55
(b)
1/3<10-10>
A A
B B
C C
Figure 4.11. High resolution TEM image from the rough surface indicate (a) the stacking fault and (b) the slip phenomenon.
Texas Tech University, Rac. G. Lee, May 2007
2 nm2 nm
(a)
(b) (c) Figure 4.12. (a) High Resolution TEM image from the rough surface. (b) Fast Fourier Transform (FFT) of the rough surface showing that the beam direction is [11-20]. (c) Sample profile shows the lattice distance of the c-direction [0001] is 0.493nm.
56
Texas Tech University, Rac. G. Lee, May 2007
57
(a)
2 nm2 nm
(b) (c) Figure 4.13. (a) High Resolution TEM image from the smooth surface. (b) Fast Fourier Transform (FFT) of the smooth surface showing that the beam direction is [11-20]. (c) Sample profile shows the lattice distance of the c-direction [0001] is 0.499nm.
Texas Tech University, Rac. G. Lee, May 2007
4.7 CBED technique and experimental data Convergent Beam Electron Diffraction (CBED) uses a convergent beam of
electrons to overcome the spatial-resolution limitations of Selected Area Diffraction
(SAD) and limit the area of the specimen which contributes to the diffraction pattern
instead of using a parallel beam which is usually used to SAD. Each spot then becomes a
disc within which variations in intensity can usually be seen. Such patterns contain a
wealth of information about the symmetry and thickness of the crystal. Figure 4.14
illustrates the Ray diagram showing CBED pattern formation. The big advantage of
CBED over all other diffraction technique is that most of the information is generated
from minuscule region beyond the reach of other diffraction method (David. B. Williams.,
Transmission Electron Microscopy, vol 2, chapter 20).
Figure 4.14. The CBED pattern formation. A convergent beam at the specimen results in the formation of disks in the BFP of the object lens.
58
Texas Tech University, Rac. G. Lee, May 2007
59
Figure 4.15 shows the CBED image taken from (a) rough surface and (b) smooth surface.
Both CBED images are acquired from [11-20] beam direction and shows different
diffraction intensity for each disk. In order to determine symmetry, computational
simulation has been performed. For the simulation, Web-emap from the Center for Micro
analysis of Materials (CMM) was used with 200 electron potential (Kv), [11-20] zone
axis of aluminum nitride, [001] x-axis and 0.3 disk radius. Because CBED diffractions
need thicker sample region then other techniques such as HRTEM, EDAX and EELS,
where the best information is obtained from the thinnest specimens, 300, 400 and 500Å
of thickness was investigated in computational simulation. Figure 4.16 shows the
comparison of experimental CBED images with simulated CBED images. From the
simulation, bright spot indicates [0002] disk, and intensities depend on the sample
thickness. The diffraction intensity of Al is much stronger than nitride because of atomic
position. [0002] disk of smooth surface shows brighter than rough surface, and it might
be assumed that smooth surface is Al-polar crystal. In the Smooth surface, the intensity of
[000-2] disk which represents the N-polarity shows much darker than the intensity of
[0002] disk. Since [000-2] disk of rough surface has Al polar, it has higher intensity than
the intensity of [000-2] from smooth surface. For the simulation formulation, the Bloch
Wave Method was used, which was developed by J. M. Zuo (J. M. Zuo., 1998).
Texas Tech University, Rac. G. Lee, May 2007
90.00 µm
(a) (b)
90.00 µm
100.00 µm
(c) (d) Figure 4.15. CBED image of aluminum nitride crystal from (a), (b) rough and (c), (d) smooth surface.
60
Texas Tech University, Rac. G. Lee, May 2007
Figure 4.16. The comparison of experimental CBED images from rough and smooth surface of aluminum nitride crystal with simulated CBED images to determine sample symmetry. 4.8 Conclusions Aluminum nitride (AlN) crystal grown on tungsten (W) substrate was investigated
by Transmission Electron Microscope (TEM) and Scanning Electron Microscope (SEM).
Bulk AlN has rough and smooth surface because of the polarity of the sample. Energy
Dispersive X-ray Analysis (EDAX) shows more nitride in smooth surface and
Convergent Beam Electron Diffraction (CBED) result shows [0002] disk of rough
surface has more intensity because of the diffraction from Al atoms. Therefore it can be
concluded that the smooth surface has aluminum polarity (Al-Polar), and rough surface
has nitride polarity (N- Polar). From the rough surface, second order intrinsic dislocation
was observed.
61
Texas Tech University, Rac. G. Lee, May 2007
62
BIBLIOGRAPHY
ABAQUS/CAE Finite Element Analysis, Version 6.5, ABAQUS, Inc, 2004. W. M. Ashmawi, M. A. Zikry, K. Wang and R. R. Reeber, “Modeling of Residual Stress for Thermally Strained GaN/Al2O3 Heterostructures”, Journal of Crystal Growth, 2004, Vol. 266, Pages 415~422. J. H. Edgar, L. Liu, B. Liu, D. Jhuang, J. Chaudhuri, M. Kuball, S. Rajasingam, “Bulk AlN crystal growth”: self-seeding and seeding on 6H-SiC substrates, Journal of Crystal Growth, 2002, Vol. 246, Pages 187~193. Helmut Foll, University of Kiel, Partial Dislocations and Stacking Fault: Internet Source; http://www.tf.unikiel.de/matwis/amat/def_en/kap_5/backbone/r5_4_1.html J.J. Gilman, W. G. Johnston, G. W. Sears, “Dislocation Etch Pit Formation in Lithium Fluoride”, Journal of Applied Physics, 1958, Vol 29, Number 5. Z. Gu, J. H. Edgar, D. W Coffey, J. Chaudhuri, L. Nyakiti, R. G. Lee, J. Wen, “Defect-Selective Etching of Scandium Nitride Crystals”, Journal of Crystal Growth, 2006, Vol. 293, No. 2, Pages 242~246. Z. Gu, L. Du, J. H. Edgar, N. Nepal, I. Y. Lin, H. X. Jiang, R. Witt, “Sublimation Growth of Aluminum Nitride Crystals”, Journal of Crystal Growth, 2006, Vol.297, Pages 105~110. H. B. Huntington, The Elastic Constants of Crystals, Academic Press, Reprinted From Solid State Physics, 1958, Page 70. H. M. Ledbetter, S. Chevacharoenkui, R. F. Davis, “Monocrystal Elastic Constants of NbC”, J. Appl. Phys, 1986, Vol. 60, No. 5.1. R. G. Lee, A. Idesman, L, Nyakity, J. Chaudhuri, “Modeling of residual stress for aluminum nitride crystal growth by sublimation”-to be submitted B. Liu, J. H. Edgar, Z. Gu, D. Zhuang, B. Raghothamachar, M. Dudley, A. Saura, M. Kuball and H. M. Meyer, “The Durability of Various Crucible Materials for Aluminum nitride Crystal Growth by Sublimation”, Mat. Res. Soc. Internet Jour. MIJ-NSR , 2004, Vol. 9, Art. 6. M. Loretto, And R. Smallman, Defect analysis in electron microscopy, appendix 2, 1975, Chapman and Hall, London.
B. Monemar, “III-V nitrides-important future electronic materials”, Journal of Materials Science, Materials in Electronics, 1999, Vol.10, No. 4, Pages 227. J. F. Nye, Physical Properties of Crystals, Clarendon Press, Oxford, 1951, Pages 131~149. Robert R. Reeber, Kai Wang, “High Temperature Elastic Constant Prediction of Some Group ІІІ-Nitrides”, MRS Internet J. Nitride Semicond, 2001, Res. 6,3. Robert R. Reeber, Kai Wang, “Lattice Parameters and Thermal Expansion of Important Semiconductors and Their Substrates”, Mat. Res. Soc. Symp, 2000, Vol.622, T6.35.1~T6.35.6 J. Carlos Rojo, L. J. Schowalter, R. Gaska, M. Shur, M. A. Khan, J. Yang, D. D. Koleske, “Growth and Characterization of Epitaxial Layers on Aluminum Nitride Substrates Prepared from Bulk, Single Crystals”, Journal of Crystal Growth, 2002, Vol. 240, Page 508. R. Sclesser, R. Dalmau, R. Yakimova, Z. Sitar, Mater, “Growth of AlN Bulk Crystals from the Vapor Phase”, Res. Soc. Symp. Proc. 693, I9.4.1, 2002. G. A. Slack, T. F. McNelly, “AlN Single Crystals”, Journal of Crystal Growth, 1977, Vol.42, Page 560. G. A. Slack, T. F. McNelly, “ Growth of high purity AlN crystals”,Journal of Crystal Growth, 1976, Vol. 34, Page 263. S. Strite and H. Morkoc, “GaN, AlN and InN: A Review”, J. Vac. Sci. Technol, 1992, B 10, 1237. L. Lopez de la Torre, B. Winkler, J. Schreuer, K. Knorr, M. Avalos-Borja, “Elastic Properties of Tantalum Carbide (TaC)”, Solid State Communications, 2005, Vol 134, Issue 4, Pages 245~250. Y. S. Touloukian, R. K. Kirby, R. E. Taylor and T. Y. R. Lee (eds.), Thermo Physical Properties of Matter, Plenum Press, New York, 1977, Vol. 13. David. B. Williams, C. Barry Carter, Transmission Electron Microscopy, Plenum Press, New York, 1996, Vol 2, chapter 19. David. B. Williams, C. Barry Carter, Transmission Electron Microscopy, Plenum Press, New York, 1996, Vol 2, chapter 20. Ralph W. G. Wyckoff, Crystal Structures, Second Edition, 1963, Vol. 1, Interscience Publishers, Pages 85~93.
Texas Tech University, Rac. G. Lee, May 2007
64
I. Yonenaga, “Hardness of bulk single-crystal GaN and AlN”, Mat. Res. Soc. Internet Jour. MIJ-NSR, 2002, Vol. 7, Art. 6. J. M. Zuo, Annual Reports of the HVEM Laboratory Kyushu University, 1998, No 22, Pages 3-10.
Texas Tech University, Rac. G. Lee, May 2007
65
APPENDIX A
THERMAL STRESS DATA OF AlN USING W SUBSTRATE
Appendix A contains the completed result in graphical form for all the normal stress and in-plane stress of aluminum nitride grown on tungsten substrate.
Texas Tech University, Rac. G. Lee, May 2007
A.1 2-D result of aluminum nitride grown on tungsten substrate
-1 0 1 2 3 4 50.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
S-S
11 (G
Pa)
X1/dH
0 1 2 3 4-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 1 2 3 4-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
S-S
33 (G
Pa)
X1/dH
0 1 2 3 4-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
Figure A. 1 : The distribution of the stress component for H
L
dd 1 = 4 for 2-D case
66
Texas Tech University, Rac. G. Lee, May 2007
0 1 2 3 4 5 60.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
S-S
11 (G
Pa)
X1/dH
0 1 2 3 4 5 6
0.0
0.2
0.4
0.6
0.8
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 1 2 3 4 5 6-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
S-S
33 (G
Pa)
X1/dH
0 1 2 3 4 5 6-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
Figure A. 2 : The distribution of the stress component for H
L
dd 1 = 6 for 2-D case
67
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 80.30
0.35
0.40
0.45
0.50
0.55
0.60
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (a) σ 12
Figure A. 3 : The distribution of the stress component for H
L
dd 1 = 8 for 2-D case
68
Texas Tech University, Rac. G. Lee, May 2007
A.2 3-D result of AlN for the effect of the size
0 1 2 3 4
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
S-S
11 (G
Pa)
X1/dH
0 1 2 3 4
0.4
0.6
0.8
1.0
1.2
1.4
1.6
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 1 2 3 4
0.0
0.5
1.0
1.5
2.0
2.5
3.0
S-S
33 (G
Pa)
X1/dH
0 1 2 3 4-0.0016
-0.0014
-0.0012
-0.0010
-0.0008
-0.0006
-0.0004
-0.0002
0.0000
0.0002
S-S
12 (G
Pa)
X Axis Title
(c) σ 33 (d) σ 12
0 1 2 3 4-2.0
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
S-S
13 (G
Pa)
X1/dH
0 1 2 3 4
-0.0044
-0.0042
-0.0040
-0.0038
-0.0036
-0.0034
-0.0032
-0.0030
-0.0028
-0.0026
-0.0024
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 4 : The distribution of the stress for H
L
dd 1 = 4 for 3-D case, fixed. Hd
69
Texas Tech University, Rac. G. Lee, May 2007
0 1 2 3 4 5 60.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
S-S
11 (G
Pa)
X1/dH
0 1 2 3 4 5 60.4
0.6
0.8
1.0
1.2
1.4
1.6
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 1 2 3 4 5 6-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
S-S
33 (G
Pa)
X1/dH
0 1 2 3 4 5 6-0.0005
-0.0004
-0.0003
-0.0002
-0.0001
0.0000
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 1 2 3 4 5 6-2.0
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
S-S
13 (G
Pa)
X1/dH
0 1 2 3 4 5 6
-0.0014
-0.0012
-0.0010
-0.0008
-0.0006
-0.0004
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 5 : The distribution of the stress for H
L
dd 1 = 6 for 3-D case, fixed. Hd
70
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 80.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8
0.6
0.8
1.0
1.2
1.4
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8
0.0
0.5
1.0
1.5
2.0
2.5
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.00020
-0.00015
-0.00010
-0.00005
0.00000
0.00005
0.00010
0.00015
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8
0.00018
0.00020
0.00022
0.00024
0.00026
0.00028
0.00030
0.00032
0.00034
0.00036
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 6 : The distribution of the stress for H
L
dd 1 = 8 for 3-D case, fixed. Hd
71
Texas Tech University, Rac. G. Lee, May 2007
A.3 3-D result of AlN for the effect of the thickness of AlN
0 2 4 6 80.5
0.6
0.7
0.8
0.9
1.0
1.1
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
1.05
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8
-0.00008
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8
0.00008
0.00009
0.00010
0.00011
0.00012
0.00013
0.00014
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 7 : The distribution of the stress for H
L
dd 1 = 32 for 3-D case, fixed. 1Ld
72
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 8
0.55
0.60
0.65
0.70
0.75
0.80
S-S
11 (G
Pa)
X1/dH
0 2 4 6 80.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.00008
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
0.00008
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8-1.0
-0.8
-0.6
-0.4
-0.2
0.0
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8
-0.000015
-0.000010
-0.000005
0.000000
0.000005
0.000010
0.000015
0.000020
0.000025
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 8 : The distribution of the stress for H
L
dd 1 = 16 for 3-D case, fixed. 1Ld
73
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 8
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
S-S
11 (G
Pa)
X1/dH
0 2 4 6 80.5
0.6
0.7
0.8
0.9
1.0
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.00010
-0.00008
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
0.00008
0.00010
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8-0.00002
-0.00001
0.00000
0.00001
0.00002
0.00003
0.00004
S-S
23 (G
Pa)
X1/dH
(a) σ 13 (a) σ 23
Figure A. 9 : The distribution of the stress for H
L
dd 1 = 8 for 3-D case, fixed. 1Ld
74
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 80.4
0.5
0.6
0.7
0.8
0.9
1.0
S-S
11 (G
Pa)
X1/dH
0 2 4 6 80.4
0.5
0.6
0.7
0.8
0.9
1.0
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8
0.0
0.5
1.0
1.5
2.0
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.00015
-0.00010
-0.00005
0.00000
0.00005
0.00010
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8
-0.00002
-0.00001
0.00000
0.00001
0.00002
0.00003
0.00004
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 10 : The distribution of the stress for H
L
dd 1 = 4 for 3-D case, fixed. 1Ld
75
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 8
0.4
0.5
0.6
0.7
0.8
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8
0.4
0.5
0.6
0.7
0.8
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8
0.000
0.001
0.002
0.003
0.004
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8-1.0
-0.8
-0.6
-0.4
-0.2
0.0
X1/dH
S-S
13 (G
Pa)
0 2 4 6 8-0.006
-0.005
-0.004
-0.003
-0.002
-0.001
0.000
0.001
X1/dH
S-S
23 (G
Pa)
(e) σ 13 (f) σ 23
Figure A. 11 : The distribution of the stress for H
L
dd 1 = 2 for 3-D case, fixed. 1Ld
76
Texas Tech University, Rac. G. Lee, May 2007
A.4 3-D result of AlN for the effect of the interaction of islands
0 2 4 6 8 10 120.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8 10 120.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8 10 12-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8 10 12
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8 10 12
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8 10 12
-0.00002
0.00000
0.00002
0.00004
0.00006
0.00008
S-S
23 (
GPa
)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 12 : The distribution of the stress component for H
L
dd 2 = 0 for 3-D case
77
Texas Tech University, Rac. G. Lee, May 2007
0 2 4 6 8 10 12 14
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8 10 12 140.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8 10 12 14
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8 10 12 14-0.00010
-0.00008
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8 10 12 14-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8 10 12 14-0.00010
-0.00008
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
0.00008
0.00010
S-S
23 (G
Pa)
X1/dH
(d) σ 13 (e) σ 23
Figure A. 13 : The distribution of the stress component for H
L
dd 2 = 0.2 for 3-D case
78
Texas Tech University, Rac. G. Lee, May 2007
-2 0 2 4 6 8 10 12 140.4
0.6
0.8
1.0
1.2
1.4
S-S
11 (G
Pa)
X1/dH
-2 0 2 4 6 8 10 12 14
0.5
0.6
0.7
0.8
0.9
1.0
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
-2 0 2 4 6 8 10 12 14
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
S-S
33 (G
Pa)
X1/dH
-2 0 2 4 6 8 10 12 14-0.00003
-0.00002
-0.00001
0.00000
0.00001
0.00002
0.00003
0.00004
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
-2 0 2 4 6 8 10 12 14-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
S-S
13 (G
Pa)
X1/dH
-2 0 2 4 6 8 10 12 14-0.00010
-0.00008
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
0.00008
0.00010
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23
Figure A. 14 : The distribution of the stress component for H
L
dd 2 = 1 for 3-D case
79
Texas Tech University, Rac. G. Lee, May 2007
A.5 3-D result of AlN for the effect of the mis-oriented grain
-2 0 2 4 6 8 10 12 14 16 18 200.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
S-S
11 (G
Pa)
X1/dH
-2 0 2 4 6 8 10 12 14 16 18 20
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
-2 0 2 4 6 8 10 12 14 16 18 20-0.4
-0.2
0.0
0.2
0.4
0.6
S-S
33 (G
Pa)
X1/dH
-2 0 2 4 6 8 10 12 14 16 18 20
-1.0
-0.5
0.0
0.5
1.0
1.5
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
-2 0 2 4 6 8 10 12 14 16 18 20-0.30
-0.25
-0.20
-0.15
-0.10
-0.05
0.00
S-S
13 (G
Pa)
X1/dH
-2 0 2 4 6 8 10 12 14 16 18 20
-0.0020
-0.0015
-0.0010
-0.0005
0.0000
0.0005
0.0010
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23 Figure A. 15 : The distribution of the stress component for mis-oriented grain.
80
Texas Tech University, Rac. G. Lee, May 2007
81
APPENDIX B
THERMAL STRESS DATA OF AlN USING POSSIBLE CRUCIBLES
Appendix B contains the completed result in graphical form for all the normal stress and in-plane stress of aluminum nitride grown on sapphire, tantalum carbide, niobium carbide, silicon carbide and tungsten substrate.
Texas Tech University, Rac. G. Lee, May 2007
B.1 The distribution of the thermal residual stress of AlN grown on sapphire substrate
0 2 4 6 8-7
-6
-5
-4
-3
-2
-1
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8-5.0
-4.5
-4.0
-3.5
-3.0
-2.5
-2.0
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8
-8
-7
-6
-5
-4
-3
-2
-1
0
1
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8
-0.0002
-0.0001
0.0000
0.0001
0.0002
0.0003
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8
0
1
2
3
4
5
6
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8
-0.00055
-0.00050
-0.00045
-0.00040
-0.00035
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23 Figure B.1 : The distribution of the thermal residual stress of AlN grown on sapphire substrate
82
Texas Tech University, Rac. G. Lee, May 2007
B.2 The distribution of the thermal residual stress of AlN grown on TaC substrate
0 2 4 6 8-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8-1.2
-1.1
-1.0
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-2.0
-1.5
-1.0
-0.5
0.0
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8
-0.00006
-0.00004
-0.00002
0.00000
0.00002
0.00004
0.00006
0.00008
0.00010
0.00012
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8-0.00023
-0.00022
-0.00021
-0.00020
-0.00019
-0.00018
-0.00017
-0.00016
-0.00015
-0.00014
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23 Figure B.2 : The distribution of the thermal residual stress of AlN grown on TaC substrate
83
Texas Tech University, Rac. G. Lee, May 2007
B.3 The distribution of the thermal residual stress of AlN grown on NbC substrate
0 2 4 6 8
-2.6
-2.4
-2.2
-2.0
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8
-2.0
-1.8
-1.6
-1.4
-1.2
-1.0
-0.8
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.00015
-0.00010
-0.00005
0.00000
0.00005
0.00010
0.00015
0.00020
0.00025
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8
0.0
0.5
1.0
1.5
2.0
2.5
S-S
13 (G
Pa)
X1/dH
0 2 4 6 8
-0.00046
-0.00044
-0.00042
-0.00040
-0.00038
-0.00036
-0.00034
-0.00032
-0.00030
-0.00028
-0.00026
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23 Figure B.3 : The distribution of the thermal residual stress of AlN grown on NbC substrate
84
Texas Tech University, Rac. G. Lee, May 2007
B.4 The distribution of the thermal residual stress of AlN grown on SiC substrate
0 2 4 6 81.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
S-S
11 (G
Pa)
X1/dH
0 2 4 6 8
1.5
2.0
2.5
3.0
3.5
4.0
S-S
22 (G
Pa)
X1/dH
(a) σ 11 (b) σ 22
0 2 4 6 8-1
0
1
2
3
4
5
6
S-S
33 (G
Pa)
X1/dH
0 2 4 6 8-0.00025
-0.00020
-0.00015
-0.00010
-0.00005
0.00000
0.00005
0.00010
0.00015
S-S
12 (G
Pa)
X1/dH
(c) σ 33 (d) σ 12
0 2 4 6 8
-4
-3
-2
-1
0
S-S
13 (G
Pa)
X1/dH
0 2 4 6 80.00050
0.00055
0.00060
0.00065
0.00070
0.00075
0.00080
0.00085
S-S
23 (G
Pa)
X1/dH
(e) σ 13 (f) σ 23 Figure B.4 : The distribution of the thermal residual stress of AlN grown on SiC substrate
85
Texas Tech University, Rac. G. Lee, May 2007
86
APPENDIX C
THE DIFFRACTION PATTERN OF AlN CRYSTAL
Appendix C contains the computational simulations of the diffraction pattern of aluminum nitride in graphical form for the important zone axis.
Texas Tech University, Rac. G. Lee, May 2007
C.1 The diffraction pattern of aluminum nitride using [0001] beam direction
Figure C.1 : The diffraction pattern of aluminum nitride using [0001] beam direction
87
Texas Tech University, Rac. G. Lee, May 2007
C.2 The diffraction pattern of aluminum nitride using [11-20] beam direction
Figure C.2 : The diffraction pattern of aluminum nitride using [11-20] beam direction
88
Texas Tech University, Rac. G. Lee, May 2007
C.3 The diffraction pattern of aluminum nitride using [1-100] beam direction
Figure C.3 : The diffraction pattern of aluminum nitride using [1-100] beam direction
89
PERMISSION TO COPY
In presenting this thesis in partial fulfillment of the requirements for a master’s
degree at Texas Tech University or Texas Tech University Health Sciences Center, I
agree that the Library and my major department shall make it freely available for research
purposes. Permission to copy this thesis for scholarly purposes may be granted by the
Director of the Library or my major professor. It is understood that any copying or
publication of this thesis for financial gain shall not be allowed without my further
written permission and that any user may be liable for copyright infringement.
Agree (Permission is granted.)
____Rac Gyu Lee_________________________________ _3/ 23/ 2007_ ____ Student Signature Date Disagree (Permission is not granted.) _______________________________________________ _________________ Student Signature Date