Theodor W. Hänsch (2015) - Science with Combs of Light

Good morning. We are now moving from biology to physics. One big event in physics that happened in 1960 was the demonstration of the first laser that Maiman came up with. The first ruby laser. That had a tremendous impact on many areas of science, including, of course, physics. The idea that one could make an optical laser was planted in 1958 in a seminal paper by Art Schawlow and Charlie Townes. The impact of lasers on science and technology can perhaps be judged by realizing that no less than 26 Nobel Prizes have now been awarded for work around the laser, including last year's chemistry prize for super resolution microscopy. Sadly, Charlie Townes passed away earlier this year. He has written a wonderful book, How the Laser Happened, which gives very valuable insights into how scientific discoveries are made. In 1970, I joined, as a postdoc, the laboratory of the brother-in-law of Charlie Townes, Art Schawlow, at Stanford University. Art gave good advice to young coworkers on how do you discover something new. He said, "Try to look where no one has looked before." In 1970, that was actually easy for me because I had come up with a simple way of making dye lasers not only widely tuneable but highly monochromatic. With these lasers, one could use new types of spectroscopy, Doppler-free spectroscopy. One of the first things we looked at with this laser was a spectral line of atomic hydrogen, the Balmer-alpha line that had played an important role in 1970s and atomic physics because people were wondering if Dirac was right in predicting the fine structure of that line. With our laser, for the first time we were able to resolve individual fine spectral components, in particular to resolve the Lamb shift directly in the optical spectrum, which had shown that Dirac is not right and that led to the development of quantum-electrodynamics. This was the beginning of a long quest for ever higher accuracy and resolution in spectroscopy of hydrogen. Before lasers, people could measure spectral lines to seven significant digits. We are now at 15 digits. In order to improve these kinds of measurements by a factor of 10, it's not enough to work harder. One really has to come up with new ideas. This has been really stimulating quite a few new ideas in Doppler-free spectroscopy. The idea of laser cooling was born in this quest and in the 1990s, the idea of the femtosecond comb. With the help of this comb, having measured the frequency of a very sharp two photon resonance in hydrogen now to 15 significant decimal digits. Have we discovered anything surprising? Perhaps. By comparing different transitions in the hydrogen atom and believing in the theory of quantum electrodynamics, one can figure out how large is a proton. What is the root mean square charge radius of the proton? A few years ago, there was another accurate measurement of the proton charge raduis based on manmade muonic hydrogen where the electron is replaced by a 200 times heavier muon that lives only for a microsecond or two. These muonic experiments gave a very good value for the proton raduis but about 4% smaller than that obtained with hydrogen. So we have to see if there is a mistake somewhere or if we really have discovered a dent in the armor of quantum electrodynamics. This year I expect to see the first results on precision laser spectroscopy of antihydrogen atoms, where the proton is replaced by an antiproton and the electron by a positron. Of course, most theorists believe in CPT symmetry and that there should be absolutely no difference between matter and antimatter, but as experimentalists, we want to find out if that's true. The tiniest discrepancy would lead to major new insights in to how the microscopic world works. The frequency comb technique that made these precise measurements feasible was celebrated in newspapers and journals around the turn of the century. The laser frequency comb, for the first time, provided a simple tool for counting the ripples of a light wave and determining optical frequencies of hundreds or even thousands of terahertz. It offered a phase coherent link between optical frequencies and radio frequencies. In the past, to establish such a link, you needed a huge laboratory and a very complex harmonic frequency chain and a big team of scientists to operate it. These big harmonic chains typically were engineered to measure just one single particular optical frequency. With a frequency comb, in a small scale laboratory on a tabletop, you now had a tool to measure not just one but any optical frequency. These things were simple enough that they could act as clockwork mechanisms for optical atomic clocks. So, what's the trick? How does a frequency comb work? At the heart of it, at least in the beginning, was a mode-locked laser, a femtosecond laser, which in its most basic form consists of two mirrors, an optical resonator, an amplifying medium, and a short light pulse bouncing back and forth. Because of amplification, one can couple out a train of other short pulses. These lasers were not new. They were used in hundreds of laboratories for ultra-fast science to study fast relocation processes in molecules and solids. Why do we talk about a frequency comb? If you have a laser that emits a short pulse and we send this pulse into a spectrometer, we find that we get a broad spectrum. There's no or very little structure. If you take the same laser but send two pulses into the spectrometer, they will interfere. We get interference we just said look like a frequency comb. The time separation between the pulses determines the spacing of this comb line so the farther apart the pulses are, the closer together the comb lines move. The only thing is these comb lines are not very sharp. If you have mode-locked lasers that emits many pulses, then the spectrometer, after a while, shows sharp comb lines. Lines that can be as sharp as of any continuous wave laser. For this to happen, we need a very precise controlled timing of our laser. One can look at it in a different way. I can look at it in terms of frequencies. People in precision spectroscopy typically would use a single mode laser, so you also have a cavity, two mirrors, but there's only one single mode of this resonator oscillating a standing wave. The frequency is determined by the number of nodes of the standing wave. Out comes a pure sinusoidal traveling wave. If you had not one, but two modes oscillating at the same time of two different frequencies, then we expect a beat note. If you look at the superposition, it's like having two tuning forks. If they oscillate in phase, they reinforce each other. If they oscillate out of phase, they cancel each other through interference. If you have not two, but three, or let's say five modes, then we really have a laser that emits short pulses separated by a long time of dark. Inside the resonator, we have a pulse of light bouncing back and forth but if you look at the modes of the frequencies, they form a comb precisely spaced. The spacing between comb lines has to be exactly equal to the repetition frequency of the laser. If you see that in the dark period between pulses the comb lines are still on, they're on all the time, but they can cancel each other through mutual interference. It is a very simple principle. Nonetheless, it was surprising to most experts how far this could be pushed. That you could take deep red pulses from the titanium-sapphire laser, send them into a microstructured fiber, and broaden the light through self-phase modulation and other nonlinear processes so that you get white light. But not ordinary white light, but light that consists of a comb of evenly spaced lines, still precisely spaced by the repetition frequency. Now we can talk about 100 thousand or a million comb lines. We know the spacing precisely. We don't know the position precisely, but there are nonlinear tricks how we can find out the position. The problem is successive pulses do not necessarily look identical. There can be a slippage of the carrier wave relative to the pulse envelope and so as a result, the entire comb is displaced by a frequency that we call the carrier envelope offset frequency. This was not known at first. But if we have a broad comb, if we produce second harmonic and sum frequencies in a non-linear crystal, then the sum frequency comb will be displaced by twice the unknown frequency. If you look at a beat note, original comb versus sum frequency comb, it can reveal this unknown carrier envelope offset frequency. If you know it, you can take it into account or you can use several controls to make it go away. Frequency combs have become a common feature in many laboratories. There are hundreds in use nowadays. Mostly based on telecommunications fiber laser technology, but many laboratories are working towards alternative frequency comb sources, including comb sources you can microfabricate on a chip. You can make microfabricated toroidal resonators that can act as optical parametric oscillators by four-wave mixing. This is a picture taken from the laboratory of Tobias Kippenberg in Lausanne. One reason why there is so much interest in developing new and smaller frequency combs is that the whole evolutionary tree of applications has evolved and is still evolving. Applications that go way beyond the original purpose of this comb as a tool for measuring the frequency of light. One application that I mentioned in the beginning that has come very far is the evolution of optical clocks. If you look at the development of clocks, from sundials to pendulum clocks to quartz clocks to caesium atomic clocks, we see that major progress of this became possible if one went to a faster pendulum. Now with our frequency comb, we can use atoms or ions that oscillate with the frequency of light, There are many possibilities. You can have laser-cooled trapped ions. You can have cold neutral atoms held in an array of laser tweezers in an optical lattice. People are also experimenting with molecules. If you look how far we have come, caesium clocks have evolved over a long time. They are approaching relative accuracy of better than a part in 10 to the 15. But optical clocks are now touching a part in 10 to the 18. The best present clock has been demonstrated in Boulder in the laboratory of Jun Ye, a strontium lattice clock. These clocks are now so accurate that if you take two of them and lift one just by one centimeter, you can tell a difference in tick rate due to Einstein's general theory of relativity. One can say this is terrible, when it's not enough to tell time, but I have to tell you precisely where my clock is. Of course, that is true, but that's this fabric of space-time as we understand it today. To compare two clocks, of course, is easy if they're right next to each other but, of course, metrology laboratories far apart would like to compare their clocks. In the past they have done clock comparisons using satellite signals, but this runs out of steam at maybe 15, 16 digits of precision. So to compare optical clocks, one needs a different way. One thing that works amazingly well is frequency transfer via optical fibers, via telecom fibers. You cannot use unmodified telecom networks because they don't allow coherent light to travel through. So one needs a special fiber link using bidirectional fiber amplifiers. The first really long distance link has been demonstrated a few years ago between the German metrology lab of the PTB in Braunschweigh and our laboratory in Garching. Katharina Predehl, she gained her doctorate by realizing this link, by noticing many fiber amplifiers along the way. This link, almost 1,000 kilometers long, has more recently been used in a loop of 2,000 kilometers of distance. It was demonstrated that it can give you an accuracy on the order of 10 to the minus 19 within just a few minutes. It's very well suited for clock comparisons. If you want to compare clocks between Munich and Braunschweig, you need to take into account the gravitation of redshift due to the height difference of four times 10 to the minus 14. But, of course, if you really would compare to 10 to the minus 18, tiny changes in local gravitational potential could be detected which is bad if you want to make good clocks, but it could be interesting if you want to learn about our planet. I should say that right now there are many links being built or already in operation in Europe. There is one link that has recently become operational between the PTB and the Observatoire in Paris. There will be a link to the National Physics Laboratory in Teddington. There is a link between the Italian National Metrology Institute, INRIM, and LENS. All these will be connected so we will have a network of clocks. What one will be able to do is to learn about the local gravitational potential so we can do precision relativistic geodesy. If you want to know what would be the time that is not affected by a local potential, then we really need to put a reference clock in space. Is it possible? Can we put an optical atomic clock in space? There has been a lot of interest because of applications of the space clock. So the reference clock for, that's immune to terrestrial potential, but also for astronomical long baseline interferometry, for tests of general relativity, tests of the equivalence principle, detection of gravitational wave, search for possible variations of fundamental constants, space navigation, formation flights of spacecraft. In April, two months ago, a frequency comb, together with two clocks, was actually tested in space in a sounding rocket during six minutes of microgravity. It's an experiment designed to compare rubidium microwave clocks. These clocks are small and they're already in use in the GPS system and satellite navigation. An optical clock, a rubidium optical clock, not one of the most precise optical clocks, but an optical clock nonetheless, and a frequency comb to compare the optical frequency and the microwave frequency. This is an experiment that was first scheduled to launch two years ago. Every six months, a team went up to install it into the rocket, only to be sent home again because there were difficulties. Finally, this April it worked. That's why I even made up a little movie to show the first frequency comb in space. It's not really space. It's a rocket that reaches 260 kilometers. The whole package is here, about 20 kilograms in weight. So here is the launch in Sweden at the Esrange Center. There's even some sound but that's not connected now. You can track the rocket with a telescope from the ground. Eventually you can use the onboard camera from the payload to look down. It's over northern Sweden and one has to launch it during a time when the lakes are still frozen. The most exciting thing of this was that everything was locked and operating during this phase. So here are some pictures that show that it's almost space, and everything kept working nicely until the end when the parachute opened for the recovery of the payload. Things fell out of luck. It's a collaborate experiment under the guidance of the Deutsches Zentrum für Luft- und Raumfahrt It shows that it's a viable technique also for space applications. The nice thing is that the system was still in working order after recovery. It was on display last week at the laser fair in Munich. What other applications are there? There are many, but because I don't have time, but let me mention one that we didn't think about at all. Astronomers could be interested in frequency combs as calibration tools for astronomical spectrographs. There is, for instance, now an astro-comb in operation and a solar telescope on the island of Tenerife. Just a few months ago, a paper was published in New Journal of Physics by Rafael Probst, one of our doctoral students who will have his doctoral exam next week. He explains the solar comb in a video abstract that was published together with this article. If you are curious, you can watch it. Since January, there are four astro-combs in operation at astronomical observatories around the world. One is at the 3.6 meter ESO telescope at La Silla, Chile. This is a telescope with which many exoplanets have been discovered. It uses a very precise optical spectrograph, the HARPS spectrograph, High Accuracy Radial velocity Planet Searcher. The idea of this kind of planet search is that you look for small Doppler shifts in the spectral lines from the star. As a planet orbits around the star, the star also wobbles around the common center of gravity. If the star moves towards the earth, its lines will be blueshifted. If it moves away from the earth, it's lines will be redshifted. These are very tiny shifts. The lines are broad and one needs to look not at one line but at many together in order to have a chance. Almost 1,000 planets have now been discovered, mostly big, Jupiter-like gas planets, but also smaller planets. The smallest requires repeatability corresponding to a Doppler velocity of one meter per second. If you wanted to look for an Earth-like planet around a sun-like star, you would need a repeatability better than five centimeters per second and the current calibration techniques do not offer that, but the astro-comb can. The idea is that you take a frequency comb and you feed the light into the spectrograph together with the star light. It's easily said, not so easily done because this comb, of course, has to have the right line spacing to be compatible with the detectors of the spectrograph. In a test run on a remote mountain flight, day and night without experts in laser physics next to it, so this has been done. Here is an example of what you get on the echelle spectrograph when one compares thorium-argon Lamb calibration lines and calibration lines from the astro-comb. The astro-comb is much nicer. It gets calibration lines everywhere. The lines can be all of equal intensity, but best of all, they don't drift because of pressure shifts or things like that, but the accuracy of these lines can be known with the precision of an atomic clock. Here's an example where you have the astro-comb together with star light, with, in this case, iron and nickel absorption lines of the star. This can be used to hunt for planets. The astro-comb in La Silla has been installed in January. Now we are waiting to see what will be discovered. Now, astronomers start to dream about even more ambitious things. We know that the space between galaxies is filled with hydrogen gas clouds. How do we know? If you look at the light of this and quasars, we see a forest of absorption lines. It's all the same lines, the hydrogen Lyman-alpha line or Lyman-beta line, but shifted due to the Hubble shift. If the universe continues to expand at an accelerating speed, as Saul Perlmutter told us yesterday, then we should be able to see some evidence by looking at the positions of these lines today and maybe again 10 years from now. Interesting aspect of frequency combs for precision spectroscopy. There are many other things I could talk about. Let me just quickly at the end tell you about something that I didn't think about at all but that has become intriguing. That you can use a frequency comb with its hundred thousand or million lines, use all the lines at once to look at complex molecular spectra. Highly multiplex technique of spectroscopy, of course, has been known for a long time which is called Fourier transform spectroscopy. This technique has not changed much since the beginnings. Essentially use Michelson interferometer, a light source, typically an incoherent light source. You send a light through a sample onto a photo detector and if you move an arm you get something like the autocorrelation function interferogram, and you can Fourier transform it to get a spectrum. These kinds of things are used everywhere but with frequency combs, you can make something like a Fourier spectrometer that can work enormously faster and can be enormously more sensitive. Nathalie Picqué, a visiting scholar from the French CNRS, has been perfecting these techniques in our laboratory for a number of years. One approach is you take not a mechanical interferometer, but two frequency combs tuned to two slightly different frequencies. You simulate the Doppler shift of a moving mirror simply by changing the repetition frequency and in doing so you can simulate a Michelson interferometer. A mirror moves ten kilometers per second, an escape velocity from Earth. If you do that, you get signals that are not in the acoustic frequency regime where there are lots of perturbations, but in the radio frequency region at the same time because you use laser light, you gain over an incoherent light source. Typically, the detector sees big bursts when the two pulses overlap and in between interferograms that you can Fourier transform to get a spectrum, in this case, some intercombination brand of acetylene. This is not a particularly impressive spectrum. You can do that with a conventional spectograph but in a conventional spectrograph it would take a few minutes before you reach this kind of resolution, whereas the spectrum shown here has been recorded within 42 microseconds. That's of interest, if, for instance, you wanted to record spectra of transient species that don't live long enough. One way of understanding this dual comb technique is to say I have two optic combs with slightly different comb spacing and I have these pairs of comb lines that interfere to produce a radio frequency comb. I translate optical frequencies from terahertz units to radio frequencies in megahertz. You've shown that you can record spectra very rapidly. You can also, if you are willing to wait for a few seconds, get very high resolution. Here is a movie that shows you how we can zoom in into such a spectrum with acetylene lines and to see how much we zoom for comparison, there is a movie, Powers of Ten, stolen from YouTube. Let's see. Okay, so now we start to zoom. There must be many comb lines because we still don't recognize them. Here we start to see individual comb lines and one can zoom in to see how sharp are they. Zoom, zoom. They have each a magnification of two million, so in a recording time of 2.7 seconds, we have 268 million data points. We can look at each of the 120 thousand comb lines and the resolution of the optical array is 200 kilohertz. We need two frequency comb sources, a single detector, and a computer, and we get very short acquisition time, extreme sensitivities from low to extreme resolution. We can have extreme accuracies because the frequency of each comb line can be known with the accuracy of an atomic clock. We can look at both absorption and dispersion. The whole method works from terahertz to the vacuum ultraviolet. Recently, Nathalie Picqué and her team have demonstrated nonlinear two-photon spectroscopy. We have no time, I will just show some dream that you can use dual comb Raman spectroscopy to do label-free bioimaging. The first steps have been published in a paper in Nature. There are, again, horizons what you might do with this technique ranging from benchmarks for quantum chemistry to nanophotonic on-chip chemical sensing and label-free bioimaging. I could give many more talks on other aspects including attosecond science. It's maybe good to remember that all of this was not planned. It was the outcome of some curious research, we were curious whether quantum mechanics describes a simple hydrogen atom precisely. In the process, we came up with tools that are now valuable in many other endeavors. In the end I have to thank my sponsors, in particular the European Research Council, the Max Planck Förderstiftung, and the Carl Friedrich von Siemens Stiftung, that have made it possible that we can continue to work, even though otherwise I would have long retired in Germany.

Theodor W. Hänsch (2015)

Science with Combs of Light

Theodor W. Hänsch (2015)

Science with Combs of Light

Abstract

The spectrum of a frequency comb, commonly generated by a mode-locked femtosecond laser, consists of several hundred thousand precisely evenly spaced spectral lines. Such laser frequency combs have revolutionized the art of measuring the frequency of light, and they provide the long-missing clockwork for optical atomic clocks. As tools for precision spectroscopy, notably of atomic hydrogen, laser combs permit stringent tests of fundamental physics laws. New applications are evolving which go far beyond the original purpose. As calibration tools in astronomy, frequency combs are facilitating the search for exoplanets, and they may lead to direct evidence for the accelerating expansion of space in our universe. Laser combs are also becoming powerful instruments for broadband molecular spectroscopy. They can dramatically improve the resolution and recording speed of Fourier spectrometers, and they are creating intriguing new opportunities for highly multiplexed nonlinear spectroscopy and microscopy.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content