# Dimitri M. Mihalas (1939 - 2013)

**Dimitri Mihalas died on Thursday the 21st of November 2013.**

*Reprinted with permission from Physics Today.*

World-renowned astrophysicist Dimitri Mihalas passed away in his sleep at his home in Santa Fe, New Mexico on November 21, 2013. Dr. Mihalas retired from the University of Illinois at Urbana- Champaign in 1999 and from the Los Alamos National Laboratory in 2012. Dimitri, to his friends and family, has donated his body to the University of New Mexico Medical School and his library to New Mexico Tech.

Dimitri was born on March 20, 1939 in Los Angeles, California where he grew up. He received his B.A., with Highest Honors, in three majors, Physics, Mathematics, and Astronomy, from the University of California at Los Angeles at age 20. Four years later he received his Ph.D. in Astronomy and Physics from the California Institute of Technology. He was a pioneer in computational astrophysics, and has remained a world leader in the fields of radiation transport, radiation hydrodynamics, and astrophysical quantitative spectroscopy for decades. His broad knowledge and immense contributions earned him election to the U.S. National Academy of Sciences in 1981 at age 42.

Dimitri had an exceptional record of work. He published more than 150 papers and technical reports, authored or co-authored eight books, and co-edited three others. Three of his books have been used as textbooks for both undergraduate and graduate students worldwide and have been translated into other languages, including Russian. His book *Foundations of Radiation Hydrodynamics* has become the “bible” of the radiation hydrodynamics community.

Dimitri wrote his Ph.D. thesis on theoretical modelings for hydrogen and helium line strengths and profiles in O-stars. He interpreted the observational data with the best theoretical analysis possible at the time. The models were computed using the then prevailing simplifying approximation of local thermodynamic equilibrium (LTE). As shown by its high number of citations, this work, primitive by today's standards, had a major impact on the field.

After joining the faculty of the Department of Astrophysical Sciences at Princeton University, Dimitri undertook new analyses of stellar abundances, devised a method for constructing a constant-flux atmosphere with convection in the full transport (not diffusion) regime, computed the first line-blanketed spectrum of a hot B star which showed that continuum- only models overestimated the effective temperature by 10% (hence luminosity by 40%), and computed a sample of H line-blanketed models accounting for the distortion of the continuum by the confluence of the hydrogen Balmer lines near 3650Å.

Dimitri realized that the widely accepted approximation of LTE in stellar atmospheres is inadequate, so he undertook a major initiative to solve the analogous non-LTE problem. This is an exceptionally difficult problem owing to the need to solve self-consistently a set of coupled, highly nonlinear effects. Only the highly idealized case of a two-level atom with a single line and two continua had been solved previously.

After several preliminary explorations of methodology with Auer, they reformulated the non-LTE stellar atmosphere problem in a completely original manner; applying the Newton- Raphson technique to the full nonlinear system, and iterating to convergence. This method was a breakthrough in the field – it revolutionized all further work on computing stellar atmospheres. Through his idea of “variable Eddington factors,” the method was made even more efficient.

With this new methodology, Dimitri constructed an extensive, widely cited, sample of non-LTE models for hot stars, and used them to evaluate the effects of departures from LTE on observable stellar line and continuum indices. These results allowed Mihalas and Auer to perform several critical studies which were not previously possible, achieving for the first time good agreement between the computed and observed strengths of the hydrogen and helium lines, and surface gravities consistent with the stars' observed masses and element abundances obtained

from nebular analyses. This work has been widely cited, both for the quality of fit to observations, as well as a “gold standard” for verification of present-day calculations using much faster iterative methods. Work with Hummer led to the discovery of the physical mechanism producing the emission the N III emission lines at wavelengths 4634 – 4650Å in O((f)) stars.

In the 1970s, Dimitri devised the now standard method for solving the line-transfer problem in expanding spherical atmospheres in the comoving frame. This work was summarized in the heavily cited 2nd edition of his textbook *Stellar Atmospheres*, which remains the standard in the field even after 25 years. A third edition (with I. Hubeny) which describes modern fast, iterative methods to solve the transfer equation, including realistic line-blanketing, is in press.

In the period 1981 – 1998, as a consultant to LANL, Dimitri gave lectures on radiation hydrodynamics, for which, he received a number of “Certificate of Appreciation” by X-Division for “Outstanding Service to the Applied Theoretical Division”. In 1984 Dimitri completed the above-mentioned *Foundations of Radiation Hydrodynamics*. In 1999 Dover Publications reprinted it in inexpensive paperbound form at the urging of many scientists from LANL, LLNL, NRL, and the academic community.

In the period 1987 – 1994 Dimitri worked with Anderson, Hummer, and B. Mihalas on the development of a modern EOS code for stellar envelopes, in support of the large British – American Opacity Project (OP) led by Professor M. J. Seaton. It emerged that the OP results are in excellent agreement (better than 10%) with the independent LLNL OPAL computations. These new opacity data significantly impact stellar evolution calculations, and have helped resolve several previous discrepancies between calculation and observation, especially with the interpretation of Cepheid variable pulsations.

Jointly with Stone, Dimitri showed that the computational oscillations or diffusion, typically found in numerical simulations of a propagating radiation front in vacuum, could be eliminated by using upwind monotonic interpolation methods. They (and M. Norman) also collaborated on incorporating 2-D radiative transfer in the radiation magnetohydrodynamics code ZEUS 2-D developed at the University of Illinois and extensively used by astrophysicists.

In collaboration with K.-H. Winkler and M. Norman, Dimitri developed novel implicit adaptive grid methods to solve 1-D radiation hydrodynamics problems having multiple time and space scales ranging over many orders of magnitude. Dimitri derived the “adaptive-grid transport theorem” rigorously. In 1994 – 1997, Dimitri developed TITAN, a 1-D implicit adaptive-grid code. In collaboration with Cheng at Los Alamos, Dimitri used TITAN to compute the best-ever numerical solution of the infamous Noh stagnating shock problem in planar, cylindrical, and spherical geometry. In 1999, in collaboration with Gehmeyr, and Sincell, Dimitri studied the famous Zel'dovich and Heaslet/Baldwin supercritical shock problem and was able to compute the first numerical solutions for this problem. Also in the same year, Dimitri and Auer wrote “An X-6 Radiation Hydrodynamics Primer” intended for LANL use, and in 2001 they published an incisive discussion on exact relativistic laboratory-frame radiation hydrodynamics.

Dimitri was honored for his many scientific contributions at the *International Conference in Honor of Dimitri Mihalas for his Lifetime Scientific Contributions on the Occasion of his 70th Birthday held in Boulder, CO in late March 2009.*

In addition to his many scientific contributions, he also published several collections of his poetry and other writings. These other writings include "A Primer on Depression and Bipolar Disorder" and "Depression, Bipolar Disorder, and Spiritual Growth." These books have had as profound an impact as his professional textbooks and articles.

Throughout his long career, Dimitri gave generously of himself to all with whom he interacted. As an advisor, role model, confidant, and friend, he saw each person as an individual, acknowledging strengths, helping overcome weaknesses, giving encouragement, and enthusiastically praising their success. He touched the lives and careers of many students and colleagues and has left a lasting legacy to be cherished by those who knew him.

Baolian Cheng, LANL, Los Alamos, NM John Castor, LLNL, Livermore, CA

Jim Stone, Princeton University, Princeton, NJ

**BAAS:**BAAS, 2013, 45, 021