Author(s) In-Kwon Kim, Y. Nagendra Prasad, Tae-Young Kwon, Hyuk-Min Kim, Ahmed A. Busnaina, and Jin-Goo Park

Transcription

1 Author(s) InKwon Kim, Y. agendra Prasad, TaeYoung Kwon, HyukMin Kim, Ahmed A. Busnaina, and JinGoo Park This article is available at IRis:

2 H152 Journal ofthe Electrochemical Society, 158 (1) H152 H156 (11) /11/158(1)/HI52/5/$28. The Electrochemical Society Citric Acid and al4 Based Alkaline Cleaning Solution for Particle Removal during PostRu CMP Cleaning InK won Kim, a Y. agendra Prasad, a TaeYoung Kwon,a HyukMin Kim,b Ahmed A. Busnaina,c,* and JinGoo Parka,h,*,z a Department of Materials Engineering and b Bio anotechnology, Hanyang University, Ansan , Korea Center for Microcontamination Control, ortheastern University, Boston, Massachusetts 2115, USA Post Ru CMP cleaning solution for particle (alumina) removal was successfully developed by controlling surface zeta potentials and choosing the proper Ru etchant. The zeta potentials of both alumina particles and Ru surface were measured as a function of ph. It was found that 1 ppm of citric is good enough to modify the zeta potentials of both alumina particles and Ru surface at ph l. At this condition, both Ru and alumina particles will have negative zeta potentials which will cause repulsion between them. This effect was confirmed by measuring the adhesion force between Ru and alumina particle by using force distance method at different ph values. The lowest adhesion force between an alumina particle and the Ru surface was observed in citric acid solu tion at ph 1. However, this effect can be magnified by adding an etchant to the cleaning solution which can lift off the particles. ai4 was chosen as Ru etchant and added to citric acid solution at ph 1. The highest PRE of 97% was achieved in citric acid solution at and slightly below.1 M ofai4 concentration at ph The Electrochemical Society. [DOI: lo.ll49/l ] All rights reserved. Manuscript submitted April 4, 11; revised manuscript received June 28, 11. Published August 5, 11. Ruthenium (Ru), included in the platinum group of the periodic table, is one of the rarest metals on earth. It has a number of valua ble properties including a very high melting and boiling point, low electrical resistivity, and a very high work function. Also it is not tarnished at room temperature. Due to these properties, demand for Ru has been increasing in the electronic, chemical and semiconduc tor industry. It has been suggested as a potential bottom electrode material for capacitors with Ta 5 as a dielectric since high dielec tric constants, low leakage currents and compatibility have been reported for Ta 5 grown on ruthenium electrodes.1 2 It is also being studied for FRAM capacitors with PZT as the ferroelectric? Ru fihn has also been investigated as a Cu diffusion barrier metal and a Cu seed layer in Cu interconnection due to its good electrical conduc tivity, inuniscibility with Cu, good adhesion property to Cu layer, and low solid solubility in copper up to 9 C. 4 7 Chemical mechanical planarization (CMP) is one of the fabrica tion processes for electrode formation and barrier layer removal.8 Af ter the CMP process, the polished surface may have many defects such as slurry particles, chemical residues, and physical damages. The abrasive particles can be easily contaminated on the top surface dur ing the CMP process? This can induce adverse effects on subsequent patterning and fihn deposition processes with a lower breakdown voltage of devices.1 Since Ru has only been recently introduced in semiconducor device fabrication, there are few studies on post Ru CMP cleaning. In this study, a post Ru CMP cleaning solution was formulated by using sodium periodate as an etchant and citric acid to modify the zeta potential of particles and Ru surfaces. Experimental Materials and Procedure Ru fihn (15 nm thickness) was deposited on tetraethylorthosili cate (TEOS) films by the atomic layer deposition method. Ru wafers were cut into 2. x 2. em pieces for the surface analysis and used for estimating PRE. Citric acid (Sigma Aldrich, 99%, USA) was added to modify the zeta potentials of both slurry particles and Ru surfaces during cleaning. Sodium periodate (ai4, Sigma Aldrich, 99.8%, USA) was chosen as an Ru etchant111 2 in preparing the cleaning solution. Slurry was prepared using y Alumina particles contained in DI water (Degussa, 99.99%, 13 nm) and used to con taminate Ru wafers. To evaluate the PRE of cleaning solutions, the Ru wafer was dipped in alumina slurry for 1 min and dried by blow ing 2. Then, the contaminated Ru wafer was dipped in the cleaning solutions for 1 min and then rinsed in DI water, and followed by 2 * Electrochemical Society Active Member. ' jgpark@hanyang.ac.kr blow. HCl (38 wt %) OH (28 wt %) aoh and KOH were used as ph adjustors. A laser zeta potential analyzer (LEZA 6, Otsuka Electronics Co., Japan) was used to obtain the zeta potentials of alumina par tides and the Ru surface. Especially, the monitor particle used is HPC (Hydroxyl Propyl cellulose) coated PSL (Poly Styrene Latex) particle which has zero charge for measurement of Ru surface zeta potential. This particle will be affected by charged surface of Ru and shows the electro osmosis behavior in the solution with applied outside potential. The etch rate was calculated by measuring the sheet resistance of Ru films by a 4 point probe (CMT SR1, Changmin Tech. Co., Ltd., Korea). A contact angle analyzer (Phoe nix 4, SEO, Korea) was used to measure the contact angle of the Ru surface. The adhesion force between an alumina particle and Ru wafer surface was measured by an atomic force microscope (AFM, XE 1, Park Systems, Korea) A 4 J..Lm diameter sized spheri cal alumina (Micron Co., Japan) was attached on the AFM tipless cantilever for the measurement of the adhesion force. The PRE on the Ru wafer was estimated in the dark field using the optical micro scope (LV1D, ikon Co., Japan). Results and Discussion Ru metal has a negative surface charge The zeta potential depends on the ph of solutions and plays a vital role in the adhesion and removal of particles from a surface. The zeta potentials of Ru surface and alumina abrasive particles were measured as a function of ph and the result is shown in Fig. 1. The zeta potential of Ru sur face decreased from+4 mv to near zero when ph was increased from 2 to 4, then reached a slightly negative value and remained almost constant at higher phs. The positive zeta potential of Ru at ph 2 is attributed to the oxidation of Ru to Ru2. Ru wafer con tained negative zeta potentials at phs above 4 due to the non forma tion of Ru2. The zeta potential of Ru in neutral region showed sim ilar values as reported in the literature In contrast, alumina particles show positive zeta potentials for a wide range of ph values except for a strong alkaline region.1 2 Ru wafer and alumina particles can have the same sign of the zeta potentials either at strong acidic or alkaline ph region. However, if the post Ru cleaning solution is formulated at these ph regions, severe corrosion of Ru can be generated on the surface by the disso lution of Ru oxide, which is formed due to the chemicals of the cleaning solution.1718 In addition to this, it causes problems during waste treatment. Therefore, the ph range of cleaning solution for Ru

3 Journal ofthe Electrochemical Society, 158 (1) H152 H156 (11) H E ns ;::; 1 c: (I) 1 c.. ns (I) ph Ionic Strength : KCI1 3 M ph Control : HCI and aoh E ns 1 ;::; c: (I) c.. 1 (I) ns t Diwater ya1p3 5 ppm Ionic Strength : KCI 1 o 3M Citric Acid 1 ppm ph ph Control : HCI and aoh 1 12 Figure 1. (Color online) The variation of zeta potential of Ru wafer and alu mina particles as a function of ph. Figure 3. (Color online) The variation of zeta potential of alumina particle in solution with and without citric acid as a function of ph. surface is limited to neutral or alkaline except the acid and strong alkaline region to prevent severe corrosion. Zeta potential of alumina particles can be modified to a negative charge by adding an organic acid. When organic acids, such as citric acid and oxalic acid, are added to the solution containing particles, the surface charge of metal oxide particles is modified due to the specific adsorption of carboxylate ions on their surface.19 2 How ever, citric acid adsorbs on alumina surface better than oxalic acid? Figure 2 shows the zeta potential of alumina particles as a function of citric acid concentration at ph 6. The zeta potential of alumina rapidly decreased as a function of citric acid concentration up to 1 ppm and then became saturated. Figure 3 shows the change of zeta potential of alumina particles in DI water with and without addition of citric acid as a function of ph. When citric acid was added to the solution containing alumina particles, the zeta potential of alumina was changed to negative value at acidic and neutral ph conditions as citrate ions were easily adsorbed on alu mina particles? IEP of alumina particles shifted from ph 9 to 3. From Figs. 1 and 3 it can be expected that the repulsive force would occur between Ru wafer and alumina particles in neutral to alkaline solution with citric acid due to their similar negative polarity. Initially, it was examined to find the nature of surfaces of Ru and alumina particles present at neutral condition. Figure 4 shows the Eh ph diagram of the Ru H system17 indicating the Ehs of clean ing solutions studied at different ph conditions. In solutions with citric acid at ph 6 and 8, Ru(OHh could be formed on Ru surface compare with ph 4 and 1. It was reported that Ru(OHh can be eas ily oxidized to Ru22H by the reaction with 2 and the positive zeta potential of Ru22H In Fig. 5, the zeta potentials of Ru and Ru2 were measured in solution of ph 6. Ru2 was obtained on Ru wafer by dipping Ru film in ai4 solution at ph 6. This finding was reported already in our previous work11 where thicker Ru oxide was formed on Ru surface in ai4 solution at ph 6. From these results it can be understood that at ph 6, the Ru surface has a negative zeta potential of 5 m V, whereas Ru oxide has a pos itive value of 3 m V. Figure 6 shows the zeta potential of Ru surface before and after the treatment in 1 ppm citric acid solutions as a function of ph. In citric acid solutions at ph 6 and 8, the zeta 4 yaj3 5 ppm ph 6 ph Control : KOH E ca r:::: a a. ca a Concentration of atric Acid (ppm) Figure 2. (Color online) The variation of zeta potential of alumina particles as a function of citric acid concentration ,2 2,2 E(V) RuO, gas? 1.6 1,6.4 Ru4 H1Au5 HRu5,2,2 o _,2,2,4 ',6,6.8 o.e 1 1 1,2 1,2 1,4 1,4 1,6 1,6 1,8 1, pH16 Figure 4. (Color online) The Eh ph diagram of the Ru H system with the Eh and ph of the solution with citric acid as a function of ph (from Ref. 17). 1,8 1,4 1,2,8.6,4.4

4 .. H154 Journal ofthe Electrochemical Society, 158 (1) H152 H156 (11) 8 6 at ph 6 E 4 m 2 c.. 2 m Ru Ru Oxide (Ru2) Figure 5. (Color online) The zeta potential values of Ru and Ru2 at ph 6. potential of Ru wafer was increased to a slightly positive value. It can be seen that the Ru oxide is formed on Ru surface in citric acid solution at both ph 6 and 8, as shown in Eh ph diagram of Fig. 4. Consequently, at ph 6 and 8, lower cleaning efficiency can be expected due to the attractive force between the negative alumina and the positive Ru oxide. Hence, it was considered that the clean ing solution has to be prepared at ph conditions above 8 because there is a repulsive force between the negative alumina and Ru surface. To further understand the role of surface charges, the adhesion force between the abrasive particle and the Ru wafer was measured by the atomic force microscope (AFM) force distance method in solution with citric acid as a function of ph. Figure 7 shows the adhesion force between the Ru surface and alumina particle in the citric acid solutions as a function of ph. The lowest adhesion force was measured at ph 1 and a relatively high adhesion force was measured at ph 6 and 8. These results confirm that the attractive force acted between the positive Ru oxide and negative alumina par ticle due to the formation of hydrous Ru oxide on the Ru surface in solution with citric acid at ph 6 and 8. This process suggests that post Ru cleaning solution with citric acid should be formulated at alkaline ph conditions 1. The dark field optical micrographs of Ru wafer before and after alumina particle contamination are shown in Figs. Sa and b, respec 3 c::::j wafer Ahnina Partide Mlesion In Otric Acid 1 ppm 25 z CJ u. 15 tj) 1.. ' 5 <( n oh4 oh6 oh8 oh1 Figure 7. (Color online) The adhesion force between Ru surface and alu mina particle in citric acid solution as a function of ph. tively. Contaminated Ru wafers were cleaned with DI water and sol utions with 1 ppm citric acid as a function of solution ph. DI water cleaning resulted in only 22% PRE, while citric acid added solution provided a higher PRE of 9% at ph 1, as shown in Fig. 9. Lower PRE at ph 6 and 8 than at ph 1 might be attributed to the 8.. ph Control: H4H 6 E 4 treated Ru Wafer in Citric Acid 1 ppm m 2 c.. 2 m bare Ru Wafer 8LLL ph Figure 6. (Color online) The zeta potential of Ru surface before and after the treatment with citric acid of 1 ppm as a function of ph. (a) (b) Figure 8. (Color online) The dark field optical micrographs of Ru wafer (a) before (reference) and (b) after alumina particle contamination.

5 Journal ofthe Electrochemical Society, 158 (1) H152 H156 (11) H155. o 1 c:::::j Citric Acid 1 (J 9 t: 8 (J 7!E w 6 cu 5 4 E 3 :::.!! (J 1 :e cu a.. ph4 ph6 phs ph 1 Figure 9. (Color online) The particle removal efficiency after cleaning in DI water and solution with citric acid as a function of ph. different zeta potential polarity of the surface and particles which caused attractive interaction between them. However, the modified zeta potential could not offer enough re moval force to clean the contaminated particles effectively due to the small magnitude of the zeta potential values. Hence, it is neces "2 1 E E 8 t: 6 cu ::: J: 4 (J w (J 2 :;: cu en KI J wateir at ph 1lo aio 4 Concentration (M) 8 6 E 4 ns 2._ c:: C1)._ a. ns._ 4 C1) 6 Citric Acid 1 ppm at ph 1 RuWafer 2.2 M 8 OM.1 M aio 4 Concentration Figure 11. (Color online) The zeta potential of Ru wafer at ph of cleaning solution after the treatment in citric acid of 1 ppm at ph 1 as a function ofai4 concentration. sary to improve the cleaning efficiency with the addition of an etch ant chemical. The cotaminated particles can be lifted off from the Ru surface by slightly etching the surface. 2 1 In this study, ai4 was examined as an etchant for the preparation of post Ru CMP cleaning.u 1 2 The effect of Ru under an etching process with zeta potential modification on post Ru CMP cleaning consisting of 1 ppm citric acid was evaluated as a function of etchant concentration at ph 1. The static etch rate of Ru and the contact angle on the Ru surface were measured as a function of al4 concentration at ph 1, as shown in Fig. 1. The static etch rate slightly increased but less than to 1 nm/min and the contact angle decreased significantly with the increase of al4 concentration. This result indicates that the etching and oxidation of Ru occurred simultaneously by al4. Whereas the roughness of Ru surface is slightly decreased and smoothed with the increase of al4 concentration that was observed in the previous study Figure 11 shows the zeta poten tial of Ru wafer at ph 1 after treating them in citric acid added cleaning solutions of ph 1 at different al4 concentrations to determine the effect of al4 concentration of zeta potentials. The (a) c, 5 t: <( 4 (J 3 cu t: (.) 1,.. i D,... 4'... i.. tt ph aio 4 Concentration (M) (b) (J t:: ctl c aio 4 Concentration (M).2 Figure 1. (Color online) (a) The static etch rate of Ru and (b) the contact angle on Ru surface as a function ofai4 concentration. Figure 12. (Color online) The particle removal efficiency in cleaning solu tion as a function ofai4 concentration.

6 H156 Journal of The Electrochemical Society, 158 (1) H152 H156 (11) zeta potential increased and gradually shifted to a positive value because Ru oxide was formed on Ru surface with the increase of ai4 concentration It can be expected that the cleaning solu tion has to be designed at low ai4 concentration to maintain neg ative polarity of Ru surface. To maintain the negative zeta poten tials,.1 M ai4 solutions were prepared for the post Ru CMP cleaning solutions. Figure 12 shows the PRE in citric acid based cleaning solution as a function of ai4 concentration. The addition of ai4 increased PRE to 97% when the ai4 concentrations remained lower than.1 M. However, lower PRE was observed at ai4 concentrations higher than.1 M. These different results might be attributed to the formation of positively charged hydrous Ru oxide. It is predict that the Ru surface zeta potential is similar for the increase of ai4 concentration form.1 M to.1 M. However the oxidation rate on Ru was reported to be faster than the etching rate of Ru at higher ai4 concentration. This positively charged surface causes attractive interaction between particles and Ru surfa ces which results in lower PRE. Therefore, efficient alumina particle removal from Ru surface can be achieved using an Ru cleaning so lution at ph 1 consisting of citric acid of 1 ppm concentration and sodium periodate of less than.1 M concentration. Conclusion Post Ru CMP cleaning solution was developed at alkaline condi tions using citric acid solutions by adding.1 M ai4 as an etch ant. In a solution with citric acid, the zeta potential of the alumina surface was changed to a negative value due to the adsorption of negative citrate ions. However, the hydrous Ru oxide, which has positive surface charge, could be formed on Ru surface in citric acid solution at ph 6 and 8. At ph 6 and 8, relatively low particle re moval efficiency was observed in citric acid solution due to the attractive force between the Ru surface and particles. At ph 1, the lowest adhesion force and highest cleaning efficiency were meas ured due to the repulsive force between the contaminated alumina particle and the Ru surface. When ai4 was added in citric acid solution at ph 1, PRE was improved due to the detachment of con taminants by the etching of the Ru surface. The highest PRE was achieved in citric acid solution with ai4 below.1 M at ph 1. The effective cleaning of contaminants was achieved by optimizing the electrostatic force and under etching of the contaminated Ru surface. Acknowledgment This work was supported by the research fund of Hanyang University (HY 1 G). References 1. K. K.i&h.iro,. Inoue, S. C. Chen, and M. Yoshlmaru, Jpn. J. Appl. Plrys. Part I, 37, 1336 (1998). 2. S. Y. Kang, K. H. Choi, S. K. Lee, C. S. Hwang, and H. J. Kim, J. Korean Phys. Soc., 37, 14 (). 3. Y. Park and J. T. Song, Mater. Lett., 58, 2128 (4). 4.. Chyan, T.. Anmagiri, and T. Ponnuswamy, ]. Electrochem. Soc., 15, C347 (3). 5. S. K. Cho, S. K. Kim, H. Han, J. J. Kim, and S.M. Oh, J. Vac. Sci. Techno/. B, 22, 2649 (4). 6. M. W. Lane, C. E. Murray, F. R. McFeely, P.M. Vereecken, and R. Rosenberg, Appl. Plrys. Lett., 83, 233 (3). 7.. K. Kwon, J. H. Kim, H. S. Park, and S. W. Kang, J. Electrochem. Soc., 151, G19 (4). 8. W. J. Lee and H. S. Park, Appl. Suif. Sci., 228, 41 (4). 9. L. Zhang, S. Raghavan, and M. Weling, ]. Vac. Sci. Techno/. B, 17, 2248 (1999). 1. F. Tardif, Semicond. Semimetals, 63, 186 (). 11. I. K. Kim, B. G. Cho, J. G. Park, J. Y. Park, and H. S. Park, ]. Electrochem. Soc., 156, H188 (9). 12. I. K. Kim, Y. J. Kang, T. Y. Kwon, B. G. Cho, J. G. Park, J. Y. Park, and H. S. Park, Electrochem. SolidState Lett., 11, H15 (8). 13. S. Y. Lee, S. H. Lee, and J. G. Park, J. Electrochem. Soc., 15, G327 (3). 14. Y. K. Hong, D. H. Eom, S. H. Lee, T. G. Kim, J. G. Park, and A. A. Busnaina, ]. Electrochem. Soc., 151, G756 (4). 15. B. Zhang, C. Zhang, H. He, Y. Yu, L. Wang, and J. Zhang, Chern. Mater., 22, 456 (1). 16. J. H. Jang, K. Machlda, Y. Kim, and K. aoi, Electrochim. Acta, 52, 1733 (6). 17. M. Pourbaix, Atlas of Electochemica/ Equilibria in Aqueous Solutions, ational Association of Corrosion Engineers, Houston, TX (1974). 18. H. T. Seo, J. Y. Park, T. Liang, and G. A. Somorjai, J. Electrochem. Soc., 157, H414 (1). 19. T. A. Ring, Fundamentals of Ceramic Power Processing and Synthesis, Academic Press, San Diego, CA (1996).. D. H. Eom, J. S. Ryu, J. G. Park, J. J. Myung, and K. S. Kim, Key Eng. Mater., , 389 (4). 21. P. B. Zantye, A. Kumar, and A. K. Sikder, Mater. Sci. Eng.,R, 45, 89 (4).