We report an optical single-frequency synthesizer at the 1.55 μm telecommunications band. Output from a continuous-wave external cavity diode laser is frequency doubled and phase locked to a predetermined component of a Ti:S laser frequency comb. The synthesizer is capable of generating a single user-specified frequency from an atomic time base within the 192–196 THz gain bandwidth of an erbium-doped fiber amplifier. By tuning the repetition rate of the femtosecond laser the synthesized optical frequency can be swept with sub-kilohertz step size. Frequency sweeps of several GHz are realized by automatically re-locking the diode laser to adjacent comb components during frequency sweep. We demonstrate the operation of the device by presenting results of Doppler-free spectroscopy on acetylene using synthesized frequencies.
©2009 Optical Society of America
The introduction of a full-octave optical frequency comb generator at the turn of millennium enabled practical absolute optical frequency measurements by providing a direct link between microwave and optical frequencies [1,2]. Since then ever sophisticated optical frequency metrology tools have been developed, one example being continuous-wave (cw) optical frequency synthesizers [3–10]. Such devices are indispensable, as unlike conventional frequency comb generators that produce a multitude of rather weak frequency components, a cw synthesizer provides a single selectable frequency. Such a light source is directly suitable for, e.g., spectroscopy, high precision characterization of optical components, or as a frequency reference in calibration of stabilized lasers.
Here we report on a cw synthesizer capable of generating a single user-specified frequency from an atomic time base on wavelengths around 1.55 μm within the 192-196 THz gain bandwidth of an erbium-doped fiber amplifier (EDFA). This frequency range is important for telecommunication networks as most of the long-haul data transmission takes place on this so called optical C-band due to the convenience provided by EDFAs. The synthesizer also directly provides an output at the second harmonic of these frequencies, i.e., at wavelengths around 775 nm.
The construction and characteristics of the synthesizer are presented below. We demonstrate the operation of the device in an experiment where we have used the synthesizer to conduct Doppler-free spectroscopy on P(16) line of acetylene (13C2H2) at the v 1+ v 3 overtone band.
2. Synthesizer construction
The general structure of the experimental setup resembles that of some radio frequency (rf) synthesizers; an optical frequency comb is used to phase coherently down-convert an optical frequency to the rf-regime, where conventional rf-techniques are employed to phase lock the frequency to a reference using a digital phase-locked loop-circuit (PLL). Figure 1 shows a schematic block diagram of the synthesizer.
A tunable external cavity diode laser (ECDL) (Photonetics Tunics-PRI) provides >5 mW of output power over the 192–196 THz gain bandwidth of an EDFA. The signal is amplified with a polarization maintaining EDFA (Pritel PMFA-27) to 250 mW for efficient second harmonic generation (SHG). The frequency-doubled signal is heterodyned with a predetermined component of an optical frequency comb generator, whose repetition rate and carrier-envelope-offset (CEO) are referenced to a hydrogen-maser (Kvarz CH1-75A). The resulting beat-note frequency is band-pass filtered and fed to a H-maser referenced digital PLL-circuit, which phase locks the ECDL frequency to the comb output, and hence, to the H-maser.
The optical frequency comb generator is based on a Kerr-lens mode-locked Ti:S laser with 1 GHz repetition rate and its offset frequency is stabilized using self-referencing . The output spectrum of the frequency comb generator extending from 500 nm to 1150 nm is relatively strongly structured and exhibits several minima 15 – 35 dB below the highest power spectral density. Unfortunately, the original configuration had a power minimum at 770 nm around which the spectral power was found to be too low to obtain sufficient beat-note signal for reliable phase locking. Because of this, 15 % of the output power of the Ti:S-laser is taken to a microstructure fiber (Crystal Fibre, NL-2.3-790-02) specifically selected to provide high-intensity output around 775 nm.
The second-harmonic generation of the ECDL is carried out using a fiber-pigtailed periodically poled lithium niobate waveguide (PPLN WG, HC-Photonics). The poling of the device has a chirped structure along the length of the 30 mm long chip, allowing wide phase-matched bandwidth without any adjustment of the chip. The device provides approximately 1 mW of power at the second harmonic when 250 mW of power from the output of the polarization maintaining EDFA is taken to its input.
The beat-frequency signal between the second harmonic signal and the comb is obtained by overlapping the laser modes in a single-mode fiber using a 2×1 fiber coupler. A Si PIN-photodetector is used for detection. A beat-frequency signal with more than 30 dB signal-to-noise ratio (SNR) using a 300-kHz resolution bandwidth is routinely achieved. This is sufficient for reliable phase locking when filtered with a relatively narrow (25 MHz) bandpass filter.
The frequency stability of the synthesizer follows that of the microwave frequency reference  for integration times longer than the time constant of the servo-loops. The H-maser used in the experiment has a relative frequency stability of 2×10-12 at an integration time of 1 s. Frequency stabilities below 10-14 are reached after an integration time of a few hundred seconds. An upper limit for short-term (2 ms) linewidth of the synthesizer was measured to be 800 kHz.
3. Frequency control
Prior to phase locking to the frequency comb, the optical frequency of the ECDL is set approximately to the targeted value. A coarse frequency tuning of the ECDL steps the output frequency with 100-MHz resolution over the entire tuning range. Continuous fine-tuning of the frequency is allowed over about ± 3 GHz range around the coarsely set operating frequency. A calibrated wavelength meter (HP86120B) is used to provide frequency information with an accuracy of 100 MHz.
After the ECDL is set close to the targeted frequency, the comb spectrum is tuned by the repetition rate until the beat-frequency signal appears inside the 25 MHz band of the bandpass filter. Then the PLL is switched on and the frequency of the ECDL attains a value defined by the frequency comb as (nfrep±fCEO±fbeat)/2 where n is the number of the comb component to which the ECDL is locked. The frequency information from the wavelength meter is sufficient for direct determination of n.
The PLL-block consists of a digital charge pump phase-locked loop circuit (Analog devices, LMX1501A) and an operational amplifier integrator circuit as a loop filter. Digital dividers integrated in the PLL circuit allow phase comparison at kHz level. This ensures sufficient averaging of signal jitter to avoid cycle slips. The feedback to the ECDL is realized by controlling the laser cavity with a piezoelectric actuator.
The locking range at the output of the integrator is limited to ± 5 MHz. Due to the rather narrow locking range, a DC-bias voltage is added to the integrator output using a data acquisition card (NI PCI-6035E) and a summing operational amplifier circuit. The output voltage of the integrator is read to the computer and its deviation from the range center is measured. A corresponding voltage is summed to the control signal for the integrator output voltage to return to the center level. This is needed to compensate slow drifts of the ECDL locking point during long measurements or when the synthesized frequency is tuned over the range allowed by the integrator output. The frequency sweeps are realized using the approach proposed in Ref. 12. The comb repetition rate is stepped in small increments (δff), which steps the frequency of the nth comb component by nδfr, and hence, also the second harmonic frequency of the ECDL locked to this component by the same amount. The repetition rate is synthesized from the H-maser with a rf-signal generator (Agilent E4424B), which allows stepping the frequency with millihertz-resolution. When mapped to target optical frequencies (i.e. multiplied by n/2 ≈ 195 000), this corresponds to a few hundred hertz resolution.
With our Ti:S-laser based comb the practical tuning range of a single frequency component by the repetition rate is limited to around 1 GHz. To achieve larger frequency sweeps, the second harmonic of the ECDL frequency locked to a certain comb component can be handed-over on the fly to the adjacent component during a sweep. The principle is depicted in Fig. 2. The repetition rate is stepped until the comb component n to which the ECDL is locked reaches the initial position of the adjacent component (n+1). Then the repetition rate is set back to the starting value and the PLL locks the laser to this (n+1)th component, after which the sequence can be repeated. Although the lock is momentarily lost during the handover process, the DC-bias voltage applied to the output of the PLL loop filter keeps the ECDL-frequency within the locking range in respect to the (n+1)th component. The speed of the sweep is mainly limited by a few hundred millisecond settling time of the PLL after each frequency step. With this approach the synthesized frequency can be swept over 6 GHz, which is the range allowed by the ECDL fine-tuning. However, frequency sweeps over several THz would be possible in principle.
4. Doppler-free spectroscopy on acetylene
We demonstrate the operation of the synthesizer by using it for Doppler-free spectroscopy of acetylene. We studied the P(16) line of 13C2H2 at the v1 + v3 overtone band, which serves as a wavelength reference at 1.54 μm . The experimental arrangement is realized using a slightly modified version of a fiber-based setup that we have previously applied with our acetylene-stabilized laser scheme .
Figure 3 shows a general structure of the complete measurement arrangement. The spectroscopy setup is connected to the synthesizer between the ECDL output and the EDFA input. In the figure, the blocks that are part of the synthesizer are shaded in gray whereas the blocks belonging specifically for the acetylene measurement are unshaded. As the details of the spectroscopy setup are given in Ref. 14, we give here only a brief description.
The output signal from the ECDL is amplified to approximately 125 mW in a 10 meter long erbium-doped fiber, pumped with a distributed feedback (DFB) laser diode at 1480 nm. Fiber-optic isolators are placed at the outputs of the ECDL and the erbium fiber to suppress possible feedbacks. Small amount of power is branched off after the erbium fiber with a 90:10 fiber coupler and taken to a photodiode, which provides a feedback signal to the DFB laser for intensity stabilization. Attenuation is added to the open arms of the fiber coupler to reduce the intensity of the signal to a level suitable for detection and to suppress the reflections from the open fiber ends. The rest of the amplified signal is directed to an acetylene cell using a three-port fiber-optic circulator.
The spectroscopic arrangement is realized in a collinear geometry using reflected pump beam to probe the absorption. An aspheric lens at the output of the circulator collimates the pump beam and couples the reflected beam back into the fiber. The collimation of the beam is carefully adjusted in order to locate the beam waist at the end mirror right behind the acetylene cell. The 1/e2-diameter of the waist is approximately 0.7 mm. This leads to one-way power density of 19 W/cm2 at the waist where 75 mW of optical power is left after insertion losses due to the fiber optic components. Pressure in the 50-cm long acetylene cell is 2.8 Pa.
After going through the circulator the probe signal is divided with a 50:50 fiber coupler. One output arm directs the signal to the synthesizer arrangement and the other arm to a temperature stabilized InGaAs photodiode for detection. As with the first fiber coupler, also here attenuation is added to the output arms of the coupler. The signal from the photodiode is amplified with a low-noise current preamplifier (SR570) and measured with a digital volt meter (Agilent 34410A).
The reflections from the interfaces of the setup are minimized by fusion splicing the fiber connections. However, the inevitable reflection from the open end of the fiber at the acetylene cell, although minimized using an angled fiber connector, is of the order of 10-4 and causes approximately one percent interference term in the detected signal. To cancel the interference through averaging, the mirror used to reflect the pump beam is dithered with a piezoelectric transducer. When the signal is appropriately low pass filtered at detection the interference term is effectively averaged out.
We studied the intensity stability of the measurement system while scanning the frequency. Figure 4(a) shows a result of 23 consecutive measurements where the synthesized frequency was swept over an 8 MHz frequency range with 100 steps at a region well outside the acetylene absorption. The signal was low pass filtered with 10 Hz band. The delay between successive measurement points within a frequency sweep was 480 ms and at each measurement point the signal was averaged 200 ms. The dead-time between sweeps was typically 30 s, which was needed for sweep initialization. The overall measurement time was 1800 s. The relative standard deviation of the 2300 intensity measurement points is 9×10-6.
Figure 4(b) shows a result of a measurement where the frequency was swept 6 GHz over the Doppler-broadened transition of P(16)-line. The frequency step size was 2.5 MHz. At each measurement point the signal was averaged 20 ms. Due to the frequency resolution, the Doppler-free line shape is not properly shown. Figure 4(c) shows measurement of the Doppler-free line shape at higher resolution. The graph is an average of 31 similar frequency sweeps over the line center. The frequency step is 79 kHz and integration time used at each point is 200 ms. The graphs in Figs. 4(b) and 4(c) are plotted on absolute frequency scale with the scale center at 194 369 569.384 MHz, which is the value adopted by International Committee for Weights and Measures (CIPM) for this transition when measured with the third-harmonic technique using 1-MHz modulation amplitude . A Lorentzian fitted to the Doppler-free line shape gives a center frequency of 9 kHz below the CIPM value. We performed four equivalent sets of measurements over the Doppler-free line shape on two separate days. The standard deviation of the line center frequencies given by Lorentzian fits was 0.8 kHz. The deviation from the CIPM value is assumed to be primarily due to line shape asymmetries arising from imperfections of the beam alignment in the spectroscopic arrangement . This effect can also be observed with stabilized lasers employing the third-harmonic technique. However, in that case the effect is less pronounced as usually only the peak of the line is examined.
We have presented an optical single-frequency synthesizer operating within the gain bandwidth of an EDFA. The construction is based on a wavelength tunable ECDL whose output is phase-locked to a predetermined component of a H-maser referenced Ti:S laser frequency comb. The synthesized frequency can be swept over several GHz limited only by the continuous tuning range of the ECDL. In principle, the frequency tuning by adjustment of the Ti:S laser repetition rate combined with the comb component handover process demonstrated here would allow frequency sweeps over the entire gain bandwidth of an EDFA.
We have demonstrated the performance of the synthesizer in high precision spectroscopy on acetylene. The capability of repeating the measurements on absolute frequency scale with the relative intensity stability at 10-5 level enabled recording the line shapes with great accuracy. Even a Doppler-free absorption feature, as small as 8×10-4 of total detected power, could be observed at high signal-to-noise ratio without modulation techniques. A Lorentzian fit to the result gives a line center frequency of 9 kHz below that of the CIPM recommendation. These values can not, however, be directly compared without further investigation of line shape asymmetries, which might cause a shift to the line center given by the fit. The line shape asymmetries are also a potential source of error when determining the line center frequencies using the conventional third-harmonic technique.
Here we demonstrated sweeping the frequency by frequency comb’s repetition rate. An alternative approach could be to vary the comb’s offset frequency by altering the pump power. This would leave the Ti:S laser cavity untouched and hence lead to more stable operation during frequency scans. However, the obtainable tuning range would in practice be restricted to less than 100 MHz.
In addition to the applications in spectroscopy, the stabile frequency and intensity properties of the developed system makes the synthesizer directly suitable, e.g., for high precision characterization of components used in optical telecommunications networks or as a frequency reference when calibrating acetylene-stabilized lasers. The output at wavelengths around 775 nm could be useful also in atomic physics experiments.
V. Ahtee is grateful for the financial support of the Academy of Finland. The authors would like to thank T. Hieta for experimental help with the fiber optics.
References and links
1. D. J. Jones, S. A. Diddams, J. K. Ranka, A. Stentz, R. S. Windeler, J. L. Hall, and S. T. Cundiff, “Carrier-envelope phase control of femtosecond mode-locked lasers and direct optical frequency synthesis,” Science 288, 635–639 (2000). [CrossRef] [PubMed]
2. R. Holzwarth, Th. Udem, T. W. Hänsch, J. C. Knight, W. J. Wadsworth, and P. St. J. Russell, “Optical frequency synthesizer for precision spectroscopy,” Phys. Rev. Lett. 85, 2264–2267 (2000). [CrossRef] [PubMed]
3. J. D. Jost, J. L. Hall, and J. Ye, “Continuously tunable, precise, single frequency optical signal generator,” Opt. Express 10, 515–520 (2002). [PubMed]
4. T. R. Schibli, K. Minoshima, F.-L. Hong, H. Inaba, Y. Bitou, A. Onae, and H. Matsumoto, “Phase-locked widely tunable optical single-frequency generator based on a femtosecond comb,” Opt. Lett. 30, 2323–2325 (2005). [CrossRef] [PubMed]
5. H. Inaba, T. Ikegami, F.-L. Hong, Y. Bitou, A. Onae, T. R. Schibli, K. Minoshima, and H. Matsumoto, “Doppler-free spectroscopy using a continuous-wave optical frequency synthesizer,” Appl. Opt. 45, 4910–4915 (2006). [CrossRef] [PubMed]
6. H. S. Moon, E. B. Kim, S. E. Park, and C. Y. Park, “Selection and amplification of modes of an optical frequency comb using a femtosecond laser injection-locking technique,” Appl. Phys. Lett. 89, 181110 (2006). [CrossRef]
7. S. E. Park, E. B. Kim, Y.-H. Park, D. S. Yee, T. Y. Kwon, C. Y. Park, H. S. Moon, and T. H. Yoon, “Sweep optical frequency synthesizer with a distributed-Bragg-reflector laser injection locked by a single component of an optical frequency comb,” Opt. Lett. 31, 3594–3596 (2006). [CrossRef] [PubMed]
8. T. M. Fortier, Y. Le Coq, J. E. Stalnaker, D. Ortega, S. A. Diddams, C. W. Oates, and L. Hollberg, “Kilohertz-resolution spectroscopy of cold atoms with an optical frequency comb,” Phys. Rev. Lett. 97, 163905 (2006). [CrossRef] [PubMed]
9. Y.-J. Kim, J. Jin, Y. Kim, S. Hyun, and S.-W. Kim, “A wide-range optical frequency generator based on the frequency comb of a femtosecond laser,” Opt. Express 16, 258–264 (2008). [CrossRef] [PubMed]
10. H. Y. Ryu, S. H. Lee, W. K. Lee, H. S. Moon, and H. S. Suh, “Absolute frequency measurement of an acetylene stabilized laser using a selected single mode from a femtosecond fiber laser comb,” Opt. Express 16, 2867–2873 (2008). [CrossRef] [PubMed]
11. M. Merimaa, K. Nyholm, M. Vainio, and A. Lassila, “Traceability of laser frequency calibrations at MIKES,” IEEE Trans. Instrum. Meas. 56, 500–504 (2007). [CrossRef]
12. B. R. Washburn, R. W. Fox, N. R. Newbury, J. W. Nicholson, K. Feder, P. S. Westbrook, and C. G. Jørgensen, “Fiber-laser-based frequency comb with a tunable repetition rate,” Opt. Express 12, 4999–5004 (2004). [CrossRef] [PubMed]
13. R. Felder, “Practical realization of the definition of the metre, including recommended radiations of other optical frequency standards (2003),” Metrologia 42, 323–325 (2005). With revision by the Working Group on the Mise en Pratique (MEP2005). [CrossRef]
14. V. Ahtee, M. Merimaa, and K. Nyholm, “Fiber-based acetylene-stabilized laser,” IEEE Trans. Instrum. Meas. (In press, doi: 10.1109/TIM.2008.2008476)
15. C. J. Bordé, J. L. Hall, C. V. Kunasz, and D. G. Hummer, “Saturated absorption line shape: Calculation of the transit-time broadening by a perturbation approach,” Phys. Rev. A 14, 236–263 (1976). [CrossRef]