Open Access
Add to BibSonomyAbstract
A two-stage optical parametric amplifier generating 5 ns 208 kW peak power pulses in the spectral region at 1.535 µm in a diffraction-limited beam was realized in a single periodically poled KTP crystal. The maximum small-signal gain for the two stages reached 75dB and the total conversion efficiency was 30%. An analysis of the small-signal gain dependence on the M2 of the pump beam is presented for the collinear and noncollinear OPA. Efficient spectral broadening of the signal was demonstrated in short pieces of single-mode telecommunication fiber.
©2003 Optical Society of America
1. Introduction
High peak power nanosecond pulses in the eye-safe region around 1.55 µm are needed for several different applications, including LIDAR, spectroscopy, range finding and also in biomedicine, due to high absorption in tissue at these wavelengths. Flash-lamp pumped Q-switched Er:glass lasers or laser diode pumped actively Q-switched Er-Yb:glass lasers can be used to generate mJ pulses, but only at low repetition rates because of poor thermal properties of the gain medium, which make it difficult to maintain fundamental transversal mode operation [1]. Optical parametric oscillators (OPO), especially the ones employing noncritically phase-matched KTiOPO4 (KTP) or KTiOAsO4, can generate mJ nanosecond pulses in the eye-safe spectral region with high efficiency and high repetition rates [2–4]. However, the spatial and spectral quality of the OPO signal is essentially determined by those of the pump, which requires the use of injection seeded Q-switched lasers and additional spectrally selective components in the OPO cavity [5].
Phase-sensitive amplification in a travelling wave optical parametric amplifier (OPA) can arguably provide a simpler route for generating mJ-level nanosecond pulses in the eye-safe and mid-infrared spectral regions with the spatial mode properties which can be tailored by using an appropriate seed beam. This property together with the possibility of noise reduction which OPA provides can be used for optical image amplification [6]. Birefringence phase-matched nonlinear crystals are not very suitable for mid-infrared OPA using nanosecond pumping at 1064 nm due to the low effective nonlinearity and/or Pointing vector walk-off. Quasi-phase-matched (QPM) nonlinear crystals eliminate Pointing vector walk-off and have typically much higher effective nonlinearities, which make them the materials of choice in nanosecond OPA pumped by near-infrared Q-switched lasers [7–9].
In this work we demonstrate a double-pass nanosecond OPA, realized in periodically poled KTP (PPKTP) and seeded with a compact, diode-pumped actively Q-switched Er-Yb:glass laser. The main goal was the investigation of the PPKTP as a gain medium for mJ-level nanosecond OPA as well as design of high-gain OPA configuration which could generate diffraction-limited signal beam without spectral broadening produced by the parametric superfluorescence. The peak power of the diffraction limited narrowband OPA signal, in excess of 200 kW was generated demonstrating small signal gain of 75 dB. At the same time, the generated peak power was sufficiently large for strong spectral broadening in short pieces of single-mode telecommunication and dispersion-flattened optical fibers, thus providing a broadband source, for instance, for spectroscopy applications.
The high nonlinearity in PPKTP, although indispensable for efficient OPA, at the same time, decreases the parametric fluorescence threshold. The amplified fluorescence has to be avoided in order to preserve the spectral content and the spatial coherence of the amplified signal. In order to reach higher signal output powers, we have designed a two-stage OPA realized by two consecutive passes through the same PPKTP crystal. The first stage provided a high-gain superfluorescence-free amplification, while the second stage was operated at much lower saturated gain but at substantially higher pump energies and played the role of a power amplifier.
2. Experiment
The experimental setup is shown in Fig. 1. The OPA was seeded by a home-built actively Q-switched Er-Yb:glass laser, generating 36 ns long 6 µJ pulses in a nearly diffraction limited beam (M2=1.1) at 1535 mn [10]. The OPA was pumped by a commercial, actively Q-switched Nd:YAG-laser, generating 5 ns long 30 mJ pulses in an essentially non-diffraction limited beam (M2=7). A high precision delay-generator, triggered from a separate pulse generator, synchronized the pump with the seed. The system was operated at 20 Hz, the limiting frequency of the Nd:YAG. It should be noted that the seed laser, with minor adjustments, could be operated at repetition rates of up to 2 kHz while keeping very similar pulse length and energy [10]. The relative pulse-to-pulse jitter between the seed and the pump was 5 ns, primarily caused by the uncertainty in the pulse build up time of the Er-Yb:glass laser.
The OPA was realized by a 15- mm-long and 1 -mm-thick PPKTP crystal with a domain inversion period of 35.4 µm, resulting in a parametric gain maximum at 1535 nm for the 1064 nm pump wavelength. The effective nonlinearity, deff=9.2 pm/V, was extracted from a separately measured OPO threshold dependence on the cavity length using the method described in Ref. [11]. The optical surfaces were polished at an angle of 3° relative to each other, in order to avoid parametric oscillation, and were left uncoated. The pump was split with a ratio of 30/70 into two beams by beamsplitter BS1. The weaker beam was focused into the PPKTP crystal using lens L1 and provided energy for the first-stage OPA. The dichroic mirror BS2 was used to combine the pump beam with the seed beams with a good spatial overlap. Both the pump and the seed had beam waist radii of 160 µm in PPKTP. After the first OPA stage the generated signal is separated from the remaining pump by the dichroic mirror BS3. The signal is routed through a delay line containing lens L6 and a roof-top retro-reflector. This arrangement performed as a slightly misaligned telescope and was adjusted to achieve mode matching between the retro-reflected signal and the second stage pump. With a total signal loss in the return path of 21%, the remaining power was large enough to operate the second OPA stage in a saturated power amplifier mode. The second OPA stage was realized in the same crystal, but in opposite direction to that of the first stage. The second OPA stage employed the largest portion of the initial pump routed by the mirrors M1 and M2. The input signal and beam waist radii in the second OPA stage were 275 µm. The relative angle between the beams in the first and the second OPA stage was 1.1° inside the crystal, allowing for easy separation of the OPA outputs.
3. Results and discussion
The parametric fluorescence threshold in PPKTP without input seed was reached at a pump peak intensity of 252 MW/cm2. A maximum signal energy of 116 µJ was generated for the first OPA stage with a maximum seed energy (peak power of 153 W) and a maximum pump energy of 1.19 mJ. The resulting signal conversion efficiency was then 10%, and by taking the idler into account, the total conversion efficiency reached 14%, which is close to the measured pump depletion of 15%. The pump energy could not be increased further in the first OPA stage due to the onset of parametric fluorescence. The pulse-length of the generated signal was 4 ns or nine-times shorter than the seed pulse. The dependence of the first-stage OPA gain on the seed peak power measured at maximum pump energy (shown in Fig. 2 by the data points) varied between 55.2 dB and 22dB for the seed peak powers between 0.7mW (3.5 pJ) and 145W (725 nJ).
It is well known, that phase sensitive amplification in OPA reproduces the temporal and spatial phase of the seed in the output signal, while the phase modulation which is present in the pump is compensated for by the idler beam. However, for pump beams with large M2 and relatively tight focusing, the OPA gain can be reduced due to limitations of the angular phase-matching bandwidth. Calculations presented in Ref. [12] for birefringence phase-matched crystals showed substantial decrease of the gain due to angular bandwidth limitations in the critically phase-matched case. QPM crystals employing the nonlinear coefficient d33 are noncritically phase matched in the traditional sense, i.e., the refractive indices do not depend on the angle in the x-y plane. However, in the case of noncollinear QPM OPA the angular sensitivity of the gain to the deviation of the pump angle θp relative to the signal propagation direction, ∂|Δk|/∂θp , can be significantly larger than in the collinear case. Here |Δk|=|k p -k s -k i -k g | and k j ,j=p,s,i,g, are the wavevectors for the pump, signal, idler and QPM grating, respectively. Indeed, when the phase matching is achieved for collinear pump and signal propagation, then ∂|Δk|/∂θp =0 and only second derivative ∂2|Δk|/∂ contributes to the angular sensitivity of the gain. Now if the phase-matching is achieved for noncollinear signal and pump propagation, ∂|Δk|/∂θp ≠0 then the angular gain bandwidth can be significantly narrowed, which is tantamount to a net gain decrease for pump beams with large M2 values. We calculated the small signal gain using an approach similar to the one in Ref. [12]. The small signal power gain can be expressed as
Fig. 2. Measured single-pass peak power gain in the first OPA stage as a function of the seed power (data points). Calculated small-signal gain in the PPKTP OPA as a function of the pump M2 for different QPM periods (solid lines); 1–35.4 µm, 2–35.34µm, 3–35.2µm.
where
L is the crystal length, Ip - pump intensity, nj - the refractive indices and λs,p -signal and pump wavelengths. For small deviations of the pump angle from the mean direction of the beam propagation, the phase mismatch in the noncollinear OPA,
while in the collinear OPA,
Fig. 3. Two-stage OPA signal (solid squares) and idler (open circles) energy as a function of the second stage pump.
where M 2 is that of the pump beam, which is focused to a beam waist radius w. The phase-mismatch given by the expressions above is attributed to the Fourier component of the pump beam with the largest spatial frequency. The solid lines in Fig. 2 show the M2 dependence of the PPKTP OPA small-signal gain, integrated over the spatial spectrum of the pump for three different QPM periodicities and assuming the pumping conditions of the experiment. For a collinear OPA (curve 1) the gain is essentially insensitive to the M2 of the pump (maximum gain decrease of 0.3 dB for M2=7). However, for shorter QPM periods (35.34 µm-curve 2, and 35.2 µm-curve 3) phase-matching is achieved for noncollinear signal and pump propagation (0.3 and 0.5 degrees, respectively) and the small-signal gain dependence on M2 is much stronger (gain suppression of 3.2 dB and 9 dB obtained at M2=7). The theoretical gain of 57.3 dB in a 15 mm-long collinear PPKTP OPA corresponds quite well to the experimental small-signal gain of 55.2 dB. The small gain sensitivity to the M2 value of the pump beam for the collinear pump and signal propagation can be easily understood if we compare the OPA pump angular acceptance bandwidth of 11.2 mrad with respect to the propagation direction of the diffraction-limited signal, to the pump divergence of 8.2 mrad in the PPKTP crystal, considering the pump beam waist radius of 160 µm and M2=7.
The second OPA stage, was pumped at a maximum energy of 4 mJ, the limit, again, determined by the appearance of parametric superfluorescence. Although the pump beam size could still be increased slightly in order to reach the second stage operation, limited by the optical damage threshold, we chose to operate the OPA at the pump intensity no exceeding 400 MW/cm2. For the seed peak power above 2 W injected into the first OPA stage, the second OPA operated in a saturated gain mode and boosted the signal peak power by an order of magnitude. For the lowest seed power of 0.7 mW, the second stage was adding additional 20 dB of gain to the 55 dB already reached in the first OPA stage, thus bringing the total small-signal gain to 75 dB. The measured output signal energy dependence on the second stage OPA pump at the maximum seed level is shown in Fig. 3 (solid squares). The generated idler energy, shown in Fig. 3 (open circles) was deduced from the Manley-Rowe condition. At maximum pump the OPA signal energy reached 1.03 mJ, corresponding to a peak power of 208 kW. The conversion efficiency in the second stage OPA was about 37%, while the overall two-stage OPA efficiency was close to 30%. The corresponding signal power extraction efficiencies where 26% and 20%, respectively. The signal output beam had an M2=1.1 and a spectrum which was very similar to that of the seed.
Fig. 4. OPA signal spectrum after propagation in 5m of single-mode telecommunication fiber at different input energies.
The generated signal peak power was high enough to achieve substantial spectral broadening in a 5-m-long piece of conventional telecommunication fiber (see Fig. 4). At the maximum energy of 1 mJ the spectrum was more than 250 nm wide. By increasing the input energy above 0.1 mJ, the amplitude of the peak at the original OPA wavelength actually decreases slightly, with all the excess energy being transferred into the broad spectral background. Stimulated Raman scattering, self-phase modulation and four-wave-mixing are the most likely processes contributing to the observed spectral broadening. An even broader spectrum (500 nm) was observed using 5-m-long dispersion-flattened fiber. This is a simple and inexpensive technique to produce broadband nanosecond pulses for spectroscopic applications.
4. Conclusions
In conclusion we have demonstrated a two-stage PPKTP OPA in the eye-safe spectral region, realized in a single crystal. The maximum signal peak power reached 208 kW in a diffraction-limited and parametric-fluorescence-free beam, with a total signal power extraction efficiency of 20%. Analysis of the QPM OPA small-signal gain dependence on the M2 of the pump beam shows a much stronger gain reduction in noncolinear OPA configuration due to increased angular sensitivity of the phase mismatch. The OPA signal peak power is large enough to generate a broadband spectrum even in standard single-mode telecommunication or dispersion flattened fibers. Further broadening, if needed, can be achieved by employing increased beam confinement in tapered fibers or holey fibers.
References and links
1. K. V. Yuamashev, “Saturable absorber Co2+:MgAl2O4 crystal for Q Switching of 1.34-µm Nd3+:YAlO3 and 1.54-µm Er3+:Glass Lasers,” Appl. Opt. 38, 6343–6346 (1999). [CrossRef]
2. M. S. Webb, P. Moulton, J. J. Kasinski, R. L. Burnham, G. Loiacono, and R. Stolzenberg, “High-averagepowerKTiOAsO4 optical parametricoscillator,” Opt. Lett. 23, 1161–1163 (1998). [CrossRef]
3. J. A. C. Terry, Y. Cui, Y. Yang, W. Sibbett, and M. H. Dunn, “Low-threshold operation of an all-solid-state KTP optical parametric oscillator,” J. Opt. Soc. Am. B 11, 758–769 (1994). [CrossRef]
4. R. Dabu, C. Fenic, and A. Stratan, “Intracavity pumped nanosecond optical parametric oscillator emitting in the eye-safe range,” Appl. Opt. 40, 4334–4340 (2001). [CrossRef]
5. G. W. Baxter, Ph. Schlup, I. McKinnie, J. Hellström, and F. Laurell, “Single-mode near-infrared optical parametric oscillator amplifier based on periodically poled KTiOPO4,” Appl. Opt. 40, 6659–6662 (2001). [CrossRef]
6. S.-K. Choi, M. Vasilyev, and P. Kumar, “Noiseless optical amplification of images,” Phys. Rev. Lett. 83, 1938–1941 (1999). [CrossRef]
7. P. E. Powers, K. W. Aniolek, T. J. Kulp, B. A. Richman, and S. E. Bisson, “Periodically poled lithium niobate optical parametric amplifier seeded with the narrow-band filtered output of an optical parametric generator,” Opt. Lett. 23, 1886–1888 (1998). [CrossRef]
8. K. W. Aniolek, R. L. Schmitt, T. J. Kulp, B. A. Richman, S. E. Bisson, and P. E. Powers, “Microlaserpumped periodically polsed lithium niobate optical parametric generator-optical parametric amplifier,” Opt. Lett. 25, 557–559 (2000). [CrossRef]
9. J. Hellström, G. Karlsson, V. Pasiskevicius, and F. Laurell, “Optical parametric amplification in periodically poled KTiOPO4 seeded by an Er-Yb:glass microchip laser,” Opt. Lett. 26, 352 (2001). [CrossRef]
10. G. Karlsson, V. Pasiskevicius, F. Laurell, and J. A. Tellefsen, “Q-switching of an Er-Yb:glass microchip laser using an acousto-optical modulator,” Opt. Commun. 217, 317–324 (2003). [CrossRef]
11. J. Hellström, V. Pasiskevicius, H. Karlsson, and F. Laurell, “High-power optical parametric oscillation in large-aperture periodically poled KTiOPO4,” Opt. Lett. 25, 174–176 (2000). [CrossRef]
12. I. Jovanovic, B. J. Comaskey, and D. M. Pennington, “Angular effects and beam quality in optical amplification,” J. Appl. Phys. 90, 4328 (2001). [CrossRef]
