Fig nbsp xA TG and DTG curves for

It should be pointed out that the electronic spectra of individual light fullerenes in 1-hexanol (Fig. 3) demonstrate that there are no solvatochromic effects, i.e., no strong change in the spectrum induced by the variation of the solution concentration or solvent composition (in the case of using binary solvent mixtures) [1] and [2]. Therefore, there is a full analogy among the spectra we recorded in 1-hexanol and the corresponding spectra in a series of one-component aromatic solvents (o-xylene, benzene, toluene, and o-dichlorobenzene), so that the use of empirical relationships (2) and (3) is quite reasonable in our case. Eqs.  (2) and (3) are a result of the Firordt method application to o-xylene solutions of fullerenes [1] and [2]:equation(2)CC60=0.0131D335−1.808D472equation(3)C(C70)=0.0425(D472−0.0081D335)CC70=0.0425D472−0.0081D335where D335 and D472 are the optical densities of the mixtures referred to the SNS-314 layer of 1 cm width, and C(C70) and C(C60) represent the corresponding fullerene concentration (g · l− 1). 1-Hexanol mixtures were preliminarily diluted, the reference system being pure 1-hexanol. Negligible admixtures of the heavy fullerenes with carbon numbers higher than 70 were ignored. Table 1 shows the new solubility values determined with the spectrophotometer together with literature data of individual light fullerenes (C60 and C70) as well as the industrial fullerene mixture in 1-hexanol. Analysis of new and literature data reveals that: (i) the experimental data on solubility of individual light fullerenes obtained previously in our scientific group [32] and [33] agree with the new values. The deviations of the solubility values are due to the recalculation of the experimental data obtained in volume concentration (see ref. [33]) into mass fractions (we did not take into account the densities of the saturated solutions, but those of the pure 1-hexanol) as well as to the different conditions of saturation — the procedure described in ref. [32] and [33] included saturation in ampoules using thermostatic shaker, whereas in this study we have used a magnetic stirrer; moreover, the swept volume was different, as in the present study we used small quantities (up to 5 ml) of the solvent. The latter fact leads to the decreasing of the solubility determination accuracy; (ii) we can see some deviation between the solubility values for the binary system C70–1-hexanol at 298.15 K measured in the current study and the one obtained by Heyman [12] and [35]. We suppose that this is due to different conditions of experiment; in the last case the authors did not use the constant shaking of the heterogeneous mixtures, and added new portions of the solvent to the system. Thus, the values presented in ref. [12] and [35] are not equilibrium solubility values of individual light fullerenes. Several independent conditions must be satisfied in obtaining correct solubility values: (1) the solid phase should reach an equilibrium by itself (it is necessary to have sufficient time for the recrystallization and redistribution of fullerene components between surface and volume layers of fullerene solid solution crystals); (2) the solid phase should be an equilibrium fullerene crystal solvate (experimental duration should be sufficient for the solvation of the solid phase); (3) the liquid phase should be a liquid solution of fullerenes at equilibrium (the duration of saturation should be sufficient for the complete organization of the fullerene liquid solution including the formation of the superstructure); (iii) analysis of experimental data in Table 1 shows that solubility of C70 is always higher than solubility of the C60 fullerene. The latter fact is caused by higher polarizability of C70 due to the lower symmetry of the molecule structure  [1]; (iv) analysis of the experimental data devoted to the solubility of the industrial fullerene mixture (39% C70, 60% C60 and 1% Сn (n > 70)) ( Table 1) shows that in the temperature range of 293.15–303.15 K the liquid phase is enriched by the C70 fullerene (the enrichment is equal to 10–15% in comparison with the composition of the industrial fullerene mixture). A method of pre-chromatographic separation of the industrial fullerene mixtures can be developed based on such a difference in the content of the fullerene components in liquid and solid phases. We can also mention that the solubility of the fullerene mixture in 1-hexanol is higher than solubility of individual fullerenes due to the salting-in effect [1].