Fullerene molecules are formed in the laboratory from carbon-rich vapors which can be obtained in a variety of ways, e.g., resistive heating of carbon rods in a vacuum, ac or dc plasma discharge between carbon electrodes in He gas, laser ablation of carbon electrodes in He gas, and oxidative combustion of benzene/argon gas mixtures. Most methods for the production of large quantities of fullerenes simultaneously generate a mixture of stable fullerenes (C60, C70, ...), impurity molecules such as polyaromatic hydrocarbons, and carbon-rich soot. Therefore, the synthesis of fullerenes must be followed by procedures to extract and separate fullerenes from these impurities according to mass, and for the higher fullerenes, separation according to specific isomeric forms may also be required. In this chapter, we describe the laboratory methods commonly used to synthesize, extract, and purify fullerenes.
Fullerenes have also been reported to occur naturally in some forms of common carbon soot [5.1]; in shungite, a carbon-rich mineral [5.2]; and in fulgurite, which is formed by lightning strikes on carbon-containing rocks [5.3]. Ordinary materials such as gasoline [5.4] and coal [5.5] have also been reported as source materials for producing fullerenes. The field of fullerene synthesis, extraction, and purification methods is still developing rapidly. The most recent developments have also focused on the production and separation of higher-mass fullerenes and endohedral metallofullerenes, which are also discussed here. In this chapter we review the status of methods for the synthesis or generation of fullerenes (§5.1), extraction of fullerenes from fullerene-containing material (§5.2), and purification or separation of one fullerenes species from another (§5.3). A brief description of the preparation of endohedral fullerenes is reviewed in §5.4. A few comments on health and safety issues are given in §5.5.
Fullerenes can be synthesized in the laboratory in a wide variety of ways, all involving the generation of a carbon-rich vapor or plasma. All current methods of fullerene synthesis produce primarily C60 and C70, and these molecules are now routinely isolated in gram quantities and are commercially available. Higher-mass fullerenes and endohedral complexes can also be made and isolated, albeit in substantially reduced amounts. At present the most efficient method of producing fullerenes involves an electric discharge between graphite electrodes in ~200 torr of He gas. Fullerenes are embedded in the emitted carbon soot and must then be extracted and subsequently purified. A variation of the arc technique is used to synthesize graphene tubules (see §19.2.5). However, it appears that it will be difficult to extend the chemical methods now used to isolate particular fullerene isomers to separate the carbon tubules according to diameter and chiral angle (see §19.2.5).
Early approaches to fullerene synthesis used laser vaporization techniques, which produced only microscopic amounts of fullerenes, and these fullerenes in the gas phase were studied in an adjoining molecular beam apparatus (see §1.3). Using these techniques, early experiments by Kroto, Smalley, and co-workers in 1985 [5.6] indicated an especially high stability for C60 and C70 and stimulated a great deal of interest in the research community to find ways to produce and isolate macroscopic amounts of fullerenes. The breakthrough came in 1990 with the discovery by Kratschmer, Huffman, and co-workers [5.7] that carbon rods heated resis-tively in a helium atmosphere could generate gram quantities of fullerenes embedded in carbon soot, which was also produced in the process. This discovery, therefore, also necessitated research to discover efficient methods for the extraction of C60 and higher fullerenes from the carbon soot, as well as methods for the subsequent isolation of the extracted fullerenes according to mass and, in some cases, even according to their isomeric form (see §3.2).
An ac or dc arc discharge in ~200 torr He between graphite electrodes [5.8] has evolved as an efficient way to make gram quantities of fullerenes on the laboratory bench top. It is believed that fullerene formation requires a minimum gas pressure of ~25 torr [5.9], It has also been reported that an increase in He gas pressure enhances the production of higher-mass fullerenes [5.10], The heat generated in the discharge between the electrodes evaporates carbon to form soot and fullerenes, which both condense on the water-cooled walls of the reactor.
The arc discharge technique can also be used to make carbon nanotubes, if a dc, rather than an ac, voltage source is used to drive the discharge (see §19.2.5). Apparently, a variety of carbonaceous solids can be used as electrode materials in the arc process, and fullerenes have in fact been produced in this way from coal [5.5].
Laser vaporization is also used for fullerene production. In a typical apparatus a pulsed Nd:YAG laser operating at 532 nm and 250 mJ of power is used as the laser source and the graphite target is kept in a furnace (1200°C) [5.11-13]. Finally, it should be mentioned that fullerenes have also been produced in sooting flames involving, for example, the combustion of benzene and acetylene [5.14], although the yields are low.
The fact that fullerenes can be produced in electric arcs with temperatures well above 4000° C in inert atmospheres is testimony to their inherent stability. In experiments to produce fullerenes by high-frequency (500 kHz) inductive heating of carbon cylinders in an He atmosphere (150 kPa), it was noticed that although carbon begins to sublime from the cylinders at 2500° C, no fullerenes were produced until the cylinder temperature rose to 2700°C [5.15]. This temperature is noticeably higher than the flame temperature of ~1800°C used to produce fullerenes in premixed, low-pressure (~40 torr) benzene/oxygen/Ar flames [5.1], where about 3% of this gas mixture was converted to soot and about one third of the resulting soot was extractable as C60 and C70. Subsequent searches for fullerenes in other carbon soots and powders (e.g., diesel exhaust, carbon black, activated charcoal, copy machine toner and an acetylene black produced with a torch without oxygen admixture) did not reveal any fullerenes [5.1]. Interestingly, the combustion of benzene/oxygen/Ar mixtures was reported [5.1] to yield a ratio for (C70/C60)~ 0.4, about a factor of three times higher than obtained typically from the electric arc method. However, high current dc contact arc discharges between graphite electrodes in He gas with an estimated electrode tip temperature of ~4700°C have been reported to generate C70/C6() ratios as high as ~ 0.5, and a total yield of fullerenes of 7-10% was obtained [5.16], significantly higher than that reported for flames.
100 - 200 Amps
SS Heat shields
1/4" Braided flat Cu cable
Pyrex glass sleeve
Upper gravity fed electrode
Cu guide Lower electrode
100 - 200 Amps
SS Heat shields
1/4" Braided flat Cu cable
Pyrex glass sleeve
Upper gravity fed electrode
Fig. 5.1. Schematic diagram of a "contact-arc" apparatus used to produce fullerene-rich soot. Rather than incorporating a means of translating one or both of the carbon electrodes in the apparatus to maintain a fixed electrical discharge gap, the upper electrode in this design maintains contact with the lower electrode by the influence of gravity [5.17].
A simple, bench-top reactor used to produce fullerenes was reported by Wudl and co-workers [5.17], and yields about ~4% extractable C60/C70 from the soot. As shown in Fig. 5.1, electrical feedthroughs are introduced through a flange at the top of the chamber, and an inexpensive ac arc welding power supply is used to initiate and maintain a contact arc between two graphite electrodes in ~200 torr of He gas. The inert gas lines and a vacuum line are not shown in Fig. 5.1. The apparatus is lowered into a tank of water to cool the cylindrical wall in contact with the cooling water. The initial design by Wudl and co-workers [5.17] used a glass cylinder for the vacuum shroud. Although this provides a convenient means for visual inspection of the electrodes through the wall, the Pyrex cylinder is easily broken, and it is advisable to replace it with a stainless steel cylinder containing a window. Since the arc discharge produces ultraviolet (UV)
radiation harmful to the eyes and skin, a piece of filter glass from a welder's face shield is placed over the window if the arc process is to be observed.
A unique and useful feature of the design in Fig. 5.1 is that it does not require the electrodes to be translated together as the electrodes are consumed. Instead, the upper electrode is connected to the electrical feedthrough by a flexible braided cable, and gravity is used to maintain the contact between the electrodes automatically. Thus, this technique is termed the "contact arc method." Typically, about 1 g of 6-mm-diameter graphite electrodes is consumed in approximately 10 min at a current of 120 A (rms) [5.18], Several electrodes 10-20 cm in length are successively mounted in the chamber and consumed before the soot is scraped from the walls and removed for the extraction step described in §5.2. Normally, inexpensive graphite electrodes with ~1.1% natural abundance of 13C are used.
In Fig. 5.2 we show a scale diagram of a fullerene generator which uses a plasma arc in a fixed gap between horizontal carbon electrodes. A very high yield of 44% solvent extractables (26% by benzene and the rest by higher
boiling point solvents) was reported for this apparatus [5.19]. The authors attributed the high yield to (1) the presence of the arc gap, as compared, for example, with resistive heating (1-10%) [5.7,20,21] or a contact arc (4%) [5.17], and (2) a static He atmosphere, in which no fullerenes leave the chamber in flowing He gas. The term "static He atmosphere" means that the inlet and outlet gas valves are closed during the arc synthesis. A dc power supply was used in Fig. 5.2 to maintain the arc discharge between a translatable 6-mm-diameter graphite electrode (30 cm maximum length) on the left and a fixed 12-mm-diameter electrode on the right [5.19]. Once the arc is initiated, a gap of ~ 4 mm was maintained between electrodes, and this was judged to produce the maximum brightness in the arc. Six millimeters of electrode are consumed per minute in a static 200 torr He atmosphere for a constant current of 60 A; the arc gap determines the voltage developed across the arc (<20 V). The chamber is constructed from a conventional four-way 8-inch-diameter stainless steel cross with knife-edge seals, and the carbon soot generated in the arc is condensed onto a removable, water-cooled stainless steel cylinder. A good summary of the considerations pertinent to high-volume fullerene soot formation is given by Lamb and Huffman [5.9],
For the carbon arc discharge technique, several parameters are known to affect the conversion of the electrodes to fullerenes [5.9,21], Although fullerenes will form in a variety of inert gas atmospheres such as argon and molecular nitrogen, high-purity helium is usually used and is the preferred gas-quenching species. Most soot-generating chambers use a quenching-gas pressure in the 100-200 torr range. It is reported that the optimum pressure is highly sensitive to design specifications and should be determined for each soot generator. Although purity of the initial rod does not affect the soot production rate, higher purity rods are favored for producing soot with lower impurity content. Regarding the diameters, it is reported that smaller-diameter rods (e.g., 1/8 inch diameter) have higher yields of soot [5.9]. While it is believed that the dimensions, geometry, and convection of soot generators bear importantly on soot generation, these design parameters have not been optimized. Regarding contaminants in the generation chamber, small concentrations of hydrogen or HzO are sufficient to suppress fullerene generation seriously, so that initial degassing of the system and constituents is highly desirable. The removal of impurities is also important for reducing the formation of polycyclic aromatic hydrocarbons [5.22,23], which may be carcinogens. Other parameters affecting fullerene generation include the density, the binder present in the electrodes, the power dissipated in the arc, and the arc gap. Further research on the optimization of the arc parameters for the production of particular fullerenes, other than C60 and C70, needs to be carried out.
In the process of fullerene formation from the carbon vapors generated from various carbonaceous materials or hydrocarbon gases, insoluble nanoscale carbon soot is generated together with soluble fullerenes and potentially soluble impurity molecules. Two distinct methods are employed to extract the fullerenes from the soot. In the most common method, the so-called solvent method, toluene, or some other appropriate solvent, is used to dissolve the fullerenes primarily, but the solvent also brings along other soluble hydrocarbon impurities. The soot and other insolubles are therefore easily separated from this solution by filtration, or simply by decanting the solution. In a second, or "sublimation method," the raw soot containing the fullerenes is heated in a quartz tube in He gas, or in vacuum (~10-5 torr), to sublime the fullerenes, which then condense in a cooler section of the tube, leaving the soot and other nonvolatiles behind in the hotter section of the tube. Both of these extraction methods also have a tendency to bring impurity molecules along with the most stable fullerenes (e.g., C60 and C70). If a very pure fullerene solution or microcrystalline powder is desired, then a final chemical purification step must be carried out, which is discussed in the next section.
Regarding the separation of high-mass fullerenes, it is believed that as a by-product of the preparation of large amounts of C60 and C70 commercially, there will be available larger amounts of high-mass fullerenes. Improved separation methods hold promise for yielding isolated higher-mass (M) fullerene samples in a purity range M ± 2% [5.9].
"Soxhlet" extraction in a hot solvent was used in early research to remove fullerenes from the carbon soot generated in the synthesis step (see §5.1.2). More recent solvent techniques discussed below are more applicable to large batch processing. The Soxhlet technique is used in a variety of extractions where the molecular species to be removed from the solid phase are soluble in organic solvents, e.g., polyaromatic hydrocarbons (PAHs) from coal. A schematic view of a Soxhlet extraction unit is shown in Fig. 5.3. The solid sample containing fullerenes, soot, and other materials is loaded into a thimble in the center section of the apparatus. The solvent is boiled in the flask at the bottom, and as a result solvent vapors rise through the side arm on the left and condense near the bottom of the condenser unit, dripping hot, distilled solvent into the thimble, or series of concentric thimbles, which are permeable to the solution. The solution containing the extracted molecules returns to the boiling flask via the side arm on the right. This closed-loop system is normally operated for several hours, during which
Water inlet _
Water inlet _
51)0 ml flask
Magnetic stirrer heater
51)0 ml flask
Magnetic stirrer heater
Fig. 5.3. Schematic view of a Soxhiet extraction unit used to separate fullerenes from soot.
time the fullerenes that are soluble in the solvent collect in the flask, and the portion of the solid sample that is not soluble in the solvent remains in the thimble. A reportedly efficient alternative to this method has been proposed [5.24] in which the carbon soot is simply dispersed in tetrahydro-furan (THF) (10 ml/g) at room temperature and sonicated in an ultrasonic bath for 20 min. After filtration to remove the insoluble matter, the THF is removed from the fullerene extract in a rotary evaporator at 50° C, leaving a powder containing fullerenes and soluble impurities in the bottom of the evaporator flask. Some authors have advised that potential PAH impurities should then be eliminated by first washing the fullerene extract in diethyl ether prior to the subsequent liquid chromatography purification step, which is discussed below.
The success of every efficient method for the extraction and purification of fullerenes is tied to the solubility of the fullerene molecules in solvents (see Table 5.1 and §5.2.3). More polar and higher boiling point solvents have been reported to be selective in extracting higher-mass fullerenes. A sequence of Soxhlet extractions was employed [5.19], first using benzene to remove primarily C60/C70 (26%), then pyridine (4%), and finally 1,2,3,5-tetramethylbenzene or tetralin (14%), to achieve a total extracted yield of 44% from the carbon soot produced in the arc discharge apparatus shown in Fig. 5.2. Time-of-flight (TOF) mass spectroscopy was used to analyze the extracted material in the extraction sequence. It was found that the higher boiling point solvent, tetralin, isolates the higher-mass fullerenes, and it was proposed on the basis of mass spectroscopy data that fullerenes with masses up to C200 constitute over half of the soluble material extracted from the soot.
Microcrystalline C60 and C70 powders are known to sublime in vacuum at relatively low temperatures, i.e., Ts ~ 350°C (C60) [5.25] and Ts = 460°C (C70) [5.26], and this fact can be used directly to separate C60 and C70 from arc-generated carbon soot without introducing solvents such as hexane, benzene, carbon disulfide, or toluene (see §5.2.1). For some experiments which are particularly sensitive to solvent contamination in the samples, this technique might provide a useful alternative to solvent extraction. In this solvent-free approach [5.19,27], the soot is placed in one end of an evacuated and sealed (or dynamically pumped) quartz tube and the whole apparatus is then placed in a furnace with a temperature gradient. Dynamical pumping (i.e., the pump is left attached to the quartz tube during the sublimation process) is preferred, since it is possible that the soot may also contain polyaromatic hydrocarbons or other volatile impurities. The tube end containing the raw arc soot is maintained at the highest temperature T ~ 600-700° C, so that CgQ, C7q, and possibly higher-mass fullerenes will sublime from the soot and drift in the temperature gradient toward the colder regions of the tube, condensing there on the walls. C70 and the relatively small amounts of higher fullerenes (e.g., C76 and C84) will condense closer to the soot, since their sublimation temperatures are higher than that of C60.
The difference in sublimation temperatures between Cm and C70 has been used to produce a C60 molecular beam from a microcrystalline mixture of C60 and C70 [5.26]. In this case, the microcrystalline mixture of CM and C70 is first heated in a dynamic vacuum to ~ 250-300° C for an extended time (> 4-6 h) to drive off most of the residual solvent, and then the
Solubility of C60 in various solvents [5.29,30].
Solvent Parameter 8" Solubility
Was this article helpful?