Open Access
Add to BibSonomyAbstract
An 8 W continuous wave linearly-polarized single-frequency 1014.8 nm fiber amplifier working at room temperature is developed with commercial double-clad single-mode Yb-doped silica fiber. Re-absorption at the laser wavelength and amplified spontaneous emission at longer wavelength are managed by optimizing the amplifier design. The laser has a linewidth of ~24 kHz without noticeable broadening after amplification. Using two resonant cavity frequency doublers, 1.03 W laser at 507.4 nm and 75 mW laser at 253.7 nm are generated with 4 W 1014.8 nm laser. Both absorption and saturated absorption spectra of the 1S0 – 3P1 transition of atomic mercury are measured with the 253.7 nm laser.
©2013 Optical Society of America
1. Introduction
Atom cooling and trapping is of great importance in precision measurement, quantum simulation and quantum computation with neutral atoms or ions. Mercury, being the heaviest nonradioactive atom that has been laser-cooled and trapped, can be used for the research on the variation of fundamental constants and CP-violation (CP stands for charge parity) of permanent electric dipole moments [1, 2]. Furthermore, the low sensitivity of black body radiation makes the mercury atom probably the best candidate of optical lattice clock, which has the potential to achieve uncertainty of 10−18 level, much more accurate than the microwave atomic clocks [2].
Laser cooling of mercury atoms requires narrow linewidth tunable lasers at 253.7 nm, which corresponds to the 1S0 - 3P1 transition. 253.7 nm is also one of the fundamental wavelengths for four-wave sum frequency mixing in mercury vapor, to generate continuous wave vacuum-UV radiation around 121 nm for future laser cooling of trapped anti-hydrogen [3]. Several 1014.8 nm lasers based on crystal or semiconductor materials have been reported to generate the 253.7 nm laser by frequency quadrupling [1, 4, 5]. The highest power reported at 1014.8 nm to date is a 7 W Yb:YAG disk lasers by Petersen et al. [2].
Compared with bulk solid state lasers, fiber lasers usually have better beam quality, higher efficiency and better robustness. The invention of double-clad fibers and progresses in diode lasers give fiber lasers excellent power scalability in the past decade [6]. In 2006, Seifert et al. reported a two-stage amplifier at 1014.8 nm with Yb-doped double-clad fiber, which generated up to 5 W laser with a linewidth of less than 3 MHz [7]. However, the Yb-doped fibers were cooled to liquid-nitrogen temperature, because it has significant absorption at 1014.8 nm at room temperature. The absorption facilitates the amplified spontaneous emission (ASE) at longer wavelength (about 1030 ~1040 nm) where the emission cross section is larger. The ASE would deplete the population in excited state and limit the efficient amplification of the signal light. A high power 1014.8 nm fiber amplifier working at room temperature is highly favored for reduced complexity.
Furthermore, laser around 1015 nm has several other applications. For example, laser cooling of Yb-doped solids by means of anti-Stokes fluorescence [8, 9] requires laser at 1010 ~1025 nm. Narrow linewidth laser at 1014 nm, when frequency doubled to 507 nm, is the clock laser in optical lattice clocks based on the 1S0 - 3P2 transition of ytterbium atoms [10].
In this paper, we report all-fiber-connectorized narrow linewidth 1014.8 nm amplifiers with Yb-doped polarization-maintaining (PM) double-clad (DC) single-mode fibers as gain media working at room temperature. Up to 8.06 W 1014.8 nm continuous-wave linearly-polarized single-transverse-mode laser output is achieved by single-stage amplification. The linewidth is measured to be ~24 kHz by the self-heterodyne method, without noticeable broadening compared with the seed. The laser is then frequency quadrupled using two resonant cavity frequency doublers. Up to 1.03 W green laser at 507.4 nm and further 75 mW ultraviolet laser at 253.7 nm are generated from 4 W fundamental laser input. Both absorption and saturated absorption spectra of the 1S0 – 3P1 transition of atomic mercury have been measured using the laser, verifying the applicability of the room temperature 1014.8 nm Yb-doped fiber amplifier for mercury cooling.
2. Experimental setup
The configuration of the 1014.8 nm diode master oscillator fiber power amplifier is shown in Fig. 1. An external cavity diode laser with a tapered amplifier at 1014.8 nm (Toptica TA Pro) works as a seed laser, which is linearly polarized and has an output spot of oval shape. Before the light is coupled into a PM single mode fiber (PM 980), two cylindrical lens are used for beam reshaping. Yb-doped PM DC fiber with a core diameter of 10 μm, a numerical aperture of 0.075, and an inner clad diameter of 125 μm is used as gain media, which has a nominal absorption of 4.8 dB/m at 976 nm. The pump diode lasers around 975 nm are injected into the inner clad of the gain fiber by a (2 + 1) × 1 pump and signal combiner. The unabsorbed pump light is damped with a pump stripper, and the fiber output end is angle cleaved to avoid back reflection. The whole amplifier is fiber-connectorized and polarization-maintaining. The output spectra are analyzed with an optical spectrum analyzer (AQ6370, YOKOGAWA). A bandpass filter is used to remove the ASE from the amplifier output, whose transmission spectrum is shown in the inset in Fig. 1. The amplifier output is then frequency doubled to 507.4 nm and quadrupled to 253.7 nm by two resonant cavity frequency doublers with LBO and BBO crystal, respectively. After that, the generated UV laser is used to measure the absorption and saturated absorption spectra of the 1S0 – 3P1 transition of atomic mercury.
Fig. 1 A schematic diagram of the experimental setup for the 1014.8 nm fiber amplifier, two resonant cavities for frequency doubling and quadrupling, and measurements of absorption and saturated absorption spectra. The inset at up-right corner shows the ASE filter’s transmission spectrum.
3. Results and discussion
Usually Yb-doped silica fibers are considered as homogeneously broadened gain medium at room temperature. In homogeneously broadened gain medium, gain at one wavelength is determined by the gain at two other wavelengths [11].The relation depends on the emission and absorption cross sections of three wavelengths, as well as the fiber geometry. According to the room temperature emission and absorption spectra provided by the fiber manufacturer, the emission cross sections at 976, 1015, and 1035 nm are 1.71, 0.68, and 0.70 pm2, respectively, and the absorption cross sections are 1.77, 0.102, and 0.0407 pm2, respectively. According to [11], the gain at 1015 nm, G1015, is related to gain at 1035 nm, G1035, and absorption (negative gain) at 976 nm, αP, by
The fiber amplifier is limited by ASE at ~1035 nm, which is believed to become strong at ~40 dB of gain and will grow rapidly for higher gain. Therefore, to achieve a higher gain at 1015 nm with ASE under control, fiber with a smaller clad-to-core area ratio is favorable. We have chosen Yb-doped fiber with a core diameter of 10 μm and cladding diameter of 125 μm, because single mode operation for laser at 1 μm is not guaranteed in commercial double-clad fibers with a core diameter larger than 10 μm.As seen in Eq. (1), another way to achieve higher gain at laser wavelength is to use shorter fiber, because the pump absorption αp is less. It is a valid solution if the laser efficiency is not a top priority as in our case. We have tested amplifier performance with varying gain fiber length, by shortening the fiber gradually. Figure 2(a) shows the direct output spectra without ASE filtering for gain fiber length of 1.45, 1.2, and 0.85 m, respectively, at a pump power of 40 W and a seed power of 30 mW. With shorter fiber, the pump absorption is less, therefore, the total output is lower. But the proportion of ASE in the total output decreases. Figure 2(b) shows the total output and the laser output without ASE as functions of the fiber length. While the total output power decreases as the gain fiber gets shorter, the signal power increases instead and peaks around 1 m.
Fig. 2 (a) Spectra of the amplifier’s direct output with gain fiber length of 1.45, 1.2, and 0.85 m, respectively, at 40 W pump and 30 mW seed. (b) The total output and the laser output with ASE being filtered out as functions of the fiber length.
The influence of seed power on fiber amplifier output was investigated as well. The output spectrum doesn’t change significantly as the seed power increases from 30 to 60 and 100 mW, while the laser output power rises from 6.73 to 7.72 and 8.06 W with a gain fiber length of 1.05 m and pump power of 44.8 W. The output power saturates as the seed power is higher than the saturation power of the fiber at 1014.8 nm, which is calculated to be ~31 mW.
Figure 3(a) shows the laser output and forward ASE power as functions of the pump power, when the Yb fiber length is 1.05 m and the seed laser power is 100 mW. The laser power at 1014.8 nm is measured by inserting a bandpass filter after the output, and the forward ASE power is calculated by subtracting the total output with the laser output power. Both increase nonlinearly due to the red shift of the pump diode wavelength with power. The diode wavelength approaches to the Yb absorption peak of 976 nm at the highest power. Maximum 8.06 W laser output is obtained with 44.8 W pump, which corresponds to an optical efficiency of 18%. The low efficiency is due to the low pump absorption for the short 1.05 m fiber length and the strong ASE. The forward ASE reaches about 6.6 W at the highest pump power. And one could expect even higher ASE power at the backward direction.
Fig. 3 (a) Laser output and forward ASE power as functions of the pump power. (b) The spectrum of laser output after the bandpass filter. The inset shows the corresponding spectrum of the direct output. The gain fiber is 1.05 m long, and the seed power is 100 mW.
The forward ASE peaks at ~1035 nm and is ~24 dB lower than the signal, measured with a resolution of 0.02 nm, as shown in the inset of Fig. 3(b). By inserting the bandpass filter, the ASE becomes ~37 dB lower than the laser [Fig. 3(b)] and negligible in power. The ASE level could be reduced further by adding more filters. Moreover, the following two resonant cavity frequency doublers are effective bandpass filters. Therefore, only the local ASE at the laser wavelength of 1014.8 nm cannot be removed, which is ~47 dB below the signal.
Because all the fiber and fiber components are polarization maintaining, the amplifier output is linearly polarized. The polarization extinction ratio of the laser output is measured to be 15 dB by a Glan-Laser calcite polarizer. The linewidth of the seed and amplified laser are measured by the self-heterodyne method, using a fiber pigtailed acousto-optic modulator with a carrier frequency shift of 150 MHz and 10 km HI1060 fiber as the delay line. The resulting heterodyne rf-spectra are shown in Fig. 4. The linewidth of the amplifier output is ~24 kHz, determined by Lorentzian fitting. No linewidth broadening during amplification is observed.
Fig. 4 Delayed self-heterodyne rf-spectra of the seed and amplified radiation (100 times averaged), and their Lorentzian fitting curves.
The amplified 1014.8 nm laser is frequency doubled to 507.4 nm in a LBO crystal and quadrupled to 253.7 nm in a BBO crystal with two external cavity frequency doublers (Wavetrain from Spectra-Physics). Both resonant cavities have similar configuration and are formed by two mirrors and one prism, as schematically illustrated in Fig. 1. Standard Pound-Drever-Hall method is used to lock the resonant cavities. Electro-optic phase modulators are inserted between the fiber amplifier and the doubling cavity and the quadrupling cavity to generate error signals. With a maximum allowed input fundamental power of 4 W (specified by the manufacturer), 1.03 W green laser at 507.4 nm is generated with a conversion efficiency of 25.7%, and 75 mW ultraviolet laser at 253.7 nm is obtained with a conversion efficiency of 7.3%, as shown in Fig. 5. We believe the power and efficiency can be improved in the future by customizing the resonant cavities.
Fig. 5 (a) The second harmonic (507.4 nm) power and conversion efficiency as functions of the fundamental infrared laser power. (b) The forth harmonic (253.7 nm) power and conversion efficiency as functions of the 507.4 nm laser.
To verify the applicability of the fiber based 253.7 nm laser for laser cooling of mercury atoms, absorption experiments of atomic mercury are performed [12] at room temperature, using a 2 mm thick quartz mercury vapor cell with natural abundance of the isotopes. As illustrated in Fig. 1, two laser beams with a power of ~0.05 mW are used as probe and reference beam. A ~1 mW counter-propagating pump laser beam overlaps with the probe beam in the cell, and the angle between the probe and pump beam is kept as small as possible. With the reference beam only, the absorption spectrum of the 1S0 – 3P1 transition in the mercury vapor cell is measured (with a power of ~0.5 mW), as shown in Fig. 6(a). Five Doppler broadened absorption peaks for the mercury isotopes are clearly visible. By subtracting the two signals detected from the probe and reference beams passed through the mercury cell, the Doppler-free spectrum of the transition could be obtained if the power of two beams are identical. Figure 6(b) shows the saturated absorption spectrum including the 1S0 – 3P1 transitions of 204Hg, 199Hg(1/2) and 201Hg(5/2). The Doppler background (the crest base) isn’t completely removed, due to the ~8% imbalance between the probe and pump beam because of the mirror blank, but three absorption peaks corresponds to three isotopes are resolved. The measurements show that the linewidth of the 253.7 nm laser is less than the natural linewidth of the mercury 1S0 – 3P1 transition, which is 1.3 MHz, and the fiber amplifier based laser is applicable for laser cooling of mercury atoms.
Fig. 6 (a) The absorption spectrum of the 1S0 – 3P1 transition of atomic mercury. (b) The saturated absorption spectroscopy of the 1S0 – 3P1 transition of atomic mercury. The Doppler background isn’t completely removed. Each peak is marked with the corresponding isotope(s).
4. Conclusion
We have investigated 1014.8 nm Yb-doped fiber amplifiers working at room temperature for laser cooling of mercury atoms after frequency quadrupling to 253.7 nm. Yb-doped fiber has a significant absorption at 1014.8 nm at room temperature, and amplification at this wavelength is limited by the amplified spontaneous emission at longer wavelength. Therefore, in the past, it is considered necessary to cool the Yb doped fiber to liquid-nitrogen temperature for efficient amplification of a low power laser at 1014.8 nm. We report that, by properly designing the fiber amplifier, 8.06 W single-frequency single-mode linearly-polarized laser output was obtained at room temperature with a seed laser power of 100 mW. The linewidth was measured to be ~24 kHz by the self-heterodyne method without noticeable broadening compared with the seed. The residual ASE light was 37 dB below the laser spectrally when measured with a resolution of 0.02 nm. By two resonant cavity frequency doublers, 4 W infrared laser was frequency doubled to 1.03 W green laser, and quadrupled to 75 mW UV laser at 253.7 nm. The UV power was limited by the allowed input of the doubling cavities. With the developed laser, both absorption and saturated absorption spectra of the 1S0 – 3P1 transition of atomic mercury were obtained. This proves that the fiber amplifier based laser is suitable for laser cooling of mercury atoms.
Acknowledgments
The work is supported by Hundred Talent Program of Chinese Academy of Sciences, Research Project of Shanghai Science and Technology Commission (Grant NO. 09DJ1400700), and the National Basic Research Program of China (Grand NO. 2011CB921504).
References and links
1. H. Hachisu, K. Miyagishi, S. G. Porsev, A. Derevianko, V. D. Ovsiannikov, V. G. Pal’chikov, M. Takamoto, and H. Katori, “Trapping of neutral mercury atoms and prospects for optical lattice clocks,” Phys. Rev. Lett. 100(5), 053001 (2008). [CrossRef] [PubMed]
2. M. Petersen, R. Chicireanu, S. T. Dawkins, D. V. Magalhães, C. Mandache, Y. Le Coq, A. Clairon, and S. Bize, “Doppler-free spectroscopy of the 1S0-3P0 optical clock transition in laser-cooled fermionic isotopes of neutral mercury,” Phys. Rev. Lett. 101(18), 183004 (2008). [CrossRef] [PubMed]
3. G. Gabrielse, R. Kalra, W. S. Kolthammer, R. McConnell, P. Richerme, D. Grzonka, W. Oelert, T. Sefzick, M. Zielinski, D. W. Fitzakerley, M. C. George, E. A. Hessels, C. H. Storry, M. Weel, A. Müllers, J. Walz, and ATRAP Collaboration, “Trapped antihydrogen in its ground state,” Phys. Rev. Lett. 108(11), 113002 (2012). [CrossRef] [PubMed]
4. J. Paul, Y. Kaneda, T. L. Wang, C. Lytle, J. V. Moloney, and R. J. Jones, “Doppler-free spectroscopy of mercury at 253.7 nm using a high-power, frequency-quadrupled, optically pumped external-cavity semiconductor laser,” Opt. Lett. 36(1), 61–63 (2011). [CrossRef] [PubMed]
5. M. Scheid, F. Markert, J. Walz, J. Wang, M. Kirchner, and T. W. Hänsch, “750 mW continuous-wave solid-state deep ultraviolet laser source at the 253.7 nm transition in mercury,” Opt. Lett. 32(8), 955–957 (2007). [CrossRef] [PubMed]
6. D. J. Richardson, J. Nilsson, and W. A. Clarkson, “High power fiber lasers: current status and future perspectives [Invited],” J. Opt. Soc. Am. B 27(11), B63–B92 (2010). [CrossRef]
7. A. Seifert, M. Sinther, T. Walther, and E. S. Fry, “Narrow-linewidth, multi-Watt Yb-doped fiber amplifier at 1014.8 nm,” Appl. Opt. 45(30), 7908–7911 (2006). [CrossRef] [PubMed]
8. M. Sheik-Bahae and R. I. Epstein, “Optical refrigeration,” Nat. Photonics 1(12), 693–699 (2007). [CrossRef]
9. D. V. Seletskiy, S. D. Melgaard, S. Bigotta, A. Di Lieto, M. Tonelli, and M. Sheik-Bahae, “Laser cooling of solids to cryogenic temperatures,” Nat. Photonics 4(3), 161–164 (2010). [CrossRef]
10. S. G. Porsev and A. Derevianko, “Hyperfine quenching of the metastable 3P0,2 states in divalent atoms,” Phys. Rev. A 69(4), 042506 (2004). [CrossRef]
11. J. Nilsson, J. D. Minelly, R. Paschotta, A. C. Tropper, and D. C. Hanna, “Ring-doped cladding-pumped single-mode three-level fiber laser,” Opt. Lett. 23(5), 355–357 (1998). [CrossRef] [PubMed]
12. W. Demtröder, Laser Spectroscopy: Vol. 1: Basic Principles (Springer, 2008).
