0% found this document useful (0 votes)
20 views6 pages

Structural Investigation of The Bias-Enhanced Nucleation and Growth of Diamond Films by Microwave Plasma Chemical Vapor Deposition

111

Uploaded by

wangqisicc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
20 views6 pages

Structural Investigation of The Bias-Enhanced Nucleation and Growth of Diamond Films by Microwave Plasma Chemical Vapor Deposition

111

Uploaded by

wangqisicc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Structural Investigation of the Bias-Enhanced Nucleation and

Growth of Diamond Films by Microwave Plasma Chemical


Vapor Deposition
Do-Geun Kim° and Tae-Yeon Seong
Department of Materials Science and Engineering, Kwangju Institute of Science and Technology (K-JIST),
Kwangju 506-712, Korea

Young-Joan Baik
Division of Ceramics, Korea Institute of Science and Technology, Seoul, 136-791, Korea

ABSTRACT
Transmission electron microscopy (TEM), transmission electron diffraction (TED), atomic force microscopy (AFM),
and scanning electron microscopy have been used to investigate the initial nucleation process and growth behavior of dia-
mond films by microwave plasma chemical vapor deposition. TED examination revealed epitaxial relations between the
-SxC and the Si, and the diamond and the p-SiC, which depended on the bias-enhanced nucleation (BEN) time and
methane (CH4) concentration. The highly oriented (001) diamond films were obtained after 25 mm BEN for 4% CH4 and
20 mm BEN for 8% CH4. TEM revealed the f3-SiC crystallites 2—25 nm across and the diamond crystallites 3—40 nm in
size, which depended on the CH4 concentration and the BEN time. As the BEN time increased, the density of the 3-SiC
cyrstallites increased from —2.7 >< 1011 to —3.4 X 10u cm2, while that of the diamond crystallites varied from —2.0 x 100
to —4.0 X 10" cm2. Discrepancy between the densities obtained using TEM and AFM is discussed. It is shown that the
heteroepitaxially oriented diamond crystallites are critically important for the growth of the highly (001)-oriented dia-
mond films, although the heteroepitaxially oriented 3-SiC crystallites could serve as nucleation sites for the growth of the
diamond films.

Introduction perature 820°C; time 20 h. In order to explore the effects of


Diamond films are of increasing importance because of nucleation time and CH4 concentration on the nucleation
their potential technological applications, ranging from behavior and subsequent growth of diamond films, three
wear-resistance coatings to high-temperature electronic different sets of samples were prepared. The first set was
devices.1'2 An advance in electronic technologies, however, BEN-treated wafers (using 4% CH4). An ac bias voltage
requires the growth of highly (001)-oriented diamond was applied for different times of 20, 25, 30, and 40 mm,
films on commercially available low-cost substrates such (termed here "20 BEN wafer," etc.). The second set was
as silicon. Bias-enhanced nucleation (BEN) has proved to also BEN-treated wafers (using 8% CH4): ac bias voltage
be extremely effective in the growth of heteroepitaxially was applied for different times of 15, 20, 25, and 40 mm,
oriented diamond films on Si wafers.37 A number of stud- (termed also "15 BEN wafer," etc.). The third set was dia-
ies811 performed to understand nucleation mechanisms mond films grown at 820°C for 20 h on these bias-nucleat-
involved in a BEN process have shown that an amorphous ed Si wafers. [0011 plan-view thin foils were prepared by
carbon and an SiC buffer layer, or an increase in surface chemical etching using a solution of HNO2/HF. <110> cross
mobility,6 play an important role in the growth of diamond section thin foils were prepared by mechanical polishing
films. Gerber et al.,12 investigating the role of the surface and Ar ion milling using a liquid N2 cold stage. TEM and
diffusion process during BEN of diamond, suggested that TED studies were performed in a JEM2O1O instrument
the flux of energetic ions was the critical factor for bias operated at 200 kV. Atomic force microscopy (AFM) (PSI)
enhancement of the nucleation density of diamond on sil- and scanning electron microcopy (SEM) (JSM5 800) were
icon. Reinke et al.,1° investigating nucleation phenomena used to characterize surface topography of the BEN
in a negative dc bias-enhanced diamond deposition, wafers and the diamond films.
reported that diamond particles formed during the pre- Results and Discussion
treatment served as nucleation sites for subsequent dia-
mond growth. In this work, we describe detailed transmis- Effects of 4% CH4 on BEN—Figure 1 shows [001] TED
sion electron microscopy (TEM) and transmission electron patterns from the wafers which were bias-nucleated with
diffraction (TED) examination of the initial phases formed 4% CH for 20—40 mm. The patterns reveal the main sili-
during a BEN process and how the BEN time and methane con structure spots and additional diffracted features
(CH4) affects the density of the phases and the epitaxially which correspond to 3-SiC and diamond. The characteristic
oriented growth of diamond. features are dependent on the BEN time. For the 20 BEN
Experimental
The BEN and growth of diamond films have been car- Table I. The experimental parameters used for BEN and growth of
ried out in a microwave plasma chemical vapor deposition
diamond films.
(MPCVD) system (ASTeX). The substrates were mirror-
polished p-type (001) Si wafers, which were cleaned in situ Bias-enhanced
by H2 plasma etching. The nucleation and growth condi- Diamond
Parameters nucleation
tions are summarized in Table I. The nucleation param- growth
eters are: CH4 concentration, 4% and 8% in H2; ac bias CH4 concentration in H2 (%) 4 2
voltage, 200 V, (60 Hz); microwave power 1 kW; pressure 8
20 Torr; substrate temperature —-850°C; time, 20—40 mm. ac bias voltage (Vrm) 200 0
The growth parameters are: CH4 concentration 2% in H2; Time (mm) 20, 25, 30, 40 1,200
microwave power 1.5 kW; pressure 30 Torr; substrate tern- 15, 20, 25, 40
Also with: Division of Ceramics, Korea Institute of Science and Microwavve power (kW) 1 1.5
Technology, Seoul, Korea.
Pressure (Torr) 20 30
Substrate temperature (°C) —845 820
J. E/ectroc/im. Soc., Vol. 145, No. 6, June 1998 The Electrochemical Society, Inc.
2095

Downloaded on 2015-05-08 to IP 128.122.253.228 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
vo J. tiecrrocnem. soc., vol. i q, NO. b, dune 1 c) I lie hiectrocriemical Society, Inc.

Fig. 1. [0011 TED patterns


from the wafers which were
bias-nucleated (with 4% CH4) at
—850°C for (a) 20, (b) 25, (c)
30, and (d) 40 mm.

wafer, Fig. la, the pattern shows {22O).SC and (200}.S.C that the (11 1}-oriented 3-SiC crystallites virtually disap-
diffuse diffracted intensities (marked A and B, respective- peared in the 25 and 30 BEN wafers and appeared again in
ly) with texture maxima parallel to the (22O} and (4O0} the 40 BEN wafer (Fig. ld). The explanation for this, how-
spots. This indicates that some of the 3-SiC crystallites are ever, is not completely clear at the moment.
heteroepitaxially oriented to the Si substrate, although Diffracted features associated with the diamond crystal-
some are slightly rotated (± 3—5°) about the 10011 direction lites are shown in the patterns in Fig. lb—d. Absence of
with regard to Si(220}, as suggested by the elongation of such features in the 20 BEN wafer (Fig. la) indicates that
diffuse intensity. There also exists a {lll), diffracted 20 mm is not long enough to nucleate the diamond crys-
ring with maximum intensity (marked C), implying that tallites. For the 25 BEN wafer (Fig. lb), there are well-
some of the 3-SiC crystallites are randomly oriented. defined (220), spots (marked D) parallel to the {220(, and
Comparison of the (lll}. and {22O}., spots shows that hence the {22O).. spots, implying that some of the dia-
the (111)-oriented 13-SiC crystallites are predominant over mond crystallites are heteroepitaxially oriented to the
the (110)-oriented ones. For the 25 and 30 BEN wafers, 3-SiC, although few are slightly rotated. For the 30 BEN
Fig. lb and c, respectively, the (220). and (200} in- wafer (Fig. lc), there are also (220), spots (marked D)
tensities (marked A and B, respectively) with maxima par- roughly parallel to the {22O}. spots, but they are ill-
allel to the {220} and {400} spots are well defined, indi- defined as compared to those of the 25 BEN wafer. For the
cating that most of the 13-SiC crystallites are 40 BEN wafer, Fig. id, there are {ll1), {220)d,, and
heteroepitaxially oriented to the Si substrate. It is noted {3ll}d,. diffracted rings, indicating the presence of ran-

Fig. 2. (a) {220} . and (b)


{220}d. DF images trom the 25
BEN wafer showing the small
bright blobs corresponding to
the 13-SiC and diamond crystal-
lites, respectively.

Downloaded on 2015-05-08 to IP 128.122.253.228 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
,j. tiecrrocnem. ,oc., voi. iq, NO. ID, ,june I ne hlectrocflemlcal socIety, Inc. ujl
A:\BEN25-2, HDF
25 BEN wafer (Fig. 2a) shows bright blobs corresponding
to the 13-SiC crystallites. The crystallites vary in size from
—3 to --6 nm and are randomly distributed.
Analogous (22O)d,. DF examination showed that as the
BEN time increased from 25 to 40 mm, the crystallite den-
sity varied from —2.0 X iO to —4.1 X 10" cm2 and the
crystallites changed in size from —5 to —30 nm. For exam-
ple, the {2201d,. DF imagej from the 25 BEN wafer (Fig. 2b)
shows bright blobs 5—20 nm across, corresponding to the
diamond crystallites.
AFM results showed the formation of a number of small
hillocks. As the BEN time increased, the hillock density
decreased from —7.3 X 10' to —2.0 x iO cm 2 and the
hillocks varied in size from —15 to —300 nm. For example,
in Fig. 3 an AFM image of the 25 BEN wafer is shown.
There are a number of small hillocks 15—45 nm in size and
--2.0 nm in height. The hillock density was estimated to be
X 10'° cm2, which corresponds to the value between
those of the diamond and 13-SiC crystallites.
Fig. 3. AFM image obtained from the 25 BEN wafer.
Effects of 8% CH4 on BEN.—Figure 4 shows [0011 pat-
terns obtained from the wafers which were bias-nucleated
domly oriented diamond crystallites. The precise mechan- with 8% CH4 for 15—40 mm. The patterns reveal diffracted
ism why the crystallographic orientation of the diamond features which correspond to 3-SiC and diamond. For the
crystallites depends on BEN time is not clear as yet. How- 15 BEN wafer (Fig. 4a), the pattern shows (220)13sc and
ever, it may be explained in terms of either defect-induced {200}130c diffuse intensities (marked A and B, respectively)
tilt or the overgrowth of the diamond on the (3-SiC with with texture maxima parallel to the (22O}, and {400)S
mixed orientations. spots. This indicates that some of the 13-SiC crystallites are
To directly reveal the (3-SiC and diamond crystallites, heteroepitaxially oriented to the Si substrate, although
10011 plan-view TEM dark field (DF) images were obtained some are slightly rotated 2—5° about the [001] direction
using the {22O}S,C and {220}d,. spots of the BEN wafers. with reference to Si{220}. For the 20 BEN wafer (Fig. 4b),
(220},(. DF results showed that the (110)-oriented 13-SiC the {22O) and {200) intensities (marked A and B,
crystallites were randomly distributed over the Si sub- respectively) with maxima parallel to the (220} and {400}
strate. The crystallite density increased progressively from spots are well defined, indicating that most of the 3-SiC
—2.7 x 10i to —1.6 x 10 cm2, and the crystallites varied crystallites are heteroepitaxially oriented to the Si sub-
from —2 to —10 nm in size as the BEN time increased from strate. There are also {ili) intensities (marked C). For
20 to 40 mm. For example, the {220) DF image from the the 25 BEN wafer (Fig. 4c) there is the (220} diffuse

Fig. 4. [001] TED patterns


from the wafers which were
bias-nucleated (with 8% CH4) at
—850°C for (a) 15, (b) 20, (c)
25, and (d) 40 mm.

Downloaded on 2015-05-08 to IP 128.122.253.228 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
2098 J. Electrochem. Soc., Vol. 145, No. 6, June 1998 The Electrochemical Society, Inc.

Fig. 5. (a) {22O}. and (b)


{220},.. DF images from the 20
BEN wafer showing the small
bright blobs corresponding to
the n-SiC and diamond crystal-
lites, respectively.

intensity (marked A). There are also weak (111 dif- ed the presence of hillocks (—45 nm across and 10—90 nm
fracted rings with maximum intensity (marked C), imply- high) with a density of 7.5 X iO cm2 containing SiC, dia-
ing the presence of the randomly oriented 3-SiC crystal- mond, and amorphous components. Gerber et al.'2 also
lites. The {1l1},c intensity has virtually disappeared in reported the presence of hillocks in dc bias-nucleated
the 40 BEN wafer. wafers consisting mainly of carbon. In the present work,
Features associated with the diamond crystallites are however, plan-view and cross-sectional TEM and TED
shown in the patterns (Fig. 4b-d). For the 20 BEN wafer examination showed no evidence for the presence of amor-
(Fig. 4b), there are well-defined {220},,,, spots (marked D) phous carbon or graphite.
parallel to the (220}, spots, implying that some of the
diamond crystallites are heteroepitaxially oriented to the Effects of CH4 concentration on the growth of diamond
n-SiC, although few are rotated 2-5°. It is noted that films—In order to assess the evolution of surface mor-
{lli}d spots (marked D,,,) also start to appear. For the 25 phology as a function of the BEN time, SEM images were
and 40 BEN wafers (Fig. 4c and d, respectively), there are obtained from the diamond films deposited on the wafers
{lll}d,, {220}d, and {3ll}d. diffracted rings, indicating the which were bias-nucleated with 4% CH4 for 20—40 mm
presence of randomly oriented diamond crystallites. (Fig. 7). Comparison of the images shows that the growth
However, for the 25 BEN wafer, there are well-defined behavior of the diamond films is strongly dependent upon
(220}da spots (marked by the arrow) parallel to the {220} the BEN time. For the film on the 20 BEN wafer, there are
spots, implying that some of the diamond crystallites are a number of diamond particles ranging from —6 to —9 im
heteroepitaxially oriented to the 3-SiC. in diam. The surface is about 70% covered with the dia-
[001] TEM DF images were obtained using the {22Ol mond particles. For the film on the 25 BEN wafer, the
and {22O}d spots of the BEN wafers to reveal the )3-SiC crystallinity of the film is greatly improved, i.e., the film
and diamond crystallites. (220} DF results showed that surface is highly (001)-oriented (Fig. 7b). However, with
the (110)-oriented 3-SiC crystallites were randomly dis- additional increase in the BEN time (30 mm), the crys-
tributed over the Si surface. The crystallite density tallinity of the film is to some extent degraded (Fig. 7c).
decreased slowly from —4.5 x 10" to —3.4 x 10" cm2 and Eventually, a BEN time of 40 mm results in the film with
the crystallites varied from —3 to —25 nm in size as the ill-defined surface which completely lost the alignment
BEN time increased from 15 to 40 mm. For example, the of grains (Fig. 7d). The percentage of the (001)-oriented
{220}, DF image from the 20 BEN wafer (Fig. 5a) shows diamond grains is —6, -'-90, —45, and —0% for 20, 25, 30,
bright blobs corresponding to the 3-SiC crystallites. The and 40 mm, Fig. 7a, b, c, and d, respectively.
crystallites change in size from —3 to —14 nm. In Fig. 8 are shown similar SEM images from the dia-
Similar DF results showed that as the BEN time mond films deposited on the wafers which were bias-nucle-
increased from 20 to 40 mm, the crystallite density ated with 8% CR, for 15—40 mm. Overall growth behavior
increased from —3.6 x iO' to —1.8 x 10° cm2 and the is similar to that of the 4% CH4 diamond films (Fig. 7). It is
crystallites varied in size from —3 to —40 nm. For exam- noted that the highly (001)-oriented film is obtained at a
ple, the DF image from the 20 BEN wafer (Fig. 5b) shows BEN time of 20 mm, which is shorter than that of the 4%
bright blobs 3—20 nm across, corresponding to the dia- CH4. As for the films deposited on the 20 BEN wafer, the
mond crystallites. It is noted that there are planar defects percentage of the (001)-oriented diamond grains is —90%.
such as microtwins and stacking faults within the dia-
mond crystallites, as indicated by the arrow.'3 Such defects
were introduced to relieve the large lattice-mismatch
between the diamond and the l3-SiC. 13,14
AFM images of the BEN wafers revealed a number of BEN 8% 2Orn
small hillocks 35—170 nm in size and 1.6—34 nm in height.
As the BEN time varied, the hillock density decreased
from —6.2 x 1010 to —7.5 x i0 cm . For example, an AFM
image of the 20 BEN wafer (Fig. 6) reveals a number of
small hillocks —100 nm in size and —23 nm in height. The
hillock density was estimated to be —9.0 X iO cm2, which
is reasonably comparable to that of the diamond crystal-
lites but much less than that of the 3-SiC crystallites.
It should be stressed that the AFM images cannot differ-
entiate the diamond crystallites from the 3-SiC crystal-
lites. Furthermore, the size of the hillocks (Fig. 3 and 6) is
much larger than those of the diamond and 3-SiC crystal-
lites (Fig. 2 and 5). This discrepancy strongly suggests that
the hillocks are not individual crystallites but agglomera-
tions consisting of the diamond and 13-SiC crystallites.
Wurzinger et al.,' investigating a dc BEN process, report- Fig. 6. AFM image obtained from the 20 BEN wafer.

Downloaded on 2015-05-08 to IP 128.122.253.228 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
.i. tiecrrocnem. soc., vol. i 4, Plo. i, dune 1 d c I ne lectrocnemical society, Inc.

Fig. 7. SEM images from the


diamond films deposited on the
(a) 20, (b) 25, Cc) 30, and (d) 40
BEN wafers which were biased
with 4%CH4.

TED and TEM results showed that the 13-SiC crystallites Si substrate in all the BEN wafers. There is also epitaxial
are always formed during the BEN process and there is to relation between the diamond and 13-SiC crystallites in
some extent epitaxial relation between the 3-SiC and the some of the BEN wafers. The epitaxial relation is most

p ____ Fig. 8. SEM images from the


diamond films deposited on the
(a) 15, (b) 20, (c), 25, and (d) 40
-. BEN wafers which were biased
with 8% CH4.

, ..b' .•.
•d.
-

Downloaded on 2015-05-08 to IP 128.122.253.228 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
2100 J. Electrochem. Soc., Vol. 145, No.6, June 1998 The Electrochemical Society, Inc.

well defined in the 25 BEN (4% Cli,) and 20 BEN (8% Manuscript submitted September 29, 1997.
CH4) wafers (Fig. lb and 4b) in which growth of the high-
ly (001)-oriented diamond films is achieved (Fig. 7b and Korea Institute of Science and Technology assisted in
8b). As for both CH4 concentrations, the densities of the meeting the publication costs of this article.
heteroepitaxially oriented 3-SiC crystallites are far
greater than those of the diamond crystallites. For exam- REFERENCES
ple, for 4% CH4, the density of 13-SiC crystallites in the 25
BEN wafer is —7.0 X 1011 cm2, whereas that of the dia- 1. W. Zhu, B. R. Stoner, B. E. Williams, and J. ¶1'. Glass,
mond crystallites is —2.0 x l0 cm2. Thus, the difference Proc. IEEE, 79, 621 (1991).
in the densities of the crystallites and the crystallinity of 2. E. Rohrer, C. F. 0. Graeff, R. Janssen, C. E. Nebel, M.
the diamond films may suggest that although the het- Stutzmann, H. Guttler, and R. Zachai, Phys. Rev. B,
54, 7874 (1996).
eroepitaxially oriented 13-SiC crystailites could serve as 3. 5. Yugo, T. Kanai, T. Kimura, and T. Muto, Appl. Phys.
nucleation sites for the growth of the diamond films, the Lett., 58, 1036 (1991).
heteroepitaxially oriented diamond crystallites are criti- 4. B. R. Stoner and J. T. Glass, Appl. Phys. Lett., 60, 698
cally important for the growth of the highly (001)-orient- (1992).
ed diamond films. 5. X. Jiang, C. P. Klages, M. Rosler, R. Zachai, M.
Hartweg, and H.-J. Fusser, Appl. Phys. A, 57, 483
Summary and Conclusions (1993).
TEM, TED, AFM, and SEM have been used to study 6. X. Jiang, K. Schiffmann, and C. P Klages, Phys. Rev. B,
BEN and diamond films grown by MPCVD to investigate 50, 8402 (1994).
the initial phases formed during a BEN process and the 7. 5. D. Woltei T. H. Borst, A. Vescan, and E. Kohn, Appl.
BEN time dependence of diamond films. The il-SiC crys- Phys. Lett., 68, 3558 (1996).
8. J. Gerber, M. Weiler, 0. Sohr, K. Jung, and H. Ehrhardt,
tallites were always formed during the BEN process and Diamond Relat. Mater., 3, 506 (1994).
they served as nucleation sites for the diamond crystal- 9. B. R. Stoner, G.-H. M. Ma, S. D. Wolter, and J. T. Glass,
lites. TED examination showed that there exist epitaxial Phys. Rev. B, 45, 11067 (1992).
relations between 13-SiC and Si, and diamond and Il-SiC, 10. P Reinke, P Kania, P. Oelhafen, and P. Guggenheim,
which were dependent on BEN time and CR, concentra- Appl. Phys. Lett., 68, 22 (1996).
tion. The growth of highly oriented (001) diamond films 11. P Wurzinger, P. Pongratz, J. Gerber, and H. Ehrhardt,
depended on the BEN time and CR, concentration; it was Diamond Relat. Mater., 5, 345 (1996).
obtained after 20 and 15 mm BEN for 4% and 8% CH4, 12. J. Gerber, J. Robertson, S. Sattel, and H. Ehrhardt,
Diamond Relat. Mater., 5, 261 (1996).
respectively. It is concluded that the heteroepitaxially ori- 13. P Tachibana, K. Hayashi, and K. Kobahi, Appl. Phys.
ented diamond crystallites nucleated during the BEN Lett., 68, 1491 (1996).
process play a crucial role in the growth of the highly 14. X. Jiang and C. L. Jia, Appl. Phys. Lett., 67, 1197
(001)-oriented diamond films. (1995).

Wear-Contact Problems and Modeling of


Chemical Mechanical Polishing
0. G. Chekina and 1. M. Keer
Department of Civil Engineering, Northwestern University, Evanston, Illinois 60208-31 09, USA

H. Liang
Cabot Corporation, Microelectronics Materials Division, Aurora, Illinois 60504, USA

ABSTRACT
Wafer shape and contact pressure evolution during chemical mechanical polishing, and the characteristics of the steady-
state regime are analyzed on the basis of approaches developed in contact mechanics. Nonpianarity caused by the geometri-
cal nonuniformity (erosion) and by the presence of different material on the surface (recess) is considered. The possibility of
the process optimization and the determination of system parameters based on the polished surface profiles is discussed.

Infroduction in this process.2'3 Mainly, the polishing of SiO, by aqueous


The needs in the semiconductor industry demand tech- silica is studied4'5; however, CMP is also used successfully
niques for high quality wafer surface planarization to for the polishing of ceramics,6 metals (Cu, Al, Ti, W), and
make multilevel interconnects and 3D packaging possible. polymers.7' Another approach to the CMP modeling is
Insufficient planarity can result in difficulties with subse- based on the consideration of abrasive action of a single
quent lithography and dry etching.' solid particle." However, the effect of global planariza-
Chemical mechanical polishing (CMP) is the leading tion, which is the main advantage of. CMP in comparison
method of global wafer planarization. The CMP process is with other planarization techniques, e.g., chemical polish-
achieved by sliding a wafer surface on a relatively soft ing (etching), cannot be explained by chemical reaction or
polymeric pad with a polishing slurry containing chemical abrasive action scale models. Explanation of global pla-
compounds and ultrafine abrasive particles between these narization is possible only if some nonlocal effects are
two surfaces. involved. A phenomenological model, taking into account
In spite of the wide application of CMP, its complete the nonlocal nature of CMP was proposed,'1 however, the
model is not yet developed.' Some investigations are de- parameters of the model were not connected with real
voted to chemical processes taking place during CMP, parameters of the system wafer-slurry-pad.
since chemical composition of slurries and the material of The necessity of considering nonlocal effects in CMP
the abrasive particles are reported to be an essential factor makes the mechanical aspects of the process the question

Downloaded on 2015-05-08 to IP 128.122.253.228 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

You might also like