From 13 scans obtained with a double-band pass spectrograph at the Snow telescope at Mount Wilson, interpreted by the method of spectral synthesis, the abundance of gold turns out to be log [N(Au)/N(H)] + 12 = 0.70, assuming loggf= --0.57.
A single line attributable to gold, 23122.79, is found in the solar spectrum. It lies in an extremely crowded region. Hence an interpretation of its intensity profile requires an application of the method the spectrum synthesis.
An earlier determination, based on spectral scans secured with the University of Michigan's spectrometer at the Snow telescope, and utilizing the spectral synthesis method has been reported, (Ross, 1971) and included in a compilation by one of us (Aller, 1968). We did not publish the results in detail because (a) we felt that better observational data could be obtained, and (b) the then employed spectral synthesis procedures utilized what appears to be a wrong model of turbulent velocities. All these synthesis procedures required excessively large damping constants.
A new double-band pass monochromator system has been developed and put into operation with the Michigan spectrometer at the Snow telescope by Walter Mitchell. This system proved eminently satisfactory for our purposes. It gives a spectral resolution 2/A2 = 300000. We made scans both in the direction of increasing and decreasing wavelength at an approximate rate of 0.1 A/rain. The observational points represented in Figure 1 represent the mean of 13 scans secured at Mount Wilson between June 25 and July 1, 1971. The final mean tracing does not differ very much from the earlier result (Ross, 1971).
In the method of spectrum synthesis, one determines log [gfN(el)/N(H)] for each line in the spectral domain under consideration. As input parameters, one must adopt a model of the atmosphere, including values of the macroturbulent velocity, and the microturbulent velocity, 4. For each line you must also select values of the damping constants. We include radiative and van der Waals' damping (calculated by classical procedures), but not Stark broadening. In earlier work large values of van der Waal's damping were required. We neglected micro-turbulence and set the most probable macro-turbulence--2.5 km/s.
Figures 2 and3 show the agreement between observed and calculated spectral scans. We show the effects of neglecting all micro and macro turbulence whatever, and also results of fitting line cores versus line wings. Table I summarizes the results of calculations.
Column (1) gives the identification of the feature whose wavelength is tabulated in the second column. Column (3) gives the lower excitation potential of the line involved,