UC Berkeley: H2CO
On my return I concentrated on another large survey. My idea was that a project requiring long observing times was ideal for a telescope dedicated to the work of a single laboratory, because obtaining such observing time on a large national facility is virtually impossible. The disadvantage lay in the relatively small size of our 85 ft. dish and its consequent low angular resolution. The trick was to find a project in which the resolution would not be too limiting.
Interstellar formaldehyde (H2CO) had been observed in the spectra of bright radio sources, but our low resolution made this study impractical. On the other hand formaldehyde had been observed in large dust clouds in absorption against the 2.7 o K. cosmic background radiation. This surprising discovery led to the conclusion that the excitation of the molecule must be by some non-equilibrium process. Fortunately in further study of this phenomenon high-resolution would not be so critical. My project was the survey 1) of a large number of dust clouds to investigate any relationship between the presence of H2CO and the opacity of the clouds, and 2) of a smaller number of clouds to study the line profiles in detail. First, 381 large, dark dust clouds were examined for the presence of H2CO. For 307 of them no H2CO was detected after 2 1/2 hours of observing, or a total of about 800 hours. For 74 clouds the search was successful, and for each of these clouds details of the line profiles were determined after much longer integration times - altogether a major use of telescope and computer time.
The investment of such a large amount of observing time was necessary because in all cases the H2CO line is very weak (none exceeds an antenna temperature of -0.4oK). Two facts are apparent in the data of this survey. One, the formaldehyde lines are all relatively weak and two, the stronger ones occur in darker dust clouds. First, the uniformly small line depths arise primarily because of the correlation between H2CO optical depth in the cloud and its observed excitation temperature. If we were to observe a cloud with high optical depth and low temperature (that is, far below 2.7 o K.) we would measure a strong, deep line. In this sample, at least, that combination does not occur. The correlation yields only weak lines.
Second, the dependence of line depth on the opacity of the dust cloud is due not (on the average) to the presence of more formaldehyde, but rather to "colder" formaldehyde. The model for a typical formaldehyde cloud has a radial gradient in both density and excitation temperature; the density decreases outward, while the temperature increases outward. (It is reminiscent of the model I derived for conditions in galactic nebulae). The suggestion that the central temperature is lower in darker clouds, where the total density of gas is probably higher, argues that collisions are responsible for the anomalous formaldehyde excitation.
During the course of this survey I found one cloud, not a very dark one, that defied all attempts at analysis. The H2CO line profile, of low intensity and narrow width, is distinctly asymmetrical. The observed formaldehyde transition has six hyperfine components, blended in the observed profiles, and included at their theoretical intensity in the analysis of the large survey. Profiles were observed at 10 different positions within this cloud (L1436) and three of them were represented satisfactorily with the theoretical line profile. This profile shows a slight asymmetry due to the presence of the 0 -- 1 transition at about 20 kHz below the central frequency and at a strength of about 0.3 times the main component. At the other positions no reasonable fit to this profile could be found. The only change that would lead to a satisfactory solution was to allow the hyperfine intensities to be anomalous. The intensity of the 0 -- 1 transition derived is ten times its theoretical value!
Such an unexpected conclusion could well qualify as a "zebra" while there are other possible "horses". One possible explanation is that there are two formaldehyde clouds with nearly the same velocity so that they produce a single, apparently abnormal line profile. (Incidentally, one of my colleagues at Berkeley published a paper in the same issue of the Astrophysical Journal supporting this conclusion -- to my surprise.) In my view the configuration required is very unlikely to occur, and I could find no function with physically meaningful parameters for the two clouds. Another possible model involves a radial gradient in intensity and temperature in the cloud. Line profiles computed on this basis are approximately symmetrical and, therefore, do not represent the observations. On the other hand, no proposal for the nonequilibrium excitation of formaldehyde in dark clouds predicts abnormal intensity of the hyperfine levels.
It was at this time that the search for larger and more complex interstellar molecules became intense. The dream was, I think, to find the presence of "benzene rings" with its implication as an essential precursor of life. I must say it seemed to me rather fanciful. I found that I did not understand the complexities of molecular structure, and (worse) did not care. Still worse, I realized in a discussion with a student that I was pretending to know something that I didn't. It was a shock. Intellectual dishonesty is fatal in any scientific endeavor.