Continuous-wave broadly tunable Cr2+:ZnSe laser pumped by a thulium fiber laser

Optics Communications 268 (2006) 115–120 www.elsevier.com/locate/optcom

Continuous-wave broadly tunable Cr2+:ZnSe laser pumped by a thulium fiber laser Alphan Sennaroglu a,*, Umit Demirbas a, Nathalie Vermeulen b, Heidi Ottevaere b, Hugo Thienpont b a b

Laser Research Laboratory, Department of Physics, Koc¸ University, Rumelifeneri Yolu, Sariyer 34450, Istanbul, Turkey Vrije Universiteit Brussel (VUB), Department of Applied Physics and Photonics (IR-TONA), B-1050 Brussels, Belgium Received 24 February 2006; received in revised form 25 May 2006; accepted 22 June 2006

Abstract We describe a compact, broadly tunable, continuous-wave (cw) Cr2+:ZnSe laser pumped by a thulium fiber laser at 1800 nm. In the experiments, a polycrystalline ZnSe sample with a chromium concentration of 9.5 · 1018 cm3 was used. Free-running laser output was around 2500 nm. Output couplers with transmissions of 3%, 6%, and 15% were used to characterize the power performance of the laser. Best power performance was obtained with a 15% transmitting output coupler. In this case, as high as 640 mW of output power was obtained with 2.5 W of pump power at a wavelength of 2480 nm. The stimulated emission cross-section values determined from laser threshold data and emission measurements were in good agreement. Finally, broad, continuous tuning of the laser was demonstrated between 2240 and 2900 nm by using an intracavity Brewster cut MgF2 prism and a single set of optics.  2006 Elsevier B.V. All rights reserved. PACS: 42.55.f; 42.55.Rz; 42.60.Pk; 42.60.Lh

1. Introduction When divalent transition metal ions such as Cr2+ or Fe2+ are introduced into chalcogenide hosts, strong absorption and emission bands are formed in the mid-infrared and depending on the ion-host combination, lasing action can be obtained in the 2–4 lm wavelength range. One of the most important members of this class of infrared solid-state lasers is Cr2+:ZnSe. As was first demonstrated by De Loach et al. [1,2], broadly tunable laser emission can be obtained from Cr2+:ZnSe between 2000 and 3100 nm in cw case [3], and between 1880 and 3100 nm during pulsed operation [4]. To date, gain-switched [1], continuous-wave (cw) [5], and mode-locked [6] operations have been demonstrated. Important applications of Cr2+:ZnSe lasers include medicine [7], atmospheric imaging [8], vibrational spectroscopy

*

Corresponding author. Tel.: +90 212 3381429; fax: +90 212 3381559. E-mail address: [email protected] (A. Sennaroglu).

0030-4018/$ - see front matter  2006 Elsevier B.V. All rights reserved. doi:10.1016/j.optcom.2006.06.064

[9–11], and optical excitation of mid-infrared lasers and optical parametric oscillators [12]. The broad absorption band of the Cr2+:ZnSe gain medium extending from 1500 to 2100 nm makes it possible to use several alternative laser systems as pump sources (absorption peak = 1775 nm, width (FWHM) = 365 nm [13–15]. Some of the pump lasers that have been used to date include Co2+:MgF2 [1,2,16–18], Tm:YALO [5], Tm: YLF [19], NaCl:OH [20], Ho:YAlO [21], Er:YAG [10], laser diodes [22–24], Er-doped fiber lasers [8,9,25–28], Ba(NO3)2 and BaWO4 Raman lasers [29,30], and KTiOPO4(KTP) optical parametric oscillators (OPO) [31]. A number of important criteria need to be met in choosing the optimum pump source for practical Cr2+:ZnSe systems, such as commercial availability, possibility of room temperature and cw operation, and long-term stability. In addition, ion–ion interaction at high doping concentrations also places constraints on the choice of the optimum pumping wavelength. To understand this further, we prepared several Cr2+:ZnSe samples by diffusion doping [15] and

116

A. Sennaroglu et al. / Optics Communications 268 (2006) 115–120

investigated the influence of the doping concentration on the lasing parameters. In the synthesis experiments, polycrystalline ZnSe samples were used. The dopant (Cr or CrSe powder) and the host (ZnSe) were placed in different compartments of silica ampoules, which were sealed under high vacuum. Samples were then subjected to thermal diffusion at temperatures between 800 and 1100 C and for diffusion times from several hours to tens of days. Based on the absorption spectra of more than 10 polycrystalline Cr2+:ZnSe samples, we found that the differential loss coefficient at 2500 nm (aloss, in the units of cm1) increases as a function of the average Cr2+ concentration NCr (cm3) according to the empirical formula aloss  ð0:04  0:02Þ þ fð0:02 þ 0:01Þ  1018 gN Cr :

ð1Þ

Furthermore, time-dependent fluorescence measurements showed that the fluorescence lifetime sF monotonically decreases with increasing active ion concentration: sF ¼ 1þ

sF0  2 :

In this paper, we describe a broadly tunable, cw, Cr2+:ZnSe laser pumped by a Tm-fiber laser at 1800 nm. In the experiments, cw power performance of the laser was characterized by using different output couplers. With the 15.3% transmitting output coupler and with 2.5 W of pump power, as high as 640 mW of output power was obtained at 2480 nm. A detailed analysis of the laser performance was also carried out. Resonator losses were determined by using the threshold, power efficiency, and absorption data. Numbers obtained from three different methods were in good agreement with each other. From the threshold data, the stimulated emission cross-section was further determined to be 4.2 · 1023 m2. This was also in good agreement with the value obtained by using the emission and lifetime data. Finally, an intracavity Brewster cut MgF2 prism and a single set of optics were used to obtain continuous tuning between 2240 and 2900 nm. 2. Experimental

ð2Þ

N Cr N0

In Eq. (2), the empirically determined parameters have the values sF0 = 5.56 ls and N0 = 17 · 1018 ions/cm3. The concentration dependence of the fluorescence lifetime reduces the fluorescence efficiency at high doping concentrations. This also leads to higher thermal gradients in the gain medium and lowers the power efficiency as the active ion concentration is increased. In light of the above results, we see that if the pump wavelength is close to the peak absorption wavelength of 1775 nm, Cr:ZnSe samples with relatively low ion concentration can be used and the deleterious effects mentioned above can all be minimized. As a concrete example, let us compare the performance of two hypothetical Cr:ZnSe systems one pumped at 1800 nm (thulium pump laser) and the other at 1550 nm (erbium pump laser). We assume that the sample length is 2 mm and the desired pump absorption is 80% (absorption coefficient of about 8 cm1). At 1800 nm, this requires a sample with an active ion concentration of 7 · 1018 ions/cm3. The corresponding lifetime and the passive loss are 4.8 ls and 4%, respectively. On the other hand, to obtain the same absorption at 1550 nm, the average ion concentration should be approximately 22 · 1018 ions/cm3,with a corresponding lifetime of 2.1 ls and passive loss of 10%. Notice that in this case, the dramatic reduction in the fluorescence lifetime adversely affects the lasing efficiency and increases the amount of thermal loading. Because of this reason, lasers based on the 2-lm transitions of the trivalent thulium ion (Tm3+) are among the most preferred pump alternatives. Furthermore, Tm3+-doped fiber pump lasers offer the additional advantage of reduced thermal management requirement for the pump. In previous studies, Peterson et al. used 1.9-micron Tm-fiber laser to pump a Cr2+:ZnSe laser and obtained as high as 335 mW with 4 W of pump power [28].

A schematic of the cw Cr2+:ZnSe laser is shown in Fig. 1. A commercial thulium fiber laser (IPG Photonics) capable of delivering cw powers up to 5 W at 1800 nm was used as the pump source. The collimated output of the fiber pump had a 1/e2 beam waist of 2.25 mm and a measured M2 of 1.03. Because pump instabilities with more than 5% fluctuation were observed at higher pumping levels, pump powers up to 3 W were used in the lasing experiments. We used a commercial 2.6-mm-long Cr2+:ZnSe sample obtained from Spectragen, Inc. It was clamped inside a copper holder maintained at 15 C and placed at Brewster’s angle inside an astigmatically compensated x-cavity between two curved high reflectors with R = 10 cm (M1 and M2). At Brewster’s angle, the Fresnel reflection loss from the sample surface was measured to be 0.2%. One arm of the cavity was terminated with a flat end high reflector (M3), and the other contained a flat output coupler (M4). In the power measurements, output couplers with 3%, 5.8%, and 15.3% transmission (2500 nm) were used. The high reflectivity of the resonator mirrors extended from 2260 to 2900 nm (reflection P 99%), and they had a measured leakage of about 0.4% at 2500 nm. Broadband coating of the high reflectors was essential in obtaining smooth, continuous tuning between 2.2 and 2.9 lm. Fig. 2a shows the measured absorption spectrum of the Cr2+:ZnSe sample used in the experiments. Inset in Fig. 2a shows the background subtracted absorption spectrum. The pump absorption coefficient at 1800 nm was determined to be 10.7 cm1. The round-trip loss at the lasing wavelength was 4.1%. This give a crystal figure of merit (FOM = a1800/a2500, a1800 = differential absorption coefficient at 1800 nm and a2500 = differential loss coefficient at 2500 nm) of 70. Shown in Fig. 2b is the time-dependent fluorescence decay curve of the Cr2+:ZnSe sample. A pulsed optical parametric oscillator (OPO) operating at 1570 nm was used for excitation. The pulsewidth and the repetition rate of the OPO were 65 ns and 1 kHz, respec-

A. Sennaroglu et al. / Optics Communications 268 (2006) 115–120

HR 1

117

Tm-Fiber Laser

L1

M2

M1

HR2

M4 OC

M3

Fig. 1. Schematic of the cw Cr2+:ZnSe laser (see the text for a description of the various components).

1

a

Fluorescence Intensity (a.u.)_

Absorption Coefficient ( cm -1)

40

12 9

30

6 3 0

20

1200 1400 1600 1800 2000 2200 2400

10

0 300

b 0.75

0.5

0.25

0 600

900

1200

1500

1800

2100 2400

2700

0

3000

5

10

15

20

25

Time (μs)

Wavelength (nm)

Fig. 2. Measured (a) absorption spectrum and (b) time-dependent fluorescence decay of the Cr2+:ZnSe sample.

3. Results and analysis Fig. 3 shows the variation of the output power as a function of the incident pump power for three different output couplers. For example, in the case of the 3% transmitting output coupler, 300 mW of output power was obtained with 2.3 W of pump power. The corresponding threshold pump power and the incident power slope efficiency were 207 mW and 16%, respectively. The free running laser output wavelength shifted slightly for the different output couplers used. While the lasing peak was at 2512 nm with the 3% OC, it shifted to 2480 nm with the 15.3% OC. The width of the laser line (FWHM) was about 0.5 nm. The best power performance was obtained with the 15.3% transmitting output coupler.

Here, as high as 640 mW of output power was obtained with 2.5 W of pump, where slope efficiency with respect to incident power was 34%. The output power began to saturate beyond 2.5 W of pump due to thermal loading

750

CrZnSe Power (mW)

tively. By doing a single-exponential fit to the experimental data, the fluorescence lifetime was determined to be 4.6 ls. By using an input lens with a focal length of 10 cm, the pump beam was focused to a waist of 36 lm inside the gain medium. The total cavity length was 90 cm, where the long arm length was 55 cm. From the standard ABCD analysis of the cavity, the beam waist at the center of the stability region was determined to be 45 lm.

3%

600

5.80% 15.30%

450

300

150

0 0

750

1500

2250

3000

Incident Power (mW) Fig. 3. Measured variation of the output power as a function of the incident pump power for three different output couplers.

118

A. Sennaroglu et al. / Optics Communications 268 (2006) 115–120

Fig. 4b). The disagreement between the Findlay–Clay and Caird analyses was also observed in previous laser studies with Cr2+:ZnSe [2,5,16,22], and attributed to the existence of residual ground state absorption at the lasing wavelength [3,16,22,34,35]. Since the existence of ground state absorption affects only the threshold power (but not the slope efficiency) [16,36] Findlay–Clay analysis gave higher L values than Caird analysis [2,5,16,22]. L was further estimated to be 6.6% from direct loss measurements taking into account absorption and reflection losses of the crystal as well as the leakage from high reflectors (2 · 2.1% gain medium passive loss, 2 · 0.2% reflection loss from the gain medium surface, 5 · 0.4% leakage from the cavity high reflectors). The stimulated emission cross-section of the gain medium at the laser wavelength can also be determined from the threshold data by using the formula [36,37]

Table 1 Threshold pump power and the slope efficiency values for each output coupler (OC) OC transmission (%)

Lasing threshold (mW)

Slope efficiency (%)

3 5.8 15.3

208 277 413

16 23 34

[19]. Table 1 summarizes the threshold pump power and the slope efficiency obtained with each output coupler. We estimated the resonator losses at the lasing wavelength by using the Findlay–Clay [32] and Caird [33] analyses. Findlay–Clay analysis is based on the fact that the threshold pump power Pth required to attain lasing is proportional to the total round trip loss and can be expressed as P th ¼ AðL þ T Þ;

ð3Þ

where A is a constant which depends on the quantum electronic characteristics of the medium, L is the passive round trip resonator loss, and T is the transmission of the output coupler. A plot of Pth versus T gives a straight line and the loss L can be determined by using the best-fit values of the slope and intercept. In Caird analysis, one uses the fact that the slope efficiency g of the laser can be expressed to a good approximation as g ¼ g0

T ; T þL

rem ðkL Þ ¼

p hmP ðx2L þ x2P Þ ðL þ LGSA þ T Þ: 4 P th sf ð1  ExpðaP hÞÞ

ð5Þ

In Eq. (5), hmP is the energy of pump photons, sf is the fluorescence lifetime of the upper laser level, Pth is the laser threshold pump power, xL and xp are the average values of the laser and pump spot sizes in the gain medium, and (1  Exp(aph) is the total absorption at the pump wavelength. Findlay–Clay value of L + LGSA (10.6%) determined above was used in the calculation of the stimulated emission cross-section. Threshold data taken with the 3%, 5.8%, and 15.3% output couplers were used. The average value of re came to 4.2 · 1023 m2 at the laser wavelength (around 2500 nm) where the deviation was 0.2 · 1023 m2. Alternatively, re can be determined from the spectroscopic emission data by using the Fuchtbauer–Ladenburg formula [30,38] given by

ð4Þ

where g0 is the maximum slope efficiency that can be obtained at high output coupling. g0 is determined by such factors as the finite quantum defect of the laser transition, mode matching between the pump and the laser beams, and luminescence quantum efficiency. A plot of 1/g as a function of 1/T hence gives a straight line and L can again be determined in a straightforward way. Fig. 4a and b show the plots of the Findlay–Clay and Caird analyses, respectively. The best-fit values of L were determined to be 10.6% (Findlay–Clay, Fig. 4a) and 6.2% (Caird,

rem ðkÞ ¼

8pcn2 srad

k5 R band

kI e ðkÞdk

I e ðkÞ:

ð6Þ

Here, c is the speed of light, n is the index of refraction, srad is the radiative lifetime, and Ie(k) is the fluorescence

a

b

8

y = 12.997x + 2.0917

6

400

y = 16.07x + 170.27 1/η

Thereshold Power (mW)_

500

300

4

200 2

100 0

0 0

4

8 T (%)

12

16

0

0.1

0.2

0.3

0.4

1/T

Fig. 4. (a) Findlay–Clay and (b) Caird analyses of the power efficiency data. L was determined to be 10.6% and 6.2% from the Findlay–Clay and Caird analyses, respectively.

119

280

20

210

15

140

10

70

0 2200

5

2400

2600

Fig. 6. Continuous-wave tuning curve of the Cr2+:ZnSe laser taken with an intracavity MgF2 prism. Solid line shows the variation of the output coupler transmission with wavelength.

2500 nm, the peak of the laser tuning curve occurred at 2320 nm due to insertion of the tuning prism. In these measurements, the laser cavity was not purged, and the relative humidity was around 67%. With a single set of optics, broad tuning could be obtained in the 2240–2900 nm wavelength range. Due to the atmospheric loses inside the cavity, the tuning curve has deeps around 2.6 lm as it was previously observed [41] in non-purged cavities. Tuning below 2.24 lm was optics limited where the reflectivity of the cavity high reflectors decreased to 90% at a wavelength of 2210 nm. 4. Conclusions In conclusion, we have described in detail the operation of a cw broadly tunable Cr2+:ZnSe laser pumped by a thuliumfiber laser at 1800 nm. Output powers as high as 640 mW

Normalized Intensity (au)

0.75

0.50

0.25

1950

2200

0 3000

Wavelength (nm)

1.00

0.00 1700

2800

OC Transmission (%)

intensity signal. Normalized emission spectrum of Cr2+:ZnSe is shown in Fig. 5. In this measurement, the gain medium was excited at 1490 nm with a Cr4+:YAG laser. A sample with a very low chromium concentration (3 · 1018 cm3) was used in the emission measurement to minimize the reabsorption of the emission between 1.7 and 2 lm due to the overlapping emission and absorption bands at that region. The fluorescence spectrum was corrected for filter, detector and grating response. The data were used to determine the lineshape function and the area under the emission spectrum. Other parameter values appearing in the above equation were taken as n = 2.45, srad = 6 ls [31]. The peak of the observed emission spectrum was at a lower wavelength (around 2050 nm), than most of the previously reported values (around 2400 nm) [1,30,39], but came close to the peak position reported by Mirov et al. (around 2150 nm) [26]. This difference is possibly due to the minimization of reabsorption of emitted photons in the lower wavelength side by using a slightly doped sample, in our case. The peak emission cross section value determined using Eq. (6) and the measured emission spectrum gave a value of 12 · 1023 m2 (at 2050 nm) in good agreement with the previously reported values in literature [1,3,29,30,40]. However, the reported peak emission wavelengths are different (around 2400 nm). At the wavelength of 2500 nm, spectroscopic data gave a value of 4.3 · 1023 m2 for re, in good agreement with the laser cross-section determined from threshold data. A Brewster-cut MgF2 prism was finally placed in the high-reflector arm of the resonator to investigate the wavelength dependence of the output power. At the pump power of 1.8 W, the output of the laser was reduced from 255 to 250 mW after the insertion of the prism with the 3% output coupler. Fig. 6 shows the variation of the output power as a function of the emission wavelength (dots), as well as the variation of the output coupler transmission (solid line). While the free running laser wavelength was around

Output Power (mW)

A. Sennaroglu et al. / Optics Communications 268 (2006) 115–120

2450

2700

2950

Wavelength (nm)

Fig. 5. Emission spectrum of the Cr2+:ZnSe medium excited at 1490 nm.

120

A. Sennaroglu et al. / Optics Communications 268 (2006) 115–120

were obtained with an input pump power of 2.5 W. Power data were analyzed to determine the passive resonator losses and the stimulated emission cross-section of the gain medium. The emission cross-section values determined from laser efficiency data and spectrum measurements were in good agreement with each other. Continuous, broad tuning of the laser was demonstrated between 2240 and 2900 nm by using an intracavity Brewster cut MgF2 prism and a single set of optics. Finally, we note that use of powerful thulium-fiber pump lasers, which are now available with output powers in excess of 150 W, offers an effective route to power scaling in Cr2+:ZnSe lasers with minimum thermal loading. Acknowledgements We thank Adnan Kurt for his assistance in the experiments. This project was supported by the Network of Excellence in Micro-Optics (NEMO) funded by the European Union 6th Framework program. U. Demirbas acknowledges the support of the Scientific and Technical Research Council of Turkey (TUBITAK) in the framework of the BAYG program. References [1] L.D. DeLoach, R.H. Page, G.D. Wilke, S.A. Payne, W.F. Krupke, IEEE J. Quantum Elect. 32 (1996) 885. [2] 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. Chen, A. Burger, IEEE J. Quantum Elect. 33 (1997) 609. [3] I.T. Sorokina, Crystalline mid-infrared lasers, in: I.T. Sorokina, K.L. Vodopyanov (Eds.), Solid-State Mid-Infrared Laser Sources, vol. 89, Heidelberg, Berlin, 2003. [4] U. Demirbas, A. Sennaroglu, Opt. Lett. 2006, in press. [5] G.J. Wagner, T.J. Carrig, R.H. Page, K.I. Schaffers, J. Ndap, X. Ma, A. Burger, Opt. Lett. 24 (1999) 19. [6] T.J. Carrig, G.J. Wagner, A. Sennaroglu, J.Y. Jeong, C.R. Pollock, Opt. Lett. 25 (1999) 168. [7] B. Jean, T. Bende, Mid-IR laser applications in medicine, in: I.T. Sorokina, K.L. Vodopyanov (Eds.), Solid-State Mid-Infrared Laser Sources, Topics in Applied Physics, vol. 89, Heidelberg, Berlin, 2003. [8] C. Fischer, E. Sorokin, I.T. Sorokina, M.W. Sigrist, Opt. Laser Eng. 43 (2005) 573. [9] N. Picque, F. Gueye, G. Guelachvili, E. Sorokin, I.T. Sorokina, Opt. Lett. 30 (2005) 3410. [10] V.A. Akimov, V.I. Kozlovskii, Y.V. Korostelin, A.I. Landman, Y.P. Podmar’kov, M.P. Frolov, Quantum Elect. 35 (2005) 425. [11] F.K. Tittel, D. Richter, A. Fried, Mid-infrared laser applications in spectroscopy, in: I.T. Sorokina, K.L. Vodopyanov (Eds.), Solid-State Mid-Infrared Laser Sources, vol. 89, Heidelberg, Berlin, 2003. [12] A. Zakel, G.J. Wagner, W.J. Alford, T.J. Carrig, High-power, rapidly tunable dual-band CdSe optical parametric oscillator, in: C. Denman, I.T. Sorokina, (Eds.), Proceedings of advanced solid-state photonics, vol. 98, OSA, Vienna, Austria, 2005, p. 433.

[13] J.T. Vallin, G.A. Slack, S. Roberts, A.E. Hughes, Phys. Rev. B 2 (1970) 4313. [14] M. Kaminska, J.M. Baranowski, S.M. Uba, J.T. Vallin, J. Phys. C Solid State 12 (1979) 2197. [15] U. Demirbas, A. Sennaroglu, M. Somer, Opt. Mater. 28 (2006) 231. [16] I.T. Sorokina, E. Sorokin, A.D. Lieto, M. Tonelli, R.H. Page, K.I. Schaffers, JOSA B 18 (2001) 926. [17] A.D. Lieto, M. Tonelli, Opt. Laser Eng. 39 (2001) 305. [18] V.I. Kozlovsky, Y.V. Korostelin, A.I. Landman, Y.P. Podmar’kov, M.P. Frolov, Quantum Elect. 33 (2003) 408. [19] K.L. Schepler, R.D. Peterson, P.A. Berry, J.B. McKay, IEEE J. Quantum Elect. 11 (2005) 713. [20] A. Sennaroglu, A.O. Konca, C.R. Pollock, IEEE J. Quantum Elect. 36 (2000) 1199. [21] V.Y. Ivanov, M. Godlewski, A. Szczerbakow, A. Omel’chuk, A. Davydov, N. Zhavoronkov, G. Raciukaitis, Acta Phys. Pol. A 105 (2004) 553. [22] A.V. Podlipensky, V.G. Shcherbitsky, N.V. Kuleshov, V.I. Levchenko, V.N. Yakimovich, M. Mond, E. Heumann, G. Huber, H. Kretschmann, S. Ku¨ck, Appl. Phys. B (2001) 253. [23] M. Mond, D. Albrecht, E. Heumann, G. Huber, S. Ku¨ck, V.I. Levchenko, V.I. Yakimovich, V.G. Shcherbitsky, V.E. Kisel, N.V. Kuleshov, M. Rattunde, J. Schmitz, R. Kiefer, J. Wagner, Opt. Lett. 27 (2002) 1034. [24] E. Sorokin, I.T. Sorokina, Appl. Phys. Lett. 80 (2002) 3289. [25] M. Mond, D. Albrecht, H.M. Kretschmann, E. Heumann, G. Huber, S. Ku¨ck, V.I. Yakimovich, V.G. Shcherbitsky, V.E. Kisel, N.V. Kuleshov, Phys. Stat. Sol. 188 (2001). [26] S.B. Mirov, V.V. Fedorov, K. Graham, I.S. Moskalev, V.V. Badikov, V. Panyutin, Opt. Lett. 27 (2002) 909. [27] I.S. Moskalev, V.V. Fedorov, S.B. Mirov, Opt. Express 12 (2004) 4986. [28] R.D. Peterson, K.L. Schepler, 1.9 lm-fiber-pumped Cr:ZnSe Laser, in: C. Denman, I.T. Sorokina (Eds.), Proceedings of Advanced SolidState Photonics, vol. 98, The Optical Society of America, Vienna, Austria, 2005, p. 236. [29] A.V. Podlipensky, V.G. Shcherbitsky, N.V. Kuleshov, V.P. Mikhailov, V.I. Levchenko, V.N. Yakimovich, L.I. Postnova, V.I. Konstantinov, Opt. Commun. 167 (1999) 129. [30] K. Graham, V.V. Fedorov, S.B. Mirov, M.E. Doroshenko, T.T. Basiev, Y.V. Orlovskii, V.V. Osiko, V.V. Badikov, V.L. Panyutin, Quantum Elect. 34 (2004) 8. [31] A. Sennaroglu, U. Demirbas, A. Kurt, M. Somer, Opt. Mater., in press. [32] D. Findlay, R.A. Clay, Phys. Lett. 20 (1966) 277. [33] J.A. Caird, L.G. DeShazer, J. Nella, IEEE J. Quantum Elect. QE-11 (1975) 874. [34] G. Goetz, H. Zimmermann, H.-J. Schulz, Z. Phys. B 91 (1993) 429. [35] C.I. Rablau, J.O. Ndap, X. Ma, A. Burger, N.C. Giles, J. Electron. Mater. 28 (1999) 678. [36] S. Ku¨ck, Appl. Phys. B 72 (2001) 515. [37] P.F. Moulton, IEEE J. Quantum Elect. 21 (1985) 1582. [38] S.B. Mirov, V.V. Fedorov, K. Graham, I.S. Moskalev, I.T. Sorokina, E. Sorokin, V. Gapontsev, D. Gapontsev, V.V. Badikov, V. Panyutin, IEE Proc. Optoelectron. 150 (2003) 340. [39] E. Sorokin, S. Naumov, I.T. Sorokina, IEEE J. Quantum Elect. 11 (2005) 690. [40] A.V. Podlipensky, V.G. Shcherbitsky, N.V. Kuleshov, V.I. Levchenko, V.N. Yakimovich, M. Mond, E. Heumann, G. Huber, H. Kretschmann, S. Ku¨ck, Appl. Phys. B (2000) 253. [41] I.T. Sorokina, Opt. Mater. 26 (2004) 395.