Astronomer Garik Israelian (watch his TEDTalk) explained the secrets of spectrography yesterday. And last night, he answered a few of our followup questions by email — including some questions from TED’s Facebook fans:

Tell me about your latest work with lithium and its significance for planet-bearing stars.

Since year 2000, our group has been working very actively to study peculiarities in the chemical composition of planet host stars, with dozens of papers in this field. We have proposed a “Li-6 test” to study planet engulfment and/or planetary matter accretion processes in sun-like stars. Up to now, we had not found a single chemical element with a different behavior in planet-bearing stars, as compared with similar stars without planets. In the new paper, which will appear soon, we present the first such case. We present results on Lithium content for a sample of about 500 stars, including 100 planet host stars, and we find that for planet host solar analogs (stars very similar to our Sun), and only for this kind of star, there is a peculiar behavior in the Li abundance. These stars have on average less than 1% of their initial Li content. Like our Sun, these stars have been very efficiently destroying Lithium. Why? We do not really know. Maybe this is related to their rotational history, strongly influenced by the presence of planetary systems. This is not the case of solar-analog stars in the comparison sample (“single” stars without known planets), where a large fraction preserves a much higher Li abundance — about 10 times more.

Our “single”-star sample has been monitored by the most precise spectrograph, called HARPS (located in Chile), for years and no planet has been detected. Let me mention that not all sun-like stars host planets — perhaps about 30% of them are planet-builders. It’s not so easy to form a planet!

Our results clearly indicate that solving the long-standing problem of the Li depletion in the Sun, which has been a puzzle for 60 years and has promoted the development of many transport and mixing theories, will require a proper understanding of the interaction of a planetary system with its planet host star.

We find that solar analogs with low Lithium content have a higher probability of hosting planets, and therefore searches of planetary systems can increase their efficiency by performing a spectroscopic observation of Lithium.

And finally, we suggest that the Sun is “Li-poor” because it hosts a planetary system.

How do you like the new Gran Telescopio Canarias so far? What are some of the questions you’re working on there?

It’s too early to speak about GTCs performance. I think we need a year or two to see the first results. The most interesting instrument for planet-hunters is the camera called CanariCam — a mid-infrared imager with spectrogroscopic, coronagraphic and polarimetric capabilities. CanariCam works in the thermal infrared between ~7.5 and 25 microns. Planet-hunters plan to conduct a search at 10 microns with CanariCam for substellar objects (brown dwarfs and massive giant planets) around many stars in the northern hemisphere. We keep our fingers crossed!

How soon will you be able to build the new 42-meter E-ELT you talked about, to detect Earth-like planets around sun-like stars?

E-ELT is planned to be operational in 2018, while the telescope site will be selected next year. One of the first light instruments may be the high-resolution spectrograph called CODEX. The high resolution and long-term stability of CODEX, coupled with the large collecting area of the E-ELT, provides an unequaled facility for measuring stellar radial velocities at the few cm/s level. This precision will allow detecting Earth-mass planets in the habitable zone around solar-type stars.

However, recent technological developments open a new window for us. We may discover Earth twins much before the E-ELT era, The so-called “laser frequency grid” (or Astro Comb) technique may allow to hunt other Earths before 2015. The super-precise HARPS-NEF (or HARPS-NORTH) spectrograph is under construction by collaboration between Harvard’s Origins of Life Initiative, New Earths Facility, and the HARPS team of the University of Geneva, and is expected to start operation soon after 2011. Our team in Canary Islands is collaborating with astronomers in Harvard and Geneva to install this instrument on the 4.2 meter William Herschel Telescope in La Palma. HARPS-NORTH will use the Doppler technique to discover and characterize Earth-like planets from candidates identified by NASA’s Kepler mission, launched on March 6 this year.

Reader Sameena asks three questions:
Is it always right to assume that other forms of life in another planet must have the same beginnings as us — tectonic plates with volcanic activity, oxygen and water?

No, I do not think that we have to always assume that. This is more a question for biologists. If they tell us, astronomers, that new life forms may exist under x conditions, breathing sulfur (for example) and somehow producing xenon (for example), then we will model the biospheres of those planets and will find out which spectral lines will/may indicate the presence of that form of life. So our task is to carry out very precise spectroscopic analysis of the atmospheres of extrasolar planets. We can infer their chemical composition, physical conditions etc. We may find then some “strange” chemicals indicating the presence of other forms of life. This will be fun!

How do you convince yourself/your peers/the world that one unusual peak on your spectroscopy is not through error?

This is more a technical question. Spectra have some noise produced by a detector and other sources. This noise can be modeled and calculated. We always know the level of the noise. Anything considerable above that level is a real signal.

How do sound waves (which need a medium) travel through space?

They don’t. But we can still study them in the atmospheres of stars. Imagine observing a volcano on the Earth from Mars. If you have a spectrograph, you can study the motion of the Earth’s upper atmosphere triggered by the explosion. Your spectrograph will give you the velocities of gas particles in the atmosphere. If you have a good knowledge of gas-dynamics, physics, etc., you can calculate the sound produced by this volcano at a given distance. This is not so difficult. But the point is this: You are far away observing a motion of gas particles (velocities), their emission (they radiate too) and therefore the state of the matter in the atmospheres. You use those observations to compute the amplitude and frequency of sound waves responsible for those motions. It’s pure classical physics.

Reader Mike Ho asks this question: I’ve read that there is spectroscopic evidence of cellulose elsewhere in the universe; is that accurate? How reliable is spectroscopy for detecting large molecules?

This is true. This was reported in Nature, in 1978. Tholins have been detected as well (I think by Carl Sagan).

There are many unidentified bands in the spectra of stars. Wide bands are produced by some complex molecules in the interstellar space. It’s really hard to identify them for two reasons:

1. observational
2. theoretical

Observational — because the spectra are full of absorption lines of stellar origin, and when you have a smooth, weak and extended absorption (such as a diffuse molecular band or DIB) in the spectrum covering some 30-50 angstroms, it’s almost impossible to see it unless you are able to carefully eliminate the spectral lines produced by a star and fit the stellar continuum (using models). Your DIBs are hidden under the stellar spectra! There are well-known diffuse molecular bands — about 200 have been identified. However, the most complex ones, produced by even more complex molecules, are still unidentified.

There is a theoretical difficulty too. One has to make quantum mechanical calculations (N-body) of very complex molecular structures and compute their spectra (thousands of lines). There are lots of approximations in this “game,” and you have to be very careful because it’s impossible to check them in the lab. (Your lab should be the vacuum between stars!)

My colleagues at the IAC in Canary Islands have detected naphthalene in the interstellar medium, a molecule that, in combination with water, ammonia and ultraviolet radiation, produces many of the amino acids fundamental to the development of life. They have also discovered fullerenes (C320 and C540). This is a terrific field of research, and I believe the interstellar medium hides a lot from your eyes. It’s full of mysteries and enigmas — free-floating planets, isolated stellar mass black holes, and all sorts of stuff.

You suggest that a planet harboring life might decide to change its chemical signature to send a message to other planets. Do you think Earth should be signaling the universe? If so, what should we do to let other planets know we exist?

Yes, technologically advanced civilizations can handle this. I don’t think we can do that (and we’d better not try — the Sun is still clean). I have not done this estimate by myself: how much of this or that chemical element we need to put in the atmosphere of the Sun so that an alien civilization will fix this as a “signal” of an intelligent (or not intelligent!) life in the solar system. Obviously, we need to use an element which is absent in the solar atmosphere (there are quite a few). But then, I can imagine that we will need few thousand tons. Let me do this calculation for an “interesting” element …