Research
Lifetime Science and Technology Achievements, Ursula Keller 2023
SESAM modelocked ion-doped solid-state lasers
Professor Keller invented and demonstrated a key new device - the semiconductor saturable absorber mirror (SESAM) [IEEE JSTQE 2, 435, 1996; Nature 424, 831, 2003] - which demonstrated the first passively mode-locked diode-pumped solid-state laser in 1992 and solved a 25-year-old challenge [Opt. Lett. 17, 505, 1992]. For more than two decades since then, her work has continued to define and push the frontier in ultrafast solid-state lasers both with theoretical models and with world-leading experimental results, demonstrating orders of magnitude improvement in key features such as pulse duration, energy, and repetition rate.
Her group at ETH Zurich could push the ultrafast solid-state laser into new performance levels once a better understand was achieved of the underlying Q-switching problem of passively modelocked solid-state lasers [JOSA B16, 46, 1999], SESAM-assisted soliton modelocking [IEEE JSTQE 2, 540, 1996] and other laser physics as summarized in an invited review article [Appl. Phys. B 100, 15, 2010]. She also pioneered SESAM Q-switched microchip lasers bridging the gap between traditional modelocking and Q-switching with high-energy picosecond pulses [Optics Letters 21, 405, 1996; J. Opt. Soc. Am. B 16, 376, 1999].
She also pioneered industrial transfer of this technology. Today most industrial ultrashort lasers are based on SESAM modelocking, with important industrial applications ranging from optical communication, precision measurements, microscopy, ophthalmology, and micromachining.
In 2022 she published an 800-pages comprehensive graduate textbook “Ultrafast Lasers” in Springer Verlag.
Ultrafast optically pumped semiconductor lasers
SESAM modelocking then was extended to optically pumped semiconductor disk lasers with unsurpassed average output power and at wavelengths which cannot be reached with diode-pumped solid-state lasers. First demonstration of passively modelocked external cavity surface emitting lasers (VECSELs) was done in collaboration with Prof. Tropper in 2000 [IEEE Pot. Tech. Lett. 12, 1135, 2000 and Phys. Rep. 429, 67, 2006]. A performance highlight is a femtosecond VECSEL with more than 1 W average output power [Opt. Exp. 19, 8108, 2011]. She invented a new class of semiconductor lasers with wafer-scale integration of both the gain and the absorber into a vertical emitting structure, allowing the technology to potentially scale into high-volume markets – referred to as the MIXSEL (Modelocked Integrated eXternal-cavity Surface Emitting Laser) [Appl. Phys. B 88, 493, 2007]. An optically pumped MIXSEL has demonstrated picosecond pulses with more than 6 W average power [Opt. Exp. 18, 19201, 2010], the pulse repetition rate could be scaled between 5 and 100 GHz [Opt. Exp. 22, 6099, 2014] and such lasers can have record-low noise performance very similar to diode-pumped solid-state lasers [IEEE Photonics Journal 6, 1500309, 2014]. More recently a femtosecond GaSb-based VECSEL has been demonstrated in the 2-µm regime [IEEE Photonics Tech. Lett. 34, 337, 2022] and the first 2-µm MIXSEL just has been demonstrated in her lab.
Few-cycle pulse generation, the carrier envelope offset and optical frequency combs
Her cutting-edge laser physics research also enabled the world’s most accurate clocks by exploring new frontiers and pushing the laser pulse duration into the one to two optical-cycle regime. Prof. Keller provided significant first experimental evidence for identifying the modelocking mechanism as “Kerr lens modelocking” using a nonlinear coupled cavity to reliably start modelocking [Opt. Lett. 16, 1022, 1991]. The better understanding allowed her group to move into the two-optical-cycle regime for the first time [Opt. Lett. 24, 631, 1999 and Science 286, 1507, 1999] with world record results.
The regime of few-cycle pulse generation confronted the ultrafast community with a new problem because the electric field within the pulse envelope was not stabilized in modelocked lasers. She defined the offset of the peak of the electric field to the peak of the pulse envelope as the carrier envelope offset phase – typically referred to as the CEO phase or even shorter as the CEP. The problem was that this CEO phase was very noisy. The solution to this problem was considered the “holy grail in ultrafast pulse generation”. Prof. Keller’s group in collaboration with Dr. Telle (PTB, Braunschweig) solved that problem in 1999 [Appl. Phys. B 69, 327, 1999] which enabled attosecond pulse generation and started the frequency comb metrology revolution (Nobel Prize in physics 2005). Her pioneering contribution to this field was also recognized by the Nobel committee for the Physics Nobel Prize in 2005 (see page 11, Ref. 37, "Advanced Information" for the Nobel Prize in Physics 2005, Link) and in many awards such as for example in the IEEE Photonics Award 2018. She and her group discussed the physical origin of the CEP fluctuations and optimized the laser cavity [IEEE JSTQE 9, 1030, 2003].
Optical frequency combs have a major impact in many new research areas ranging from precision frequency metrology, optical clocks to high field physics and attosecond pulse generation. Prof. Keller’s group continues to demonstrate compact frequency combs in the at 1.5 µm [Opt. Exp. 19, 24171, 2011], at 1 µm and gigahertz repetition rates from diode-pumped Yb-doped solid-state lasers [Optics Express 24, 11043, 2016; Nature Comm. 8, 1673, 2017] and in the mid-infrared regime with nonlinear frequency conversion [Phys. Rev. Applied 6, 054009, 2016]. Currently the lowest noise gigahertz frequency comb on the market is based on her work on SESAM-modelocked Er:Yb:glass lasers [available from Menhir Photonics].
Single-cavity dual-comb modelocking
Dual-comb laser sources generate two optical frequency combs with different comb spacings and have become a hot topic for many applications. Prof. Keller’s group invented two breakthrough methods which enables single-cavity dual-comb source generation based on polarization [Optics Express 23, 5521, 2015] and spatial [Optica 9, 713, 2022] multiplexing with an adjustable pulse repetition rate difference. The simple straight cavity of a MIXSEL inspired the first invention of polarization duplexing and the demonstration of successful dual-comb molecular spectroscopy using only one unstabilized semiconductor laser (i.e. dual-comb MIXSEL) in water [Science, 356, 1164, 2017] and acetylene [Opt. Express 27, 3190, 2019] and can be considered a paradigm shift for industrial frequency metrology applications.
A second approach was invented based on spatial multiplexing by inserting a monolithic device with two separate angles on the surface, such as a biprism [Optica 9, 713, 2022], which also can be used in transmission for shorter cavities and gigahertz pulse repetition rates [Optics Express 31, 7103, 2023] First demonstrated with a 80-MHz Yb-doped diode-pumped solid-state laser generating sub-140 fs and 2.4 W of power per comb with record-low sub-cycle relative timing jitter for an integration interval of [20 Hz, 100 kHz] using a new noise characterization technique [Optics Express 30, 5075, 2022].
For optical comb line resolved measurements a pulse repetition rate in the few gigahertz regime is ideal and most recently a 1-GHz single-cavity dual-comb laser with sub-80-fs pulses and 3-W average power per comb was demonstrated, which allows for coherent averaging the full interferometric dual-comb signal on long time scales [Optics Express 31, 7103, 2023]. The pulse repetition rate difference can be continuously tuned up to 27 kHz. No active stabilization is required even if there is some noise because of the coherent averaging.
Many proof-of-principle dual-comb applications with these free-running dual-comb lasers have been demonstrated by our group: optical spectroscopy in the near infrared [Science, 356, 1164, 2017; Optics Express 27, 3190, 2019] and at longer wavelength [Optics Express 30, 199904, 2022; Optics Express 31, 6475, 2023], equivalent time sampling [Appl. Phys. B, 128, 24, 2022], picosecond ultrasonics [Optics Express 29, 35745, 2021; Photoacoustics 29, 100439, 2023], THz spectroscopy [arXiv 2302.10526, 2023] and lidar [Optics Express 30, 39691, 2022]. Work is in progress to commercialize these lasers in a new spin-off company (K2 Photonics).
Invention of the attoclock technique
The frontier laser research with the CEO phase stabilization enabled Prof. Keller to invent the attoclock [Science 322, 1525, 2008], a powerful and unconventional tool to study fundamental processes in quantum mechanics – with attosecond accuracy using 1000-times longer laser pulses. She established the attoclock to measure the electron tunneling time for the first time which is a highly debated topic in theoretical physics for the last 60 years. Recent attoclock experimental measurements [Optica 1, 343, 2014] found a finite tunneling time over a wide intensity range and therefore a large variation of tunnel barrier width. Only two theoretical predictions are compatible within experimental error: the Larmor time, and the peak of the probability distribution of tunneling times constructed using a Feynman Path Integral (FPI) formulation. The FPI theory matches the observed qualitative change in tunneling time over a wide intensity range, and predicts a broad tunneling time distribution with a long tail. A more recent review confirmed this result taking into account nonadiabatic effects and making sure all other approximations are valid [Journal of Modern Optics 66, 1052, 2019]. The attoclock was also used to reveal the tunnel geometry [Nat. Phys. 8, 76, 2012], the timing of each photoelectron in double ionization of argon [Nat. Phys. 7, 428, 2011], and how much momentum is transferred to the photoelectron in the direction of laser propagation with sub-femtosecond resolution at high laser intensities, where multiple photons are involved in the ionization process. This has been an unresolved question [Nature Communication 10, 5548, 2019].
Attosecond photoemission time delays and ionization dynamics beyond the dipole approximation
The group implemented coincidence detection [Phys. Rev. Lett. 115, 133001, 2015; IEEE JSTQE 21, 8700307, 2015] and angular dependence [Phys. Rev. A 94, 063409, 2016] of attosecond photoemission time delays for the first time, which revealed that a more complex continuum with interfering quantum paths can strongly affect this ionization time delay even in atoms [Phys. Rev. Lett. 115, 133001, 2015 and Nature Comm. 9, 955, 2018] and molecules [Nature Phys. 14, 733, 2018 and Science 360, 1326, 2018]. Her group also explored how long it takes for an excited electron in a metal to “feel” its effective mass [Optica 4, 1492, 2017].
Her group also made fundamental contributions to a better understanding of how the breakdown of the electric dipole approximation is affecting ionization dynamics in strong field ionization, as summarized in an invited review article [J. Phys. B :At. Mol. Opt. Phys. 54, 094001 2021]. In the regime of strong field ionization for example the maximum of the photoelectron distribution is shifted opposite to the laser beam propagation direction, which is counter-intuitive within the framework of the radiation pressure [Phys. Rev. Lett. 113, 243001, 2014]. More detailed studies as a function of ellipticities [Phys. Rev. A 97, 013404, 2018; J. Phys. B: At. Mol. Opt. Phys. 51, 114001, 2018] also shows how the attoclock measurement [Journal of Modern Optics 66, 1052, 2019; Nature Communication 10, 5548, 2019] is affected.
PHz electronics with attosecond transient absorption spectroscopy (ATAS)
The frontier laser research with the CEO phase stabilization also enables petahertz electronics. In the 1980s, rapid progress in picosecond and femtosecond ultrafast lasers started to bridge the gap between electronics and optics with terahertz frequencies. With the recent progress in few-cycle femtosecond and attosecond pulse generation with full electric field control, the frequency regime can be extended into the petahertz regime for the investigation of ever faster physical processes and device performance. Attosecond transient absorption spectroscopy (ATAS) allows for element and carrier specific probe of excited states dynamics when an attosecond XUV pulse is used to probe the transition between core and excited states. With the probe being resonant with a characteristic core level, ATAS becomes inherently element specific while its broad bandwidth simultaneously reveals the dynamics of both, pump-excited electrons and holes with attosecond time resolution.
Her group performed pioneering attosecond field-induced dynamics in diamond [Science 353, 916, 2016] and GaAs [Nature Phys. 14, 560, 2018]. Surprisingly for the resonant excitation in GaAs, they found that the transient response is still dominated by intra-band motion. They revealed a new ultrafast response around transition metal atoms based on real-space electron localization into 3d-orbitals following the femtosecond excitation pulse envelope in elementary transition metals [Nature Phys. 15, 1145, 2019], which is not present in Al [Phys. Rev. X 12, 021045, 2022] and only happens around molybdenum in a 2D semiconductor MoSe2 (a transition metal dichalcogenide, TMDC). Only the spectral response of the transition metal Mo is dominated by screening effects with ultrafast d-orbital localization, while the selenium is not affected by this collective response and can be described by an independent particle model with band filling, thermalization and lattice heating [PNAS 120, e2221725120, 2023]. This can be considered surprising given that the near-infrared pump excited carriers are well above the Mott-transition and valence electrons from both Mo and Se contribute to the conduction band. This means the overall dynamics in such materials with transition metal elements cannot be described with an independent particle model. This may have implications for the applicability of ubiquitous effective mass approximation in such semiconductors.