Fullerene-silicon polymerization evidence

We report experimental results for C60-Si deposition by simultaneous thermal vaporization of fullerene source and chemical vapor deposition from silane source. The samples were characterized by Scanning Electron Microscopy, Energy-dispersive X-ray spectroscopy, Micro-Raman spectroscopy, Wide-angle X-ray scattering, X-ray photoelectron spectroscopy and its thermal stability was studied, discussed and compared with pure C60 deposited by the same method. A crystalline material was obtained and results suggest that a polymerization of fullerenes bridged by silicon atoms was achieved. Correspondence to: Di Liscia EJ, Gerencia de Física, GIyA, Comisión Nacional de Energía Atómica, Avda. General Paz 1499, (1650) Gral. San Martín, Pcia. Buenos Aires, Argentina, E-mail: diliscia@tandar.cnea.gov.ar Received: October 28, 2017; Accepted: December 05, 2017; Published: December 09, 2017 Di Liscia EJ (2017) Fullerene-silicon polymerization evidence Interdiscip J Chem, 2017 doi: 10.15761/IJC.1000123 Volume 2(2): 2-4 anode at 300 W. The measurements were made on a 2 mm2 area over the sample and at different depths, removing material by sputtering. For the data analysis, the spectra were fitted by Gaussian/Lorentzian convolution functions with a previous optimization of the background parameters. The background was modeled by a Shirley background function. Some deposited samples were annealed in vacuum (5.10-7 mbar) at 600o C after deposition in order to study the material thermal stability. Results SEM images for pure fullerene and C60-Si samples are shown in Figure 1. In the later, hexagonal structures of a few μm are clearly observed, a major morphologic difference with pure C60 samples. EDS measurements over several points at different crystals on C60-Si samples grown on Ge (100) substrates, mainly detected both C and Si presence at an atomic ratio less than 2% Si respect to C. Also, a weak signal of oxygen was measured, which could have been absorbed after deposition when exposed to air. Germanium corresponding to the substrate was also observed. Raman spectra from pristine, deposited fullerene and from C60Si samples are shown in Figure 2. If a power density above 4 W/ mm2 is used a clear change in the spectra is observed. The spectra show no significant differences between the samples. The breathing frequencies of pristine fullerene were observed in both deposited C60 and C60-Si samples, indicating that fullerenes were not broken during deposition. Si-Si signal was not observed, suggesting that there was not Si segregation. The spectra for pure silicon samples show the characteristic bands of amorphous silicon [20]. WAXS results are shown in Figure 3. Peaks corresponding to cubic and hexagonal (marked by arrows) fullerene structures are observed in both deposited C60 and C60-Si samples [21]. Broader peaks than in the case of pristine fullerene (our source material) could indicate smaller grain sizes. Peak positions for the cubic symmetry are the same for all samples and this implies that silicon atoms among fullerenes are not significantly changing the lattice parameters nor the distance between fullerenes. WAXS pattern’s peaks after 600o C in vacuum annealing become a little thinner, but both cubic and hexagonal phases are still present. As cubic phase was the predominant one, numerical calculations were made using the semiempirical many-body Tersoff potentials, which had been used and achieved a good description of carbon fullerenes, including several polymeric phases [14,22]. A zincblende C60-Si structure was found to be stable. Results for XPS measurements for annealed C60-Si samples are shown in Figure 4. The peaks can be assigned to Si2p, Si2s, C1s and O1s. Figure 1. SEM images for pure deposited C60 (a) and deposited C60-Si (b). Figure 2. Raman spectra for pure C60 (a) and C60-Si (b) deposited samples. Figure 3. WAXS patterns from pristine C60, deposited C60 and C60-Si on Si(100) wafer and C60-Si samples deposited on Si(100) and Ge(100) wafers and annealed to 600o C after deposition. The peaks position and width shown in Fig 4. lightly change at different depths. An atom proportion of 53.4% carbon, 27.2% silicon and 19.4% oxygen was found (after 60 seconds of argon sputtering at 1000 V), which ensures an enough quantity of silicon in order to conform a zincblende structure as proposed in the simulation results. In Figure 4 (b) and (c) the analysis for Si2p and O1s XPS signals is shown. The envelop of each Si2p contribution was the result of 2p 3/2 Di Liscia EJ (2017) Fullerene-silicon polymerization evidence Interdiscip J Chem, 2017 doi: 10.15761/IJC.1000123 Volume 2(2): 3-4 and 2p 1/2 orbitals summation where the doublet separations energy and the branching ratio were fixed in 0.6 eV and 0.5 eV respectively. From the Si2p signal analysis, a proportion of 21.7% Si-C, 1.6% SiO2 and 76.7% Si-O was calculated. For O1s signal, mostly Si-O was found, which could be attributed to some species like CSiOn (n<3) [23], and around 1% of SiO2. It is hard to fit correctly for C1s signal, as C-C bonds are indistinguishable from C-Si ones. But the peak is consistent with a mixture of signals from C-Si, C-C (sp3 carbons) and C=C (sp2 carbons) bonds. By analyzing O1s and C1s results, it can be ensured that a negligible amount of C=O and C-O bonds are present in the sample [24-28]. The thermal stability is the most outstanding result. While deposited C60 samples are completely wiped out when annealed in vacuum at 600o C, leaving the substrate totally clean, C60-Si samples morphology and Raman spectra are unchanged. Even for the samples where pure fullerene and C60-Si were simultaneously deposited in different zones of the substrate, the C60 was completely removed while the mixed material remained unaltered. This result implies that covalent bonds were formed between fullerenes. Conclusion It was observed that crystalline structures are formed when C60 and silicon are simultaneously deposited by the described method. This is clearly suggested by the shapes observed by SEM and confirmed by WAXS. Looking at the diffractograms is possible to affirm that cubic fcc and hexagonal structures are present in the sample, the former in agreement with our simulation results. While both could be assigned to a pure C60 material, the presence of silicon was confirmed by EDS and XPS also no evidence of segregation was found. Moreover, XPS analysis evidenced a considerable proportion (27.1%) of silicon atoms bonded to carbon atoms, but measuring over individual crystals by EDS a relation lesser than 2% was found between silicon and carbon atoms. Even more significant, the samples show no perceptible alterations after an annealing at 600o C, a temperature at which pure C60 samples deposited by the same method were completely cleaned. This suggests that covalent bonds were formed between fullerenes when C60 deposition is carried simultaneously with silicon deposition. Moreover, calculations employing the semiempirical many-body Tersoff potential found a zincblende structure to be stable and such configuration is in concordance with most of our experimental results.


Introduction
Many works have been published about fullerenes and in particular about C 60 or buckyminsterfullerene, the most stable of fullerenes. Its properties are of great value in a diversity of applications like organic solar cells, hydrogen gas storage, metals strengthening/hardening, optical limiters, solid sensors, drug carriers, etc [1][2][3][4][5][6][7][8]. The fabrication of novel nanostructured materials from fullerenes is promising for several of these applications. However weak van der Waals bond between fullerenes is a serious limitation for some applications and polymerization has been a vastly studied solution. A variety of fullerene polymers were reported, using photopolymerization, pressure-induced polymerization, charge-transfer polymerization mediated by metals, electron beam-induced and plasma-induced polymerization as some strategies to prepare the so-called all-carbon polyfullerenes [9][10][11][12][13][14][15]. In particular, using Group IV elements in order to bind fullerene molecules has been investigated and both theoretical and experimental works about fullerenes molecules bonded by silicon bridges have been published by , which motivates the possibility to construct a polymerized fullerene-silicon material.
We report experimental results for deposition of C 60 -Si material by simultaneous thermal vaporization of fullerene source and chemical vapor deposition from silane source. The materials produced were characterized by Scanning Electron Microscopy, (SEM), Energydispersive X-ray spectroscopy (EDS), Micro-Raman spectroscopy, Wide-angle X-ray scattering (WAXS), X-ray photoelectron spectroscopy (XPS) and their thermal stability was also studied.

Experimental
Fullerene and Si were deposited on Ge (100) and Si (100) wafers in a vacuum chamber. Fullerene source was pristine C 60 powder (99.9% C 60 ) evaporated at 550º C and silane (SiH 4 ) was introduced in the chamber and dissociated on a tungsten hot filament in order to incorporate silicon atoms to the C 60 lattice. Fullerene source was degassed in situ at 300º C and 5.10 -7 mbar for several hours in order to release the oxygen that could be trapped in it. The deposition process was carried at 5.10 -5 mbar with both evaporated fullerene and dissociated silane mixing in

Abstract
We report experimental results for C 60 -Si deposition by simultaneous thermal vaporization of fullerene source and chemical vapor deposition from silane source. The samples were characterized by Scanning Electron Microscopy, Energy-dispersive X-ray spectroscopy, Micro-Raman spectroscopy, Wide-angle X-ray scattering, X-ray photoelectron spectroscopy and its thermal stability was studied, discussed and compared with pure C 60 deposited by the same method. A crystalline material was obtained and results suggest that a polymerization of fullerenes bridged by silicon atoms was achieved.
Correspondence to: Di Liscia EJ, Gerencia de Física, GIyA, Comisión Nacional de Energía Atómica, Avda. General Paz 1499, (1650) Gral. San Martín, Pcia. Buenos Aires, Argentina, E-mail: diliscia@tandar.cnea.gov.ar anode at 300 W. The measurements were made on a 2 mm 2 area over the sample and at different depths, removing material by sputtering. For the data analysis, the spectra were fitted by Gaussian/Lorentzian convolution functions with a previous optimization of the background parameters. The background was modeled by a Shirley background function.
Some deposited samples were annealed in vacuum (5.10 -7 mbar) at 600º C after deposition in order to study the material thermal stability.

Results
SEM images for pure fullerene and C 60 -Si samples are shown in Figure 1. In the later, hexagonal structures of a few μm are clearly observed, a major morphologic difference with pure C 60 samples. EDS measurements over several points at different crystals on C 60 -Si samples grown on Ge (100) substrates, mainly detected both C and Si presence at an atomic ratio less than 2% Si respect to C. Also, a weak signal of oxygen was measured, which could have been absorbed after deposition when exposed to air. Germanium corresponding to the substrate was also observed.
Raman spectra from pristine, deposited fullerene and from C 60 -Si samples are shown in Figure 2. If a power density above 4 W/ mm 2 is used a clear change in the spectra is observed. The spectra show no significant differences between the samples. The breathing frequencies of pristine fullerene were observed in both deposited C 60 and C 60 -Si samples, indicating that fullerenes were not broken during deposition. Si-Si signal was not observed, suggesting that there was not Si segregation. The spectra for pure silicon samples show the characteristic bands of amorphous silicon [20].
WAXS results are shown in Figure 3. Peaks corresponding to cubic and hexagonal (marked by arrows) fullerene structures are observed in both deposited C 60 and C 60 -Si samples [21]. Broader peaks than in the case of pristine fullerene (our source material) could indicate smaller grain sizes. Peak positions for the cubic symmetry are the same for all samples and this implies that silicon atoms among fullerenes are not significantly changing the lattice parameters nor the distance between fullerenes. WAXS pattern's peaks after 600º C in vacuum annealing become a little thinner, but both cubic and hexagonal phases are still present.
As cubic phase was the predominant one, numerical calculations were made using the semiempirical many-body Tersoff potentials, which had been used and achieved a good description of carbon fullerenes, including several polymeric phases [14,22]. A zincblende C 60 -Si structure was found to be stable.
Results for XPS measurements for annealed C 60 -Si samples are shown in Figure 4. The peaks can be assigned to Si 2p , Si 2s , C 1s and O 1s .   The peaks position and width shown in Fig 4. lightly change at different depths. An atom proportion of 53.4% carbon, 27.2% silicon and 19.4% oxygen was found (after 60 seconds of argon sputtering at 1000 V), which ensures an enough quantity of silicon in order to conform a zincblende structure as proposed in the simulation results.
In Figure 4 (b) and (c) the analysis for Si 2p and O 1s XPS signals is shown. The envelop of each Si 2p contribution was the result of 2p 3/2 and 2p 1/2 orbitals summation where the doublet separations energy and the branching ratio were fixed in 0.6 eV and 0.5 eV respectively. From the Si 2p signal analysis, a proportion of 21.7% Si-C, 1.6% SiO 2 and 76.7% Si-O was calculated. For O 1s signal, mostly Si-O was found, which could be attributed to some species like CSiO n (n<3) [23], and around 1% of SiO 2 . It is hard to fit correctly for C 1s signal, as C-C bonds are indistinguishable from C-Si ones. But the peak is consistent with a mixture of signals from C-Si, C-C (sp 3 carbons) and C=C (sp 2 carbons) bonds. By analyzing O 1s and C 1s results, it can be ensured that a negligible amount of C=O and C-O bonds are present in the sample [24][25][26][27][28].
The thermal stability is the most outstanding result. While deposited C 60 samples are completely wiped out when annealed in vacuum at 600º C, leaving the substrate totally clean, C 60 -Si samples morphology and Raman spectra are unchanged. Even for the samples where pure fullerene and C 60 -Si were simultaneously deposited in different zones of the substrate, the C 60 was completely removed while the mixed material remained unaltered. This result implies that covalent bonds were formed between fullerenes.

Conclusion
It was observed that crystalline structures are formed when C 60 and silicon are simultaneously deposited by the described method. This is clearly suggested by the shapes observed by SEM and confirmed by WAXS. Looking at the diffractograms is possible to affirm that cubic fcc and hexagonal structures are present in the sample, the former in agreement with our simulation results. While both could be assigned to a pure C 60 material, the presence of silicon was confirmed by EDS and XPS also no evidence of segregation was found. Moreover, XPS analysis evidenced a considerable proportion (27.1%) of silicon atoms bonded to carbon atoms, but measuring over individual crystals by EDS a relation lesser than 2% was found between silicon and carbon atoms. Even more significant, the samples show no perceptible alterations after an annealing at 600º C, a temperature at which pure C 60 samples deposited by the same method were completely cleaned. This suggests that covalent bonds were formed between fullerenes when C 60 deposition is carried simultaneously with silicon deposition. Moreover, calculations employing the semiempirical many-body Tersoff potential found a zincblende structure to be stable and such configuration is in concordance with most of our experimental results.