Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

High-power ultra-broadband frequency comb from ultraviolet to infrared by high-power fiber amplifiers

Open Access Open Access

Abstract

A high-power ultra-broadband frequency comb covering the spectral range from ultraviolet to infrared was generated directly by nonlinear frequency conversion of a multi-stage high-power fiber comb amplifier. The 1030-nm infrared spectral fraction of a broadband Ti:sapphire femtosecond frequency comb was power-scaled up to 100 W average power by using a large-mode-area fiber chirped-pulse amplifier. We obtained a frequency-doubled green comb at 515 nm and frequency-quadrupled ultraviolet pulses at 258 nm with the average power of 12.8 and 1.62 W under the input infrared power of 42.2 W, respectively. The carrier envelope phase stabilization was accomplished with an ultra-narrow line-width of 1.86 mHz and a quite low accumulated phase jitter of 0.41 rad, corresponding to a timing jitter of 143 as.

©2012 Optical Society of America

1. Introduction

High-power ultraviolet femtosecond pulses with high repetition rate are desired for a variety of applications such as precision spectroscopy [13], biomedicine [4], laser material processing [5], and nanotechnology [6]. Tunable femtosecond pulses with milliwatt-level average power in the ultraviolet region have been demonstrated by intracavity doubling of optical parametric oscillator [7,8]. Using an enhancement resonant cavity with the cavity length fine controlled by a servo loop, enhanced frequency conversion efficiency could be achieved without the increase of pump power [9]. However, both the intracavity and enhancement resonant methods suffer from the complexity of crucial alignment and electronic feedback control. Due to the high peak intensities caused by femtosecond pulses and minor limitation of phase matching bandwidth, nonlinear frequency conversions are typically very efficient for ultrashort pulses. Therefore, extracavity frequency-quadrupling by direct nonlinear optical conversion of high-power infrared femtosecond pulses provides a straightforward way to achieve ultraviolet laser pulses [10].

Fiber lasers have attracted much attention for the generation of high-power carrier-envelope-phase-controlled ultrashort pulses as practical alternatives to solid-state lasers on account of the excellent beam quality, capability of high energy-extraction efficiency and low phase noise [1113]. Cladding pumped large-mode-area photonic crystal fiber (PCF) enables power scaling for ultrashort pulses with high pump absorption efficiency and low detrimental nonlinear effects such as stimulated Raman and Brillouin scattering of which the thresholds are relatively high in the large mode area (LMA) [14]. A mode-locked Yb-fiber similariton laser oscillator has been recently power-scaled by a cladding-pump chirped-pulse fiber amplifier for realization of a record high-peak-intensity frequency comb to ionize noble gases [15]. The linear amplification process in the LMA fiber chirped-pulse amplification system introduces very little amplitude and phase noise and is of great benefit for scaling optical frequency combs to high average powers [16,17].

In this paper, we demonstrate the generation of ultraviolet and green laser pulses by extracavity frequency quadrupling and doubling the high-power infrared pulses from the cascade fiber amplifiers. The nonlinear conversion process of the fourth harmonic generation is efficient due to the femtosecond pulse-width and high nonlinear optical coefficient of beta-Barium-Borate (BBO) crystals. The stable carrier envelope phase (CEP) of infrared pulses was achieved by modulating the pump power of the Ti:sapphire femtosecond laser oscillator via acoustic optical modulator (AOM). Comparison between the conventional self-referenced scheme and our cross-referenced f-2f scheme was made to characterize the noise properties of the locked offset frequency during the amplification process. The results of these two measurements had very little difference, indicating negligible amplitude-to-phase noise in the linear fiber chirped-pulse amplifiers.

2. Experimental setup

The experimental setup for frequency quadrupling of a high-power fiber amplifier and CEP controlling is shown in Fig. 1 , which consisted of a Ti:sapphire femtosecond oscillator, four-stage fiber amplifiers, CEP controlling and frequency quadrupling scheme. The Ti:sapphire oscillator used for this work generated 10-fs pulses at 79.5 MHz repetition rate with an average output power of 200 mW. The oscillator had a broadband spectrum that extended from 650 to 1100 nm around the center wavelength of 800 nm. The infrared spectral fraction of the oscillator was filtered and coupled into a guiding fiber by microscope objective with an average power of 0.32 mW, ranging from 980 to 1080 nm as the seed for the cascaded fiber amplifiers. The pre-amplifier consisted of two-stage Yb-doped single-mode fiber (YbDF 350, OFS) amplifiers pumped by fiber-pigtailed laser diodes of 400 mW at 976 nm. The gain fiber used for the first and second stage amplifier was 0.6 and 1.5 m in length, respectively. Behind the fiber pre-amplifier, the average power of the infrared laser pulses reached 120 mW with the pulse duration of 10 ps. The power amplifier was constructed with 1.0 and 1.6 m LMA Yb-doped PCFs for the first and second, respectively. The gain fiber had a core diameter of 40μm (NA = 0.03) and inner cladding diameter of 200 μm (NA = 0.55) with the pump absorption of 10 dB/m at 976 nm. This double-clad structure ensured the high pump coupling efficiency of 75% with the seed coupling efficiency of 60%. Both power amplifiers were backward-pumped by high-power diode laser at 976 nm with a fiber pigtail core of 400 μm. Each end of the gain fiber was sealed and polished at an angle of 8° so as to protect the micro-structure and suppress parasitic lasing. Optical isolators were used before each fiber amplifier to avoid possible distortion or damage caused by any reflection of cascade amplifiers.

 figure: Fig. 1

Fig. 1 Experimental setup: IF, infrared filter; WDM, wavelength division multiplexing; YDF, ytterbium-doped ðber; OI, optical isolator; λ/2, half-wave plate; LD, laser diode; LMA-YDCF, large-mode-area Yb-doped double-clad fiber; APD, silicon avalanche photodiode.

Download Full Size | PDF

The output power of the first-stage power amplifier was 5 W with 30 W pump power, and it was divided into two parts: one part was for CEP stabilization with an average power of 1 W, while the other part was sent to the subsequent main power amplifier for power scaling. A pair of transmission gratings was employed for dispersion compensation after the fiber amplifier, which had a grating period of 1250 lines/mm and was designed to have maximum diffraction efficiency under Littrow conditions at a wavelength around 1030 nm. By fine tuning the distance between the grating pairs, the best dispersion compensation could be achieved, resulting in a laser pulse width of 220 fs.

The compressed pulses were delivered into a self-referenced nonlinear f-2f interferometer. It begun with the first beam splitter (BS-1) used to separate the infrared pulses into two parts. One part was focused into a 20-cm-long single-mode PCF (NL-PM-750, CRYSTAL FIBER) to acquire super-continuum, the other part was frequency-doubled by a 2-mm-thick BBO, with focusing and collimating spherical lenses in focal diameter of 100 and 50 mm, respectively. Laser pulses from both super-continuum and second harmonic generation (SHG) were combined with the second beam splitter (BS-2). By inserting an optical delay line in the arm of frequency doubling, the temporal delay of two laser pulses was fine controlled within the transit cycle of repetitive pulse train. The combined beams were incident to a reflection grating, where the spectral fraction at 515 nm was selected and later filtered by an aperture. Then it was detected by a silicon avalanche photodiode. By fine tuning the delay line, the beat signal was observed.

The CEP stabilization was achieved by using a phase locking loop (Menlosystems) with the beat frequency locked to a 20-MHz stable RF synthesizer, which was referenced to a 10 MHz Rubidium frequency standard (Stanford Research) with less than 2 × 10−11 short-term frequency stability. The error signal was calculated and converted to a voltage signal for modulating the pump power from the AOM of the Ti:sapphire femtosecond laser oscillator.

As for further power scaling of the infrared ultrashort laser pulses, the output power of the second-stage power amplifier was 100 W under 210-W laser diode pump power with 3-W input seeds from the first power amplifier. There was no obvious variation in the spectrum of the amplified high-power laser pulses with the output power rising from 10 W to 100 W, indicating the amplified spontaneous emission noise was effectively suppressed. We used the same kind of transmission gratings after the second-stage power amplifier for dispersion compensation, with the total compression efficiency of about 60%.

In order to extend the spectrum of high-power infrared ultrashort pulses to ultraviolet range, two BBO crystals were used as the nonlinear crystals for frequency quadrupling because of their great performance in the ultraviolet regime. Both BBO crystals had the same size of 5 × 5 × 4 mm3 and were all type-Ι phase matched. The crystal cut at 23.3° was used for SHG of the 1030-nm pulses and that for SHG of the 515-nm pulses was cut at 49.6°. The high-power infrared femtosecond laser pulses were focused onto the first SHG BBO by an f = 50 mm spherical lens to obtain green light and then collimated to parallel beam using the same kind of lens. A half-wave plate at 515 nm was added to change the polarization state of the generated green pulses. After the wave-plate, a coated lens with a focal length of 50 mm was used to focus the green light on the other SHG BBO crystal. Both of the two SHG crystals were mounted on three dimensional translation stages. By fine adjusting the angle and position of the crystal, the ultraviolet light was observed.

3. Results and discussion

As shown in Fig. 2(a) , the super-continuum from the amplified infrared pulses after the first-stage amplifier extended from 1030 to 475 nm, which provided an octave spectrum available for the self-referenced f-2f measurement. A beat signal was observed around 20 MHz with a signal-to-noise ratio about 33 dB at a resolution bandwidth of 100 kHz, as shown in Fig. 2(b).

 figure: Fig. 2

Fig. 2 (a) The super-continuum from the amplified infrared pulses after the first-stage amplifier; (b) The raw RF beat signal at the resolution bandwidth of 100 kHz.

Download Full Size | PDF

The linewidth of the free-running beat signal was 11.5 kHz as shown in Fig. 3(a) . Such a narrow linewidth was due to the excellent passive phase noise properties of the Ti:sapphire femtosecond oscillator. A digital counter with 10-bit resolution (Agilent 53131A) was employed to measure the instantaneous frequency fluctuation of the locked fCEP with a gate time of 1 s. The standard deviation of the CEP frequency at 5 W was estimated to be 1.69 mHz, which already included both system and detection noise within a recording time of 30 minutes, as shown in Fig. 3(b).

 figure: Fig. 3

Fig. 3 (a) The averaged free-running RF beat signal with a span of 800 kHz; (b) CEP offset frequency deviation measured by a high-accuracy counter with 1-s gate time.

Download Full Size | PDF

In order to measure the power spectral density, the beat signal was analyzed by a dynamic fast Fourier transform (FFT) analyzer (Stanford research SR760). We employed a frequency down-conversion method to transfer the fluctuation of locked fCEP from 20 MHz to DC. The power spectral density was integrated from 8 mHz to 100 kHz. The accumulated phase noise was calculated to be 0.41 rad at the output power of 5 W as shown in Fig. 4(a) , corresponding to a timing jitter of 143 as. We used the similar frequency down-conversion method to measure the linewidth of locked beat signal. The span of FFT was set to be 382 mHz, corresponding to a resolution linewidth of 0.9 mHz. The spectrum of beat signal was integrated within 1020 s. It revealed a linewidth of 1.86 mHz, as shown in Fig. 4(c).

 figure: Fig. 4

Fig. 4 The power spectral density and phase noise of locked offset frequency measured by (a) self- and (b) cross-referenced f-2f method. Linewidth of locked offset frequency measured by (c) self- and (d) cross-referenced f-2f method.

Download Full Size | PDF

We also used cross-referenced f-2f method as a comparison to show the noise characteristic of the locked offset frequency. In this method, a modified Mach-Zehnder interferometer was built. It started with a dichroic mirror, which was used as the input coupler to separate the long-wavelength part from Ti:sapphire laser. The infrared laser pulses centered at 1030 nm were filtered, amplified and frequency-doubled as one arm of the interferometer, while the rest laser pulses at 800 nm were coupled into a 20-cm-long PCF to extend the spectrum to short wavelength around 515 nm as the other arm of the interferometer. The output pulses from the two interference arms were combined by a beam splitter at 515 nm. The phase noise was 0.49 rad, corresponding to 170 as timing jitter, as shown in Fig. 4(b), while the linewidth was 2.06 mHz as shown in Fig. 4(d). The experimental results measured by the cross-referenced f-2f method were a little larger than those from the self-referenced one, which was mainly caused by the residual phase noise induced in the modified interferometer. The phase noise measured by either self- or cross- referenced f-2f method was relatively small in our experiment, indicating the high-power fiber amplifier added negligible phase noise. The accumulated phase noise was mainly caused by mechanical variation in fiber transmission, mode competition in super-continuum generation and diode pump fluctuation from the fiber amplification. Due to the excellent single-mode operation and large-mode-area of PCF in the power amplifiers, the ASE noise was sufficiently restrained. Also, gain competition was eliminated and stimulated Raman and Brillouin scattering were avoided because of the reduced peak power density in the gain fiber. The residual output pulses of the Ti:sapphire oscillator with a broadband spectra ranging from 650 to 900 nm could also be directly used as femtosecond comb source.

The average infrared power used for frequency quadrupling was 42.2 W with beam waist focused to 27 μm (focal diameter: 50 mm), corresponding to a peak intensity in the focus point of 48.3 GW/cm2, which resulted in 12.8 W at 515 nm. The optical-to-optical efficiency from infrared to green laser pulses was estimated to be 30%, with a variation less than 1% when the input infrared power increased. Then the green light was collimated and re-focused to the second SHG BBO crystal to obtain ultraviolet laser pulses.

In order to avoid the simultaneous existence of three wavelengths and select the ultraviolet laser pulses, three high-reflective mirrors at 266 nm and a short-wavelength-pass filter with a cut-off wavelength at 450 nm were used after the frequency-quadrupling system. A short-wavelength-pass filter with a cut-off wavelength at 400 nm was also placed before measurement to suppress the unnecessary Fresnel reflection.

Figure 5(a) shows the clean ultraviolet spectrum of femtosecond laser pulses at 258 nm, from which we can see the residual fundamental-wave parts at 1030 nm and green SHG at 515 nm were almost removed from the ultraviolet laser pulses. Experimentally, we measured the output power of the ultraviolet light as a function of the input infrared power. Figure 5(b) shows the frequency-quadrupling output power and conversion efficiency versus the input infrared power. The maximum average output power of the ultraviolet light obtained was 1.62 W at the input infrared power of 42.2 W. Therefore, the corresponding overall optical-to-optical efficiency was 3.85%. The ultraviolet output power was limited by a number of factors. The BBO crystals used for frequency quadrupling were not temperature-controlled. At a high repetition rate as high as 79.5 MHz, high-power infrared laser pulses generated unavoidable thermal load on the BBO crystal. Two-photon absorption of the frequency-quadrupled pulses also produced additional thermal effects in the BBO crystals. In our experiments, we need to strike a balance between conversion efficiency and operation convenience. Moreover, the conversion efficiency was reduced by the decrease of the peak power of the green laser pulses due to the broadened pulse width caused by the uncompensated dispersion. With further optimization of the optical efficiency and pulse width, higher ultraviolet output powers could be expected.

 figure: Fig. 5

Fig. 5 (a) The clean blue spectrum at 258 nm, a zoomed spectrum was shown in the inset. (b) The output ultraviolet power and corresponding conversion efficiency versus the input infrared power.

Download Full Size | PDF

3. Conclusions

In conclusion, we have demonstrated 1.62-W 79.5-MHz ultraviolet femtosecond pulses by the fourth harmonic generation of a high-power infrared LMA fiber chirped-pulse amplifier with an optical-to-optical efficiency of 3.85%. By careful elimination of the phase noise in the multi-stage fiber amplifiers, the spectral-fraction fiber amplifier exhibited negligible additional phase noise in comparison with the seed Ti:sapphire femtosecond frequency comb, as monitored by the self- and cross-referenced f-2f measurements. The spectral fraction amplified infrared comb as well as its frequency doubling and quadrupling together with the broadband Ti:sapphire seed comb provided an ultra-broadband high-power frequency combs covering the spectral range from ultraviolet to infrared.

Acknowledgment

This work was supported by National Natural Science Fund (11004061), National Key Project for Basic Research (2011CB808105), Special Research Fund for the Doctoral Program of Higher Education (20090076120004), projects from Shanghai Science and Technology Commission (10ZR1409000), Shanghai Educational Development Foundation (09CG18).

References and links

1. A. Cingöz, D. C. Yost, T. K. Allison, A. Ruehl, M. E. Fermann, I. Hartl, and J. Ye, “Direct frequency comb spectroscopy in the extreme ultraviolet,” Nature 482(7383), 68–71 (2012). [CrossRef]   [PubMed]  

2. Th. Udem, R. Holzwarth, and T. W. Hänsch, “Optical frequency metrology,” Nature 416(6877), 233–237 (2002). [CrossRef]   [PubMed]  

3. D. Z. Kandula, C. Gohle, T. J. Pinkert, W. Ubachs, and K. S. Eikema, “Extreme ultraviolet frequency comb metrology,” Phys. Rev. Lett. 105(6), 063001 (2010). [CrossRef]   [PubMed]  

4. J. Xu, D. Toptygin, K. J. Graver, R. A. Albertini, R. S. Savtchenko, N. D. Meadow, S. Roseman, P. R. Callis, L. Brand, and J. R. Knutson, “Ultrafast fluorescence dynamics of tryptophan in the proteins monellin and IIAGic,” J. Am. Chem. Soc. 128(4), 1214–1221 (2006). [CrossRef]   [PubMed]  

5. C. Hnatovsky, V. G. Shvedov, W. Krolikowski, and A. V. Rode, “Materials processing with a tightly focused femtosecond laser vortex pulse,” Opt. Lett. 35(20), 3417–3419 (2010). [CrossRef]   [PubMed]  

6. G. Vaschenko, F. Brizuela, C. Brewer, M. Grisham, H. Mancini, C. S. Menoni, M. C. Marconi, J. J. Rocca, W. Chao, J. A. Liddle, E. H. Anderson, D. T. Attwood, A. V. Vinogradov, I. A. Artioukov, Y. P. Pershyn, and V. V. Kondratenko, “Nanoimaging with a compact extreme-ultraviolet laser,” Opt. Lett. 30(16), 2095–2097 (2005). [CrossRef]   [PubMed]  

7. M. Ghotbi, A. Esteban-Martin, and M. Ebrahim-Zadeh, “Tunable, high-repetition-rate, femtosecond pulse generation in the ultraviolet,” Opt. Lett. 33(4), 345–347 (2008). [CrossRef]   [PubMed]  

8. J. H. Sun, B. J. Gale, and D. T. Reid, “Composite frequency comb spanning 0.4-2.4 microm from a phase-controlled femtosecond Ti:sapphire laser and synchronously pumped optical parametric oscillator,” Opt. Lett. 32(11), 1414–1416 (2007). [CrossRef]   [PubMed]  

9. X. Zhou, D. Yoshitomi, Y. Kobayashi, and K. Torizuka, “1 W average-power 100 MHz repetition-rate 259 nm femtosecond deep ultraviolet pulse generation from ytterbium fiber amplifier,” Opt. Lett. 35(10), 1713–1715 (2010). [CrossRef]   [PubMed]  

10. O. Kuzucu, F. N. Wong, D. E. Zelmon, S. M. Hegde, T. D. Roberts, and P. Battle, “Generation of 250 mW narrowband pulsed ultraviolet light by frequency quadrupling of an amplified erbium-doped fiber laser,” Opt. Lett. 32(10), 1290–1292 (2007). [CrossRef]   [PubMed]  

11. T. Schibli, I. Hartl, D. Yost, M. Martin, A. Marcinkevicius, M. Fermann, and J. Ye, “Optical frequency comb with submilihertz linewidth and more than 10 W average power,” Nat. Photonics 2(6), 355–359 (2008). [CrossRef]  

12. E. Baumann, F. R. Giorgetta, J. W. Nicholson, W. C. Swann, I. Coddington, and N. R. Newbury, “High-performance, vibration-immune, fiber-laser frequency comb,” Opt. Lett. 34(5), 638–640 (2009). [CrossRef]   [PubMed]  

13. Z. W. Barber, F. R. Giorgetta, P. A. Roos, I. Coddington, J. R. Dahl, R. R. Reibel, N. Greenfield, and N. R. Newbury, “Characterization of an actively linearized ultrabroadband chirped laser with a fiber-laser optical frequency comb,” Opt. Lett. 36(7), 1152–1154 (2011). [CrossRef]   [PubMed]  

14. W. Li, Q. Hao, M. Yan, and H. Zeng, “Tunable flat-top nanosecond fiber laser oscillator and 280 W average power nanosecond Yb-doped fiber amplifier,” Opt. Express 17(12), 10113–10118 (2009). [CrossRef]   [PubMed]  

15. I. Hartl, T. R. Schibli, A. Marcinkevicius, D. C. Yost, D. D. Hudson, M. E. Fermann, and J. Ye, “Cavity-enhanced similariton Yb-fiber laser frequency comb: 3 x 1014 W/cm2 peak intensity at 136 MHz,” Opt. Lett. 32(19), 2870–2872 (2007). [CrossRef]   [PubMed]  

16. W. Li, K. Yang, M. Yan, H. Zhou, J. Ding, and H. Zeng, “Long-term carrier-envelope-phase stabilized fiber-bulk hybrid laser with millihertz linewidth and 50 W average power,” Laser Phys. 21(3), 531–535 (2011). [CrossRef]  

17. A. Ruehl, A. Marcinkevicius, M. E. Fermann, and I. Hartl, “80 W, 120 fs Yb-fiber frequency comb,” Opt. Lett. 35(18), 3015–3017 (2010). [CrossRef]   [PubMed]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (5)

Fig. 1
Fig. 1 Experimental setup: IF, infrared filter; WDM, wavelength division multiplexing; YDF, ytterbium-doped ðber; OI, optical isolator; λ/2, half-wave plate; LD, laser diode; LMA-YDCF, large-mode-area Yb-doped double-clad fiber; APD, silicon avalanche photodiode.
Fig. 2
Fig. 2 (a) The super-continuum from the amplified infrared pulses after the first-stage amplifier; (b) The raw RF beat signal at the resolution bandwidth of 100 kHz.
Fig. 3
Fig. 3 (a) The averaged free-running RF beat signal with a span of 800 kHz; (b) CEP offset frequency deviation measured by a high-accuracy counter with 1-s gate time.
Fig. 4
Fig. 4 The power spectral density and phase noise of locked offset frequency measured by (a) self- and (b) cross-referenced f-2f method. Linewidth of locked offset frequency measured by (c) self- and (d) cross-referenced f-2f method.
Fig. 5
Fig. 5 (a) The clean blue spectrum at 258 nm, a zoomed spectrum was shown in the inset. (b) The output ultraviolet power and corresponding conversion efficiency versus the input infrared power.
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.