Preparation and characterization of ruthenium films via an electroless deposition route

(1)

Preparation and characterization of ruthenium

ﬁlms via an electroless

deposition route

Jing-Yu Chen, Li-Yeh Wang, Pu-Wei Wu

⁎

Department of Materials Science and Engineering, National Chiao Tung University, Hsin-Chu 300, Taiwan, ROC

a b s t r a c t

a r t i c l e i n f o

Available online 7 May 2010 Keywords:

Ruthenium Ruthenium oxide Electroless deposition

A metallic Rufilm was prepared by an electroless deposition method, followed by hydrogen reduction treatment. The electroless deposition formulation produced a solidfilm on a Cu-coated Si substrate at 40 °C preactivated by PdCl2solution. Chemicals including K2RuCl5·xH2O, NaNO2, NaOH, and NaClO were mixed in a proper ratio that enabled heterogeneous nucleation andfilm growth. Results from X-ray photoelectron spectroscopy (XPS) on the as-depositedfilms confirmed the presence of RuOxand Ru, while X-ray diffraction (XRD) pattern suggested an amorphous nature. Planar images from a scanning electron microscope revealed a rather smooth surface at thickness less than 250 nm. Above that formation of surface crack and partial detachment from the substrate were observed. After hydrogen reduction at 200 °C for 2 h, we obtained a metallic Rufilm, as confirmed by XPS and XRD. In addition, the surface roughness was increased due to the formation of pinholes that was caused by the volume contraction associated with RuOxreduction to Ru.

1. Introduction

In the family of platinum group metals (PGM), Ru is a relatively inexpensive element that exhibits low electrical resistivity and superb chemical stability. Hence, Ru has attracted signiﬁcant attention recently for possible applications in electrocatalysis and semiconductor devices [1,2]. The oxidized form of Ru, RuOx, is also of particular interest because

of its unusually low electrical resistivity. For example, RuOxis widely

studied as the active material for pseudocapacitors, as well as electrocatalysis to promote the oxygen evolution reaction in a water electrolyzer[3,4].

To date, a variety of deposition methods have been explored to prepare Ru and RuOx ﬁlms. For example, vacuum-based approaches

including physical vapor deposition, chemical vapor deposition, and atomic layer deposition have been demonstrated with various successes [5_–7]. In general, these techniques involve expensive setups, long processing time, and excessive material waste. These drawbacks render them impractical to produce Ru and RuOxﬁlms in a commercial scale.

Therefore, considerable efforts have been devoted to identify an alternative route in solution-based depositions such as electroplating and sol–gel synthesis[8,9].

Electroplating is a known practice to fabricateﬁlms of metals and oxides[10]. However, to obtain a uniform deposit the substrate to be electroplated requires adequate electrical conductivity and simple surface contour. These criteria limit the applicability of electroplating

to planar conductive substrates. In contrast, electroless deposition is able to operate on both conductive and insulating platforms in various shapes [11]. This is because the driving force for the electroless deposition comes from the reducing agent in the plating solution, which allows preferential nucleation and growth on the substrates. To date, among many metals derived from the electroless method, the Ni–P system has received the most scrutiny [12]. Typical steps for the electroless deposition entail substrate sensitization and activation, followed by electroless plating. Variables including pH, reducing agent, temperature, and additive are critical in determining theﬁlm qualities.

Among the PGM elements, the electroless depositions of Pd and Pt have been widely documented and their practices are rather common [10]. In contrast, there are much fewer studies on the remaining PGM metals including Ru, Os, Rh, and Ir[13,14]. Our experiences indicated that the electroless deposition for a solid Ru film is extremely challenging because in solution the Ru cation often exists in multiple oxidation states. As a result, possible disproportionation reactions would take place that renders homogeneous precipitation everywhere instead of desirable heterogeneousfilm growth on the substrate. In this work, we develop a practical Ru electroless deposition formulation and demonstrate the formation of Ru/RuOxcompositefilm on a Cu-coated Si

substrate. After H2 reduction, we convert the compositeﬁlm to a

crystalline Ruﬁlm. 2. Experimental details

An 8-inch Si wafer predeposited with SiO2(500 nm), TaN (20 nm),

and Cu (60 nm) was used as the substrate. The wafer was broken into small pieces in 2 × 2 cm2, and rinsed with acetone and water to

Thin Solid Films 518 (2010) 7245–7248

⁎ Corresponding author.

E-mail address:[email protected](P.-W. Wu).

Contents lists available atScienceDirect

Thin Solid Films

(2)

remove surface residues. Next, the sample underwent an activation treatment by submerging in an aqueous solution including 0.1 wt.% PdCl2(Aldrich) and 1 wt.% HCl (SHOWA) for 10 s. Subsequently, the

activated substrate was immersed in the plating bath which contained 30 mL of aqueous solution consisting of 0.0186 g K2RuCl5·xH2O (Alfa

Aesar), 0.0704 g NaNO2(SHOWA), 0.04 g NaOH (Mallinckrodt), and

1.8858 g NaClO (SHOWA). We carried out the electroless deposition at various times to estimate the deposition rate. Both the activation step and electroless deposition were conducted at 40 °C. Afterward, the sample was kept in the plating solution for 40 min and dried in air. To prepare a pure Ruﬁlm, we also performed a reduction treatment on the sample at 200 °C for 2 h in a mixture of 50% H2and 50% Ar.

The morphology and thickness for the as-depositedfilms were observed by afield emission scanning electron microscope (FESEM: JSM 6700). An atomic force microscope (AFM: Vecco Dimension 5000 Scanning Probe Microscopy) was used to determine their surface roughness. Phase and crystallinity for the samples were identified by a high-resolution X-ray diffractometer (HRXRD: Bede D1). The oxida-tion states of Ru and O were con_{firmed by an X-ray photoelectron} spectrometer (XPS: Thermo Microlab 350).

3. Results and discussion

The electroless formation of Ru nanoparticles is very straightfor-ward provided suitable reducing agents are employed. Unfortunately, due to the presence of multiple oxidation states for the Ru cations, it becomes rather challenging to deposit a continuous Ru_{ﬁlm via the} electroless route. To achieve preferential growth of Ru on the activated substrate, we designed a chemical bath with both reducer and oxidizer. In our formulation, the K2RuCl5·xH2O provided the

source for Ru3+_{cation while the NaNO}

2acted as the reducing agent.

Because the NaNO2was unable to reduce Ru3+directly, we adopted

an indirect route in which the NaClO was added in the presence of NaOH to oxidize the Ru precursor to RuO4. Subsequently, RuO4was

reduced by the NaNO2to Ru and RuOx, respectively. The thickness for

the as-depositedﬁlm at various times is provided inFig. 1. As shown, the deposit thickness grew faster at the beginning but became slower at a later stage. This behavior is expected as available Ru ions in the plating bath decreased with time.

To verify the chemical nature for the deposits, we employed XPS to determine the oxidation states for Ru and O. To obtain sufﬁcient signal strength, we selected the sample after 480 min of electroless deposition.Fig. 2(a) displays the XPS pro_{ﬁles for the Ru 3p}3/2before

and after hydrogen reduction. Although the Ru 3d signals are the strongest in XPS response, the binding energies for the metallic Ru

and oxidized Ru are close to each other. Hence, it is difficult to determine the exact oxidation state for the Ru in the as-deposited film. Therefore, we decided to adopt the Ru 3p signals because they provided a better distinction between Ru and oxidized Ru. The other reason for not using the Ru 3d signals was to minimize possible interference from carbon. From the XPS pro_{files, apparently, the} as-deposited film revealed a strong peak at 464.2 eV, confirming the presence of Ru. However, its position was slightly shifted relative to the metallic Ru at 462.2 eV[15], which suggested that the Ru existed both at metallic and oxidized states. Moreover, XPS results from the as-depositedfilms at various plating times demonstrated negligible variation, indicating reasonable consistence in thefilm composition. On the other hand, after hydrogen reduction the peak position was located at 462.7 eV. This inferred that a metallic Ru state was obtained. According to Matsui et al., in a hydrogen atmosphere, 200 °C was adequate to reduce RuOxto Ru[16]. This substantiates our

XPS results of Ru formation after hydrogen reduction.

Fig. 2(b) provides the XPS proﬁles from O 1s signal on identical samples. As shown, before hydrogen treatment the peak position was 531.2 eV. From the XPS database, the O from RuO2 and H2O was

expected to be 529.4 and 533.3 eV, respectively [15]. Hence, our signals inferred that Ru was likely in a hydrous form. This possibility was also suggested by Chang and Hu when they determined the O 1s signal from Ru–OH at 531.2 eV[17]. Interestingly, the O 1s signal after hydrogen reduction exhibited a slight shift to 532.1 eV, a value closer to what appeared to be surface H2O. These evidences agreed well with

those from the Ru 3p3/2, indicating that the reduction of RuOxto Ru

was achieved. We did not perform XPS curveﬁtting for O 1s signal because the oxygen can exist in states of RuOx, adsorbed H2O, and

hydrated RuOx. Previously, Shen et al. discussed similar issues and

determined that the O 1s is difﬁcult to assign[18]. In addition, the adsorbed H2O can still be detected for theﬁlm after H2reduction. As a

result, the O 1s signal was shifted less than 1 eV after H2reduction.

Fig. 3presents the XRD patterns for the substrate, as well asﬁlms before and after hydrogen reduction. As shown, the as-depositedﬁlm revealed a broad peak from 30 to 40° amid considerable background noises. However, this pattern is inconsistent with those from standard diffraction patterns of Ru (JCPDS: 060663) and RuO2(JCPDS: 401290)

but agreed well with that of the substrate. Hence, we surmised that the as-depositionfilm was amorphous in nature. In contrast, the XRD pattern after hydrogen reduction exhibited notable diffraction peaks and they were identified from the metallic Ru at 38° (100), 42° (002), and 44° (101), respectively. Similarly, those minor peaks between 30 and 40° were attributed to the Si substrate. However, we realized that the crystallinity for the Ru_{film was compromised moderately judging} from the broad and relatively noisy background.

At this stage, evidences from XPS and XRD indicated that we deposited an amorphous Ru/RuOx composite ﬁlm and it was

converted to a metallic Ru ﬁlm after hydrogen reduction. Fig. 4 presents the SEM images in planar view for the as-deposited Ru/RuOx

ﬁlms. As shown in Fig. 4(a), the Ru/RuOx ﬁlm after 120 min of

electroless deposition revealed a smooth morphology with a thickness of 100 nm. In addition, there was presence of minute particles on the surface. In our SEM observations, the as-deposited films maintained reasonable surface uniformity within 240 min of deposition time. Above that, we noticed minor cracks around surface protrusions, and those cracks grew larger at increasing plating time. At 480 min, thefilm was broken into smaller pieces detaching from each other, as shown inFig. 4(b). According to Chen et al.[19], the development of internal stress for a deposit during electroless deposition is attributed to both extrinsic and intrinsic factors. The extrinsic stress is known as“thermal stress” and it is associated with a mismatch in thermal expansion coefficient between the substrate and deposit. In contrast, the intrinsic stress is attributed to the processing parameters that lead to microstructural inhomogeneity such as grain boundaries, vacancies, and voids. Since the as-depositedfilms were

Fig. 1. Thickness variation for the as-depositedﬁlm as a function of plating time.

(3)

amorphous mixture of Ru and RuOx, the presence of vacancies and

voids was certainly possible to produce_{film cracking. On the other} hand, thermal stress was not considered to be a factor because we performed the electroless deposition at 40 °C and the as-deposited film was cooled to room temperature slowly. Lastly, the lattice parameter mismatch between the deposit and substrate was unlikely to produce film cracks because the as-deposit film revealed an amorphous mixture.

The SEM images for the Ruﬁlms after hydrogen reduction are provided inFig. 5. The sample inFig. 5(a) was obtained from 120 min of electroless deposition and 2 h of hydrogen reduction. Clearly, there appeared many pinholes on the surface but theﬁlm still maintained its integrity. We understood that the conversion of RuOx to Ru

involved considerable volume contraction that might be responsible

Fig. 2. XPS spectra for (a) Ru 3p3/2line obtained from the as-depositedﬁlm and after hydrogen reduction, and (b) O 1s line from the same ﬁlm before and after hydrogen reduction.

Fig. 3. XRD patterns for the substrate, as-depositedﬁlm after 480 min electroless deposition, and the sameﬁlm after hydrogen reduction at 200 °C for 2 h.

Fig. 4. SEM images in plain view for the Ru/RuOxﬁlms after (a) 120 and (b) 480 min of

electroless deposition.

7247 J.-Y. Chen et al. / Thin Solid Films 518 (2010) 7245–7248

(4)

for the observed pinholes. As expected, when the plating time was increased, the number of surface pinholes in the reducedﬁlm was increased as well. In addition, we also witnessed the increase of surface roughness after the hydrogen treatment. The roughness value (Rq) from AFM for the as-depositedﬁlm was 4.59 nm, and it became

14.85 nm after hydrogen reduction.Fig. 5(b) displays the image from the sample of 480 min electroless deposition and 2 h hydrogen reduction. Apparently, there were severe structural alterations on the surface with protrusion of crystallites and cracks that penetrated to the substrate. This might explain the substantial noises recorded in earlier XRD and XPS analysis.

4. Conclusions

A composite Ru/RuOx ﬁlm was prepared on a Cu-coated Si

substrate by an electroless deposition method in which activation and electroless plating were carried out consecutively at 40 °C. The plating bath contained chemicals including K2RuCl5·xH2O, NaNO2,

NaOH, and NaClO that enabled heterogeneous nucleation andfilm growth on the activated substrate. Results from XPS and XRD on the as-depositedfilms confirmed the presence of RuOx and Ru in an

amorphous structure. The images from SEM indicated a solidfilm with smooth surface at thickness below 250 nm. Above that, surface cracking was observed that lead to film detachment from the substrate. After hydrogen reduction at 200 °C for 2 h, we obtained a metallic Rufilm with surface pinholes caused by volume contraction associated with RuOxreduction.

Acknowledgment

Financial support from the National Science Council of Taiwan (NSC 98-2221-E-009-040-MY2) is greatly appreciated.

References

[1] J.S. Spendelow, A. Wieckowski, Phys Chem Chem Phys 6 (22) (2004) 5094. [2] S.K. Cho, S.K. Kim, H. Han, J.J. Kim, S.M. Oh, J Vac Sci 22 (6) (2004) 2649. [3] J.H. Kim, D.S. Kil, S.J. Yeom, J.S. Roh, N.J. Kwak, J.W. Kim, Appl Phys Lett 91 (5)

(2007) 052908.

[4] H.C. Ma, C.P. Liu, J.H. Liao, Y. Su, X.Z. Xue, W. Xing, J Mol Catal A Chem 247 (1-2) (2006) 7.

[5] J.G. Lee, Y.T. Kim, S.K. Min, S.H. Choh, J Appl Phys 77 (10) (1995) 5473. [6] J.H. Shin, A. Waheed, W.A. Winkenwerder, H.W. Kim, K. Agapiou, R.A. Jones, G.S.

Hwang, J.G. Ekerdt, Thin Solid Films 515 (13) (2007) 5298.

[7] S.K. Kim, S.Y. Lee, S.W. Lee, G.W. Hwang, C.S. Hwang, J.W. Lee, J. Jeong, J Electrochem Soc 154 (2) (2007) D95.

[8] Y.H. Lee, J.G. Oh, H.S. Oh, H. Kim, Electrochem Commun 10 (7) (2008) 1035. [9] Y.S. Kim, H.I. Kim, J.H. Cho, H.K. Seo, G.S. Kim, S.G. Ansari, G. Khang, J.J. Senkevich,

H.S. Shin, Electrochim Acta 51 (25) (2006) 5445.

[10] M. Paunovic, M. Schlesinger, R. Weil, in: M. Schlesinger, M. Paunovic (Eds.), Modern Electroplating, John Wiley & Sons, New York, 2000, p. 1.

[11] H. Matsubara, T. Yonekawa, Y. Ishino, N. Saito, H. Nishiyama, Y. Inoue, Electrochim Acta 52 (2) (2006) 402.

[12] K.I. Popov, S.S. Djokic, B.N. Grgur, Kluwer Academic/Plenum Publishers, New York, 2002.

[13] Y. Okinaka, C. Wolowodiuk, in: G.O. Mallory, J.B. Hajdu (Eds.), Electroless Plating : Fundamentals and Applications, American Electroplaters and Surface Finishers Society, Orlando, Fla, 1990, p. 421.

[14] C.R.K. Rao, D.C. Trivedi, Coord Chem Rev 249 (5-6) (2005) 613.

[15] NIST X-ray Photoelectron Spectroscopy Database, Version 3.5 (National Institute of Standards and Technology, Gaithersburg, 2003); http://srdata.nist.gov/xps [16] Y. Matsui, M. Hiratani, S. Kimura, J Mater Sci 35 (16) (2000) 4093. [17] K.H. Chang, C.C. Hu, J Electrochem Soc 151 (7) (2004) A958. [18] J.Y. Shen, A. Adnot, S. Kaliaguine, Appl Surf Sci 51 (1–2) (1991) 47.

[19] Z. Chen, X.D. Xu, C.C. Wong, S. Mhaisalkar, Surf Coat Technol 167 (2-3) (2003) 170. Fig. 5. SEM images in plain view for the Ruﬁlms obtained after (a) 120 and (b) 480 min

of electroless deposition followed by hydrogen reduction at 200 °C for 2 h.