Enhanced Nucleation and Growth of Diamond Film on Si by CVD Using a Chemical Precursor

(1)

Published: July 05, 2011

ARTICLE

pubs.acs.org/JPCC

Enhanced Nucleation and Growth of Diamond Film on Si

by CVD Using a Chemical Precursor

Rajanish N. Tiwari,*

,†,§

Jitendra N. Tiwari,

‡

Li Chang,

§

and M. Yoshimura

†

†_{Surface Science Laboratory, Toyota Technological Institute, 2-12-1 Hisakata Tempaku, Nagoya 468-8511, Japan} ‡_{Center for Superfunctional Materials, Department of Chemistry, Pohang University of Science and Technology, San 31,}

Hyojadong, Namgu, Pohang 790-784, Korea

§_{Department of Materials Science and Engineering, National Chiao Tung University, 1001 Ta Hsueh Road, Hsinchu, Taiwan,} 30050, R.O.C.

1. INTRODUCTION

Diamond has a variety of outstanding properties, such as wide band gap, chemical inertness, high carrier mobility, excellent biological compatibility, high propagation speed of acoustic wave, good optical transparency, high thermal conductivity, and the greatest hardness, which make it a promising candidate for a wide range of applications, for instance, microelectronics, optics, tribological, thermal management, biomedical, DNA-based sensors, manufacturing engineering, and so on.1 7 Plati-num has been used as a substrate for the growth of chemical vapor deposited (CVD) diamondfilm. Recently, platinum (Pt) is used for high-pressure, high-temperature diamond synthesis as a catalytic solvent.8It is also known that Pt is a strong catalyst for hydrogen and hydrocarbons. Therefore, there is definite interest in studying the effects of Pt on diamond growth. The lattice constant of diamond and platinum (fcc) is 3.56 and 3.92 Å, respectively. Therefore, the lattice difference between diamond and Pt is approximately ∼10%.9 Tachibana et al. have studied diamond growth on Pt substrate by a bias-enhanced nucleation method. They had noticed that the density of diamond nuclea-tion was enhanced with bias time.10Most recently, the growth of diamondfilms on a Pt layer was realized by microwave plasma chemical vapor deposition (MPCVD) without bias treatment. In this technique, they used diamond seeding particles to enhance the diamond nucleation.11Although diamondfilms on a Pt/non-Si substrate, for example, single-crystal SrTiO3(111) or sapphire

substrates have been studied.12,13However, very few studies of diamond growth on a Pt/Si substrate have been reported at partially low temperature. The prevention of silicidation is the key to growing high-quality diamond on Pt/Si.

In this paper, we report for the ﬁrst time chemical vapor deposition of diamondﬁlms on Pt/SiO2/Si by using adamantane as a seeding layer to enhance the diamond nucleation at partially low temperature. Adamantane (C10H16) is one of a series of carbon structures, a very stable crystalline compound, and a highly symmetric molecule with point group symmetry, Td. Adamantane is the smallest possible diamondoid (chemical formula C(4n + 6)H(4n + 12), where n = 0, 1, 2, 3...), consisting of 10 carbon atoms arranged as a single diamond cage sur-rounded by 16 hydrogen atoms.14 16 A cagelike structure is formed with six CH2 and four CH groups giving rise to a molecular structure with four cyclohexane rings in chair form. Its structure has zero strain as all C C C bond angles are 109.45 with a corresponding bond length of 1.54 Å, and the C H bond length is 1.1 Å. The density of adamantane is 1.07 g/cm3. Adamantane does not melt at ambient pressure but sublimates. Adamantane can sublime easily and has a relatively high vapor pressure. Partial breakdown of adamantane is known to yield Received: May 4, 2011

Revised: June 21, 2011 ABSTRACT:Nucleation of diamond is of great importance for

its growth, and a detailed understanding of the nucleation process is, therefore, desired for many applications. The pre-treatment of the substrate surface may impact the initial growth period. This study demonstrates the synthesis of diamondﬁlms by microwave plasma chemical vapor deposition on a Pt/SiO2/ Si substrate. The Pt particles were deposited on the SiO2/Si surface at room temperature, whereas adamantane was seeded on the SiO2/Si surface by ultrasonic treatment. The Pt particles on the SiO2surface behave as a catalyst, which adsorb hydro-carbons from adamantane contained plasma and give distinct

features of carbon in the early stage deposition. The adamantane transforms not only in the nanodiamond phase but also in other carbon phases, which can then act as nuclei for diamond growth. These kinds of phases enhanced the diamond density at partially low temperature deposition. The presence of an oxide intermediate layer between Pt and Si prevents silicidation as well as SiC in diamond deposition.

(2)

carbon clusters (CnHx), where n = 3, 5, 6, 7, 8, and 9, of signiﬁcant abundance.17,18In addition, adamantane easily dissolves in or-ganic solvents and is widely commercially available. Therefore, in this study, we used hexane (C6H14) as a solvent. The evapo-ration properties of hexane are high. After adamantane is seeded on the Pt/Si surface by ultrasonication, the hexane will be evapo-rated from the Si surface and only pure will be adamantane left on the Si surface. The Pt interlayer on the Si substrate adsorbs the hydrocarbons from carbon source materials (adamantane), which can play a major role in diamond nucleation and growth. The nucleation and growth of diamond on the Pt/Si substrate will be discussed. Moreover, the rate ofﬁlm growth in MPCVD with 1% methane (CH4) in hydrogen (H2) was 0.45μm h 1.

2. EXPERIMENTAL METHODS

The schematic diagram for the deposition of diamond on Pt/ SiO2/Si is shown in Figure 1. The synthesis processes of the diamondﬁlms on Pt/SiO2/Si are described as follows: mirror-polished p-type (100) silicon wafers with dimensions of 1 1 cm2without any mechanical pretreatment were used as the sub-strates. The substrates were ultrasonically cleaned with acetone and alcohol for 10 min, respectively, and then dried with high-pressure nitrogen gas to remove any contamination particles left on the surface. The oxide layer prevented silicidation between Pt and Si; therefore, we did not use BOE solution to remove the native oxide (SiO2) from the Si surface. Further, the platinum particles/ﬁlm were coated on SiO2/Si substrates by a sputtering process for 210 s and then again cleaned with acetone and alcohol for 10 min, respectively. The Pt particles were not uniformly distributed on the Si surface. The average thickness/size of Pt particles is∼50 nm. Further, adamantane was seeded on cleaned Pt/SiO2/Si samples by an ultrasonic method. In this method, the solution was prepared by using (100 mg) adamantane + (5 mL) hexane and stirring for 1 min (the commercial adamantane powders with 99+% purity were obtained from Sigma-Aldrich Chemie GmbH (CAS: 281-23-2)). After manually stirring for 60 s at room temperature, the adamantane was fully dissolved into the hexane solvent and became a clear (colorless) solution. The Pt/SiO2/Si samples were then immersed into the solution for 10 min of ultrasonication. After ultrasonication, the adamantane was seeded on the Pt/SiO2/Si surface. We did not use any kind of process to dry the sample. Because the evaporation property of hexane is very high, hexane evaporated from the sample and

adamantane was left on the Pt/SiO2/Si surface. Hexane is a colorless liquid, and it has considerable vapor pressure at room temperature. Hexane is relatively safe, largely unreactive, and easily evaporated. Finally, the adamantane-seeded Pt/SiO2/Si substrates were then placed on a Mo-disk holder for diamond growth in a 1.5 AsTeX-type MPCVD system. The detailed processing parameters for MPCVD are as follows: the total pressure was 20 Torr, the microwave power was 600 W, the temperature wase700 C, the total ﬂow rate was 200 sccm (1% CH4in H2), and the deposition time was varied from 2 to 120 min. The surface temperature was measured by an optical pyrometer. Finally, the samples were allowed to cool down to ambient temperature in the presence of hydrogen gas at 10 Torr. To know the role of Pt, we also prepared Si substrates without a Pt coating and treated with the same experimental conditions.

The surface morphology of the sample was examined using ﬁeld-emission scanning electron microscopy (SEM, JEOL JSM-6700F), and the chemical composition of the surface was char-acterized by X-ray photoelectron spectroscopy (XPS, Thermo VG 350F, Mg K X-ray source). Elemental mapping in cross section was performed using a PHI700 scanning Auger nano-probe (ULVAC-PHI Inc.). Examination of the internal micro-structures of the deposited materials was carried out using transmission electron microscopy (TEM, Philips Tecnai20 operated at 200 kV). The cross-sectional TEM specimen was prepared using a focused ion beam (FIB, FEI Nova 200 dual-beam FIB). For the protection of the TEM specimen against damage from the high-energy ion beam (30 keV Ga+), the specimen was coated with platinum. Raman spectroscopy was Figure 1. Schematic diagram showing diamond synthesis inﬁve steps:

(a) silicon substrate with native oxide layer, (b) Pt coated on the SiO2/Si

substrate by a sputtering process, (c) Pt/SiO2/Si immersed into

the hexane + adamantane solution, (d) adamantane deposited on the Pt/SiO2/Si surface by ultrasonication, and (e) diamond growth by

MPCVD.

Figure 2. (A) Plan-view SEM image (inset is high-magniﬁcation image), (B) cross-sectional view, (C) FTIR spectrum, (D) Raman spectrum, and (E) X-ray diﬀraction pattern of adamantane on the Pt/ SiO2/Si after the ultrasonic process.

(3)

The Journal of Physical Chemistry C _ARTICLE

performed using a LABRAM HR800 system with an Ar laser (wavelength = 514.5 nm), which could be focused to a diameter of 1 mm for micromode operation. The crystallinity of the diamond was evaluated using an X-ray diﬀractometer (D2 XRD Bruker). 3. RESULTS AND DISCUSSION

After adamantane seeding on the Pt/SiO2/Si surface by the ultrasonic treatment, the surface morphology was characterized by scanning electron microscopy (SEM). The plan-view SEM image is shown in Figure 2A. From the SEM image, it is clear that the adamantane particles are homogeneously dispersed on the Pt/SiO2/Si surface by ultrasonic treatment. The inset high-magnification SEM image in Figure 2A shows that the adaman-tane particle sizes are varied from 10 nm to 1.5μm. The cross-sectional SEM image in Figure 2B shows the thickness of adamantane particles. From the cross-sectional SEM image, it is clear that the adamantane particles∼100 nm in thickness are deposited on the Pt/SiO2/Si surface after 10 min of the ultra-sonic process. Further, the bonding characteristics of the ada-mantane film were identified by Fourier transform infrared spectroscopy (FT-IR) and Raman spectroscopy, as shown in Figure 2C,D. There are several weak and strong peaks in Figure 2C. The peak at 1037 cm 1 was assigned to the CC stretch/CH bend, whereas the band at around 1107 1375 cm 1 was attributed to the sp2(C H) group. In the FTIR spectrum, the peak due to CH2deformation appears as a strong absorption band at 1455 cm 1. The strong band at 2572, 2665, 2860, 2883, and 2921 cm 1was attributed to CH stretching.19,20In addition, the adamantane-seeded on the Pt/SiO2/Si substrate was eval-uated using Raman spectroscopy (LABRAM HR800). Intense Raman peaks in the range of 600 1800 cm 1 are shown in Figure 2D. All the multiple Raman peaks at 635, 758, 949, 970, 1095, 1220, 1240, 1311, 1432, and 1472, cm 1are the typical characteristics of adamantane molecules.21We observed that the hydrocarbon signals from adamantane/Pt/SiO2/Si are partially similar in the FTIR and Raman spectra, as shown in Figure 2C,D. We also noticed that there is not any kind of bond formation between adamantane and hexane. It is clear that the hexane is evaporated and only adamantane is left on the Pt/SiO2/Si surface for diamond growth. In addition, we used D2 XRD to evaluate

the adamantane/Pt/SiO2/Si substrate. The XRD pattern in Figure 2E shows that the diﬀraction peaks at 2θ = 16.3, 18.8, 31.5, and 32.9 correspond to the interplanar spacings of adamantane {111}, {200}, {220}, {222}, and {640}, respectively. In addition to Si peaks, the diﬀraction peaks of Pt{111} and Pt{200} are seen at 39.8 and 46.3, respectively,22

indicating that Pt on the SiO2/Si substrate has survived after the solution treatment of hexane and adamantane with ultrasonication.

After adamantane seeding on Pt/SiO2/Si, we placed the sample in the MPCVD reactor for diamond growth. We used 1% CH4in H2for diamond growth. The growth time was varied from 2 to 15 min in the presence of 1% CH4in H2. Figures 3 and 4 show the SEM images after different growth times. Plan-view SEM images in Figure 3A show that there is a homogeneously distributedfilm on the Pt/SiO2/Si surface. The high-magnifica-tion SEM images in Figure 3B,C show that thefilm consists of networking islands. In general, a similar morphology has been observed in the carburization process for diamond nucleation and growth.23After 5 min of deposition, the shape of the particles are changed, as shown in Figure 3D F, into the disconnected clusters with a density of∼1011 12cm 2. After 10 min of deposi-tion, the SEM images in Figure 4A C show that most of the particles are formed in the spherical shape with an average size of less than 100 nm. The SEM image in Figure 4F shows that some oriented carbon particles in square and triangle shapes have been formed after 15 min growth. The density of the carbon particles is ∼107 8

cm 2.

Further, the cross-sectional SEM images after 2, 5, 10, and 15 min of deposition are shown in Figure 5. After 2 min of deposi-tion, the ﬁlm thickness is ∼100 nm, similar to that of the adamantane seeding ﬁlm, suggesting that most of the frac-tions of adamantane molecules may have undergone a transfor-mation into other forms of carbon species in MPCVD as no adamantane evidence can be obtained from Raman data shown later. After 5 and 10 min of deposition, the average particle height is∼115 and 130 nm, whereas it is ∼150 nm after 15 min. It is clear that the height increases with the deposition time.

In addition, we used Auger electron mapping in cross section to identify the presence of elements and their distribution on the surface after 2 min of treatment. The elemental mapping in cross section was performed in a PHI 700 scanning auger nanoprobe Figure 3. Top-view SEM images of (A C) after 2 min and (D F) after 5 min growth on adamantane/Pt/SiO2/Si.

(4)

(ULVAC-PHI Inc.). The cross-sectional specimen was prepared simply by cleavage. Figure 6A shows an SEM micrograph in cross-sectional view with the corresponding maps of C, Pt, and Si in Figure 6B D. From the image contrast in SEM and the elemental maps, one can clearly identify two regions that correspond to carbon and Si phases, while the Pt interlayer shows low contrast. There are two possibilities to show low contrast: (i) the platinum particles were not uniformly distributed on the surface and (ii) the Pt may overlap with hydrocarbons. Moreover, to understand the role of the Pt interlayer in diamond

deposition, we prepared adamantane-coated samples on SiO2/Si substrates without the Pt coating under the same experimental conditions. After 2 min of deposition, we can see that the carbon ﬁlm is not uniformly distributed on the Si surface, as shown in Figure 7A. It is likely that most of the fractions of adamantane are either etched by hydrogen plasma or evaporated. The high-magniﬁcation SEM image (Figure 7B) shows that the particles are several nanometer sizes. After 5, 10, and 15 min of deposition, particles are uniformly distributed. The sizes of the carbon particles are changed with deposition time. In comparison, the Figure 4. Top-view SEM images of (A C) after 10 min and (D F) after 15 min growth on adamantane/Pt/SiO2/Si.

(5)

density of particles became less in the absence of the Pt interlayer. Furthermore, the samples after deposition on adamantane seeded on Pt/SiO2/Si and without Pt (adamantane seeded/Si) were characterized by Raman spectroscopy, as shown in Figure 8. In both cases, similar features are observed [broad range (1250 1370 cm 1) and G band]. The broad range peak (1250 1370 cm 1) shows the presence of sp2/sp3. In the broad peak range (1250 1370 cm 1), we observed the high-intensity peak at 1325 and 1327 cm 1along with the G band in both cases (after 15 min of deposition), as shown in Figure 8A,B. Previous studies have shown that these peaks (1325 and 1327 cm 1in both cases) show the presence of a nanodiamond phase and the peak shifting due to the eﬀect of the nanodiamond size.24

On the other hand, the second-order G band does not require an elastic related scattering process and has been shown for defect-free sp2carbons.25Figure 8 shows that the intensity of peaks is increasing with deposition times. If the bands (D and G) have similar intensities, this may imply a high quantity of structural defects.25,26 Up to 10 min, the intensity of Raman signals is almost similar in both cases. However, we noticed that the one additional peak (presence of Pt) at 1339 cm 1corresponds to diamond. After 15 min (Figure 8A), a sharp Raman peak at 1598 cm 1is observed. However, we did not notice any such kind of peak in the case of without Pt (Figure 8B). Therefore, it is likely that more graphite particles or other sp2phases along with the nanodiamond (sp3) phase formed in the presence of the Pt interlayer than without Pt. The weak Raman signals at 1480, 1519, 1416, 1441, and 1716 cm 1come from diﬀerent carbon phase. Previously we reported that the Raman peak at 1480 cm 1 may come from nanodiamond. Therefore, it is possible that, in a short deposition time (2 min), some nanodiamond (sp3) may form on Pt, and further deposition results in the change of the sp3 phase to other carbon phases.

It is well known that pure hydrogen plasma can etch the hydrocarbon layer. Therefore, to evaluate the adsorption of hydrocarbon by Pt, we prepared two adamantane seeded on Pt/SiO2/Si samples that were treated with pure hydrogen gas plasma for 5 and 15 min, as shown in Figure 9. The SEM image in Figure 9A shows that the carbonfilm (after 5 min of etching) was uniformly distributed on the Pt/SiO2/Si surface. The high-magnification SEM image in Figure 9B shows the surface morphology of carbonfilm, which formed a network structure after 5 min of treatment. After 15 min of treatment, the carbon network is broken by hydrogen plasma (Figure 9D). However, when we treated adamantane/Si (without Pt) with pure hydro-gen plasma at the same experimental condition, most of thefilm disappeared and few particles were left on the Si surface. There-fore, it strongly suggests that the Pt particles/interlayer can effectively adsorb the hydrocarbons and significantly reduce the evaporation rate of adamantane (hydrocarbon) from the surface,27which then might transform into other forms of carbon phases.

In addition, XPS was used to identify the presence of chemical compositions on Pt/SiO2/Si surface after 2, 5, 10, and 15 min depositions. Figure 10A shows the survey spectrum of adaman-tane seeded on Pt/SiO2/Si treated with 1% CH4in H2for various times (2, 5, 10, and 15 min), whereas Figure 11A shows the survey spectrum of adamantane seeded on Pt/SiO2/Si treated (5 min) with pure hydrogen (200 sccm). In addition to XPS signals due to Pt particles/ﬁlm and the carbon overlayer, the XPS analysis also detected O(1s), Si(2p), and Si(2s) photoelectrons in both cases (Figures 10A and 11A). The XPS high-resolution spectra of carbon (C 1s) is shown in Figure 10B E. It is reasonable to expect that the adamantane composed into CHx species after various time treatments (2, 5, 10, and 15 min). Each of the ﬁtting components is almost similar in all cases. The Figure 6. (a) Cross-sectional SEM of deposited adamantane/Pt/Si with the corresponding nanobeam Auger electron mapping of (b) carbon, (c) platinum, and (d) silicon.

(6)

carbon (C 1s) was assigned as follows: sp3carbon with C C bonds (C C sp3 carbon, 283.7 283.8 eV), sp2 carbon with C C bonds (C C sp2carbon, 284.4 284.45 eV), sp2carbon

with H C bonds (H C sp2carbon, 284.8 eV), sp3carbon with H C bonds (H C sp3carbon, 285.2 285.3 eV), and CO at ∼287.1 and 287.7 eV.28,29

Interestingly, the carbon signals from the pure hydrogen-treated sample (after 5 min) in Figure 11B shows the doublet peaks of carbon at 284.5 and 287.6 eV. The Figure 7. Plan-view SEM images of (A, B) after 2 min, (C, D) after 5 min, (E, F) after 10 min, and (G, H) after 15 min growth on adamantane/SiO2/Si

(without Pt).

Figure 8. Raman spectra of (a) adamantane seeded Pt/SiO2/Si (with

Pt) and (b) adamantane seeded SiO2/Si (without Pt) after 2, 5, 10, and

15 min treated with 1% CH4in H2at the same experimental conditions.

Figure 9. Plan-view SEM images of (a, b) after 5 min and (c, d) after 15 min, treated with pure hydrogen plasma on adamantane/Pt/SiO2/Si.

(7)

peak is approximately at 284.5 eV, which is the typical character-istics of graphite sp2bonding.30,31The signal at 284.5 eV shows the presence of the graphite phase, while 287.6 eV is from carbonyl or C O groups.29 The XPS carbon signal at 289.1 eV also corresponds to carbonyl or C O groups.32It is most likely that carbon bonded with native layer oxygen (SiO2). Similar observations are seen in the spectra for the cases of depo-sition with CH4, as shown in Figure 10B E. Figure 10B E shows that the intensity of the carbon peak is increasing with treatment times. Therefore, the surface concentration of the carbon (sp2) phase may increase with deposition time. Never-theless, the sp3carbon might be always present on the surface in all cases. In previous studies, Han et al. and Lambrecht et al. have shown that graphitic carbon is an effective precursor for diamond nucleation.33,34Kompoulos et al. and Feng et al. have noticed diamondfilm on the thin amorphous carbon-coated Si substrate, while Barnes et al. had noticed the high density of diamond nucleation on the Si substrate, which was first scratched with diamond paste and then deposited with a non-diamond amor-phous carbonfilm.35 37

The Pt(4f) XPS spectrum shown in Figures 10F and 11C exhibited a broad doublet peak with the Pt(4f7/2) peak maximum at 70.3 eV, which negatively shifted from that for bulk Pt by∼0.9 eV. We ascribe the negative binding energy shift of the Pt(4f) doublet peak to the size eﬀect of Pt particles. The oxygen signal at 532.7 eV is associated with native oxide on the Si surface. The oxygen peak at 532.7 eV is attributed to the O Si bond in SiO2.38,39The signals of silicon Si(2p) and Si(2s) at 99 and 151 eV indicate that the substrate surface is uncovered by Pt and carbonﬁlms.40

Diamond Formation after 120 min Deposition Time.Further, diamond growth for 120 min was carried out on adamantane-seeded Pt/SiO2/Si substrates and adamantane-seeded SiO2/Si sub-strates without the Pt overcoating. The structure and surface morphology of synthesized diamond and their characteristics are shown in Figure 12. Figure 12A C shows the SEM images of the deposition on Pt/SiO2/Si, revealing that the diamonds have been synthesized uniformly on the Pt/SiO2/Si substrate in a cauliflower shape with the appearance of some {100} diamond facets. For the sample without the Pt interlayer, the synthesized diamonds in the {100} facet are clearly seen in Figure 12D. However, the density of diamond in the sample with the Pt interlayer is∼108cm 2, much higher than 106cm 2on that with-out the Pt interlayer, as shown in Figure 12A,D. Thus, it is clear that the coating of adamantane on the Pt interlayer can enhance the diamond density. The thickness of diamond on Pt/Si is∼900 nm, as shown in Figure 12C, giving the growth rate of diamond of ∼0.45 μm/h. In addition, the XRD pattern supports the formation of the microcrystalline diamond film on Pt/SiO2/Si, as shown in Figure 12E. In addition to Si peaks, the XRD pattern in Figure 12E presents sharp and well-defined peaks at 2θ = 39.8, 43.9, 44.6, and 46.39 corresponding to the interplanar spacings of Pt {111}, diamond {111}, graphite {101}, and Pt {200}, respectively. Figure 10. XPS spectra: (A) survey spectra, (B E) high-resolution

spectra of the C (1s), and (F) high-resolution spectra of the Pt (4f) region of adamantane/Pt/SiO2/Si after various treatment times.

Figure 11. (A) XPS survey spectrum of adamantane/Pt/SiO2/Si, (B)

high-resolution spectrum of the C (1s), and (C) high-resolution spectrum of the Pt (4f) region after 5 min of treatment with pure hydrogen plasma.

(8)

Moreover, the micro-Raman spectrum of the diamond film on Pt/ SiO2/Si is shown in Figure 12F. The characteristic feature of diamond at 1332 cm 1is clearly observed, the full width at half-maximum of the peak is determined to be 7.2 cm 1. The film also contains dis-ordered graphitic carbon components or amorphous carbon, as indicated by D and G bands. We also used XPS to characterize the diamond Pt/SiO2/Si, as shown in the survey spectrum of Figure 13A. Apart from the expected peaks from elements of diamond (C) and Pt, the oxygen signal has also been detected. Figure 13B shows the XPS C(1s) spectrum of the diamond sample. The deconvolution of the spectrum has shown that the C 1s peaks are composed of three peaks at∼283.7 284.4, 284.8, 285.3, and 287.1 eV corresponding to C C (sp3), C C and C H (sp2), and C O bondings.31 43 The negative binding energy (0.5 eV) shift, as shown in the high-resolution spectrum of Pt(4f) (Figure 13C), is due to the size effect of Pt particles. However, the Pt(4f) peak at 73.4 eV corresponds to Pt C bonding.44

Furthermore, the cross-sectional TEM specimen of a sample (adamantane seeded on Pt/SiO2/Si) deposited for 2 h was prepared by the focused ion beam (FEI Nova 200 Dual beam FIB). For the protection of the TEM specimen against damage from the high-energy ion beam (30 keV Ga+) in FIB, the specimen was coated with platinum. A cross-sectional bright-ﬁeld TEM image of diamond/Pt/SiO2/Si is shown in Figure 14A. From Figure 14A, the dark contrast regions are actually corre-sponding Pt particles (as shown in the following EDX) with a size of∼5 to ∼50 nm. Clearly, the Pt is not a continuous interlayer. Also, it is seen that the thickness of the diamond particle is

∼900 nm. The HRTEM image in Figure 14B is obtained from an interfacial region without the Pt particles, showing diamond and Si lattice fringes. The inset fast Fourier transform (FFT) of the lattice image of Si and diamond shows the diffraction character-istics from which the measured interplanar spacing of∼2.06 Å can be identified to be the diamond {111} one (Figure 14B). Between diamond and Si, it shows amorphous SiO2 with a thickness of∼4 nm. The SAD pattern obtained from around the marked region of the particle (Figure 14B) shows the single crystalline diamond characteristics in the Æ110æ zone axis. To further identify the Pt phase at the interface, we acquired EDX from different regions, as shown by the numbers in Figure 14A. In Figure 15, the EDX spectrum from region 1 (Figure 14A), where we can see some particles (in black contrast) at the interface, shows Pt, Si, and O peaks, whereas at region 2 on the interface (where no black particles exist), it shows Si, O, and Cu peaks. Therefore, it is clear that the Pt signal comes from black particles, while O and Si signals are from SiO2 on Si substrate. The Cu signal comes from the Cu grid. The EDX from region 3 (from Si substrate) shows the Si and O, and that from region 4 (diamond particles) shows the high concentration of carbon. It is also clear that the SiO2layer between Pt and Si prevents the silicidation. Hayashi et al have observed high crystalline (111) oriented diamond on Pt/TiO2/SiOx/Si. The oxide layers prevented the silicidation formation between Pt and Si, and it may also prevent the formation of SiC.

Figure 12. (A) Plan-view SEM, (B) high-magniﬁcation SEM image, (C) cross-sectional SEM image of diamond (growth time, 2 h) on Pt/ SiO2/Si, (D) plan-view (the inset shows a high-magniﬁcation image of

diamond on substrate without Pt), (E) XRD pattern, and (F) Raman spectra of diamond (growth time, 2 h) on Pt/SiO2/Si.

Figure 13. XPS spectra: (A) survey spectrum, (B) high-resolution spectrum of the C (1s), and (C) high-resolution spectrum of the Pt (4f) region of diamond on adamantane/Pt/SiO2/Si.

(9)

The schematic diagram of the growth of diamondﬁlm on the Pt/SiO2/Si substrate is shown Figure 16. The bonding charac-teristics depend on their structure, such as saturated hydrocar-bons, which are weakly adsorbed. Methane can desorb intact from the Pt (111) surface by heating.45Unsaturated hydrocar-bons can make covalent bonds with the Pt surface. Jacob et al. have shown adsorption of various CHxand C2Hyhydrocarbons with the Pt surface. They noticed that, in each case, the structures on the surface are comparable to gas-phase analogues in which some CH bonds are replaced with C Pt o bonds.45 The hydrogen is highly soluble in transition metals. Therefore, we predicate that, during ultrasonication, adamantane would have been adsorbed on the Pt/SiO2/Si surface. In the CVD, the hydrocarbon radicals from the plasma (1% CH4in H2plasma) with the seeded adamantane might form the Pt C H complex. Hydrogen ions (H(, H2(, H3() in H/C plasma (1% CH4/H2 mixture in mw plasma CVD reactor) are converted into hydro-carbon ions, CxHy(, in ion conversion reactions and concentra-tions [Hx(], [CxHy(]46as a result; the eﬀect of hydrogen ions seems to be negligible in comparison with CxHy(ions. From the C 1s signals in XPS spectra (Figures 10, 11, and 13), it is shown that the CO bond is present on the surface. It seems that Pt behaves as a catalyst and support to CO bond formation in early stage (2 min) deposition, as shown in Figure 10B. Therefore, CO bonding can overcome the evaporation rate of adamantane and enhance the density. In general, hexane has evaporation proper-ties. However, Pt adsorbs the hydrocarbons; therefore, a chemi-cal reaction not only has taken place with adamantane but also may have taken place with hexane (C6H14). Previous studies have shown that the hexane can react with Pt.47The melting point of Pt (∼1772 C) is much higher than the growth temperatures used in this study. The solid solubility of C in Pt is less than 1 atom % at 650 C according to the Pt C binary phase diagram.48

Therefore, it is unlikely that the dissolution medium is a pure form of Pt or a Pt C solid solution. It is rather appropriate to consider the Pt C H ternary system in order to understand the SEM observations as the present CVD growth of diamond was Figure 14. (A) Cross-sectional BF-TEM image of diamond/Pt/Si.

(B) HRTEM image of diamond/Si. The d-spacing between the fringes was∼2.06 Å of the diamond {111} plane. The insets show FFT patterns of Si and diamond lattice images (indicated by arrow). (C) SAED from marked (rectangle) region of diamond along theÆ110æ direction.

Figure 16. Schematic diagram of diamond growth on adamantane seeded on Pt/SiO2/Si substrates.

Figure 15. EDX spectra of inset mark in BF image in Figure 14A: (1) from Pt interface particles, (2) without Pt particles in the interfacial region, (3) Si substrate, and (4) from diamond particles.

(10)

plasma).

4. CONCLUSIONS

We have confirmed that diamond films are successfully grown on adamantane-seeded Pt/SiO2/Si substrates by MPCVD from the gaseous mixture of methane and hydrogen. The seeded adamantane wasfirst transformed into nanodiamond and some unidentified carbon particles in the early stage of deposition within 2 min. After further deposition of 5 15 min, the existence of the graphite phase is evident, while no adamantane and nonodiamond can be identified. With a long time deposition of 120 min, diamond formation is observed. The SiO2 layer between Pt and Si not only prevents the silicidation reaction but also avoids the formation of silicon carbide. The density of synthesized diamond on Pt/SiO2/Si was higher than that with-out Pt/SiO2/Si.

’ AUTHOR INFORMATION Corresponding Author *E-mail: rajanisht@gmail.com. ’ ACKNOWLEDGMENT

The authors would like to thank the National Science Council of the Republic of China, Taiwan, forﬁnancially supporting this research under contract NSC96-2622-E-009-002-CC3 and 98-2221-E-009-042-MY3.

’ REFERENCES

(1) Dearnaley, G.; Arps, J. H. Surf. Coat. Technol. 2005, 200, 2518–2524. (2) Yan, J. K.; Chang, L. Nanotechnology 2006, 17, 5544–5548. (3) Chen, H. G.; Chang, L. Diamond Relat. Mater. 2009, 18, 141–145.

(4) You, M. S.; Hong, F. C. N.; Jeng, Y. R.; Huang, S. M. Diamond Relat. Mater. 2009, 18, 155–159.

(5) Walkowiak, B.; Okroj, W.; Jerczynska, H.; Pawlowska, Z. Diamond Relat. Mater. 2009, 18, 651–656.

(6) Wolter, S. D.; McClure, M. T.; Glass, J. T.; Stoner, B. R. Appl. Phys. Lett. 1995, 66, 2810–2812.

(7) Tiwari, R. N.; Chen, W.-C.; Wang, W.-L.; Chang, L. J. Appl. Crystallogr. 2010, 43, 883–889.

(8) Saito, T.; Tsuruga, S.; Ohya, N.; Kusakabe, K.; Morooka, S.; Maeda, H.; Sawabe, A.; Suzuki, K. Diamond Relat. Mater. 1998, 7, 1381–1384.

(9) Tachibana, T.; Yokota, Y.; Hayashi, K.; Miyata, K.; Kobashi, K.; Shintani, Y. Diamond Relat. Mater. 2000, 9, 251–255.

(10) Tachibana, T.; Yokota, Y.; Miyata, K.; Kobashi, K.; Shintani, Y. Diamond Relat. Mater. 1997, 6, 266–271.

(11) Shintani, Y. J. Mater. Res. 1996, 11, 2955–1956.

(12) Tachibana, T.; Yokota, Y.; Miyata, K.; Onishi, T.; Kobashi, K.; Tarutani, M.; Takai, Y.; Shimizu, R.; Shintani, Y. Phys. Rev. B 1997, 56, 15967–15981.

May, P. W.; Filik, J.; Dahl, J. E.; Liu, S.; Carlson, R. M. K. J. Mol. Spectrosc. 2006, 238, 158–167.

(20) Corn, R. M.; Shannon, V. L.; Synder, R. G.; Strauss, H. L. J. Chem. Phys. 1984, 81, 5231–5238.

(21) Filik, J.; Harvey, J. N.; Allan, N. L.; May, P. W.; Dahl, J. E. P.; Liu, S.; Carlson, R. M. K. Spectrochim. Acta, Part A 2006, 64, 681–692. (22) Zheng, T.; Nishiyama, N.; Egashira, Y.; Ueyama, K. Stud. Surf. Sci. Catal. 2006, 162, 561–567.

(23) Tarakci, M.; Korkmaz, K.; Gencer, Y.; Usta, M. Surf. Coat. Technol. 2005, 199, 205–212.

(24) Sun, K. W.; Wang, C. Y. J. Phys.: Conf. Ser. 2007, 92, 012031. (25) Costa, S.; Palen, E. B.; Kruszynska, M.; Bachmatiuk, A.; Kalenczuk, R. J.Mater. Sci.-Pol. 2008, 2, 433–441.

(26) Scheibe, B.; Palen, E. B.; Kalenczuk, R. J. Mater. Charact. 2010, 61, 185–191.

(27) Kobashi, K. Diamond Films: Chemical Vapor Deposition for Oriented and Heteroepitaxial Growth; Elsevier Science: Amsterdam, 2005.

(28) Takabayashi, S.; Okamoto, K.; Shimada, K.; Motomitsu, K.; Motoyama, H.; Nakatani, T.; Sakaue, H.; Suzuki, H.; Takahagi, T. Jpn. J. Appl. Phys. 2008, 47, 3376–3379.

(29) Dagli, G.; Sung, N.-H. Polym. Compos. 1989, 10, 109–116. (30) Smirnova, T. P.; Badalyan, A. M.; Borisov, V. O.; Yakovkina, L. V.; Kaichev, V. V.; Shmakov, A. N.; Nartova, A. V.; Rakhlin, V. I.; Fomina, A. N. High Energy Chem. 2003, 37, 303–309.

(31) Kvon, R. I.; Boronin, A. I.; Shaikhutdinov, S. K.; Buyanov, R. A. Appl. Surf. Sci. 1997, 120, 239–242.

(32) Delamar, M.; Dbsarmot, G.; Fagebaume, O.; Hitmi, R.; Pinson, J. Carbon 1997, 35, 801–807.

(33) Han, Y. X.; Ling, H.; Sun, J.; Zhao, M.; Gebre, T.; Lu, Y. F. Appl. Surf. Sci. 2008, 254, 2054–2058.

(34) Lambrecht, W. R. L; Lee, C. H.; Segall, B.; Angus, J. C.; Li, Z.; Sunkara, M. Nature 1993, 364, 607–610.

(35) Komvopoulos, K.; Xu, T. Diamond Relat. Mater. 2000, 9, 274–282.

(36) Feng, Z.; Komvopoulos, K.; Bogy, D. B.; Ager, J. W.; Anders, S.; Anders, A.; Wang, Z.; Brown, I. G. J. Appl. Phys. 1996, 79, 485.

(37) Barnes, P. N.; Wu, R. L. C. Appl. Phys. Lett. 1993,62, 37–39. (38) Schaller, E. M.; Kuettel, O. M.; Schlapbach, L. Phys. Status Solidi A 1996, 153, 415–429.

(39) Gsell, S.; Berner, S.; Brugger, T.; Schreck, M.; Brescia, R.; Fischer, M.; Greber, T.; Osterwalder, J.; Stritzker, B. Diamond Relat. Mater. 2008, 17, 1029–1034.

(40) Boukherroub, R.; Wayner, D. D. M.; Sproule, G. I.; Lockwood, D. J.; Canham, L. T. Nano Lett. 2001, 1, 713–717.

(41) Saw, K. G.; du Plessis, J. Mater. Lett. 2004, 58, 1344–1348. (42) Ong, S.-E.; Zhang, S.; Du, H.; Sun, D. Diamond Relat. Mater. 2007, 16, 1628–1635.

(43) Ahmed, M.; Byrne, J. A.; McLaughlin, J. A. D. e-J. Surf. Sci. Nanotechnol. 2009, 7, 217–224.

(44) Parravano, G. J. Catal. 1970, 16, 1–15.

(45) Jacob, T.; Goddard, W. A., III J. Phys. Chem. B 2005, 109, 297–311.

(46) Mankelevich, Y. A.; Ashfold, M. N. R.; Ma, J. J. Appl. Phys. 2008, 104, 113304.

(11)

The Journal of Physical Chemistry C _ARTICLE (48) Massalski, T. B.; Okamoto, H.; Subramanian, P. R.; Kacprzak,

L., Eds.; Binary Alloy Phase Diagrams; ASM International: Materials Park, OH, 1990; p 794.

(49) Roy, R.; Dewan, H. S.; Ravindranathan, P. Mater. Res. Bull. 1993, 28, 861–866.