research

Background: The chemical composition of the Interstellar Medium (ISM) depends largely on what evolved stars (asymptotic giant branch or, AGB, stars), inject into space at the end of their lifetimes. Much of the mass returned by stars to the ISM is probably lost by intermediate-mass stars (1 — 3 M_sun) during their tenure on the AGB. The processed elements in their atmospheres and convective envelopes eventually will fuel the next generation of star formation. Carbon stars are prodigious mass losers, and although only ~10% of stars on the AGB are carbon-rich they provide ~50% of the carbon in the interstellar medium. How mass loss occurs and evolves over a star's tenure on the AGB is not well understood, although atmospheric pulsations are likely to play an important role. My dissertation research probed the link between atmospheric structure and the mass loss process of carbon stars. I obtained high resolution optical spectra of bright carbon stars in the Solar Neighborhood over a four year period using the Hamilton Echelle spectrograph on the 3 m telescope at Lick Observatory. The resulting spectra produced broad wavelength coverage (from 5085 to 7800 Å) of 87 carbon stars, 67 with multi-epoch observations. My dissertation research addressed the atmospheric kinematics of these large, tenuous stars by considering the velocities of atomic transitions, both in absorption and emission. Comparing these atmospheric velocities with center-of-mass velocities (as measured by radio CO rotational emission lines that arise in the dusty circumstellar envelope) I showed that optical velocities of carbons stars are not reliable measures of center-of-mass motion. This results cast doubt on experiments that have used optical velocities of carbon stars to infer high mass-to-luminosity ratios in nearby dwarf spheroidal galaxies. My work also found fundamental differences in atmospheric structure between Mira and Semi-Regular carbon stars. During my postdoctoral year at the University of California at Berkeley, my efforts concentrated on finishing a spectral atlas of carbon stars that was an offshoot of my dissertation. Atmospheres of AGB stars, both carbon- and oxygen-rich have been difficult for theorists to model due to the very complicated opacities and tenuous atmospheres of these stars. Recently, new efforts have promised vastly improved results. A test of these models atmospheres, in part, will be direct comparison with high resolution optical spectra. The atlas provides sample spectra for six carbon stars over all 40 echelle orders, with atomic and molecular features identified, and is a fundamental tool that will permit detailed, comparative abundance analyses among the different types of carbon stars in the sample; for example, the occurrence of s-process enhancement and carbon isotope ratios (see below). I have made these data available to the scientific community.

As a Jansky Fellow at NRAO, I expanded my research interests to include radio and millimeter observations of circumstellar envelopes and also of interstellar material in the outer Galaxy. I have also begun a collaboration with a multi-national group that will combine both radio and optical observations of cool, evolved stars. While at STScI, I collaborated with an NRAO engineer to build an interference-excision radio receiver. The sections below describe the specific strategies for both current and future projects.

Technetium and s-Process Elements– The atmospheres of AGB carbon stars are unique windows into the stellar interior, since it is during this stage of stellar evolution that material normally locked away from view is dredged-up to the surface. The unusual chemical abundance in the atmosphere provides clues to nucleosynthetic processes and to the internal structure of the stars. Elements synthesized by neutron capture (s-process) in the stellar interior are dredged up to the surface by thermal pulses. For example, ⁹⁹Tc, a short-lived isotope, is believed to indicate recent enrichment of s-process elements and is, therefore, a signature of recent dredge-up. The blue resonance lines of ⁹⁹Tc have been observed previously in some carbon and oxygen-rich AGB stars. However, since AGB stars are very faint in this spectral region (near 4200 Å), integration times are very long and observations are limited to only the brightest stars. I have investigated a more easily observable transition of ⁹⁹Tc in the red, near 6085 Å. Although this transition is heavily blended by molecular lines, the presence of the feature in the red is correlated with detections and non-detections of Tc in the blue made by Little et al. (1987) for 10 stars common with my sample. I am currently attempting to categorize the 87 carbon stars in my sample by the presence or absence of this feature. In addition, I have identified and measured other s-process elements in the spectra of my sample of carbon stars and plan to compare the statistics with those of oxygen-rich AGB stars. Initial results show that as many as 15% of carbon stars enhanced in other s-process elements lack Tc, which suggests that these stars have experienced mass transfer from a companion.

¹²C/¹³C Abundance Ratio – The carbon isotope ratio in evolved stars is a reflection of the nuclear processes in the stellar interior. This enriched material brought up from inner depths and is then deposited into the cool circumstellar envelope where carbon-rich molecules and dust grains form. Carbon isotope ratios, for example, are affected by the relative roles of the triple-a and CNO tri-cycle that alter the amount of ¹³C produced; understanding the relative contribution of these processes can help constrain the mass range of these AGB stars, as well as reveal details of the element-building processes dependent on the ¹³C neutron source. I am currently developing a cross-correlation analysis to determine the ¹²C/¹³C ratio of the 87 carbon stars in my sample. It has been shown that carbon stars typically have enhanced ¹³C (¹²C/¹³C ~ 30) relative to solar abundance (¹²C/¹³C ~ 89), yet a distinct group of carbon stars shows greatly enriched ¹³C, with ¹²C/¹³C ~ 3; their origin and evolution are unclear. Determining whether ¹³C is enriched is relatively straightforward, but obtaining accurate abundance ratios in these stars is difficult due to their very complicated molecular optical spectra. Curve-of-growth analysis is unreliable in this case since: 1) the estimated placement of the continuum is very uncertain, and 2) the wings of absorption lines are heavily blended due to very closely spaced rotational transitions; both of these cause large errors in measuring equivalent widths. The difficulty of these measurements is reflected in the large range of reported values for the same stars. I plan to employ an alternative method that uses a coherence function to measure the relative similarity of two spectra within given frequency ranges. The coherence function plays the role of a correlation coefficient at each frequency. The average coherence is calculated for all frequencies over a chosen range and results in one number characterizing the similarity between two spectra. This method is independent of the continuum level, does not depend on individual lines, and uses the entire spectrum. The echelle data I have obtained are ideal for this technique given their high resolution (Dl ~ 60,000) and broad wavelength coverage (5000 to 7800 Å).

Cold IRAS Stars with Gas-Deficient Circumstellar Envelopes– IRAS has provided astronomers with a valuable tool to study the dusty envelopes of evolved stars. Infrared and radio measurements combined provide details of the dust and gas contributions, respectively, to circumstellar envelopes. Until recently, the correlation of the flux at 60 mm to that of circumstellar CO (J=1-0) appeared to be well established for oxygen-rich, evolved stars with intermediate mass loss rates. However, on the basis of more sensitive searches for CO with the 30 m telescope at IRAM, my collaborators and I have found an unexpected group of objects, apparently deficient in circumstellar molecular gas. As many as 245 of 400 oxygen-rich objects in our sample were not detected in CO, although their IRAS flux densities predicted that an observable amount of gas should be present. As part of a multi-national collaboration to investigate the nature of these interesting stars (with A. Omont, former Director of l'Institut d'Astrophysique de Paris, as lead investigator), we have submitted two of three papers to Astronomy and Astrophysics that investigate the nature of these gas-deficient objects. As a complimentary effort, I observed 90 of these stars at Lick Observatory with the Kast spectrograph to determine their luminosity class and spectral type. Initial results show that this group contains a large number of dusty supergiants. Since most IRAS sources have no visible spectroscopy, the proportion of red supergiants to AGB giants in the sample was hitherto unknown. Our third paper in the series will address the optical spectroscopy of the sample. We have since begun a collaboration with Josef Hron of the University of Vienna and D. J. MacConnell of STScI to help with the project.

The Diffuse Interstellar Bands (DIB's) are a set of over one hundred absorption features extending from 4430 Å, in the blue to the near infrared and are observed in many different regions of the interstellar medium. Although they have been known for over 60 years, their origin is still a mystery. One of the most important questions is whether the carriers are formed in circumstellar envelopes or in the Interstellar Medium. Le Bertre & Lequeux (1993) have reported observations of diffuse absorption bands in the spectra of a series of stars surrounded by circumstellar shells. They find evidence for absorption bands in a variety of sources, both carbon-rich (C/O > 1) and oxygen-rich (C/O < 1) and claim that diffuse band carriers are being produced in both types of objects. However, many of their sources are found in the Galactic plane where it is very difficult to distinguish between interstellar and circumstellar material. Although at least some carriers of DIBs most likely originate in circumstellar outflows, they are not expected to be seen there in absorption since the column density should be insufficient for a detection. Under conditions like in the red rectangle, a strong ultraviolet flux from the central source can excite the carriers into emission, but otherwise, they are not seen. We observed 10 stars at a variety of galactic latitudes that have both an infrared excess and smooth continuum (i.e., early spectral type). This small sample is supplemented by 391 stars observed for DIBs by Duke (1951). Statistical arguments based on these results suggest that the strength of DIBs toward these stars is correlated with galactic latitude rather than with the amount of circumstellar material. For rhe next proposal cycle for HST, A. Omont will join us to propose STIS observations of the stars of LeBertre & Lequeux.

The Unusual Circumstellar Environment of U Equ – In the course of the optical spectral survey of the cold IRAS objects described in the paragraph above, we discovered a star with a highly unusual optical spectrum, characterized by anomalously thin, deep molecular absorption and by bright emission lines of oxygen-rich molecular species such as TiO, VO and AlO. The spectacular optical spectrum is unlike any of the 100 target stars and standards that we have observed. U Equ is likely to have a cold disk of dusty material, since the IRAS flux ratios indicate optically thin material, yet the LRS spectrum shows silicate absorption, usually seen only in optically thick envelopes. We have observed unusual water and OH masers with profiles that vary on a timescale of a few months. I am now modeling the circumstellar environment that might give rise to the very cold TiO and AlO absorption lines. Initial results predict that the temperature of these lines might be as low as 700 K, well below the dust condensation temperature. It is possible that some disruptive event, such as the evaporation of a large planet, might have taken place to return Ti, a highly refractory element, to the gaseous phase. UKIRT spectra near 2µ show CO absorption, but like the optical spectra, the higher rotation levels are not populated. Keith Noll of STScI, Tom Geballe of UH, and I are in the process of writing up the infrared results.

Buckminsterfullerene: a Search for C₆₀ at Low Frequencies Buckminsterfullerene, a soccer-ball construction of 60 carbon atoms, has been found to be extremely stable and is suspected of being produced in the dusty circumstellar envelopes of carbon stars. Laboratory experiments have shown that C₆₀ forms spontaneously in a hot, oxygen-depleted environment and is physically stable even under violent conditions, although it is chemically reactive. The physical stability is attributed to the closed form of the molecule. The warm, oxygen-depleted atmospheres of carbon stars might plausibly be conducive to the formation of C₆₀. Since Buckminsterfullerene apparently survives in UV radiation fields, it is possible that a significant fraction of the C₆₀ injected into interstellar space by generations of carbon stars still survives in the interstellar medium. The question has been where in the electromagnetic spectrum to look for C₆₀. A hyperfine transition of hydrogenated Buckminsterfullerene (HC₆₀) was recently found experimentally at 94 MHz by chemists Morton et al. 1993. The hydrogenated form has mirror symmetry, and it is the spin flip of the associated hydrogen that causes the transition, similar to the 21 cm line of H I. The chemists suggest that radio astronomers search for it near this frequency.

A New Approach to RFI–the First Adaptive Filter Receiver in Radio Astronomy Since our window of interest for the C₆₀ project , 94 MHz, lies in the FM broadcast band, we are faced with overcoming severe radio frequency interference (RFI). My collaborators are chemist S. Balm of UCLA and electrical engineer R. Bradley of NRAO. Our approach to the problem takes advantage of relatively new digital filtering technology never before used in radio astronomy (adaptive filtering). The success of our adaptive filter will essentially open up new observing windows. Also, more and more, RFI is becoming a serious problem even in protected bands at higher frequencies (especially at L-band) and must be addressed in the next decade if we are to continue useful astronomical observations at these frequencies. We published a seminal paper describing preliminary results. At every RFI conference since, our work has been cited and heralded as an important new direction to pursue.

See a Summary of my Research at VSU

Details of My Research

	Background: The chemical composition of the Interstellar Medium (ISM) depends largely on what evolved stars (asymptotic giant branch or, AGB, stars), inject into space at the end of their lifetimes. Much of the mass returned by stars to the ISM is probably lost by intermediate-mass stars (1 — 3 M_sun) during their tenure on the AGB. The processed elements in their atmospheres and convective envelopes eventually will fuel the next generation of star formation. Carbon stars are prodigious mass losers, and although only ~10% of stars on the AGB are carbon-rich they provide ~50% of the carbon in the interstellar medium. How mass loss occurs and evolves over a star's tenure on the AGB is not well understood, although atmospheric pulsations are likely to play an important role. My dissertation research probed the link between atmospheric structure and the mass loss process of carbon stars. I obtained high resolution optical spectra of bright carbon stars in the Solar Neighborhood over a four year period using the Hamilton Echelle spectrograph on the 3 m telescope at Lick Observatory. The resulting spectra produced broad wavelength coverage (from 5085 to 7800 Å) of 87 carbon stars, 67 with multi-epoch observations. My dissertation research addressed the atmospheric kinematics of these large, tenuous stars by considering the velocities of atomic transitions, both in absorption and emission. Comparing these atmospheric velocities with center-of-mass velocities (as measured by radio CO rotational emission lines that arise in the dusty circumstellar envelope) I showed that optical velocities of carbons stars are not reliable measures of center-of-mass motion. This results cast doubt on experiments that have used optical velocities of carbon stars to infer high mass-to-luminosity ratios in nearby dwarf spheroidal galaxies. My work also found fundamental differences in atmospheric structure between Mira and Semi-Regular carbon stars. During my postdoctoral year at the University of California at Berkeley, my efforts concentrated on finishing a spectral atlas of carbon stars that was an offshoot of my dissertation. Atmospheres of AGB stars, both carbon- and oxygen-rich have been difficult for theorists to model due to the very complicated opacities and tenuous atmospheres of these stars. Recently, new efforts have promised vastly improved results. A test of these models atmospheres, in part, will be direct comparison with high resolution optical spectra. The atlas provides sample spectra for six carbon stars over all 40 echelle orders, with atomic and molecular features identified, and is a fundamental tool that will permit detailed, comparative abundance analyses among the different types of carbon stars in the sample; for example, the occurrence of s-process enhancement and carbon isotope ratios (see below). I have made these data available to the scientific community. As a Jansky Fellow at NRAO, I expanded my research interests to include radio and millimeter observations of circumstellar envelopes and also of interstellar material in the outer Galaxy. I have also begun a collaboration with a multi-national group that will combine both radio and optical observations of cool, evolved stars. While at STScI, I collaborated with an NRAO engineer to build an interference-excision radio receiver. The sections below describe the specific strategies for both current and future projects. Technetium and s-Process Elements– The atmospheres of AGB carbon stars are unique windows into the stellar interior, since it is during this stage of stellar evolution that material normally locked away from view is dredged-up to the surface. The unusual chemical abundance in the atmosphere provides clues to nucleosynthetic processes and to the internal structure of the stars. Elements synthesized by neutron capture (s-process) in the stellar interior are dredged up to the surface by thermal pulses. For example, ⁹⁹Tc, a short-lived isotope, is believed to indicate recent enrichment of s-process elements and is, therefore, a signature of recent dredge-up. The blue resonance lines of ⁹⁹Tc have been observed previously in some carbon and oxygen-rich AGB stars. However, since AGB stars are very faint in this spectral region (near 4200 Å), integration times are very long and observations are limited to only the brightest stars. I have investigated a more easily observable transition of ⁹⁹Tc in the red, near 6085 Å. Although this transition is heavily blended by molecular lines, the presence of the feature in the red is correlated with detections and non-detections of Tc in the blue made by Little et al. (1987) for 10 stars common with my sample. I am currently attempting to categorize the 87 carbon stars in my sample by the presence or absence of this feature. In addition, I have identified and measured other s-process elements in the spectra of my sample of carbon stars and plan to compare the statistics with those of oxygen-rich AGB stars. Initial results show that as many as 15% of carbon stars enhanced in other s-process elements lack Tc, which suggests that these stars have experienced mass transfer from a companion. ¹²C/¹³C Abundance Ratio – The carbon isotope ratio in evolved stars is a reflection of the nuclear processes in the stellar interior. This enriched material brought up from inner depths and is then deposited into the cool circumstellar envelope where carbon-rich molecules and dust grains form. Carbon isotope ratios, for example, are affected by the relative roles of the triple-a and CNO tri-cycle that alter the amount of ¹³C produced; understanding the relative contribution of these processes can help constrain the mass range of these AGB stars, as well as reveal details of the element-building processes dependent on the ¹³C neutron source. I am currently developing a cross-correlation analysis to determine the ¹²C/¹³C ratio of the 87 carbon stars in my sample. It has been shown that carbon stars typically have enhanced ¹³C (¹²C/¹³C ~ 30) relative to solar abundance (¹²C/¹³C ~ 89), yet a distinct group of carbon stars shows greatly enriched ¹³C, with ¹²C/¹³C ~ 3; their origin and evolution are unclear. Determining whether ¹³C is enriched is relatively straightforward, but obtaining accurate abundance ratios in these stars is difficult due to their very complicated molecular optical spectra. Curve-of-growth analysis is unreliable in this case since: 1) the estimated placement of the continuum is very uncertain, and 2) the wings of absorption lines are heavily blended due to very closely spaced rotational transitions; both of these cause large errors in measuring equivalent widths. The difficulty of these measurements is reflected in the large range of reported values for the same stars. I plan to employ an alternative method that uses a coherence function to measure the relative similarity of two spectra within given frequency ranges. The coherence function plays the role of a correlation coefficient at each frequency. The average coherence is calculated for all frequencies over a chosen range and results in one number characterizing the similarity between two spectra. This method is independent of the continuum level, does not depend on individual lines, and uses the entire spectrum. The echelle data I have obtained are ideal for this technique given their high resolution (Dl ~ 60,000) and broad wavelength coverage (5000 to 7800 Å). Cold IRAS Stars with Gas-Deficient Circumstellar Envelopes– IRAS has provided astronomers with a valuable tool to study the dusty envelopes of evolved stars. Infrared and radio measurements combined provide details of the dust and gas contributions, respectively, to circumstellar envelopes. Until recently, the correlation of the flux at 60 mm to that of circumstellar CO (J=1-0) appeared to be well established for oxygen-rich, evolved stars with intermediate mass loss rates. However, on the basis of more sensitive searches for CO with the 30 m telescope at IRAM, my collaborators and I have found an unexpected group of objects, apparently deficient in circumstellar molecular gas. As many as 245 of 400 oxygen-rich objects in our sample were not detected in CO, although their IRAS flux densities predicted that an observable amount of gas should be present. As part of a multi-national collaboration to investigate the nature of these interesting stars (with A. Omont, former Director of l'Institut d'Astrophysique de Paris, as lead investigator), we have submitted two of three papers to Astronomy and Astrophysics that investigate the nature of these gas-deficient objects. As a complimentary effort, I observed 90 of these stars at Lick Observatory with the Kast spectrograph to determine their luminosity class and spectral type. Initial results show that this group contains a large number of dusty supergiants. Since most IRAS sources have no visible spectroscopy, the proportion of red supergiants to AGB giants in the sample was hitherto unknown. Our third paper in the series will address the optical spectroscopy of the sample. We have since begun a collaboration with Josef Hron of the University of Vienna and D. J. MacConnell of STScI to help with the project. The Diffuse Interstellar Bands (DIB's) are a set of over one hundred absorption features extending from 4430 Å, in the blue to the near infrared and are observed in many different regions of the interstellar medium. Although they have been known for over 60 years, their origin is still a mystery. One of the most important questions is whether the carriers are formed in circumstellar envelopes or in the Interstellar Medium. Le Bertre & Lequeux (1993) have reported observations of diffuse absorption bands in the spectra of a series of stars surrounded by circumstellar shells. They find evidence for absorption bands in a variety of sources, both carbon-rich (C/O > 1) and oxygen-rich (C/O < 1) and claim that diffuse band carriers are being produced in both types of objects. However, many of their sources are found in the Galactic plane where it is very difficult to distinguish between interstellar and circumstellar material. Although at least some carriers of DIBs most likely originate in circumstellar outflows, they are not expected to be seen there in absorption since the column density should be insufficient for a detection. Under conditions like in the red rectangle, a strong ultraviolet flux from the central source can excite the carriers into emission, but otherwise, they are not seen. We observed 10 stars at a variety of galactic latitudes that have both an infrared excess and smooth continuum (i.e., early spectral type). This small sample is supplemented by 391 stars observed for DIBs by Duke (1951). Statistical arguments based on these results suggest that the strength of DIBs toward these stars is correlated with galactic latitude rather than with the amount of circumstellar material. For rhe next proposal cycle for HST, A. Omont will join us to propose STIS observations of the stars of LeBertre & Lequeux. The Unusual Circumstellar Environment of U Equ – In the course of the optical spectral survey of the cold IRAS objects described in the paragraph above, we discovered a star with a highly unusual optical spectrum, characterized by anomalously thin, deep molecular absorption and by bright emission lines of oxygen-rich molecular species such as TiO, VO and AlO. The spectacular optical spectrum is unlike any of the 100 target stars and standards that we have observed. U Equ is likely to have a cold disk of dusty material, since the IRAS flux ratios indicate optically thin material, yet the LRS spectrum shows silicate absorption, usually seen only in optically thick envelopes. We have observed unusual water and OH masers with profiles that vary on a timescale of a few months. I am now modeling the circumstellar environment that might give rise to the very cold TiO and AlO absorption lines. Initial results predict that the temperature of these lines might be as low as 700 K, well below the dust condensation temperature. It is possible that some disruptive event, such as the evaporation of a large planet, might have taken place to return Ti, a highly refractory element, to the gaseous phase. UKIRT spectra near 2µ show CO absorption, but like the optical spectra, the higher rotation levels are not populated. Keith Noll of STScI, Tom Geballe of UH, and I are in the process of writing up the infrared results. Buckminsterfullerene: a Search for C₆₀ at Low Frequencies Buckminsterfullerene, a soccer-ball construction of 60 carbon atoms, has been found to be extremely stable and is suspected of being produced in the dusty circumstellar envelopes of carbon stars. Laboratory experiments have shown that C₆₀ forms spontaneously in a hot, oxygen-depleted environment and is physically stable even under violent conditions, although it is chemically reactive. The physical stability is attributed to the closed form of the molecule. The warm, oxygen-depleted atmospheres of carbon stars might plausibly be conducive to the formation of C₆₀. Since Buckminsterfullerene apparently survives in UV radiation fields, it is possible that a significant fraction of the C₆₀ injected into interstellar space by generations of carbon stars still survives in the interstellar medium. The question has been where in the electromagnetic spectrum to look for C₆₀. A hyperfine transition of hydrogenated Buckminsterfullerene (HC₆₀) was recently found experimentally at 94 MHz by chemists Morton et al. 1993. The hydrogenated form has mirror symmetry, and it is the spin flip of the associated hydrogen that causes the transition, similar to the 21 cm line of H I. The chemists suggest that radio astronomers search for it near this frequency. A New Approach to RFI–the First Adaptive Filter Receiver in Radio Astronomy Since our window of interest for the C₆₀ project , 94 MHz, lies in the FM broadcast band, we are faced with overcoming severe radio frequency interference (RFI). My collaborators are chemist S. Balm of UCLA and electrical engineer R. Bradley of NRAO. Our approach to the problem takes advantage of relatively new digital filtering technology never before used in radio astronomy (adaptive filtering). The success of our adaptive filter will essentially open up new observing windows. Also, more and more, RFI is becoming a serious problem even in protected bands at higher frequencies (especially at L-band) and must be addressed in the next decade if we are to continue useful astronomical observations at these frequencies. We published a seminal paper describing preliminary results. At every RFI conference since, our work has been cited and heralded as an important new direction to pursue. See a Summary of my Research at VSU

		Department of Phyics, Astronomy & Geosciences Nevins Hall Valdosta State University Valdosta, GA 31698-0055 tel: (229) 333-5752