6 Discussion

6.1 Does IRAS 16293-2422 have a hot core?

Hot cores have been defined for high-mass sources as small (<0.1 pc), dense ( $n_{{\rm H}_2} > 10^7$ cm^-3) and warm (T >100 K) regions (Walmsley 1992; Kurtz et al. 2000). The chemical signatures of hot cores are high abundances of a wide variety of complex organic molecules and fully hydrogenated molecules such as H₂O, NH₃ and H₂S. These chemical characteristics are thought to arise from thermal evaporation of ice mantles from the grains close to the protostar, followed by rapid high-temperature gas-phase reactions for a period of 10⁴-10⁵yr (Charnley et al. 1992). In this scenario, the fully hydrogenated species are "first generation'' molecules produced by surface chemistry on the grains, whereas complex organic molecules like HC₃N, CH₃OCH₃ and HCOOCH₃ are "second generation'' produced by gas-phase chemistry between evaporated species. To what extent do these physical and chemical characteristics also apply to the low-mass object IRAS 16293-2422?

  Table 7: Summary of derived abundances in the envelope of IRAS 16293-2422.

$^{\rm a}$ Abundance in warm and dense inner part of the envelope $\leq$150 AU in radius around IRAS 16293-2422.
$^{\rm b}$ Abundance in cooler, less dense outer part of the envelope around IRAS 16293-2422.
$^{\rm c}$ From Ohishi et al. (1992), updated using values from (2000) for L134N pos. C .
$^{\rm d}$ From van Dishoeck & Blake (1998) (and references cited); otherwise from Sutton et al. (1995).
$^{\rm e}$ From Bachiller & Pérez Gutiérrez (1997) at position B2 assuming CO/H ₂ = 10^-4.
$^{\rm f}$ Based on Ehrenfreund & Charnley (2000) assuming H₂O(ice)/H $_2 = 5\times10^{-5}$. The ranges reflect the values found for different sources. The value for HNCO is derived from OCN^-, assuming this ice evaporates as HNCO.
$^{\rm g}$ From Bockelée-Morvan et al. (2000) for comet Hale-Bopp at 1 AU, assuming H₂O/H $_2 = 5\times10^{-5}$.
$^{\rm h}$ Assuming that the observed emission or upper limits apply to the inner warm region only.

The dust radiative transfer modelling performed in Sect. 4 clearly indicates that there is evidence for the presence of hot and dense material within $\sim$150 AU of the protostar. Moreover, this material is most probably in a state of collapse towards the protostar. The need for introducing drastic jumps in the abundances of some "first generation'' species like H₂CO and CH₃OH at the evaporation temperature of the ices ($\sim$90 K) further suggests that the ice mantles are liberated in the hot inner regions. Table 7 compares the abundances $f_{{\rm in}}$ of these molecules with those found in interstellar ices and in high-mass hot cores. The table also includes the abundances found in comet Hale-Bopp, which may be representative of interstellar ices. Although the IRAS 16293-2422 abundances are still up to an order of magnitude lower than those found in typical ices, they are comparable to those observed in high-mass hot cores. The high degree of deuterium fractionation measured for these molecules (van Dishoeck et al. 1995; Ceccarelli et al. 2001) is consistent with this interpretation, and probably results from a combination of gas-phase deuterium fractionation in the cold pre-collapsing cloud and grain-surface reactions (Tielens 1983; Roberts & Millar 2000).

The high abundances of the sulfur-bearing species of $\sim$10^-7may also fit with this scenario. Molecules like SO, SO₂, OCS and H₂CS are all predicted to be drastically enhanced by the injection of H₂S into the hot core region (Charnley 1997), but may also be present in the grain mantles themselves. H₂S is detected toward IRAS 16293-2422, but because only a single line is observed, the jump model cannot be uniquely constrained. If all H₂S emission is assumed to come from the hot inner region, its inferred abundance is ${\sim}1\times10^{-7}$. This relatively high H₂S abundance is consistent with values found in a survey of massive hot cores (Hatchell et al. 1998) but an order of magnitude lower than the estimates for Orion (Minh et al. 1990).

The physical characteristics in IRAS 16293-2422 are thus consistent with those of a typical "hot core'' as observed towards many high mass protostars, except for a large change in physical scale. The remaining question is whether the evaporated ices have driven a similarly complex organic chemistry in this low-mass protostar. Among potential "second generation'' products, HC₃N and CH₃CN are observed in fairly large amounts with some evidence for jumps in their abundances. However, the IRAS 16293-2422 spectra do not show the wealth of spectral features due to other complex organics - although the confusion limit is clearly not yet reached. The limits on the abundances of molecules such as CH₃OCH₃ and HCOOCH₃, derived if it is assumed that these molecules are located only in the warm inner part of the envelope, are higher than those found for high-mass protostars (Table 7). Deep integrations down to $\sim$5-10 mK are needed to verify the presence of these complex organic species.

Typically, species such as SO₂, OCS, HC₃N and CH₃CN are assumed to be "second generation'' products, in which case their abundance ratios can be used as "chemical clocks'' to determine the time since the evaporation of the ice mantles. For the density and temperature prevailing in the inner parts of the envelope around IRAS 16293-2422 an age of $\sim$10⁴ yr is inferred from such chemical models, a value that is rather uncertain and sensitive to the cosmic ray ionization rate (e.g., Charnley 1997; Hatchell et al. 1998). However, if the velocity field derived from the dust modelling for the Shu infall model is correct, the transit time for grains and molecules through the warm, dense region surrounding IRAS 16293-2422 is only several hundred years. This is not sufficient time for extensive second generation processing to result, and would naturally explain the lack of large, complex organic species toward this source if their abundances are indeed found to be very low by subsequent deep searches. In such a scenario, the observed abundances would provide a unique opportunity to trace the chemical richness derived from previous stages of grain mantle and gas phase chemistry at sensitivities much higher than those achieved via infrared spectroscopy of icy grains.

Figure 9: Observed line widths ($\Delta V$) of different molecules as functions of their estimated rotational excitation temperature ( $T_{{\rm rot}}$).

6.2 Alternative scenarios

The above discussion demonstrates that the IRAS 16293-2422 data can be consistently interpreted with a "hot core'' in which the ices are liberated from the grains by thermal evaporation above 90 K. However, one problem is that the size of the "hot core'' is only $\sim$150 AU in radius, comparable to the size(s) of the circumstellar disk(s), and much smaller than the binary separation. On this scale, departures from spherical symmetry can be expected to play a role and the disk can shadow part of the inner envelope, keeping it much colder. Moreover, the IRAS 16293-2422 envelope is known to be traversed by outflows. To what extent could these processes affect the interpretation? The data clearly indicate jumps in the abundances of "first generation'' molecules known to be present in interstellar ices, but could these icy mantles also be removed by alternative mechanisms?

Blake et al. (1994) and van Dishoeck et al. (1995) proposed that grain-grain collisions in the turbulent shear zones where the outflow interacts with the envelope can also be effective. This mechanism is observed for the class 0 protostar L1157, where the outflow interaction can be spatially separated from the immediate protostellar environment with single-dish telescopes. The inferred abundances by Bachiller & Pérez Gutiérrez (1997) are included in Table 7 and are seen to also be close to the values found in the inner region of IRAS 16293-2422, especially for the sulfur-bearing species.

An interpretation in which the ices are liberated by mild shocks or turbulence rather than thermal evaporation has two observational consequences. First, the lines of the "first generation'' ice mantle species are expected to be wider and have a different velocity structure than those of molecules located predominantly in the quiescent outer envelope. Figure 9 shows the observed line widths of different molecules as functions of their excitation temperature, where the values are taken from Blake et al. (1994) and van Dishoeck et al. (1995). Molecules with clear "jumps'' in their abundances (SO₂, CH₃OH, CH₃CN) have larger line widths and higher excitation temperatures than molecules which trace the outer envelope (CN, C₂H, DNC, DCO⁺). The line widths observed towards IRAS 16293-2422 are significantly wider than what is typically observed for class 0 sources ($\la$1 km s^-1 for C¹⁸O and C¹⁷O; Jørgensen et al. 2002) The observed widths up to 8 km s^-1could, however, possibly be explained by the infall model with thermal evaporation (Fig. 5) and are not clear-cut evidence for association with the molecular outflow, illustrating the difficulty in disentangling the contributions from various velocity components. Second, the spatial distribution of the "first generation'' molecules will be different. Molecules produced by thermal evaporation should be located within a $\sim$150 AU (1$\arcsec$) radius, whereas in the shock scenario they are expected to coat the walls of the outflow(s). Such an "X-type'' geometry can extend over a much larger region, even though the total mass of warm gas may be comparable to that in the first model. Chemically, there is expected to be little difference between the two scenarios, except perhaps in the "second generation'' products if the liberation by shocks occurs in lower density or temperature gas. Observations at sub-arcsec resolution with the Smithsonian SubMillimeter Array (SMA) and the Atacama Large Millimeter Array (ALMA) are needed to distinguish these scenarios.

Recently, Viti et al. (2001) have suggested, based on chemical modelling, that some molecular abundance ratios are strongly affected by the presence of a shock. In these models, the shocks not only liberate the ice mantles but also drive high-temperature ($\sim$2000 K) reactions. In particular the combination of HCO/H₂CO and NS/CS ratios may be suited for tracing the dynamical history of a hot core. For the HCO/H₂CO ratio we derive an upper limit of $\sim$1, about an order of magnitude larger than any model predictions, whereas no NS data are available. Thus, the abundances derived from the data set cannot be used to constrain their origin in the inner hot region of IRAS 16293-2422.