The aqueous polycondensation of resorcinol with formaldehyde proceeds through a sol-gel transition and results in the formation of highly crosslinked, transparent gels. If the solvent is simply evaporated from the pores of these gels, large capillary forces are exerted and a collapsed structure known as a xerogel is formed. In order to preserve the gel skeleton and minimize shrinkage, the aforementioned solvent or its substitute must be removed under supercritical conditions. The microporous material that results from this operation is known as an aerogel. Because resorcinol-formaldehyde aerogels and xerogels consist of a highly crosslinked aromatic polymer, they can be pyrolyzed in an inert atmosphere to form vitreous carbon monoliths. The resultant porous materials are black in color and no longer transparent, yet they retain the ultrafine cell size (< 50 nm), high surface area (600-800 m2 /g), and the interconnected particle morphology of their organic precursors. The thermal, acoustic, mechanical, and electrical properties of carbon aerogels/xerogels primarily depend upon polymerization conditions and pyrolysis temperature. In this paper, the chemistry-structure-property relationships of these unique materials will be discussed in detail.

Carbon aerogels are synthesized via the polycondensation of resorcinol and formaldehyde, followed by supercritical drying and pyrolysis at 1050 °C in nitrogen. Because of their interconnected porosity, ultrafine cell structure and high surface area, carbon aerogels have many potential applications, such as in supercapacitors, battery electrodes, catalyst supports, and gas filters. The performance of carbon aerogels in the latter two applications depends on the permeability or gas flow conductance in these materials. By measuring the pressure differential across a thin specimen and the nitrogen gas flow rate in the viscous regime, we calculated the permeability of carbon aerogels from equations based upon Darcy's law. Our measurements show that carbon aerogels have apparent permeabilities on the order of 10−12 to 10−10 cm2 for densities ranging from 0.44 to 0.05 g/cm3. Like their mechanical properties, the permeability of carbon aerogels follows a power law relationship with density and average pore size. Such findings help us to estimate the average pore sizes of carbon aerogels once their densities are known. This paper reveals the relationships among permeability, pore size and density in carbon aerogels.

Low density foams are used for a variety of applications, including catalytic supports, battery anodes, microporous membranes, and laser fusion targets. The technique for making replica carbon foams described in this paper has been previously reported[I] and involves a process in which an inorganic substrate (sodium chloride) is infused with a carbonizable polymer. After carbonization, the substrate is removed by a leaching process and the wet foam is dried; the resultant foam is referred to as replica carbon. This paper describes improvements in the processing which result in a smaller pore size and improved foam homogeneity.

The original substrate is the single most important factor affecting the resultant structure. Techniques to improve the uniformity of the substrate and the translation of substrate anomalies into the final product are described.

The extraordinary mechanical properties of commercial carbon fiber are due to the unique graphitic morphology of the spun filaments. Contemporary advanced structural composites exploit these properties by creating a disconnected network of graphitic filaments held together by an appropriate matrix. Carbon foam derived from a blown mesophase pitch precursor can be considered to be an interconnected network of graphitic ligaments. As such interconnected networks, they represent a potential alternative reinforcing phase for structural composite materials. Based on this notion, consideration is given to novel forms of graphitic carbon and the processing routes to create hybrid composites, such as net shape fabrication and fiber placement processes.

Mathematical equations have been formulated to guide an experimental effort to produce an open-celled mesophase pitch foam. The formulation provides an analytical description of homogeneous bubble nucleation and growth, diffusion of the blowing gas through the liquid to the bubble surface, and the average material thickness between bubbles. Implications of the formulation for the experimental production of mesophase pitch foam are discussed.

Open-celled foam is modeled as a network of struts formed by the nucleation, growth and intersection of uniform spherical bubbles arranged in a regular three-dimensional array. The number of struts per unit cell and their characteristic geometry and orientations depend on the bubble-centering algorithm chosen to define the array and on the degree of bubble intersection. A foam model based on a body-centered cubic arrangement of uniform spherical bubbles is used to compare strut geometries over a range of bubble interferences. The model predictions are compared to the microstructures observed in amorphous carbon foam. The significance of these idealized foam results for graphitic carbon foams is discussed.

The ligament structure of several open-cell carbon foams was examined by optical and electron microscopy. The arrangement, sizes, and shapes of the ligaments were measured and analyzed according to the cell sizes. The ligament lengths and cross-sections vary with the cell sizes in a simply scaled fashion. A models based on the observed dodecahedral-like arrangement of ligaments was constructed consisting of 12-, 14-, and 15-faced polyhedra with five-edged faces dominating.

Recently we reported the chemistry-structure-property relationships of organic aerogels, which are synthesized by the polycondensation of resorcinol and formaldehyde in a slightly basic medium, followed by supercritical drying. These materials can be pyrolyzed in an inert atmosphere to form vitreous carbon aerogels. As measured by gas adsorption techniques, the BET surface area and pore size distributions of micro and meso pores of the carbon aerogels are affected both by the pyrolysis temperature and the formulation. Definite trends are observed in our preliminary measurements; for example, the surface area decreases with increasing pyrolysis temperature until a plateau is reached at about 900°C. This paper explores the effects of pyrolysis temperature and aerogel density on the BET surface area and pore size distributions.

The two most important material properties for foam process modeling are liquid viscosity and surface tension. Several different pitches were examined including Ashland petroleum pitches and a synthetic mesophase pitch (Mitsubishi AR). Viscosities of isotropic pitch were measured at various temperatures above their softening points using an oscillating disk rheometer and found to be independent of shear rate and closely match the manufacturer's data. Only a very weak dependence of surface tension on temperature was found from calculations made from pendent drop measurements. The density of these materials, which was needed for the surface tension calculations, was measured using an elevated temperature mercury dilatometer. Transitions indicated by the volumetric results were also investigated by DSC.

Commercially available carbon black contains oxygen complexes on its surface that affect the surface properties of the carbon. Water adsorption on the surface of carbon black is influenced by the amount and type of oxygen complexes present. When carbon black is heated in vacuum at sufficiently high temperatures, removal of the oxygen complexes occurs and the surface of the carbon particles is modified. The amount of water adsorbed by the carbon is dependent on the vacuum heat treatment temperature. As the heat treatment temperature increases, water adsorption on the carbon decreases.

Commercially available electrically conductive carbon black adsorbs from 1.25% to 2.50% water when exposed to 50% relative humidity for 24 hours at 25°C. This variation in water adsorption is due to a difference in the amount of oxygen complexes on the surface of the carbon. The carbon with more oxygen complexes adsorbs more water. However, when this carbon black is heat treated at 1200°C for 4 hours in a vacuum of 1 × 10−5 torr or better, the water adsorbed by the carbon is 0.4% when exposed to 50% relative humidity. Data showing the dependence of water adsorption on vacuum thermal processing are presented and discussed.

Thermal fatigue resistance of carbon-carbon composites with three different fiber surface treatments is studied in thermal cycles between 100 °C and 1,700 °C up to a number of cycles of 100 in free and restricted expansion conditions. The effects of thermal cycles were studied by SEM paying attention especially to structural damage and interfacial debonding between fibers and matrix. Bending tests subsequent to 10 thermal cycles were used to study the effect of thermal cycling on mechanical properties of the composites. The effect depends on the surface treatment of fibers prior to pyrolization. In some cases, the bending strength decreased due to the thermal cycling, whereas a suitable surface treatment minimized the damaging effect and increased the pseudo-ductility of the composite.

Carbon fibers are finding increasing application in making various composites with special properties. It is found that such composites have properties that can be markedly effected by the carbon fiber surface. The effects of surface modification by electrochemical and plasma oxidation is discussed and the surface chemical changes described. The interaction of the matrix with the fiber surface is a subtle mixture of physical and chemical effects. The paper discusses work in the author's laboratory using surface studies with core and valence band X-ray photoelectron spectroscopy (XPS) and other techniques.

Thin films of pure germanium-carbon alloys (GexC1−x with x ≈ 0.0, 0.2, 0.4, 0.5, 0.6, 0.8, 1.0) have been grown on Si(100) and A12O3 (0001) substrates by pulsed laser ablation in a high vacuum chamber. The films were analyzed by x-ray θ-2θ diffraction (XRD), x-ray photoelectron spectroscopy (XPS), Auger electron spectroscopy (AES), conductivity measurements and optical absorption spectroscopy. The analyses of these new materials showed that films of all compositions were amorphous, free of contamination and uniform in composition. By changing the film composition, the optical band gap of these semiconducting films was varied from 0.00eV to 0.85eV for x = 0.0 to 1.0 respectively. According to the AES results, the carbon atoms in the Ge-C alloy thin film samples has a bonding configuration that is a mixture of sp2 and sp3 hybridizations.

Neutron and X-ray diffraction techniques have been applied to the study of two samples of a-Si:C:H. Both samples were prepared using conventional glow discharge methods, but the hydrocarbon/silane precursor gas was diluted with hydrogen in one case. Analysis of the X-ray diffraction data gives a clear picture of the silicon network, since the scattering profile is dominated by the Si-Si correlations. The high real-space resolution neutron diffraction data, however allows one to comment on the effect of this dilution on the silicon-carbon bonding morphology, and in particular on the degree to which the additional hydrogen enhances hetero-coordination. In addition we present the results of a preliminary computer simulation study of the structure of a-C:H and a-Si:H using an approximate molecular dynamic density functional theory, and discuss its viability in the study of the more complex a-Si:C:H ternary alloy.

We have used a first principles molecular dynamics technique to find the relaxed atomic geometry and the corresponding electronic structure of a simple novel form of threecoordinated solid carbon, which we name polybenzene. The polybenzene structure is found to have a substantially lower energy per atom (by 0.23 eV) than C60, yet contains only 24 atoms per unit cell. In addition we have calculated the zero wavevector vibrational spectrum for this novel carbon structure.

Submonolayer coverages of carbon adsorbed on highly-oriented pyrolytic graphite were examined by scanning tunneling microscopy under ultra-high vacuum condition. Linear carbon wires were found on atomically flat graphite surfaces. The wires had different thicknesses, from single atomic width to about lnm. The long wires extended to over several hundred nanometers. Two directions, graphite β-β direction and 30° rotated, were preferred for the long wire orientation. Parallel wire alignment, with several nanometers of inter-wire spacings were observed. Carbon particles, from 0.7 to 2 nm in diameter were found to be attached to the carbon wires. Particles from different wires formed parallel linear chains about perpendicular to the wire direction.

The new technique of ion structural chromatography is applied to carbon clusters. The results indicate that C5+ and C6+ are purely linear but C7+, C8+, C9+ and C1O+ have both linear and monocyclic ring structures. From C11+ to C20+ only monocyclic ring structures are observed. At C21+, a new family of planar ring structures appears. The first 3 dimensional structure occurs at C29+ and the first fullerene at C30+. Isomer structure is verified by the comparison of experimental mobilities with those derived from theory for the various structures. For C20+ only the monocyclic ring is observed experimentally but electronic structure calculations suggest more compact structures might be lower in energy. The results are discussed in terms of possible growth mechanisms for C60.

Methods have been developed for the systematic generation of all fullerene isomers and for screening them to identify potentially stable structures. Some structural features associated with stable fullerenes are identified.

Prompted by earlier work showing that graphite intercalation compounds form superconducting ternary compounds in which thallium plays an important role C60 has been doped with Tl-alloys. Results of dc-magnetization, ac-susceptibility, X-ray analysis and NMR investigations will be presented.

Molecular cluster calculations within the local density approximation have been performed in a study of the electronic structure of the C60 molecule - “Buckminsterfullerene” doped with K, B and N. Calculations for the KC60 molecule, with the K atom located at the centre of the cage as well as at different positions inside or outside the cage, show how the valence 4s electron is transferred to the LUMO state of the bare C60 molecule. Doping with a B or N atom located at the centre of the cage creates a molecule with a partly occupied level of 2p character in the HOMO and LUMO gap, similar to donor and acceptor levels in the band gap of traditionally doped semiconductors. Doping by substitution of one or two of the carbon atoms in the cage with X = B or N, as modelled with the C59 X1 or C58X2 clusters, gives a different structure with a splitting of the HOMO and LUMO levels in the pure C60 molecule and with the creation of acceptor and donor levels with the substitution of B and N, respectively.