chemical vapor deposition
Cheng-Li Lin, Peng-Sen Chen, Chun-Li Chang, and Mao-Chieh Chen
Citation: Journal of Vacuum Science & Technology B 20, 1947 (2002); doi: 10.1116/1.1502697 View online: http://dx.doi.org/10.1116/1.1502697
View Table of Contents: http://scitation.aip.org/content/avs/journal/jvstb/20/5?ver=pdfcov
Published by the AVS: Science & Technology of Materials, Interfaces, and Processing
Articles you may be interested in
Evaluation of integrity and barrier performance of atomic layer deposited W N x C y films on plasma enhanced chemical vapor deposited Si O 2 for Cu metallization
Appl. Phys. Lett. 89, 081913 (2006); 10.1063/1.2338768
Structural and morphological control of nanosized Cu islands on SiO 2 using a Ti underlayer
J. Appl. Phys. 94, 3492 (2003); 10.1063/1.1597972
Nucleation of copper on TiN and SiO 2 from the reaction of hexafluoroacetylacetonate copper(I) trimethylvinylsilane
Appl. Phys. Lett. 80, 2604 (2002); 10.1063/1.1469687
Effectiveness of Ti, TiN, Ta, TaN, and W 2 N as barriers for the integration of low-k dielectric hydrogen silsesquioxane
J. Vac. Sci. Technol. B 18, 221 (2000); 10.1116/1.591176
Electromigration study of Al/low dielectric constant polymer and Al / SiO 2 dioxide line structures
This work investigates the chemical vapor deposition Cu films deposited on the TaN substrate with and without an H2-plasma treatment prior to Cu film deposition and the effect of postdeposition
thermal annealing. The Cu films deposited on the H2-plasma-treated TaN substrate have a number
of favorable properties over the films deposited on the TaN substrate without the plasma treatment. These include an increased共111兲-preferred orientation, smoother film surface, and a larger effective deposition rate. However, the Cu films deposited on the H2-plasma-treated substrate have a higher
electrical resistivity, presumably due to the smaller grain size. The postdeposition thermal annealing enhanced the共111兲-preferred orientation and decreased the resistivity of the as-deposited Cu films. We presume that the H2-plasma treatment, resulted in a dense and uniform distribution of hydrogen radicals adsorbed on the substrate surface, leading to the shortening of incubation time and the formation of Cu films with a smoother surface and enhanced 共100兲-preferred orientation. A combined process including H2-plasma substrate treatment prior to Cu film deposition and postdeposition thermal annealing at an appropriate temperature in N2 ambient, is proposed for the
advantage of low-resistivity and high 共111兲-oriented Cu film deposition. © 2002 American
Vacuum Society. 关DOI: 10.1116/1.1502697兴
I. INTRODUCTION
The application of copper to multilevel interconnection offers a number of advantages over the conventional Al-based metallization, including lower bulk resistivity 共1.67
⍀ cm兲, superior electromigration resistance, and higher re-sistance to stress-induced voids.1,2 There are various tech-niques of copper film deposition, such as chemical vapor deposition 共CVD兲,3,4 conventional as well as ionized metal plasma 共IMP兲 physical vapor deposition,5 and electrochemi-cal deposition共ECD兲 including electroplating and electroless plating;6,7among these, the CVD method has the advantages of superior step coverage and excellent gap filling capability for high-aspect-ratio vias and trenches,4making it the most promising technique of Cu film deposition for future inte-grated circuits 共ICs兲 application. To implement copper into metal lines and high-aspect-ratio vias interconnection, the damascene process is developed to cope with the difficult Cu dry etching problem. Although the Cu ECD combined with IMP of a thin Cu seed layer and barrier layers provides a suitable solution for IC technologies above 0.13 m, depo-sition techniques for a conformal and continuous thin barrier as well as a conformal and void-free Cu film filling into deep subquarter-micron vias, such as CVD, are eventually un-avoidable for future generation devices.8A number of barrier materials have been used as substrates for CVD of Cu films, such as W, Ti, TiN, Ta, and TaN,3,9–11 among these, TaN exhibits a superb barrier capability against Cu diffusion.12 However, the chemically vapor-deposited Cu films on a TaN substrate exhibited a fairly low peak-ratio of Cu共111兲/
Cu共200兲 preferred orientation in the x-ray diffraction 共XRD兲 spectrum,3,11which is unfavorable for electromigration resis-tance. Since the nucleation process and the microstructure of Cu films are very sensitive to the substrate surface conditions,13,14there have been a number of studies concern-ing the plasma treatment on the substrate surface prior to the Cu film deposition, resulting in the deposited Cu film with enhanced 共111兲-preferred orientation and superior film property.15–17
In this work, we investigate the effects of TaN substrate pretreatment by hydrogen (H2) plasma on copper CVD
us-ing a multichamber low-pressure CVD system. The effects of postdeposition thermal annealing are also investigated. II. EXPERIMENT
Figure 1 shows the schematic diagram of the multicham-ber Cu CVD apparatus built for this study. The apparatus consists of a cluster of four chambers and a direct liquid injection 共DLI兲 system for precursor delivery. The chamber cluster is composed of a sample-loading chamber 共for loading/unloading samples兲, a pretreatment chamber, a reac-tion chamber, and a transfer chamber. The pretreatment chamber is used to preclean and/or modify the substrate sur-face by plasma treatment. CVD of Cu films is to be carried out in the reaction chamber. The transfer chamber, which houses a robot arm, is designed to handle the transfer of a substrate wafer to and from each chamber. In the reaction chamber, there is a shower-head injector, through which the Cu precursor is introduced into the reaction chamber in a stream of carrier gas. Under the injector, there is a substrate susceptor that can be heated by a resistive heating element up to a maximum temperature of 400 °C. The susceptor is a兲Electronic mail: [email protected]
also rotatable for better uniformity of film deposition. The shower-head injector is moveable in the vertical direction, so that the distance between the injector and the substrate wafer can be adjusted. The side wall of the reaction chamber and the precursor injector are kept at a temperature of 45 °C by the circulation of warm water to prevent Cu deposition of precursor condensation.
The Cu precursor used in this study is Cu 共1,1,1,5,5,5-hexafluoroacetylacetonate兲trimethylvinylsilane 关Cu共hfac兲T-MVS兴 with 2.4 wt % TMVS additive, which enhances the stability of the precursor.18 –20The liquid Cu precursor is de-livered by the DLI system consisting of a liquid flow con-troller 共LFC兲 and a controlled evaporation mixer 共CEM兲 共Fig. 1兲. Initially, the liquid precursor is propelled by N2 gas
through the LFC. It is then vaporized in the CEM and mixed with the carrier gas. Helium共He兲 is used as the carrier gas in this study. The precursor-saturated carrier gas is introduced into the reaction chamber through the gas injector.
In this work, TaN was used as the substrate for the CVD of Cu films. The TaN layer of 50 nm thickness was sputter deposited on a thermal-oxide共500 nm thick兲-covered Si wa-fer. A dc magnetron sputtering system with a base pressure of 1.5⫻10⫺6Torr was used to reactively sputter a Ta target 共99.99% purity兲 in an Ar/N2 gas mixture at a pressure of 7.6
mTorr for TaN film deposition. The flow rates of Ar and N2
were 24 and 6 sccm, respectively, for making the gas mix-ture, and the TaN film was sputter deposited at a power of
150 W. The TaN-coated substrate wafer was loaded into the multichamber Cu CVD system. When the pressure of the sample-loading chamber reached 10⫺6 Torr, the substrate wafer共together with the substrate holder兲 was transferred to the pretreatment chamber or reaction chamber via the trans-fer chamber depending on the process requirement of whether the plasma pretreatment on substrate was to be per-formed or not. In this study, H2 plasma treatment was
per-formed at 50 W power for 10 min under the following con-ditions: H2 flow rate 15 sccm, gas pressure 40 mTorr,
substrate temperature 80 °C, and self-dc bias⫺259 V. After the plasma pretreatment, the substrate wafer was transferred to the reaction chamber for Cu film deposition. Prior to Cu film deposition, the substrate wafer was heated to the desired deposition temperature with He carrier gas flowing at 25 sccm and the chamber pressure maintained at 150 mTorr. Usually, approximately 1 h is required for the substrate wafer to reach the present temperature. In this study, Cu CVD was performed over a temperature range of 120 °C to 240 °C at a pressure of 150 mTorr with a precursor flow rate of 0.4 mL/ min and a He carrier gas flow rate of 25 sccm. Major pro-cessing conditions and the parameters of the Cu CVD system used in this study are summarized in Table I. At the end of Cu film deposition, the substrate wafer was cooled in the ambient of He at a pressure of 150 mTorr.
The thickness of Cu films was measured using a DekTek profiler on the patterned Cu films and was verified by cross-FIG. 1. Schematic diagram of multichamber Cu CVD apparatus built for this study.
TABLEI. Processing conditions and parameters of the Cu CVD system. Pretreatment chamber
共H2-plasma operating conditions兲
Reaction chamber 共Cu film deposition conditions兲
Substrate temperature共°C兲 80 Substrate temperature共°C兲 120–240
Operating pressure共mTorr兲 40 Operating pressure共mTorr兲 150
H2gas flow rate共sccm兲 15 Cu precursor flow rate共mL/min兲 0.4
rf power共W兲 50 CEMatemperature共°C兲 70
Self-dc bias共V兲 ⫺259 Carrier gas共He兲 flow rate 共sccm兲 25
Pretreatment time共min兲 10 Substrate holder rotation speed共rpm兲 10
Gas-injector/susceptor distance共cm兲 2
Delivery line temperature共°C兲 72
Reactor wall temperature共°C兲 45
a
CEM indicates controlled evaporation mixer.
sectional scanning electron microscopy 共SEM兲. SEM was also used to observe the surface morphology of the deposited Cu films. A four-point probe was employed to measure the sheet resistance. Auger electron spectroscopy共AES兲 and sec-ondary ion mass spectrometry共SIMS兲 were used to analyze the impurity content in the Cu films. The crystal structure of Cu film was identified by x-ray diffraction 共XRD兲 analysis. The surface roughnesses of the Cu and substrate TaN films were evaluated by atomic force microscopy共AFM兲. Ruther-ford backscattering spectroscopy 共RBS兲 was used to deter-mine the composition of the TaN substrate layer.
III. RESULTS AND DISCUSSION
A. Chemical vapor deposition Cu films on TaN substrates without plasma pretreatment
Copper films were chemically vapor deposited on reac-tively sputtered TaN substrates, which have a resistivity of 0.6 m⍀ cm and a composition of TaN1.2, as determined by
Rutherford backscattering. The chemical reaction of Cu CVD using Cu共hfac兲TMVS as a precursor with He as a car-rier gas proceeds on the substrate surface by a facile dispro-portionation as follows:21,22
2Cu⫹1共hfac兲TMVS共g兲→2Cu⫹1共hfac兲TMVS共s兲, 共1兲 2Cu⫹1共hfac兲TMVS共s兲→2Cu⫹1共hfac兲共s兲⫹2TMVS共g兲,
共2兲 2Cu⫹1共hfac兲共s兲→Cu共s兲⫹Cu⫹2共hfac兲2共s兲, 共3兲
Cu⫹2共hfac兲2共s兲→Cu⫹2共hfac兲2共g兲, 共4兲
where共g兲 denotes ‘‘gas phase’’ and 共s兲 denotes ‘‘adsorbed on substrate surface.’’ The reaction step 3 关Eq. 共3兲兴 is the key step of Cu nucleation on the substrate surface, which in-volves a process of electron exchange between the adsorbed Cu⫹1(hfac) and the substrate surface. Thus, it is easier to deposit Cu films on the conducting substrate than the insu-lating substrate. Since the chemical reaction of Cu CVD in-volves a thermal dissociation of the Cu precursor关Eq. 共2兲兴, a higher temperature would result in a higher rate of deposi-tion. Figure 2 shows the effective deposition rate of Cu films
as a function of substrate temperature 共Arrhenius plot兲 at a constant pressure of 150 mTorr with a He carrier gas flow rate of 25 sccm and a liquid Cu precursor flow rate of 0.4 mL/min. The effective deposition rate of Cu film was calcu-lated using the measured thickness of a Cu film deposited for 10 min. The activation energy Ea was determined to be 7.35 kcal/mol by the Arrhenius equation
R⫽R0exp共⫺Ea/kT兲, 共5兲
where R is the deposition rate, R0 is the Arrhenius
pre-exponential constant or frequency factor, k is the Boltzmann constant, and T is the absolute temperature. This value of Ea is smaller than the value of 17.90 kcal/mol共deposited on TiN substrate at 0.5 Torr兲 reported in literature.20 The difference in the values of the activation energy is presumably due to different deposition conditions. The resistivity of Cu films was calculated using the measured sheet resistances and films thicknesses. Figure 3 illustrates the resistivity of Cu films as a function of deposition temperature at a deposition pressure of 150 mTorr. The resistivity of CVD Cu films is
closely related to the impurity content and
microstructure.23,24 The slightly higher resistivity at low deposition temperatures is presumably due to higher con-tamination of residual impurities from the reaction byprod-ucts, while the high resistivity at high deposition tempera-tures results from the higher contamination of impurities in the film as well as the porous film structure. The optimal temperature for Cu film deposition for minimization of resis-tivity under the present deposition conditions appears to be around 160 °C 共2.30 ⍀ cm兲. Figure 4 shows SEM micro-graphs of the surface morphology of Cu films deposited at various temperatures for a deposition time of 10 min. The grain size of Cu increases with the deposition temperature. At higher deposition temperatures 共200 °C–240 °C兲, the Cu films exhibit stacked grains featuring clear boundaries as well as voids between large grains. Figure 5 illustrates the Auger electron spectroscopy 共AES兲 depth profiles of CVD Cu films deposited on TaN substrates. The most notable im-purities contaminated in CVD Cu films are carbon 共C兲 and oxygen共O兲, which may result from incomplete desorption of hfac ligand during the CVD process.24Although fluorine共F兲 FIG. 2. Effective deposition rate vs substrate temperature共Arrhenius plot兲 at
a constant pressure of 150 mTorr for TaN substrate.
FIG. 3. Resistivity of Cu film vs deposition temperature for Cu films
impurity was not detected in the AES depth profile, the SIMS analysis shown in Fig. 6 reveals obvious F contamination in Cu film.
B. Nucleation and surface morphology observation The nucleation process and the microstructure of Cu films are very sensitive to the substrate surface conditions, which play an important role in the CVD of Cu films. Figure 7 shows atomic force microscopy 共AFM兲 images of the TaN substrate before and after H2-plasma treatment at 50 W for 10 min. The H2-plasma treatment slightly improved the sur-face smoothness of the TaN substrate. In order to study the effect of substrate pretreatment by H2plasma, the nucleation
process of Cu films was investigated. Figure 8 shows SEM micrographs for the nucleation process of Cu films deposited at 160 °C on a TaN substrate with and without a H2-plasma
treatment. On the TaN substrate without H2-plasma
treat-ment, the Cu-containing adspecies were sparsely nucleated on the substrate surface, and the subsequent adspecies tend to nucleate on the existing Cu nuclei rather than the TaN
sub-strate. As a result, the Cu nuclei grew into Cu grains sparsely distributed on the substrate, while a few new Cu nuclei may also randomly nucleated directly on the TaN substrate, form-ing smaller Cu grains关Fig. 8共a兲兴. After 2 min deposition, all grains grew larger, though a few smaller new grains are
scat-FIG. 5. AES depth profiles of Cu films deposited on TaN substrate at 160 °C
for 10 min共without plasma treatment prior to Cu deposition兲.
FIG. 6. SIMS depth profiles of Cu films deposited on TaN substrate at
160 °C for 10 min共without plasma treatment prior to Cu deposition兲.
FIG. 7. AFM images showing surface morphology of TaN substrates共a兲 without plasma treatment, and共b兲 with H2-plasma treatment at 50 W共with
40 mTorr pressure for a plasma time of 10 min兲. FIG. 4. SEM micrographs showing surface morphology of Cu films
depos-ited on TaN substrates at a temperature of共a兲 120 °C, 共b兲 160 °C, 共c兲 200 °C, and 共d兲 240 °C. The films were deposited at a constant pressure of 150 mTorr for 10 min共without plasma treatment prior to Cu deposition兲.
tered among the larger ones 关Fig. 8共b兲兴. On the H2-plasma-treated TaN substrate, the Cu-containing
adspe-cies 关Cu共hfac兲兴 was easily and uniformly nucleated on the substrate surface, forming dense and small Cu nuclei after 1 min deposition关Fig. 8共c兲兴. The Cu-containing adspecies con-tinuously adsorbed on the surface of the Cu nuclei to proceed with the growth of copper nuclei, resulting in larger Cu grains and a continuous Cu film after 2 min deposition关Fig. 8共d兲兴. A similar result was reported in literature that the nucleation of Cu on the H2-plasma-treated TiN substrate led
to a dense distribution of Cu grains.25 The nucleation of Cu on the TaN substrate is closely related to the substrate sur-face conditions. On the H2-plasma-treated TaN substrate
sur-face, there are many uniformly adsorbed hydrogen radicals and/or atoms.25,26 This surface hydrogen may enhance the chemisorption of Cu-containing adspecies 关Cu共hfac兲兴 on the substrate surface through the direct bonding between the sur-face hydrogen and the CH radical in the Cu precursor.26 Thus, Cu can be easily and uniformly nucleated on the H2-plasma-treated TaN substrate surface, resulting in
short-ened incubation time and densely distributed Cu grains. The dense and uniform distribution of Cu grains on the H2-plasma-treated TaN substrate indicates that the substrate
has a higher surface energy or lower interfacial energy, lead-ing to nucleation of Cu with a smaller wettlead-ing angle共contact angle兲.16,27,28 The smaller wetting angle would enhance the growth of Cu film in two dimensions共layer growth兲, forming the most stable and共111兲 closely packed configuration.16The SEM micrograph in Fig. 9 shows the surface morphology of a Cu film deposited at 160 °C for 10 min on the H2-plasma-treated TaN substrate. In comparison with the
corresponding Cu film deposited on the TaN substrate with-out H2-plasma treatment关Fig. 4共b兲兴, the surface morphology
of the Cu film deposited on the H2-plasma-treated TaN
sub-strate is characterized by regularly shaped smaller grains. As a result, superior surface smoothness is expected for the Cu film deposited on the H2-plasma-treated TaN substrate. This
is indicated by the results of AFM analysis shown in Fig. 10.
The Cu film deposited on the TaN substrate without H2-plasma treatment shows irregular Cu grains with an
av-erage surface roughness root-mean square 共rms兲 of 56 nm, while the Cu film deposited on the H2-plasma-treated TaN substrate shows regularly shaped smaller grains with an av-erage surface roughness 共rms兲 of 20 nm.
The thicknesses of the Cu films deposited on the TaN substrate with and without an H2 plasma treatment were measured to be 509 and 408 nm, respectively, for a 10 min deposition at a substrate temperature of 160 °C. The thicker Cu film on the H2-plasma-treated TaN substrate is
presum-ably due to the shorter incubation time and the much denser Cu nuclei during the nucleation stage. The resistivity of the Cu film deposited on the H2-plasma-treated TaN substrate
was determined to be 2.82⍀ cm, which is higher than that of the Cu film deposited on the substrate without the plasma FIG. 8. SEM micrographs showing Cu nucleation at 160 °C for共a兲 1 min
and共b兲 2 min deposition on TaN substrate without plasma treatment, and 共c兲 1 min and共d兲 2 min deposition on TaN substrate with H2-plasma treatment.
FIG. 9. SEM micrograph showing surface morphology of Cu film deposited at 160 °C and 150 mTorr for 10 min on H2-plasma-treated TaN substrate.
FIG. 10. AFM images showing surface morphology of Cu films deposited on TaN substrates共a兲 without plasma treatment and 共b兲 with H2-plasma
treatment 共2.30 ⍀ cm兲, presumably due to the smaller Cu grains and thus a higher boundary density.
C. Preferred orientation and postdeposition thermal annealing
The H2-plasma treatment on the TaN substrate also affects
the preferred orientation of deposited Cu films. Figure 11 shows the XRD spectra for Cu films deposited on TaN sub-strates with and without an H2-plasma treatment. It can be seen that the peak ratio of Cu共111兲 to Cu共200兲 reflection increased from 2.80 to 4.15 resulting from the H2-plasma treatment on the TaN substrate. With the H2-plasma treat-ment, there are many uniformly adsorbed hydrogen radicals on the TaN substrate surface, which enable the chemisorption of Cu-containing adspecies on the substrate surface, leading to the formation of Cu film with enhanced 共111兲-preferred orientation because the 共111兲 texture is the most stable configuration.16,28
Copper films deposited on the TaN substrate were ther-mally annealed at 400 °C for 30 min in an N2 ambient.
Fig-ure 12 shows the surface morphology of the thermally an-nealed Cu film deposited on the H2-plasma-treated TaN
substrate. In comparison with the as-deposited Cu film共Fig. 9兲, the thermally annealed Cu film reveals a closer contact between Cu grains, similar to those reported in literature for CVD Cu films annealed in Ar and Ar/H2 ambients.
29
More-over, thermal annealing also resulted in a decrease in film resistivity as well as an increase in the peak ratio of Cu共111兲 to Cu共200兲 reflection, as shown in Table II. The decrease in film resistivity is presumably due to the closer contact
be-tween the Cu grains. In addition, we presume that the Cu film deposited on the H2-plasma-treated substrate possesses a
uniform surface energy and little variation of film stress be-cause of the better regular arrangement of Cu grains. Ther-mal annealing would reduce the grain boundary and surface energy of the Cu film, resulting in the recrystallization of Cu grains to forming the most stable 共111兲 texture.16,30On the other hand, Cu grains of different sizes are irregularly ar-ranged for the Cu film deposited on the TaN substrate with-out a plasma treatment; thus, there are nonuniform surface energy and nonuniform films stress, and this nonuniformity would be reduced in order to reduce the total system energy during thermal annealing.30As a result, there is not enough driving force to recrystallize the Cu grains for the formation of the most stable共111兲 texture, and the improvement of the Cu共111兲/Cu共200兲 reflection peak ratio was relatively moder-ate as compared with the Cu film deposited on the H2-plasma-treated substrate. A similar observation was also
reported for CVD Cu films thermally annealed at 450 °C in Ar and Ar/H2 ambients.29 With regard to adhesion of Cu
films, a Scotch tape pulling test was used to qualify the ad-hesion between the CVD Cu films and the TaN substrate. All samples with the Cu film deposited at 160 °C and 150 mTorr, irrespective of substrate pretreatment by H2-plasma and/or
postthermal annealing, passed the Scotch tape test. In sum-mary, we conclude that the TaN substrate surface with uni-form and dense hydrogen adatom resulting from the H2-plasma treatment, is responsible for the improvement of
various Cu film properties.
FIG. 12. SEM micrograph showing surface morphology of Cu film deposited on H2-plasma-treated TaN substrate followed by thermal annealing at
400 °C for 30 min in N2ambient.
FIG. 11. XRD spectra of Cu films deposited on TaN substrates with and
without H2-plasma treatment prior to Cu deposition.
TABLEII. Effects of thermal annealing共400 °C/30 min兲 on Cu film resistivity and peak ratio of Cu共111兲/Cu共200兲
reflection. H2-plasma treatment on TaN substrate Film resistivity 共⍀ cm兲 Cu共111兲/Cu共200兲 peak ratio
As deposited After annealing As deposited After annealing
No 2.30 2.15 2.80 3.10
Yes 2.82 2.25 4.15 5.52
number of favorable properties over the films deposited on the TaN substrate without the plasma treatment. These in-clude an increased共111兲-preferred orientation, smoother film surface, and a larger effective deposition rate. However, the Cu films deposited on the H2-plasma-treated substrate have a
higher electrical resistivity, presumably due to the smaller grain size, and thus higher grain-boundary density. Postdepo-sition thermal annealing resulted in the reduction of electri-cal resistivity and the increase of Cu共111兲/Cu共200兲 reflection ratio. We presume that the H2-plasma treatment resulted in a
dense and uniform distribution of hydrogen radicals共and/or atoms兲 adsorbed on the substrate surface, leading to the shortening of incubation time and the formation of Cu films with a smoother surface and enhanced共111兲-preferred orien-tation. A combined process including H2 plasma substrate
treatment prior to Cu film deposition and postdeposition ther-mal annealing at an appropriate temperature共e.g., 400 °C兲 in an N2 ambient, is proposed for the advantage of
low-resistivity and high共111兲-oriented Cu film deposition. In this way, Cu film with a resistivity of 2.25⍀ cm and a Cu共111兲/ Cu共200兲 peak ratio of 5.52 was obtained.
ACKNOWLEDGMENTS
The authors wish to thank the Semiconductor Research Center of National Chiao-Tung University and the National Nano Device Laboratory for providing an excellent process-ing environment. This work was supported by the National Science Council, ROC, under Contract No. NSC-90-2215-E-009-096.
1A. Jain, T. T. Kodas, R. Jairath, and M. J. Hampden-Smith, J. Vac. Sci.
Technol. B 11, 2107共1993兲.
2N. Awaya, H. Inokawa, E. Yamamoto, Y. Okazaki, M. Miyake, Y. Arita,
and T. Kobayashi, IEEE Trans. Electron Devices 43, 1206共1996兲.
3R. Kroger, M. Eizenberg, D. Cong, N. Yoshida, L. Y. Chen, S.
Ra-maswami, and D. Carl, J. Electrochem. Soc. 146, 3248共1999兲.
Y. Shacham-Diamand and S. Lopatin, Microelectron. Eng. 37, 77共1997兲.
8P. Motte, M. Proust, J. Torres, Y. Gobil, Y. Morand, J. Palleau, R. Pantel,
and M. Juhel, Microelectron. Eng. 50, 369共2000兲.
9
N. I. Cho and Y. Sul, Mater. Sci. Eng., B 72, 184共2000兲.
10S. Voss, S. Gandikota, L. Y. Chen, R. Tao, D. Cong, A. Duboust, N.
Yoshida, and S. Ramaswami, Microelectron. Eng. 50, 501共2000兲.
11R. Kroger, M. Eizenberg, D. Cong, N. Yoshida, L. Y. Chen, S.
Ra-maswami, and D. Carl, Microelectron. Eng. 50, 375共2000兲.
12
M. T. Wang, Y. C. Lin, and M. C. Chen, J. Electrochem. Soc. 145, 2538 共1998兲.
13D. H. Kim, R. H. J. Wentorf, and W. N. Gill, J. Appl. Phys. 74, 5164 共1993兲.
14
K. Hanaoka, H. Ohnishi, and K. Tachibana, Jpn. J. Appl. Phys., Part 1 34, 2430共1995兲.
15Y. S. Kim, D. Jung, and S. K. Min, Thin Solid Films 349, 36共1999兲. 16K. Kamoshida and Y. Ito, J. Vac. Sci. Technol. B 15, 961共1997兲. 17C. L. Lin, P. S. Chen, and M. C. Chen, Jpn. J. Appl. Phys., Part 1 41, 280
共2002兲.
18J. A. T. Norman, B. A. Mutamore, P. N. Dyer, D. A. Roberts, and A. K.
Hochberg, J. Phys. IV 1, C2-271共1991兲.
19J. C. Chiou, Y. J. Chen, and M. C. Chen, J. Electron. Mater. 23, 383 共1994兲.
20
T. Nguyen, L. J. Charneski, and S. T. Hsu, J. Electrochem. Soc. 144, 2829 共1997兲.
21J. A. T. Norman, D. A. Roberts, A. K. Hochberg, P. Smith, G. A. Petersen,
J. E. Parmeter, C. A. Apblett, and T. R. Omstead, Thin Solid Films 262, 46共1995兲.
22N. Awaya and Y. Arita, Thin Solid Films 262, 12共1995兲.
23S. S. Yoon, J. S. Min, and J. S. Chun, J. Mater. Sci. 30, 2029共1995兲. 24P. J. Lin and M. C. Chen, Jpn. J. Appl. Phys., Part 1 38, 4863共1999兲. 25
J. H. Lee, J. H. Lee, K. J. Hwang, J. Y. Kim, C. G. Suk, and S. Y. Choi, Thin Solid Films 375, 132共2000兲.
26A. E. Kaloyeros, C. Dettelbacher, E. T. Eisenbraun, W. A. Lanford, and P.
J. Toscano, Mater. Res. Soc. Symp. Proc. 229, 123共1991兲.
27R. J. Stokes and D. F. Evans, Fundamentals of Interfacial Engineering 共Wiley-VCH, New York, 1997兲, p. 59.
28H. Toyoda, T. Kawanoue, S. Ito, M. Hasunuma, and H. Kaneko, in
Stress-Induced Phenomena in Metallization, Third International Workshop, Palo Alto, CA, June 1995, edited by P. S. Ho, J. Bravman, C. Y. Li, and J. Sanchez 共American Institute of Physics, New York, 1996兲, No. 373, p. 169.
29S. K. Rha, W. J. Lee, S. Y. Lee, D. W. Kim, C. O. Park, and S. S. Chun,
Jpn. J. Appl. Phys., Part 1 35, 5781共1996兲.
