Department of Materials Science and Engineering, National Chiao-Tung University, Hsinchu, Taiwan 30050 The study prepared Ru/RuNxbilayer barriers on mesoporous SiO2(mp-SiO2) dielectric layers for direct Cu electroplating applica-tions using in situ two-step plasma-enhanced atomic chemical vapor deposition (PEALD). For the 5 nm thick Ru/RuNxbilayer de-posited at 200C, obvious thermal decomposition begins at temperatures lower than 400C. Copper can be successfully electroplated on the as-deposited Ru/RuNxbilayer, and the Cu/Ru/RuNx/mp-SiO2film stack can withstand thermal treatment at temperatures up to 500C without significant physical and chemical degradations according to TEM and SIMS analyses. The study shows that the electroplated Cu layer behaves like a passivation layer that improves the thermal stability of the Ru/RuNxbarrier during the thermal annealing. Pull-off tensile test shows that interfaces in the Cu/Ru/RuNx/mp-SiO2film stack have good adhesion strength, but delamination occurs at the interface between the Ru/RuNxbilayer and themp-SiO2layer at 600C, resulting in Cu and Ru diffusion into the dielectric layer. The study has demonstrated that the PEALD Ru/RuNxbilayer structure prepared using the in situ two-step approach is suitable for the seedless Cu electroplating process in nanometer scale interconnect technology.
VC2011 The Electrochemical Society. [DOI: 10.1149/1.3554734] All rights reserved.
Manuscript submitted October 18, 2010; revised manuscript received December 30, 2010. Published March 2, 2011.
Traditional Cu interconnect technology requires a Cu seed layer grown on the diffusion barrier for the Cu electroplating process. A seed/glue/barrier tri-layer structure, such as the Cu seed/Ta/TaN film stack, is generally used in the dual-damascene process for better adhesion and microstructure. Physical vapor deposition (PVD) is the most widely used method to deposit the Cu seed layer and the bar-rier layer. However, the PVD approach may present scaling difficul-ties for sub-32 nm technology nodes because of intrinsic drawbacks of the PVD process, such as poor conformality and thickness uni-formity in nanometer scale. To avoid these intrinsic shortcomings associated with the PVD process, atomic layer chemical vapor depo-sition (ALD) has long been regarded as a promising alternative to the PVD process for the dual-damascene technology. The feature of self-limiting surface reactions in the ALD process provides accurate thickness control in the atomic scale as well as excellent conformity and uniformity for patterns with a high aspect ratio over a large area.
Because a diffusion barrier with a thickness <5 nm is required for sub-32 nm technology nodes,1various diffusion barriers deposited by ALD have been studied in the past decade.2–4 Besides, PVD-deposited Cu seed layers have a minimum thickness limit of30– 40 nm, and are, therefore, inappropriate for the Cu dual-damascene technique for sub-32 nm nodes because of the poor confomality in-herent to PVD.5Much effort has been done in order to replace the Cu PVD seed layer with materials that can be properly processed in the dual-damascene technique for future technology nodes. Some Cu-plateable metals, such as Ru, Ir, and Os,6–8have attracted con-siderable attention because they also have good barrier properties for copper. Among various candidate metals, Ru receives the most study because of its low electrical resistivity and negligible solubil-ity in copper at 900C.9Ru is also a stable transition metal in air and has good wettability with Cu. However, polycrystalline Ru is unsuitable for serving as a Cu diffusion barrier because its columnar grain structure provides a fast diffusion path for Cu atoms at 450C.10On the other hand, amorphous Ru demonstrates better dif-fusion barrier properties than the polycrystalline Ru barrier, and has a higher adhesion strength with Cu.11When a high level of nitrogen atoms is dissolved in an Ru film, the nitrided Ru film usually has a nanocrystalline structure with a very short range ordering,12or even exhibits an amorphous form. Several recent studies demonstrated that nitrided Ru barriers prepared by PVD had satisfactory barrier characteristics.11–14However, RuNxhas a relatively high electrical
resistance, making direct Cu electrodeposition difficult on the RuNx
surface. In this study, we deposited Ru/RuNx bilayer barriers on
mesoporous SiO2(thereafter denoted bymp-SiO2) dielectric layers
by an in situ two-step plasma-enhanced ALD (PEALD) process, and copper could be directly electrodeposited on the as-deposited Ru/RuNx barrier. The as-deposited Ru/RuNx bilayer is thermally
degraded at temperatures below 300C as a result of decomposition of the RuNxlayer. However, if thermal annealing is carried out after
the Ru/RuNx bilayer is electroplated by a Cu layer, the RuNx
decomposition is suppressed, and the Cu/Ru/RuNx/mp-SiO2
multi-layer stack can withstand thermal treatment at temperatures up to 500C without significant degradation in mechanical and electrical properties.
Experimental
PEALD RuNxthin films were deposited on plasma-treated
mp-SiO2thin films, which were spin-coated on 6-in. p-type Si wafers, at
200C using bis(ethylcyclopentadienyl)ruthenium [Ru(EtCp)2] and
a gas mixture of H2/N2as the Ru precursor and the reducing
reac-tant, respectively. The preparation of themp-SiO2thin film and the
pore sealing of the porous dielectric by oxygen plasma treatment have been described previously.15,16The PEALD system had a base pressure of 5 104Torr. One cycle of the PEALD RuNx
deposi-tion process consisted of four consecutive gas pulses, including the Ru(EtCp)2precursor injection pulse, the Ar purge pulse with a flow
rate of 50 sccm, the H2/N2plasma nitridation pulse, and another Ar
purge pulse. During the PEALD RuNxdeposition, the stainless steel
tube containing the precursor was maintained at 75C to produce an adequate amount of vapor, which was carried by argon gas at a flow rate of 50 sccm. The gas line delivering Ru(EtCp)2was heated at
80C by heating tapes to prevent the precursor from condensation. For the deposition of metallic Ru thin films, an H2 gas pulse was
used instead of the H2/N2pulse. Cu electroplating was carried out at
room temperature using a custom plating system with an electrolyte bath comprised of H2SO4, CuSO45H2O and HCl. The study used a
current density of 10 mA/cm2 for the Cu electroplating. Thermal anneal treatment of the samples was carried out in a vaccum funace at 107Torr.
The surface morphology of the RuNxthin film was examined by
scanning electron microscopy (SEM, JEOL JSM-6500 F) and atomic force microscopy (AFM, Digital Instruments. NanoScope E). The surface chemical composition was characterized by x-ray photoelectron spectroscopy (XPS, thermo VG 350) using the Mg Ka
x-ray radiation. The XPS spectra presented in the report have been calibrated with the binding energy of the Pt 4f7/2electron. A grazing
incident angle X-ray diffraction (XRD) system (Bede D1) was used to characterize the crystallinity and the chemical phase of the PEALD and the electroplated thin films, using Cu Karadiation with
2h ranging from 30 to 60. The microstructure of the Cu/Ru/RuNx/
mp-SiO2film stack was studied by transmission electron microscopy
z
(TEM, Philips tacnai 20). Auger electron spectroscopy (AES, thermo VG 350) and secondary ion mass spectroscopy (SIMS, Cameca ION-TOF ) were used to examine the elemental distribution in the film stacks.
The adhesion strength of interfaces in the Cu/Ru/RuNx/mp-SiO2
film stack was measured by the pull-off tensile test. An Al rod (2.7 mm diameter) was glued onto the Cu surface with epoxy resin and cured at 150C for 1 h. The rod was pulled at a progressively
increasing load of 0–100 kg until delamination occurred to the film stack. Electrical measurement was performed with an HP 4145B semiconductor parameter analyzer using the metal-insulator-semi-conductor (MIS) capacitor measurement structure with an Al thin film sputter-deposited on the Si wafer as the backside electrode. The MIS capacitor structure has an area of 0.25p mm2. The sheet
resist-ance of the electroplated Cu thin film was measured with a four-point probe measurement system (Napson RT-80).
Results and Discussion
The PEALD RuNxdiffusion barrier was deposited at 200C on
the oxygen-plasma treatedmp-SiO2layer of 100 nm in thickness,
which had a dense surface oxide layer and showed good dielectric properties as reported previously.15The pore sealing treatment can prevent Ru precursor species from penetrating into the porous dielectric layer during the PEALD RuNx barrier deposition. The
growth rate per cycle of the PEALD RuNxfilm increases with the
Ru(EtCp)2 pulse time and becomes saturated at about 0.038 nm/
cycle when the Ru(EtCp)2pulse time exceeded 5 s. The saturation
of the growth per cycle is characteristic for a self-limited reaction.17 Direct Cu electroplating could not be performed on the 4 nm thick RuNxthin film because RuNxhas a high electrical resistivity. A
pre-vious study showed that a sputter-deposited RuNxfilm had a sheet
resistance ten times higher than a metallic Ru thin film.6To reduce the sheet resistance of the barrier layer, we deposited metallic Ru on the RuNxthin film using the in situ two-step PEALD process. The
preparation of the Ru/RuNxbilayer used 90 PEALD cycles for the
RuNxbottom layer and ten PEALD cycles for the Ru top layer. The
metallic Ru layer was deposited under process conditions like the RuNx layer except that the H2/N2mixture-gas pulse was replaced
with an H2gas pulse. The growth rate of the PEALD Ru thin film
under the process condition is about 0.03 nm/cycle according to TEM analysis.
Figure 1 shows AFM images of the mp-SiO2 substrate before
and after the deposition of the PEALD Ru/RuNxbilayer. The in situ
two-step PEALD process produced a smooth Ru/RuNxbilayer
bar-rier of good conformality. The root-mean-square (rms) surface roughness of the mesoporous silica substrate and the Ru/RuNx
bilayer is 0.5 and 0.8 nm, respectively. The smooth surface is crucial for a seedless Cu electroplating process, which requires a continuous
barrier layer of ultrathin thickness. Ruthenium nitride is thermally unstable at temperatures higher than 275C, at which it decomposes into metallic Ru and nitrogen.11The low decomposition temperature cannot meet the process temperature requirement for the Cu dual-damascene process. To evaluate the thermal stability of the PEALD Ru/RuNx bilayer on the porous silica dielectric, we annealed the
samples at various temperatures in vacuum (107Torr) for 30 min, and examined the microstructure and chemical states of the an-nealed Ru/RuNxfilm stack. Figure2shows SEM images of the
ther-mally annealed Ru/RuNxbilayer. When the sample is annealed at
400C, slight film rupture occurs to the sample surface (see Fig.2b),
indicating thermal decomposition of RuNx has already led to film
degradation. For the sample annealed at 500C, the SEM image
(Fig.2c) shows that particles, with a size ranging from a few nano-meters to one tenth of a micrometer, scatter on the sample surface. The particle growth becomes much more severe for the sample annealed at 600C as shown by Fig.2d. XPS and XRD analyses
dis-cussed later suggest that these particles are discrete Ru islands. Figures 3aand3bshow the Ru(3p3/2) and N(1s) XPS spectra,
respectively, of the Ru/RuNxbilayer as a function of the annealing
temperature. To determine the Ru(3p3/2) electron energy for metallic
Ru and nitrided Ru, we first performed XPS analysis separately for the as-deposited Ru and the as-deposited RuNxthin films. The
mas-ured Ru(3p3/2) energy is 462.0 and 463.0 eV for metallic Ru and
nitrided Ru, respectively. Nonlinear least square curve fitting of the Ru(3p3/2) peak, assuming a Gaussian peak shape, shows that the
Figure 2. SEM images of (a) the as-deposited, (b) the 400C-annealed, (c) the 500C-annealed, and (d) the 600C-annealed Ru/RuNxbilayers deposited on the mesoporous silica film.
composition of metallic Ru in the bilayer sample increases with the annealing temperature. In contrast, the N(1s) signal decreases with increasing the annealing temperature. Compared with the as-depos-ited bilayer, the 400C-annealed bilayer has a larger metallic Ru composition and a smaller N(1s) signal, suggesting thermal decom-position already occurs to the underlying RuNxlayer at 400C. For
the 500C-annealed bilayer, thermal degradation is more severe as indicated by the larger decrease in the Ru(3p3/2) signal ratio of Ru
to RuNx. The XPS result is in agreement with the SEM observation
discussed above, which shows the presence of voids and particles on the annealed bilayer. The N(1s) signal of the bilayer annealed at 600C is barely detected, indicating that RuNxis likely completely
decomposed into metallic Ru phase.
Figure4ashows XRD spectra of the Ru/RuNxbilayer annealed
at different temperatures. The XRD spectra of the as-deposited and the 400C-annealed samples show a very broad peak centered at 2h¼ 33.5. The broad peak is likely due to the presence of RuN
x
nanocrystalline clusters in the amorphous matrix of the Ru/RuNx
bilayer.11 For the bilayer annealed at 500C, the XRD spectrum becomes nearly featureless, suggesting that thermal degradation of the Ru/RuNxbilayer continues to proceed when the annealing
tem-perature is increased. In the XRD spectrum of the bilayer annealed at 600C, a very weak and broad peak, which is just above the noise level, is situated around 43.1. The presence of the weak peak can be perceived by comparison, between the as-deposited and the 400C-annealed samples, of the spectrum profile within the range from 40 to 47. The peak may result from the overlap of the
diffrac-tion peaks due to the (002) (42.15) and (101) (44.00) lattice planes
of the hexagonal Ru crystal structure.17,18 Because the 600
C-annealed bilayer comprises only metallic Ru according to the XPS and XRD analyses, particles observed in the SEM image of Fig.2d
are most likely Ru particles.
The thermal degradation of the RuNxlayer in the bilayer
struc-ture has adverse effects on direct Cu electroplating. Under the elec-troplating condition of the study, Cu could be electrodeposited on the as-deposited Ru/RuNx bilayer, but Cu electrodeposition on a
pure RuNx layer and the Ru/RuNx bilayer annealed at 400 and
500C was unsuccessful. Noted that the 400C-annealed Ru/RuN x
bilayer still keeps a continuous film surface according to the SEM image in Fig.2c. The failure in the Cu electroplating indicates that the 400C-annealed bilayer has a significantly higher electrical
re-sistance than the as-deposited one. It is likely that the partial thermal decomposition of the underlying RuNxlayer at 400C might drive
nitrogen out-diffuse from the underlying RuNxlayer into the upper
metallic Ru layer, thereby increasing the sheet resistance of the bilayer structure. To ensure successful Cu electroplating, we per-formed the thermal treatment only after Cu was electroplated on the as-deposited Ru/RuNxbilayer. Figure 4bshows XRD spectra of a
100 nm thick Cu film electroplated on the Ru/RuNx bilayer as a
function of the annealing temperature. The Cu(111) and Cu(200) peaks are the only two diffraction peaks detected in the spectra of the annealed samples, and the electroplated Cu has a high degree of the (111) texture. The preferential texture is desirable for Cu inter-connects because Cu films with the (111) preferential orientation are four times better in electromigration resistance than those with the (200) orientation.19The as-deposited Cu thin film has a sheet resis-tivity of 21.0 mX/h, and the sheet resisresis-tivity of the electroplated Cu layer decreases to 14.1 mX/h for all the Cu/Ru/RuNxfilm stacks
annealed at temperatures between 400 and 600C. The sheet
resis-tivity of a thin film is an indirect indication of the crystallinity and the chemical purity of the thin film. Although, as discussed later, thermal annealing at 600C leads to chemical and microstructure
breakdowns for the Cu/Ru/RuNxfilm stack, the breakdowns result
in an insignificant change in the sheet resistivity of the film stack, indicating that the barrier failure has little adverse effect on the ma-terial properties of the electroplated Cu thin film.
Figure5shows cross-sectional SEM images of the Cu/Ru/RuNx/
mp-SiO2multilayer stack, as a function of the annealing
tempera-ture. For samples annealed at 500C and below, the interface
between the Ru/RuNxbarrier layer and the Cu layer is free from
par-ticles and delamination (Figs.5aand5c). On the other hand, local-ized interface delamination occurs to the sample annealed at 600C
(Fig.5d). Auger depth profiling and ESCA analyses reveal that the delamination develops at the interface between the Ru/RuNxbarrier
layer and themp-SiO2layer. Figure6shows the Auger depth profile
of the 600C-annealed multilayer sample with the delaminated layer being mecahnically removed. The depth profile clearly indicates that SiO2is the matrix and Ru has diffused into themp-SiO2layer.
Cu is also present on the surface of the separated substrate. The dif-fusion of Ru and Cu into the mesoporous dielectric layer can be more clearly seen in the SIMS depth profile discussed later. Figure 3. (Color online) (a) Ru(3p3/2) and (b) N(1s) XPS spectra of the
Ru/RuNxbilayer sample as a function of the annealing temperature. The two curve-fitted Ru(3p3/2) peaks at 462.0 and 463.0 eV represent metallic Ru and nitrided Ru, respectively.
Figure 4. XRD spectra of (a) the Ru/RuNxbilayer deposited on themp-SiO2 layer, and (b) the Cu/Ru/RuNx/mp-SiO2 multilayer stacks as a function of the annealing temperature.
To furher clarify the integrity of the interfaces in the Cu/Ru/ RuNx/mp-SiO2 mutilayer stack after the thermal anneal, we used
TEM to examine the annealed samples. Cross-sectional TEM images of Figs.7aand7bclearly show that the interfaces of the Ru/ RuNxbarrier with the Cu layer and with the mesoporous SiO2
sub-strate are physically intact after the sample is annealed at temperatures below 500C. The high resolution TEM image in Fig. 7c demon-strates that the Ru/RuNx layer in the 500C-annealed sample is
amorphous without obvious flaw, and has a uniform thickness close to that of the as-prepared multilayer stack. Unlike the bare Ru/RuNx
layer, the Cu capped barrier bilayer is free from crevice and particle at the interface after the thermal anneal. For the 600C-annealed Cu/Ru/RuNx/mp-SiO2 mutilayer stack, the TEM image (Fig. 7d)
shows the presence of a10 nm wide band of darker contrast in the surface region of themp-SiO2layer. The dark band is absent in the
TEM images of the 400- and the 500C-annealed samples, and it must result from Ru and Cu diffusion into the mp-SiO2 layer as
shown by the Auger depth profile in Fig.6. SIMS was also used to study if the Ru/RuNxbilayer was an effective Cu diffusion barrier at
temperatures500C. Figure 8shows SIMS depth profiles of the
thermally annealed Cu/Ru/RuNx/mp-SiO2mutilayer stack. The
400-and the 500C-annealed samples show little difference in the depth profile feature for all the analyzed elements. Noted that the Ru and N signals are entirely present in the signal regime of themp-SiO2
layer, and the Cu signal also extends into the dielectric layer with a
rapid drop at the same depth as the Ru and N signal drops. However, the signal penetration cannot represent the true elemental distribu-tion at the interface of the ultrathin Ru/RuNxlayer with themp-SiO2
layer, but is likely a result of artifacts of the SIMS analysis, such as ion mixing and Gibbsian segregation.20 The concurrent and rapid drop for the Cu, Ru and N signals at the interface suggests that the Ru/RuNxbarrier can effectively retard Cu diffusion into the
dielec-tric layer at temperatures as high as 500C. Combined with the TEM study discussed above, the SIMS analysis suggests that the electrodeposited Cu film behaves like a passivation layer that can effectively impede thermal degradation of the PEALD Ru/RuNx
barrier at 500C and below. For the multilayer stack annealed at 600C, the profile shape of the Ru, N and Cu signals is markedly changed and the nitrogen content is greatly reduced, strongly indi-cating the occurrence of chemical breakdown of the barrier layer. The breakdown results in metal diffusion into themp-SiO2layer as
revealed by the gradual signal penetration of Ru and Cu as well as the relatively high background of the Cu signal in the SiO2region.
We performed leakage current measurements to study the effect of thermal annealing on the electrical properties of the Cu/Ru/RuNx/
mp-silica multilayer stack. The electroplated copper was used as the top electrode and an Al thin film sputter-deposited on the back side of the Si wafer was the bottom electrode. Figure9shows the leak-age current density versus the electric field for the annealed multi-layer samples. Prior to the thermal treatment, the multimulti-layer sample has a leakage current density of 7.2 106A/cm2at the stress field of 1 MV/cm. The leakage current density slightly increases to 8.1 106A/cm2after the sample is annealed at 400C. For the 500 C-annealed sample, a much larger leakage current density (2.1 105 A/cm2) was measured, indicating that the Ru/RuNxbarrier did suffer
greater thermal degradation at 500C than at 400C. The leakage current density of the multilayer samples annealed at 500C is
comparable to that of some previously reported Cu interconnect structures with an Ru based barrier.21–23The 600C-annealed sam-ple has a leakage current density as high as 4.7 104 A/cm2 at Figure 5. Cross-sectional SEM images of (a) the as-prepared, (b) the
400C-annealed, (c) the 500C-annealed, and (d) the 600C-annealed Cu/Ru/RuNx/mp-SiO2multilayer stacks.
Figure 6. (Color online) Auger depth profiles of the 600C-annealed Cu/Ru/ RuNx/mp-SiO2multilayer sample with the delaminated top layer being removed.
Figure 7. Cross-sectional TEM images of (a) the 400C-annealed, (b) the 500C-annealed, (c) the 500C-annealed (in high resolution), and (d) the 600C-annealed Cu/Ru/RuNx/mp-SiO2multilayer stacks. To prepare the TEM specimen for the 600C-annealed sample, the top delaminated layer was removed from the substrate.
1 MV/cm, suggesting that metallic impurities may diffuse into the dielectric layer as a result of the extensive thermal breakdown of the Ru/RuNxbilayer.
Although the Ru/RuNxbilayer capped with the electroplated Cu
can physically and chemically withstand the thermal anneal at
tem-peratures as high as 500C, the material and electrical characteriza-tions discussed above do not provide information about the mechan-ical strength of the multilayer stack. We thus studied the interface adhesion of the Cu/Ru/RuNx/mp-silica mutilayer stack by pull-off
tensile test, and the result is shown in Fig.10. The adhesion strength of the the mutilayer is within the range between 46 and 48 kg/cm2 for samples annealed at 400 and 500C, while the adhesion strength of the sample annealed at 600C drastically drops to2 kg/cm2
. The adhesion strength of the Cu/Ru/RuNx/mp-SiO2sample annealed
at temperatures500C is comparable to that of the Cu/TaN x
/mp-SiO2 structure reported in our previous study.24For the 400- and
500C-annealed samples, the mechanical failure during the pull-off test takes place at the interface between the Ru/RuNxbarrier layer
and the electroplated Cu layer according to ESCA analysis. This is opposite to the 600C-annealed sample, in which the delamination occurs at the interface between the barrier layer and the mp-SiO2
layer. This adhesion test indicates the in situ two-step PEALD pro-cess can deposit an Ru/RuNxbarrier layer strongly adhering to the
mesoporous dielectric layer.
Conclusion
We prepared Ru/RuNx diffusion barriers on mesoporous SiO2
thin films by PEALD for the application of seedless Cu electroplat-ing. The 4 nm thick Ru/RuNxbilayer was deposited on the porous
dielectric substrate at 200C using the in situ two-step PEALD pro-cess. Obvious thermal degradation occurs to the bare Ru/RuNx
bilayer at temperatures400C, and direct Cu electroplating is
suc-cessful only on the as-deposited bilayer barrier. The electroplated Cu layer behaves like a passivation layer that can impede thermal degradation of the Ru/RuNxbarrier. The Cu/Ru/RuNx/mp-SiO2
mul-tilayer structure can withstand the thermal treatment without signifi-cant physical and chemical degradations at temperatures 500C
according to TEM and SIMS analyses. The 400- and the 500 C-annealed samples have a low leakage current density. When the Cu capped multilayer sample is annealed at 600C, thermal decomposi-tion of the RuNxlayer results in metal diffusion into themp-SiO2
layer and delamination at the interface between the Ru/RuNxbilayer
and the dielectric layer. The pull-off tensile test shows that, at tem-peratures 500C, interfaces in the Cu/Ru/RuN
x/mp-SiO2
multi-layer stack have a good adhesion strength. The study has demon-strated that the PEALD Ru/RuNx bilayer structure prepared using
the in situ two-step approach is suitable for the seedless Cu electro-plating process in nanometer scale interconnect technology.
Acknowledgments
This work was supported by the National Science Council of R.O.C. under Contract No. NSC97-2221-E-009-016-MY3. Techni-cal supports from National Nano Device Laboratories is gratefully acknowledged.
National Chiao Tung University assisted in meeting the publication costs of this article.
Figure 8. (Color online) SIMS depth profiles of (a) the 400C-annealed, (b) the 500C-annealed, and (c) the 600C-annealed Cu/Ru/RuNx/mp-SiO2 mul-tilayer stacks. The profile data in the Cu layer are mostly truncated so that the interface region can be clearly examined.
Figure 9. The leakage current density of the Cu/Ru/RuNx/mp-SiO2/Si/Al multilayer samples as a function of the annealing temperature.
Figure 10. (Color online) Bar diagrams of the adhesion strength of the Cu/Ru/RuNx/mp-SiO2 multilayer stack as a function of the annealing temperature.
7. D. Josell, C. Witt, and T. P. Moffat,Electrochem. Solid-State Lett., 9, C41 (2006). 8. D. Josell, J. E. Bonevich, T. P. Moffat, T. Aaltonen, M. Ritala, and M. Leskela,
Electrochem. Solid-State Lett., 9, C48 (2006).
9. Binary Alloy Phase Diagrams, 2nd ed., T. B. Massalski, Editor, p. 1467, American Society of Metals, Materials Park (1990).
10. R. Chan, T. N. Arunagiri, Y. Zhang, O. Chyan, R. M. Wallace, M. J. Kim, and T. Q. Hurdc,Electrochem. Solid-State Lett., 7, G154 (2004).
11. M. Damayanti, T. Sritharan, S. G. Mhaisalkar, and Z. H. Gan,Appl. Phys. Lett., 88, 044101 (2006).
12. M. Damayanti, T. Sritharan, S. G. Mhaisalkar, H. J. Engelmann, E. Zschech, A. V. Vairagar, and L. Chan,Electrochem. Solid-State Lett., 10, P15 (2007).
0663.
19. C. Ryu, K. W. Kwon, A. L. S. Loke, H. Lee, T. Nogami, V. M. Dubin, R. A. Kavari, G. W. Ray, S. S. Wong,IEEE Trans. Electron Devices, 46, 1113 (1999). 20. R. G. Wilson, F. A. Stevie, and C. W. Magee,Secondary Ion Mass Spectrometry,
John Wiley & Sons, New York (1989).
21. J. B. Yeh, D. C. Perng, and K. C. Hsu,J. Electrochem. Soc., 157, H810 (2010). 22. D. C. Perng, K. C. Hsu, S. W. Tsai, and J. B. Yeh,Microelectron. Eng., 87, 365
(2010).
23. D. C. Perng, J. B. Yeh, and K. C. Hsu,Appl. Surf. Sci., 256, 688 (2009). 24. C. C. Chang, F. M. Pan, and C. W. Chen, J. Electrochem. Soc., 157, G62
