Molecular dynamics simulations of oligomers of the recently reported polycarbyne network backbone polymers indicate that calculated bondlengths between adjacent carbon backbone atoms are very long compared to the bond distance of a C-C single bond. Some degree of bond cleavage is theorized to occur between adjacent carbon atoms of the polymers' network backbones, resulting in the formation of biradicals. This theory is supported experimentally by the polymers' electronic absorption spectra, their degrees of polymerization, and their ESR spectra, which show a decreasing signal as the steric bulk between adjacent carbons is decreased.The synthesis of the first member of a new class of carbon-based polymers, poly(phenylcarbyne) [PhC] n (1), has recently been reported (7,2). This polymer's stoichiometry is identical to that of poly(diphenylacetylene) (3,4) yet its structure is significantly different. Unlike the polyacetylenes, which are linear polymers whose backbones consist of alternating single and double bonds, poly(phenylcarbyne) has a random network backbone which consists of sp 3 -hybridized carbon atoms bonded via carbon-carbon single bonds to three other backbone atoms and one phenyl substituent. This backbone structure is unique in carbon-based polymers and has been found to confer novel properties and reactivity on 1, as, for example, pyrolytic conversion to diamond or diamondlike carbon at atmospheric pressures (7). Because of this pyrolytic conversion, numerous applications for this polymer can be envisioned.Although the network backbone microstructure of poly(phenylcarbyne) and its Group 14 congeners (7,2,5-9) has been confirmed by 13 C and ^9Si NMR, the macrostructures of this class of polymers have not been established. The seemingly completely random assembly of the polymer backbones defeats any spectroscopic or diffraction characterization technique, since no two polymer molecules in a given sample may display identical macrostructures and therefore no one macrostructure gives rise to enough characteristic signal to be unequivocally detected. Information about the macrostructure of the polymers is important, however, as polymer macrostructure appears to influence the materials' properties (see below). We have therefore used molecular modeling techniques to provide insights into possible macrostructures for the polycarbyne class of network backbone polymers, and to 0097-6156/95/0589-0304$12.00A)