Preparation and tribology properties of water-soluble fullerene derivative nanoball

1.1 Materials

Fullerene (C₆₀) was obtained from Wuhan University (purity 99%). Vinylpyrrolidone and benzoyl peroxide were purchased as commercial products from Acros. Vinylpyrrolidone was distilled under vacuum at 90 °C. Toluene was distilled at 110 °C. All other reagents and solvents such as dimethyl sulfoxide (DMSO) were used as received without further purification.

1.2 Preparation of water-soluble fullerene derivatives

A solution of 30–100 mg of C₆₀ in 30–100 mL of toluene or 1, 2-dichlorobenzene (or pyridine, dioxane) was mixed with 0.1–4.0 g of vinylpyrrolidone. The mixture was deoxygenated by bubbling dry nitrogen gas for ∼30 min, followed by the addition of 150 mg of BPO (benzoyl peroxide). The polymerization reaction was carried out in a glass bottle at 70–80 °C for 30 h. A change in color from violet (characteristic of C₆₀ in toluene) to red–brown was observed during the reaction. After solvent removal under reduced pressure, the residue was dissolved in THF, and unreacted C₆₀ was removed by centrifugation (5000 rpm, 30 min) and filtration through 0.45 μm membrane filters. The THF solution was concentrated under reduced pressure and subsequently poured into an excess of heptane for precipitating the copolymer (a solid sample with almost brown color). The copolymer was dissolved in water and was further purified by column chromatography (Sephadex G-25, distilled deionized water as eluent). After drying under vacuum, the final copolymer sample (water-soluble fullerene derivative) was obtained (purity 95%). A polymer of neat vinylpyrrolidone was prepared under the same conditions to be used as a reference. FTIR (KBr): ν 2959, 1660, 1458, 1422, 1295, 1120, 658, 530 (C60) cm-1; ¹H NMR (300 MHz, CDCl3, TMS): δ = 1.50–1.75 (m, 2H, CH2CH), 1.86–2.01 (m, 2H, CH2), 2.10–2.32 (m, 2H, CH2CO), 3.01–3.30 (m, 2H, CH2 N), 3.60–3.82 (m, 1H, CH) ppm; ¹³C NMR (75 MHz, CDCl3, TMS): δ = 17.2, 18.1, 31.8, 45.0, 94.2, 140–150 (C₆₀), 174(f) ppm.

1.3 Characterization

The molecular weights and their distributions of the copolymers were determined by gel permeation chromatography (GPC) utilizing a Waters model 515 pump and a model 2410 differential refractometer with three Styragel columns HT2, HT3, and HT4 connected in a serial fashion. THF was used as the eluent at a flow rate of 1.0 mL/min. Polystyrene standards with dispersity of 1.08–1.12 obtained from Waters were employed to calibrate the instrument. IR spectra were recorded with a Perkin–Elmer FT-IR spectrophotometer using KBr disks. ¹H NMR spectra were recorded in CDCl₃ at 20 °C on a Bruker AM 300 MHz apparatus. ¹³C NMR spectra were recorded in CDCl₃ at 20 °C on a Bruker AM 75 MHz apparatus. UV–vis absorption spectra were recorded on a Hitachi U-3300 spectrophotometer.

Thermogravimetric analyses (TGA) were carried out under nitrogen with a TA Instruments Q500 thermogravimetric analyzer in the 25–1000 °C range, at a 10 °C/min rate (HiRes method).

1.4 Self-assembly of water-soluble fullerene derivatives

The solvent-induced assembly of water-soluble fullerene derivatives was investigated by dissolving the fullerene derivatives in THF:water (80:20, v/v), filtering and dropping on a glass patch. After evaporation for 48 h under ambient conditions, the samples were observed by SEM.

1.5 Measurement of tribology properties

Since pure water has poor lubricity and is corrosive, a 2 wt.% of triethanolamine aqueous solution was chosen as base stock. Triethanolamine is widely used in the metal-working fluids as a multifunctional additive. The fullerene-containing polymer prepared above was used as a lubricant additive in the base stock. The tribological measurements were carried out by using an MQ-800 four-ball tribotester at a rotational speed of 1450 rpm and at a temperature of 25 °C. The maximum nonseizure load was obtained by GB3142–82, similar to ASTMD2783; the wear scar diameter was measured under a load of 200 N and test duration of 30 min; friction coefficient was measured under a load of 200 N and test duration of 10 s. The steel balls used in the tests were made of GCr15 (AISIE52100) bearing steel with the 64–66 surface HRC hardness and 0.012 μm of surface roughness R_a. In order to study the effect of C₆₀ in the C₆₀−vinylpyrrolidone copolymer, the same concentration of the polyvinylpyrrolidone was used as a reference additive and measured under the same conditions.

2.1 Synthesis of water-soluble fullerene derivatives

Water-soluble fullerene derivatives were prepared according to the procedure shown in Scheme 1. They are soluble in polar solvents such as water, acetone, tetrahydrofuran and dimethyl sulfoxide. The solubility of derivatives in water is approximately 5 mg/ml (calculated as C₆₀ residue), which is larger than that of the inclusion complex of cucurbit (Braun et al., 1994) uril and C₆₀. One reasonable explanation for the increase of solubility is that the intrinsic solubility of PVP is larger than that of cucurbit(Braun et al., 1994)uril. The covalent attachment of C₆₀ to the polyvinylpyrrolidone backbone was confirmed by a variety of techniques. FTIR spectra were measured in KBr matrices, and the results were compared with that of neat polyvinylpyrrolidone samples prepared under the same experimental conditions (Fig. 1). As shown in Fig. 1, observed FTIR spectrums of the copolymers with C₆₀ contents are different from that of neat polyvinylpyrrolidone. For the copolymers with C₆₀ contents, the FTIR spectrum shows extra absorption in the 520–530 cm⁻¹ region, which is typical with respect to substituted C₆₀ (Bunker et al., 1995). UV absorption spectra were measured in CH₃OH, and the results were compared with that of polyvinylpyrrolidone (Fig. 2). As shown in Fig. 2, the observed UV absorption spectra of copolymers with C₆₀ contents are very different from that of polyvinylpyrrolidone. The somewhat structured absorption band of free C₆₀ is replaced by a steadily decreasing curve, typical for substituted C₆₀ (Bunker et al., 1995). UV absorption wavelength maximum of copolymers with C₆₀ contents is 202 nm, different from that of free C₆₀ and polyvinylpyrrolidone, which is typical with respect to substituted C₆₀. The emission wavelength can be attributed to the C₆₀-containing sites in the polyvinylpyrrolidone structure. In the ¹H NMR spectrum of the copolymers with C₆₀ contents (Fig. 3), peaks corresponding to the CH₂ aliphatic protons adjacent to the ketone group of the vinylpyrrolidone and the other CH₂ aliphatic protons adjacent to the CH group of vinylpyrrolidone, were observed at 2.10–2.32 and 1.50–1.75 ppm, respectively. The peak related to the CH aliphatic protons from vinylpyrrolidone also appeared between 3.60 and 3.82 ppm. In the ¹³C NMR spectrum of the copolymers with C₆₀ contents (Fig. 4), peak corresponding to the C₆₀ was observed at 140–150 ppm (Guhr et al., 1994). The C₆₀−vinylpyrrolidone copolymers proved to be extremely soluble in a variety of organic solvents such as H₂O, CHCl₃, etc. This solubility behavior is similar to poly-vinylpyrrolidone itself and totally unlike C₆₀, which demonstrates the dramatic effect the polymer attachment has on the processability of fullerene derivatives. The weight percent of C₆₀ incorporated into the copolymers could be determined by TGA since the polyvinylpyrrolidone backbone decomposes at 230–280 °C to leave C₆₀ which undergoes only minor weight loss up to 600 °C. Comparison with starting materials and physical mixtures confirmed the TGA results. Interestingly the weight percent of C₆₀ was found to be ca. 80% of the theoretical value. The above results are all consistent, with monoaddition being the dominant process, and demonstrate the viability of this procedure for the preparation of copolymers with high loadings of C₆₀.

Close

Scheme 1

Synthesis of water-soluble C₆₀−vinylpyrrolidone copolymers.

Export to PPT

Close

Figure 1

FTIR spectra of C₆₀−vinylpyrrolidone copolymer (a) and polyvinylpyrrolidone (b).

Export to PPT

Close

Figure 2

Comparison of C₆₀−vinylpyrrolidone copolymer’s UV absorbance (a) with that of polyvinylpyrrolidone (b).

Export to PPT

Close

Figure 3

¹H NMR spectrum of the copolymers with C₆₀ (top) and vinylpyrrolidone (bottom) taken in CDCl₃.

Export to PPT

Close

Figure 4

¹³C NMR spectrum of the copolymers with C₆₀ taken in CDCl₃.

Export to PPT

The results presented here clearly show that C₆₀ and vinylpyrrolidone can be copolymerized under different conditions (Table 1). With a constant benzoyl peroxide amount, C₆₀ contents in the copolymers increase with increasing initial C₆₀: vinylpyrrolidone reactant ratio (Table 1). There is no apparent reduction in the polymer yield at high initial C₆₀: vinylpyrrolidone reactant ratios, which is probably due to relatively small [C₆₀]/[initiator] values. In addition, solubility in water of the copolymers prepared with different initial C₆₀: vinylpyrrolidone reactant ratios is significantly different (Table 1). In a classical mechanism of radical polymerization, propagation of ${\text{C}}_6{0}^{·}$ radicals is critical to the formation of true C₆₀−vinylpyrrolidone copolymers. However, it has been suggested that propagation of ${\text{C}}_6{0}^{·}$ radicals is slow in general because of their relatively high stability. It has also been suggested that C₆₀ competes effectively with vinylpyrrolidone (VP) for initiator radicals and, as a logical extrapolation, for VP^⋅ radicals as well. In this regard, C₆₀ may act as a radical inhibitor in the copolymerization reaction. Evidence for the inhibiting effect includes the observation that at high [C₆₀]/[initiator] ratios polymerization yields are low in comparison with those in neat vinylpyrrolidone polymerization. Thus, the mechanism of C₆₀−vinylpyrrolidone copolymerization may be different from the classical mechanism of radical polymerization of vinylpyrrolidone such that there is a significant population of C₆₀ radicals and their propagation is relatively slow. However, even with the suggested inhibiting effect of C₆₀ in the copolymerization reaction, contributions of polyvinylpyrrolidone structures are substantial even in the copolymers with high C₆₀ contents, as shown in the observed FT-IR spectra of the copolymers. Gel permeation chromatography (GPC) results indicate that the materials do not contain oligomers, and the C₆₀−vinylpyrrolidone copolymer and the polyvinylpyrrolidone reference have almost the same molecular weight distributions (weight- average molecular weight of 52,000–78,000 and polydispersity of 3–6). The results could be used to support a proposed copolymer mechanism as follows.

Close

Export to PPT

2.2 Self-assembly of water-soluble fullerene derivatives

The assembly behavior of water-soluble fullerene derivatives was investigated by SEM. The solvent-induced assembly of water-soluble fullerene derivatives was investigated by dissolving the fullerene derivatives in THF:water (80:20, v/v), filtering and dropping on a glass patch. After evaporation for 48 h under ambient conditions, the samples were observed by SEM. SEM images showed the macrostructure of water-soluble fullerene derivatives was spheres (Fig. 5a), which is different from the irregular morphologies found for polyvinylpyrrolidone (Fig. 5b). The average sphere size was calculated from the SEM picture and the sphere sizes were homogeneous. The samples have sphere sizes ranging from 30 nm to 80 nm. The quantity of water in the incorporated solution performs an important role in the assembly and morphological shifts in the fullerene derivatives. SEM analysis shows that the average particle diameter of fullerene derivative nanoball is about 55 nm.

Close

Figure 5

SEM images of C₆₀−vinylpyrrolidone copolymers (A) and polyvinylpyrrolidone (B).

Export to PPT

2.3 Tribology properties of water-soluble fullerene derivatives

The maximum nonseizure load (P_B value) represents the load-carrying capacity of the lubricant. The P_B value was measured and the results are given in Fig. 6(A). The P_B value of the base stock is 150 N, and the base stock with the C₆₀−vinylpyrrolidone copolymer showed a higher maximum nonseizure load than that of the base stock. In other words, the fullerene containing polymer could strengthen the load-carrying capacity of the base stock. When the fullerene derivative content reaches 0.3 wt.%, the P_B value is a maximum. Then, increased fullerene containing polymer resulted in a decrease in load-carrying capacity of the base stock.

Close

Figure 6

Effect of the fullerene copolymer content on: (A) maximum nonseized load; (B) wear scar diameter and (C) friction coefficient.

Export to PPT

The wear scar diameter data are given in Fig. 6(B). It is seen that the addition of the fullerene derivative can decrease the wear scar diameter of the base stock. When the fullerene containing polymer content reaches 0.3 wt.%, the wear scar diameter is minimum. But excessive fullerene copolymer content results in a larger wear scar diameter than that of the base stock. Decrease in load-carrying capacity and wear resistance at excessive fullerene-containing polymer content may be attributed to the corrosive wear, since the fullerene-containing polymer may react with the surface of the metal.

The friction coefficient measurements were carried out and the results are given in Fig. 6(C). It is indicated that with the increase of the additive content, the friction coefficient reduces rapidly from 0.233 of base stock to minimum 0.091 of that containing 0.8 wt.% fullerene copolymer.

The reduction in friction coefficient implies that the fullerene copolymer has the possibility to cause a microcosmic rolling effect between two rubbing surfaces as nanometer tiny ball. The dependence of wear scar diameter on load is shown in Fig. 7(A).

Close

Figure 7

Effect of load on: (A) wear scar diameter and (B) friction coefficient. (A) Base stock with 0.3 wt.% polyvinylpyrrolidone; (B) base stock with 0.3 wt.% C₆₀−vinylpyrrolidone copolymer.

Export to PPT

Under testing loads, the wear scar diameter of base stock with 0.3 wt.% fullerene copolymer is smaller than that with 0.3 wt.% polyvinylpyrrolidone. It means that, the presence of fullerene-containing polymer can strengthen the wear resistance of base stock.

Fig. 7(B) shows that, during the testing loads, the friction coefficient of base stock with 0.3 wt.% fullerene copolymer is lower than that with 0.3 wt.% polyvinylpyrrolidone, especially at the load of 200 N, the difference in friction coefficient between the two additives is larger, which may be attributed to the lower load-carrying capacity of the polyvinylpyrrolidone. The direct touch between the two friction surfaces then occurred, which in turn led to a rapid rise in friction coefficient.

The worn surface in four-ball machine testing, which was obtained under a load of 200 N and a testing time of 30 min, was observed by SEM. The wear scar of base stock with 0.3 wt.% polyvinylpyrrolidone and that with 0.3 wt.% fullerene-containing polymer are shown in Figs. 8 and 9, respectively. It indicated that the wear scar obtained with the fullerene-containing polymer additive is obviously smaller and exhibits mild scratches, but in comparison, larger wear scar and sharp tracks were observed in the presence of base stock with polyvinylpyrrolidone additive. In other words, the fullerene-containing polymer can improve the microcosmic wear condition.

Close

Figure 8

Morphology of the wear scar lubricated by base stock with 0.3 wt.% polyvinylpyrrolidone (100x).

Export to PPT

Close

Figure 9

Morphology of the wear scar lubricated by base stock with 0.3 wt.% C₆₀−vinylpyrrolidone copolymer (100x).

Export to PPT

Based on the above lubrication measurements and worn surface analyses, the lubrication mechanism of the fullerene-containing polymer additive can be deduced. The fullerene-containing polymer plays the role of a solid lubricant. Since fullerene has a very high load-carrying capacity and the fullerene copolymer is nanometer tiny balls with core–shell structure, which can penetrate into rubbing surfaces and deposit there, it is reasonable to speculate that the fullerene copolymer nanometer balls may be more effective than polyvinylpyrrolidone to support and isolate two relative motion surfaces, and therefore, the load-carrying capacity and anti-wear performance of the base stock would be improved. Moreover, the nanometer balls are expected to roll between two relative motion surfaces to reduce the friction coefficient. In addition, the shell is a polymer of neat vinylpyrrolidone, which is relatively soft but very elastic, thus improving the microcosmic condition of wear.