Optics Communications 281 (2008) 4411–4414

Contents lists available at ScienceDirect

Optics Communications

journal homepage: www.elsevier .com/locate /optcom

Efficient tunable Yb:YAG ceramic laser

Shinki Nakamura a,*, Hiroaki Yoshioka a, Yu Matsubara a, Takayo Ogawa b, Satoshi Wada b

a Department of Media and Telecommunications, Faculty of Engineering, Ibaraki University, 4-12-1 Nakanarusawa, Hitachi, Ibaraki 316-8511, Japanb Solid-State Optical Science Research Unit, Discovery Research Institute, The Institute of Physical and Chemical Research (RIKEN), 2-1 Hirosawa, Wako, Saitama 351-0198, Japan

a r t i c l e i n f o

Article history:Received 21 March 2008Received in revised form 2 May 2008Accepted 2 May 2008

Keywords:Yb:YAGCeramic laserHigh efficiencyHigh powerTunable

0030-4018/$ - see front matter � 2008 Elsevier B.V. Adoi:10.1016/j.optcom.2008.05.009

* Corresponding author. Fax: +81 294 38 7148.E-mail address: nakamura@mx.ibaraki.ac.jp (S. Na

a b s t r a c t

A high-power efficient ceramic Yb:YAG laser was demonstrated at a room temperature of 20 �C with anYb concentration of 9.8 at.%, a gain medium of 1 mm, a pumping power of 13.8 W, an output coupler ofT = 10%, and a cavity length of 20 mm. A 6.8 W cw output power was obtained with a slope efficiency of72%. The ceramic Yb:YAG laser exhibited a continuous tunability in the spectral range of 63.5 nm from1020.1 to 1083.6 nm for T = 1% at a maximum output power of 1.6 W. To the best of our knowledge, thisis the first study of the tunability of ceramic Yb:YAG lasers, except crystal Yb:YAG studies.

� 2008 Elsevier B.V. All rights reserved.

1. Introduction

Ceramic laser media fabricated by vacuum sintering [1–3] andnanocrystalline [3] technology are very attractive materials be-cause they have several remarkable advantages compared withsingle crystal laser materials. Ceramic samples with a large sizecan be easily fabricated, whereas this is extremely difficult for sin-gle crystals; multiplayer and multifunctional ceramic laser materi-als are possible because of the polycrystallinity of ceramics [4].Potentially, because of their short fabrication period and becausethey can be mass-produced, the cost of ceramic laser materialscould be much lower than that of single crystals. Furthermore,no complex facilities and critical techniques are required for thegrowth of ceramics. Since 1995, Ikesue and coworkers have beendeveloping several types of ceramic laser material [1,2], and theyfound in 2000 that the output power from a 3.4 at.% Nd:YAG cera-mic microchip laser is twice that from a Nd:YAG crystal microchiplaser of the same size [5]. At a low doping concentration, it wasfound that the efficiency of a diode-end-pumped Nd:YAG ceramiclaser is even higher than that of a Nd:YAG single crystal laser. Since1998, Yanagitani and coworkers have been developing severaltypes of ceramic lasers, and Lu et al. reported the Nd:YAG ceramiclaser as one of them in 2001 [6]. The mechanical properties of YAGceramics were reported in Ref. [7]. YAG ceramics had a 10% higherhardness than a YAG single crystal, and the fracture toughness ofthe YAG ceramics was more than threefold that of the YAG singlecrystal. Therefore, the ceramics had a higher resistance to thermal

ll rights reserved.

kamura).

shock than the single crystal. Ytterbium (Yb3+)-doped materials arevery attractive for diode-pumped solid-state lasers (DPSSLs) [8].The Yb3+-doped materials have high quantum efficiency and exhi-bit no concentration quenching simply because the Yb3+ ion hasonly two manifolds, namely, the ground state 2F7/2 and the upperlevel 2F5/2. Thus far, many articles about Yb:YAG crystal lasers havebeen published [9–11]. Yb:YAG has broad absorption and emissionbands. The broad absorption band in the near-IR region is suitablefor laser-diode (LD) pumping, and the broad emission band enablesthe generation of ultrashort pulses [10]. However, an Yb:YAG laseris known as a quasi-three-level laser or a quasi-four-level laser,and a finite population exists at the Stark level of the lower mani-fold 2F7/2, where laser transition terminates, which requires high-intensity pumping, a high-brightness pump source, and an efficientheat removal technique [12–14] to prevent reabsorption from thelower level of the laser. Takaichi et al. reported the absorption andemission spectra of a Yb:YAG ceramic (CYb = 1 at.%) and demon-strated laser oscillation, which was the first diode-end-pumpedYb:YAG ceramic laser (not Nd:YAG) with a 345 mW cw outputpower and a slope efficiency of 26% [15]. Recently, Tsunekaneand Taira have demonstrated a high-power diode-edge-pumpedsingle crystal Yb:YAG/ceramic undoped YAG composite microchiplaser [16,17]. Early in 2007, a diode-edge-pumped, composite all-ceramic Yb:YAG (CYb = 10 at.%) microchip laser was demonstratedby Tsunekane and Taira, and a 414 W cw output power was ob-tained with a slope efficiency of 47% [18]. Very recently, Dong etal. have demonstrated a 2.7 W heavily doped (20 at.%) Yb:YAGceramic laser with a slope efficiency of 52% [19]; however, itstwo-pass-pumping miniature laser configuration was more com-plex than a simple conventional end-pumping configuration and

0 10 150

1

2

3

4

5

6

7

8

0 10 150

1

2

3

4

5

6

7

8

Absorbed Pump Power (W)

Out

put P

ower

(W

) T = 10%

T = 1%

T = 5%

slope = 72%η10%T = for

(a) (b)

5 5

Fig. 2. Input–output power dependence of efficient Yb:YAG (CYb = 9.8 at.%) ceramiclaser: (a) with output couplers of T = 1, 5, and 10%, and (b) only T = 10%.

4412 S. Nakamura et al. / Optics Communications 281 (2008) 4411–4414

its output power was not markedly high. Nakamura et al. demon-strated a 5.5 W cw Yb:YAG (9.8 at.%) ceramic laser with a slopeefficiency of 52% using a simple end-pumping scheme [20] witha 400 lm fiber-coupled LD. Dong et al. demonstrated a highly effi-cient (a slope efficiency of 79%) Yb:YAG ceramic laser [21] with a100 lm fiber-coupled LD using an end-pumping scheme, but itsoutput power was 1.7 W.

In this paper, we report a high-power (6.8 W) and high-effi-ciency tunable Yb:YAG ceramic laser demonstrated using an end-pumping scheme with a slope efficiency of 72% at room tempera-ture (20 �C). In the previous reports of Yb:YAG ceramic lasers, nodescription of the tunability of the lasers is given. However, thereare some reports about the tunability of Yb:YAG crystal lasers[22–24]. In 2000, the widest tunability range from 1024.1 to1108.6 nm was demonstrated with a 160 mW Yb:YAG crystal laserusing a 0.1% output coupler and a birefringent filter by Saikawa etal. [23]. Subsequently, Saikawa et al. reported a 180 mW Yb:YAGcrystal laser with a tunability in the spectral range of 59 nm from1021.9 to 1081.2 nm in 2002 [24]. In this study, we investigatedthe tunability of a 1.5 W Yb:YAG ceramic laser using a 1% outputcoupler and a prism, because the ceramic laser has a low cost, isa large size, and enables rapid mass production. The tunable rangeof the watt class Yb:YAG ceramic laser was 63.5 nm from 1020.1 to1083.6 nm.

2. Experimental setup

The experimental setup for the Yb:YAG ceramic laser is shownin Fig. 1. A 940 nm fiber-coupled LD (JENOPTIK Laserdiode, JOLD-30-FC-12) was used as a pumping source, the core diameter ofthe fiber was 200 lm, and the numerical aperture (NA) of the fiberwas 0.22. The pumping beam was focused onto the ceramic with aratio of 1:1 using the lenses L1 (f = 25 mm) and L2 (f = 25 mm). Thediameter of the focused spot on the ceramic was �200 lm. To ob-tain high efficiency and high power, a laser cavity consisting of aflat dichroic mirror (DM) and a flat output coupler (OC) as a linearresonator (without a mirror M and an SF10 prism) was used. TheDM was antireflection (AR)-coated at 940 nm and had a highreflectivity at 1030 nm. The OC was partially-reflection-coatedwith a transmittance of T = 1, 5, and 10% at 1030 nm. An AR-coatedceramic Yb:YAG (CYb = 9.8 at.%, Konoshima Chemical) with dimen-sions of 5 � 10 � 1 mm3 was used. A 1-mm-thick Yb:YAG ceramicplate was wrapped with indium foil and mounted in a water-cooled copper block that acted as a heat sink. Water was main-tained at a room temperature of 20 �C during laser oscillation.The cavity length was 20 mm, which was optimized, as shown inthe later part of this letter. Fig. 2 shows the dependence of the out-put power on the absorbed pump power, which was determined byconsidering the absorption efficiency difference between non-las-ing and lasing cases, i.e., absorption efficiency remains constantwith an increase in pump intensity in the lasing case [12,13]. Final-ly, a tunable laser setup with a V-shape cavity including a concave

LD

Yb:YAG ceramics

DML1 L2

(AR coated)

@9 40 nm

Fiber

OCSF10 Prism

M

Fig. 1. Experimental setup for tunable Yb:YAG ceramic laser. LD: fiber-coupled d-iode laser; L1, L2: focusing lenses; DM: flat dichroic mirror; M; a concave mirror(ROC = 250 mm); OC: output coupler.

mirror M (radius of curvature, ROC = 250 mm) and an SF10 disper-sive prism is shown in Fig. 1. The SF10 dispersive prism is insertedinto the V-shape resonator as the tuning element between thefolded mirror M and the output coupler OC at Brewster’s angle.The cavity length was 315 mm.

3. Experimental results

Fig. 2a shows the output power as function of the absorbedpump power in the cases for the three transmittances of the outputcouplers T = 1, 5, and 10%, and Fig. 2b shows the output power asfunction of the absorbed pump power only for the case ofT = 10%. The absorbed pump powers at the lasing threshold were1.2, 2.0, and 2.3 W, and the maximum output powers of 6.9, 6.9,and 6.8 W for T = 1, 5, and 10%, respectively, were obtained atthe absorbed pump power of 13.8 W. The round trip loss L in theresonator was estimated to be 0.09 by the lasing thresholds andthe reflectivity of the output couplers [25], which resulted in asmall signal gain g0 of 2.0 cm�1, and a single pass gain G of 1.2 withthe 1 mm thick gain medium. Each linear line was fit in Fig. 2a forT = 1, 5, and 10%. The slope efficiencies gslope were 60, 64, and 72%for T = 1, 5, and 10%, respectively. Since we considered that T = 10%is best for obtaining the highest slope efficiency of 72%, we filledthe data for the T = 10% case to Fig. 2b. The maximum outputpower of 6.8 W for T = 10% was obtained at the absorbed pumppower of 13.8 W, indicating that the efficiency of convertingpumping optical power to output optical power, gopt–opt, was49%. The line of the best fit is shown in Fig. 2b. The slope efficiencygslope was 72% for T = 10%. The maximum output power of 6.8 Wwas determined to be fourfold higher than 1.7 W and the slope effi-ciency was determined to be 7% lower than 79% using the 100 lmfiber-coupled LD reported by Dong et al. [21]. Our 6.8 W laser withthe slope efficiency of 72% is expected to have a higher slope effi-ciency than the present result if the pumping source is replacedwith a 100 lm fiber-coupled 25 W LD, for example, LIMO25-F100-DL940 (Lissotschenko Mikrooptick) while maintaining thehigh output power, because the pumping intensity would increaseto a value of fourfold higher than that of a 200 lm fiber-coupledLD. In comparing our laser with the edge-pumped compositeYb:YAG ceramic laser [18] developed by Tsunekane and Taira, welimit our discussion to the cw case; the laser power of 414 W ob-tained by Tsunekane and Taira is much higher than our result,but their slope and optical–optical conversion efficiency were 47and 44%, which were 25 and 5% lower than our slope and opti-cal–optical conversion efficiency of 72 and 49%, respectively. The

Fig. 3. Transverse intensity profile of the Yb:YAG ceramic laser beam. The intensitydistribution indicates that the beam is a Gaussian beam (a TEM00 mode beam).

0

1

2

3

4

5

6

1000 1020 1040 1060 1080 1100

Laser Wavelength (nm)

Out

put

Pow

er (

W)

T = 0.1% T = 1% T = 5% T = 10%

Fig. 5. Output power of ceramic Yb:YAG tunable laser as function of oscillationwavelength for various output couplers.

S. Nakamura et al. / Optics Communications 281 (2008) 4411–4414 4413

transverse intensity profile of the Yb:YAG ceramic laser beam isshown in Fig. 3. The intensity distribution indicates that the beamis a Gaussian beam (a TEM00 mode beam). The beam image in Fig. 3was as stable as the pumping LD and we found no amplitudeinstability.

These results of high output power, high efficiency, and goodbeam quality were obtained after the optimization of the cavitylength. The cavity length was varied to obtain an optimum valuefor the highest efficiency and highest output power, and the focallength of the thermal lens for designing a tunable laser cavity con-figuration. Fig. 4 shows the maximum output power as a functionof the cavity length. Fig. 4 shows that the optimum cavity length isless than 20 mm. This value is the appropriate cavity length for ourlaser, because there is no space to reduce the cavity length lessthan 20 mm. When we used the 400 lm fiber-coupled LD [20],the optimum cavity length with the highest output power andhighest slope efficiency was 25 mm, and reducing the length lessthan 25 mm yielded a worse result.

The focal length of the thermal lens in the ceramic Yb:YAG platewas considered for designing a tunable laser cavity configuration.

0

1

2

3

4

5

6

7

8

0 20 40 60 80 100 120 140

Cavity Length (mm)

Out

put

Pow

er (

W)

T = 1%

T = 5%

T = 10%

Fig. 4. Maximum output power as function of cavity length of efficient Yb:YAGceramic laser.

Fig. 4 also shows that the focal length of the thermal lens is109 mm [120 mm (the cavity length) minus 11 mm (the distanceof the ceramic Yb:YAG and the DM)], because the cavity becomesunstable, terminating the laser oscillation when the Fabry–Perotcavity length exceeds the thermal lens focal length. By consideringthis thermal lens, a tunable laser with a v-shape cavity including aconcave mirror M (radius of curvature, ROC = 250 mm) and anSF10 dispersive prism was obtained, as shown in Fig. 1. The SF10dispersive prism is inserted into the V-shape resonator as the tun-ing element between the folded mirror M and the output couplerOC at Brewster’s angle. The cavity length is 315 mm. Fig. 5 showsthe dependence of the output power versus the laser oscillationwavelength for the output couplers of T = 0.1, 1, 5, and 10%, whenabsorbed pump power was 13.8 W. There are two separate steeppeaks (1031.7 and 1049.0 nm) for T = 10% and there is no oscilla-tion wavelengths from 1037 to 1048 nm, indicating the loss inthe resonator overcomes the relatively small gain of the Yb:YAGceramics in the region. The tunable ranges from 1022.2 to1036.8 nm and from 1048.6 to 1051.2 nm with a maximum outputpower of 5.23 W for T = 10% were obtained. This tendency is alsoobserved in the case of T = 5% for the same reason. There are twopeaks, but these are continuously connected with a small dip be-tween them. The tunable range was 41.1 nm from 1020.1 to1061.2 nm with a maximum output power of 4.14 W for T = 5%.In the cases of T = 1 and 0.1%, the tuning curves are continuous,smooth and flat, which is due to the small cavity loss, though a gainis relatively small in the spectral region from 1037 to 1048 nm. Thetunable range was 63.3 nm from 1027.4 to 1090.7 nm with a max-imum output power of 0.12 W for T = 0.1%. We achieved a quasi-continuous and smooth tuning range of 63.5 nm, from 1020.1 to1083.6 nm, with a maximum output power of 1.61 W near1050 nm for T = 1%. To the best of our knowledge, this is the firststudy of the tunability of an Yb:YAG ceramic laser. The shortestwavelength of the tuning range in Fig. 5 is limited to the dichroiccoating range of the pumping mirror (high-reflection coating from1020 to 1200 nm, Layertec, No. 103542). This 63.5 nm tuning rangeof the Yb:YAG ceramic laser at 20 �C or 293 K is 1.76-fold broaderthan the 36.0 nm tuning range, from 1018 to 1054 nm, which wasproduced from the Yb:YAG crystal laser with a three-plate birefrin-gent filter at 218 and 248 K [22]. Furthermore, our tuning range of63.5 nm from 1020.1 to 1083.6 nm with the high-power ceramicYb:YAG laser at 20 �C is broader than that of a low-power crystalYb:YAG laser by Saikawa et al. [23], which has the tuning rangeof 59 nm from 1022 to 1081 nm at 18 �C. The widely tunableYb:YAG crystal laser with birefringent filters reported by Saikawa

4414 S. Nakamura et al. / Optics Communications 281 (2008) 4411–4414

et al. [24] had the tuning range of 84.5 nm from 1024.1 to1108.6 nm; however, the highest output power was 180 mW,which is much lower than the maximum output power of 1.6 Win our ceramic laser with an SF10 prism in Fig. 5. By comparingthe output powers of the cases with and without a dispersive tun-ing element, the maximum output power of the laser was deter-mined to be 1.6 W in Fig. 5 and 6.8 W in Fig. 2. The decrease inlaser output power is due to the insertion loss of the prism andthe mode-volume mismatch loss between pump beam and laserbeam in the 315 mm long cavity configuration with the armincluding prism, comparing the linear 20 mm short cavity laserused to measure the output power in Fig. 2, which can be improvedby optimizing the resonator design. For example, the laser beamwaist comes to the place of the ceramic Yb:YAG even using a longcavity when the resonator type is improved from the V-shape cav-ity to a symmetric Z-shape or X-shape cavity.

4. Summary

A diode-end-pumped high-efficiency high-power Yb:YAG cera-mic laser was demonstrated at a room temperature of 20 �C withan Yb concentration of 9.8 at.%, a gain medium thickness of1 mm, a pumping power of 13.8 W, an output coupler of T = 10%,and a cavity length of 20 mm. A 6.8 W cw output power was ob-tained with a slope efficiency of 72%. This is the relatively high effi-ciency of ceramic Yb:YAG lasers at room temperature. The beamquality was shown as a transverse intensity distribution indicatinga Gaussian beam (a TEM00 mode beam). The tunable range of63.5 nm from 1020.1 to 1083.6 nm for T = 1% was also obtainedat room temperature and the highest output power was 1.6 W.To the best of our knowledge, this is the first study of the tunabilityof ceramic Yb:YAG lasers, except crystal Yb:YAG studies. This tun-ability could be very attractive for femtosecond laser applicationsand suitable for a laser amplifier (the maximum single pass gainof 1.2 with the gain medium of 1 mm thickness). The cost ofYb:YAG ceramic laser materials could be much lower than that ofsingle crystals because of their high-speed and large-size produc-tion, and mass-production, which could be tremendously attrac-tive for industrial applications.

Acknowledgements

The authors thank Dr. Masaki Tsunekane of Japan Science andTechnology Agency (JST) and the Institute for Molecular Science

(IMS) for fruitful discussions. This work was partially supportedby Nippon Sheet Glass Foundation for Materials Science and Engi-neering, and by Mr. Eiichi Matsui.

References

[1] A. Ikesue, T. Kinoshita, K. Kamata, K. Yoshida, J. Am. Ceram. Soc. 78 (1995)1033.

[2] A. Ikesue, T. Kinoshita, K. Yoshida, J. Am. Ceram. Soc. 79 (1996) 1921.[3] J. Lu, M. Prabhu, J. Song, C. li, J. Xu, K. Ueda, A.A. Kaminskii, H. Yagi, T.

Yanagitani, Appl. Phys. B71 (2000) 469.[4] T. Taira, IEEE J. Sel. Top. Quantum Electron. 13 (2007) 798.[5] I. Shoji, S. Kimura, Y. Sato, T. Taira, A. Ikesue, K. Yoshida, Appl. Phys. Lett. 77

(2000) 939.[6] J. Lu, T. Murai, K. Takaichi, T. Uematsu, K. Misawa, M. Prabhu, J. Xu, K. Ueda, T.

Yanagitani, H. Yagi, A.A. Kaminskii, A. Kudryashov, Appl. Phys. Lett. 78 (2001)3586.

[7] A.A. Kaminskii, M.Sh. Akchurin, V.I. Alshits, K. Ueda, K. Takaichi, J. Lu, T.Uematsu, M. Musha, A. Shirakawa, V. Gabler, H.J. Eichler, H. Yagi, T. Yanagitani,S.N. Bagayev, J. Fernandez, R. Balda, Crystallogr. Rep. 48 (2003) 515.

[8] W.F. Krupke, IEEE J. Sel. Top. Quantum Electron. 6 (2000) 1287.[9] P. Lacovara, H.K. Choi, C.A. Wang, R.L. Aggarwal, T.Y. Fan, Opt. Lett. 16 (1991)

1089.[10] C. Hönninger, G. Zhang, U. Keller, A. Giesen, Opt. Lett. 20 (1995) 2402.[11] D.S. Sumida, H. Bruesselbach, R.W. Byren, M. Mangir, R. Reeder, Proc. SPIE 3265

(1998) 100.[12] S. Matsubara, T. Ueda, S. Kawato, T. Kobayashi, Jpn. J. Appl. Phys. 46 (2007)

L132.[13] Y. Sato, T. Taira, IEEE J. Quantum Electron. 40 (2004) 270.[14] T. Shoji, S. Tokiwa, J. Kawanaka, M. Fujita, Y. Izawa, Jpn. J. Appl. Phys. 43 (2004)

L496.[15] K. Takaichi, H. Yagi, J. Lu, A. Shirakawa, K. Ueda, T. Yanagitani, Phys. Status

Solidi A 200 (2003) R5.[16] M. Tsunekane, T. Taira, Jpn. J. Appl. Phys. 44 (2005) L1164.[17] M. Tsunekane, T. Taira, Opt. Lett. 31 (2006) 2003.[18] M. Tsunekane, T. Taira, Appl. Phys. Lett. 90 (2007) 121101.[19] J. Dong, A. Shirakawa, K. Ueda, H. Yagi, T. Yanagitani, A.A. Kaminskii, Opt. Lett.

32 (2007) 1890.[20] S. Nakamura, Y. Matsubara, T. Ogawa, S. Wada, Jpn. J. Appl. Phys. 47 (2008)

2149.[21] J. Dong, A. Shirakawa, K. Ueda, H. Yagi, T. Yanagitani, A.A. Kaminskii, Appl.

Phys. Lett. 89 (2006) 091114.[22] U. Brauch, A. Giesen, M. Karszewski, Chr. Stewen, A. Voss, Opt. Lett. 20 (1995)

713.[23] J. Saikawa, S. Kurimura, I. Shoji, T. Taira, Opt. Mater. 19 (2002) 169.[24] J. Saikawa, S. Kurimura, N. Pavel, I. Shoji, T. Taira, Advanced solid-state lasers,

in: H. Ingeyan, U. Keller, C. Marshall (Eds.), OSA Trends in Optics andPhotonics, vol. 34, OSA, Washington DC, 2000, p. 106.

[25] W. Koechner, Solid-State Laser Engineering, third ed., Springer, Berlin, 1992.