I have never written an autobiography before; I always feel uncomfortable to talk about myself. But this time, thanks to Hai-Lung Dai and Tim Lian, I was persuaded to give it a try as it would give me an opportunity to reflect on my academic life and express my sincere gratitude to my teachers, colleagues, associates, and friends whose influences on me have made my life most fulfilling. I am a believer of hard work, but also a believer of luck. I have been fortunate throughout my life that at important junctions, I always seemed to get help from the right person(s).
I was born in Shanghai and moved to Taiwan near the end of the civil war in China. I received my B.S. degree in electrical engineering from National Taiwan University in 1956. Electrical engineering was the most popular major in Taiwan universities in the early 1960s; physics became popular only after Yang and Lee won the Nobel Prize. The academic community of Taiwan at the time was very much isolated from the world. There was barely any academic information about the West. For an engineering degree in Taiwan, no advanced science courses beyond general physics and partial differential equations were required or recommended. After receiving my B.S. in EE and spending 18 months on military training and services, I had the choice of either finding a job in the industry or going abroad for graduate studies. In the late 1950s, the Chinese student population in US universities was very small; it was nearly impossible for students from Taiwan to be admitted to a top university with financial help. I was extremely lucky to have won a fellowship sponsored by a Chinese-American foundation that would pay the tuition and living expenses for a student to study abroad in any US university for a year. This allowed me to get into Stanford to earn an M.S. in EE.
I arrived at Stanford in the summer of 1958. It was the time semiconductor devices started booming and Silicon Valley began to form. I decided to go into modern electronics but quickly realized that I needed to patch up my physics background, particularly on quantum mechanics and linear algebra. I took an elective course on semiconductor devices from John Moll, who just moved from Bell Labs to Stanford. Moll was a prominent star in the field of semiconductor devices, highly respected in the electronics industry. He was also a very friendly, nice person, always willing to help. I was truly indebted to Moll for his kindness and willingness to take me under his wings and provide me with all sorts of help: advising me on research, finding me a beginning research job at Hewlett-Packard for two summers, and co-authoring with me on my first publication.
I could continue to work with Moll on silicon devices after the first year at Stanford, but I was then more attracted to physics. I was not sure I had enough background for graduate studies in physics. So, I applied to the Division of Engineering and Applied Physics at Harvard. I was fortunate to be admitted with a fellowship. Upon leaving Stanford, Moll advised me that I should look for research opportunities to work for Nicolaas Bloembergen at Harvard. Apparently, Bloembergen was already very famous at the time for his nuclear magnetic resonance (NMR) work and his three-level maser scheme, but I was so ignorant that I did not even have any idea who Bloembergen was. It turned out that on my first day at Harvard, I found surprisingly Bloembergen was already assigned to be my academic adviser. Fellowships allowed me to spend my first two years at Harvard all on course work. By the end of the second year, Bloembergen told me that I ought to pass the qualifying exam and move on to thesis research. I took the exam in mid-June. Having taken the final exams on courses only two weeks ago, I was quite confident going into the oral qualifying exam. However, this was the first time I ever talked in front of a group of professors. As the exam went on, I got more and more nervous; toward the end of the 2-h exam, I was hardly able to speak and had to rely on hand gestures to answer questions. I did pass the exam. After congratulating me, Bloembergen told me to meet with him in his office the next day. Next morning, I went to his office expecting that he would instruct me how to find a research group to carry out Ph.D. thesis work. Instead, he simply told me what my thesis topic would be. Thus, without any hassle, I took the advice of Moll and became a member of the Bloembergen group.
I started working in Bloembergen’s lab in the summer of 1961. It was an exciting time. The first (ruby) laser invented by Maiman was announced in May 1960. The first nonlinear optical effect (second harmonic generation in quartz) was reported by Franken et al. in August 1961. Bloembergen had set his mind on shifting his research from NMR to nonlinear optics. However, commercial lasers were not yet available and university labs generally were not capable of building a laser. The first ruby laser Bloembergen purchased for his lab from Trion and Raytheon Corp could only be scheduled to arrive some time in 1962. Nonetheless, the excitement over the opening of an unfathomable brand-new field was there; even students as ignorant as I could feel it. Every afternoon around 4 p.m., Bloembergen would come to the lab and gather his group in the corridor, describing to us the most recent advances in laser sciences and his insightful comments on them. He repeatedly pointed out that he perceived nonlinear optics to be such a wide-open field that to find a meaningful research problem, one only had to open a page of Born and Wolf and ask how the optical effect described on the page would change if the relevant optical coefficient became nonlinear. Without a laser available for experiment at the time, the group discussions focused mainly on the basic physics underlying nonlinear optical effects. In September 1962, Bloembergen, together with Armstrong, Ducuing, and Pershan, published the classical paper of “Interactions between light waves in a nonlinear dielectric,” establishing the theoretical foundation of nonlinear optics. At this early stage, Bloembergen’s works on nonlinear optics were all theoretical. My Ph.D. thesis was an experimental project studying “Faraday rotation of rare earth ions in CaF2.” To make connection with nonlinear optics, we started to think about the inverse Faraday effect that would generate a magnetic field through optical rectification (OR). I was very grateful to Bloembergen for letting me work with him on two specific theoretical topics. One was on nonlinear response coefficients of material systems under strong and/or weak excitations of multiple laser frequencies. This became an important subject later with the arrival of tunable lasers. The other was on the coupled wave approach to properly describe stimulated Raman scattering (involving coherent Stokes and anti-Stokes Raman generation) by treating molecular vibrations as waves. Unlike other theories at the time, our work predicted that the coherent anti-Stokes radiation should be suppressed in the phase-matched direction and therefore would appear with a dark ring in the output angular distribution. This prediction was later experimentally verified. Naturally, I learned a great deal from Bloembergen. He was extremely sharp. He had very broad and deep knowledge in physics. He was gifted with tremendous physical insights. He had clear physical pictures in mind, but he could also carry out complex theoretical calculations. Furthermore, he was most efficient at work. I witnessed him completing a draft of our paper for Physical Review Letters in a single morning. It was certainly the best of my luck to have him as my mentor. Bloembergen wrote the first book on nonlinear optics (published by Benjamin Inc. in 1964) in record time. I picked up nonlinear optics mainly through proofreading the book sentence by sentence and equation by equation when he handed us chapter by chapter nearly every week. I learned from him how physical pictures are most important in discussing physics. My first discussion with him lasted only 5 min because I was not able to answer his query physically. I was very proud of myself that by the end of my second year in his group, I could lengthen our discussion to half an hour.
In 1964, I was ready to move on to start my own career. I received an offer from Purdue University, and then, Bloembergen informed me that there was a possible opening at Berkeley. I was asked to meet with Professor Walter Knight, Professor Alan Portis, and Professor Michael Tinkham from Berkeley at an APS conference. The meeting was held in the lobby of a hotel and lasted about half an hour. A few weeks later, I received the offer of assistant professorship from Berkeley. I have been employed by Berkeley ever since. Now, in retrospect, it was indeed remarkable that I could get into Berkeley so easily. But the ensuing life at Berkeley was not. I was naïve about academic life at the time, not knowing what would be expected from me as a young faculty member. I did not ask what would be provided for me to start research at Berkeley. Upon arrival at Berkeley, I realized that I had no office, no research fund, no student—nothing but a cabinet of lenses normally used for classroom demonstration of classical optics. I was advised to write a research proposal to the Office of Naval Research and hopefully get funded the next year. In the meantime, I could borrow money from William Nierenberg’s group to start research. Fortunately, Bloembergen happened to have already planned to spend a sabbatical year at Berkeley, and his coming to Berkeley together with me made my first year at Berkeley easier and more rewarding. I shared the office with Bloembergen and gained the real privilege of being able to have daily discussion of physics with him. We continued working on theories of stimulated light scattering in various media.
In the early 1960s, the Berkeley physics department was prominent in the world for its high energy physics research. Solid state physics was also very strong, particularly in nuclear and electronic magnetic resonances. I was, however, the only person hired to work on modern optics and could hardly find anyone in the department to talk to. (George Pimentel in chemistry just started his chemical laser work and John Whinnery in electrical engineering just started his quantum electronics group.) I was very appreciative when Sumner Davis, an optical spectroscopist known for his work on molecular emission spectroscopy from the sun, invited me to participate in their weekly group meetings.
I was not familiar with the tenure system at the time, but I was told that I must publish to get promotion. Without a lab to work on for the first two years at Berkeley, I had to think about possibly publishing theoretical papers. I was lucky that in my last year at Harvard, I had the opportunity to audit the lectures on quantum optics by Roy Glauber, who just completed his pioneering work that sets the theoretical foundation of quantum optics. I thought I could extend his formalism to nonlinear optical processes and succeeded in publishing a paper in Physical Review on “Quantum statistics of nonlinear optics.” Glauber happened to like it. When he organized the International School of Physics, Enrico Fermi, on quantum optics, at Varenna, Italy, in the summer of 1967, he very kindly invited me to give a lecture there on my work. This was my first ever trip to Europe. On the weekend at the summer school, Marlan Scully and two others planned to have an excursion to the Swiss Alps by car and asked if I would like to join them. I was not sure I could because I did not have a visa. (This was before the time of the European Union.) They said: “Don’t worry. The border check would be casual.” At the border entering Switzerland, a border patrol asked who we were. Scully replied: “We are physicists.” The border patrol said cheerfully: “Ah, physicists. Go!” We believed that he must have misunderstood that we were either physicians or athletes.
I was only able to start setting up a lab in late 1965 after I received the first ruby laser from Quantronics. With only one beginning student, I had to pick one problem to work on. I chose to investigate a reported mysterious effect of stimulated Raman scattering in liquids, which turned out to result from self-focusing of light. We began to learn to build Q-switched ruby lasers and mode-locked Nd:glass lasers ourselves. In 1967, fortune once again descended on me. Leo Brewer in the Chemistry Department convinced the Atomic Energy Commission (AEC, now the Department of Energy) to let him expand his chemical materials center in the Lawrence Berkeley National Laboratory (LBNL) into a Materials Science Division (MSD) and invited several physics faculty members, including me, to join the Division. MSD provided each of us a yearly budget that allowed us to keep a sizable research group and carry out a relatively flexible research program. Thus, I was no longer under pressure to submit proposals to different agencies to get research funding. For this, I am forever grateful to Brewer. He made my academic life so much more pleasant. I remained as a member of MSD in LBNL until a few years after my retirement.
For an experimental research group, the success depends critically on research funding and, more importantly, people working in the group. With the stable support of MSD, I could set up a sustainable group of reasonable size. Throughout my academic career, I always have had the fortune to have a group of talented and motivated students, postdocs, and visiting scientists working with me. I am indebted to them for the success of our group. Without their help, my life would be very miserable. Although I am tempted to describe their contributions and thank them individually here, limited by space, this seems impossible. In the following, I will focus on selective episodes that will touch upon some, but not all, of our group members. For this, I apologize in advance.
Our experimental nonlinear optics research at Berkeley began with studies of self-focusing. Self-focusing/self-trapping of light was a hot topic in the late 1960s that had attracted the attention of many eminent scientists, including Charles Townes and Alexander Prokhorov. It was observed that when a Q-switched laser pulse passed through a liquid cell like benzene over a critical cell length, multiple bright dots in the beam would appear at the exit plane. These dots had a constant diameter of a few microns. Townes believed that due to optical Kerr effect in liquid, the laser pulse would first self-focus and then self-trap into light filaments. Weak irregularities in the beam profile could lead to breaking of the beam into multiple filaments. There was, however, the problem that self-trapping of a nanosecond laser pulse is not a stable, but metastable, phenomenon. On the other hand, it was clear that the focal position of self-focusing should depend on the laser power, and because the power of a laser pulse varies with time, the focal point(s) should move with time. A single-mode laser pulse would create a single moving focal spot along the beam axis that would appear as a bright track of light like a filament; a multimode laser pulse would break into multiple moving focal spots appearing as multiple filaments. To verify the conjecture beyond doubt, we needed to carry out an experiment with a single-mode Q-switched laser. Michael Loy, an outstanding beginning student, joined my group and picked up the project. Despite his initial unfamiliarity of lasers, he successfully converted our ruby laser into single-mode and proved the moving-focus picture in quantitative agreement with theory. In the VI International Quantum Electronics Conference in Kyoto in 1970, I had the luck to present a contributed paper on self-focusing immediately after the invited paper by Prokhorov on the same subject. Prokhorov used a lot of mathematics to describe possible moving focus. By describing the moving focus picture in physical terms supported by experimental result, I probably attracted more attention from the audience.
Far-infrared spectroscopy (now renamed as THz spectroscopy) is important for material studies, but in the 1970s, it was plagued by weak light sources and insensitive detectors. Paul Richards at Berkeley was among the very few renowned experts in the field at the time. I convinced him that difference frequency generation (DFG) and optical rectification (OR) could provide him with relatively intense far-IR pulses, and we ought to give it a try. With Dillard Faries from my group and Keith Gehring from Richards’s group, we embarked on the project using two temperature-tuned ruby lasers synchronously Q-switched to generate a tunable far-IR output from a LiNO3 crystal. The laser pulses were generated on demand; this means that each shot had to be triggered manually. The far-IR detector was a doped InAs at 1.5 K. The far-IR output spectrum was measured using a Michelson interferometer or Fabry–Pérot. The experiment was a success, verifying our prediction. Later, with my student Jim Morris and Richards’s student Patrick Yang (who was the father of Andrew Yang, one of the democratic presidential candidates in 2020) joining us, we extended the experiment to far-IR generation by optical rectification of picosecond mode-locked laser pulses. It was then obvious to us that such a system would be impractical to use in practice unless the laser pulses could have a high repetition rate. Subsequently, Richards gave up on possible improvement in far-IR sources. He turned his attention to improvement in far-IR detection using Josephson junctions and shifted his research interest to the study of cosmic blackbody radiation (CBR). His student, John Mather, later received the Nobel Prize for mapping out the anisotropic CBR. The comment on our project at the time was “what a stupid idea,” but years later with the arrival of cw mode-locked lasers with MHz repetition rates, far-IR generation by DFG or OR soon became very popular. What would take us one long day to collect a far-IR spectrum now takes less than one second, and the far-IR or THz spectroscopy field has blossomed. This is a wonderful example of how technology advance can drastically change a field.
In my early years at Berkeley, I got to be close friends with Erwin Hahn and Marvin Cohen. Hahn was the inventor of spin echoes, which are of crucial importance in high resolution nuclear magnetic resonance (NMR) spectroscopy and NMR medical imaging. He was extremely innovative, always full of interesting ideas. He had become interested in extending coherent transient effects in NMR to the optical regime. He often spent long hours describing to me his thoughts about possible transient effects associated with optical excitations. I learned a lot from him on how to convert physical effects between multiple rotating frames and the lab frame although I easily got lost after one rotating transformation. Cohen and I started in Berkeley together in 1964. We are of the same age with birth dates only a few weeks apart, but he was much more knowledgeable than I in all respects. He taught me what I ought to know about the tenure system and how to survive in the academic world. His advice always seemed to be right to the point. Even now, I will go to him when I run into problems. Cohen is a top theoretical solid-state physicist, exceptionally intelligent with very broad interest. He was already well known in the 1960s for his works on superconductivity and electronic band structure calculations. Being supported by MSD of LBNL, I was under some pressure to do research on materials. Cohen and I frequently talked about possible collaborations. The optical spectroscopy of crystals was a possibility; the electronic band structure of a crystal could be deduced from the electronic spectrum of the crystal. One day, Mel Klein, a biophysicist from Melvin Calvin’s lab, showed me a preprint sent to Calvin by Stacy French at Stanford. It described a wavelength modulation spectrometer that allowed derivative molecular spectroscopy. This was the time when modulation spectroscopy on crystals by different means of modulation was getting popular because they could improve spectral resolution of reflection spectra from crystals. Wavelength modulation spectra would be the simplest to interpret as it does not involve modification of any material property; the calculation to fit the observed spectra could lead to more accurate band structures for crystals. So, I decided to start a project on wavelength modulation spectroscopy of solids. (Many years later, I learned French was an old friend of my father-in-law. They worked in the same lab at Harvard when they were young. What a coincidence!) I was fortunate to have Ricardo Zucca, a superb, mature, and experienced student from Argentina, willing to work on the project. He already had a family with several children at the time and was strongly motivated to earn his Ph.D. as quickly as possible. Following French’s design, he quickly constructed our version of a wavelength modulation spectrometer and started measuring differential spectra of different crystals, which Cohen’s group would use to calculate their band structures. They were able to obtain, for example, the most accurate band structure of Si and Ge at that time. Our experimental measurements were efficient. We soon ran out of interesting crystals to measure. This was a rare case that in a theory–experiment collaboration, experiment could move faster than theory.
As a member of MSD, I was supposed to do some research on material characterization. We began working on laser Raman spectroscopy. Luckily, in the early 1970s, three exceptionally capable postdocs/visiting researchers came to join our group: Yves Petroff, Arnold Schmidt, and Peter Yu. [Petroff later served as the director of the European Synchrotron Radiation Facility (ESRF) when ESRF was established. Schmidt became a professor at the Technical University of Vienna and served as the president and chair of the supervisory board of the Austrian Science Fund. He was the founding father of the Institute of Science and Technology of Austria and the mentor of Ferenc Krausz, the 2023 Nobel laureate in physics. Yu was recruited back to Berkeley from IBM. He and Manuel Cardona wrote the most informative, well-circulated book on semiconductor physics, Fundamentals of Semiconductors.] Together with a group of outstanding students [Don Bethune (winner of APS McGroddy Prize for New Materials), Tai Chiang (winner of APS Davisson-Germer Prize for surface physics), and others], they helped me expand our research to Raman scattering and luminescence spectroscopy on crystalline solids and built a reasonably productive program on linear and nonlinear optical studies of materials (gases, liquids, and solids) at MSD. We also enjoyed having Francesco DeMartini (from Rome) spend a sabbatical leave with us working on surface polaritons.
One day in 1970, Nabil Amer from LBNL showed up in my office and introduced “liquid crystals (LC)” to me. I was fascinated by this new material that I had never heard about. Amer would like to collaborate with us on Raman studies of liquid crystals (LC), which I readily agreed, but later after I learned more about LC, I was more attracted by its correlated molecular behavior and associated giant Kerr nonlinearity. The material has a refractive index that can be appreciably changed by a field strength of about 1 V/μm, and field-induced reorientation of correlated molecules is the underlying mechanism. (LC displays based on the effect were developed and commercialized in the mid-1980s.) In this respect, dc and optical fields are equally effective. We carried out a series of nonlinear optical studies on LC, including harmonic generation, self-focusing, nonlinear Fabry–Pérot, and optical-field-induced phase transitions, over many years. Knowing that MSD was not interested in LC, I applied to NSF for research support on LC. I was fortunate to get a regular-size grant that I renewed successively for more than 20 years. Only for one year around the mid-2000s, I was told that the grant could not be renewed because the description of the outreach section in the renewal proposal was not good enough. I talked to the grant monitor, Fred Stafford, when he came to visit Berkeley. He asked, “Did you have female students on the project?” I said, “yes, I had a former female student now on the Harvard faculty, another one in NIH, and a female postdoc on the faculty of ETH, Zurich.” He said, “That’s it. You should revise the outreach section.” Indeed, the revised proposal went through with no problem. Until the late 1990s, our group was the only group working on soft condensed matter in the physics department of Berkeley. Our research effort on LC, later extended to polymers, was small but quite successful. We were able to attract many outstanding students and postdocs/visitors. Among them, Hui Hsiung received the second Glenn Brown Prize of the International Liquid Crystal Society for the best Ph.D. thesis; he later became the Vice President of the world’s second largest liquid crystal display company, AUO Inc.; Wei Chen is a current vice president of Apple Inc.; Xiaowei Zhuang, famous for her invention of super-resolution microscopy, STORM, is a distinguished professor at Harvard and an NAS member; Marla Feller is a highly recognized professor of neurobiology at Berkeley and an NAS member; Viola Vogel is a professor at ETH, Zurich, a world leader in bioscience, and a recipient of numerous international awards and honors, including the foreign membership of NAS.
The simplest experimental research project I have ever experienced is on optically induced spatial self-phase modulation of light in a nematic LC film. In the experiment, Steven Durbin and Serge Arakelian (visiting from the Soviet Union) inserted an aligned nematic LC thin film cell in the path of a single-mode argon laser beam and observed the beam diffracted into multiple rings on a screen at a distance. By measuring the ring diameters and spacings between rings with a ruler, they could quantitatively explain the observation as resulting from self-phase modulation in space quantitatively due to giant optical Kerr nonlinearity of LC. The shape of the rings directly exhibited the mode pattern of the laser beam. We used to joke that for us to keep high-tech information from a Russian scientist, we allowed him to use only a ruler in measurements. The work was published in Optics Letters and has garnered a citation over 500. Not bad! From such an experiment, it was clear to us that nematic LC probably had the strongest optical Kerr nonlinearity of all materials although its response was slow because of strong molecular correlation; we should expect to observe extraordinary nonlinear optical phenomena in nematic LC. I was grateful to Francesco DeMartini for sending his best student, Enrico Santamato, to Berkeley to spend a year with us. Santamato was exceptionally strong in both theory and experiment. He performed several fascinating nonlinear optical experiments in LC well supported by theory: laser-induced collective molecular rotation, direct measurement of orbital angular momentum of light, self-induced stimulated light scattering, laser-induced nonlinear dynamics, and others. He is currently a professor at the University of Napoli, Italy.
My only research work directly related to industrial interest is also on LC. In the late 1980s, we shifted our LC research to studies of LC interfaces and surface-induced LC bulk alignment after we had initiated second-harmonic and sum-frequency spectroscopy as a surface-specific technique (to be described later). It was the time when the LC display industry was booming. The LC display panel was an array of field-biased LC cells, each of which was constructed by sandwiching an aligned nematic LC film between two polymer-coated windows. To prepare an initially well-aligned LC film along a certain direction, the polymer surfaces were rubbed along that direction. Not knowing how macroscopic mechanical rubbing could lead to LC bulk alignment, this was considered black magic. Most people believed that rubbing created grooves on polymer surfaces, and since LC molecules next to the polymer would fall alongside into the grooves, they would in turn align the bulk LC molecules along the grooves through strong LC molecular correlation. Using second harmonic and sum frequency spectroscopy in a series of experiments over several years, we (Wei Chen, Marla Feller, Hui Hsiung, and others) were able to find the correct answer: from measurements of orientation distributions of LC monolayers on different rubbed and unrubbed surfaces, we found that the monolayers on selected rubbed polymer surfaces are oriented along the rubbing direction through the LC–polymer interaction. Xiaowei Zhuang and Lorenzo Marrucci (visiting from Italy) later showed experimentally that the oriented LC surface monolayer led to the LC bulk alignment in quantitative agreement with theory. There were quite a few other students, postdocs, and visiting scientists working on some other interesting surface and nonlinear optical studies of LC, including optical bistability, optical transistor, surface memory, surface order/disorder transition, orientation wetting of LC, and enhancement of Kerr nonlinearity of LC by dye doping. I am grateful to them for their tremendous contributions to our LC research at different stages and keeping it moving forward healthily.
I was very fortunate to have the privilege to collaborate with Yuan T. Lee for many years and become a lifetime friend with him. Yuan Lee, world-renowned for his work on chemical reaction dynamics using molecular beams, was recruited back to Berkeley in 1974. I occasionally played ping-pong and talked about science with him. In the early 1970s, Ambartzumian and colleagues of the then Soviet Union reported a discovery that created a great deal of excitement in the laser science community. By irradiating an SF6 gas cell with a CO2 laser beam tuned to the proper frequency, isotope-selective dissociation of SF6 was observed. This suggested a simple attractive method for isotope separation, particularly separation of uranium if UF6 were used. The isotope selectivity came from selective vibrational excitation of isotopic molecules, but the dissociation mechanism was not clear. It could be due to multiphoton excitation of individual molecules above the dissociation level or excitations of molecules to high vibrational levels followed by collision-induced molecular dissociation. Many laboratories around the world, including Lawrence Livermore and Los Alamos National Laboratories, tried to solve the mystery. It was clear to us that if the experiment were carried out in a molecular beam, molecular collisions would no longer be a concerned factor. Yuan Lee and I went to the director of MSD and persuaded him to let us purchase a CO2 TEA laser and set up the experiment using Lee’s crossed beam machine. Coggiola and Peter Schulz from the Lee group and Aasmund Sudbo (now at U. Oslo, Norway) from my group formed a team to work on the project. We had a breakthrough on our first try. Soon after we turned on the CO2 laser, we were surprised by a loud alarm, indicating that the beam machine had suffered a vacuum leak. It happened that we had installed a CaF2 window on the machine to let the laser beam through but forgotten to check the transparency of the window at the CO2 laser frequency. The laser beam broke through the window in just a few shots. I was the obvious person to blame for the blunder because I was supposed to take care of the optical side of the experiment. After fixing the problem, the experiment was an immediate success; we confirmed that multiphoton excitation and dissociation of individual molecules were the mechanism, and we understood how stepwise multiphoton excitations could excite molecules above the dissociation level. We were naturally very happy to be able to win the competition against labs that had orders of magnitude larger budgets working on the same topic. In subsequent years, Yuan Lee and I continued our collaboration using the scheme of crossed laser and molecular beams on a host of molecular physics and reaction dynamics problems. Quite a few talented researchers participated in the projects. I enjoyed interacting with Edward Grant, Jim Lisy, Hoisin Kwok, Andy Kung, and others in the Lee group. I had the great pleasure of learning about reaction dynamics and other scientific matter from Yuan Lee. He is an exceptionally knowledgeable scientist. I also admire him as a fair, decent, conscientious, and righteous person. He worked extremely hard and seemed to need only a few hours of sleep every day. I had a hard time to keep up with him on a daily basis. Our collaboration ended a few years after he received the Nobel Prize, and I focused more on nonlinear optics for surface studies.
In 1974, Fleischmann et al. reported the observation of surface enhanced Raman scattering (SERS) from molecules adsorbed on a roughened silver electrode in an electrochemical cell, a discovery that stimulated great interest in the science community. An enhancement of 106 was found. Local field enhancement from local surface plasmon resonance on roughened silver and resonance enhancement from the excitation of adsorbate–silver charge transfer band were believed to be the underlying mechanisms, but their relative importance could not be determined. It dawned on me that the outputs of both SERS and SHG had a fourth power dependence on the local field correction factor and should have nearly the same local field enhancement, but the latter could be measured without the presence of adsorbed molecules. Chenson Chen and Ruben de Castro (visiting from Brazil) performed the SHG measurement on an electrochemically roughened silver surface and found a local field enhancement of ∼104. They could also readily observe the adsorption and desorption of a pyridine monolayer on a silver electrode during an electrochemical cycle as in the case of SERS. [Here, let me insert an episode that I find amusing. In 1984, Ted Hansch and I were charged to organize the VII International Laser Spectroscopy Conference in the summer. We selected the Maui Surf Hotel in Maui, Hawaii, as the conference site. Three months before the conference, Ted and I had to go to Maui for a day to check out the conference venues and settle some details with the hotel. We finished our business in the morning and could spend the afternoon relaxing in a beach bar. Our conversation unavoidably touched upon science. Learning about our SHG experiment on the silver electrode of an electrochemical cell, Ted said: “I just got a red diode laser from a company that has a 20-mW cw output (among the highest at the time). Will you be able to use it in your experiment?” A quick back-of-the-envelope calculation indicated YES despite the weak laser power. With the diode laser, the experimental setup would be small and easily movable. We then decided that if the experiment was indeed feasible, it would be fun to present a post-deadline paper at the conference showing the real experiment on site. I had no difficulty persuading Gary Boyd in my group to pick up the project after promising him the Maui trip if he could make it work. It was a successful presentation in the post-deadline poster session. Attendees could unmistakably observe the SHG response from the adsorption and desorption of molecules on the silver electrode as the electrochemical cell underwent oxidation and reduction cycles. We entitled the paper “Maui Surface Experiment.”]
The SHG signal from a roughened silver surface was so strong that it prompted us to realize that even without local-field enhancement, we should be able to detect surface molecular monolayers using Q-switched or mode-locked lasers. This then suggested that we could use SHG as a general surface analytical tool. We were fortunate to have had several most talented and dedicated people joining our group at the time: Tony Heinz (now at Stanford, NAS member), Harry Tom (now at UC Riverside), X. D. Zhu (now at UC Davis), and Daniel Ricard (visiting from France). They conducted a series of experiments, measuring the orientations and electronic spectra of adsorbed molecules, adsorption isotherm of molecules from solutions, and so on. They also worked out the basic theory of surface SHG to establish SHG as a viable surface probe. But we soon noticed that it was always difficult for newcomers to break into an established field. Surface science at that time was dominated by research interest on well-defined surfaces in ultrahigh vacuum although the existence of a “pressure gap” problem between surfaces in vacuum and in real environment was well known. To attract the attention of the surface science community, we needed to demonstrate that the technique works for samples in UHV. We also needed the support of a renowned surface scientist to enhance our credibility. I persuaded Gabor Somorjai in chemistry to collaborate with us. I borrowed a UHV system from him over the 1982–1983 Christmas/New Year holidays. Harry Tom and Zhu, together with Mathew Mate from Somorjai’s group, spent nearly the entire holidays to fix the leak of the system and successfully demonstrated that SHG could monitor in situ the kinetics of molecular adsorption and desorption on Rh(111). This started my collaboration with Somorjai for nearly 20 years.
With surface SHG passing the test, we turned to possible applications. It was my luck to have another group of remarkably capable students and visitors joining the effort: Besides Wei Chen, Hui Hsiung, Marla Feller, and Viola Vogel mentioned earlier, there were Theo Rasing (now at Radboud Univ.; winner of the Netherlands’ highest award, the Spinoza Prize), Mahnwon Kim (visiting from Exxon, now at KAIST; then president of Korea Institute of Advanced Study), Richard Superfine (now at U. North Carolina), Xudong Xiao (now at Wuhuan U., China), Jung Huang and Rupin Pan (now at National Chiao Tung Univ., Taiwan), and Garry Berkovic (visiting from Israel), among others. They demonstrated that surface SHG could be employed to studies of characteristics of surfactant and organic monolayers, phase transitions of Langmuir films, monolayer polymerization, surface ordering of molecules, surface magnetization, etc. Wei Chen and Philippe Guyot-Sionnest (now at U. Chicago) worked out a more detailed theory on surface SHG.
Surface diffusion was a subject of interest in the 1980s. Knowing that surface SHG had submonolayer sensitivity to map out the surface distribution of adsorbates, I envisioned that we could use SHG to monitor and measure the surface diffusion of molecules on substrates. We would first prepare a molecular monolayer on a substrate and then use laser desorption by crossed laser beams to create a 2D grating on the monolayer. Increasing the temperature would speed up surface diffusion to wipe out the grating, and SHG diffraction from the grating could be used to monitor how fast the grating would disappear, from which the diffusion constant could be deduced. Preparing the grating along different directions would allow the measurement of anisotropic diffusion. Zhu and Theo Rasing tried out the idea on CO on Ni(111) in ultrahigh vacuum. They struggled very hard to create the desired grating because the Q-switched Nd:YAG laser in hand had too much shot-to-shot variation. Once the laser was under control, the experiment was successful. But we soon realized that the grating could be easily monitored by linear optical diffraction as well; Xudong Xiao demonstrated the scheme in a study of anisotropic diffusion of CO on Ni(110). I should not have forgotten that in measurements, the linear optical technique is always more sensitive than the nonlinear one if both are applicable. Anyway, we had found a viable technique for studies of surface diffusion on a large variety of surfaces and interfaces. Unfortunately, interest in macroscopic surface diffusion seems to have waned since the 1990s; researchers these days are more interested in surface diffusion of atoms/molecules on the atomic scale monitored by scanning tunneling microscopy (STM).
Although we appreciated the simplicity of SHG as a surface probe, it was obvious to us that we needed vibrational spectroscopy information for better characterization of surfaces and interfaces. This was not possible with SHG because of the lack of sensitive infrared detectors. We had to resort to infrared–visible sum frequency generation (SFG). To demonstrate surface sum frequency vibrational spectroscopy (SFVS), Zhu and Hajo Suhr (visiting from Germany) used a discretely tunable CO2 TEA laser as the IR input and the second harmonic of a synchronized Nd:YAG laser as the visible input overlapping on a coumarin dye monolayer on silica. The TEA lasers were delicate, constantly having problems with various components. Nevertheless, after consuming four CO2 TEA lasers retired from our earlier experiment on the multiphoton dissociation of SF6 molecules with Yuan Lee’s group, they were able to record the vibrational spectra of the coumarin monolayer. After this initial success, we realized that to further develop SFVS, we needed a continuously tunable coherent IR source generated from an optical parametric (OP) system. As I recall, such systems were not yet commercially available in the 1980s. We had to build our own OP system. Harry Tom did have a simple OP system (pumped by a Q-switched Nd:YAG laser) built in Yuan Lee’s lab at the time that was meant for laser spectroscopy studies of molecular beams. He used the system to try out SFVS on a monolayer of p-nitrobenzoic acid (pNBA) on silica. Unexpectedly, he could observe strong CH stretch spectra even in the absence of pNBA. Only years later, we realized that the spectra must have come from the surface contamination of silica from atmosphere. Mechanical pumps were used to pump molecular beam chambers in Lee’s lab so that the air must have been polluted with exhausted oil vapor.
We decided to build a dedicated picosecond laser-OP system for SFVS. Without funds for equipment, we had to build everything from scratch, including the mode-locked Nd:YAG laser, the power supply, and the OP oscillator and amplifier designed to be tunable from 2 to 4 μm. Jeffrey Hunt (now a senior technical fellow at Boeing) was charged with the project. We were fortunate to get expert help from two visiting Chinese scholars, Huanan Zhu and Zuhe Yu, from the Institute of Physics in Beijing. In a year or so, the system was running and ready to be used for exploration of SFVS.
That we could have the pleasure of hosting Huanan and Zuhe could be traced back to my visit to China in the summer of 1972. Soon after President Nixon opened the door to China, I was invited to join the first Chinese American Delegation of Scientists and Academicians led by Chi-Kung Jen of John Hopkins University and Chia-Chao Lin of MIT to visit China. Most of the members of our delegation were highly accomplished senior scientists; many prominent academicians in China were their old friends. The visit was extremely exciting. I had the unique opportunity to meet most of the leading Chinese scientists at the time. It was also a unique experience for me to witness how the Chinese scholars were feverishly aspired to devour new knowledge. Later, in 1978, I spent a three-month sabbatical leave at the Institute of Physics in Beijing, allowing me to establish warm friendship with many scientists of the younger generation. When scientific exchange between the US and China started in 1980, our lab was among the first to receive Chinese scholars. Besides Huanan and Zuhe, we also had Pexian Ye from the Institute of Physics come to help us on theory and Nanmin Zhao from Tsinghua University on liquid crystals. Guozeng Yang, who later became the director of the Institute of Physics, visited us for three months and wrote an impactful paper on the theory of laser super-continuum generation. Since then, we had the pleasure of continuously hosting visiting scholars from China until my retirement. Among them were Lei Xu and Xiaofeng Jin, academic leaders from Fudan University.
Our home-made laser-OP system required daily maintenance, and everyone working with the system hated it. Nevertheless, we relied on it to successfully carry out experiments that demonstrated the unique capabilities and applications of SFVS to surface science in different fields. For example, Hunt and Guyot-Sionnest showed that SFVS could detect adsorbed molecules with their specific orientations and conformations at air/solid and liquid/solid interfaces and characterize 2D phase transitions of Langmuir films. Superfine and Jung Huang recorded the first surface vibrational spectra of a neat liquid; they also reported the first phase sensitive SFVS measurement. These experiments firmly established SFVS as a powerful and versatile surface spectroscopic tool for any surfaces and interfaces accessible by light. Several years later, Ekspla Co. in Lithuania made their picosecond SFVS system commercially available. At a laser science conference, I visited the exhibition and saw the system displayed by Ekspla. I commented to the person attending the booth: “Your OP system looks very much the same as our home-made unit.” He answered: “Of course, we copied from you.” But undeniably, their laser system was much more stable than our home-made laser. We eventually had to purchase a laser from them.
People often asked how I decided to write the book on The Principle of Nonlinear Optics. Starting from the birth of nonlinear optics in 1961, there had been a lot of exciting development in the field during the 1960s and 1970s, and, fortunately, I was able to follow most of the developments. I created a graduate course on modern optics at Berkeley. Because there was no suitable book for the course, I had to prepare my own lecture notes. After teaching the course many times, I naturally wondered if I should compile the lecture notes into a book, knowing that no book on nonlinear optics yet existed other than the one by Bloembergen published in 1965. I asked Bloembergen if he cared to join me writing the book, but he kindly refused and said I should go ahead by myself. It took me several years to finally complete the book draft. After it was published, I discovered quite a few mistakes and typos in it. I had them marked, preparing to have them corrected in a future revision. However, in subsequent years when I focused more on surface SHG and SFVS, nonlinear optics had kept growing fast and branching out in many new directions. I soon realized that it would no longer be possible for me to keep track of all the interesting developments and new applications of nonlinear optics while still pursuing research in surface science. I had to give up the idea of having a new, revised edition of the book.
On further exploring applications of surface SFVS, we found seemingly endless cases where SFVS could yield information about surfaces and interfaces no other surface probes could. We were lucky to have continuously a string of unusually talented and motivated students and visitors joining us over many years: Notably, Quan Du and Eric Freysz (visiting from France) recorded the first vibrational spectra of water/air and water/silica interfaces; Rodney Chin observed the first surface phonon spectra of solids; Paulo Miranda (now at U. Sao Paulo, Brazil) discovered that the long hydrocarbon chains of a surfactant monolayer at solid/liquid interfaces could have very different conformations depending on the liquid; Xing Wei (an exceptional all-round talent) demonstrated that mechanical rubbing of a polymer surface could align polymer chains at the surface along the rubbing direction. He also produced the first vibrational spectra of ice surfaces and observed the surface melting of ice starting from the libration of dangling OH bonds at the surface. Our group also noted an interesting episode. Xiaowei Zhuang went to Stanford in 1996 to do postdoctoral work with Steven Chu; Chu won the Nobel Prize the year after. Paulo Miranda went to UC Santa Barbara in 1997 to work with Alan Heeger; Heeger won the Nobel Prize the year after. So, when Xin Wei was ready to get his Ph.D. in 1998, we all wondered where he would choose to go for his postdoctoral work. It turned out that although Xin is a superb scientist, he has little ambition. When Bell Labs came to recruit, he happily accepted the staff position offer without hesitation.
SFVS apparently appealed more to Somorjai than SHG and our collaborations with Somorjai’s group were intensified. Dana Zhang, Xingcai Su, Paul Cremer, Ken Chou, Zhan Chen, David Gracias, and Steve Baldelli from Somorjai’s group collaborated with us closely; in particular, Cremer, Chou, Chen, and Baldelli would come over to participate in our weekly group seminars and discuss problems. Our collaborations focused mainly on SFVS studies of catalytic reactions in real atmosphere and polymer surfaces and interfaces in different environments. The works clearly demonstrated that SFVS is a unique and effective tool for surface studies in these two areas. A special piece of work of great physics interest was the discovery of surface-induced ferroelectric ice by Xingcai Su. As the interest of Somorjai’s group in SFVS grew, I was amazed to learn that in just a few years, their group already had more SFVS setups and more people working on SFVS than our group. Perhaps because physicists and chemists often have different emphasis in research, the newer members of Somorjai’s group stopped coming over to discuss and our collaboration ended.
Around the turn of the century, the NSF Center of Advanced Materials for Purification of Water with Systems (CAMPWS) was founded at the University of Illinois in Champaign. Presumably because our work demonstrating SFVS as the only viable spectroscopic technique to probe water interfaces was being noticed, I was invited by Mark Shannon, the director of the Center, to join the Center. This provided us an opportunity to renew our SFVS studies on water interfaces. In the subsequent ten years with the Center, I was again fortunate to have a group of very capable and dedicated students and visitors working on water interfaces and related problems: Na Ji (now in neuroscience at UC Berkeley), Weitao Liu (now at Fudan Univ., Shanghai), Luning Zhang (now at Tongji Univ., Shanghai), Victor Ostroverkhov (now at GE Research Lab), Chuanshan Tian (now at Fudan Univ.), Jaeho Sung (from South Korea), and Glenn Waychunas (collaborator from Lawrence Berkeley Lab). We obtained the first phase-sensitive SF vibrational spectra of water interfaces, from which more detailed information on structures of water interfaces were deduced. During this period, SFVS studies of water interfaces became a hot topic and attracted worldwide theoretical and experimental research effort. The activities have not yet abated.
In 2000, learning from Janice Hicks’ work on surface SHG studies of adsorbed chiral molecules on substrates, we realized that SFG could be a more sensitive tool to probe chiral media than linear optical rotatory power and circular dichroism. Luckily, I had a superb Russian student, Mikhail Belkin (now at the Technical U. of Munich) with us. He picked up the project and with help from Timothy Kulakov (a postdoc from Russia) and Karla Ernst (visiting from Switzerland), and later Na Ji and Song-Hee Kim (from South Korea), soon succeeded in showing that the technique indeed had very high sensitivity as expected; even detection of chiral vibrational spectra of a monolayer was possible if SF resonant enhancement was imposed. Belkin also worked out the theory of chiral SFG in collaboration with Christos Flytzanis (at Ecole Normale, Paris) and Robert Harris (at UC Berkeley). Na Ji later conducted theory and experiment on chirality of amino acids induced by chiral centers. She, together with Kai Zhang from Hao Yang’s group (Chemistry, UC Berkeley), also demonstrated SF chiral microscopy and took chiral images of a HeLa cell.
I retired in 2005. Around that time, I still had an active research group of students, postdocs, and visitors working on a diverse range of topics. Other than works mentioned above, we had Thai Troung on liquid crystals, Weitao Liu and Luning Zhang on liquid interfaces, Raschke (from Germany) on solid surfaces, Seok-Cheol Hong, Chao-Yuan Chen (from Taiwan), Masahito Oh-e (from Japan), Francois Lagugne-Labarthet (from France), and Pasquale Pagliusi (from Italy) on chirality and surface modification of polymers, John McGuire (now at Shanghai Tech Univ.) on ultrafast surface dynamics of water interfaces, and David Cho and Evgenia Kim (from Russia) on metamaterials. In particular, I should thank Tony Heinz for sending his best student, Feng Wang, to us as a Miller postdoctoral fellow in 2005. Feng brought with him his expertise on carbon nanotubes, graphene, and other 2D materials. I had great pleasure learning from and working with him and witnessing his rapid growth as a scientist. After two years, he accepted the offer to join our physics faculty. I then gladly transferred my laboratory to him and became a super-postdoc in Feng Wang’s group. Feng has done extremely well at Berkeley and has become a prominent world leader in physics of 2D materials.
Without the burden of taking care of a research group, I had more freedom to travel around. I accepted invitations to serve as a visiting professor in Taiwan, Hong Kong, China, and South Korea. I signed a long-term visiting professor contract with Fudan University in Shanghai. I collaborated with Weitao Liu and Chuanshan Tian who had gone back to Fudan to establish their own research groups. I wrote two books on SHG and SFG (one being effectively a new edition of the other), intending them to serve as standard references on the spectroscopy techniques and to symbolize the near ending of my professional life. Recently, Andy Kung and I helped Yudan Su and Jiaming Le in Tian’s group try out successfully a new scheme to generate continuously tunable, femtosecond THz pulses from diamond that could be used to extend SFVS to the THz range. I also worked with Yuxuan Wei, a student from Tian’s group, to find a solution for the long-standing surface local field problem; the paper is published as a contribution to this special issue here. I anticipate that there will be more advances of SFVS in both technical aspects and scientific applications. In particular, I believe that the technique could find a wide range of applications in bioscience although I am not capable myself of pursuing in that direction. I also trust that the technique could find many practical applications in industry. Unfortunately, the current setups of SF spectroscopy are too complex for non-experts. For many workers, the setups are one beam too many (borrowing from what Art Schawlow used to say: “diatomic molecules are one atom too many”). Hopefully, the situation will change in the future when fiber lasers can be used as the pump source so that the SF spectroscopy setup can be turned into a small, movable, turn-key system.
Looking back, I feel that I have had a most pleasant, satisfying, and fulfilling life. I am not an ambitious person. I have never had a large research group and a large research budget I need to worry about. Thanks to the support of the Materials Science Division of the Lawrence Berkeley National Laboratory, I had enough freedom to choose interesting projects to work on. The support of the National Science Foundation should also be acknowledged; though smaller, it allowed me to pursue research areas outside the interest of the MSD. I have been extremely fortunate to always have good teachers and friends by my side. I also have the luck of always having a group of excellent researchers to work with. I am deeply indebted to them; without their help, I would not have any real success. Many of them are now very successful in their own careers. They have become my lifelong friends. We have had a group reunion every five years since 2005.
I should acknowledge our fruitful collaborations with many researchers outside Berkeley: Christos Flytzanis (France), Ching Fong (UC Davis), Michitoshi Hayashi (Taiwan), Sheng Lin (Taiwan), and Ping Sheng (Hong Kong) on various theoretical problems; Mark Shannon and David Cahill (UIUC) on water/solid interfaces; Vladimir Agranovich (Russia) and Xiang Zhang (currently, Hong Kong) on metamaterials; Stephen Cheng (U. Akron) on polymers; H. Seki and Tung Chuang (IBM, Almaden) on diamond surfaces; Valeri Krongauz (Israel) on quasi-liquid crystals; Gerd Marowsky (Germany) and Mahnwon Kim (Exxon, New Jersey) on adsorbed monolayers; Vesse Petrova-Koch on porous silicon; and Kohe Uosaki (Japan) on electrochemical interfaces. I would like to express my deep appreciation to many of my good friends for hosting me and my family on some most enjoyable occasions: Herbert Walther on my two half-year sabbaticals at the Max Planck Institute in Quantum Optics in Garching, Germany; Arnold Schmidt on my half-year sabbatical at the Technical University of Vienna in Vienna; Doseok Kim on my repeated visits over a year to Sogang University in Seoul, South Korea; Guo-Zeng Yang, Xiaofeng Jin, and Jie Zhang on my numerous visits to China; Chia-wei Woo, Michael Loy, Kok-Wei Cheah, and Michel van Hove on my visits to Hong Kong; and Hai-Lung Dai for his kindness, generosity, and help in many circumstances. Last, but not least, I must express my utmost sincere appreciation to my wife, Hsiaolin Shen; without her taking care of all the details at home, including solving problems relating to modern electronic devices, I certainly would not be able to devote myself as much to academic works and my success would hardly be possible. She also helps keep communications with our past group members going. She now knows many of them better than I do.
Finally, I would like to thank with all my heart Eric Borguet, Hai-Lund Dai, Tony Heinz, Jer-Lai Kuo, and Wei-Tao Liu, who edited this special issue of the Journal of Chemical Physics, and all authors who contributed to it.
