The maps presented in Paper I are here used to infer the variation of the column densities of HCO+, DCO+, N2H+, and N2D+ as a function of distance from the dust peak. These results are interpreted with the aid of a crude chemical model which predicts the abundances of these species as a function of radius in a spherically symmetric model with radial density distribution inferred from the observations of dust emission at millimeter wavelengths and dust absorption in the infrared. Our main observational finding is that the N(N2D+)/N(N2H+) column density ratio is of order 0.2 towards the L1544 dust peak as compared to N(DCO+)/N(HCO+) = 0.04. We conclude that this result as well as the general finding that N2H+ and N2D+ correlate well with the dust is caused by CO being depleted to a much higher degree than molecular nitrogen in the high density core of L1544. Depletion also favors deuterium enhancement and thus N2D+, which traces the dense and highly CO-depleted core nucleus, is much more enhanced than DCO+. Our models do not uniquely define the chemistry in the high density depleted nucleus of L1544 but they do suggest that the ionization degree is a few times 10^{-9} and that the ambipolar diffusion time scale is locally similar to the free fall time. It seems likely that the lower limit which one obtains to ionization degree by summing all observable molecular ions is not a great underestimate of the true ionization degree. We predict that atomic oxygen is abundant in the dense core and, if so, H3O+ may be the main ion in the central highly depleted region of the core.

We have mapped the dense dark core L1544 in H13CO+(1-0), DCO+(2-1), DCO+(3-2), N2H+(1-0), NTH+(3-2), N2D+(2-1), N2D+(3-2), C18O(1-0), and C17O(1-0) using the IRAM 30-m telescope. We have obtained supplementary observations of HC18O+(1-0), HC17O+(1-0), and D13CO+(2-1). Many of the observed maps show a general correlation with the distribution of dust continuum emission in contrast to C18O(1-0) and C17O(1-0) which give clear evidence for depletion of CO at positions close to the continuum peak. In particular N2D+(2-1) and (3-2) and to a lesser extent N2H+(1-0) appear to be excellent tracers of the dust continuum. We find that the tracers of high density gas (in particular N2D+) show a velocity gradient along the minor axis of the L1544 core and that there is evidence for larger linewidths close to the dust emission peak. We interpret this using the model of the L1544 proposed by Ciolek & Basu (2000) and by comparing the observed velocities with those expected on the basis of their model. The results show reasonable agreement between observations and model in that the velocity gradient along the minor axis and the line broadening toward the center of L1544 are predicted by the model. This is evidence in favour of the idea that amipolar diffusion across field lines is one of the basic processes leading to gravitational collapse. However, line widths are significantly narrower than observed and are better reproduced by the Myers & Zweibel (2001) model which considers the quasistatic vertical contraction of a layer due to dissipation of its Alfvenic turbulence, indicating the importance of this process for cores in the verge of forming a star.

We have computed the dust temperature distribution to be expected in a pre-protostellar core in the phase prior to the onset of gravitational instability. We have done this in the approximation that the heating of the dust grains is solely due to the attenuated external radiation field and that the core is optically thin to its own radiation. This permits us to consider non spherically symmetric geometries. We predict the intensity distributions of our model cores at millimeter and sub-millimeter wavelengths and compare with observations of the well studied object L1544. We have also developed an analytical approximation for the temperature at the center of spherically symmetric cores and we compare this with the numerical calculations. Our results show (in agreement with Evans et al. 2001) that the temperatures in the nuclei of cores of high visual extinction (> 30 magnitudes) are reduced to values of below ~8 K or roughly half of the surface temperature. This has the consequence that maps at wavelengths shortward of 1.3 mm see predominantly the low density exterior of pre-protostellar cores. It is extremely difficult to deduce the true density distribution from such maps alone. We have computed the intensity distribution expected on the basis of the models of Ciolek & Basu (2000) and compared with the observations of L1544. The agreement is good with a preference for higher inclinations (37 degrees instead of 16) than that adopted by Ciolek & Basu (2000). We find that a simple extension of the analytic approximation allows a reasonably accurate calculation of the dust temperature as a function of radius in cores with density distributions approximating those expected for Bonnor-Ebert spheres and suggest that this may be a useful tool for future calculations of the gas temperature in such cores.