ELECTRON-POSITRON PAIR PRODUCTION AND ANNIHILATION SPECTRAL FEATURES FROM COMPACT SOURCES

We study the efficiency of e(+/-) pair production in compact gamma-ray sources, as measured by the fraction of power supplied to a source that is converted into pair annihilation radiation emitted by the source (the pair yield). We focus mainly on thermal plasmas, in contrast to previous studies of the pair yield in nonthermal plasmas, We calculate the pair yield in thermal plasmas where bremsstrahlung is the only source of soft photons as well in plasmas in which there is a copious source of soft photons, We find that the pair yield reaches its maximum of similar to 0.3 in Comptonized bremsstrahlung plasmas. The corresponding maximum fraction of source power converted into pair rest mass is similar to 0.2. Addition of ambient soft photons to the source reduces the pair yield. The maximum yield is still close to similar to 0.3 in plasmas where the soft photons are Comptonized into hard power laws of energy spectral index alpha similar or equal to 0.1-0.2, but decreases with increasing a and becomes < 0.01 for a greater than or similar to 0.7, Even when the pair yield is large, we never find an observable annihilation feature from a thermal plasma. To have a high pair yield, typically either the plasma temperature must be relativistic or photons must undergo many Compton scatterings before escaping the source. Hence, the emergent annihilation spectrum is either intrinsically broad or strongly smeared by repeated Compton scatterings and is difficult to distinguish from the underlying Compton scattering and bremsstrahlung continuum. This is in contrast to the nonthermal case where strong, narrow annihilation features can be produced by pairs which have cooled in the source. We summarize the conditions when such strong features are produced, and calculate the pair yield when primary electrons are injected with power-law energy distributions. While observable annihilation features are not produced directly in thermal sources, strong pair outflows, e.g., electron-positron jets, may be created in sources with hard spectra. The outflowing pairs may cool and annihilate outside the source region, in which case a visible annihilation feature is possible. However, our results applied to Nova Muscae show that neither thermal nor nonthermal pair models can explain the large strength of the annihilation feature observed from that source (unless a complex obscuration scenario is invoked). On the other hand, the annihilation feature of the 1979 March 5 gamma-ray burst can be explained by our models provided the source is located at a distance much less than that to the LMC.