- Email: [email protected]
The vibrational Raman spectra of C60 and C70
0584-8539191s3.OO+o.al Q 1991 Petgamotl Press pk
Spccnochimifa Acta, Vol. 47A, No. 9/10, pp. 1289-1292.1991 F’rintedin Great Britain
The vibrational Raman spectra of Co and Cm T. JOHN DENNIS,JONATHAN P. HARE, HAROLDW. KROTO,
ROGERTAYLORand DAVID R. M. WALTON School of Chemistry and Molecular Sciences, University of Sussex, Brighton BNl9QJ, Sussex, U.K.
and PATRICK J. HENDRA Chemistry Department, Southampton University, Southampton So9
SNH,U.K.
(Received 10 April 1991; in final form 23 May 1991; accepted 24 May 1991) Ahstraet-Vibrational Raman spectra have been obtained at 1 cm-’ resolution, for C, and Cm For C, the positions of the bands were in good agreement with theoretical predictions. For C, 27 bands were observed out of the total of 53 that were theoretically predicted. These data will permit refinement of calculations of vibrational lines appropriate to closed-cage carbon clusters. The technique for producing appreciable quantities of high purity C, is described.
INTRODUCTION THE isolation, separation, and characterization of the closed-cage structures Cm and C, [l, 21 has initiated unprecedented activity aimed at investigating the chemical and physical properties of these novel molecules. Amongst the latter properties the NMR [2], UV/vis [2-41, and IR [2,5,6] spectra have been reported. Vibrational Raman spectra have been obtained for mixtures of Cso and Cn, and some tentative assignments were made, based on the differing intensities of lines in samples containing different concentrations .of Cx, [7]. Theoretical spectra were reported by several workers, prior to the availability of experimental data [8-141. Our access to both pure C, and G, and Fourier transform high resolution Raman spectrometry, made it appropriate to obtain good quality spectra so that the calculations may be subsequently refined, and parameters appropriate to closed-cage structures deduced.
EXPERIMENTAL Pure Co, was made as previously described [2,3] and was entirely tree from any tracesof C,,,. For the present work, substantialquantities of pure CN)(the minor component of the soot extract)were required. Attempts to scale up our previously described procedure [2,3] were not wholly
successful, and led to C, of cu 90% purity, the remainder being Cso. This was due to three factors. First, deposition of larger quantities of the C&Cm mixture on silica at the top of the column resulted in not all of the fullerenes dissolving in the hexane eluent within a reasonable time. Traces of C, (invisible) were thus being leached into the hexane as the C, fraction was being developed. Secondly, use of anhydrous alumina as the stationary phase resulted in very substantial on-column decomposition of C,,, to give a product which remained very firmly bound to the alumina. Thirdly, the fairly strong binding of Cn, to the column and the need to strip the column reasonably rapidly in order to minimize the above decomposition necessitated elution of this fraction with carbon tetrachloride. However, this procedure also stripped the column of bound traces of C, which, at the concentration levels being employed, became significant. Two strategies were developed to overcome these difficulties. First, it was found that use of neutral alumina containing 0.5% water, whilst still giving very satisfactory separation, resulted in negligible decomposition, more rapid elution, and less contamination from C, on final stripping of the column with carbon tretrachloride; this result could be attributed to the weaker binding between C, and the stationary phase. Secondly, the problem of incomplete solution of the fullerene in hexane was overcome by carrying out the separation in a manner orthogonal to that which is normally regarded as good practice: the extract from soot, which consists mainly of 1289
T. JOHN DENNIS et al.
1290 Hi
A~
Hi
Hi
Hi
“9
H9
A9
l-
I #I’
:
:c \ I
.’
i
:
I
I
t
+!~..~,.,~z 1600
1400
1200
000
4F
600
Roman shift km-‘) Fig. 1. Raman vibrational
spectrum for C, showing provisional obtained in Ref. [9].
assignments
based on the values
was heated under reflux with hexane for 2 h. The cooled solution was then filtered and passed through the column in the usual way. In this manner approximately 1.5 1 of hexane extract could be processed through a column measuring 500 x 15 mm i.d., producing C,,, of 97.5% average purity, as confirmed by NMR analysis. This procedure was repeated a number of times in order to obtain the required amount of material. The hexane extracts were concentrated to dryness using a rotary evaporator, redissolved in a little benzene, and the benzene extract concentrated to dryness in a 5 dram vial; this procedure was also adopted for Cso Because both materials tenaciously retain benzene, each extract was then heated at 170°C at 3 mm Hg for 2 h, and this removed any other volatile material that may have been concentrated from the hexane. [The hexane (HPLC grade) was continuously recycled using
fullerenes,
‘I
rl
,,,,,,, I600
1400
I200
1000
600
600
Raman shift km-‘) Fig. 2. Raman vibrational
spectrum for G.
?rx’/ 400
2
The vibrational Raman spectra of C, and C,
1291
Table 1. Calculated and observed Raman vibrational frequencies for C,
Ref.
PI
[91
218 435 388 654 743 1004
284 463 660 730 830 1325
1221 1468
1442 1696
1409
1798
1601
1816
Calculated (cm-‘)* PO1 1111 [I21
1131
P41
Observed (cm-‘)? P51
274 413
272 (266) 429 495 (485) 570,711$ 772
1099 1250 1422 1467 (1479) 1573.5
272 428 548 552 780
263 447 610 771 924
332 410 422 676 764
510
526 828
258 440 513 691 801
1160 1399 1688 1627 1831
1261 1407 15% I667 1722
937 1149 1300 1423 1377
1292 1575 1910 1875 2085
1154 1265 1465 1442 1644
* Italicized values are for A, symmetry; all others arc for H8 symmetry. t Values in parentheses are for side bands. $ This is a very weak feature. grease-free equipment, not only on cost grounds but also because this minimized the possibility of concentrating high-boiling impurities; it must be borne in mind in this work that the solubility of the fullerenes .is comparable to the impurity levels found in many solvents.] The solids so obtained were packed into a 1 mm diameter hole drilled through the centres of separate 10 mm (diameter) X 25 mm brass cylinders. The material (cu 3 mm length) was held in place by 1 mm stainless steel wires that were a sliding fit. Just prior to obtaining the spectra, the material was gently eased gush to the surface of one end of the cylinder, and placed so as to face the laser beam in the cell compartment of the FI-Raman spectrophotometer (Perkin-Elmer FT IR/Raman Model 1720). To obtain the Raman spectrum of a black solid is potentially very difficult because of heating by the laser beam, and this caused problems in preliminary experiments, especially with the more volatile and darker C&. However, by reducing the laser power to a very low value of 20 mW for Cm and 40 mW for Cr,,, quite excellent spectra were obtained at 1 cm-’ resolution (Figs 1 and 2). The quality of the spectra obtained at these very low power levels shows that these fullerenes are very strong Raman scatterers.
DISCUSSION
The spectra which we have now obtained are sufficiently detailed to encourage theoreticians to evaluate and refine with some confidence calculations of vibrational spectra for these closed-cage structures. Some indication of the difficulty with existing data is indicated by the results for Cm in Table 1, obtained by seven different groups. A measure of consistency is present in that each predicts that 10 bands should be observed, but there is some disagreement as to the relative frequencies of the two symmetries. Experimentally we discern a total of 14.bands, but one of these is very weak and another two, which occur at the sides of strong bands, may be due to resonance. Of the calculated values, overall those of NECHUet al. [14] correlate most closely with our data, and are Table 2. Observed Raman vibrational frequencies for C, Frequency (cm-‘) 226 252 259 394 410 417 430 456 489
I S
m S W W W W VS VW
Frequency (cm-‘) 520 534 547 566 700 736 1061
1180 1221
I VW VW VW W
m m s ms W
Frequency (cm-‘) 1226 1294 1329 1348 1365 1423 1465 1509 1562
I s W VW VW VW
m m mw m
1292
T.
JOHN
DENNIS eral.
depicted on Fig. 1. Those of WV et al. [lo] are particularly good at lower wave numbers, but overestimate the values at higher wave numbers. The presence of the two sharp bands at 266 and 272cm-’ is a curious feature. Interestingly, we note that in the less well-resolved Raman spectrum of partially pure C, obtained by BETHUNE et al. [7], a shoulder was clearly visible on the band which they observed at 273 cm-‘. Possibly this band is due to the presence of traces of an oxygen derivative of C, (formed quite readily in air [15] and possibly accelerated by laser heating). An interesting difference between the BETHUNEet al. [7] results and ours is that we find the 272 cm-’ band to be more intense than the 495 cm-’ band, whereas they found the reverse. This may be a function of the resolution, which was somewhat higher in our spectra. In the absence of polarization data and reliable intensity calculations it is not yet possible to unequivocally assign the bands that we have obtained. It seems reasonably probable, however, that the 1467 cm-’ band is due to the A, cage-breathing mode, whilst the 495 cm-’ band is due to the A, cage-squashing mode. For C, there have been two calculations of the positions of the Raman bands, and both agree in the expectation that some 53 bands should be observable [9,11]. We have detected 27 (Table 2 and Fig. 2) and some of these are very weak. At this time, however, there seems no advantage in attempting to correlate our results with calculations which are likely to be less reliable than those for Cboin view of the greater difficulty of the computations for C 70. Once calculations are able to accurately predict the results for C,,, the parameters employed should then produce meaningful data for C, (and other fullerenes) . There are significant differences between our spectrum for C, and that obtained by BETHUNEet al. [7]. Part of this is due to the presence of C, in their sample. Nevertheless, the strong/very strong peaks at 226 and 456 cm-‘, respectively, are absent in their spectrum, which also shows the peak at 1568 cm-’ as the most intense, whereas it is the fifth largest in our spectrum. Acknowledgements-We information.
are pleased to acknowledge the help of Dr
SYDNEY
LEACH in providing valuable
REFERENCES [l] [2] [3] [4) [5] [6] [7] [S] [9] [lo] [ll] [12] [13] [14] [15]
W. Kratschmer, L. D. Lamb, K. Fostiropoulos and D. R. Huffman, Nature 347, 354 (1990). R. Taylor, J. P. Hare, A. K. Abdul-Sada and H. W. Kroto, J. Chem. Sot., Chem. Commun. 1423 (1990). J. P. Hare, H. W. Kroto and R. Taylor, C/tern. Phys. Lett. 177,394 (1991). H. Ajie, M. M. Alvarez, S. J. Anz, R. D. Beck, F. Diederich, K. Fostiropoulos, D. R. Huffman, W. Kr&schtner, Y. Rubin, K. E. Schriver, D. Sensharma and R. L. Whetten, J. Phys. C/rem. 94,8639 (1998). D. S. Bethune, G. Meijer, W. C. Tang, H. J. Rosen, W. G. Golden, H. Seki, C. A. Brown and M. S. de Vries, C/rem. Phys. Lett. in press (1991). J. P. Hare, T. J. Dennis, H. W. Kroto, R. Taylor, A. W. Allaf, S. Balm and D. R. M. Walton, J. Chem. Sot., Chem. Commun. 412 (1991). D. S. Bethune, G. Meijer, W. C. Tang and H. J. Rosen, Chem. Phys. Lett. 174,219 (1990). S. J. Cyvin, E. Brendsdai, B. N. Cyvin and J. Brunvoll, Chem. Phys. Lett. 143,377 (1988). Z. Slanina, J. M. Rudzinski, M. Togasi and E. Osawa, J. Molec. Struct. Theo&em. 202,169 (1989). Z. C. Wu, D. A. Jelski and T. F. George, Gem. Phys. Lett. 137,291 (1987). R. E. Stanton and M. D. Newton, J. Phys. C/rem. 92,214l (1988). C. Coulumbeau and A. Rassat, J. C/rim. Phys. 84, 875 (1987). D. E. Weeks and W. G. Harter, J. C/rem. Phys. 90,4744 (1989). F. Negri, G. Orlandi and F. Zerbetto, Chem. Phys. Lett. 144, 31 (1988). R. Taylor, J. P. Parsons, A. G. Avent, S. P. Rannard, T. J. Dennis, J. P. Hare, H. W. Kroto and D. R. M. Walton, Nature 351,277 (1991).
Spccnochimifa Acta, Vol. 47A, No. 9/10, pp. 1289-1292.1991 F’rintedin Great Britain
The vibrational Raman spectra of Co and Cm T. JOHN DENNIS,JONATHAN P. HARE, HAROLDW. KROTO,
ROGERTAYLORand DAVID R. M. WALTON School of Chemistry and Molecular Sciences, University of Sussex, Brighton BNl9QJ, Sussex, U.K.
and PATRICK J. HENDRA Chemistry Department, Southampton University, Southampton So9
SNH,U.K.
(Received 10 April 1991; in final form 23 May 1991; accepted 24 May 1991) Ahstraet-Vibrational Raman spectra have been obtained at 1 cm-’ resolution, for C, and Cm For C, the positions of the bands were in good agreement with theoretical predictions. For C, 27 bands were observed out of the total of 53 that were theoretically predicted. These data will permit refinement of calculations of vibrational lines appropriate to closed-cage carbon clusters. The technique for producing appreciable quantities of high purity C, is described.
INTRODUCTION THE isolation, separation, and characterization of the closed-cage structures Cm and C, [l, 21 has initiated unprecedented activity aimed at investigating the chemical and physical properties of these novel molecules. Amongst the latter properties the NMR [2], UV/vis [2-41, and IR [2,5,6] spectra have been reported. Vibrational Raman spectra have been obtained for mixtures of Cso and Cn, and some tentative assignments were made, based on the differing intensities of lines in samples containing different concentrations .of Cx, [7]. Theoretical spectra were reported by several workers, prior to the availability of experimental data [8-141. Our access to both pure C, and G, and Fourier transform high resolution Raman spectrometry, made it appropriate to obtain good quality spectra so that the calculations may be subsequently refined, and parameters appropriate to closed-cage structures deduced.
EXPERIMENTAL Pure Co, was made as previously described [2,3] and was entirely tree from any tracesof C,,,. For the present work, substantialquantities of pure CN)(the minor component of the soot extract)were required. Attempts to scale up our previously described procedure [2,3] were not wholly
successful, and led to C, of cu 90% purity, the remainder being Cso. This was due to three factors. First, deposition of larger quantities of the C&Cm mixture on silica at the top of the column resulted in not all of the fullerenes dissolving in the hexane eluent within a reasonable time. Traces of C, (invisible) were thus being leached into the hexane as the C, fraction was being developed. Secondly, use of anhydrous alumina as the stationary phase resulted in very substantial on-column decomposition of C,,, to give a product which remained very firmly bound to the alumina. Thirdly, the fairly strong binding of Cn, to the column and the need to strip the column reasonably rapidly in order to minimize the above decomposition necessitated elution of this fraction with carbon tetrachloride. However, this procedure also stripped the column of bound traces of C, which, at the concentration levels being employed, became significant. Two strategies were developed to overcome these difficulties. First, it was found that use of neutral alumina containing 0.5% water, whilst still giving very satisfactory separation, resulted in negligible decomposition, more rapid elution, and less contamination from C, on final stripping of the column with carbon tretrachloride; this result could be attributed to the weaker binding between C, and the stationary phase. Secondly, the problem of incomplete solution of the fullerene in hexane was overcome by carrying out the separation in a manner orthogonal to that which is normally regarded as good practice: the extract from soot, which consists mainly of 1289
T. JOHN DENNIS et al.
1290 Hi
A~
Hi
Hi
Hi
“9
H9
A9
l-
I #I’
:
:c \ I
.’
i
:
I
I
t
+!~..~,.,~z 1600
1400
1200
000
4F
600
Roman shift km-‘) Fig. 1. Raman vibrational
spectrum for C, showing provisional obtained in Ref. [9].
assignments
based on the values
was heated under reflux with hexane for 2 h. The cooled solution was then filtered and passed through the column in the usual way. In this manner approximately 1.5 1 of hexane extract could be processed through a column measuring 500 x 15 mm i.d., producing C,,, of 97.5% average purity, as confirmed by NMR analysis. This procedure was repeated a number of times in order to obtain the required amount of material. The hexane extracts were concentrated to dryness using a rotary evaporator, redissolved in a little benzene, and the benzene extract concentrated to dryness in a 5 dram vial; this procedure was also adopted for Cso Because both materials tenaciously retain benzene, each extract was then heated at 170°C at 3 mm Hg for 2 h, and this removed any other volatile material that may have been concentrated from the hexane. [The hexane (HPLC grade) was continuously recycled using
fullerenes,
‘I
rl
,,,,,,, I600
1400
I200
1000
600
600
Raman shift km-‘) Fig. 2. Raman vibrational
spectrum for G.
?rx’/ 400
2
The vibrational Raman spectra of C, and C,
1291
Table 1. Calculated and observed Raman vibrational frequencies for C,
Ref.
PI
[91
218 435 388 654 743 1004
284 463 660 730 830 1325
1221 1468
1442 1696
1409
1798
1601
1816
Calculated (cm-‘)* PO1 1111 [I21
1131
P41
Observed (cm-‘)? P51
274 413
272 (266) 429 495 (485) 570,711$ 772
1099 1250 1422 1467 (1479) 1573.5
272 428 548 552 780
263 447 610 771 924
332 410 422 676 764
510
526 828
258 440 513 691 801
1160 1399 1688 1627 1831
1261 1407 15% I667 1722
937 1149 1300 1423 1377
1292 1575 1910 1875 2085
1154 1265 1465 1442 1644
* Italicized values are for A, symmetry; all others arc for H8 symmetry. t Values in parentheses are for side bands. $ This is a very weak feature. grease-free equipment, not only on cost grounds but also because this minimized the possibility of concentrating high-boiling impurities; it must be borne in mind in this work that the solubility of the fullerenes .is comparable to the impurity levels found in many solvents.] The solids so obtained were packed into a 1 mm diameter hole drilled through the centres of separate 10 mm (diameter) X 25 mm brass cylinders. The material (cu 3 mm length) was held in place by 1 mm stainless steel wires that were a sliding fit. Just prior to obtaining the spectra, the material was gently eased gush to the surface of one end of the cylinder, and placed so as to face the laser beam in the cell compartment of the FI-Raman spectrophotometer (Perkin-Elmer FT IR/Raman Model 1720). To obtain the Raman spectrum of a black solid is potentially very difficult because of heating by the laser beam, and this caused problems in preliminary experiments, especially with the more volatile and darker C&. However, by reducing the laser power to a very low value of 20 mW for Cm and 40 mW for Cr,,, quite excellent spectra were obtained at 1 cm-’ resolution (Figs 1 and 2). The quality of the spectra obtained at these very low power levels shows that these fullerenes are very strong Raman scatterers.
DISCUSSION
The spectra which we have now obtained are sufficiently detailed to encourage theoreticians to evaluate and refine with some confidence calculations of vibrational spectra for these closed-cage structures. Some indication of the difficulty with existing data is indicated by the results for Cm in Table 1, obtained by seven different groups. A measure of consistency is present in that each predicts that 10 bands should be observed, but there is some disagreement as to the relative frequencies of the two symmetries. Experimentally we discern a total of 14.bands, but one of these is very weak and another two, which occur at the sides of strong bands, may be due to resonance. Of the calculated values, overall those of NECHUet al. [14] correlate most closely with our data, and are Table 2. Observed Raman vibrational frequencies for C, Frequency (cm-‘) 226 252 259 394 410 417 430 456 489
I S
m S W W W W VS VW
Frequency (cm-‘) 520 534 547 566 700 736 1061
1180 1221
I VW VW VW W
m m s ms W
Frequency (cm-‘) 1226 1294 1329 1348 1365 1423 1465 1509 1562
I s W VW VW VW
m m mw m
1292
T.
JOHN
DENNIS eral.
depicted on Fig. 1. Those of WV et al. [lo] are particularly good at lower wave numbers, but overestimate the values at higher wave numbers. The presence of the two sharp bands at 266 and 272cm-’ is a curious feature. Interestingly, we note that in the less well-resolved Raman spectrum of partially pure C, obtained by BETHUNE et al. [7], a shoulder was clearly visible on the band which they observed at 273 cm-‘. Possibly this band is due to the presence of traces of an oxygen derivative of C, (formed quite readily in air [15] and possibly accelerated by laser heating). An interesting difference between the BETHUNEet al. [7] results and ours is that we find the 272 cm-’ band to be more intense than the 495 cm-’ band, whereas they found the reverse. This may be a function of the resolution, which was somewhat higher in our spectra. In the absence of polarization data and reliable intensity calculations it is not yet possible to unequivocally assign the bands that we have obtained. It seems reasonably probable, however, that the 1467 cm-’ band is due to the A, cage-breathing mode, whilst the 495 cm-’ band is due to the A, cage-squashing mode. For C, there have been two calculations of the positions of the Raman bands, and both agree in the expectation that some 53 bands should be observable [9,11]. We have detected 27 (Table 2 and Fig. 2) and some of these are very weak. At this time, however, there seems no advantage in attempting to correlate our results with calculations which are likely to be less reliable than those for Cboin view of the greater difficulty of the computations for C 70. Once calculations are able to accurately predict the results for C,,, the parameters employed should then produce meaningful data for C, (and other fullerenes) . There are significant differences between our spectrum for C, and that obtained by BETHUNEet al. [7]. Part of this is due to the presence of C, in their sample. Nevertheless, the strong/very strong peaks at 226 and 456 cm-‘, respectively, are absent in their spectrum, which also shows the peak at 1568 cm-’ as the most intense, whereas it is the fifth largest in our spectrum. Acknowledgements-We information.
are pleased to acknowledge the help of Dr
SYDNEY
LEACH in providing valuable
REFERENCES [l] [2] [3] [4) [5] [6] [7] [S] [9] [lo] [ll] [12] [13] [14] [15]
W. Kratschmer, L. D. Lamb, K. Fostiropoulos and D. R. Huffman, Nature 347, 354 (1990). R. Taylor, J. P. Hare, A. K. Abdul-Sada and H. W. Kroto, J. Chem. Sot., Chem. Commun. 1423 (1990). J. P. Hare, H. W. Kroto and R. Taylor, C/tern. Phys. Lett. 177,394 (1991). H. Ajie, M. M. Alvarez, S. J. Anz, R. D. Beck, F. Diederich, K. Fostiropoulos, D. R. Huffman, W. Kr&schtner, Y. Rubin, K. E. Schriver, D. Sensharma and R. L. Whetten, J. Phys. C/rem. 94,8639 (1998). D. S. Bethune, G. Meijer, W. C. Tang, H. J. Rosen, W. G. Golden, H. Seki, C. A. Brown and M. S. de Vries, C/rem. Phys. Lett. in press (1991). J. P. Hare, T. J. Dennis, H. W. Kroto, R. Taylor, A. W. Allaf, S. Balm and D. R. M. Walton, J. Chem. Sot., Chem. Commun. 412 (1991). D. S. Bethune, G. Meijer, W. C. Tang and H. J. Rosen, Chem. Phys. Lett. 174,219 (1990). S. J. Cyvin, E. Brendsdai, B. N. Cyvin and J. Brunvoll, Chem. Phys. Lett. 143,377 (1988). Z. Slanina, J. M. Rudzinski, M. Togasi and E. Osawa, J. Molec. Struct. Theo&em. 202,169 (1989). Z. C. Wu, D. A. Jelski and T. F. George, Gem. Phys. Lett. 137,291 (1987). R. E. Stanton and M. D. Newton, J. Phys. C/rem. 92,214l (1988). C. Coulumbeau and A. Rassat, J. C/rim. Phys. 84, 875 (1987). D. E. Weeks and W. G. Harter, J. C/rem. Phys. 90,4744 (1989). F. Negri, G. Orlandi and F. Zerbetto, Chem. Phys. Lett. 144, 31 (1988). R. Taylor, J. P. Parsons, A. G. Avent, S. P. Rannard, T. J. Dennis, J. P. Hare, H. W. Kroto and D. R. M. Walton, Nature 351,277 (1991).