EVOLUTION OF THE 1-4 MU-M SPECTRUM OF NOVA PW VULPECULAE 1984

We present infrared spectroscopy of the slow nova PW Vul (Nova Vul 198 4 #1) between 13 and 328 days after outburst. The hydrogen line spectr um was initially accompanied by lines of C I, O I and N I, which were strongest around 80 days after outburst. The 1.129-mu m O I line excit ed by Ly beta fluorescence dominated our spectra until at least 138 da ys after outburst and was still visible similar to 100 days later. The CNO lines were gradually replaced by the coronal [Si VI], [Mg VIII] a nd [Al VI] lines, first evident similar to 130 days after outburst, an d [Si VII] and [Alv] as the spectrum evolved. The appearance of corona l lines while the O I fluorescence persisted suggests that the ejecta were clumpy. The 1.083-mu m He I line strengthened relative to the H I lines, from similar to 0.4 Pa gamma on Day 78 to similar to 27 Pay on Day 272, at a phase when the coronal lines were conspicuous. This He I line came to dominate the spectrum and significantly affect photomet ry through the J filter. The hydrogen line strengths are in agreement with recombination theory, at low temperature (1000-3000 K) and high d ensity (10(9-10) cm(-3)) on Day 78, and subsequently at higher tempera tures and lower densities. The relative emission measure determined fr om the ratios of observed and theoretical hydrogen line strengths meas ured from our 1984 December and 1985 observations declined as t(-3), a s expected for recombination of a fixed mass in a freely expanding she ll. During the first similar to 130 days of the evolution of the nova characterized by irregular light variations prior to this, the emissio n measure did not decline, indicating continued feeding of the ejecta by the stellar wind. A comparison of infrared hydrogen lines with cont emporaneous optical data yields an extinction of A(V) = 1.78. The stro ngest expected He II and Fe II lines were visible for a while, but for bidden [Fe Ir] was not seen. Before and at maximum the optical-infrare d spectral energy distribution followed F(nu)proportional to nu(1.1), as predicted by slow nova wind models. The spectrum flattened to nu(0. 2) during the initial fading and varied between these values during th e irregular light variations.