Q-switched 2-�m lasers by use of a Cr2�:ZnSesaturable absorber

Tzong-Yow Tsai and Milton Birnbaum

Flash-lamp-pumped Ho:YAG �2090-nm� and Tm:YAG �2017-nm� lasers were, for the first time to ourknowledge, passively Q switched by use of a Cr2�:ZnSe saturable absorber. A Q-switched Ho laser with1.3-mJ pulse energy and �90-ns pulse duration and a Q-switched Tm laser with �3.2-mJ pulse energyand 90-ns pulse duration were demonstrated. Compared with the free-running output energies at theQ-switching threshold pump levels, the Q-switching efficiencies were approximately 5% for the Ho:YAGlaser and 16% for the Tm:YAG laser. © 2001 Optical Society of America

OCIS codes: 140.0140, 140.3540, 140.3580.

1. Introduction

Interest has increased with regard to 2-�m lasersthat use Tm3� or Ho3� as active ions. In particular,high-intensity pulses at this eye-safe spectral regionare quite useful for medical applications and laserranging. Actively Q-switched Tm and Ho laserswere developed during the early 1990s.1–4 The ad-vantages of saturable absorber Q switches �SAQSs�compared with active Q switches are well known andconsist primarily of simplicity, reliability, and econ-omy. At 2 �m, it has been difficult to obtainSAQSs.5 Ho3� doped crystals were used to Q switchTm3�:YAG lasers.6–8 Good repeatable Q-switchingresults ��10 mJ, 60 ns� could be obtained by Ho:YLF�YVO4�CaF2 SAQSs for Tm3�:YAG lasers. However,intracavity focusing was required for these Q-switched lasers. The Cr2�:ZnSe SAQS described inthis paper details the state-of-the-art performancewithout the requirement of intracavity focusing forHo3�:YAG lasers at 2090 nm and for Tm3�:YAG la-sers at 2017 nm.

Transition-metal-ion-doped II–VI compounds haverecently been used to develop tunable lasers in themid-IR.9,10 Because of the noncentrosymmetric

structure in tetrahedral coordination provided byII–VI compounds, the strength of the crystal field issmaller than that in the octahedrally coordinatedcrystals �by a factor of 4:9�, resulting in a redshift ofabsorption and emission spectra to the mid-IR range.In addition, the unshielded 3d orbit of TM ions intetrahedral sites strongly interacts with the outer 4porbit, also resulting in high cross sections and broadspectra of absorption and emission.10,11 In a Cr2�:ZnSe crystal, Cr2� ions replace the Zn2� ions in thetetrahedral sites provided by the ZnSe host, thus notrequiring charge compensation. The inset in Fig. 1shows that the 5D energy level of a free 3d4 Cr2� ionin the tetrahedral site is split into two levels, 5E and5T2, which provide the 1.5–2.2-�m absorption bandand the 1.8–3-�m emission band. Figure 1 alsoshows the background-subtracted absorption spec-trum of a Cr2�:ZnSe sample that we used in the laterQ-switching experiment and that we measured withthe Cary 2400 spectrophotometer.

Table 1 lists some spectral data of Ho3� dopedmaterials and Cr:ZnSe at two Q-switching wave-lengths, 2090 and 2017 nm. The absorption crosssection �a of Cr:ZnSe is �100 times the emissioncross section �e of Tm:YAG at 2017 nm ��2 � 10�21

cm2� and �7 times the �e of Ho:YAG at 2090 nm��1.4 � 10�20 cm2�.12 Compared with the previousHo3� SAQSs, Cr:ZnSe has a much lower saturationenergy fluence �Es h��a� resulting in a reducedvulnerability to damage during the Q-switching op-eration. Furthermore, because of the short decaylifetime of �8 �s, Cr:ZnSe is capable of providing amuch faster repetition rate for Q-switched pulsesthan Ho3� materials ��a � 10 ms�. To our knowl-edge, this is the first report of a chalcogenide SAQS

At the time of this research, T. Y. Tsai and M. Birnbaum werewith the Center for Laser Studies, University of Southern Califor-nia, SSC 313, University Park, Los Angeles, California 90089-1112. M. Birnbaum is now with the Center for Laser Studies,Department of Electrical Engineering, 30332 Via Borica, RanchoPalos Verdes, California 90275-0000.

Received 1 October 2000; revised manuscript received 5 Febru-ary 2001.

0003-6935�01�366633-05$15.00�0© 2001 Optical Society of America

20 December 2001 � Vol. 40, No. 36 � APPLIED OPTICS 6633

for Ho lasers at 2090 nm. Thus, Cr2�:ZnSe is astate-of-the-art 2-�m SAQS for Ho:YAG lasers aswell as for Tm:YAG lasers, because they do not re-quire intracavity focusing.

2. Experiments

A. Bleaching Experiment

Bleaching experiments are essential to obtain neces-sary spectral data. The absorption cross section �aand the doping concentration Na0 can be obtained bymatching the experimental results with the theoret-ical predictions of the Avizonis–Grotbeck equation.13

In our experiment, a Tm:YAG laser Q switched byCr2�:ZnSe at 2017 nm was used as a pulse sourcewith a pulse energy of �3.5 mJ and a pulse durationof �70 ns. A set of neutral density filters and a 7-cmfocal-length lens were employed to obtain incidentenergy fluences in the 0.1–7-J�cm2 range. The Cr:ZnSe sample was 0.9 mm thick, dark brown, and notantireflection coated. The bleaching experimentalresults are shown in Fig. 2. Before trying to matcha theoretical curve to the experimental data, the ab-sorption levels of Cr:ZnSe must be considered. Bothabsorption and emission bands originate from thesame transition levels 5E–5T2. The emission band isredshifted because of the Stokes shift but still par-tially overlaps the absorption band from 1.8 to 2.2�m.10 The emission cross section �e of Cr:ZnSe is

�2.7 times �a at 2017 nm and �6 times �a at 2090nm. It was anticipated that Cr:ZnSe should func-tion like a two-level saturable absorber. In fact, thesimulation result shows agreement with experimen-tal data of Q-switching performance �see Subsection2.B�, when a model of an ideal two-level absorber isused. The Avizonis–Grotbeck equation was modi-fied to consider the energy level structure and theexcited-state absorption �ESA� cross section �ESA of aSAQS as follows:

E�z�

z� �

hNa0

pa�1 �

�ESA

pa�a��1 � exp��

pa�a

hE� z���

� �Na0

�ESA

pa� �a�E� z�, (1)

where E�z� is the energy fluence �energy per unitarea�; h is the photon energy; �a is the ground-stateabsorption cross section; �ESA is the ESA cross sec-tion; �a corresponds to nonsaturable loss in the sat-urable absorber; and pa is the population reductionfactor defined as 1 for an ideal three-level saturableabsorber and as �1 � g1�g2� for a two-level saturable

Fig. 1. Background-subtracted absorption spectrum of a Cr2�:ZnSe 5.6-mm thick crystal. The optical density at 2.09 �m isapproximately 0.28.

Fig. 2. Bleaching experimental data and theoretical curves of themodified Avizonis–Grotbeck equation. The middle curve was ob-tained assuming an ideal two-level absorber �pa 2� with �a 2 �10�19 cm2. The upper curve is based on pa 2 and �a 2.3 �10�19 cm2. The lower curve shows pa 1 and �a 2.3 � 10�19

cm2 ��a is �2.3 � 10�19 cm2 according to information reported inRefs. 10 and 11�.

Table 1. Characteristics of Saturable Absorbers at Corresponding Wavelengths for a Tm:YAG Laser and a Ho:YAG Laser

Saturable AbsorberAbsorption Cross Section

��10�20 cm2�

DecayLifetime at

�300 KSaturation FluenceEs h��a �J�cm2� Reference

Ho:YLiF4 �a, 0.19 for � polariza-tion at 2017 nm

�12 ms 52 6

�0.16 for � polarization�Ho:YVO4 �a, 0.6 at 2017 nm — 16 7Ho:CaF2 �a, 0.27 at 2017 nm �24 ms 37 8Cr:ZnSe �a, 23 at 2017 nm 8 �s 0.43 10Cr:ZnSe �a, 10 at 2090 nm 8 �s 0.98 10

6634 APPLIED OPTICS � Vol. 40, No. 36 � 20 December 2001

absorber �g is the number of level degeneracies�.The original Avizonis–Grotbeck equation could be ob-tained if one assumes that pa 2 and �ESA 0.Because of the lack of resonance from 5E to a higherlevel, the ESA was ignored in the simulation, �ESA �0. The matching bleached curve in Fig. 2, obtainedby use of a two-level model with pa 2, reveals that�a is approximately �2 � 0.2� � 10�19 cm2 at 2017 nm,which is close to the �a value of �2.3 � 10�19 cm2 asreported by Page et al.10 and by DeLoach et al.11

According to the absorption spectrum �Fig. 1�, �a at2090 nm could be calculated to be �1 � 10�19 cm2.

B. Q-Switching Experiment

Because intracavity focusing was not required, theQ-switching experimental arrangement was simple�shown in Fig. 3�. The flash-lamp-pumped resona-tor was 33 cm long and consisted of a 100% reflectiv-ity rear mirror and an 80% reflectivity outputcoupler. The Cr:ZnSe samples were dark brownwith parallel-plane uncoated surfaces. It was in-ferred that the Cr2� ions were not uniformly dopedinto the ZnSe samples because of the color variationof the samples. We determined the nonsaturablelosses in the samples by measuring the single-passinternal transmissions at 2500 nm using a Cary 2400spectrophotometer. Each of our samples exhibitedsubstantial nonsaturable loss, as listed in Table 2.A 4 mm � 76 mm Tm3� �6%�, Cr3� �0.6%�:YAG and

a 5 mm � 83 mm Ho3��0.5%�, Tm3��6%�, Cr3��0.5%�:YAG were obtained from the Crystal Products Divi-sion of Union Carbide Corporation. Free-runningperformance of the flash-lamp-pumped lasers wasfirst tested with an output coupler with 80% reflec-tivity, as shown in Fig. 4. The Tm:YAG laser exhib-ited a higher slope efficiency than the Ho:YAG laser.

After the Cr:ZnSe SAQS was aligned in the reso-nator, the Tm:YAG laser and the Ho:YAG laser werepassively Q switched at 2017 and 2090 nm, respec-tively. The experimental data and the correspond-ing simulation results are listed in Table 2. Thesystem was operated at a lasing threshold conditionwith the threshold pump energy Epth. More intenseQ-switched pulses could be obtained with higherpump energies ��Epth�. The crystal Cr:ZnSe �sam-ple 4� is the same sample that was tested in thebleaching experiment. In the simulation we as-sumed that pa 2, Ag Aa 0.01 cm2, �a,2017nm �

2 � 10�19 cm2, and �a,2090nm � 1 � 10�19 cm2, asobtained from the previous bleaching experiment.The simulations show close agreement with the ex-perimental results.

Typical Q-switched Ho:YAG and Tm:YAG laserpulses obtained by use of Cr:ZnSe are shown in Fig.5. Compared with free-running output energies atthe same pump level energy, we obtained Q-switching efficiencies of �16% for the Tm:YAG laserand �5% for the Ho:YAG laser. The surfaces of the

Fig. 3. Experimental arrangement of a Cr:ZnSe saturable ab-sorber Q-switched Ho:YAG and Tm:YAG laser.

Fig. 4. Free-running performance of a Ho:YAG laser and a Tm:YAG laser with an output coupler of 80% reflectivity.

Table 2. Characteristics of Lasers Q-Switched by Cr:ZnSe Samples without Intracavity Focusinga

Q-Switching Sample

Characteristics of CrZnSe SamplesExperimental

OutputsSimulation

Results

ODNonsaturable Loss

��Loss Coefficient, cm�1�Na0

�cm�3�Eout

�mJ�Wp

�ns��

�%�Eout

�mJ�Wp

�ns�

1, Tm:YAG at 2017 nm 0.05 0.05 �0.25� 1.3 � 1018 3.2 90 16 4.1 922, Ho:YAG at 2090 nm 0.2 0.17 �1.7� 1.8 � 1019 1.3 90 5 1.4 723, Ho:YAG at 2090 nm 0.28 0.22 �0.4� 4.8 � 1018 1.5 70 2.5 1.7 404, Ho:YAG at 2090 nm 0.3 0.24 �2.4� 3 � 1019 1.6 50 1 1.8 37

aOD, optical density; Na0, Cr2� doping concentration; Eout, pulse energy; Wp, pulse width �FWHM�; �, Q-switching efficiency.

20 December 2001 � Vol. 40, No. 36 � APPLIED OPTICS 6635

SAQS were damaged when the output energy washigher than 4 mJ. It is reasonable to have higherefficiency from Tm:YAG than from Ho:YAG, becausethe cross-sectional ratio �a��g of Cr:ZnSe to Ho:YAGis much smaller than that of Cr:ZnSe to Tm:YAG.In the experiment, the high surface reflectivity of theCr:ZnSe sample �n � 2.47� contributed to cavity lossand degraded the Q-switching performance. Anantireflection-coated Cr:ZnSe crystal should reducedamage vulnerability and result in a higher outputpulse energy.

3. Simulation

The SAQS performance can be modeled by use of themodified rate equations of Siegman14:

dndt

� �Kg Ng � Ka Na � �c � �m�n, (2)

dNg

dt� �pg Kg nNg, (3)

dNa

dt� �pa Ka nNa, (4)

where n is the number of photons in the resonancemode, Ng is the effective gain population, Na is theeffective absorption population. �c is the cavity loss,

and �m is the useful loss caused by transmission ofthe output coupler. Kg is the coupling coefficientdefined as �e��Agt1�, where �e is the emission crosssection of the laser, Ag is the cross-sectional beamarea through the laser, and t1 is the one-way trip timeof the resonator. Ka is �a��Aat1�, �a is the absorp-tion cross section, Aa is the beam area through theSAQS. pa and pg are the population reduction fac-tors related to the level degeneracies of the SAQS andthe laser. Because Tm and Ho are classified asthree-level lasers, pg 2 was assumed. In addition,spontaneous emission in the lasers, population relax-ation of Cr:ZnSe and pump rate, were ignored in therate equations. Computer simulation starts fromt 0 when the initial condition of Ng is assumed to beKgNgi KaNaT � �c � �m, where NaT is the initialabsorption population. Simulation results and thecomparison with experimental data are given in Sub-section 2.B.

4. Conclusions

Cr:ZnSe has been demonstrated to be a state-of-the-art saturable absorber Q-switch for Ho:YAG andTm:YAG lasers. Intracavity focusing was not re-quired. The Q-switching efficiencies of Q-switchedTm:YAG and Ho:YAG lasers were 16% and 5%, re-spectively. Cr:ZnSe samples of high optical qual-ity �reduced nonsaturable loss� and antireflectioncoating would surely provide enhanced Q-switchedperformance.

References1. B. T. McGuckin, R. T. Menzies, and H. Hemmati, “Efficient

energy extraction from a diode-pumped Q-switched Tm,Ho:YLiF4 laser,” Appl. Phys. Lett. 59, 2926–2928 �1991�.

2. S. R. Bowman, M. J. Winings, S. K. Searles, and B. J. Feldman,“Short-pulsed 2.1 �m laser performance of Cr,Tm,Ho:YAG,”IEEE J. Quantum Electron. 27, 1129–1131 �1991�.

3. T. S. Kubo and T. J. Kane, “Diode-pumped lasers at five eye-safe wavelengths,” IEEE J. Quantum Electron. 28, 1033–1040�1992�.

4. S. R. Bowman, J. G. Lynn, S. K. Searles, B. J. Feldman, J.McMahon, W. Whitney, C. Marquardt, D. Epp, G. J. Quarles,and K. J. Riley, “High-average-power operation of aQ-switched diode-pumped holmium laser,” Opt. Lett. 18,1724–1726 �1993�.

5. B. A. Ermakov, A. V. Lukin, and L. M. Sobolev, “PassiveQ-switching of a 2-�m laser,” Opt. Spectrosc. �USSR� 63, 137�1987�.

6. Y.-K. Kuo, M. Birnbaum, and W. Chen, “Ho:YLiF4 saturableabsorber Q-switch for the 2-�m Tm,Cr:Y3Al5O12 laser,” Appl.Phys. Lett. 65, 3060–3062 �1994�.

7. Y.-K. Kuo and M. Birnbaum, “Ho:YVO4 solid-state saturable-absorber Q switch for a 2-�m Tm,Cr:Y3Al5O12 laser,” Appl.Opt. 35, 881–884 �1996�.

8. Y.-K. Kuo, M. Birnbaum, F. Unlu, and M.-F. Huang, “Ho:CaF2

solid-state saturable-absorber Q switch for the 2-�m Tm,Cr:Y3Al5O12 laser,” Appl. Opt. 35, 2576–2579 �1996�.

9. J. Mckay, D. Krause, and K. L. Schepler, “Optimization ofCr2�:CdSe for efficient laser operation,” in Advanced Solid-State Lasers, H. Injeyan, U. Keller, and C. Marshall, eds., Vol.34 of OSA Trends in Optics and Photonics Series �OpticalSociety of America, Washington, D.C., 2000�, pp. 218–224.

10. R. H. Page, K. I. Schaffers, L. D. DeLoach, G. D. Wilke, F. D.Patel, J. B. Tassano, Jr., S. A. Payne, W. F. Krupke, K. T.

Fig. 5. Output pulses of a Cr2�:ZnSe saturable absorberQ-switched �a� Ho3�:YAG laser and �b� Tm3�:YAG laser at 2090and 2017 nm, respectively.

6636 APPLIED OPTICS � Vol. 40, No. 36 � 20 December 2001

Chen, and A. Burger, “Cr2�-doped zinc chalcogenides as effi-cient widely tunable mid-infrared lasers,” IEEE J. QuantumElectron. 33, 609–619 �1997�.

11. L. D. DeLoach, R. H. Page, G. D. Wilke, S. A. Payne, andW. F. Krupke, “Transition metal-doped zinc chalcogenides:spectroscopy and laser demonstration of a new class of gainmedia,” IEEE J. Quantum Electron. 32, 885–895 �1996�.

12. R. C. Powell, Physics of Solid-State Laser Materials �Springer-Verlag, New York, 1998�, Chap. 9.

13. P. V. Avizonis and R. L. Grotbeck, “Experimental and theoret-ical ruby laser amplifier dynamics,” J. Appl. Phys. 37, 687–693�1966�.

14. A. Siegman, Lasers �University Science, Mill Valley, Calif.,1986�, Chap. 26.

20 December 2001 � Vol. 40, No. 36 � APPLIED OPTICS 6637