VOLUME 45, NUMBER 1) PHYSICAL REVIEW LETTERS 29 SEPTEMBER 1980 Solitons in Polyacetylene: Magnetic Susceptibility S. Ikehata, '' ' J. Kaufer, ' T. Woerner, ) A. Pron, M. A. Druy A. Sivak, ~ A. J. Heeger, ~' and A. G. MacDiarmid~- Laboratory for Research on the Structure of Matter, University of Pennsylvania, Philadelphia, Pennsylvania 19104 (Received 23 July 1980) The absoulte spin susceptibility of AsF&-doped trans-polyacetylene, [CH(AsF&)~] „, was determined by spin resonance techniques for 0.0004~y 0. 138. The Curie-law contribu- tion decreases above y = 0.001 to a level less than 1 ppm in the highly conducting regime. The temperature-independent Pauli term is small ((5&10 emu/mole) for y &0. 01, re- mains small up to y =0. 05, and shows an abrupt increase near y =0. 07 to a value of 3 ~10 emu/mole at higher concentrations. The results are consistent with the soliton doping mechanis m. PACS numbers: 75. 20.En, 71. 30. + h, 71. 50. Jr The possibility of doping through the generation of nonlinear solitons" in polyacetylene, (CH)„, has attracted considerable attention in recent studies. Experimental results from measure- ments of the magnetic, " infrared, '" optical, ' and phototransport' properties of lightly doped samples are in agreement with the sol. iton mecha- nism. On the other hand, Tomkiewicz et al. ' have presented evidence against solitons in (CH)„. They argue, from measurements of the magnetic susceptibility (y) at a concentration near y = 0.009, that the properties of doped (CH)„are determined by an inhomogeneous mixture of metallic islands even at light doping levels, and that the semicon- ductor-metal (SM) transition is caused by the percolation of such metallic islands. The impor- tance of the magnetic properties has caused us to reexamine the behavior of the susceptibility in a series of samples carefully prepared to achieve maximum uniformity. In this Letter, we present our results from direct measurements of spin susceptibilities of AsF, -doped trans-(CH)„, i.e. , [CH(AsF, )„]„, over the full range from 0.0004 &y (0. 138, well into the high-conductivity re- gime. Our results are fully consistent with the soliton doping mechanism and indicate that the SM transition is not due to percolation of metallic regions. (CH)„ films were prepared at — 78'C"; the dop- ing utilized a technique developed to achieve a homogeneous distribution of dopant ions. A large flat trans-(CH)„ film (-130 mg after doping) was placed in the center of a 500 ml bulb to which two cold fingers were attached, one entering at the top, the other at the bottom. The film rested on small glass supports, thus permitting both sides of the film to be exposed evenly to the AsF, vapor Afixed . amount of purified AsF„™ (10- 20)g in excess of that required for a given doping level, was transf erred to the upper cold f inger and the apparatus was evacuated and sealed. Dop- ing was carried out by first holding the upper cold finger at — 131'C for 12 h. Independent mass spectroscopic studies showed that the vapor of AsF, (vapor pressure(0. 1 Torr) over the AsF, contained no HF or AsF, . After 12 h the lower cold finger was held at — 196 C for 12 h during which time the upper cold finger was grad- ually permitted to warm to approximately room temperature. After this time all the AsF, had been cryopumped into the lower cold finger. The lower cold finger was then held at -131'C and the upper coM finger at room temperature for 12 h, thus repeating the doping process a second time. This entire procedure was repeated a to- tal of 4 times, and after final cryogenic pumping for 12 h, the films were weighed and loaded into either an NMR tube (-100 mg of sample) or an EPR tube (-2 mg of sample). The use of large samples of film permitted an accurate determina- tion of sample composition. The large volume of the reaction bulb and the small vapor pressure of AsF, (which would decrease still further during the reaction) leads to very slow doping. For the film having a 7' nominal AsF, content, an ele- mental analysis gave the composition [CH, „- (AsF, ~), ~,], . The sum of the analytical per- centages of the constituent elements was 99. 13%. It is therefore assumed that all films contained the AsF, species. Although different anions can be generated by AsF, doping, this method as- sures the presence of the AsF, species. " For lightly doped samples (y ( 0.0055), the spin susceptibilities were obtained using a standard 10-GHz microwave reflection spectrometer. Ab- solute susceptibilities were obtained through cali- bration with a ruby (Cr'+ in Al, o, ) standard. In the higher conductivity regime (y) 0. 0055), the technique of Schumacher and Slichter (SS) was used. The 'H and '~F NMR and the spin reso- 1980 The American Physical Society 1123