Method to improve chemical-mechanical-planarization polishing rate of low-k methyl-silsesquiazane for ultralarge scale integrated interconnect application

(1)

Method to improve chemical-mechanical-planarization polishing rate of low-k

methyl-silsesquiazane for ultralarge scale integrated interconnect application

T. C. Chang, T. M. Tsai, P. T. Liu, S. T. Yan, Y. C. Chang, H. Aoki, S. M. Sze, and T. Y. Tseng

Citation: Journal of Vacuum Science & Technology B 22, 1196 (2004); doi: 10.1116/1.1755218 View online: http://dx.doi.org/10.1116/1.1755218

View Table of Contents: http://scitation.aip.org/content/avs/journal/jvstb/22/3?ver=pdfcov

Published by the AVS: Science & Technology of Materials, Interfaces, and Processing

Articles you may be interested in

Yield improvement of 0.13 m Cu/low- k dual-damascene interconnection by organic cleaning process

J. Vac. Sci. Technol. B 25, 1819 (2007); 10.1116/1.2794049

Plasma enhanced chemical vapor deposited SiCOH dielectrics: from low-k to extreme low-k interconnect materials

J. Appl. Phys. 93, 1785 (2003); 10.1063/1.1534628

Characterization of low permittivity (low-k) polymeric dielectric films for low temperature device integration

J. Vac. Sci. Technol. A 20, 1149 (2002); 10.1116/1.1463083

Highly reliable chemical–mechanical polishing process for organic low-k methylsilsesquioxane

J. Vac. Sci. Technol. B 19, 1212 (2001); 10.1116/1.1385684

Effect of oxygen plasma exposure of porous spin-on-glass films

(2)

of low-k

methyl-silsesquiazane for ultralarge scale integrated

interconnect application

*

T. C. Changa)

Department of Physics and Institute of Electro-Optical Engineering, National Sun Yat-Sen University, Kaohsiung, Taiwan and Center for Nanoscience and Nanotechnology, National Sun Yat-Sen University, Taiwan

T. M. Tsai

Institute of Electronics, National Chiao Tung University, Hsin-Chu, Taiwan

P. T. Liu

National Nano Device Laboratory, 1001-1 Ta-Hsueh Road, Hsin-Chu 300, Taiwan

S. T. Yan

Y. C. Chang and H. Aoki

Life Science and Electronic Chemicals Division, Clariant Corp., Bunkyo Green Court 2-28-8 Honkomagome, Bunkyo-ku, Tokyo 113-8662, Japan

S. M. Sze

Institute of Electronics, National Chiao Tung University, Hsin-Chu, Taiwan and National Nano Device Laboratory, 1001-1 Ta-Hsueh Road, Hsin-Chu 300, Taiwan

T. Y. Tseng

共Received 29 September 2003; accepted 5 April 2004; published 28 May 2004兲

In this work, characteristics of low-k methyl-silsesquiazane 共MSZ兲 for the chemical-mechanical-planarization 共CMP兲 process using oxygen plasma pretreatment were investigated in detail. The low-dielectric-constant 共low-k兲 MSZ was prepared by a spin-on deposition process. The resultant wafers were followed by an oxygen (O2) plasma treatment. After

oxygen plasma treatment, the CMP process was implemented. Electrical and material analyses were utilized to explore the characteristics of post-CMP MSZ. Experimental results showed that the polish rate of MSZ film with O2 plasma pretreatment was increased as much as two times in

magnitude, as compared to that of the MSZ without O2 plasma pretreatment. In addition, the

post-CMP MSZ exhibited superior electrical properties. These results clearly indicated that the modification surfaces that resulted from O2-plasma treatment facilitated CMP MSZ. After CMP

关DOI: 10.1116/1.1755218兴

I. INTRODUCTION

As integrated circuit dimensions continue to shrink, highly packed multilevel interconnections with low-resistance metal and low-dielectric constant materials have attracted much attention as a method for increased ultralarge scale integrated 共ULSI兲 circuits operating speed.1,2 Because continued miniaturization of device dimensions and the re-lated need to interconnect an increasing number of devices on a chip is a trend in ULSI technology, this has led to building multilevel interconnections on planarized levels.3,4 Obviously, surface planarization is a key technology during the manufacture of multilevel interconnects.5–7 Because the chemical-mechanical-planarization 共CMP兲 process satisfies the requirement of global topography planarization during integrated circuit fabrication, this method has been

intro-duced to form interconnections between devices and between devices and outside.8,9 In the development of CMP low-k dielectrics, Forester et al.10found that the polish rate of alkyl siloxane-based spin-on glass 共SOG兲 was lower than that of plasma enhanced chemical vapor deposition oxide or thermal oxide using only conventional silica-based slurry. When us-ing conventional oxide slurries, the polish rate of alkyl siloxane-based SOG is dependent on the organic content. A higher Si–R/Si–O ratio in the SOG films induces a lower hydrolysis reaction rate, leading to a lower polish rate. Sev-eral reports11,12 indicate that the use of alkaline cerium oxide-based slurry and the introduction of additives could greatly improve the removal rate for organic spin-on materi-als. However, before adopting consumables into production, much experimental work should be completed to qualify these consumables.

In this study, we propose oxygen plasma pretreatment on low-k methyl-silsesquiazane共MSZ兲 film to improve the pol-*No proof corrections received from author prior to publication.

(3)

ish rate of CMP MSZ film. The low-k MSZ film is provided by Clariant Corp. in Japan and is one of methyl-silsesquioxane-like organic polymers derived from MSZ pre-cursor solution. Moreover, the investigation of post-CMP characteristics such as electrical performance and desorption of constitution water also were emphasized.

II. EXPERIMENT

The substrates used in this study were 150 mm p-type

共11–25 ⍀ cm兲 single-crystal silicon wafers with 共100兲

orien-tation. Before film deposition, Si wafers were boiled in H2SO4⫹H2O2 solution and heated to 120 °C for 20 min to

remove particles on the surfaces. Then, these wafers were spin coated with a MSZ solution at a spin speed of 2000 rpm for 30 s on a model 100CB spin coater. This was followed by a series of sequential thermal baking steps on the hot plate at 150 and 280 °C for 3 min. The wafers coated with MSZ precursor solution then received a hydration treatment. The wafers were left in a clean room for 48 h and the precursor structure of MSZ films will be transformed to methyl-silsesquiazane through hydrolysis and condensation process, as follows:

Hydrolysis reactions:

⬅Si–NH–Si⬅共s兲⫹H2O共g兲→⬅Si–NH2共s兲⫹HO–Si⬅共s兲 ⬅Si–NH2共s兲⫹H2O共g兲→⬅Si–OH共s兲⫹NH3共g兲.

Condensation reactions:

⬅Si–OH共s兲⫹HO–Si⬅共s兲→⬅Si–O–Si⬅共s兲⫹H2O共g兲.

Afterward, the resultant wafers were thermally cured in a quartz furnace at 400 °C for 30 min under N2 ambient. The

final MSZ films 共called as-cured MSZ, marked as sample STD兲 were formed to a thickness of 400 nm. The final struc-ture is shown in Fig. 1.

After film formation, the as-cured MSZ was treated with O2 plasma for 60 s prior to the CMP process 共marked as

sample O兲. The O2plasma was operated at a pressure of 650

mTorr and with an oxygen gas flow rate of 900 sccm. A radio frequency power of 110 W was applied to the upper elec-trode. Next, the wafers were placed on the grounded bottom electrode, which can be rotated for improving uniformity. The wafers had a substrate temperature up to 250 °C. Then the CMP process was applied to the plasma-treated MSZ films for 2 min. These post-CMP MSZ films were marked as ‘‘sample C.’’ The CMP experiment was carried out on an IPEC/Westech 372M CMP processor with a Rodel IC 1400 pad on the primary polishing platen and Rodel Politex

Regu-lar embossed pad on the final buffering platen. A Rodel R200-T3 carrier film was used to provide a buffer between the carrier and wafer. A single 6 in. wafer was mounted on a template assembly. During the polishing experiment, the slurry was commercial CABOT™ SS-25 diluted by de-ionized water with the ratio of 1:1, which is typically used to polish SiO2. The resultant solution pH value is in the range

of 10–11. The polishing parameters, such as down force, backpressure, platen and carrier rotation speeds, and slurry flow rate, were set to be 3, 2, 50, 60 rev/min, and 150 ml/ min, respectively. By means of light interference effects in films, the thickness of all MSZ films in this experiment was measured using an N&K 1200 analyzer. The structure prop-erties of the MSZ films were studied using Fourier-transform infrared spectroscopy共FTIR兲. The infrared spectrometry was performed from 4000 to 400 cm⫺1 using a Bio-Red QS300 FTIR spectrometer calibrated to an unpatterned wafer and their data were collected in the absorbance mode for study-ing the chemical structure of films. The surface morpholo-gies of the polished films were investigated by atomic-force microscopy共AFM兲. Thermal desorption spectroscopy 共TDS兲 was carried out to monitor the desorbed elements from post-CMP MSZ films during the high temperature process. In the duration of TDS analysis, samples were heated from room temperature to 600 °C at a rate of 20 °C/min in a vacuum chamber. In addition, the outgassing species were collected through the mass spectrometer. In this work, M/e 共mass-to charge ratio兲⫽18 peak attributed to H₂O was monitored. Electrical characteristics of post-CMP MSZ films were per-formed on the metal–insulator–semiconductor capacitor with metallic–aluminum deposition as the top electrode and backside electrode. Leakage current–voltage (I – V) and capacitance–voltage characteristics were also used to ana-lyze the leakage current behaviors and measure the dielectric constants of post-CMP MSZ films, respectively. In addition, I – V measurements were also conducted for these specimens at different stable temperatures during the temperature rising and cooling procedures to evaluate the practicability of

im-FIG. 1. Formation mechanism diagram of MSZ films. FIG. 2. FTIR spectra of MSZ during formation procedure.

1197 Changet al.: Method to improve CMP polishing rate 1197

(4)

proving the polishing rate of MSZ films with O2 plasma

pretreatment for the CMP process.

III. RESULTS AND DISCUSSION

For the applications for multilevel interconnects, low-k dielectrics must be carefully characterized for their material

and electrical properties. Figure 2 represents the FTIR spec-tra of MSZ films during formation processes. After 10 h of hydration reaction proceeded in a clean room, it was ob-served that the Si–N peaks almost disappeared, while the Si–O network-like bonds and Si–OH bonds grew gradually. This was due to the hydrolysis process. This result implied that the hydrolysis and condensation process occurred simul-FIG. 3. AFM micrographs of 60 s O2plasma-treated MSZ films共a兲 without CMP process 共b兲 with CMP process.

(5)

taneously during the period. In the periods ranging from 10 to 48 h, the condensation process resulted in a network-like structure through the bulk MSZ film. Next, the water content was eliminated and the standard MSZ film was obtained after thermal curing in a furnace. AFM images of O2

plasma-treated MSZ films before and after CMP process are shown in Figs. 3共a兲 and 3共b兲, respectively. The roughness (Ra) of

MSZ films with O2 plasma treatment was 0.341 nm, while

the Ravalue of O2plasma-treated MSZ after being subjected

to a CMP process was reduced to 0.227 nm. In addition, the CMP polish rate of MSZ with O2 plasma pretreatment was

increased as much as twice in magnitudes compared to MSZ without O₂ plasma treatment共the average polishing rate of MSZ is increasing from 16 to 35 nm/min兲. The results indi-cated that the removal rate of MSZ can be improved by O₂ plasma treatment, even with SS-25 slurry used typically. Fig-ure 4 shows FTIR spectra of O₂ plasma-treated MSZ films before and after the CMP process. Prior to the CMP process, both intensities of Si–OH and H2O groups共at 993 and 3400

cm⫺1兲 increased, whereas the intensities of C–H 共2974 cm⫺1兲 and Si–CH3共at 781 and 1273 cm⫺1兲 groups decreased

when MSZ film underwent O2 plasma treatment. Moreover,

the peak of the Si–O bond at 1070 cm⫺1was slightly formed in O2plasma-treated MSZ. After the CMP process, however,

all intensities of functional groups in MSZ films maintained a high level again. These observed phenomena are clearly interpreted as follows.

The decomposition of functional groups in MSZ films, due to O2 plasma pretreatment, would lead to forming

Si–OH bonds, which easily induced moisture uptake and modified the MSZ surfaces from hydrophobic into hydro-philic ones. It was believed that oxygen radicals generated from O2 plasma could react with a large amount of Si–CH3

groups on MSZ films, which caused the decreasing intensi-ties of Si–C and C–H groups. Hence, the oxidation reaction would convert Si–CH3 groups into Si–OH groups via the

following processes:

⬅Si–CH3共s兲⫹2O2共g兲→⬅Si–OH共s兲⫹CO2共g兲⫹H2O共g兲.

Because the Si–OH group was hydrophilic, it was easy to induce moisture uptake. As a result, after the MSZ film un-derwent O2 plasma treatment, the intensities of the Si–OH

and H2O signals would increase. In addition, a partial

amount of Si–OH groups might react with each other via a dehydration reaction during O2 plasma treatment

⬅Si–OH共s兲⫹HO–Si⬅共s兲→⬅Si–O–Si⬅共s兲⫹H2O共g兲.

The intensity of the absorption band 共at 1070 cm⫺1兲, which was characteristic for the Si–O–Si vibration in silica, was thereby increased. The hydrophilic surfaces consisting of oxide facilitated CMP of MSZ films and a rapid CMP polish rate was obtained, just with only the conventionally used CMP oxide slurry CABOT™ SS-25. The hydrophilic Si–OH bonds would be removed after the CMP process. Thus, Si–OH bonds disappeared in FTIR spectra. After the removal of the oxide layer on the surface of O₂ plasma-treated MSZ, organic function groups such as Si–C and C–H bonds maintained a high level of peak intensity.

The temperature dependence of moisture desorption is shown in Fig. 5. After being subjected to the O2 plasma

treatment, the moisture content of MSZ film was increased, while it decreased after the CMP process. This was consis-tent with our inference that oxide layers on MSZ surfaces could induce moisture and be easily polished away just with oxide slurry used typically. Once the surface oxide layers were absent, the surfaces of MSZ would return to being hydrophobic-like, resulting in less moisture content. Further-more, the electrical characteristics were investigated to evaluate the impacts of the CMP process on MSZ films. Fig-ures 6共a兲 and 6共b兲 show the leakage current and dielectric constant of O2 plasma-treated MSZ before and after the

CMP process. The electrical properties of MSZ with O2

plasma treatment were degraded. After the CMP process, however, the leakage current and dielectric constant of MSZ

FIG. 4. FTIR spectra of O2plasma-treated MSZ films before and after CMP

process. FIG. 5. Moisture-desorption spectra of O2plasma-treated MSZ films before

and after CMP process.

(6)

were recovered significantly. In order to explore the leakage behaviors of MSZ films with O2 plasma pretreatment before

and after CMP process, we tried to conduct I – V measure-ment at different temperatures during the temperature rising and cooling procedure. Figure 7 shows the leakage current density of sample O and sample STD measured at 25 °C

共curves I, II, and IV兲 and 150 °C 共curve III兲, respectively.

Owing to O2plasma treatment, the leakage current of sample

O is larger than that of sample STD. The O2 plasma could

modify the surface of MSZ film, leading to formation of defects and inducing moisture uptake. Both the hydrophilic defects and the defect-induced moisture often result in an increase of leakage current. In order to recognize the effect of moisture uptake on the O2 plasma-treated MSZ film,

leakage-current measurement is performed before and after the 150 °C bake. In comparison with sample STD, after O2

plasma ashing, the leakage current density increases about 1–2 orders of magnitude due to defects-induced moisture uptake, as shown for sample O 共curve II兲. After the 150 °C baking process共curve III兲, an amount of water molecules are desorbed from the sample O so that the leakage current of the sample O 共measured at the 150 °C baking temperature兲 decreases about one order of magnitude. Nevertheless, when the measured temperature of sample O is cooling from 150 to 25 °C, the leakage current of sample O increases signifi-cantly again共curve IV兲, which results from the moisture re-uptake during the temperature-cooling processing. In addi-tion, the moisture reuptake may be due to the remainder of hydrophilic defects caused by O₂ plasma damage in the sur-face of the MSZ film. Figure 8 shows the leakage current density of sample C and sample STD measured at 25 °C

共curves I, II, and IV兲 and 150 °C 共curve III兲, respectively.

Because the modification surface layer of O2 plasma-treated

MSZ was polished away by the CMP process, the leakage current of sample C 共curves II and IV兲 was close to that of as-cured MSZ film共curve I兲 at 25 °C. Moreover, the leakage current of sample C 共curve III兲 increased by one order of magnitude compared to that of the as-cured MSZ film共curve I兲 at 150 °C. This indicates that the leakage current mecha-nism could be dominated by a thermionic field emission pro-cedure at the high temperature I – V measured condition. Our results reveal that the hydrophilic surface layer made due to the O2plasma treatment would result in the increase of

leak-age current of MSZ film. However, the leakleak-age current of O2

plasma-treated MSZ could be recovered after polishing the most of hydrophilic layer by CMP process. These electrical results were consistent with aforementioned FTIR and TDS analysis data.

FIG. 6. Dielectric properties of O2 plasma-treated MSZ before and after

CMP process 共a兲 leakage-current density of MSZ films as a function of

electrical field.共b兲 Variation in dielectric constant of O2plasma-treated MSZ films.

FIG. 7. Leakage-current density of sample O before and after the 150 °C bake关curve I, sample STD measured at 25 °C; curve II, sample O measured at 25 °C 共before 150 °C bake兲; curve III, sample O measured at 150 °C; curve IV, sample O measured at 25 °C共after 150 °C bake兲兴.

(7)

IV. CONCLUSIONS

In this study, we have proposed an effective method to improve the CMP polish rate of organic MSZ films. The oxygen plasma was used to make the surface of the MSZ film more hydrophilic and to facilitate the CMP of MSZ films. As a result, the polish rate of O2 plasma-treated MSZ

was as large as twice the magnitude of MSZ without O2

plasma pretreatment. In addition, the electrical properties of O2plasma-treated MSZ films could be recovered almost to a

similar state as the as-cured MSZ after CMP process. This indicated that most of damage surfaces of O2plasma-treated

MSZ films could be removed during the CMP process. Also, the dielectric properties of MSZ could maintain the low-k quality. The results were consistent with those of material analyses. Therefore, O2 plasma pretreatment will be a

prom-ising method to increase the polish rate of organic MSZ films.

ACKNOWLEDGMENTS

This work was performed at the National Nano Device Laboratory and was supported by Clariant Corporation, Ja-pan, and the National Science Council of the Republic of China under Contract. Nos. NSC92-2112-M-110-020 and NSC92-2215-E-009-019.

1

P. K. Vasudev, Solid-State Electron. 39, 481共1996兲.

2_{Y. Hayashi, J. Kawahara, K. Shiba, M. Tagami, S. Saito, T. Onodera, K.}

Kinoshita, and M. Hiroi, NEC Res. Dev. 42, 51共2001兲.

3_{P. Wrschka, J. Hernandez, G. S. Oehrlein, and J. King, J. Electrochem.}

Soc. 147, 706共2000兲.

4_{A. K. Sikder and A. Kumar, J. Electron. Mater. 31, 1016}_共2002兲. 5_{B. Trednick, J. Lee, K. Holland, and T. Bibby, Proceedings of the First}

International Chemical Mechanical Polish (CMP) for VLSI/ULSI Multi-level Interconnection Conference, 1996, p. 107.

6

The National Technology Roadmap for Semiconductors共Semiconductor Industry Association, San Jose, CA, 2001兲.

7_{P. T. Liu et al., Jpn. J. Appl. Phys., Part 1 40, 3143}_共2001兲.

8_{S. Lakshminarayanan, P. Wright, and J. Pallinti, IEEE International}

In-terconnect Technology Conference, 2002, p. 99.

9_{J. Chen, S. Parikh, T. Vo, S. Rengarajan, T. Mandrekar, P. Ding, L. Chen,}

and R. Mosely, IEEE International Interconnect Technology Conference, 2002, p. 185.

10

L. Forester, D. K. Choi, and R. Hosseini, VMIC Proceedings of the Inter-national Conference, 1995, p. 482.

11_{Y. Homma, T. Furusawa, K. Kusuka, M. Nagasawa, Y. Nakamura, M.}

Saitou, H. Moroshima, and H. Sato, IEEE VLSI Multilevel Interconnec-tion, Proceedings of the International Conference, 1995, p. 457.

12

Y. L. Wang, C. Liu, S. T. Chang, M. S. Tsai, M. S. Feng, and W. T. Tseng, Thin Solid Films 308–309, 550共1997兲.

FIG. 8. Leakage-current density of sample C before and after the 150 °C bake关curve I, sample STD measured at 25 °C; curve II, sample C measured at 25 °C 共before 150 °C bake兲; curve III, sample C measured at 150 °C; curve IV, sample C measured at 25 °C共after 150 °C bake兲兴.