- Email: [email protected]
Diamond nucleation and growth on zeolites
Diamond and Related Materials 12 (2003) 1647–1652
Diamond nucleation and growth on zeolites E. Titusa, M.K. Singha, K.N.N. Unnia, P.K. Tyagia, A.K. Duab, Mainak Royb, D.S. Misraa,* a Department of Physics, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India Novel Materials and Structural Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India
b
Abstract In this work, we report the use of zeolites as substrates for the deposition of porous diamond films. Films were deposited in a hot-filament chemical vapor deposition (HFCVD) apparatus. The HFCVD system was fed with a mixture of methane (0.8%) with the balance being hydrogen. A series of depositions were done in the pressure range 20–120 Torr and at substrate temperature 880 8C. The morphologies of the as-deposited films were analyzed by scanning electron microscopy and show isolated diamond grains in the initial nucleation stages, which develop into a microporous film in the next stage and form a continuous film after long time deposition. Raman spectroscopy was used to investigate the crystal morphology, structure and non-diamond impurities in the films deposited at various growth conditions. The nature of the hydrogen bonding with sp3 and sp2 network and the quantitative analysis were done by Fourier transform infrared spectroscopy. 䊚 2003 Elsevier B.V. All rights reserved. Keywords: Hot-filament chemical vapor deposition; Scanning electron microscopy; Fourier transform infrared spectroscopy
1. Introduction Diamond has many attractive properties and therefore its applications are unlimited w1–4x. Chemical vapor deposition (CVD) of diamond has the advantage of producing diamond of tuned thickness and geometry, at low pressure and moderate temperature on substrates of different geometries. Today intense research activity is directed towards the production of dense uniform diamond films useful for many different applications in optical, mechanical, thermal and electronic domains. However, for many applications such as micro-filters w5x, field-emission components w6x, photonic band gap devices w7x, fuel-cell electrodes w8x, etc., it is of interest to obtain porous diamond films, or diamond membranes, with large effective areas and a specified porosity. Out of the above applications, micro-filters made of CVD diamond will be extremely useful. This is mainly because of the fact that diamond is chemically inert and thermally stable w9x. One important qualification for a material suitable for micro-filter would be the uniform size and the distribution of the pores. However, getting uniform size and the distribution of grains on silicon *Corresponding author. Tel.: q91-22-576-7561; fax: q91-22-5783480. E-mail address: [email protected] (D.S. Misra).
substrate is difficult mainly because the substrate has to be nucleated with diamond powder. This generally results in damage of the substrate surface and the nonuniform sizes of the nuclei. The search for a substrate, other than diamond itself, onto which diamond can be grown without any treatment, has attracted considerable attention. However, not much success has been achieved in this direction. A substrate onto which diamond nucleation is not required may enable us to develop the films with uniform grain sizes and also pave the way for depositing single crystal films on non-diamond substrates. A few reports have also indicated that porous diamond samples are of special interest for field emission due to its extreme hardness, high thermal conductivity and high resistance to corrosion. The field emission properties of porous diamond were reported earlier w10x. The photoemission electron microscopy images from the emitting area around a pore showed that there is indeed enhanced electron emission occurring in the pore regions. The studies of other groups w11–13x showed that the residual gas inside the field emission displays is perhaps the most important issue related to the device lifetime. Increasing the display area while maintaining the display thickness unchanged results in the decrease of the lifetime, since the pressure gradient is enhanced. Mam-
0925-9635/03/$ - see front matter 䊚 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0925-9635(03)00307-8
1648
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
mana et al. w14x demonstrated that the porous diamond membranes show good performance in vacuum, with low emitter switching voltage. Their experimental results showed a correlation between the emitted current and the number of pores and suggest that the emission sites are located at the pore edges in both polyamide and diamond membranes. Here, we report the first work of deposition of porous diamond films on zeolite pellets. We find that the nucleation of diamond takes place directly on zeolite pellets without any treatment with the diamond powder. This is a significant development, in our opinion, because it implies that the pre-generated nucleation sites are not essential for diamond deposition. The zeolite we have selected is a high temperature material with decomposition temperature in excess of 950 8C. The films are deposited at different pressure and temperature to understand the growth process. The morphology of the films indicates the granular growth of the films dominated by (1 1 1) oriented grains. Raman spectroscopy reveals the presence of 1332 cmy1 peak accompanied with a graphite band at 1560 cmy1. We conclude from the morphological studies that the zeolite may be an important substrate for depositing films with uniform grain size. 2. Experimental procedure
Fig. 1. SEM micrographs of the diamond films deposited for 3 h on (a) p-type Si (1 0 0) substrates treated with 2-mm diamond powder (sample I); (b) zeolite substrate (sample II) and (c) zeolite substrate (sample III). The deposition pressures for the samples (a and b) was 20 Torr and for (c) was 120 Torr, respectively.
Zeolites used as substrates in the present study are alumino silicates represented by the chemical formula M2ynØAl2O3ØySiO2ØwH2O where y is 20 (y is the SiyAl ratio in the structure). M is the charge balancing cation, such as sodium, potassium, barium, magnesium, calcium, etc., which balances the charge on the zeolite structure, n is the cation valency and w represents the moles of water contained in the zeolite voids. The zeolite powder of particle size 10 mm and pore size 5.8 ˚ was pressed in the form of a pellet of thickness 1 A mm. This system consists of a three-dimensional channel. The straight channels along the b-axis are defined by 10-member oxygen rings with an elliptical crosssection with major and minor axes, respectively, of 5.7– ˚ and 5.1–5.2 A. ˚ These channels are interconnected 5.8 A by sinusoidal channels along the a-axis defined 10member oxygen rings with nearly circular cross-section ˚ of 5.4"0.2 A. Depositions were carried out in hot-filament chemical vapor deposition system whose details are discussed elsewhere w15x. The gas precursors for the depositions were CH4 and H2 at the flow rates of 1.6 and 200 sccm, respectively. These flow rates correspond to 0.8% of CH4 and 99.2% of H2. The deposition pressure was varied between 20 and 120 Torr and the temperature was 880 8C. Polished p-type silicon (1 0 0) substrates were also used for the comparison. The silicon substrates were treated with 2-mm diamond powder and cleaned ultrasonically in an acetone bath prior to the deposition.
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
1649
Fig. 2. SEM micrographs of (a) continuous diamond films on silicon substrate deposited at 20 Torr (sample IV); (b) microporous diamond films on zeolite deposited at 20 Torr (sample V); (c) microporous diamond film on zeolite deposited at 120 Torr (sample VI) and (d) a continuous film (sample VII) deposited on zeolite substrate at 20 Torr after 24-h deposition.
As mentioned earlier, no treatment of zeolite pellet is necessary for the deposition of diamond films. The morphologies of the films were studied using scanning electron microscopy (SEM). For infrared (IR) analyses the zeolite was dissolved completely by dipping the film in concentrated HF solution. The analysis was done in the range 400–4000 cmy1 with a resolution of 2 cmy1. Laser Raman spectra were recorded in the range 1200–1700 cmy1 with step sizes of 2 cmy1 to estimate the non-diamond content in the samples. An Arq laser of 50 mW was used for recording the spectra. 3. Results and discussion Fig. 1a–c show the SEM micrographs of the diamond deposited for 3 h on (a) p-type Si (1 0 0) substrates, treated with 2-mm diamond powder (sample I), (b) zeolite substrate (sample II) and (c) zeolite substrate
(sample III), respectively. The deposition pressure was 20 Torr for samples (a and b) and 120 Torr for (c), respectively. The films show isolated grains in all the three cases. An interesting feature of the grains on the zeolite substrates is the uniformity in the structure and sizes. All diamond crystals are of approximately 25-mm size and have perfect uniform shape. Majority of crystal facets deposited on zeolites seem to have (1 1 1) orientation. In contrast, the diamond films deposited on silicon substrates produce crystals of hugely varying shapes and sizes. The contrasting results on the two substrates are due to the fact that zeolites are untreated and silicon substrates are treated with diamond powder. Treatment with diamond powder generates nuclei of all shapes and sizes on silicon. On the other hand, there is a perfect uniformity of shape and size on zeolite substrate because every nuclei start on an identical site. This symmetry is lost on silicon due to the treatment.
1650
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
Fig. 4. XRD pattern of sample V showing a dominant diamond N1 1 1M peak. Other XRD peaks correspond to the zeolite substrate.
Fig. 2a–c show respective SEM micrographs of (a) continuous diamond films on silicon substrate deposited at 20 Torr (sample IV), (b) microporous diamond films on zeolite deposited at 20 Torr (sample V) and (c) microporous diamond film on zeolite deposited at 120 Torr (sample VI). The deposition time for the samples in Fig. 2a–c was for 12 h. A continuous film (sample VII) formed on zeolite substrate at 20 Torr only after 24-h deposition (Fig. 2d). For comparison, we deposited diamond films on untreated silicon substrates also. The untreated silicon substrates showed no deposition, while the treated silicon substrate produced diamond films with dominantly (1 1 1) oriented grains as evident by X-ray diffraction (XRD) analyses w16x. The non-diamond content in the diamond films was analyzed using Raman spectroscopy technique. The spectra recorded in the full range of samples IV, V and VI are shown in Fig. 3a–c. A sharp Raman line is observed at 1332.5 cmy1 in the samples IV and V. A broad band corresponding to the graphite w17x also appears at approximately 1560 cmy1 for sample V. It is
Fig. 3. Raman spectra of diamond films on samples (a) IV (deposited on silicon substrate at 20 Torr); (b) V (deposited on zeolite pellet at 20 Torr and (c) VI (deposited on zeolite pellet at 120 Torr).
Fig. 5. FTIR spectra of samples IV (deposited on silicon substrate at 20 Torr) and VII (zeolite pellet at 20 Torr) in the range 2700–3100 cmy1.
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
known that Raman scattering coefficient is significantly higher for graphite than that for diamond w18x. Thus, even very small concentration of sp2 phase could be easily detected using Raman spectroscopy. Therefore, the samples IV and V in which the graphite signal is weak could be considered good quality crystalline diamonds w19x. Sample VI shows only the band (;1580 cmy1) corresponding to the graphite. The line width of the diamond Raman line is high in sample V. The FWHM values of the diamond line of sample V (;10.5 cmy1) is higher than sample IV (7 cmy1). This signifies that the films deposited on zeolites may have more defects. Or else it may also be related to the stress in the films due to the presence of the impurities. XRD pattern of the films (Fig. 4) shows a dominant N1 1 1M peak corresponding to cubic diamond phase along with the other peaks corresponding to the zeolite structure w20x. Fig. 5 shows the typical Fourier transform infrared (FTIR) spectra of the samples IV and VII in the range 2700 and 3100 cmy1. As discussed earlier, sample IV is deposited on silicon substrate, whereas sample VII is deposited on zeolite at 20 Torr. The spectra in this region are due to the superposition of the CH vibrations from spm CHn, where m, ns1, 2 and 3. The experimental spectrum was fitted using Gaussian peaks after background correction. The peak positions were fixed initially according to the values quoted in Ref. w21x. The half width and amplitude were taken as fitting parameters. The frequency of the vibration is shifted slightly for C–H bonds in different local environments, allowing us to differentiate between CHn groups. Since each individual spm CHn configuration is characterized by a specific IR absorption peak, one can use these spectral peaks to analyze the relative hybridization of the carbon atoms. The spectra of the sample IV contain a broad band containing various peaks corresponding to multiple CHn modes. In contrast, the spectra of sample VII contain two dominant peaks at 2860 and 2930 cmy1 that correspond to symmetric and asymmetric stretch bands of CH2 group. A weak band corresponding to hydrogen bonded to sp2 carbon (3025 cmy1) is also visible. The SEM result shows that the crystals are very uniformly arranged on zeolite substrate and it is interesting to note that in FTIR spectra, the sp3 CH2 peaks of this sample are very prominent. FTIR analysis of our previous paper w22x also shows dominant sp3 CH2 peaks for uniform (1 0 0) textured diamond films unlike in ordinary films. The integrated absorbance of each band also can be used to estimate the hydrogen concentration in a particular mode. As discussed by MacNamara et al. w23x, the concentration of the oscillating species is proportional to the integrated intensity of the absorption band. The total hydrogen content is
| a(v)dv v
NHsAns
1651
(1)
where An is the proportionality factor and a(v) is the absorption coefficient at frequency v. An is proportional to the inverse of oscillator strength. We have chosen An to be 1=1020 and assumed it to be constant for all the modes and samples w24x. a(v), the frequencydependent absorption coefficient, is taken to be a Gaussian function. The concentration of the bonded hydrogen determined according to Eq. (1) is significantly lower (0.076 at.%) in sample IV compared to that (0.092 at.%) in sample VII. This correlates well with the Raman results that show that the samples deposited on zeolite pellet contain higher concentration of graphitic impurities. 4. Conclusion We report here the nucleation and deposition of diamond films on zeolite pellets. On the zeolite pellets, diamond films grow without any treatment. The uniform pore size and crystalline nature of the zeolite promotes the growth of diamond grains of uniform size and welldefined orientation. The diamond crystals on zeolite are typically of 25–30 mm size and have a uniform shape all over the substrate indicating that all the nucleation sites are equally preferable. This is in contrast to the deposition on silicon substrates. The size and orientation of the crystals on silicon depend upon the residual nuclei left behind by the diamond powder. The results reported here might have interesting consequences for the mechanism of the growth of the diamond crystals. References w1x W. Yarbrough, R. Messier, Science 247 (1990) 688. w2x R.F. Davis, Diamond Films and Coatings, Noyes Publications, New Jersey, 1992, p. 1992. w3x L.S. Pan, D.R. Kania, Diamond: Electronic Properties and Applications, Kluwer Academic Publishers, Boston, 1995. w4x A. Gicquel, K. Hassouni, F. Silva, J. Achard, Curr. Appl. Phys. 1 (2001) 479. w5x V. Baranauskas, A.C. Peterlevitz, D.C. Chang, S.F. Durrant, Appl. Surf. Sci. 185 (2001) 108. w6x W. Zhu, C. Bower, G.P. Kochanski, S. Jin, Diamond Relat. Mater. 10 (2001) 1709. w7x V.P. Mammana, R.D. Manasano, P. Verdonck, A. Pavani Filho, M.C. Salvadori, Diamond Relat. Mater. 6 (1997) 1824. w8x K. Honda, M. Yoshimura, T.N. Rao, et al., J. Electroanal. Chem. 514 (2001) 35. w9x H.-K. Chung, J.C. Sung, Mater. Chem. Phys. 72 (2001) 130. w10x V. Zhirnov, G.J. Wojak, W.B. Choi, J. Jcuomo, J.J. Hren, J. Vac. Sci. Technol. A 15 (1997) 1733. w11x B.R. Chalamala, R.H. Reuss, K.A. Dean, Appl. Phys. Lett. 79 (2001) 2648. w12x R.H. Reuss, B.R. Chalamala, J. Vac. Sci. Technol. B 19 (2001) 537. w13x A. Zoulkarneev, N.S. Park, J.E. Jung, J.W. Kim, J.P. Hong, J.M. Kim, J. Vac. Sci. Technol. B 16 (1998) 741.
1652
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
w14x V.P. Mammana, L.R.C. Fonseca, A.P. Filho, O.R. Monteiro, R. Ramprasad, P. Allmen, J. Vac. Sci. Technol. B 19 (2001) 537. w15x T. Sharda, D.S. Misra, D.K. Avasthi, Vacuum 47 (1996) 1259. w16x A.K. Sikder, Thesis Submitted to the Indian Institute of Technology, Bombay, India, unpublished. w17x J. Wagner, M. Ramsteiner, C. Wild, P. Koidl, Phys. Rev. B 40 (1989) 1817. w18x N. Wada, S.A. Solin, Physica B 105 (1981) 353. w19x D.S. Knight, W.B. White, J. Mater. Res. 4 (1989) 385.
w20x E. Piera, M.A. Solomon, J. Coronas, M. Menendez, J. Santamaria, J. Membrane Sci. 149 (1998) 99. w21x C. Dischler, W. Wild, M. Serbet, P. Koidl, Physica B 185 (1993) 217. w22x E. Titus, A.K. Sikder, M.K. Singh, D.S. Misra, Diamond Relat. Mater. 11 (2002) 1403. w23x K.M. MacNamara, D.H. Levy, K.K. Gleason, C.J. Robinson, Appl. Phys. Lett. 60 (1992) 580. w24x W. Jacob, M. Unger, Appl. Phys. Lett. 68 (1996) 475.
Diamond nucleation and growth on zeolites E. Titusa, M.K. Singha, K.N.N. Unnia, P.K. Tyagia, A.K. Duab, Mainak Royb, D.S. Misraa,* a Department of Physics, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India Novel Materials and Structural Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India
b
Abstract In this work, we report the use of zeolites as substrates for the deposition of porous diamond films. Films were deposited in a hot-filament chemical vapor deposition (HFCVD) apparatus. The HFCVD system was fed with a mixture of methane (0.8%) with the balance being hydrogen. A series of depositions were done in the pressure range 20–120 Torr and at substrate temperature 880 8C. The morphologies of the as-deposited films were analyzed by scanning electron microscopy and show isolated diamond grains in the initial nucleation stages, which develop into a microporous film in the next stage and form a continuous film after long time deposition. Raman spectroscopy was used to investigate the crystal morphology, structure and non-diamond impurities in the films deposited at various growth conditions. The nature of the hydrogen bonding with sp3 and sp2 network and the quantitative analysis were done by Fourier transform infrared spectroscopy. 䊚 2003 Elsevier B.V. All rights reserved. Keywords: Hot-filament chemical vapor deposition; Scanning electron microscopy; Fourier transform infrared spectroscopy
1. Introduction Diamond has many attractive properties and therefore its applications are unlimited w1–4x. Chemical vapor deposition (CVD) of diamond has the advantage of producing diamond of tuned thickness and geometry, at low pressure and moderate temperature on substrates of different geometries. Today intense research activity is directed towards the production of dense uniform diamond films useful for many different applications in optical, mechanical, thermal and electronic domains. However, for many applications such as micro-filters w5x, field-emission components w6x, photonic band gap devices w7x, fuel-cell electrodes w8x, etc., it is of interest to obtain porous diamond films, or diamond membranes, with large effective areas and a specified porosity. Out of the above applications, micro-filters made of CVD diamond will be extremely useful. This is mainly because of the fact that diamond is chemically inert and thermally stable w9x. One important qualification for a material suitable for micro-filter would be the uniform size and the distribution of the pores. However, getting uniform size and the distribution of grains on silicon *Corresponding author. Tel.: q91-22-576-7561; fax: q91-22-5783480. E-mail address: [email protected] (D.S. Misra).
substrate is difficult mainly because the substrate has to be nucleated with diamond powder. This generally results in damage of the substrate surface and the nonuniform sizes of the nuclei. The search for a substrate, other than diamond itself, onto which diamond can be grown without any treatment, has attracted considerable attention. However, not much success has been achieved in this direction. A substrate onto which diamond nucleation is not required may enable us to develop the films with uniform grain sizes and also pave the way for depositing single crystal films on non-diamond substrates. A few reports have also indicated that porous diamond samples are of special interest for field emission due to its extreme hardness, high thermal conductivity and high resistance to corrosion. The field emission properties of porous diamond were reported earlier w10x. The photoemission electron microscopy images from the emitting area around a pore showed that there is indeed enhanced electron emission occurring in the pore regions. The studies of other groups w11–13x showed that the residual gas inside the field emission displays is perhaps the most important issue related to the device lifetime. Increasing the display area while maintaining the display thickness unchanged results in the decrease of the lifetime, since the pressure gradient is enhanced. Mam-
0925-9635/03/$ - see front matter 䊚 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0925-9635(03)00307-8
1648
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
mana et al. w14x demonstrated that the porous diamond membranes show good performance in vacuum, with low emitter switching voltage. Their experimental results showed a correlation between the emitted current and the number of pores and suggest that the emission sites are located at the pore edges in both polyamide and diamond membranes. Here, we report the first work of deposition of porous diamond films on zeolite pellets. We find that the nucleation of diamond takes place directly on zeolite pellets without any treatment with the diamond powder. This is a significant development, in our opinion, because it implies that the pre-generated nucleation sites are not essential for diamond deposition. The zeolite we have selected is a high temperature material with decomposition temperature in excess of 950 8C. The films are deposited at different pressure and temperature to understand the growth process. The morphology of the films indicates the granular growth of the films dominated by (1 1 1) oriented grains. Raman spectroscopy reveals the presence of 1332 cmy1 peak accompanied with a graphite band at 1560 cmy1. We conclude from the morphological studies that the zeolite may be an important substrate for depositing films with uniform grain size. 2. Experimental procedure
Fig. 1. SEM micrographs of the diamond films deposited for 3 h on (a) p-type Si (1 0 0) substrates treated with 2-mm diamond powder (sample I); (b) zeolite substrate (sample II) and (c) zeolite substrate (sample III). The deposition pressures for the samples (a and b) was 20 Torr and for (c) was 120 Torr, respectively.
Zeolites used as substrates in the present study are alumino silicates represented by the chemical formula M2ynØAl2O3ØySiO2ØwH2O where y is 20 (y is the SiyAl ratio in the structure). M is the charge balancing cation, such as sodium, potassium, barium, magnesium, calcium, etc., which balances the charge on the zeolite structure, n is the cation valency and w represents the moles of water contained in the zeolite voids. The zeolite powder of particle size 10 mm and pore size 5.8 ˚ was pressed in the form of a pellet of thickness 1 A mm. This system consists of a three-dimensional channel. The straight channels along the b-axis are defined by 10-member oxygen rings with an elliptical crosssection with major and minor axes, respectively, of 5.7– ˚ and 5.1–5.2 A. ˚ These channels are interconnected 5.8 A by sinusoidal channels along the a-axis defined 10member oxygen rings with nearly circular cross-section ˚ of 5.4"0.2 A. Depositions were carried out in hot-filament chemical vapor deposition system whose details are discussed elsewhere w15x. The gas precursors for the depositions were CH4 and H2 at the flow rates of 1.6 and 200 sccm, respectively. These flow rates correspond to 0.8% of CH4 and 99.2% of H2. The deposition pressure was varied between 20 and 120 Torr and the temperature was 880 8C. Polished p-type silicon (1 0 0) substrates were also used for the comparison. The silicon substrates were treated with 2-mm diamond powder and cleaned ultrasonically in an acetone bath prior to the deposition.
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
1649
Fig. 2. SEM micrographs of (a) continuous diamond films on silicon substrate deposited at 20 Torr (sample IV); (b) microporous diamond films on zeolite deposited at 20 Torr (sample V); (c) microporous diamond film on zeolite deposited at 120 Torr (sample VI) and (d) a continuous film (sample VII) deposited on zeolite substrate at 20 Torr after 24-h deposition.
As mentioned earlier, no treatment of zeolite pellet is necessary for the deposition of diamond films. The morphologies of the films were studied using scanning electron microscopy (SEM). For infrared (IR) analyses the zeolite was dissolved completely by dipping the film in concentrated HF solution. The analysis was done in the range 400–4000 cmy1 with a resolution of 2 cmy1. Laser Raman spectra were recorded in the range 1200–1700 cmy1 with step sizes of 2 cmy1 to estimate the non-diamond content in the samples. An Arq laser of 50 mW was used for recording the spectra. 3. Results and discussion Fig. 1a–c show the SEM micrographs of the diamond deposited for 3 h on (a) p-type Si (1 0 0) substrates, treated with 2-mm diamond powder (sample I), (b) zeolite substrate (sample II) and (c) zeolite substrate
(sample III), respectively. The deposition pressure was 20 Torr for samples (a and b) and 120 Torr for (c), respectively. The films show isolated grains in all the three cases. An interesting feature of the grains on the zeolite substrates is the uniformity in the structure and sizes. All diamond crystals are of approximately 25-mm size and have perfect uniform shape. Majority of crystal facets deposited on zeolites seem to have (1 1 1) orientation. In contrast, the diamond films deposited on silicon substrates produce crystals of hugely varying shapes and sizes. The contrasting results on the two substrates are due to the fact that zeolites are untreated and silicon substrates are treated with diamond powder. Treatment with diamond powder generates nuclei of all shapes and sizes on silicon. On the other hand, there is a perfect uniformity of shape and size on zeolite substrate because every nuclei start on an identical site. This symmetry is lost on silicon due to the treatment.
1650
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
Fig. 4. XRD pattern of sample V showing a dominant diamond N1 1 1M peak. Other XRD peaks correspond to the zeolite substrate.
Fig. 2a–c show respective SEM micrographs of (a) continuous diamond films on silicon substrate deposited at 20 Torr (sample IV), (b) microporous diamond films on zeolite deposited at 20 Torr (sample V) and (c) microporous diamond film on zeolite deposited at 120 Torr (sample VI). The deposition time for the samples in Fig. 2a–c was for 12 h. A continuous film (sample VII) formed on zeolite substrate at 20 Torr only after 24-h deposition (Fig. 2d). For comparison, we deposited diamond films on untreated silicon substrates also. The untreated silicon substrates showed no deposition, while the treated silicon substrate produced diamond films with dominantly (1 1 1) oriented grains as evident by X-ray diffraction (XRD) analyses w16x. The non-diamond content in the diamond films was analyzed using Raman spectroscopy technique. The spectra recorded in the full range of samples IV, V and VI are shown in Fig. 3a–c. A sharp Raman line is observed at 1332.5 cmy1 in the samples IV and V. A broad band corresponding to the graphite w17x also appears at approximately 1560 cmy1 for sample V. It is
Fig. 3. Raman spectra of diamond films on samples (a) IV (deposited on silicon substrate at 20 Torr); (b) V (deposited on zeolite pellet at 20 Torr and (c) VI (deposited on zeolite pellet at 120 Torr).
Fig. 5. FTIR spectra of samples IV (deposited on silicon substrate at 20 Torr) and VII (zeolite pellet at 20 Torr) in the range 2700–3100 cmy1.
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
known that Raman scattering coefficient is significantly higher for graphite than that for diamond w18x. Thus, even very small concentration of sp2 phase could be easily detected using Raman spectroscopy. Therefore, the samples IV and V in which the graphite signal is weak could be considered good quality crystalline diamonds w19x. Sample VI shows only the band (;1580 cmy1) corresponding to the graphite. The line width of the diamond Raman line is high in sample V. The FWHM values of the diamond line of sample V (;10.5 cmy1) is higher than sample IV (7 cmy1). This signifies that the films deposited on zeolites may have more defects. Or else it may also be related to the stress in the films due to the presence of the impurities. XRD pattern of the films (Fig. 4) shows a dominant N1 1 1M peak corresponding to cubic diamond phase along with the other peaks corresponding to the zeolite structure w20x. Fig. 5 shows the typical Fourier transform infrared (FTIR) spectra of the samples IV and VII in the range 2700 and 3100 cmy1. As discussed earlier, sample IV is deposited on silicon substrate, whereas sample VII is deposited on zeolite at 20 Torr. The spectra in this region are due to the superposition of the CH vibrations from spm CHn, where m, ns1, 2 and 3. The experimental spectrum was fitted using Gaussian peaks after background correction. The peak positions were fixed initially according to the values quoted in Ref. w21x. The half width and amplitude were taken as fitting parameters. The frequency of the vibration is shifted slightly for C–H bonds in different local environments, allowing us to differentiate between CHn groups. Since each individual spm CHn configuration is characterized by a specific IR absorption peak, one can use these spectral peaks to analyze the relative hybridization of the carbon atoms. The spectra of the sample IV contain a broad band containing various peaks corresponding to multiple CHn modes. In contrast, the spectra of sample VII contain two dominant peaks at 2860 and 2930 cmy1 that correspond to symmetric and asymmetric stretch bands of CH2 group. A weak band corresponding to hydrogen bonded to sp2 carbon (3025 cmy1) is also visible. The SEM result shows that the crystals are very uniformly arranged on zeolite substrate and it is interesting to note that in FTIR spectra, the sp3 CH2 peaks of this sample are very prominent. FTIR analysis of our previous paper w22x also shows dominant sp3 CH2 peaks for uniform (1 0 0) textured diamond films unlike in ordinary films. The integrated absorbance of each band also can be used to estimate the hydrogen concentration in a particular mode. As discussed by MacNamara et al. w23x, the concentration of the oscillating species is proportional to the integrated intensity of the absorption band. The total hydrogen content is
| a(v)dv v
NHsAns
1651
(1)
where An is the proportionality factor and a(v) is the absorption coefficient at frequency v. An is proportional to the inverse of oscillator strength. We have chosen An to be 1=1020 and assumed it to be constant for all the modes and samples w24x. a(v), the frequencydependent absorption coefficient, is taken to be a Gaussian function. The concentration of the bonded hydrogen determined according to Eq. (1) is significantly lower (0.076 at.%) in sample IV compared to that (0.092 at.%) in sample VII. This correlates well with the Raman results that show that the samples deposited on zeolite pellet contain higher concentration of graphitic impurities. 4. Conclusion We report here the nucleation and deposition of diamond films on zeolite pellets. On the zeolite pellets, diamond films grow without any treatment. The uniform pore size and crystalline nature of the zeolite promotes the growth of diamond grains of uniform size and welldefined orientation. The diamond crystals on zeolite are typically of 25–30 mm size and have a uniform shape all over the substrate indicating that all the nucleation sites are equally preferable. This is in contrast to the deposition on silicon substrates. The size and orientation of the crystals on silicon depend upon the residual nuclei left behind by the diamond powder. The results reported here might have interesting consequences for the mechanism of the growth of the diamond crystals. References w1x W. Yarbrough, R. Messier, Science 247 (1990) 688. w2x R.F. Davis, Diamond Films and Coatings, Noyes Publications, New Jersey, 1992, p. 1992. w3x L.S. Pan, D.R. Kania, Diamond: Electronic Properties and Applications, Kluwer Academic Publishers, Boston, 1995. w4x A. Gicquel, K. Hassouni, F. Silva, J. Achard, Curr. Appl. Phys. 1 (2001) 479. w5x V. Baranauskas, A.C. Peterlevitz, D.C. Chang, S.F. Durrant, Appl. Surf. Sci. 185 (2001) 108. w6x W. Zhu, C. Bower, G.P. Kochanski, S. Jin, Diamond Relat. Mater. 10 (2001) 1709. w7x V.P. Mammana, R.D. Manasano, P. Verdonck, A. Pavani Filho, M.C. Salvadori, Diamond Relat. Mater. 6 (1997) 1824. w8x K. Honda, M. Yoshimura, T.N. Rao, et al., J. Electroanal. Chem. 514 (2001) 35. w9x H.-K. Chung, J.C. Sung, Mater. Chem. Phys. 72 (2001) 130. w10x V. Zhirnov, G.J. Wojak, W.B. Choi, J. Jcuomo, J.J. Hren, J. Vac. Sci. Technol. A 15 (1997) 1733. w11x B.R. Chalamala, R.H. Reuss, K.A. Dean, Appl. Phys. Lett. 79 (2001) 2648. w12x R.H. Reuss, B.R. Chalamala, J. Vac. Sci. Technol. B 19 (2001) 537. w13x A. Zoulkarneev, N.S. Park, J.E. Jung, J.W. Kim, J.P. Hong, J.M. Kim, J. Vac. Sci. Technol. B 16 (1998) 741.
1652
E. Titus et al. / Diamond and Related Materials 12 (2003) 1647–1652
w14x V.P. Mammana, L.R.C. Fonseca, A.P. Filho, O.R. Monteiro, R. Ramprasad, P. Allmen, J. Vac. Sci. Technol. B 19 (2001) 537. w15x T. Sharda, D.S. Misra, D.K. Avasthi, Vacuum 47 (1996) 1259. w16x A.K. Sikder, Thesis Submitted to the Indian Institute of Technology, Bombay, India, unpublished. w17x J. Wagner, M. Ramsteiner, C. Wild, P. Koidl, Phys. Rev. B 40 (1989) 1817. w18x N. Wada, S.A. Solin, Physica B 105 (1981) 353. w19x D.S. Knight, W.B. White, J. Mater. Res. 4 (1989) 385.
w20x E. Piera, M.A. Solomon, J. Coronas, M. Menendez, J. Santamaria, J. Membrane Sci. 149 (1998) 99. w21x C. Dischler, W. Wild, M. Serbet, P. Koidl, Physica B 185 (1993) 217. w22x E. Titus, A.K. Sikder, M.K. Singh, D.S. Misra, Diamond Relat. Mater. 11 (2002) 1403. w23x K.M. MacNamara, D.H. Levy, K.K. Gleason, C.J. Robinson, Appl. Phys. Lett. 60 (1992) 580. w24x W. Jacob, M. Unger, Appl. Phys. Lett. 68 (1996) 475.