Generation of 45 femtosecond pulses at 3 μm with a KNbO3 optical parametric amplifier
Introduction
Time-resolved spectroscopy of molecular vibrations requires intense femtosecond pulses of mid-infrared light that have pulse durations shorter than the dynamics of interests, typically <100 fs. While short mid-IR pulses can be generated in free electron lasers [1] and color-center lasers [2], parametric amplification, often coupled with difference frequency generation (DFG), has emerged as the most ideal option for experiments requiring mid-infrared pulses with femtosecond pulse lengths [3], [4], [5], [6], [7], [8]. Optical parametric oscillators (OPO) and amplifiers (OPA) pumped by the output of Ti:sapphire amplifiers can routinely generate sub-picosecond pulses that span the wavelength range of interest for studying molecular vibrations (2–20 μm).
The 3 μm (3300 cm−1) region is of particular interest to spectroscopists due to the absorption of hydride stretches in this region. Not only are the frequencies of O–H, N–H and C–H stretching vibrations sensitive to their molecular surrounding, but also the vibrations are more localized than other functional groups, making them excellent probes of local structure, hydrogen bonding, and molecular dynamics in condensed phase systems. However, generating transform-limited sub-100 fs IR pulses in this spectral region has proven more challenging than other mid-IR wavelengths. This is particularly the case for the microjoule level energies suitable for nonlinear spectroscopy of vibrations with small transition dipole moments. The most common crystal used in the parametric generation of femtosecond infrared pulses is beta barium borate, β-BaB2O4 (BBO), because it supports large bandwidths and has a high damage threshold [9]. When pumped by the 800 nm output of a Ti:sapphire amplifier, parametric amplification in BBO can be used to efficiently generate pulses in the 1.1–2.8 μm near-infrared region [10]. BBO absorbs at longer wavelengths, sharply decreasing its efficiency and preventing its use to directly produce pulses with wavelengths longer than about 3.2 μm. The 3–12 μm region is most commonly accessed by DFG of the signal and idler waves from parametric amplification in BBO in another nonlinear crystal, such as silver gallium sulfide, AgGaS2[3], [4], [8], or gallium selenide, GaSe [6]. The efficiency of the commonly used Type II mixing in AgGaS2 drops precipitously near 3 μm because the steep phase matching curve for the signal beam [11]. These crystals can not be used to directly convert 800 nm pulses to the mid-IR because their insufficient transparency leads to damage by two-photon absorption.
A limited number of crystals are both transparent beyond 3 μm and have a UV band edge that does not permit two-photon absorption of strong 800 nm pulses, including KTP [5], LiNbO3[12], [13], and KTA [14]. Potassium niobate, KNbO3 (KNB), shows the most promise for short mid-IR pulse generation, because its nonlinearity is the largest, allowing for short crystal lengths, and its steep phase-matching curve near 3 μm results in a large acceptance bandwidth [15], [16], [17], [18], [19], [20], [21]. It has the additional advantage of a high damage threshold, allowing its use for the generation of pulses energetic enough to perform nonlinear experiments. Several previous studies have demonstrated the generation of mid-IR pulses in KNB under various conditions, and most notably sub-100 fs pulses were achieved in two OPAs involving a single pass through KNB. Kafka and Watts generated 3–4 μJ pulses near 3 μm, which were reported to have a 50 fs duration but were not well characterized [17]. Gruetzmacher and Scherer [20] generated nJ level pulses, which were thoroughly analyzed by cross-correlating them with the pump, finding a duration under 50 fs.
Building on this earlier work, we have designed and constructed a two-stage BBO/KNB OPA that produces sub-50 fs, 3–4 μJ pulses suitable for performing nonlinear vibrational spectroscopy. This paper provides design and operation details for the OPA, focusing on the selection of the nonlinear crystals and the full characterization of the resulting pulse spectra, time-dependent amplitude and phase profiles, and stability.
Section snippets
Design of the 3 μm OPA
At room temperature, KNB is a negative biaxial material that is optically transparent beyond 4 μm. Its crystals are orthorhombic with point-group symmetry mm2, whose smallest refractive index lies along the two-fold rotation symmetry axis [15], [22]. We use the convention nx < ny < nz and consider only propagation within the principal planes. Type I phase-matched propagation of the appropriate wavelengths can be achieved in the xy and xz planes, but the latter maximizes the nonlinear coefficient d
Compression and characterization of 3 μm pulses
The 3 μm pulses exiting the OPA are not transform-limited and therefore require compression before use in an experiment. It has recently been shown that pulse compression to second order in the optical phase can be achieved by using pairs of optical materials with normal and anomalous dispersion in the mid-IR [26]. This technique is not possible for visible pulses, since the sign of the GVD is the same for all materials in that region of the electromagnetic spectrum. However, the sign of GVD
Conclusion
We have designed and built a two-stage, white light seeded OPA to generate short pulses of 3 μm light. The first stage employs BBO to generate a 1.1 μm seed for subsequent parametric amplification of the 3 μm pulse in KNB. By passing the mid-IR beam through materials chosen to eliminate GVD, the pulses are compressed to a duration less than 50 fs. These pulses are ideal for studies of vibrational dynamics on hydride stretching vibrations such as O–H, N–H and C–H.
Acknowledgements
This work was supported by Basic Energy Sciences of the US Department of Energy (DE-FG02-99ER14988), the Laser Research Facility at MIT (NSF CHE-0111370), and the David and Lucile Packard Foundation. J.J.L. thanks the DOD for an NDSEG fellowship.
References (29)
- et al.
Opt. Commun.
(1990) - et al.
Opt. Commun.
(2000) - et al.
IEEE J. Quantum Electron.
(1991) - et al.
Opt. Lett.
(1989) - et al.
Opt. Lett.
(1994) - et al.
J. Opt. Soc. Am. B
(1997) - et al.
J. Opt. Soc. Am. B
(1997) - et al.
J. Opt. Soc. Am. B
(2000) - et al.
J. Opt. A: Pure Appl. Opt.
(2001) - et al.
Opt. Lett.
(2000)
J. Appl. Phys.
Opt. Lett.
Opt. Lett.
Opt. Lett.
Cited by (51)
Ultrashort pulse generation in the mid-IR
2015, Progress in Quantum ElectronicsStrong H-bonding from Zeolite Bro̷nsted Acid Site to Water: Origin of the Broad IR Doublet
2023, Journal of Physical Chemistry BAmplification of mid-IR continuum for broadband 2D IR spectroscopy
2023, Optics LettersFrom Networked to Isolated: Observing Water Hydrogen Bonds in Concentrated Electrolytes with Two-Dimensional Infrared Spectroscopy
2022, Journal of Physical Chemistry BResearch progress of mid-infrared rare earth ion-doped fiber lasers at 3 fim
2019, Laser and Optoelectronics ProgressAttosecond light sources in the water window
2018, Journal of Optics (United Kingdom)