Laboratory Data Needs in UV, IR, and X-Ray Astrophysics

J. Michael Shull

University of Colorado, CASA and JILA
Boulder, CO 80309

I provide a brief review of laboratory data needs for UV, IR, and X-ray astrophysics. The scientific areas center primarily on interstellar and intergalactic matter and on active galactic nuclei. From the vantage point of observational astronomers, abundance measurements, ionization corrections, and physical diagnostics usually require accuracies better than 20%. The current level of sophisticated astrophysical modeling in these fields requires increasingly accurate rate coefficients for collisional ionization, photoionization, recombination, electron-impact excitation, resonant-line absorption, and charge exchange. Laboratory measurements and theoretical computations should both be supported. Self-consistency checks between inverse processes are desirable; for example, among photoionization, radiative recombination, and dielectronic recombination. I also comment on the need for firmer identification of the constituents and physical properties of the solid particles (dust grains, PAHs) in the interstellar medium. Because these particles potentially dominate UV absorption and photoelectric heating of the gas, and are key diagnostics in infrared emission, we need more reliable models than the simple PAH paradigm currently used.

Introduction

To this audience, there is little need to dwell on the importance of atomic data in astrophysics. Suffice it to say that nearly all astronomical inference from emission-line and absorption-line spectroscopy relies on the existence of accurate cross sections, rate coefficients, line frequencies, and related parameters about the abundant atoms, ions, and molecules in space. In this paper, I will mention a few of the most important areas for understanding galactic interstellar and intergalactic science. I will focus, in particular, on two related scientific problems that are ``drivers'' for several current and future NASA missions:

• The origin and evolution of heavy elements in the first stars and galaxies
• The cycles of baryons, heavy elements, and dust in the interstellar medium

These scientific issues motivate many of the key projects on major NASA missions, particularly FUSE (the Far Ultraviolet Spectroscopic Explorer), the spectrographs on Hubble Space Telescope (STIS and COS), AXAF (Advanced X-ray Astrophysics Facility), SOFIA (Stratospheric Observatory for Infrared Astronomy), and SIRTF (Space Infra-Red Telescope Facility).

Scientific Issues

Astronomers are generally inclined to study the most abundant elements, whose scientific importance is governed by a combination of scientific factors such as abundance, accessible spectral lines, transition strength, excitation mechanism, and diagnostic importance. The foremost consideration is often the astrophysical abundance, which implicates 20-25 elements spread over a number of ionization stages. One of the virtues of the organization of this conference is that we may discuss a broad range of wavelength bands. However, within any given band (optical, UV, IR, X-ray) there are often certain fundamental factors that limit the accessible elements and ion stages. For instance, carbon is generally inaccessible in the optical and requires either ultraviolet or infrared measurements for direct detection. Similarly, excitation conditions or atomic/molecular parameters often make it difficult to detect key atoms and molecules. In the interstellar or intergalactic gas, most atoms and ions have their resonance asbsorption lines in the ultraviolet, particularly the 900-1200 Å region studied by FUSE. The dominant probes of baryons in the intergalactic medium, H I and He II, are most sensitively detected through their Ly$\alpha$ absorption lines, with rest wavelengths of 1215.670 Å and 303.781 Å, and higher Lyman-series lines. In many cases, these lines are redshifted to longer wavelengths accessible to the HST spectrographs, as well as to FUSE. Another example occurs in molecular gas, in which the dominant species, H₂, is best studied through its far-ultraviolet Lyman and Werner absorption bands or through its near-IR vibrational emission lines. However, current (ISO) and future (SIRTF, SOFIA) instruments may be able to probe warm H₂ through its weak quadrupole emission lines at 28, 17, and 12 $\mu$m.

The elements of significant abundance include H and He, the dominant repositories of baryons, the abundant molecules (H₂ and CO), and the heavy elements C, N, O, Ne, Mg, Si, S, and Fe (all greater than 10^-5 abundance relative to hydrogen by number). Other heavy elements of lower abundance (Na, Al, P, Cl, Ar, K, Ca, Ti) are important because of the accessibility of their lines, while many heavier elements in the iron-group (Cr, Mn, Co, Ni, Cu, Zn) have taken on renewed importance because their utility for diagnosing stellar nucleosynthetic sources. The production of selected heavy elements by certain mass ranges of stars is also significant, as astronomers attempt to relate chemical evolution to stellar progenitors, including the rates of massive star formation, supernovae, and starlight. The ``alpha-capture'' elements (O, Si, S) are especially important ties to massive star formation and Type-II supernovae, while the iron-group elements and carbon are well-tied to low-mass star formation and Type-I supernovae. The stellar source of Zn is currently is not well understood, although its ratio to Ni, Cr, and Fe is widely used as a diagnostic of the presence of interstellar dust (Zn appears not to deplete strongly onto grains).

Molecular studies are also important for UV and IR astronomy. Although H₂ is the dominant molecule in interstellar space, and CO is the most frequently used species to infer molecular cloud masses, these are by no means the only molecular diagnostics. Trace molecules and molecular ions often provide excellent diagnostics of dense cloud cores, outflows from star-forming regions, and cloud chemical networks.

Finally, increasingly important diagnostics of interstellar matter and its radiative environment comes from the solid particles or dust grains. Composed of silicates, graphite, SiC, amorphous carbon, and ice mantles, these grains provide important energy transfer mechanisms from the radiation field to the gas, since they can dominate the UV absorption cross section and photoelectric heating rate of diffuse gas.

Data Needs: Ionization and Recombination

Accurate models of the ionization conditions in astrophysical plasmas, both photoionized or collisionally ionized, require an array of basic atomic data. The critical cross sections or rate coefficients include:

• Photoionization cross sections, $\sigma_{\rm ph}(\nu)$
• Radiative recombination rates, $\alpha_{\rm rad}(T)$
• Dielectronic recombination rates, $\alpha_{\rm di}(T)$
• Collisional ionization rates, $C_i(T) = \langle \sigma_i v \rangle$
• Charge-transfer rates with H^o, He^o, He⁺
• Inner-shell photoionization and Auger electron cascades

Details on how these rate coefficients are used may be found in many books and review articles on atomic processes in astrophysical nebulae and plasmas. (e.g., Osterbrock 1989; Brickhouse 1996; Shull 1993, 1996) Together with photoionization or collisional ionization models, these data are used to derive the ionization fractions, f_Z,z, at a prescribed temperature T or photoionization parameter, $U = n_{\gamma}/n_H$.The fraction of element Z in ion stage z then provides the overall elemental abundance. Although such calculations usually assume ionization equilibrium (Shull & Van Steenberg 1982; Arnaud & Rothenflug 1985; Sutherland & Dopita 1993), many situations involve transient plasmas out of equilibrium, such as hot, post-shock plasmas or recombining plasmas. In equilibrium, the dominant ionization stage is determined primarily by the ionization potential and Boltzmann factors, and the ionization recombination rates need only to be determined to about 20% accuracy. However, disequilibrium plasmas depend more directly on the normalization constant in the rate coefficients, and 10% accuracy would be desirable to achieve the diagnostic goals.

The most critical data needs lie primarily on the recombination side of the equation. Dielectronic recombination rates are not adequately known, particularly at low temperatures, $(1-3) \times 10^4$ K, as are typical of photoionized plasmas. This will require understanding the importance of $\Delta n = 0$ transitions, as well as fine-structure effects (relativistic and spin-orbit spilitting). In the panel discussion, we had a consensus that, for both X-ray and UV studies, a high priority should be placed on a comparison of experimental and theoretical measurements of radiative and dielctronic recombination rates. The plasma experiments may require a combination of storage ring and ion-beam facilities. A plasma experiment can produce a range of ions and a UV/X-ray spectrum under controlled conditions, while an ion-beam experiment can focus on a single species at aparticular energy. Each of these experiments should then be tested against the extensive close-coupling calculations and compared to observations of spectra from hot and photoionized plasmas.

Another important need is accurate calculations and measurements of charge-transfer rates of ions with H I. In many cases, these processes dominate the plasma neutralization. Currently, modelers must rely on theoretical calculations for charge-exchange rates. Some experimental benchmarks would be valuable.

Data Needs: Lines

Once the ionization state has been characterized, the next step is to understand the line formation physics. Since spectroscopic diagnostics come from emission lines and absorption lines, further atomic data are needed to characterize and interpret the line spectra. For absorption lines, accurate abundances usually require oscillator strengths for several lines over a range of strengths, to resolve the ambiguities of saturation (curve-of-growth). The key line parameters include:

• Transition wavelengths, usually accurate to 10^-6 in the UV and 10^-3 for X-rays. Future high-resolution spectrographs will require even higher accuracy.
• Absorption oscillator strengths, f_ij and Einstein transition probabilities, A_ij
• Collisional-excitation cross sections or Maxwellian-averaged collision strengths, $\Omega_{ij}(T)$

Of particular interest are line identifications in of third-row elements (Ne, Mg, Si, S) and heavy elements in the iron-group, e.g., Fe II and Fe III. For the elements that are important for radiative cooling and spectral diagnostics (e.g., C, O, Ne, Si, S, Fe) the electron-impact excitation cross sections and absorption oscillator strengths, resonance structure near threshold need careful attention.

Recent laboratory work by groups at Wisconsin and Toledo have clarified some of the f-value inconsistencies in Si II and Fe II and have made accurate measurements for Cr II, Ni II, Zn II, and Co II. In contrast to the situation five years ago (Shull 1993), the situation has improved considerably. In my talk, I mentioned a few lines that probably need more accurate f-values and collisional-excitation rates, owing to their importance in key science projects with FUSE and HST.

• Accurate f-values are needed for the weak O I and N I lines around 950 Å. These lines will be key for FUSE measurements of the D/H ratio, using O and N as proxies for hydrogen, which is often difficult to determine due to saturated H I lines.
• Numerous far-UV lines of Fe II and Fe III will be studied with FUSE. Line identifications and f-values are needed.
• EUV lines of highly ionized CNO will be studied by HST and FUSE in redshifted absorption-line systems (galactic halos). Accurate f-values are needed
• Studies of the strong far-UV and EUV resonance lines of O VI, S VI, Ne VIII, and Mg XII are key FUSE science. Excitation rates are needed, as well as f-values for the Mg and Si lines below 912 Å.

Data Needs: Molecular Astrophysics

Studies of molecular hydrogen will be among the new results of FUSE. The Lyman and Werner bands below 1126 Å will appear in nearly every spectrum of O stars, quasars, and other AGN. Measuring the abundances, N(J), in the rotational states j = 0,1,2... of H₂ will provide information relevant to the gas kinetic temperature (from J = 0 and 1), the gas density, n_H, and the ultraviolet radiation field, J₁₀₀₀ near 1000 Å that radiatively pumps these states. Thus, with proper modeling, we can infer the run of gas pressure and UV radiation field above the Galactic disk plane, and in the environments of massive star formation. Abundance ratios of CO/H₂, combined with measurements of heavy-element abundances and UV extinction curves, can test photo-chemical models of the transition regions between diffuse clouds and dark star-forming clouds. The ratio, HD/H₂, can be used to infer the cosmic-ray intensity that penetrates these regions. In addition, trace molecules (N₂, O₂, etc.) are often useful diagnostics of the photo-chemical models of photo-dissociation regions and translucent clouds. The data needs for these science problems include:

• Accurate excitation and de-excitation cross sections for the H₂ (J = 1-7) rotational states by H^o.
• Accurate CO photoabsorption cross sections for the (C-X) and (E-X) bands.
• Better theoretical and experimental work on the rates of H₂ catalysis on dust-grain surfaces. Of particular interest is the dependence of the H₂ formation rates on gas and grain temperatures.
• Oscillator strengths for trace molecules (N₂, O₂).

Data Needs: Dust Grains and PAHs

Quantitative studies of solid particles and small clusters are still in their infancy. However, the many opportunities for IR studies, as well the accurate UV spectroscopic studies now available with FUSE and HST, suggest the need for far-IR dust emissivities, absorption cross sections, and UV and IR feature identifications. Among the key programs that I mentioned in the talk are:

• Dust emissivities in the far-IR and sub-mm bands, from $10 \mu$m to 1000 $\mu$m. These emissivities determine both quantitative measures of grain mass, but also control the thermal history and far-IR spectra from heated dust.
• UV and IR diagnostics of small clusters and PAHs are strongly dependent on the size and charge of the particles in the range N = 5-50 atoms, as well as larger clusters.
• Accurate oscillator strengths for the predicted UV adn optical diffuse interstellar bands that accompany PAHs.
• Photoabsorption cross sections and effective f-values for specific clusters.
• Gas-phase studies of PAHs, leading to more accurate quantitative models for the carbon budgets of interstellar clouds and nebulae.

Summary: Long-Term Data Needs

In the earlier sections, I have emphasized key laboratory measurements and theoretical calculations in atomic/molecular astrophysics. These parameters were tied to current and scheduled NASA missions in UV, IR, and X-ray space astronomy. One of the most important overall needs that pervades all these areas is for accurate (15%) parameters to interpret the ionization equilibrium and dis-equilibrium state of hot or photoionized plasmas. Probably the greatest uncertainty here is for radiative and dielectronic recombination rate coefficients. Careful attention should be given to self-consistency among inverse reactions, such as photoionization cross sections and recombination rates from a Maxwellian velocity distribution.

On a longer timescale, NASA and ESA will undoubtedly push the frontier of spectroscopy with higher throughput (higher S/N), higher spectral resolution, and new wavelength bands (especially the far-UV, the far-IR, and the sub-millimeter). In the X-ray, the Constellation-X new-mission study offers the chance for high-throughput X-ray spectrscopy, with attendant requirements for more accurate wavelengths, oscillator strengths, and collision strengths for higher-n transitions of elements other than the usual handful (O, Si, S, Fe). A future far-UV spectrograph on a 4-6 meter space telescope following Hubble will provide the opportunity for higher spectral resolution ($R \geq 10^5$) and unprecedented signal-to-noise for the detection of weak features of atoms, ions, molecules, and grains. It will also require wavelengths accurate to better than 10^-7, oscillator strengths and collision strengths for higher ion stages than the first and second ions, and new data on isotopes of the abundant elements (e.g., C, S, Fe).

Here too, these new data will accentuate the need for the basic ionization equilibrium parameters noted above. If there were one major outcome of this meeting, I would ask NASA to support a long-term program to define the ionization and recombination rates for all ionization stages of $\sim20$ astrophysically abundant elements. This program should be a combination of theoretical computations, like the Opacity Project and its successor, and careful benchmark experiments using ion traps and atomic beams.

References:

Arnaud, M., & Rothenflug, R. 1985, Astr. Ap. Suppl., 60, 425.

Brickhouse, N. S. 1996, EUV Spectroscopy of Stellar Coronae, in Atomic Processes in Plasmas, AIP Press, eds. A. L. Osterheld & W. H. Goldstein, Vol. 381, 31-38.

Osterbrock, D. E. 1989, Astrophysics of Gaseous Nebulae and Active Galactic Nuclei, University Science Books.

Shull, J. M. 1993, Physica Scripta, Vol. T47, 165-170.

Shull, J. M. 1996, Atomic Physics for Modeling of Astrophysical Plasmas, in Atomic Processes in Plasmas, AIP Press, eds. A. L. Osterheld & W. H. Goldstein, Vol. 381, 47-55.

Shull, J. M., & Van Steenberg, M. 1982, ApJS, 48, 95.

Sutherland, R. S., & Dopita, M. A. 1993, ApJS, 88, 253.

Acknowledgments:

My astrophysical research at the University of Colorado is supported by grants from the NASA Astrophysical Theory Program (NAG5-4063), NSF extragalactic astrophysics (AST96-17173), and the Space Telescope Science Institute (GO-06593.01-95A).