- Email: [email protected]
Effect of synthesis temperature on hardness of carbon phases prepared from C60 and nanosized diamonds under pressure
Carbon 44 (2006) 2027–2031 www.elsevier.com/locate/carbon
Effect of synthesis temperature on hardness of carbon phases prepared from C60 and nanosized diamonds under pressure N.V. Surovtsev a,*, A.A. Kalinin b, V.K. Malinovsky a, Yu.N. Pal’yanov b, A.S. Yunoshev c a
Institute of Automation and Electrometry, Russian Academy of Sciences, Novosibirsk 630090, Russia Institute of Mineralogy and Petrography, Russian Academy of Sciences, Novosibirsk 630090, Russia c Lavrentyev Institute of Hydrodynamics, Russian Academy of Sciences, Novosibirsk 630090, Russia
b
Received 23 November 2005; accepted 21 January 2006 Available online 15 March 2006
Abstract Vicker’s hardness and Raman scattering spectra have been studied for carbon phases prepared from C60 fullerene and nanosized diamonds at high temperatures and a pressure of 6 GPa. It was found that the hardness dependence on annealing temperature has a maximum near 1100 K for both fullerene and nanosized diamonds as initial materials. This temperature is only slightly higher than the temperature at which the C60 cage collapses, and appears to correspond to the termination of intercluster bonding in the case of nanosized diamonds. The hardness maximum is interpreted as a result of competition between an increase in intercluster/intercage bonding and local instability for graphitic-like ordering. 2006 Elsevier Ltd. All rights reserved. Keywords: Fullerene; High pressure; Raman spectroscopy; Mechanical properties
1. Introduction Investigations of new carbon phases, synthesized from the fullerite powder under high-temperature and high-pressure conditions, attract significant attention nowadays [1– 3]. The phase diagram of the carbon materials and their properties is extremely rich, giving rise to the fundamental topics in this research area. Moreover, the unique features of the carbon phases, which are achievable, define significant interest from applied physics. In some cases, the carbon phases reach very high values of hardness (these phases often called as ‘‘hard carbon’’), which are combined with a very high plasticity of these materials [3–6]. A critical region of the phase diagram close to the temperature, at which fullerene cages are destroyed, is especially interesting (see, for example, the phase diagram in [3,7]). Some of car-
*
Corresponding author. Tel.: +7 383 3307978; fax: +7 383 3333863. E-mail address: [email protected] (N.V. Surovtsev).
0008-6223/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.carbon.2006.01.011
bon phases, synthesized in this region, have a ferromagnetism behavior [8–10]. This feature is found for the phases, prepared at pressure of 6–9 GPa and T = 800–1100 K. It is interesting that according to [11] this region corresponds to a jump in hardness, if the hardness of the carbon phases is considered versus the synthesis temperature. In the work [12] it was found that there is a hardness maximum near the synthesis temperature of 900 K for samples, prepared at pressure of 9 GPa. In this work, a significant dependence on annealing time at high temperatures was reported. The annealing time was varied from 0.5 min to 30 min. In [12], it was suggested that the hardness maximum is related to an intermediate state with a time-dependent structure. The statement, that the hardness maximum has kinetic origin and is absent for samples well equilibrated at high temperatures, should be checked. Another question—what is a relation between the hardness maximum and the temperature, at which C60 cages are destroyed? Is the hardness maximum a specific feature for fullerene molecules, or carbon phases, synthesized from other nanometer structured
2028
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031
carbons, like nanosized diamonds, also have a similar maximum? The present study is devoted to these topics. 2. Experiment details Carbon phases synthesis was realized by a high-pressure apparatus of a ‘‘split sphere’’ type (BARS) [13] and rapid temperature quenching to the ambient value. Pressure calibration was realized as described in [14]. Stabilized modification of ZrO2 was used as a material for high-pressure cell, being in a shape of a tetragonal prism. A graphite heater was used [14]. Temperature was measured in each experiment with Pt6%Rh/Pt30%Rh thermocouple whose junction was placed in the central part of the cell, just above a sample. The initial material was packed into platinum ampoule and separated from the heater by MgO or CsCl medium. This allowed us to exclude ingress of graphite. The samples were synthesized from a fullerite powder with C60 content not less 99.9% and from nanosized diamonds, extracted and refined from the explosion soot. The initial material was compacted to cylinders of B 7.1 mm and height of 1.7–2.4 mm. Synthesis of carbon phases was realized at P = 6 GPa and at T = 520, 870, 970, 1070, 1170, and 1470 K. The annealing time was 5 h, which excludes the kinetic effects like described in [12]. Raman scattering spectra of the samples were recorded using a triple spectrograph TriVista 777 from S& I/PiActon and a line 532 nm from a solid-state laser. A multichannel regime by a silicon CCD matrix with a spectral resolution of 2.5 cm1 was used. Nominally right angle Raman scattering experiment was carried out at room temperature with a grazing of the laser beam on the surface (about 60 from the normal). The spherical–cylindrical lens was used in order to focus the laser beam of a power 300 mW into a rectangle 10 · 0.2 mm on the sample surface, being parallel to a spectrometer entrance slit. Such illumination of the samples results to a relatively low pump power density together with a high integral pump intensity. No polarization selection was applied in Raman experiment. Experimental spectra of inelastic light scattering contained a significant photoluminescence background, which in a limited spectral range was considered as a frequencyindependent background and was subtracted from the experimental spectra. The hardness was measured by the standard Vicker’s indentation technique. The value of indenter load was 2 N. In order to overcome the difficulty of high elastic recovery, which was typical for our carbon samples, the true indentation surface was estimated through the sizes of diagonals, clearly seen in microscope with crossed polarizations. 3. Experimental results and discussion Raman spectra of the carbon phases, synthesized from the fullerite powder, are shown in Fig. 1 in a spectral range
Fig. 1. Raman spectra of carbon phases, synthesized from fullerite C60. (a) Initial material (solid line) and for synthesis temperature 870 K (dashed line). (b) The synthesis temperature 970 K (thin line), 1170 K (thick line), and 1470 K (dashed line).
1250–1670 cm1. This spectral range provides the most intense lines and is quite informative for the carbon phase identification [15]. Raman spectra of initial C60 material and 2D-polymerized phase is shown in the top of Fig. 1. The maximum temperature, at which the synthesized carbon phase being a 2D-polymer, was 870 K. According to [7], the 2D-polymer synthesized at 870 K is mainly rhombohedral (1406 cm1 line) with a minor content of the tetragonal phase (the group of lines near 1450 cm1). (Note that according to the previous interpretation of [16], our sample is almost pure rhombohedral.) As it is seen from Fig. 1b, the pentagonal pinch mode (relatively narrow lines in 1400–1500 cm1 range [15]) or its analogies are absent for samples, synthesized at 970 K or higher temperatures. It means that the fullerene cages are destroyed for these carbon phases. Raman spectra of the carbon phases, synthesized at 1170 K and higher, are similar as for disordered graphite. A peculiarity of Raman spectrum for a phase, synthesized at 970 K, is a relatively narrow line near 1581 cm1, which is very close to the crystalline graphite line [15]. Presence of the so-called second-order G 0 band, observed near 2600–2700 cm1 for green-blue laser excitation, is a common feature to carbon materials, containing condensed aromatic structures ([17] and references therein). The intensity of this band is rather high for a second-order spectrum. Thus, existence of the significant G 0 band is the evidence that the carbon phase has a graphite-like ordering, resulting in the appearance of aromatic structures. Raman spectra, including the spectral range of G 0 band, are shown in Fig. 2 (the same intensity scale is used for the lower and higher frequency spectral ranges). The second-order G 0 band is well seen for Raman spectrum of the sample with the synthesis temperature 970 K and
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031
Fig. 2. Raman spectra of carbon phases, synthesized from fullerite C60. The synthesis temperature 870 K (thick line), 970 K (thin line), and 1470 K (dotted line).
absent for the spectrum of 2D-polymer. Thus, Fig. 2 proves that the carbon phases, synthesized at 970 K and higher temperatures, contain aromatic structures, corresponding to a graphite-like ordering on nanometer scale. Raman scattering spectra of the carbon phases, synthesized from nanosized diamonds, are shown in Fig. 3. An analog of the bulk diamond line is seen near 1325 cm1 in the spectrum of the initial material (the Raman line frequency of a nanosized diamond is lower in comparison with a bulk sample due to small size of crystallites [18]). Also, a Raman peak near 1600 cm1, an analog of G-mode of disordered graphite, is seen in the spectrum of the initial material. The presence of this peak means that the significant part of carbon atoms in the nanosized diamond material have a graphite-like bonding.
Fig. 3. Raman spectra of carbon phases, synthesized from nanosized diamonds. (a) Initial material (dotted line) and for synthesis temperature 870 K (solid line). (b) The syntheses temperature 970 K (dotted line), 1170 K (thick line), and 1470 K (thin line).
2029
As the synthesis temperature increases, Raman spectra of the carbon phases, prepared from nanosized diamonds, become similar to the spectra of disordered graphite (Fig. 3). At that, the position of the peak near 1300 cm1 increases, transforming to an analog of D-mode [15] for the sample with the synthesis temperature 1470 K. The peak position versus the synthesis temperature is shown in Fig. 4a. An interesting feature of this dependence is the plateau in the range 870–1170 K, which corresponds to the spectral position only slightly above the bulk diamond Raman line, but significantly below the position of D-mode. It is expected that the basic mechanism for the transformation of the nanosized diamond powder under high pressure–temperature conditions is related to rebounding of carbon atoms, placed on nanocrystallite surfaces. New bonds link neighbor nanocrystallites, giving rise to a 3D network. A photoluminescence background in experimental spectra arises mainly from defect states of carbon atoms on nanocrystallite surfaces. So, in a rough approximation, the intensity of the photoluminescence background reflects the density of defect states or number of carbon atoms on nanocrystallite surfaces, not bonding to neighbor nanocrystallites or nanoclusters. The ratio between photoluminescence and Raman intensity versus the synthesis temperature is shown in Fig. 4b (the strongest Raman line was taken as Raman intensity). It is seen that this ratio sharply decreases as the synthesis temperature increases up to about 900–1000 K. For samples with higher synthesis
Fig. 4. Carbon phases properties versus synthesis temperature. (a) Spectral position of the line 1300 cm1 for the phases synthesized from nanosized diamonds. Arrow points the bulk diamond position. (b) Ratio between the photoluminescence and Raman scattering intensities: circles for nanosized diamonds as an initial material and triangles for fullerites C60. (c) Hardness: circles for nanosized diamonds as an initial material and triangles for fullerites C60. Line is a guide for eyes.
2030
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031
temperatures, the ratio magnitude is the same and independent of the synthesis temperature. According to the interpretation proposed, this result means that all carbon atoms on nanocluster surfaces have already reacted. For the sake of comparison, the ratio in the case of the carbon phases, prepared from C60 powder, is also shown in Fig. 4b. In the case of the last carbon phase, the ratio is approximately independent of the synthesis temperature. Hardness data of the samples studied are shown in Fig. 4c. There is a sharp increase of hardness of the carbon phases, prepared from C60 powder, in the range of 970– 1170 K for the synthesis temperature. At the higher synthesis temperature, 1470 K, the hardness decreases two times in comparison with the maximum, being nevertheless much higher than the hardness of the 2D-polymer phase or starting material. The use of the long synthesis time in our sample preparation excludes the interpretation of the hardness maximum in the spirit of [12]. The hardness maximum reflects rather the hardness property of metastable carbon phases. As it is seen from Figs. 1 and 2, the hardness maximum is reached for the carbon phases, where the fullerene cages are destroyed and the significant part of carbon atoms belong to aromatic structures. The hardness maximum can be interpreted as a consequence of intermolecular bonding, resulting to the cage destruction, when the majority of carbon atoms is involved into intermolecular bonding. The corresponding 3D network of carbon atoms is characterized by a high hardness. It is clear that the carbon bonds are enough flexible at the hardness maximum temperature. The synthesis temperature increase above 1100 K favors the carbon bonds to organize more extended graphitic-like structures, since graphite is the most stable phase at the P T parameters considered. However, these energetically favorable structures are less hard. Thus, the hardness maximum is the result of a competition between the increase of the intercage bonding and the local instability for graphitic-like ordering. Let us consider the nanocrystallites of nanosized diamonds as an analog of C60 molecules, but having different chemical order. It is seen in Fig. 4c, that the carbon phases, prepared from nanosized diamonds, obey the hardness versus the synthesis temperature dependence, being very similar to the case of fullerenes as initial material. Thus, the hardness maximum is not specific of fullerene molecules, but corresponds to the common feature of nanostructured carbon phases. This result evidences that the hardness maximum reflects the competition between the increase of the intercage/intercluster bonding and the tendency for graphitic-like ordering. 4. Conclusions Existence of the maximum for the hardness versus the synthesis temperature was experimentally demonstrated for the carbon phases, prepared from fullerite and nanosized diamonds under P = 6 GPa. This maximum corresponds to metastable states of the carbon phases and not to a transi-
tional kinetic effect on the level one-tens minutes. In the case of fullerite as initial material, the hardness maximum is reached for the carbon phases with destroyed fullerene cages. It was shown that the hardness maximum is not a privilege feature of the carbon phases, prepared from fullerite, but also exists for the carbon phases, synthesized from nanosized diamonds. In the last case, the hardness maximum corresponds to the minimum concentration of internal defects, as was monitored by photoluminescence. The hardness maximum is interpreted as the result of the competition between the increase of the intercluster/intercage bonding and the local instability for graphitic-like ordering. Acknowledgement This work was supported by the Interdisciplinary Science Fund at the Russian Foundation for Basic Research of the Siberian Branch of the Russian Academy of Sciences (project no. 140). References [1] Blank VD, Buga SG, Dubitsky GA, Serebryanaya NR, Popov MY, Sundqvist B. High-pressure polymerized phases of C60. Carbon 1998;36:319–34. [2] Sundqvist B. Fullerenes under high pressures. Adv Phys 1999;48(1): 1–134. [3] Brazhkin VV, Lyapin AG, Popova SV, Bayliss SC, Varfolomeeva TD, et al. Interplay between the structure and properties of new metastable carbon phases obtained under high pressures from fullerite C60 and carbine. JETP Lett 2002;76(11):681–92. [4] Blank VD, Buga SG, Serebryanaya NR, Denisov VN, Dubitsky GA, et al. Ultrahard and superhard carbon phases produced from C60 by heating at high-pressure—structural and Raman studies. Phys Lett A 1995;205(2–3):208–16. [5] Brazhkin VV, Lyapin AG, Antonov YV, Popova SV, Klyuev YA, et al. Amorphization of fullerite (C-60) at high-pressures. JETP Lett 1995;62(4):350–5. [6] Brazhkin VV, Lyapin AG, Popova SV, Klyuev YA, Naletov AM. Mechanical properties of the 3D polymerized, sp2–sp3 amorphous, and diamond-plus-graphite nanocomposite carbon phases prepared from C60 under high pressure. J Appl Phys 1998;84(1):219–25. [7] Davydov VA, Kashevarova LS, Rakhmanina AV, Senyavin VM, Ce´olin R, et al. Spectroscopic study of pressure-polymerized phases of C60. Phys Rev B 2000;61(18):11936–45. [8] Makarova Tl, Sundqvist B, Ho¨hne R, Esquinazi P, Kopelevich Y, Scharff P, et al. Magnetic carbon. Nature 2001;413:718–20. [9] Wood RA, Lewis MH, Lees MR, Bennington SM, Cain MG, Kitamura N. Ferromagnetic fullerene. J Phys.: Condens Matter 2002;14(22):L385–91. [10] Narozhnyi VN, Mu¨ller KH, Eckert D, Teresiak A, Dunsch L, et al. Ferromagnetic carbon with enhanced Curie temperature. Physica B 2003;329–323:1217–8. [11] Lyapin AG, Brazhkin VV, Gromnitskaya EL, Popova SV, Stal’gorova OV, Voloshin RN, et al. Hardening of fullerite C60 during temperature-induced polymerization and amorphization under pressure. Appl Phys Lett 2000;76(6):712–4. [12] Wood RA, Lewis MH, West G, Bennington SM, Cain MG, Kitamura N. Transmission electron microscopy, electron diffraction and hardness studies of high-pressure and high-temperature treated C60. J Phys.: Condens Matter 2000;12(50):10411–21. [13] Pal’yanov YN, Khokhryakov AF, Borzdov YM, Sokol AG, Gusev VA, Rylov GM, et al. Growth conditions and real structure of
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031 synthetic diamond crystals. Russ Geol Geophys 1997;38(5):920– 45. [14] Pal’yanov YN, Sokol AG, Borzdov YM, Khokhryakov AF. Fluidbearing alkaline carbonate melts as the medium for the formation of diamonds in the Earth’s mantle: an experimental study. Lithos 2002;30(3–4):145–59. [15] Dresselhaus MS, Pimenta MA, Eklund PC, Dresselhaus G. Raman scattering in fullerenes and related carbon-based materials. In: Weber WH, Merlin R, editors. Raman scattering in material science, vol. 42. Berlin: Springer; 2000. p. 314–64.
2031
[16] Rao AM, Eklund PC, Hodeau J-L, Marques L, Nunez-Requeiro M. Infrared and Raman studies of pressure-polymerized C60s. Phys Rev B 1997;55(7):4766–73. [17] Matthews MJ, Pimenta MA, Dresselhaus G, Dresselhaus MS, Endo M. Origin of dispersive effects of the Raman D band in carbon materials. Phys Rev B 1999;59(10):R6585–8. [18] Gorelik VS, Igo AV, Mikov SN. Light Raman scattering in limited crystals. Zhurnal Experimentalnoi I Teoreticheskoi Fiziki 1996; 109(6):2141–9.
Effect of synthesis temperature on hardness of carbon phases prepared from C60 and nanosized diamonds under pressure N.V. Surovtsev a,*, A.A. Kalinin b, V.K. Malinovsky a, Yu.N. Pal’yanov b, A.S. Yunoshev c a
Institute of Automation and Electrometry, Russian Academy of Sciences, Novosibirsk 630090, Russia Institute of Mineralogy and Petrography, Russian Academy of Sciences, Novosibirsk 630090, Russia c Lavrentyev Institute of Hydrodynamics, Russian Academy of Sciences, Novosibirsk 630090, Russia
b
Received 23 November 2005; accepted 21 January 2006 Available online 15 March 2006
Abstract Vicker’s hardness and Raman scattering spectra have been studied for carbon phases prepared from C60 fullerene and nanosized diamonds at high temperatures and a pressure of 6 GPa. It was found that the hardness dependence on annealing temperature has a maximum near 1100 K for both fullerene and nanosized diamonds as initial materials. This temperature is only slightly higher than the temperature at which the C60 cage collapses, and appears to correspond to the termination of intercluster bonding in the case of nanosized diamonds. The hardness maximum is interpreted as a result of competition between an increase in intercluster/intercage bonding and local instability for graphitic-like ordering. 2006 Elsevier Ltd. All rights reserved. Keywords: Fullerene; High pressure; Raman spectroscopy; Mechanical properties
1. Introduction Investigations of new carbon phases, synthesized from the fullerite powder under high-temperature and high-pressure conditions, attract significant attention nowadays [1– 3]. The phase diagram of the carbon materials and their properties is extremely rich, giving rise to the fundamental topics in this research area. Moreover, the unique features of the carbon phases, which are achievable, define significant interest from applied physics. In some cases, the carbon phases reach very high values of hardness (these phases often called as ‘‘hard carbon’’), which are combined with a very high plasticity of these materials [3–6]. A critical region of the phase diagram close to the temperature, at which fullerene cages are destroyed, is especially interesting (see, for example, the phase diagram in [3,7]). Some of car-
*
Corresponding author. Tel.: +7 383 3307978; fax: +7 383 3333863. E-mail address: [email protected] (N.V. Surovtsev).
0008-6223/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.carbon.2006.01.011
bon phases, synthesized in this region, have a ferromagnetism behavior [8–10]. This feature is found for the phases, prepared at pressure of 6–9 GPa and T = 800–1100 K. It is interesting that according to [11] this region corresponds to a jump in hardness, if the hardness of the carbon phases is considered versus the synthesis temperature. In the work [12] it was found that there is a hardness maximum near the synthesis temperature of 900 K for samples, prepared at pressure of 9 GPa. In this work, a significant dependence on annealing time at high temperatures was reported. The annealing time was varied from 0.5 min to 30 min. In [12], it was suggested that the hardness maximum is related to an intermediate state with a time-dependent structure. The statement, that the hardness maximum has kinetic origin and is absent for samples well equilibrated at high temperatures, should be checked. Another question—what is a relation between the hardness maximum and the temperature, at which C60 cages are destroyed? Is the hardness maximum a specific feature for fullerene molecules, or carbon phases, synthesized from other nanometer structured
2028
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031
carbons, like nanosized diamonds, also have a similar maximum? The present study is devoted to these topics. 2. Experiment details Carbon phases synthesis was realized by a high-pressure apparatus of a ‘‘split sphere’’ type (BARS) [13] and rapid temperature quenching to the ambient value. Pressure calibration was realized as described in [14]. Stabilized modification of ZrO2 was used as a material for high-pressure cell, being in a shape of a tetragonal prism. A graphite heater was used [14]. Temperature was measured in each experiment with Pt6%Rh/Pt30%Rh thermocouple whose junction was placed in the central part of the cell, just above a sample. The initial material was packed into platinum ampoule and separated from the heater by MgO or CsCl medium. This allowed us to exclude ingress of graphite. The samples were synthesized from a fullerite powder with C60 content not less 99.9% and from nanosized diamonds, extracted and refined from the explosion soot. The initial material was compacted to cylinders of B 7.1 mm and height of 1.7–2.4 mm. Synthesis of carbon phases was realized at P = 6 GPa and at T = 520, 870, 970, 1070, 1170, and 1470 K. The annealing time was 5 h, which excludes the kinetic effects like described in [12]. Raman scattering spectra of the samples were recorded using a triple spectrograph TriVista 777 from S& I/PiActon and a line 532 nm from a solid-state laser. A multichannel regime by a silicon CCD matrix with a spectral resolution of 2.5 cm1 was used. Nominally right angle Raman scattering experiment was carried out at room temperature with a grazing of the laser beam on the surface (about 60 from the normal). The spherical–cylindrical lens was used in order to focus the laser beam of a power 300 mW into a rectangle 10 · 0.2 mm on the sample surface, being parallel to a spectrometer entrance slit. Such illumination of the samples results to a relatively low pump power density together with a high integral pump intensity. No polarization selection was applied in Raman experiment. Experimental spectra of inelastic light scattering contained a significant photoluminescence background, which in a limited spectral range was considered as a frequencyindependent background and was subtracted from the experimental spectra. The hardness was measured by the standard Vicker’s indentation technique. The value of indenter load was 2 N. In order to overcome the difficulty of high elastic recovery, which was typical for our carbon samples, the true indentation surface was estimated through the sizes of diagonals, clearly seen in microscope with crossed polarizations. 3. Experimental results and discussion Raman spectra of the carbon phases, synthesized from the fullerite powder, are shown in Fig. 1 in a spectral range
Fig. 1. Raman spectra of carbon phases, synthesized from fullerite C60. (a) Initial material (solid line) and for synthesis temperature 870 K (dashed line). (b) The synthesis temperature 970 K (thin line), 1170 K (thick line), and 1470 K (dashed line).
1250–1670 cm1. This spectral range provides the most intense lines and is quite informative for the carbon phase identification [15]. Raman spectra of initial C60 material and 2D-polymerized phase is shown in the top of Fig. 1. The maximum temperature, at which the synthesized carbon phase being a 2D-polymer, was 870 K. According to [7], the 2D-polymer synthesized at 870 K is mainly rhombohedral (1406 cm1 line) with a minor content of the tetragonal phase (the group of lines near 1450 cm1). (Note that according to the previous interpretation of [16], our sample is almost pure rhombohedral.) As it is seen from Fig. 1b, the pentagonal pinch mode (relatively narrow lines in 1400–1500 cm1 range [15]) or its analogies are absent for samples, synthesized at 970 K or higher temperatures. It means that the fullerene cages are destroyed for these carbon phases. Raman spectra of the carbon phases, synthesized at 1170 K and higher, are similar as for disordered graphite. A peculiarity of Raman spectrum for a phase, synthesized at 970 K, is a relatively narrow line near 1581 cm1, which is very close to the crystalline graphite line [15]. Presence of the so-called second-order G 0 band, observed near 2600–2700 cm1 for green-blue laser excitation, is a common feature to carbon materials, containing condensed aromatic structures ([17] and references therein). The intensity of this band is rather high for a second-order spectrum. Thus, existence of the significant G 0 band is the evidence that the carbon phase has a graphite-like ordering, resulting in the appearance of aromatic structures. Raman spectra, including the spectral range of G 0 band, are shown in Fig. 2 (the same intensity scale is used for the lower and higher frequency spectral ranges). The second-order G 0 band is well seen for Raman spectrum of the sample with the synthesis temperature 970 K and
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031
Fig. 2. Raman spectra of carbon phases, synthesized from fullerite C60. The synthesis temperature 870 K (thick line), 970 K (thin line), and 1470 K (dotted line).
absent for the spectrum of 2D-polymer. Thus, Fig. 2 proves that the carbon phases, synthesized at 970 K and higher temperatures, contain aromatic structures, corresponding to a graphite-like ordering on nanometer scale. Raman scattering spectra of the carbon phases, synthesized from nanosized diamonds, are shown in Fig. 3. An analog of the bulk diamond line is seen near 1325 cm1 in the spectrum of the initial material (the Raman line frequency of a nanosized diamond is lower in comparison with a bulk sample due to small size of crystallites [18]). Also, a Raman peak near 1600 cm1, an analog of G-mode of disordered graphite, is seen in the spectrum of the initial material. The presence of this peak means that the significant part of carbon atoms in the nanosized diamond material have a graphite-like bonding.
Fig. 3. Raman spectra of carbon phases, synthesized from nanosized diamonds. (a) Initial material (dotted line) and for synthesis temperature 870 K (solid line). (b) The syntheses temperature 970 K (dotted line), 1170 K (thick line), and 1470 K (thin line).
2029
As the synthesis temperature increases, Raman spectra of the carbon phases, prepared from nanosized diamonds, become similar to the spectra of disordered graphite (Fig. 3). At that, the position of the peak near 1300 cm1 increases, transforming to an analog of D-mode [15] for the sample with the synthesis temperature 1470 K. The peak position versus the synthesis temperature is shown in Fig. 4a. An interesting feature of this dependence is the plateau in the range 870–1170 K, which corresponds to the spectral position only slightly above the bulk diamond Raman line, but significantly below the position of D-mode. It is expected that the basic mechanism for the transformation of the nanosized diamond powder under high pressure–temperature conditions is related to rebounding of carbon atoms, placed on nanocrystallite surfaces. New bonds link neighbor nanocrystallites, giving rise to a 3D network. A photoluminescence background in experimental spectra arises mainly from defect states of carbon atoms on nanocrystallite surfaces. So, in a rough approximation, the intensity of the photoluminescence background reflects the density of defect states or number of carbon atoms on nanocrystallite surfaces, not bonding to neighbor nanocrystallites or nanoclusters. The ratio between photoluminescence and Raman intensity versus the synthesis temperature is shown in Fig. 4b (the strongest Raman line was taken as Raman intensity). It is seen that this ratio sharply decreases as the synthesis temperature increases up to about 900–1000 K. For samples with higher synthesis
Fig. 4. Carbon phases properties versus synthesis temperature. (a) Spectral position of the line 1300 cm1 for the phases synthesized from nanosized diamonds. Arrow points the bulk diamond position. (b) Ratio between the photoluminescence and Raman scattering intensities: circles for nanosized diamonds as an initial material and triangles for fullerites C60. (c) Hardness: circles for nanosized diamonds as an initial material and triangles for fullerites C60. Line is a guide for eyes.
2030
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031
temperatures, the ratio magnitude is the same and independent of the synthesis temperature. According to the interpretation proposed, this result means that all carbon atoms on nanocluster surfaces have already reacted. For the sake of comparison, the ratio in the case of the carbon phases, prepared from C60 powder, is also shown in Fig. 4b. In the case of the last carbon phase, the ratio is approximately independent of the synthesis temperature. Hardness data of the samples studied are shown in Fig. 4c. There is a sharp increase of hardness of the carbon phases, prepared from C60 powder, in the range of 970– 1170 K for the synthesis temperature. At the higher synthesis temperature, 1470 K, the hardness decreases two times in comparison with the maximum, being nevertheless much higher than the hardness of the 2D-polymer phase or starting material. The use of the long synthesis time in our sample preparation excludes the interpretation of the hardness maximum in the spirit of [12]. The hardness maximum reflects rather the hardness property of metastable carbon phases. As it is seen from Figs. 1 and 2, the hardness maximum is reached for the carbon phases, where the fullerene cages are destroyed and the significant part of carbon atoms belong to aromatic structures. The hardness maximum can be interpreted as a consequence of intermolecular bonding, resulting to the cage destruction, when the majority of carbon atoms is involved into intermolecular bonding. The corresponding 3D network of carbon atoms is characterized by a high hardness. It is clear that the carbon bonds are enough flexible at the hardness maximum temperature. The synthesis temperature increase above 1100 K favors the carbon bonds to organize more extended graphitic-like structures, since graphite is the most stable phase at the P T parameters considered. However, these energetically favorable structures are less hard. Thus, the hardness maximum is the result of a competition between the increase of the intercage bonding and the local instability for graphitic-like ordering. Let us consider the nanocrystallites of nanosized diamonds as an analog of C60 molecules, but having different chemical order. It is seen in Fig. 4c, that the carbon phases, prepared from nanosized diamonds, obey the hardness versus the synthesis temperature dependence, being very similar to the case of fullerenes as initial material. Thus, the hardness maximum is not specific of fullerene molecules, but corresponds to the common feature of nanostructured carbon phases. This result evidences that the hardness maximum reflects the competition between the increase of the intercage/intercluster bonding and the tendency for graphitic-like ordering. 4. Conclusions Existence of the maximum for the hardness versus the synthesis temperature was experimentally demonstrated for the carbon phases, prepared from fullerite and nanosized diamonds under P = 6 GPa. This maximum corresponds to metastable states of the carbon phases and not to a transi-
tional kinetic effect on the level one-tens minutes. In the case of fullerite as initial material, the hardness maximum is reached for the carbon phases with destroyed fullerene cages. It was shown that the hardness maximum is not a privilege feature of the carbon phases, prepared from fullerite, but also exists for the carbon phases, synthesized from nanosized diamonds. In the last case, the hardness maximum corresponds to the minimum concentration of internal defects, as was monitored by photoluminescence. The hardness maximum is interpreted as the result of the competition between the increase of the intercluster/intercage bonding and the local instability for graphitic-like ordering. Acknowledgement This work was supported by the Interdisciplinary Science Fund at the Russian Foundation for Basic Research of the Siberian Branch of the Russian Academy of Sciences (project no. 140). References [1] Blank VD, Buga SG, Dubitsky GA, Serebryanaya NR, Popov MY, Sundqvist B. High-pressure polymerized phases of C60. Carbon 1998;36:319–34. [2] Sundqvist B. Fullerenes under high pressures. Adv Phys 1999;48(1): 1–134. [3] Brazhkin VV, Lyapin AG, Popova SV, Bayliss SC, Varfolomeeva TD, et al. Interplay between the structure and properties of new metastable carbon phases obtained under high pressures from fullerite C60 and carbine. JETP Lett 2002;76(11):681–92. [4] Blank VD, Buga SG, Serebryanaya NR, Denisov VN, Dubitsky GA, et al. Ultrahard and superhard carbon phases produced from C60 by heating at high-pressure—structural and Raman studies. Phys Lett A 1995;205(2–3):208–16. [5] Brazhkin VV, Lyapin AG, Antonov YV, Popova SV, Klyuev YA, et al. Amorphization of fullerite (C-60) at high-pressures. JETP Lett 1995;62(4):350–5. [6] Brazhkin VV, Lyapin AG, Popova SV, Klyuev YA, Naletov AM. Mechanical properties of the 3D polymerized, sp2–sp3 amorphous, and diamond-plus-graphite nanocomposite carbon phases prepared from C60 under high pressure. J Appl Phys 1998;84(1):219–25. [7] Davydov VA, Kashevarova LS, Rakhmanina AV, Senyavin VM, Ce´olin R, et al. Spectroscopic study of pressure-polymerized phases of C60. Phys Rev B 2000;61(18):11936–45. [8] Makarova Tl, Sundqvist B, Ho¨hne R, Esquinazi P, Kopelevich Y, Scharff P, et al. Magnetic carbon. Nature 2001;413:718–20. [9] Wood RA, Lewis MH, Lees MR, Bennington SM, Cain MG, Kitamura N. Ferromagnetic fullerene. J Phys.: Condens Matter 2002;14(22):L385–91. [10] Narozhnyi VN, Mu¨ller KH, Eckert D, Teresiak A, Dunsch L, et al. Ferromagnetic carbon with enhanced Curie temperature. Physica B 2003;329–323:1217–8. [11] Lyapin AG, Brazhkin VV, Gromnitskaya EL, Popova SV, Stal’gorova OV, Voloshin RN, et al. Hardening of fullerite C60 during temperature-induced polymerization and amorphization under pressure. Appl Phys Lett 2000;76(6):712–4. [12] Wood RA, Lewis MH, West G, Bennington SM, Cain MG, Kitamura N. Transmission electron microscopy, electron diffraction and hardness studies of high-pressure and high-temperature treated C60. J Phys.: Condens Matter 2000;12(50):10411–21. [13] Pal’yanov YN, Khokhryakov AF, Borzdov YM, Sokol AG, Gusev VA, Rylov GM, et al. Growth conditions and real structure of
N.V. Surovtsev et al. / Carbon 44 (2006) 2027–2031 synthetic diamond crystals. Russ Geol Geophys 1997;38(5):920– 45. [14] Pal’yanov YN, Sokol AG, Borzdov YM, Khokhryakov AF. Fluidbearing alkaline carbonate melts as the medium for the formation of diamonds in the Earth’s mantle: an experimental study. Lithos 2002;30(3–4):145–59. [15] Dresselhaus MS, Pimenta MA, Eklund PC, Dresselhaus G. Raman scattering in fullerenes and related carbon-based materials. In: Weber WH, Merlin R, editors. Raman scattering in material science, vol. 42. Berlin: Springer; 2000. p. 314–64.
2031
[16] Rao AM, Eklund PC, Hodeau J-L, Marques L, Nunez-Requeiro M. Infrared and Raman studies of pressure-polymerized C60s. Phys Rev B 1997;55(7):4766–73. [17] Matthews MJ, Pimenta MA, Dresselhaus G, Dresselhaus MS, Endo M. Origin of dispersive effects of the Raman D band in carbon materials. Phys Rev B 1999;59(10):R6585–8. [18] Gorelik VS, Igo AV, Mikov SN. Light Raman scattering in limited crystals. Zhurnal Experimentalnoi I Teoreticheskoi Fiziki 1996; 109(6):2141–9.