CERN Accelerating science

This website is no longer maintained. Its content may be obsolete. Please visit for current CERN information.

CERN Accelerating science

Interview with Kip Thorne

by Panos Charitos

Kip Thorne is one of the leading physicists working on Einstein's theory of relativity today. He has pioneered the scientific investigation of black holes in the universe. He was one of the founders of the LIGO project to detect gravitational waves and he has been one of the international team of physicists developing the LISA gravitational wave detector, a project of the space agency ESA which is likely to have some NASA participation. He has carried out important research in an unusually wide range of fields: general relativity, astrophysics, the quantum theory of measurement, time travel, even the experimental details of the design of gravitational wave detectors. Panos Charitos (PC) and Spyros Argyropoulos (SA) met him in Geneva and discussed with him about the new window that gravitational waves open and the cosmological implications of this discovery. 

(Image Credits: Chris Klimek)

P.C. We would like to kick off this discussion with a question about your motivation. How did you become involved with the search for gravitational waves?

I arrived at Caltech as a young professor around 1966 and created a research group working on gravitational waves, black holes and neutron stars. Initially, I was primarily interested in theory. However, Joseph Weber announced observational evidence for gravitational waves in 1969 and I followed his work with great interest.  With my student Bill Press, I wrote a paper in 1972 that envisioned the future of gravitational-wave research, both theory and experiment. When I first heard Rainer Weiss’s idea for gravitational-wave interferometers—the instruments we use in LIGO— I was very sceptical. The idea of using light to measure motions of a mirror that are 10-11 of the light’s wavelength , seemed crazy, but after long discussions with Weiss and with Vladimir Braginsky in Moscow, I became convinced it might be possible. My research on sources of gravitational waves indicated that if this idea succeeded, it would be a major tool for astronomy and cosmology. I therefore decided to devote my career to experiments on gravitational wave detection. My involvement in experiments began around 1976 and I had aleady been working on theory since about 1966, so today I count fifty years in the field, forty years of which I dedicated to experiments.

 P.C Did you expect that gravitational waves would be discovered during your lifetime?

Yes, and I thought it quite likely it would come from two colliding black holes of just the sort that we did see. I wrote a popular book called Black Holes and Time Warps: Einstein’s Outrageous Legacy, published in 1994, and I wrote a prologue to this book during my honeymoon in Chile in 1984. In that prologue, I described the observation of two black holes, both weighing 25 solar masses, spiralling together and merging and producing three solar masses of energy and gravitational waves, and that’s very close to what we’ve seen. So I was already in the 1980s targeting black holes as the most likely kind of source; for me this was not a surprise, it was a great satisfaction that everything came out the way I thought it probably would.

P.C. What are gravitational waves and how are they different from other types of waves with which we are more familiar?

Let’s recall that electromagnetic waves are oscillations of the electromagnetic field that propagate through space-time. Gravitational waves, by contrast, are oscillations of the shape of space-time, very much like waves on the surface of a lake, except instead of going up and down, they stretch and squeeze space transverse to the direction of the waves’ propagation. In other words, inertial reference frames that are separated from each other move back and forth relative to each other. Our detectors consist of mirrors hanging from overhead supports that try to remain at rest at their local inertial frames; the inertial frame of one mirror and the inertial frame of another mirror 4 km away move back and forth relative to each other due to the waves’ stretching and squeezing of space, so the mirrors also move back and forth.

P.C. What was the origin of the interferometer concept used in the LIGO and VIRGO experiments?

The first people who had this interferometer idea were two Russian theorists: Mikhail Gertsenshtein and Vladislav Pustovoit, in 1962. They did not have enough experimental background to understand the noise sources that such an experiment would encounter nor how to deal with them nor what accuracy could be achieved; but they did conceive the basic idea.  A few years later, several other people had the same idea, independently. I regard Rainer Weiss as the primary inventor of this technique because in 1972 he identified all the major noise sources that the initial LIGO and VIRGO interferometers would face. He invented ways to deal with each noise source and computed what level of accuracy one could achieve. By comparing these results with the expected strengths of the waves that my colleagues and I had predicted, Weiss concluded that a detector with arm lengths of a few km had a good chance of success.

Weiss’s analysis became a sort of “blueprint” for the initial LIGO and VIRGO detectors. Of course some changes were made to the final design due to inventions that offered additional benefits by Ronald Drever in Glasgow, Scotland, Roland Schilling in Garching Germany, and others. As a result, the final detectors were a little different from Weiss’s original invention, but they were substantially similar.

Aerial view of LIGO's Hanford, Washington, facility

P.C. Why did it take so long to detect gravitational waves?

The main problem was that the interaction of gravitational waves with matter is so weak that the motions to be detected were no larger than than 1:100 the diameter of a proton over a distance of 4 km. By 1978, we theorists had gained sufficient understanding of the sources of gravitational waves to identify— at a conference in Seattle, WA—the strength of the first waves that would be seen as 10-21 in strain, i.e. 1:100 the diameter of a proton. Indeed, that was the strength of the first waves observed, so our 1978 predictions were right on. This value is so small that it required very complex instruments, and a very long time to perfect these instruments and make them robust enough, for our measurements to succeed.

P.C. Were there other ideas or methods discussed at the time for detecting gravitational waves?

There were a number of others. The first detectors were developed by Joseph Weber; they were aluminum cylinders, whose mechanical oscillations were affected by passing gravitational waves, and they operated in roughly the same frequency band as LIGO and VIRGO do today: kilohertz frequencies.

In the 1970s, the idea of drag-free spacecraft tracking each other with laser beams—which became LISA (the Laser Interferometer Space Antenna) was proposed by Peter Bender, Rai Weiss and others.  LISA operates at frequencies 10,000 times lower than LIGO, and it became a major topic of discussion in the 1978 conference. Bender more than anyone else pushed the idea and developed its technological concepts through the 1980s and into the 1990s, when ESA and NASA finally embraced it. Thus, LIGO and LISA both developed from work done in the 1970s, 

There was also much interest, in the 1970s, in detectors for gravitational waves with frequencies much higher than LIGO’s kilohertz band. In the end it became pretty clear to us that there would not be strong enough sources for detectors at those high frequencies, so little experimental effort has been put into them except for a modest effort by Bob Baker and his colleagues

In addition, there was the idea of looking for gravitational waves by looking at a set of pulsars in the sky. As the gravitational wave sweeps over the earth, (crudely speaking) it changes the ticking rates of clocks on earth and as a result, when you look at pulses from an array of pulsars they will all speed up and slow down together. If this is seen, it will be clear that something is happening on earth; the passing of a gravitational wave. This “pulsar timing array” (PTA) technique, which is now close to success, also dates back to the late 1970s.

The so-called “B-mode” polarization of the cosmic microwave background is a very promising technique for looking for primordial gravitational waves with extremely low frequencies (periods of mllions to billions of years).  This technique was not conceived until the 1990s, while all the others have been explored since the 1970s.

To summarize: We do expect that within the next twenty years we will make gravitational wave observations in four frequency bands: with LIGO/VIRGO; with LISA; with PTAs’ and with CMB polarization. This is like creating optical astronomy, x-ray astronomy, radio astronomy and infra-red astronomy all within twenty years.

S.A. Are the ideas of Joseph Weber still pursued?

They have still been pursued in recent years in Brazil and at Leiden, but not at a very intense level. These mechanical detectors have some possibility of contributing at moderately high gravitational wave frequencies, e.g. 10,000 or 3,000 Hz. However, this technology is not competitive with interferometers in the 10 to 1,000 Hz band where their richest science is expected.

S.A. In the late 1960s, when Joseph Weber presented the first evidence for gravitational waves, was the sensitivity of the instruments not really sufficient for detection?

No, it was not.  Weber’s sensitivity  was far lower than the current LIGO and VIRGO sensitivities. Detectors similar to Weber’s were built by a number of people around the world. Vladimir Braginsky in Moscow was the first and many others followed, in an effort culminating in the beautiful bar detectors built by Amaldi and others in Italy and at CERN. That technology continued improving in parallel with interferometers until fairly recently, but nobody could replicate what Weber seemed to observe.

S.A. How did LIGO come into being?

It was clear by 1980 that for success, the gravitational wave interfereomters pioneered by Weiss would have to be not only very large (several kilometers in length) but also very complex, and so would require a big team of experimenters. To this end, in 1984 Weiss, Ronald Drever and I created the LIGO collaboration between Caltech and MIT.  In 1987, we brought on Robbie Vogt as our first LIGO director. He made the collaboration between the Caltech and MIT research groups much more effective, and he led us in writing our proposal for construction of LIGO. Once the proposal was written, extensively reviewed and accepted by the National Science Foundation (NSF), he led the effort to obtain funding in the US Congress. The project received its first construction funds in 1992.  Our plan from the outset was to build two successive generations of detectors (interferometers): initial interferometers that would be a crucial stepping stone along the way toward advanced interferometers, which would likely make the first detection of waves.

In 1994, as we were getting close to initiating construction, we brought in a new director, Barry Barish. Barish transformed LIGO from a modest-sized Caltech-MIT collaboration to the very large collaboration that was needed for success. He created two entities. One was the LIGO laboratory, which handled all the R&D for the detectors, and the design and construction of the facilities and the detectors. The other was the LIGO scientific collaboration that now includes 1000 scientists and engineers at almost 100 institutions in 16 nations and deals with the science of LIGO, the characterisation of noise in the interferometers, the searches for gravitational waves, the data analysis, etc. This organizational structure has been very successful. In the mid 2000s, after leading the construction of our initial detectors and their early searches for waves,  Barish was lured away from LIGO by high-energy physicists to lead the design of the International Linear Collider, and was replaced by two more highly skilled LIGO directors, first Jay Marx and now David Reitze.  Our four successive directors, especially Barish, have been crucial to LIGO’s success.

In addition, the collaboration with VIRGO has been crucial and is key for the future. The VIRGO team collaborates with the LIGO team on all aspects of data analysis and the search and discovery of gravitational waves in the LIGO data. This will continue in the future when the advanced VIRGO detector, and then the Japanese detector KAGRA and LIGO India join our searches.

Last but not least, I would like to emphasize that the discovery of gravitational waves by our second-generation, advanced LIGO interferometers, belongs to more than 1,000 scientists: the members of the VIRGO and LIGO collaborations. Somehow I end up getting far more credit that I deserve for this. I was involved from the beginning but not in the 2015 discovery itself; that belongs to the younger generation.

S.A. Could you explain what kind of sources produce gravitational waves?

The sources we have observed thus far are pairs of black holes that orbit around each other, spiral together, collide and merge. With advanced LIGO and VIRGO, as the instruments approach their design sensitivity in the next three years, we expect also to see gravitational waves from neutron star binaries: neutron star pairs  that spiral together, collide and merge. These are thought to be the sources of at least some gamma ray bursts; so basically we will be looking at the “engines” that drive gamma ray bursts.  Coincident electromagnetic and neutrino observations will be crucial to extracting maximum information from these events.

In addition, we are looking for gravitational waves from black holes that have a neutron star orbiting around them and tear the neutron star apart. We are searching for gravitational waves from the cores of supernova explosions, and also from cosmic strings created in the early universe. LIGO and VIRGO will searching for all these sources during the next few years.

P.C.  What can we expect to see in the future?

Two collisons of black holes have been definitely seen thus far. And there is a third one that was probably a black hole collision and not just noise, but we are not certain because its signal is so weak. These observations took place during four months of searches last winter with both detectors working well simultaneously for about two months. We therefore had about two months of good data, which means roughly one black hole merger per month. When our advanced LIGO interferometers reach their design sensitivity, in 2019 or 2020, they will be seeing three times farther into the universe, covering a volume 3x3x3 or about 30 times larger than at present, which means that they might see very roughly one event per day.

There are tentative plans for a much more sensitive interferometer called the Einstein telescope in Europe, and plans for a third and fourth generation of LIGO detectors, which could further increase the distance seen a factor of three and then four more. This is a huge discovery range. We expect the field to progress amazingly fast, just as radio astronomy did in the 1950s to the 1980s and x-ray astronomy in the 1960s to the 1990s.

Gravitational waves will be a major tool for astronomy in the coming decades and centuries. LISA, at far lower frequencies than LIGO/VIRGO, will see collisions of supermassive black holes with signal to nose ratios of thousands, and also small black holes spiraling into supermassive black holes. Over times of a few years, as each small hole orbits its huge companion, it will create very complex gravitational waves. These waves carry a detailed map of the geometry, the shape of space-time, around the supermassive black hole. With LISA, we thereby will get detailed maps of black holes, similar to maps of the surface of the earth, the moon and Mars.

LISA is the ideal instrument to observe gravitational waves produced when the universe was about 10-12 seconds old, and its cooling triggered the electroweak phase transition, giving birth to the electromagnetic force and the weak nuclear force. It is possible that this phase transition was first order, which means that the electromagnetic force was born in bubbles, like water droplets that form from water vapor. According to the theory, when bubbles with the electromagnetic force on the interior but not the exterior are formed, they expand with the speed of light, collide and produce gravitational waves. These waves shift to longer wavelengths as the universe expands and today they should be in the LISA wavelength band. It is a realistic dream that, thereby, LISA will watch the birth of the electromagnetic force. That’s truly amazing.

A pair of two merging black holes could emit gravitational waves similar to those observed by LIGO (Image credit: The SXS (Simulating eXtreme Spacetimes) Project) 

P.C. Do you think that research in this field will shed light to other open questions, for example about dark matter?

Dark matter gravitates and today we see clear signs of its gravitational effects in galaxies. There is a possibility, as Lisa Randall has pointed out, that some forms of dark matter experience strong self interactions, like the nuclear force, which could lead to compact objects made of dark matter, such as a dark matter “star”.

If there are condensed dark matter objects, they will produce gravitational waves just like ordinary stars and in addition they could lead to the formation of black holes that would also produce gravitational waves. However, an astronomer would not be able to tell the difference between a black hole made from dark matter and one made of ordinary matter.  It is not so clear at this stage how we can search for and identify gravitational waves from dark matter objects and differentiate them from those from ordinary matter.  However it is plausible that in 20 to 30 years people will have figured out ways to explore this opportunity.

Gravitational waves are an ideal tool for exploring the birth of the Universe. They are the only form of radiation that penetrates through matter so effectively that, if they were produced in the Big Bang, they would not be significantly scattered or absorbed as they travel through the extremely dense and hot matter of the very early universe.  However,  they would be amplified by the universe’s extremely rapid, inflationary expansion in its very early moments. Therefore, if we see these primordial gravitational waves as they are typically called, they will be carrying information about both whatever came out of the Big Bang, and the epoch of inflation.

Conventional theory says that what came out of the Big Bang was nothing but vacuum fluctuations. If that was the case, then inflation is predicted to have parametrically amplified the fluctuations to produce primordial gravitational waves strong enough to be seen in the polarisation of the microwave background today, and also to be seen by a possible successor of LISA called the Big Bang Observer. If something stronger that vacuum fluctuations came off the Big Bang, we would observe something stronger after inflationary amplification.

S.A. People might have heard about gravitational waves from the news about the BICEP experiment a couple of years ago. Could you explain the difference between the gravitational waves that LIGO and VIRGO detected and primordial gravitational waves?

The BICEP experiment attempted to detect and measure the polarization of the microwave background, about which I have just spoken.  The prediction that was made by several theorists in the 1990s was that primordial gravitational waves with very long wavelength would have interacted with the primordial plasma during the epoch of recombination, i.e. when electrons were combining with protons to form neutral hydrogen.

That was the era when electromagnetic waves were last scattering. The gravitational waves are predicted to have influenced the last scattering, placing onto the resulting microwave background a so-called “B-mode” pattern of polarization, which is different from the polarisation from most astrophysical processes. The challenge was to find this B-mode polarisation, but it was known that there was another important process that could also produce such a pattern: the emission of microwaves by dust. It was necessary to find a region on the sky were there was very little dust to produce such a B-mode pattern. The BICEP team thought they had found such a region, i.e. a hole in the dust, and concentrated on that. They did find the B-mode pattern and announced its discovery; basically they said “we do not think there is enough dust there, so we may have discovered the imprint of the primordial gravitational waves”. The Planck team and others argued that in fact there was more dust there than the BICEP team had thought, so the Planck team and the BICEP team did some mapping of the dust together and found that there might indeed be enough dust in that direction to produce the observed B-mode pattern. It is not clear whether all of the observed B-mode polarization is due to dust or some of it is due to gravitational waves. This is the challenge that people working in this field are facing; they are trying to separate the gravitational-wave-induced polarization from the dust-produced polarization. This is tremendously important. I am told that this effort could be successful within the next decade and will lead to the first confirmed discovery of primordial gravitational waves.

P.C. High energy physics has changed after the Higgs discovery at the LHC; CERN and other partners are now deploying plans for a future collider. From your experience in building LIGO and international collaboration, do you think that it is timely to start designing large scale infrastructure to push the energy frontiers further?

I am not an expert in high energy physics but it seems obvious that a next generation of particle colliders and detectors, beyond the LHC, is very much needed and justified. We clearly need large scale infrastructure in some areas of science; gravitational waves and high energy physics are prime examples.

P.C. Regarding future large-scale research infrastructures, such as those explored under the FCC study, what are the lessons to be learnt from LIGO?

Maybe the best thing to learn is the importance of superb management, in order for a large project like LIGO to succeed. We indeed have had superb management, especially with Barry Barish.  Barry created the modern LIGO and he is an absolutely fantastic project director. Having him lead us through our transition into the modern LIGO was essential to our success—as were a very good experiment idea and a superb team, of course.

P.C. At that time, was it easy to push for LIGO at the funding agencies and other stakeholders?

It was a remarkable tribute to the National Science Foundation and the US Congress, as well as to the INFN and CNRS and the funding processes in Italy and France, that we and VIRGO were successful in getting funded. No one had observed a gravitational wave directly, large sensitivity improvements were needed and large infrastructure was required. We very frankly said that we would probably have to build two generations of instruments —initial and then advanced LIGO, and initial and then advanced VIRGO— before we could have a high probability of seeing waves; and this prediction turned out to be correct. Because of this necessity for two generations, it was initially hard to get funding. We submitted the LIGO proposal in late November 1989; the National Science Foundation approved it in 1990; but then and it took two years, until 1992, for Congress to provide the funding. Fortunately, LIGO was much smaller in cost than the LHC or the SSC, small enough that all we needed was the support of the relevant Congressional Committees and not the entire Congress—and that we did receive. We also had a succession of four LIGO directors, Rochus Vogt, Barry Barish, Jay Marx and now David Reitze, who were very effective in explaining to the Congressional Committees what was required. It is essential to have leadership that is able to communicate with the people who make the funding decisions.

P.C. Given your involvement in outreach and in the movie Interstellar, what is the personal reward that you get from the science communication activities and your research?

One personal reward is the joy of working closely with brilliant and creative people who are not scientists. It was wonderful to collaborate on Interstellar with Christopher and Jonathan Nolan, with Paul Franklin’s visual effects team, and earlier with Steven Spielberg and Lynda Obst, my partner. Brainstorming about science with all these people was very enjoyable.

Another reward has been our success, through Interstellar, in increasing enthusiasm for science among the general public, especially young people. With this movie, I reached 100 million people with my message of the beauty and power of science. There is no other way, besides a block-buster movie like Interstellar, for a professor like me to reach so large an audience.  And for those non-scientists who became inspired by Interstellar’s science, I wrote a book, The Science of Interstellar, to teach them details of the science that they had seen so compellingly on the screen.  Young peoples’ enthusiastic response to my book has been wonderful.

I have several different collaborations: one on a second film; collaborations in a multimedia concert about sources of gravitational waves with Hans Zimmer and Paul Franckman, who did the music and visual effects for Interstellar; and collaborations with Chapman University art professor Lia Halloran on a book with her paintings and my poetry about the warped side of the universe. I am having great fun entering collaborations between scientists and artists and I think, at this point of my life, if I have a total failure with trying to write poetry, well that’s alright: I’ve had enough success elsewhere.