a2809_1.pdf CThX1.pdf
Nearly Three-Octave-Spanning Frequency Comb from a Phase-Controlled Femtosecond Ti:sapphire Laser and Synchronously Pumped Optical Parametric Oscillator J.H. Sun, B.J.S. Gale and D.T. Reid Ultrafast Optics Group, School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS
[email protected]
Abstract: A repetition-rate-stabilized frequency comb ranging from the violet to the mid-infrared is obtained by phase-locking a femtosecond Ti:sapphire laser and synchronously pumped optical parametric oscillator to a common supercontinuum reference. © 2007 Optical society of America OCIS codes: 120.5060, 140.7090, 190.4970, 320.7100.
Controlling carrier-envelope phase-slip (CEPS) and repetition frequencies of ultrafast laser pulses to achieve very broad band stabilized frequency combs is essential in coherent synthesis and optical metrology [1-3]. Frequency comb synthesis has been achieved between two independent Ti:sapphire lasers [1] and between independent Ti:sapphire and Cr:forsterite lasers [4]. However in such schemes, strict synchronization loops between lasers are required to equalize the interval of each frequency comb. By contrast, a synchronously pumped femtosecond (fs) optical parametric oscillator (OPO) is intrinsically synchronized with its pumping source, generally a fs Ti:sapphire laser. Furthermore, various phase-matched and non-phase-matched frequency mixing processes occurs simultaneously in a femtosecond OPO, generating multiple outputs which contain very broad spectral components. These advantages make a Ti:sapphire-pumped OPO an excellent candidate for broad band stabilized frequency comb generation. Previous research in the phase control of OPOs has demonstrated ways of locking the CEPS frequencies of the pump, signal and idler in a 3:2:1 ratio [5] but the frequency combs obtained in this way are moving relative to each other and are not frequency-locked to an absolute reference. In an OPO, the phases of the pump, signal and idler waves are related by the parametric equation φ p = φ s + φi − π / 2 (or φ&p = φ&s + φ&i ) [6]，where the dot indicates a time derivative and hence a phase-slip frequency. Therefore, if we controlled the CEPS frequencies of any two of the three waves, all of the other outputs, including those derivative products, such as second harmonic generation (SHG) of the signal and sum frequency generation (SFG) of pump and idler, are passively controlled. In this paper, we locked the repetition rate to a reference RF frequency and stabilized the CEPS frequency of the pumping Ti:sapphire laser. By detecting the interference between visible outputs from the OPO and the broadened pump spectrum from a photonic crystal fiber, the CEPS frequencies of the OPO outputs were also stabilized. The frequency combs of all these pulses cover nearly three octaves, considering the tunable range of the OPO. Fig. 1 shows a schematic of the frequency combs, and their offsets are listed in the inset table. 200MHz clock
1 / 4 divider
Mixer
Ti:S, 800nm 200MHz, 50fs
Verdi X
F rep
Delay1 P2
Delay2 PPLN OPO
Visible outputs
APD1
LM
P1
PBS IF2 IF1
PMT 1 / 2 divider
Fig. 1 Schematic diagram of stabilized frequency combs from a Ti:sapphire laser and an OPO. They all have fixed offsets from a reference comb with zero offset. The inserted boxes show an enlarged view of the comb offsets for the idler and SFG of pump and idler. The table (inset) shows the comb offset frequencies of each wavelength achieved experimentally.
Fig. 2 Schematic diagram of the experiment. PZT: piezoelectric transducer; AOM: acousto-optical modulator; APD: avalanche photo-diode; PMT: photomultiplier tube; PFD: phase frequency detector; IF: interference filter; LM: lower mirror; PBS: polarising beam-splitter; P: prism; Frep: pulse repetition rate.
Fig. 2 shows the experimental configuration. A custom-built Ti:sapphire laser, pumped by a Verdi X at 6W, generated 1.3W average mode-locked output power at 800nm and produced pulses with a 200 MHz repetition frequency and 50 fs duration. An AOM (IntraAction ASM-803B47-1) was used to modify the Ti:sapphire pump
a2809_1.pdf CThX1.pdf
power and so control its CEPS frequency. One end mirror of the Ti:sapphire laser cavity was mounted on PZT1 (PI P-820.10) to achieve repetition rate locking with a 200MHz clock (PRL-175NT-200). 80% of the output power was used to pump a MgO:PPLN OPO with signal output tunable from 1.2 to 1.37 µm. The remainder 20% power was launched into a 30cm photonic crystal fiber (PCF, CRYSTAL FIBRE NL-2.0-740) in a typical f to 2f nonlinear interferometer to detect the CEPS frequency of the pump laser. In the OPO cavity, PZT2 (Thorlabs, AE0203D04F; 261kHz resonant frequency) was mounted on a cavity end mirror in order to control OPO signal CEPS frequency [7]. Among the visible non-phasematched mixing outputs, the frequencies at 2ωs, (red) and ωp+ωi (yellow) are in the spectral range of the residual pump supercontinuum from the nonlinear interferometer. By using two independent delay lines for the red and yellow, we beat them against the supercontinuum simultaneously in the second interferometer (Fig. 1, right) and used APD2 and PMT to detect the heterodyne frequencies of the red and the yellow, which are 2φ&s − φ&p and (φ&p + φ&i ) − φ&p = φ&i respectively. Two phase frequency detectors (PFD) [8] were used to compare the detected frequencies from APD1 and APD2 with sub-harmonics of the repetition rate. The error signals, after amplification, were used to drive the AOM and PZT2 respectively, completing the phase-locking loops for the pump and the frequency-doubled signal pulses, and resulting in full offset-frequency control of the various frequency combs generated. The beat of the yellow output monitored by the PMT was used to assess the locking quality. -70
-20
-40
-30
-50
-30 (c)
(b)
-80
RF intensity (dBm)
RF intensity (dBm)
(a)
-90
-100 -60
-40
-110 -70
-50
-120
-80
-60 49.96
50.00
50.04
Frequency (MHz)
25.00
25.04
Frequency (MHz)
37.46
37.50
37.54
Frequency (MHz)
Fig. 3 Measured CEPS frequency of the pump (a), heterodyne beat frequencies of the SHG of signal (b) and SFG of pump and idler (c). The frequency span of each measurement is 100kHz.
(b)
6
-35 -40 -45
5 4 3
-50
-130 24.96
7
(a) Intensity (V)
-30
-10
24.996
25.000
25.004
2 0.0
unlocked locked
.2
Frequency (MHz)
.4
.6
.8
1.0
time (ms)
Fig. 4 (a) RF spectrum of the locked heterodyne beat frequencies of signal SHG. (b) Error signals from the phase frequency detector PFD2 when locked (red) and unlocked (black).
In our experiment, we locked the CEPS frequency of the pump at 50MHz, and the heterodyne beat frequency of the red at 25MHz. This means the CEPS frequency of signal is 12.5MHz, hence that of idler (the beat of yellow output with the supercontinuum), should be 37.5MHz, and this is confirmed by the measurement in Fig. 3. All beat frequencies were monitored by a RF spectrum analyser (Agilent E4411B) in turn. Compared with the CEPS frequency jitter before locking, which was 500 kHz, 8MHz and 4MHz for the pump, red and yellow respectively, the widest locking result, that of the red, showed a bandwidth of around 3kHz at -3dB (Fig. 4), which is close to the 1kHz resolution bandwidth of the spectrum analyser. This resolution bandwidth is not fine enough to measure the repetition rate locking result, because even when free running, the jitter is less than 100Hz. The quality of the repetition-rate locking was assessed from the error signal from the mixer (see Fig. 2) and we confirmed that this was almost constant during repetition-rate and CEPS frequency-locking. In summary, by locking the repetition rate to an external reference clock, and by locking the CEPS frequencies of the pulses from a femtosecond Ti:sapphire laser and OPO using a common supercontinuum, we have obtained a composite frequency comb covering from the violet to the mid-infrared. Changing the locking frequencies will achieve different relative and absolute positions of the constituent combs. This result has applications in the coherent synthesis of ultrafast pulses. Using an atomically-referenced clock to lock the laser repetition frequency should allow this technique to provide very broad and precisely controlled frequency combs for optical metrology. Reference: [1]. H. R. Telle, G. Steinmeyer, A. E. Dunlop, J. Stenger, D. H. Sutter, U. Keller, Appl. Phys. B 69, 327-332 (1999) [2]. Robert K. Shelton, Long-Sheng Ma, Henry C. Kapteyn, et al., Science 293, 1286-1289 (2001). [3]. T. Udem, R. Holzwarth, and T. W. Hänsch, Nature 416, 233 (2002). [4]. Yohei Kobayashi, Dai Yoshitomi, Masayuki Kakehata, et al., Opt. Lett. 30, 2496-2498 (2005). [5]. Y. Kobayashi, H. Takada, M. Kakehata and K. Torizuka, Opt. Lett., 28 1377-1379 (2003) [6]. Y. R. Shen, The principles of Nonlinear Optics (Wiley InterScience, New York, 1984), p 188. [7].Y. Kobayashi and K. Torizuka, Opt. Lett., 26 1295-1297 (2001) [8]. M. Prevedelli, T. Freegarde, T. W. Hänsch, Appl. Phys. B 60, S241-S248 (1995).
