The [O I] 3P1 → 3P2 fine-structure line at 63.18 μm has been observed at nine locations in a 1′ × 2′ region in the supernova remnant (SNR) IC 443. Upper limits to the line emission from [Si II] 2P3/2 → 2P1/2 at 34.8 μm and CO (J = 22-21) at 118.6 μm have also been obtained. These are the first reported observations of far-infrared line emission from an SNR. The peak [O I] line flux occurs at the same location as the shockexcited molecular hydrogen emission peak. Comparison with the IRAS far-IR emission shows that the [O I] 63 μm line emission is an important contributor to the 60 μm band, with estimates ranging from ∼40%-75% of the total band flux. This is the first region of this type to be discovered in the IRAS data set. The distribution of the [O I] line emission appears to be similar to that of the molecular hydrogen 1-0 S(1) line. We consider the possibility of X-ray or far-UV excitation of the emission, but we conclude that the [O I] line emission is shock-excited. The IR line emission is modeled using a variety of C and J shocks. We find, based on our current theoretical understanding of shock processes, that C shocks with a distribution of shock velocities from 10-45 km s-1 are present, with the lower velocity C shocks providing the [O I] 63 μm line emission and the higher velocity C shocks the 1-0 S(1) line emission. This model cannot, however, explain why the [O I] 63 μm/H2 1-0 S(1) and H2 1-0 S(1)/2-1 S(1) line ratios appear to remain relatively constant across the source, unless the ad hoc assumption is made that the distribution of shock velocities is the same in every beam. However, by relaxing certain theoretical assumptions, a partially dissociative J shock model with υs ∼ 10-20 km s-1 can explain the observations. It has to be assumed that both (1) the shock is indeed J type despite suspicions that the physical parameters of the flow are appropriate to C shocks and (2) the oxygen chemistry is suppressed so that H2O and OH are not formed. We can offer little theoretical rationale for making these assumptions. Such a shock model has, however, been found by Brand et al. to provide an excellent fit to the H2 line emission from OMC-1. The far-IR shocked line emission (O I, H2O, CO) may well dominate the total emission in the IRAS bands, with [O I] 63 μm line emission dominating the 60 μm band, and with H2O and/or CO line emission contributing appreciably to the 100 μm band. Thus, previous models involving collisionally heated grains to explain the IRAS emission from the shocked molecular region may be unnecessary.
