An interview with Mike Werner

Interview by Bo Reipurth, SFN #356 - August 2022

Your PhD from 1968 dealt with observations of interstellar molecular hydrogen. What led to this topic?

My thesis advisor was Martin Harwit, whom I met when I after graduating from Haverford College took a “gap year” and worked at the US Naval Research Laboratory in DC. Martin was there learning the craft of rocketry so that he could start a highly successful infrared rocket program at Cornell. Prior to my coming to Cornell, Tommy Gold, Ed Salpeter, and Robert Gould had written several papers about the amount of, at the time, unobservable molecular hydrogen in space and how it affected the distribution and the motions of stars perpendicular to the plane. So it was natural that Martin and I should embark upon an [unsuccessful] search for H2, focussing on the 3-0 vibrational band around 8500 Å.

Around the same time you did a theoretical study with Ed Salpeter on grain temperatures in interstellar clouds.

Yes, I had a post-doctoral fellowship at Fred Hoyle's Institute of Astronomy in Cambridge, UK. In those days, it was all arranged by a letter from Martin Harwit to Fred Hoyle, as Martin had previously worked with Fred. Ed Salpeter, whom I barely knew from Cornell, was there. I wrote a draft of this paper on grain temperatures which he kindly read....and saved me from a grievous error by pointing out that the grains at the exterior of the cloud would radiate inward and keep the grains in the interior from getting arbitrarily cold. Nevertheless, we found that for some reasonable assumptions the grains were cold enough to catalyze the formation of molecular hydrogen on their surfaces. This was followed up by a second paper headed by Dave Hollenbach, Ed's student, in which we estimated the H2 content of the sun's vicinity by looking at the Lynds catalog of dark clouds. Apart from many later papers based on work with Spitzer, this remains my most highly cited research paper.

Another early paper was a study, with Joe Silk and Martin Rees, on heating of HI regions by soft X-rays.

In those early days [late 1960's] Field, Goldsmith, and Habing in a classic paper had suggested that low energy cosmic rays heated HI regions. Joe Silk, also at Hoyle's institute, asked me to work with him on the idea that soft X-Rays could be responsible for this heating, and in a second paper we roped in Martin Rees to help us investigate the influence of discrete X-Ray sources, as opposed to the X-Ray background. My impression is that X-Rays never found much traction as a possible heating source, and also that nowadays people think a lot about photoelectric heating as a major contributor.

Among your first papers is a study of the ionization equilibrium of carbon in interstellar clouds.

By this time I had moved to Berkeley and was working with Charlie Townes. Those were the early days of molecular line studies of the ISM, and, as I recall, the excitation of the molecules depended on the electron density in HI regions, where the electrons are contributed mainly by carbon atoms. I found that a neutral cloud had an ionized skin but that the carbon became neutral within the cloud due both to extinction of ionizing radiation by dust and self-shielding of the ionizing photons. The resulting structure was, of course, something like an inside out Str\"omgren sphere. For the Habing value of the diffuse interstellar radiation field of 5000 photons/cm2/sr/s/Å at 1000 Å I found that carbon was largely neutral for densities >10,000 cm-3 and ionized throughout the cloud at densities <100 cm-3. Since it was thought at the time that the molecular clouds were high density regions, I concluded that molecules like NH3 and H2CO [among the few known back in the early 1970's] were probably excited by collisions with neutrals. This paper was also an early precursor to the idea of a photodissociation region.

In 1976 you and your team obtained what at the time were high-resolution maps of the Orion Nebula at 20, 50, and 100 μm. This is a highly cited paper which has had an important impact on our understanding of the region.

Again, by this time I had moved to Caltech and we established a collaboration with Al Harper to carry out photometry from the Kuiper Airborne Observatory. The KAO, which featured a 91 cm telescope floating in a reentrant cavity in the side of a converted C141 transport, was just starting to fly regularly. It carried the telescope above most of the water vapor in the Earth's atmosphere and provided access to the far infrared wavelengths, say 30-to-300 μm very roundly. This was basically unexplored territory, which made the early days of the KAO very exciting. Everything which we did was totally new.

Getting back to Orion, Eric Becklin's remarkable physical intuition showed us that comparing maps at 50 and 100 μm could establish the direction of temperature gradients and heat flow, and establish whether the infrared sources in the BNKL region were dust embedded or merely background objects. However, due to issues with pointing the KAO telescope, it was necessary to use a common focal plane aperture about one arcminute in diameter and separate the 50 from the 100 μm radiation with a dichroic beam splitter in order to get reliable flux ratios. When we combined the 50/100 μm data with a ground based 24 μm image we were able to show that the BNKL cluster was embedded in the molecular cloud, because it was the warmest region and the temperature decreased as one moved away from it. At the time, others were arguing that BN might be a highly reddened background object. So this was a major step in establishing the nature of embedded YSOs seen in the infrared.

In the same data set, we serendipitously detected infrared radiation from the bar to the SE of the Trapezium which marks the transition between the ionized interior of the HII region and a more neutral exterior region. We later carried out a multiwavelength study of this bar which was headed up by Craig Sarazin who came to spend a summer in Pasadena. This is of course yet another photodissociation region and it shows a peculiar broadband SED which, in retrospect, is due to PAH emission unresolved in the relatively broad filters we used for the study.

In 1983 you, with Harriet Dinerstein and Richard Capps, performed a polarimetric study of the infrared BN/KL cluster in Orion. What did you learn?

At this point, I was somewhere between Caltech and NASA-Ames in my professional hegira. I honestly can't remember what motivated this investigation, which we carried out at the NASA Infrared Telescope Facility in Hawaii, but the results were fascinating. We found extraordinarily high, more than 25%, polarization at 3.8 μm for some of the individual clumps within the BN/KL cluster. Interpreting this as scattered light, the direction of polarization incriminated a particular point source known as IRS7 as the illuminating source. This established that some of these clumps may have been density enhancements rather than self luminous objects, at least at the shorter infrared wavelengths. It is important to realize that the reason the polarization is so high is that this scattered light is the main source of radiation at 3.8 μm as the clumps are too cold for thermal emission at this wavelength.

Subsequently you have led far-infrared studies of several other high-mass star forming regions, including W3 and NGC 7538, and several regions in the Large Magellanic Cloud.

In those early days we observed regions like W3 and NGC 7538 because they were bright in the 10 μm region and proved to be bright in the far infrared, and thus easily studied with the limited sensitivity of the KAO. These other regions, being more distant but comparably bright, are of course much more luminous in the far infrared than the Orion Nebula, suggesting formation of several or many O-stars, much more dramatic than the comparatively wimpy Orion region. Although some of these studies were quite interesting and identified previously unknown embedded clusters, I think that more understanding of star formation has come from more comprehensive studies such as the Spitzer c2d program, as well as studies of individual regions like Cygnus X and W5 by members of the IRAC instrument team.

You were the Project Scientist for the Spitzer Space Telescope, and the paper in 2004 on the mission, of which you were lead author, has accumulated more than 2400 citations. What were some of the highlights from your involvement with Spitzer?

My involvement with Spitzer started in 1977 and extends to the present. Numerous highlights and lowlights illuminate this 45-year journey. I started work on what was then SIRTF [Shuttle Infrared Telescope Facility] in 1977 while I was on the faculty at Caltech, eventually ending up as [and still serving as] the Project Scientist for the elegant free-flying observatory which operated in a novel heliocentric orbit from mid-2003 through early 2020, with a 16+ year lifetime which far exceeded even the most outrageous pre-launch predictions.

From 1984, when the science team and instruments were selected, to 1993 or so we were working in a continuously evolving programmatic framework. Spitzer was liberated from the Shuttle in 1983 thanks to the scientific and technical success of IRAS and also because the Shuttle environment was seen to be very inhospitable to a sensitive, cryogenic infrared telescope. This rolled us into low Earth orbit where we had to deal with a variety of possible implementation options, none very attractive. Our next major redesign came with the realization that high Earth orbit was preferable to low Earth orbit because of the much lower heat load from the Earth. This version of Spitzer was blessed by the Bahcall committee in 1990 as the highest priority space astronomy initiative for the 1990's even as the project was being moved from Ames to JPL. The final and decisive architectural changes occurred in 1992-3 with the suggestion from JPL engineer Johnny Kwok that we should use the driftaway or heliocentric orbit. This approach meshed very well with the suggestion from Facility Scientist Frank Low that we should launch Spitzer with the telescope warm and have it cool radiatively on orbit until cold helium gas liberated by the small amount of power generated by the instruments would bring it to its final operating temperature of around 5~K.

As the Project Scientist during this turbulent period I managed to keep the science team focused on our vision of the extraordinary scientific power of a cold telescope in space instrumented with what were at the time large arrays of infrared detectors. This vision is what kept us together and moving ahead even in the face of programmatic diversity and pressure. It helped a good deal that we had funds available for the instrument teams to pursue detector and array development and other resources to study the optical and cryogenic design of the observatory.

Once we had the overall design settled in the early 1990's, the detailed design, integration, and tests went smoothly enough, although not without some heart-stopping incidents, including a near disaster in which the helium tank which cooled the instruments in the cryostat developed an ice plug and came perilously close to exploding. My role during this period expanded to include assuring that the engineering teams were aware of the importance of Spitzer science and the need for the requirements we were working towards; monitoring their work to assure that Spitzer was designed and built to do the science we had planned for; maintaining interest and enthusiasm for Spitzer and Spitzer science in the user community; making sure that NASA HQ knew of and supported what we were up to; and keeping the relevant congressional offices and staffers aware of our progress and prepared to see that we had the necessary budgetary allocations. Of course, I did not do all of this by myself but instead worked with other members of the science and management team to carry out these important tasks.

Following the launch of Spitzer in August 2003 the combined radiative-cryogenic cooling system worked perfectly and the telescope cooled according to plan. The Early Release Observations press conference and the formal renaming of SIRTF as Spitzer, in honor of space astronomy pioneer Lyman Spitzer, took place in mid-December 2003, and our cryogenic mission, which lasted until mid-2009 and provided the user community with three instruments and a range of imaging and spectroscopic capabilities covering wavelengths from 3 to 180 μm, was under way. During this period, I was not too involved in the day to day operation of Spitzer but stepped in when something was amiss, for example during our relatively rare safe-hold events. I was more occupied with my own scientific work with the Spitzer data, which included studies of X-Ray active AGN, debris disks, and PAH emission in reflection nebulae. A personal scientific highlight was the discovery, in conjunction with Kris Sellgren, of C60 in the interstellar medium of our Galaxy.

We always knew that the combination of the orbit and the cryo-thermal design meant that Spitzer should stay pretty cold following cryogen depletion. Our post-launch models based on the actual performance of the system suggested that this would be the case, and in fact when the last drop of liquid helium boiled away the telescope equilibrated at about 26.5 K, where it stayed from 2006 through the end of the Spitzer mission in January 2020. At this temperature, both the detector dark current and the telescope background were low enough to enable observations with our two shortest wavelength arrays, at 3.6 and 4.5 μm, with sensitivity and image quality essentially identical to what was achieved during the cryogenic mission.

This “warm mission”, if 26~K can be considered warm, was remarkably productive scientifically. Nature and the ingenuity of the science community cooperated to identify major science thrusts well-adapted to the constraints of our two-wavelength machine. Our ability to execute ground-breaking science while continually dialing back our operating budget saw us through five separate senior reviews - reviews at which missions which have entered the “extended mission phase” at the end of their prime mission are scrutinized for scientific merit and affordability.

Our biggest hit from this era was the identification of 7 Earth-sized planets snuggled together around the faint red star Trappist-1. This announcement got us above the fold in the New York Times and even featured on a Google Doodle! My own scientific work underwent a bit of a renaissance during the warm mission. While sitting in a rather boring meeting at the Space Telescope Science Institute I had the idea that Spitzer could usefully survey the entire 10x10 degree within which the Kepler spacecraft had measured hundreds of thousands of dwarf stars to uncover those with planets which transited as seen from Earth. Our proposal for over 400 hrs of Spitzer time to carry out this program was successful and we have deposited in the exoplanet archive precise [better than a few percent] 3.6 and 4.5~$\mu$m photometry of all the stars monitored by Kepler during its prime mission, both with and without observable transits. This in turn led to a second successful proposal, again awarded hundreds of hours of observing time, for follow up Spitzer studies of transits of stars with transiting planets discovered in the K2 mission which cleverly repurposed the Kepler hardware. This and follow on programs extending the methodology to TESS discoveries have spawned over 20 refereed publications, providing unique information about the characteristics of the exoplanetary systems.

Although I myself did not do a lot of work with Spitzer on star formation, others of course did. Among the many contributors to Spitzer studies of star formation I want to give a shoutout to Neal Evans and his Cores to Disks project, which was one of the six original Spitzer Legacy projects and led to numerous publications, among them one by Evans et al. in 2009, which was instrumental in establishing the duration of the various evolutionary stages, Class I, II, III, etc.