-
Dear Editors,
This letter, reflecting on my research career, is dedicated to Professor Qingshi Zhu for his 80th Birthday. Part of this letter is based on my comment “A 20-year journey on the invention of vibrational photothermal microscopy” published in the May 2025 Nature Methods Focus Issue on Bond-Selective Imaging [1].
In a brief summary, my research career is driven by two opposite yet complementary forces. One is having a far-reaching goal of decoding the chemistry of a cell through development of living system spectroscopy. The other is pursing persistent effort and continuous innovation towards that far-reaching goal. Both are derived from my PhD training under Professor Zhu in 1990s, i.e., 取法乎上,得乎其中;聪明在于学习,天才在于积累. Zhu’s generous support of my international visits to Paris and to Hong Kong laid my scientific career and opened my eyes to nonlinear optical spectroscopy techniques. More importantly, the solid training on vibrational spectroscopy in Zhu’s lab laid the ground for my research career on vibrational spectroscopic imaging. What I have achieved today are just some small pieces in the endless journey of uncovering the secret of life, and eventually, the meaning of life. I am grateful for the invitation from the CJCP (Chinese Journal of Chemical Phyics) editors and appreciate this opportunity to express my sincerely gratitude to Professor Qingshi Zhu for his heartful mentoring.
Early days in Bond-Selective Chemistry Lab: Back in 1993, Professor Zhu, then the youngest member of the Chinese Academy of Science, was recruited to University of Science and Technology of China (USTC). That same year, I received the Guo Moruo Scholarship, the highest honor for USTC undergraduates, and later became a national champion in the Mathematical Modeling Competition. As many undergraduates in USTC, I took TOFEL and GRE tests and received ~30 application packages for graduate study in the United States. Meanwhile, I was recommended to Professor Zhu by the Department of Chemical Physics. Fascinated by the dream of bond-selective chemistry, I chose to pursue my PhD at USTC. At the 1994 USTC graduation ceremony, speaking as the student representative, I stated that “The road to success can lie on the other side of the Pacific Ocean, but it can also be found in the motherland”. In September 1994, I joined the Bond-Selective Chemistry Laboratory as the first PhD student Professor Zhu recruited from USTC.
On the first day, Professor Zhu walked me around the eastern campus of USTC and brought me to the statue of Hua Luo-Geng (FIG. 1), a famous mathematician and the first chair of the Department of Mathematics at USTC. Under the statue, it says “Smartness comes from learning; genius comes from accumulation”. This event impacted my whole career. Many years later, I was introduced as a genius of instrumentation by the Chair of Biomedical Engineering when I was recruited to Boston University in 2017. To me, it is very clear that such recognition resulted from years of efforts in learning and accumulation.
Another motto I learned from Professor Zhu is “Aiming high, achieving medium; aiming medium, achieving low”. Many years later, I learned a similar one at Purdue University from a 2011 speech by Neil Armstrong, the first man to walk on the moon. He said: “You should shoot for a star, so that even if you fail, you still can land on the moon”. These mottos have guided my entire career development.
The Bond-Selective Chemistry Lab was friendly and international. In Zhu’s lab, we learned a lot from Russian visiting scientists in the 1990s. With help from Dr. Ulenikov, I performed analyses of AsH3 spectra and found the beautiful transition from normal mode at the fundamental state to local mode at the highly excited state [2]. With help from Dr. Kachanov, Shuiming Hu and I built an intracavity absorption setup and coupled it to a FTIR spectrometer via collaboration with Hai Lin, another PhD student in the lab [3]. This work allowed me to see the power of instrumentation to enable new measurement. Since then, I have spent my whole career building new instruments to enable groundbreaking science.
During my PhD studies, Professor Zhu generously arranged two short visits for me: six months at the Université Paris-Sud, conducting spectral analysis under the supervision of Professor Graner Garner, and four months at the Hongkong University of Science and Technology (HKUST), studying quantum dynamics theory under the supervision of Professor Yijing Yan. I was able to produce two first-author papers during each visit. At Paris, in addition to the research, I spent a lot of time enjoying learning French. At HKUST, I saw much advanced science, prompting me to go abroad after my PhD defense in December 1998. Thanks for Zhu’s mentoring and his generous arrangement of international visits, I had a PhD thesis with nine first-author publications for me to head abroad.
A transition via HKUST: After graduation, Professor Yijing Yan offered me a postdoc position at HKUST. An assistant professor hired by HKUST to develop pump-probe spectroscopy had returned to the United States due to a family matter, which left $400,000 worth of instruments idle. Without prior experience in this area, I was tasked with restarting the experiments. After one year of efforts, I built two systems that measure time-resolved Kerr effect and photoluminescence, respectively. Using these setups, I found an unexpected photoluminescence from C60 epoxide and conducted photoluminescence measurements of nanomaterials, in collaboration with the Shi-He Yang group.
In 1999, Professor Zewail received the Nobel Prize in Chemistry for his pioneering work on femtosecond chemistry using ultrafast pump-probe spectroscopy. Also in 1999, a new era of spectroscopy appeared on the horizon; Sunney Xiaoliang Xie and coworkers published coherent anti-Stokes Raman scattering (CARS) microscopy in Physical Review Letters [4]. CARS could focus all excitation energy on a single Raman band, thus generating a much larger signal than spontaneous Raman scattering. Such large signal allows vibrational spectroscopic imaging of biomolecules in live cells in a label-free or dye-free manner. This imaging method allows scientists to use spectroscopy to “watch” chemistry inside a living system, thus fundamentally changing the scope of spectroscopy-based analysis.
In the same year, Professor Xie visited HKUST for an international conference. I had the fortune of meeting Professor Xie after reading his CARS paper. Yan and Xie were close friends from when they were coworkers at UCSD, and after Xie’s visit, Yan recommended me to Xie for a short visit to Harvard University. This trip, supported by Yan’s research funding, changed my entire career. I must confess that I only knew cars in the street before I read Xie’s CARS paper. Still, I was very excited for this unexpected opportunity because it gave me an opportunity to taste the best science by visiting Harvard. In early 2000, Professor Yan helped me prepare all the documents needed to testify for my visa to visit the United States. Finally, I got a visa for my three-month visit to Harvard.
Developing CARS at Harvard: For my three-month visit from March to June 2000, Professor Xie arranged me to develop CARS microscopy, then a very new direction in his single-molecule fluorescence research group. My job was to build a picosecond-laser based CARS microscope (FIG. 2) for vibrational imaging of live cells with high speed and high spectral resolution. In previous work, Xie and coworkers used a high energy femtosecond OPA system, which generated a large signal, but with poor spectral resolution and low imaging speed. It was a high-risk project since we did not know how much pulse energy was needed to generate CARS signals from one-femtoliter volume at each pixel. Yet I had no other choice but to work it out.
I indeed worked very hard, 15 hours a day and 7 days a week, in both instrumentation and theory. When I arrived at Harvard, the room was empty except for an optical table. Two Ti:Saphire lasers arrived a few days later. With tremendous help from Xie and lab mates, I generated the first CARS signal from one-micron polystyrene particles by the end of May. It was a big relief to Xie after his investment in the very expensive laser system. In the final month, I further achieved a high-quality CARS image from my cheek cells. This preliminary data helped Xie secure a R01 grant on CARS microscopy. The three months spent in Professor Xie’s lab helped me secure a postdoc position at Harvard to continue the development of CARS microscopy.
For me, being postdoc at Harvard was a very productive period. In early 2001, I helped the Xie group publish their first journal article from Harvard [5]. Over the next two years, I published 9 first author papers and 6 co-author papers on CARS microscopy. These studies established instrumentation strategies, such as picosecond laser excitation and epi-detection. In parallel, my Green’s Function model built the foundation for CARS contrast mechanisms.
Crossing the discipline from science to engineering: My achievements in CARS microscopy put me on the job market with 10 on-site interviews in fall 2002. Finally, I chose Purdue University to start my own independent career in the fall of 2003. Back then, Purdue University had a cluster hire, in which candidates were interviewed before home department assignment. While all my training was in the physical chemistry discipline, imaging was considered a subject of engineering back then. After my interview, it was decided that Weldon School of Biomedical Engineering should be my home department. Joining Biomedical Engineering enabled me to pursue new research dramatically different from my postdoc work. This crossing of disciplines turned out to be a smart switch—since research funding in the US is largely from the National Institute of Health (NIH)—and I quickly learned how to write a R01 proposal for biomedical applications of my work. Moreover, being a faculty at a top College of Engineering gave me the opportunity to recruit elite engineering students to build advanced microscopes in my research lab.
Teaching in College of Engineering was quite challenging for me. I had to learn circuits to teach bio-electronics to undergraduate students. Fortunately, my father-in-law, who lived with us in the United States, was a professor at the Department of Electronics at USTC and helped me grasp circuitry. Undergraduate teaching also gave me the opportunity to apply a resonant circuit for lock-in free multiplex stimulated Raman scattering (SRS) imaging [6]. This idea got me a Science & Engineering Grant from the Keck Foundation, which gave me the resources to build next-generation chemical microscopes.
A new direction of chemical bond imaging inspired by an unexpected observation: Although coherent Raman scattering microscopy allows high speed imaging, its sensitivity is at the millimolar level, which was not sufficient for visualizing low-concentration molecules. An unexpected observation first made in 2005, two years after I joined Purdue University, opened a new way of bond-selective imaging with larger depth and better sensitivity. Back then, my team was investigating the mechanism of photodamage in CARS microscopy. Published in J. Opt. Soc. Am. B in 2007, we reported an unexpectedly FASTER rate of photodamage in samples in the presence of Raman resonance upon a prolonged excitation (FIG. 3). We further found that the enhanced photodamage was not due to CARS, but due to energy deposition through stimulated Raman gain and loss processes that occur simultaneously with CARS [7].
Because energy deposition can induce photothermal (PT) and photoacoustic (PA) effects, this finding triggered us to develop a stimulated Raman photoacoustic microscope with a nanosecond OPO laser, provided by my Purdue colleague Robert Luckt, built with help from Song Hu in Lihong Wang’s research group. We worked incredibly hard for two years, but we found no signal! Accidentally, Han-Wei Wang, the first author of our PRL (Phys. Rev. Lett.) paper [8], blocked the visible beam and recorded a PA signal from the near infrared beam. I immediately realized that we had generated a PA signal from overtone vibrational transitions that produce local-mode vibrations—a central research theme in the Bond-Selective Chemistry lab. This finding opened a new way of in vivo imaging lipids deposited inside a blood vessel through the development of an intravascular catheter. In 2014, my student Pu Wang and I co-founded Vibronix Inc. to translate vibrational imaging technologies toward improved diagnosis and treatment.
Because of this experiment, I came to believe that the Raman effect is not strong enough for PA or PT imaging. Over the next few years, we started to “think out of the Raman box”, i.e., exploring alternative modalities for a higher chance of success. The Raman scattering cross section is on the order of 10−30 cm2/molecule, which limits the detection sensitivity to millimolar concentrations. Outside the box, mid-infrared absorption has a much bigger cross section ~10−20 cm2/molecule.
For certain chemical groups, such as azide, their absorption cross section can be as large as electronic absorption that generates fluorescence. Yet, unlike fluorescence emission, infrared excitation only produces heat through vibrational relaxation on the picosecond time scale. Meanwhile, heat diffusion occurs on the nanosecond to microsecond time scale following Newton’s Law. By using a nanosecond infrared pulse train to repetitively excite a chemical bond, the heat accumulates, causing a transient change in refractive index and thermal expansion of a particle. This photothermal effect can be probed with extremely high sensitivity using optical detection techniques.
At that time, the challenge was that we had no laser source for infrared imaging, as the whole group was developing coherent Raman scattering microscopy. In fall 2014, while attending the SciX meeting, I was fortunate to meet the founders of Block Engineering, who were looking for life science applications of their quantum cascade laser (QCL). I was also fortunate to have a strong team (Delong Zhang, Mikhail Slipchenko, and Chi Zhang) for the initial experiments. We built a photothermal microscope using the loaned QCL (FIG. 4) in the corner of a table dedicated to coherent Raman microscopy. On Jan 18th, 2015, at 2 am, Delong got the first signal. We then demonstrated a high-performance mid-infrared photothermal (MIP) microscope, published in Science Advances in 2016 [9]. For the first time, our method allowed infrared mapping of specific chemical bonds inside a living cell with sub-micron spatial resolution.
Growing the seed into a big tree: In 2017, I was recruited to Boston University as the Inaugural Moustakas Chair Professor of Photonics and Optoelectronics at the Photonics Center, where we have substantially expanded and matured this technology (for a review, [10]) with a strong team of students and postdocs. One key project focused on developing a confocal laser-scanning MIP, in which we synchronously scan mid-IR and photothermal probe laser beams. By a method called single pulse digitization, we realized video-rate imaging speed [11]. Another key development has been wide-field MIP [12]. In this camara-based method, we match the focal size between the infrared laser and the visible probe, so that all the energy can be utilized toward high-throughput measurement. By measuring interferometric scattering, we achieved the sensitivity to chemically fingerprint a single virus in a wide-field manner. We further developed computational MIP via collaboration with my BU colleague Lei Tian. We employed intensity diffraction tomography to reconstruct a three-dimensional refractive-index map, and by modulating the refractive index with infrared pulses, achieved three-dimensional chemical-bond mapping in tissues and C. elegans.
There is a saying: ‘your opponent can be your best friend.’ Though MIP was initially developed as a dye-free technique, it turns out that fluorescence is a much more sensitive readout of the photothermal effect than scattering due to the strong temperature dependence of fluorescence emission efficiency (quantum yield) of many common dyes. Fluorescence-detected MIP microscopy was simultaneously demonstrated by my group [13], Garth Simpson’s group at Purdue [14], and Photothermal Spectroscopy Corp. This method opens a new door of life science applications using fluorescence as both guide and sensor of chemical environments.
In parallel, we have started to incorporate photothermal reporters for cell functional imaging. In the silent window (a spectral region void of biomolecular IR absorption bands), chemical bonds like azide and nitrile give very strong infrared absorption. Importantly, the diameter of a chemical bond is about one angstrom, which is much smaller than a fluorescent probe in terms of volume, and thus can be used to track biomolecules with minimal perturbation to cellular function. In a recent article [15], we designed nitrile chameleons as a substrate to map enzymatic activities inside living cells. For instance, when the enzyme phosphatase removes a phosphate group in the substrate, the peak of the nitrile vibration shifts to a new wavenumber. Thus, we can map both the substrate and the product to monitor biochemistry inside a live cell. In a separate study, we demonstrated click-chemistry-free MIP imaging of carbohydrate trafficking in live cells [16]. Using the strong MIP signal from azide, we were able to monitor cellular uptake of azido-trehalose, recycling to the membrane, and pole accumulation during cell division in real time.
Though we failed in stimulated Raman photoacoustic imaging 15 years ago, we recently showed that the photothermal effect is substantial. In stimulated Raman scattering (SRS) microscopy, one measures either intensity gain on the Stokes beam or intensity loss on the pump beam, but there is an energy difference between the gain and loss. In fact, stimulated Raman can be treated as a two-photon vibrational excitation process, depositing energy into the molecules. Stimulated Raman photothermal microscopy can visualize small features like a single virus particle or a single membrane inside a cell [17]. More recently, we developed short-wave infrared photothermal microscopy that extends the photothermal imaging depth to one millimeter [18].
Entrepreneurship for worldwide use: Back in October 2015, I presented MIP microscopy in my Craver Award speech at the SciX meeting. I was fortunate enough to have Roshan Shetty and Dean Dawson (then CEO and marketing director at Anasys Instruments) in the audience. After the speech, they invited me for coffee and told me that they wanted to convert our technique into a product. In the October 2017 SciX meeting, Anasys Instruments announced mIRage®, an optical photothermal infrared (O-PTIR) microscope. In 2018, Photothermal Spectroscopy Corp was founded to further develop and commercialize optical photothermal microscopy. FIG. 5 shows a picture of part of the leadership team taken outside the company in Santa Barbara. None of us were born in the United States, yet this immigrant team has made a global impact, with over 120 mIRage systems delivered to top universities, including Caltech, Stanford, Yale, EPFL, Lund University, and Vienna University, and to over 15 countries, as of January 2025.
This entrepreneur success has enabled broad applications far beyond my own research. For example, the Klementieva group at Lund University is using mIRage to study protein aggregation in Alzheimer’s Disease. In China, Nanjing University used mIRage to detect nanoplastics. The Davis group at Yale is using mIRage to visualize metabolism in live cells. On the industrial side, leading companies are using mIRage to characterize contaminants and defects in semiconductors, data storage, and display applications.
Photothermal spectroscopy appeared in 1970s as an ultrasensitive absorption spectroscopy [19]. In 1983, a French group proposed the concept of photothermal imaging based on the change of refractive index, i.e., the mirage effect [20]. The initial effort of photothermal infrared imaging was independently reported at a SPIE conference proceeding [21]. We were fortunate to enter the field at the right time. With over 10 years of experience in pushing the physical limits of coherent Raman microscopy, we had accumulated all the expertise needed to build a high-performance MIP microscope and demonstrate live-cell imaging with sub-micron spatial resolution. Meanwhile, the strong expertise in AFM-IR at Anasys Instruments and the maturity of QCL allowed rapid commercialization of our method into a real mIRage product for worldwide use.
I never planned to grow an unexpected observation of photodamage back in 2005 into a rising direction of chemical bond imaging [22]. The developments were exclusively driven by curiosity and passion, achieved with hard work and financial support over years. Looking back, my PhD work on local-mode based bond-selective chemistry built a solid foundation for my invention of vibrational photothermal microscopy [23], now used worldwide. Notably, the rapid vibrational relaxation, which challenges bond-selective chemistry in liquid phase, leads to a bright contrast in photothermal microscopy that allows bond-selective imaging. This Yin and Yang journey took me 30 years, from 1994 to 2025. Science is cool; life is short. I hope that young scientists can find and focus on a big problem when they are young.
Sincerely Yours,
Ji-Xin Cheng
Moustakas Chair Professor of Photonics and Optoelectronics, Boston University Photonics Center, 8 Sanit Mary’s Street, Boston, MA 02215, USA
Bond-Selective Chemistry: a Foundation for Bond-Selective Imaging
- Available Online: 27/02/2026
Abstract:
首页
登录
注册


DownLoad: