In the OSTI Collections: Fullerenes

Dr. Watson computer sleuthing scientist

Article Acknowledgement:

Dr. William N. Watson, Physicist

DOE Office of Scientific and Technical Information

Separator line

The spherelike 60-atom carbon molecule named “buckminsterfullerene” (after Buckminster Fuller[Wikipedia], whose geodesic domes[Wikipedia] the molecule resembles) is just one of a larger class of molecules known simply as “fullerenes”.  This class definitely includes other spherelike carbon molecules with more or fewer carbon atoms.  All these molecules have neighboring carbon atoms bonded together to form adjoining rings—mostly hexagonal rings except in the smallest molecules, but always including some pentagonal rings, without which the molecules could not resemble spheres, since then they couldn’t curve without strain along more than one direction.  A molecule with only hexagonal rings of carbon atoms may curve in one direction to form a tube.  Tubular carbon molecules are sometimes also called fullerenes, and in any case (if they have nanometer-scale widths) are known as carbon nanotubes.  “Fullerenes” may also be thought of as including nested carbon molecules, with smaller spherelike or tubular layers of carbon atoms within larger ones.  With this usage, one could say that fullerenes are carbon molecules with one- or two-dimensional curvature.  The other known forms of pure carbon—graphite (multiple flat layers of carbon atoms in only hexagonal rings), graphene (similar to graphite except for being a single layer of carbon atoms), and diamond—don’t curve along any dimensions unless they’re under stress. 


As interesting as fullerenes’ structures are, the physical properties that these structures give them can be even more interesting.  The arrangement of a molecule’s atoms affects how the electrons at its surface are arranged, which largely determines (1) the molecule’s mechanical and chemical interactions with other molecules and their outermost electrons, and (2) how the molecule interacts with light, x-rays, and other electromagnetic waves.  These interactions enable materials that are partly or entirely composed of fullerenes to perform various functions, among which are the transformation of energy into more useful forms.  Several efforts to produce useful fullerenes have been sponsored by the U. S. Department of Energy, as exemplified by the reports described below. 


Figure 1

Figure 1.  Two schematic models of a 60-atom buckminsterfullerene molecule.  On the left, a “ball-and-stick” model showing the relative positions of the carbon atoms and the chemical bonds among them when the molecule is in its lowest-energy state.  On the right, a “space-filling” model that roughly indicates the volume taken up by the carbon atoms.  While both types of model are commonly used to understand molecular structures, it’s important to recognize that they’re only rough indicators of a molecule’s shape.  A molecule’s atoms, which consist of nuclei surrounded by swarms of electrons, do not have sharply-defined boundaries and of course are not actually joined by sticks. 




Separator line

Material-assembly, optical, and electrical properties


Atomic arrangement affects not only the properties of single molecules, but those of larger pieces of material as well.  Thus the relative positions of a composite material’s different molecules determine what properties the composite will have.  But a fabrication problem exists if desirable composite properties depend on the components’ molecules mixing together while at least one set of molecules naturally tends to stick together instead of dispersing among the others—as is the case for fullerenes’ tendency not to disperse among polymers.  This tendency can be overcome if all the polymer and fullerene molecules are dissolved in something first before being mixed.  Accomplishing this can be problematic in itself, since few solvents work equally well on polymers and fullerenes; separate solvents may be needed for both, but then the solvents may interact detrimentally with each other.  If these problems are avoided, the mixture of different molecules achieved will need to be retained to at least some extent as the solvent(s) are evaporated to leave a fullerene-polymer composite. 


Researchers at the University of Tennessee and Oak Ridge National Laboratory studied the formation of one such composite material, monitoring the mixing of its components (buckminsterfullerene and polyacrylonitrile) in solution and then monitoring their solvents’ evaporation.  Their observations enabled them to infer how the dissolved buckminsterfullerene molecules shaped the composite material’s ultimate crystal form, as indicated by the title of their report, “The impact of fullerenes on the ordering of polyacrylonitrile during nanocomposites formation”[DoE PAGES].  They found that the chainlike polyacrylonitrile molecules arrange themselves into hexagonally packed[Wikipedia] rodlike crystals if neither buckminsterfullerene nor residual solvents are present.  If buckminsterfullerene molecules are added first, they get distributed homogeneously among the polyacrylonitrile chains and inhibit their crystallization.  But when the crystals do form, only part of the polyacrylonitrile crystallizes, the crystals are less rigid, the polyacrylonitrile molecules are arranged orthorhombically[Wikipedia] instead of hexagonally as if they formed in the presence of a residual solvent, and the buckminsterfullerene molecules get sequestered to the crystals’ outer edges.  Apparently then, when buckminsterfullerene molecules are present in a crystallizing solution, they envelop the rodlike polyacrylonitrile crystals, obstructing their further crystallization and thus limiting their size. 


The report didn’t name a specific use for this product aside from its exemplifying a general class of materials, some members of which might have exotic electrical and thermal properties[op. cit., p. 1] derived from their nanoparticle (and other) constituents—buckminsterfullerene and polymer molecules in this case.  Other fullerene-based materials, like the one discussed by Cornell University researchers in “Anisometric C60 Fullerene Colloids Assisted by Structure-Directing Agent” (“Final Report Grant No. ER46517”[SciTech Connect]), were synthesized to have particular properties—in this case, photoluminescence.  In these experiments, suitable mixtures were found in which suspended buckminsterfullerene (C60) molecules aggregated into microcrystals of fairly uniform structure and size—characteristics that were found to depend on the C60 concentration and volume.  The mixtures included a solvent, an antisolvent, and another material that acted as a “structure-directing agent” that inhibited the growth of C60 microcrystals, not unlike the way C60 itself inhibited polyacrylonitrile crystal growth in the University of Tennessee/Oak Ridge National Lab experiments.   In the Cornell experiments, the C60 molecules, originally dissolved with molecules of the structure-directing agent in solvent droplets, would be induced to aggregate into microcrystals as the solvent’s own molecules dispersed into the antisolvent surrounding the droplets.  But the structure-directing agent’s molecules apparently tended to block the access of additional molecules onto certain facets of the C60 microcrystals, thus inhibiting further aggregation there.  The buckminsterfullerene microcrystals thus formed were more uniform than those produced without the aggregation inhibitor.  They also showed less fluorescence from the same light input[Wikipedia], with a briefer average fluorescence time[Wikipedia]


Figure 2 leftFigure 2 right

Figure 2.  Left:  A schematic diagram of how buckminsterfullerene (C60) crystals apparently grow under the conditions described in “Final Report Grant No. ER46517”[SciTech Connect].  As a solvent droplet (grey) shrinks by diffusing into a surrounding antisolvent (blue), C60 molecules (white) within it are induced to aggregate into a crystal.  The solvent droplet also contains molecules of a “structure-directing agent” (green) which associate with certain facets of the C60 crystal and thus block further growth of those facets.  Right:  scanning electron microscope images of C60 crystals grown in a tetralin solvent containing 2,4,6-trimethylpyridine as a structure-directing agent.  Each image shows the results of different C60 concentrations and injection volumes.  In concentrations of millimolars[Wikipedia] (mM = millimoles[Wikipedia]/liter) and injection volumes of microliters (μL), the growth conditions for each image are:  a) 1.6 mM, 150 μL; b) 1.6 mM, 300 μL; c) 1.6 mM, 450 μL; d) 1.8 mM, 150 μL; e) 1.8 mM, 300 μL; f) 1.8 mM, 450 μL; g) 2 mM, 150 μL; h) 2 mM, 300 μL; i) 2 mM, 450 μL.  Scale bars in the images represent 1 micrometer.  (After “Final Report Grant No. ER46517”[SciTech Connect], pp. 2 and 3.)  


An interesting type of fullerene composite is possible because of fullerenes’ hollow structure:  a fullerene shell can be formed to enclose smaller atoms or molecules and interact with them.  This type of composite has been given a special chemical notation:  the molecular formula of, say, an 80-carbon fullerene enclosing a molecule of Ho3N is not Ho3NC80, but Ho3N@C80, the “@” denoting that the first set of atoms is enclosed in the second.  A fullerene so “doped” may behave quite differently from an empty one.  Ho3N@C80 was itself studied by one group of researchers interested in how interaction of composite fullerenes’ components might affect the way they respond to intense laser beams.  Their work, reported in “The Role of Super-Atom Molecular Orbitals in Doped Fullerenes in a Femtosecond Intense Laser Field”[DoE PAGES], dealt with how the electromagnetic fields of strong infrared laser pulses lasting only 30 quadrillionths of a second (30 femtoseconds) affect doped fullerenes’ structure and dynamics—a step toward optimizing such fullerenes for use in electronic devices, including solar batteries. 


A significant difference in the behaviors of C60 and Ho3N@C80 was the rate at which electrons were dissociated from these molecules by energy absorbed from the laser pulses, and how the dissociation rates varied with the pulses’ intensity.  Computations of the electron distributions based on density functional theory[Wikipedia] revealed different processes at work in the two molecules.  Laser pulses whose power flux is less than one watt per square nanometer will put the less-symmetric doped C80’s electrons into states of motion from which they can be readily dissociated from the rest of the molecule; with electrons in these states, the molecule only has to absorb the energy of about three laser-pulse photons to remove the electrons altogether.  Since these predissociational states, or “orbitals”, involve motion throughout the entire molecule, they’re known as “molecular orbitals”, but the motion furthermore happens to resemble that of low-energy electrons in single atom.  Electrons moving in this type of molecular orbital traverse the Ho3N@C80 much as if that molecule were a single large atom itself; hence these states of motion are called “super-atom molecular orbitals”, as they are in the report title.  (See Figure 3.)  While it’s also possible for electrons in the more symmetric undoped C60 molecule to be super-atom states, the molecule’s symmetry makes it impossible for a laser pulse of less than one watt per square nanometer to drive its electrons into them.  A C60 molecule’s electrons can still be dissociated from it, but only by a different sequence of events that involve the absorption of five laser-pulse photons instead of three—something much less likely to happen in the 30 femtoseconds that it took the laser pulses of the experiment to pass over a molecule. 


Figure 3

Figure 3.  (a) Contour plots of wave functions for an electron in a typical lowest-energy state of motion in a molecule of Ho3N@C80 (valence state) and in various higher-energy states (SAMO = Super-Atom Molecular Orbital) with different amounts of orbital angular momentum.  (b) Corresponding electric charge densities (electron charges e per cubic Bohr radius a0) for the wave functions at various radial distances r (in Ångström units Å = 0.1 nanometer) from the center of the Ho3N@C80 molecule.  Note that an electron in the illustrated valence state has almost zero chance of being more than about 14 Ångström units away from the molecule’s center, while electrons in the SAMO states are substantially more likely to be found at those distances.  (c) Calculated numbers of possible super-atom molecular orbital states for a Ho3N@C80 electron that has been excited to various energies between 4 and 6.5 electron-volts (eV).  The “IP” just above 6.5 eV represents the ionization potential—the excitation energy at which an electron would be completely detached from the rest of the molecule.  (After “The Role of Super-Atom Molecular Orbitals in Doped Fullerenes in a Femtosecond Intense Laser Field”[DoE PAGES], p. 4.) 


Doped fullerenes’ ability to energize electrons with light makes them potentially useful in solar batteries.  A different investigation dealt with tubular fullerenes’ ability to energize electrons with heat.  In some materials, if the temperature becomes different at different points, electrons in the material will redistribute themselves and thus set up different concentrations of positive and negative charge.  If the material conducts charge quickly and conducts heat slowly, the charge will redistribute quickly in response to the temperature difference and the temperature difference will persist, as will the new concentrations of positive and negative charge.  If an electrically-conducting path is provided between the hotter and colder portions of the material, electrons will be repelled from the negative concentrations and attracted to the positive ones.  Thus a heat source that produces a temperature difference across such a material is turned into a producer of electric current—something quite valuable if the heat is an otherwise untapped waste product of some other power source.  Some inorganic semiconducting materials can turn heat into electric power, but the most effective ones aren’t flexible enough be made into arbitrary irregular shapes.  However, composites of polymers and carbon nanotubes do have this flexibility, and some carbon nanotubes have been found able to drive electric currents when differentially heated. 


“Tailored semiconducting carbon nanotube networks with enhanced thermoelectric properties”[DoE PAGES], published last year in Nature Energy, describes how earlier explorations of carbon nanotubes’ suitability as thermoelectric energy converters, whose history was one of mixed results, were improved upon by the authors’ finding a way to separate nanotubes with different conductivities.  As it happens, some nanotubes readily conduct heat, in the form of random vibrations of the tubes’ carbon-atom lattice structures, while others don’t.  Those that conduct heat well also conduct electric charges easily, as metals do; those that don’t conduct heat well act electrically as semiconductors instead of metals.  It has been long known that whether a carbon nanotube behaves like a metal or a semiconductor depends on how its atomic lattice is oriented with respect to the tube’s length.  Of all the atomic arrangements that a carbon nanotube can have, one third of them result in metallic behavior and the other two-thirds make the tube semiconducting.  Among their achievements, the authors were able to produce carbon nanotubes of many atomic arrangements and preferentially extract semiconducting nanotubes that were useful in thermoelectric composites.  Their experiments produced composites that exhibited much larger voltage differences between differently-heated points than earlier experiments did, exceeding 100 microvolts for every degree Celsius of temperature difference.  The authors also described theoretical analyses of how their composites worked that may guide the fabrication of future thermoelectric materials and devices. 


Figure 4 leftFigure 4 right

Figure 4.  Left:  Side-view and end-view schematics of smallest repeated units for various single-walled carbon nanotubes; attaching several such units end to end would make longer tubes of the same type.  The integers in parentheses indicate how an unrolled unit would be oriented if its atoms were lined up with those of a sheet of graphene.  If the difference between the two integers is a multiple of 3, the nanotube will behave as a metal; if not, the nanotube acts as a semiconductor.[Maruyama]  Right:  Representative atomic-force micrograph of a thin polyfluorene film containing carbon nanotubes that are predominantly semiconducting.  (After “Tailored semiconducting carbon nanotube networks with enhanced thermoelectric properties”[DoE PAGES], pp. 5 and 7.) 




Separator line

Polymer-fullerene solar cells


The possibility of using polymer-fullerene composites based on spherelike fullerenes rather than tubular ones to power electric current with light instead of heat was recognized in 1992.  The first devices based on these composites, known as bulk heterojunction systems, were very inefficient, but those explored by 2015 generally had efficiencies of 8% and 9%, and one with 10.7% efficiency had been reported.[DoE PAGES, p. 4]  People have continued seeking to further improve various features of bulk heterojunction systems. 


Polymer molecules[Wikipedia] in a polymer-fullerene heterojunction can use energy absorbed from light to separate positive and negative charges.  At least some oppositely charged pairs stay bound together[Wikipedia] without recombining until they reach a fullerene molecule, where they dissociate and move toward different electrical contacts, thus building up the contacts’ potential difference.  Since the bound charge pairs tend to recombine in the time they take to travel about 10 nanometers, they can only separate entirely if they don’t have to travel far to reach a fullerene molecule.  Thus if the heterojunction is to produce a current-driving voltage difference between its contacts, it apparently must have its charge-donating material well mixed with its charge-accepting material; how the two are mixed together determines the shape of the donor-acceptor interface, and therefore how efficiently the heterojunction device converts light into electrical energy.  Completely understanding the exact role that the interface shape plays in the device’s efficiency would be useful for producing the maximally efficient shape.  Thus the authors of “Harnessing Structure-Property Relationships for Poly(alkyl thiophene)-Fullerene Derivative Thin Films to Optimize Performance in Photovoltaic Devices”[DoE PAGES] used neutrons and x-rays to examine the forms taken by the interfaces of nine different polymer-fullerene bulk heterojunctions made from three different polymers and three different fullerenes.  They found direct correlations among the interface forms and several device performance parameters, including the devices’ overall efficiencies, and demonstrated that by deliberately adding a layer of fullerenes through which electrons can be transported from the heterojunction to their electrical contact, the heterojunction’s efficiency can be increased by 5% overall. 


While much recent effort to improve bulk heterojunctions has focused on the polymers that donate electrical charge carriers, few changes to the fullerenes used have greatly improved heterojunctions’ efficiencies.  Insight into how electrons behave in the fullerene molecules might usefully inform improvements, but the complexity of fullerenes’ electronic structures makes their analysis challenging.  Researchers at Chicago State University and Argonne National Laboratory have taken on the challenge[DoE PAGES], observing how electrons in heterojunctions’ fullerenes respond to microwaves while the heterojunctions were kept in static magnetic fields[Wikipedia] and analyzing the responses in terms of density functional theory.  Every electron, besides being a unit of negative electric charge, is also a magnet.  While the magnetic fields of a molecule’s electrons tend to line up in opposing pairs and cancel each other out, any unpaired electron will give the molecule an uncancelled magnetic field that tends to align with others—such as the static fields of external magnets or the oscillating fields of microwaves.  A molecule with an unpaired electron can either absorb or amplify microwaves of the right frequency.  The exact response of a bulk heterojunction’s fullerenes to microwaves thus provides clues to how its electrons are distributed.  When four fullerenes widely used in bulk heterojunctions were examined in that environment, their unpaired electrons were found most likely to be on the molecules’ spherelike portions instead of their chainlike portions, with less symmetrical distributions on the less symmetric fullerenes (see Figure 5).  The researchers proposed that the difference is related to how the frequencies of resonant microwaves differ with the fullerenes’ orientations in an external magnetic field, and that these orientational differences result from the flexible fullerene cages’ mechanical interactions with their heterogeneous environment in the bulk heterostructure. 


Figure 5

Figure 5.  “Front-view” and “side-view” contour maps, for four different fullerenes, of the most likely locations for an “unpaired” electron whose magnetic field isn’t cancelled out by that of another oppositely-oriented electron.  Changing the alignment of the unpaired electron’s magnetic field from the same direction as that of an external magnet to the opposite direction requires energy—more energy if the external field’s direction is along the g3 axes depicted above, less energy if the external field’s direction is g2, and least if the external field’s direction is g1.[Wikipedia]  (From “Electronic Structure of Fullerene Acceptors in Organic Bulk-Heterojunctions. A Combined EPR and DFT Study”[DoE PAGES], p. 11.)




Separator line

Perovskite-fullerene solar cells


Photovoltaic devices with polymers as charge donors offer the advantages of light weight and flexibility, but in some respects other types of solar cells surpass them.  Photovoltaics whose charge donors are materials with a perovskite structure[Wikipedia], for instance, have been more efficient.  Yet perovskite-based solar cells[Wikipedia] have some disadvantages of their own.  Among the more serious ones has been their instability in the presence of oxygen and moisture—a serious problem for solar cells deployed in the open air, particularly if the weather ever includes rain or even humidity.  The reports “Enhanced ambient stability of efficient perovskite solar cells by employing a modified fullerene cathode interlayer[DoE PAGES] and “Enhancing stability and efficiency of perovskite solar cells with crosslinkable silane-functionalized and doped fullerene”[DoE PAGES] show how groups of researchers at the University of Washington and the University of Nebraska—Lincoln, respectively, have addressed this by different modifications of the charge-accepting fullerenes. 


The Washington group experimented with using a mixture of two fullerenes to transport charge carriers from the perovskite to the electrical contacts in solar cells, and found that such devices were more stable in ambient air and, at least initially, more efficient (15.5%) than similar devices in which the electrical contact was made directly with the perovskite with no fullerene layer between them (12.1%).  The superiority persisted with conditions of 20% relative humidity.  Although the solar cells with a fullerene interlayer were scarcely any more efficient 14 days later than the cells without one had been at the beginning, those other solar cells had already completely degraded completely after 7 days.  Furthermore, the cells with the new fullerene mixture also exhibited less recombination of positive and negative charge-carriers. 


The Lincoln group tried a different kind of layer between the perovskite and the electrode.  Theirs consisted of a fullerene attached by hydrogen bonds[Wikipedia] to a polymer constructed from trichloro(3,3,3-trifluoropropyl)silane molecules, with some methylammonium iodide added as a dopant to improve the layer’s electrical conductivity.  The ability of these layers to self-assemble by hydrogen bonding makes their fabrication relatively easy.  Solar cells made with the layers had an initial efficiency of 19.5%, and a week of continual exposure to light and air that was much more humid (50%-75% relative humidity) only reduced their efficiency to roughly 70% of its original value.  The doped fullerene and silane polymer layer also made the solar cells more water repellant, as shown by Figure 6. 


 Figure 6 leftFigure 6 middleFigure 6 right

Figure 6.  Left:  Schematic molecular models of the two fullerenes that researchers at the University of Washington mixed to form a charge acceptor for solar cells that had perovskite[Wikipedia] charge donors.  The new charge acceptor improved on an earlier one made only of “bis-C60” molecules.  Middle:  Schematics for perovskite-based solar cells[Wikipedia] made by researchers at the University of Nebraska—Lincoln.  The charge acceptor for these was made with different fullerene molecules joined by hydrogen bonds to a silane-based[Wikipedia] polymer—giving the charge acceptor an easily self-assembled structure.  Right:  Water droplets on films of uncoated perovskite have a smaller contact angle than water droplets on perovskite film with a fullerene-silane coating—that is, the coated perovskite layers get less wet.  (From/After “Enhanced ambient stability of efficient perovskite solar cells by employing a modified fullerene cathode interlayer[DoE PAGES], p. 1 of 7, and “Enhancing stability and efficiency of perovskite solar cells with crosslinkable silane-functionalized and doped fullerene”[DoE PAGES], pp. 22, 24, and 25.)


Humidity has not been the only limiter of perovskite-based solar cells’ usefulness.  Among those that use planar electrodes, one thing that limits them is a tendency for their separated negative and positive charge carriers to recombine before they can produce current outside the device.  Another is that electric power they produce depends not only on how much light is shining on them, but on the devices’ previous history:  two identical cells subjected to different patterns of light exposure may continue to produce different amounts of power even after their illumination has become the same.  Countermeasures have included making the devices’ surface or near-surface portions from materials whose fabrication processes are more complex—a problematic measure in itself that furthermore doesn’t fix defects in the cells’ interiors.  “Perovskite-Fullerene Hybrid Materials Eliminate Hysteresis In Planar Diodes”[DoE PAGES] reports how the history and charge-recombination problems were reduced by making solar cells from a lead iodide perovskite and one of the same fullerenes (PC61BM = phenyl-C61-butyric acid methyl ester[Wikipedia]) whose electron distribution was examined in the Chicago State University and Argonne National Laboratory research discussed above.  Experiments with the perovskite-PC61BM solar cells suggested to the experimenters that when PC61BM is incorporated at or near perovskite grain boundaries, it blocks electric charges from becoming trapped on the boundaries’ iodine-rich sites.  The experimenters also speculated that, in a solar cell whose iodine atoms weren’t tied up by PC61BM, electric fields could cause negatively charged iodine atoms to migrate and thus affect the cell’s later response to light, while tying up iodine atoms by PC61BM molecules would make the cell’s atomic arrangement more stable and the cell’s behavior more consistent. 




Separator line

Two kinds of fullerene substitutes


Although fullerenes appear to have their uses, at least some of them present problems of their own.  The PC61BM and PC71BM molecules discussed above are expensive and have limited light absorption, and their interfaces with the polymers in bulk heterojunction cells have unstable shapes.  This has motivated experiments with substitute organic materials like the ones reported in “Covalently Bound Clusters of Alpha-Substituted PDI—Rival Electron Acceptors to Fullerene for Organic Solar Cells”[DoE PAGES] by researchers at the University of Chicago and Argonne National Laboratory.  The substitute examined is made of relatively flat molecules whose backbones can be packed together closely (Figure 7).  Cells made with the substitute material had similar efficiencies similar to those made with PC71BM and the same polymer. 


Figure 7

Figure 7.  Schematics of the repeating units of the polymer PTB7-Th (lower left), and of molecules that go into various stages of production of TPBDT (right)—a nonfullerene substitute for PC71BM fullerenes in a polymer-based solar cell.  Cells made with TPBDT had efficiencies similar to those made with PC71BM.  (From “Covalently Bound Clusters of Alpha-Substituted PDI—Rival Electron Acceptors to Fullerene for Organic Solar Cells”[DoE PAGES], p. 1 of 5.) 


Attempts at a quite different kind of fullerene substitution are reported in the Journal of Applied Physics paper “The quest for inorganic fullerenes”[DoE PAGES].  The idea here was to see if elements other than carbon could form molecules similar to the spherelike fullerenes.  Since a graphene flake with some atoms (necessarily) at its edges has a higher energy than a spherelike fullerene with the same number of carbon atoms and no edges, carbon sheets tend to form spherelike molecules.  Other elements and compounds are known to form tubes and nested structures like fullerenes, but not “single-wall” spherelike molecules.  For instance, tungsten disulfide (tungstenite, WS2) and molybdenum disulfide (MoS2) are layered materials that resemble carbon in how they function as lubricants and can form nested fullerenes.  The smallest previously-known WS2 and MoS2 structures are triple-layered and octahedral[Wikipedia] rather than single-layered and spherelike. 


Whereas earlier inorganic fullerene syntheses, accomplished by methods suitable for making large amounts of material, resulted in multithousand-atom molecules much larger than carbon fullerenes, in this case the experimenters subjected WS2 and MoS2 to vaporization processes like those that yield spherelike fullerenes from graphite. Applied to carbon, these processes produce clusters of atoms whose most common masses fall among molecules of two types:  lower-mass, one-dimensional carbon chains and rings, and higher-mass fullerenes.  Applied to WS2 and MoS2, there were also two masses that were most frequently seen among the resulting clusters.  As was already known, the clusters whose masses were near the smaller most-common mass were one-dimensional chains and two-dimensional platelets (not rings) with varying proportions of metal and sulfur atoms. 


The clusters whose masses were closer to the larger most-common mass were examined by multiple techniques—mass spectrometry[Wikipedia], photoemission spectroscopy[Wikipedia], scanning transmission electron microscopy[Wikipedia], and scanning tunneling microscopy[Wikipedia]—to determine their form.  However, these clusters also turned out to be larger platelets instead of closed surfaces with spherelike or other forms.  The experimenters concluded that this difference from carbon “might be explained by the additional degree of freedom a two-element system has.  The edges of MoS2 and WS2 flakes can be stabilized by additional sulfur atoms.  It is almost impossible to avoid this stabilization in a 2-element condensation process.  Hence, it could be that carbon is the only material able to form real fullerenes.” 




Separator line








Separator line

Reports available through DoE PAGES


  • “The impact of fullerenes on the ordering of polyacrylonitrile during nanocomposites formation” [Metadata]


  • “The Role of Super-Atom Molecular Orbitals in Doped Fullerenes in a Femtosecond Intense Laser Field” [Metadata]


  • “Tailored semiconducting carbon nanotube networks with enhanced thermoelectric properties” [Metadata]


  • “Electronic Structure of Fullerene Acceptors in Organic Bulk-Heterojunctions. A Combined EPR and DFT Study” [Metadata]

Page 4 of the manuscript (page 4730 of the journal volume) mentions when charge transfer in polymer-fullerene systems was discovered and how efficient such systems had since become at producing electric current from light. 


  • “Harnessing Structure-Property Relationships for Poly(alkyl thiophene)-Fullerene Derivative Thin Films to Optimize Performance in Photovoltaic Devices” [Metadata] [DoE PAGES]


  • “Enhanced ambient stability of efficient perovskite solar cells by employing a modified fullerene cathode interlayer” [Metadata]


  • “Enhancing stability and efficiency of perovskite solar cells with crosslinkable silane-functionalized and doped fullerene” [Metadata]


  • “Perovskite-Fullerene Hybrid Materials Eliminate Hysteresis In Planar Diodes” [Metadata]


  • “Covalently Bound Clusters of Alpha-Substituted PDI—Rival Electron Acceptors to Fullerene for Organic Solar Cells” [Metadata]


  • “The quest for inorganic fullerenes” [Metadata]




Separator line

Report available through OSTI’s SciTech Connect


  • (“Final Report Grant No. ER46517”) [Metadata]

This report contains descriptions of the Crystal Engineering Communications paper “Anisometric C60 Fullerene Colloids Assisted by Structure-Directing Agent” described above and three other papers about results of additional research supported by the same Energy Department grant. 




Separator line

Additional reference