bLorsb

Single-mode, singly resonant, pulsed periodically poledlithium niobate optical parametric oscillator

Philip Schlup, Iain T. McKinnie, and Stuart D. Butterworth

We report single longitudinal mode operation of a pulsed Nd:YAG pumped periodically poled lithiumniobate optical parametric oscillator ~OPO!. The combination of a prism and an etalon provides bothcoarse and fine spectral resolution, thereby eliminating parasitic resonant oscillation of cascaded signalwavelengths, idler wavelengths, and adjacent longitudinal modes of the signal field. Optical band-widths of less than 300 MHz have been obtained at signal wavelengths between 1.47 and 1.59 mm. Incomparison with broadband operation, the single-mode OPO shows only a 9% reduction in pump deple-tion with a negligible reduction in extraction efficiency. © 1999 Optical Society of America

OCIS codes: 190.4360, 190.4410, 190.4970, 230.4320.

siot

1. Introduction

Since their recent introduction,1 quasi-phase-matched optical parametric oscillators ~OPO’s! basedon periodically poled lithium niobate ~PPLN! haveeen the subject of considerable research activity.ow- and high-power continuous-wave ~cw!, nanosec-nd, picosecond, and femtosecond devices have beeneported. Potential applications of these devices inpectroscopy and lidar in particular, require narrow-and nanosecond pulsed operation. Although in cw2

and long-pulse3 OPO’s single-frequency operationcould be achieved even with broadband pumping,previous results with nanosecond OPO’s have dem-onstrated the need for a single-mode pump laser.4–6

Narrow-band PPLN OPO’s with both passive linenarrowing6 and injection seeding7 have recently beenreported. The simplicity and inherent broad tuningrange of the passive line-narrowing technique makesit attractive for many applications. However, al-though single-mode operation was obtained with onlya single etalon as reported in Ref. 6, the spectralpurity and operating efficiency of the OPO were lim-ited by processes such as optical parametric genera-

P. Schlup ~pschlup@physics.otago.ac.nz! and I. T. McKinnie arewith the Department of Physics, University of Otago, P. O. Box 56,Dunedin, New Zealand. S. D. Butterworth is with Microlase Op-tical Systems, Ltd., West of Scotland Science Park, Kelvin Cam-pus, Maryhill Road, Glasgow G20 0SP, UK.

Received 19 March 1999; revised manuscript received 2 August1999.

0003-6935y99y367398-04$15.00y0© 1999 Optical Society of America

7398 APPLIED OPTICS y Vol. 38, No. 36 y 20 December 1999

tion and cascading. Cascading occurs because thegain in PPLN is sufficient in certain cases to allow theOPO signal to drive a second quasi-phase-matchedOPO.6,8 Cascading is a potentially useful techniquethat could be used to enhance mid-infrared OPO op-eration or generate output at otherwise inaccessiblewavelengths. However, for our purposes it is a par-asitic process that reduces the maximum attainablesignal energy and introduces undesired spectral com-ponents.

Here we report on the performance of a PPLN OPOthat incorporates intracavity Brewster prisms toeliminate cascaded oscillation and to ensure resonantoperation of only one signal field. The spectralwidth of the OPO signal was narrowed to less than300 MHz by use of a single intracavity etalon. Pumpdepletions of between 40 and 47% and 70–100-mJignal output energies were measured over the tun-ng range between 1.47 and 1.59 mm. The insertionf the intracavity etalon increased the oscillationhreshold by a factor of between 1.5 and 1.9.

2. Experiment

The OPO was pumped by an injection-seededQ-switched Nd:YAG oscillator ~modified Contin-uum Powerlite PL7000!, producing 15-ns pulses at10-Hz repetition rate and a wavelength of 1064 nmin a near-diffraction-limited beam ~M2 5 1.1!. Acombination of a half-wave plate and polarizer wasused as a variable attenuator to control the energyof the pump pulses. A maximum energy of 1 mJwas used to prevent damage to the PPLN crystal.The pump beam was focused by use of a 500-mmfocal-length achromat. Figure 1 shows a sche-

t

twoml

2etemmastrtwTwOlowa

hpr

matic of the OPO configuration. The basic broad-band OPO comprised a 19-mm-long PPLN crystal~Crystal Technology, Inc! and two plane mirrorswith wedged substrates. The PPLN crystal was0.5 mm thick and had eight adjacent poled regionswith poling periods ranging from 28.5 to 29.9 mm in0.2-mm steps. The endfaces were polished and an-ireflection coated with a single layer of SiO2 for

signal wavelengths between 1.4 and 1.7 mm. Thecrystal was mounted in a temperature-controlledoven and kept at 160 °C as a precaution againstphotorefractive effects. It was placed 40 mm be-yond the pump waist to eliminate strong single-pass optical parametric generation that wasobserved when the pump was focused in the crystal.In this configuration the pump spot size was 170mm at the front face of the crystal. The pump en-ered the OPO cavity through one end mirror, aedged glass substrate coated for high reflectivityver the signal wavelength range and 90% trans-ission at the pump wavelength. The OPO cavity

ength of 75 mm ~free spectral range of 1.5 GHz!was the minimum that could be achieved in prac-tice. Four output couplers, with reflectances of95%, 66%, 14%, and 8% at the signal wavelengths,were used in the OPO. The 66% reflectance opticyielded a threshold of 340 mJ and a signal energy of115 mJ at maximum pump energy, compared with a

40-mJ threshold and 15-mJ maximum signal en-rgy for the 95% reflectance mirror. With each ofhe low reflectance mirrors the OPO signal outputnergy was insensitive to the output coupler align-ent. In fact with the output coupler removed, asuch as 140 mJ of signal at 1-mJ pump energy andcorresponding pump depletion of 55% were ob-

erved. It is believed that residual reflections fromhe coated PPLN crystal face formed a low finesseesonator with the high reflector. On the basis ofhese results, the 66% reflectance output coupleras chosen for the remainder of this research.his mirror had a transmission of 70% at the pumpavelength. The glass mirror substrates in thePO have low transmission at the idler wave-

engths and, as the aim of this study was to dem-nstrate single-mode oscillation at signalavelengths, no results of the residual idler outputre presented.

Fig. 1. Experimental configuration of the OPO: HR, high reflec-tor; OC, output coupler.

2

3. Results

The first clues that indicate the onset of cascadedOPO operation in the linear resonator were the ob-servations of a roll-over in the signal output energyand oscillations in the signal temporal profile whenpumping at around three times the threshold level.Figure 2 shows both of these results, observed with ashort cavity length of 45 mm. The oscillations indi-cate a dynamic interplay of the signal with anotherstrong optical field. In an OPO pumped well abovethreshold, backconversion can yield a similar tempo-ral structure through a pump–signal interaction.However, the absence of a corresponding structure onthe depleted pump temporal profile indicates thatthis is not the case here. We confirmed the opera-tion of a cascaded OPO in this case by measuring thewavelength of a weak output beam corresponding tothe sum frequency of the cascaded signal and pumpfields. The cascaded signal wavelengths deducedwere in excellent agreement with predictions basedon the Sellmeier equations reported by Jundt.9 Inthe linear resonator, the cascaded output could besuppressed only by cavity misalignment. Similarmisalignments performed with the singly resonantOPO cavity discussed below indicate that this coarsemethod of cascading suppression results in a higherthreshold pump energy and lower maximum signaloutput energy.

A prism was introduced into the cavity to eliminatecascaded OPO oscillation as well as any residualpump or idler resonance. Two uncoated Brewster-cut prisms were tested: one of SF10 glass substrateand the other of ZnS. Because it was found that theOPO threshold was consistently around 200 mJ

igher over the entire tuning range when the ZnSrism was used, the SF10 prism was selected for theemainder of the measurements. When the SF10

Fig. 2. Temporal profiles ~top! and OPO signal efficiency ~bottom!indicating the presence of cascaded parametric oscillation. Withcascading present, we observe a roll-over in OPO signal efficiencyand an oscillatory signal temporal profile ~a!. Corresponding os-cillations on the depleted pump temporal profile are absent ~b!.With cascading eliminated, the signal output energy increasesmonotonically up to the maximum pump energy ~c!.

0 December 1999 y Vol. 38, No. 36 y APPLIED OPTICS 7399

namtfn

waefiiAsOpotctmt1f

7

prism was inserted into the OPO resonator, we ob-served a threshold increase of only 10 mJ with respectto the linear resonator. Spectral components corre-sponding to cascaded interactions were found to beabsent, demonstrating that the prism dispersion issufficient to obtain singly resonant operation. Allsubsequent OPO measurements were carried outwith this intracavity prism.

We obtained high-resolution spectra of the OPOemission by imaging the fringe patterns of a range ofFabry–Perot interferometers onto a CCD array. In-asmuch as InGaAs or similar detector arrays capableof detecting signal wavelengths around 1.5 mm are

ot widely available, the sum frequency of the signalnd pump beams was used for bandwidth measure-ents. As the linewidth of the pump was well below

he instrument resolution, the bandwidth of the sumrequency light could be attributed solely to the sig-al.Without an intracavity etalon the OPO operatedith a bandwidth of around 0.3 nm, with consider-ble shot-to-shot variation. A single fused-silicatalon with a free spectral range of 30 GHz and anesse of 12 over the signal wavelength range was

nserted between the prism and the output coupler.n etalon tilt of 11 mrad was sufficient to ensureingle-mode operation over the full 1.47–1.59-mmPO signal tuning range accessible by varying theoling period and crystal temperature. Single-modeperation was not sensitive to either the alignment orhe position of the PPLN crystal within the OPOavity, confirming that any residual resonant reflec-ions from the PPLN endfaces do not play a role inode selection. Figure 3 shows measurements of

he OPO linewidth for signal wavelengths of 1.47 and.55 mm that we analyzed by using 14- and 5-GHzree spectral range ~FSR! Fabry–Perot interferom-

Fig. 3. Intensity as a function of distance ~linearized! along anilluminated strip through the fringe patterns measured with theOPO illuminating the 14- and 5-GHz FSR Fabry–Perot interferom-eters. Results are shown for the OPO operating at signal wave-lengths of ~a! and ~b! 1.55 mm, ~c! 1.47 mm.

400 APPLIED OPTICS y Vol. 38, No. 36 y 20 December 1999

eters. Both measurements confirm that only onecavity mode oscillates and the 5-GHz FSR inter-ferometer measurement indicates that the OPO op-erates with a linewidth of around 240 MHz.Measurements with an additional 100-GHz FSR in-terferometer were used to confirm the absence ofother modes of the intracavity etalon in the OPOoutput spectrum.

The addition of the intracavity etalon increased thethreshold and decreased the efficiency of the OPO.Figure 4 shows for comparison the signal output en-ergy and pump depletion as a function of pump pulseenergy for the broadband and the single-mode OPO’sat a signal wavelength of 1.55 mm. The thresholdincrease from 390 to 660 mJ on insertion of the etalonis attributed to the beam walk-off effects in the eta-lon. Such effects can be reduced by the use of thin-ner, lower finesse etalons, but these were found to beineffective in obtaining single-mode oscillation. Atthe 1-mJ maximum pump energy the signal outputpulse energy dropped by 15 mJ from 115 to only 100mJ when the etalon was inserted. A correspondingdrop in pump depletion from 49% to 40% was ob-served. No measurable drop in extraction efficiency~defined as the ratio of OPO signal efficiency to pumpdepletion! was found.

The OPO output could be set to a particular desiredoperating wavelength within its gain bandwidth.The wavelength was stable with regard to shot-to-shot as well as longer term fluctuations. Single-mode tuning of the OPO wavelength was monitoredby use of the 14-GHz FSR interferometer. By ad-justment of the tilt of the intracavity etalon, the OPOfrequency could be smoothly adjusted over a range ofaround 50 GHz. Tuning was accompanied by an in-crease in threshold at a rate of around 8 mJ GHz21

and decrease in output energy at a rate of approxi-mately 2.4 mJ GHz21.

A slight wavelength dependence of the OPOthreshold and pump depletion was observed. Thistrend was also observed in the broadband OPO but is

Fig. 4. ~a! OPO signal output energy and ~b! pump depletion as afunction of pump energy, showing the effect of adding the intra-cavity etalon, at a signal wavelength of 1.55 mm.

tiMsb

ch

6

7

8

9

enhanced here, possibly as a result of the spectraldependence of the etalon coatings. The single-modeOPO threshold decreased from 700 mJ at a signalwavelength of 1.47 mm to 570 mJ at 1.59 mm. Overhe same wavelength range, a corresponding increasen pump depletion from 40 to 47% was measured.

easurements of the signal output indicate that theingle-mode OPO produces a near-diffraction-limitedeam with M2 of 1.2.

4. Summary

In summary, we have demonstrated a simple, singlyresonant single-mode PPLN OPO that operates witha bandwidth of less than 300 MHz. Parasitic cas-caded OPO operation, which can lead to unwantedspectral components and reduced efficiency, waseliminated by coarse frequency selection, allowing asingle etalon to be used to obtain single-mode opera-tion at all signal wavelengths between 1.47 and 1.59mm. A comparison with previously published data6

indicates that the coarse frequency selection ~andonsequent cavity length extension! implementedere leads to a threshold increase of 150 mJ. How-

ever, there is no reduction of beam quality or operat-ing wavelength range, and the elimination ofcascading is accompanied by a slight increase in sig-nal efficiency. In comparison with a broadbandOPO, the threshold for the single-mode OPO in-creased from 390 to 660 mJ. This was accompaniedby a 15-mJ drop in maximum signal energy and a 9%drop in pump depletion.

We thank Quick Wu at the University of Otago forcoating the PPLN crystal and Allister Ferguson ofStrathclyde University for helpful discussions. This

2

research was supported by the New Zealand Foun-dation for Research Science and Technology.

References1. L. E. Myers and W. R. Bosenberg, “Periodically poled lithium

niobate and quasi-phase-matched optical parametric oscilla-tors,” IEEE J. Quantum Electron. 33, 1663–1672 ~1997!.

2. W. R. Bosenberg, A. Drobshoff, J. I. Alexander, L. E. Myers, andR. L. Byer, “93% pump depletion, 3.5-W continuous-wave, singlyresonant optical parametric oscillator,” Opt. Lett. 21, 1336–1338 ~1996!.

3. J. F. Young, R. B. Miles, S. E. Harris, and R. W. Wallace, “Pumplinewidth requirement for optical parametric oscillators,”J. Appl. Phys. 42, 497–498 ~1971!.

4. L. B. Kreuzer, “Single mode oscillation of a pulsed singly reso-nant optical parametric oscillator,” Appl. Phys. Lett. 15, 263–265 ~1969!.

5. L. A. W. Gloster, I. T. McKinnie, Z. X. Jiang, T. A. King, J. M.Boon-Engering, W. E. van der Veer, and W. Hogervorst,“Narrow-band b-BaB2O4 optical parametric oscillator in agrazing-incidence configuration,” J. Opt. Soc. Am. B 12, 2117–2121 ~1995!.

. P. Schlup, S. D. Butterworth, and I. T. McKinnie, “Efficientsingle-frequency pulsed periodically poled lithium niobate op-tical parametric oscillator,” Opt. Commun. 154, 191–195~1998!.

. G. W. Baxter, Y. He, and B. J. Orr, “A pulsed optical parametricoscillator, based on periodically poled lithium niobate ~PPLN!,for high-resolution spectroscopy,” Appl. Phys. B 67, 754–756~1998!.

. M. Vaidyanathan, R. C. Eckardt, V. Dominic, L. E. Myers, andT. P. Grayson, “Cascaded optical parametric oscillations,” Opt.Express 1, 49–53 ~1997!.

. D. H. Jundt, “Temperature-dependent Sellmeier equation forthe index of refraction, ne, in congruent lithium niobate,” Opt.Lett. 22, 1553–1555 ~1997!.

0 December 1999 y Vol. 38, No. 36 y APPLIED OPTICS 7401