15 (2006) 1580–1584www.elsevier.com/locate/diamond

Diamond & Related Materials

Nucleation and growth of amorphous carbon film on tungsten-implantedstainless steel substrates

Ming Xu a,⁎, Liuhe Li a,b, Xun Cai a, Youming Liu a,c, Qiulong Chen a,Junxia Guo a, Paul K. Chu c

a School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200030, Chinab 702 Department, Mechanical Engineering School of Beijing University of Aeronautic and Astronautic, 100083, Chinac Department of Physics & Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong

Received 23 June 2005; received in revised form 28 December 2005; accepted 28 December 2005Available online 8 February 2006

Abstract

Amorphous carbon (a-C) films were deposited on W-implanted (20 kV, 3×1017 ions cm−2) and un-implanted steel substrates by plasmaimmersion ion implantation and deposition (PIII&D). The W implantation pretreatment changes the surface structure and impacts film nucleation.Consequently, the growth mechanism of the a-C film is altered resulting in different surface morphologies and roughnesses even though the filmsdeposited on the un-implanted steel substrates possess similar a-C structures as revealed by Raman spectroscopy. The structural differences areprobed by X-ray photoelectron spectroscopy and X-ray diffraction. Moreover, microstructural observations were carried out by transmissionelectron microscopy. A model based on the statistical formation theory is proposed to explain the growth of the a-C films on the implanted and un-implanted substrates.© 2006 Elsevier B.V. All rights reserved.

PACS: 81.05; 81.15; 68.35; 61.72Keywords: Amorphous carbon films; Nucleation; Growth; Tungsten; Plasma immersion ion implantation and deposition; Stainless steel

1. Introduction

Amorphous carbon (a-C) films have found applications inmany areas such as microelectronic and optical devices,biomedical products, corrosion resistant materials, protectiveovercoats, as well as microelectro-mechanical systems(MEMS) [1–5]. Their unique properties can be attributed tonot only the special and interesting properties of themicrostructures, but also their surface morphology. The directuse of a-C films on certain materials such as ferroussubstrates still faces some technological difficulties. Iron,the main element in stainless steel, enhances the deposition ofdisordered graphite and carbon has a high rate of diffusion inferrous substrates, thereby hampering the growth of a-C films.To our knowledge, there have been few reports investigating

⁎ Corresponding author. Tel.: +86 21 62932087; fax: +86 21 62932587.E-mail address: xml@sjtu.edu.cn (M. Xu).

0925-9635/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.diamond.2005.12.053

the a-C growth characteristics on pre-implanted steelsubstrates.

In this work, a-C films are deposited on 321 stainless steelsthat have been pre-implanted with W. Ion implantation altersthe surface chemistry of the steel substrate and the subsequentgrowth of the a-C film. The surface morphology androughness of the a-C film are also affected. The nucleationand growth mechanisms of the a-C films deposited on the pre-implanted and un-implanted substrates are investigated andcompared.

2. Experimental details

Coupons of 321 stainless steels (composition in wt.%: Fe:70.1, C: 0.11, Si: 0.90, Cr: 18.2, Ni: 9.40, Ti: 0.64, S: 0.06 andP: 0.03) were cut into dimensions of 15×15×2 mm3. Thesamples were austenized at 1500 K for 1 h to achieve a solidsolution state. They were then polished and cleaned withacetone before treatments.

0 5 10 15 20 25 300

10

20

30

40

50

60

70

80

C

Cr

C

W

Fe

Ato

mic

Co

nce

ntr

atio

n(%

)

Depth(nm)

20 kV, 3X1017 ions cm-2

W-implanted 321 stainless steel

Fig. 1. XPS depth profile of the steel substrate implanted with W ions (20 kV,3×1017 ions cm−2).

286 284 282 280 278In

ten

sity

(C

PS

) 284.4

283.3

C 1s

Binding energy (eV)

(a)

730 725 720 715 710 705 700

Inte

nsi

ty (

CP

S)

720.2

706.9

Fe 2p1/2

Fe 2p3/2

Binding energy (eV)

Inte

nsi

ty (

CP

S)

Binding energy (eV)

Fe2p

(b)

42 40 38 36 34 32 30

31.3

33.5

W 4f5/2

W 4f7/2

W 4f

(c)

Fig. 2. XPS spectra: (a) C1s, (b) Fe2p and (c) W4f.

1581M. Xu et al. / Diamond & Related Materials 15 (2006) 1580–1584

Tungsten ion implantation was carried out in a multipurposeplasma immersion ion implanter (PIII) equipped with severalplasma generating tools including RF discharge, hot filamentdischarge and vacuum arc metal plasma sources [6–8]. The basepressure in the vacuum chamber was 3×10−3 Pa. Tungsten PIIIwas conducted using the following conditions: target biasVi =20 kV, main arc average current Ia=1.0A, pulse repetitionrate f=40 Hz. Samples 1 and 2 underwent W PIII for about anhour to achieve an approximate implant fluence of 3×1017 ionscm−2. Synchronization of the target bias and vacuum arc pulsesensured pure metal PIII without significant metal deposition.After W PIII, C2H2 was introduced into the vacuum chamberand RF discharge was triggered inside the vacuum chamber toignite the plasma. In this mode, a-C thin films were depositedusing plasma immersion ion implantation and deposition(PIII&D) without breaking vacuum thereby eliminating poten-tial contamination during sample transfer. In PIII&D, the C2H2

plasma was also sustained by 1 kW hot filament discharge.Negative high voltage (−20 kV) pulses with a pulse width of400 μs were applied to the W-implanted substrate (sample 2) ata repetitive frequency of 40 Hz. The processing time was4 h and the a-C film thickness was approximately 0.5 μm.Sample 3 was prepared by conducting a-C film depositiondirectly onto the untreated steel substrate under similarPIII&D conditions as sample 2.

X-ray photoelectron spectroscopy (XPS) was employed toinvestigate the composition and chemical bonding of the W ion-implanted 321 stainless steel sample (sample 1). Elementaldepth profiles were acquired using argon ion bombardment atan approximate sputtering rate of 4.7 nm/min. The structuresand phases in the near surface were determined by X-raydiffraction and the microstructures were evaluated by transmis-sion electron microscopy (TEM). The structure and surfacemorphology of the a-C films were investigated by Ramanspectroscopy and atomic force microscopy (AFM). Meanwhile,the root mean square roughness (Rm) of the film surfaces wasdetermined by averaging the results obtained from five differentareas.

3. Results

Fig. 1 displays the elemental depth profiles acquired by XPSfrom the steel substrate after W PIII. The tungsten profile showsa typical ‘implant-like’ Gaussian distribution, implying that

30 40 50 60 70 80 90

Inte

nsi

ty (

CP

S)

α-Fe

α-Fe

untreated 321 stainless steel

2θ

Fe2TiO4

Fig. 3. XRD pattern acquired from the un-implanted 321 stainless steelsubstrate.

(a)

(b)

500nm

500nm

Fig. 4. TEM micrograph: (a) un-implanted 321 stainless steel and (b) W-implanted 321 stainless steel.

1582 M. Xu et al. / Diamond & Related Materials 15 (2006) 1580–1584

pure tungsten ion implantation has been achieved. Due to itschemical affinity with carbon, a large amount of carbon is foundon the substrate surface. The high-resolution C1s, Fe2p andW4f XPS spectra are exhibited in Fig. 2. The W4f spectrumreveals the W4f7/2 doublet corresponding to carbide-modifiedtungsten [9,10]. One of the two C1s peaks centered at 283.3 eVis associated with C–W bonding [11]. The other peak at284.4 eV indicates the presence of C–C bonding or graphiticC corresponding to sp2 [12–14]. The Fe2p peaks associatedwith α-Fe are observed at 720.1 eV and 706.9 eV [15–17].According to Porte [18], the binding energies of Fe2pelectrons are affected by tungsten atoms in the vicinity sothat the values are slightly lower than that of pure iron. Forcomparison, XRD was also performed on the untreatedsubstrate and the results are shown in Fig. 3. Only thediffraction peaks of α-Fe can be observed suggesting that nocarbide phase exists in the untreated substrate. The presenceof Fe2TiO4 results from natural surface oxidation.

Fig. 4 depicts the microstructures of the un-implanted andW-implanted 321 stainless steels as revealed by TEM.Compared to the un-implanted sample, the micrograph of W-implanted layer shows the existence of high-density disloca-tions induced by high-energy ion implantation. Being large andheavy, tungsten ions introduce substantial radiation damages tothe substrate surface resulting in defect strengthening.

Fig. 5(a) and (b) show the Raman spectra acquired from thea-C films deposited on the W PIII and untreated 321 stainlesssteel samples, respectively. By using Gaussian multipeakfitting, both Raman spectra can be deconvoluted into twosubpeaks: G band at 1560 cm−1 and D band at approximately1360 cm−1. Because of the same deposition conditions, the twoa-C films yield similar Raman spectra, which indicate that bothfilms are amorphous carbon. Our results show that, owing to thesame preparation parameters, the two samples have similar filmstructures that are not modified by the different substratesurfaces.

Atomic force microscopy (AFM) was performed to study indetails the surface morphology of the a-C films. Fig. 6 depicts

the AFM images over a 1 μm2 area on the two samples. Themorphology of the a-C film on the untreated substrate resemblesthat of the film on the pre-implanted substrate to some extent.Nonetheless, denser buds and shallower pits are observed onsample 2 (a-C film/W-implanted steel) compared to sample3 (a-C film/untreated steel). The root mean square roughness(Rm) of the film surface is also improved from 6.98 nm to4.12 nm after the pretreatment.

4. Discussion

There are three recognized thin-film growth modes at thepost-nucleation stage before the film develops a typicalstructure zone model (SZM) structure: (a) layer or Frank–van

Fig. 6. AFM images: (a) a-C film on the untreated s

800 1000 1200 1400 1600 1800 2000

Sample 2

DG

Raman shift (cm-1)

Inte

nsi

ty

(a)

488nm

488nm Sample 3

(b)

Fig. 5. Raman spectra of sample 2 (a-C film deposited on W-implanted 321stainless steel) and sample 3 (a-C film deposited on un-implanted 321 stainlesssteel).

1583M. Xu et al. / Diamond & Related Materials 15 (2006) 1580–1584

der Merwe mode, (b) layer-plus-island or Stranski–Krastanov(SK) mode and (c) island or Volmer–Weber mode. The growthmode depends upon the deposited strain energy and surfaceenergy. The growth mechanisms of the a-C films deposited byPIII&D can be explained using the statistical formation theory[19]. The growth models for the a-C films on the W-implantedsubstrate and without pretreatment are illustrated in Fig. 7.

The carbon ion density in PIII&D is usually less than that inother deposition methods such as vacuum arc deposition (VAD)and the formation of critical carbon nucleation sites is relativelydifficult. As shown in Fig. 7(a), the carbon deposit nucleates onthe substrate surface randomly or at special defect or impuritysites on the substrate surface. The nuclei then grow by theaddition of carbon adatoms and their edges merge with eachother, as illustrated by the island-growth mode [19,20]. In Fig. 7(b), a modified layer, in which carbon, tungsten and othersubstrate elements have been mixed, is delineated from thesubstrate according to our XPS results. Through W implanta-tion, tungsten carbide and C–C bonds form on the steelsubstrate prior to the a-C film deposition. Previous studies haveshown that tungsten carbide (WC) is a very stable carbide phasewith a low carbon diffusivity and high thermal stability. It hasbeen found to be most effective in enhancing the nucleationdensity of diamond coatings [21]. Compared to the untreatedsteel substrate, the presence of graphite C can provide arelatively lattice-matched template between the substrate and

teel and (b) a-C film on W-implanted substrate.

Carbonion

Carbonion

substrate

Carbonfilm

Carbonfilm

(a)

Substrate

Modified layer

(b)

Fig. 7. Postulated a-C film growth models: (a) a-C film growth on the untreatedsteel substrate and (b) a-C film growth on the W-implanted steel substrate.

1584 M. Xu et al. / Diamond & Related Materials 15 (2006) 1580–1584

a-C film so as to reduce the deposited strain energy and act asthe precursor of film nucleation. High-energy tungsten ionimplantation leads to radiation damages in the substrate surfaceleading to defect strengthening (Fig. 4). The defects existing onthe surface can be the preferred nucleation sites, which changethe film nucleation conditions after ion implantation. Conse-quently, the nucleation densities are dramatically improved.Subsequent growth stems from the very dense islands leading tothe development of larger, more closely packed buds with asmoother film surface (Fig. 6(b)). In this case, there is somesimilarity to SK growth and it is different from the a-C film/untreated substrate on which they contact each other giving riseto deep valleys (Fig. 6(a)). The substantial differences betweenthe results of the untreated substrate and W-implanted substratewith respect to the surface roughness confirm the effects of thesubstrate materials on the film structure and our observationsare consistent with those reported by others [22].

5. Conclusion

We have investigated and compared the a-C growth modes atthe post-nucleation stage with and without W pre-implantation.

Even though the Raman spectra show similar structures, thegrowth mechanisms and surface morphologies of the two filmsdeposited on the implanted and un-implanted substrates aredistinctly different. After ion implantation, the modified surfaceshows the presence of WC and the C–C bonds and defectsdramatically enhance the nucleation density. It is believed thatthe different surface chemistry results in the shifting of the post-nucleation stage from the island mode to SK mode. A smootherfilm with improved surface roughness is produced on thepretreated substrate under the same deposition conditions.

Acknowledgements

This work was financially supported by Hong KongResearch Grants Council (RGC) Competitive EarmarkedResearch Grant (CERG) No. CityU 1120/04E and NationalNatural Science Foundation of China No.50271004.

References

[1] B. Bhushan, Tribology and Mechanics of Magnetic Storage Devices, 2ndEd., Springer, New York, 1996.

[2] J. Robertson, Mat. Sci. Eng. Rep. 37 (4–6) (2002) 129.[3] X.Y. Liu, P.K. Chu, C.X. Ding, Mat. Sci. Eng. Rep. 47 (2–4) (2004) 49.[4] R.W.Y. Poon, K.W.K. Yeung, X.Y. Liu, P.K. Chu, C.Y. Chung, W.W. Lu,

K.M.C. Cheung, D. Chan, Biomaterials 26 (15) (2005) 2265.[5] S.C.H. Kwok, J. Wang, P.K. Chu, Diamond Relat. Mater. 14 (1) (2005) 78.[6] Z.M. Zeng, T. Zhang, X.B. Tian, B.Y. Tang, T.K. Kwok, P.K. Chu, Surf.

Coat. Technol. 128–129 (2000) 236.[7] P.K. Chu, S. Qin, C. Chan, N.W. Cheung, L.A. Larson, Mat. Sci. Eng. Rep.

17 (6–7) (1996) 207.[8] P.K. Chu, B.Y. Tang, L.P. Wang, X.F. Wang, S.Y. Wang, N. Huang, Rev.

Sci. Instrum. 72 (3) (2001) 1660.[9] D. Briggs, M.P. Seah, second edition, Practical Surface Analysis, vol. 1,

John Wiley and Sons, 1993.[10] C. Rincón, G. Zambrano, A. Carvajal, P. Prieto, H. Galindo, E. Martínez, et

al., Surf. Coat Technol. 148 (2001) 277.[11] K. Hamrin, G. Johansson, A. Fahlman, C. Nordling, J. Phys. Chem. Solids

30 (1969) 1835.[12] L. Ley, M. Cardona, Y. Baer, M. Campagna, W.D. Grobman, Topics in

Applied Physics, L. Ley and M. Cardona, vol. 27, Springer-Verlag, BerlinHeidelberg New-York.

[13] N. Moncoffre, G. Hollinger, H. Jaferezic, Nucl. Instrum. Methods Phys.Res., B 7/8 (1985) 177.

[14] J. Diaz, G. Paolicelli, S. Ferre, Phys. Rev., B 54 (11) (1996) 8064.[15] A.S. Lim, A. Atrens, Appl. Phys., A 51 (1990) 411.[16] J.H. Yang, T.H. Zhang, Vacuum (China) 4 (2001) 11.[17] J.H. Yang, T.H. Zhang, Mater. Sci. Eng., A 362 (2003) 200.[18] L. Porte, Phys. Rev., B 28 (1983) 3214.[19] L.H. Li, X.B. Tian, P.K. Chu, Y.H. Zhang, X.M. Cui, H.Q. Zhang, Nucl.

Instrum. Methods Phys. Res., B 206 (2003) 691.[20] M. Prutton, Surface Physics, Second Edition, 126.[21] A. Kumar, Q. You, J.S. Kapat, A. Mangiaracina, A. Catletge, Y. Vohra,

Thin Solid Films 308–309 (1997) 209.[22] H. Savaloni, M.A. Player, Thin Solid Films 256 (1995) 48.