Title: ------ Late Spectral Evolution of SN 1987A: II. Line Emission Author: ------- Cecilia Kozma & Claes Fransson (Stockholm Observatory) Abstract: --------- Using the temperature and ionization calculated in our previous paper, we model the spectral evolution of SN 1987A. We find that the temperature evolution is directly reflected in the time evolution of the lines. In particular, the IR-catastrophe is seen in the metal lines as a transition from thermal to non-thermal excitation, most clearly in the [O I]$\wll$6300, 6364 lines. The good agreement with observations clearly confirms the predicted optical to IR-transition. Because the line emissivity is independent of temperature in the non-thermal phase, this phase has a strong potential for estimating the total mass of the most abundant elements. The hydrogen lines arise as a result of recombinations following ionizations in the Balmer continuum during the first $\sim$ 500 days, and as a result of non-thermal ionizations later. The distribution of the different zones, and therefore the gamma-ray deposition, is determined from the line profiles of the most important lines, where possible. We find that hydrogen extends into the core to $\lesssim 700 \kms$. The hydrogen envelope has a density profile close to $\rho \propto V^{-2}$ from $2000 - 5000 ~\kms$. The total mass of hydrogen-rich gas is $\sim 7.7 \Msun$, of which $\sim 2.2 \Msun$ is mixed within 2000 $\kms$. The helium mass derived from the line fluxes is sensitive to assumptions about the degree of redistribution in the line. The mass of the helium dominated zone is consistent with $\sim 1.9 \Msun$, with a further $\sim 3.9 \Msun$ of helium residing in the hydrogen component. Most of the oxygen-rich gas is confined to 400 -- 2000 $\kms$, with a total mass of $\sim 1.9 \Msun$. Because of uncertainties in the modeling of the non-thermal excitation of the [O I] lines, the uncertainty in the oxygen mass is considerable. In addition, masses of nitrogen, neon, magnesium, iron and nickel are estimated. The dominant contribution to the line luminosity often originates in a different zone from where most of the newly synthesized material resides. This applies to e.g. carbon, calcium and iron. The [C I] lines, mainly arising in the helium zone, indicate a substantially lower abundance of carbon mixed with helium than stellar evolution models give, and a more extended zone with CNO processed gas is indicated. The [Fe II] lines have in most phases a strong contribution from primordial iron, and at $t \gtrsim 600 - 800$ days this component dominates the [Fe II] lines. The wings of the [Fe II] lines may therefore come from primordial iron, rather than synthesized iron mixed to high velocity. Lines from ions with low ionization potential indicate that the UV field below at least $1600 $ \AA~ is severely quenched by dust absorption and resonance scattering. To appear in: ------------- ApJ (Main Journal); scheduled for April 10, 1998, Vol. 497.