Open Access
Add to BibSonomyAbstract
We disseminated an ultra-broadband optical frequency reference based on a femtosecond (fs)-laser optical comb through a kilometer-scale fiber link. Its spectrum ranged from 1160 nm to 2180 nm without additional fs-laser combs at the end of the link. By employing a fiber-induced phase noise cancellation technique, the linewidth and fractional frequency instability attained for all disseminated comb modes were of order 1 Hz and 10−18 in a 5000 s averaging time. The ultra-broad optical frequency reference, for which absolute frequency is traceable to Japan Standard Time, was applied in the frequency stabilization of an injection-seeded Q-switched 2051 nm pulse laser for a coherent light detection and ranging LIDAR system.
© 2016 Optical Society of America
1. Introduction
Optical frequency comparisons are essential for the re-definition of the SI second by optical clocks being developed worldwide. An optical frequency transfer via optical-fiber links is a promising manner for the precise comparison of optical clocks. There are two fundamental techniques for the optical frequency transfer: optical carrier transfer and optical frequency comb transfer [1]. Optical carrier transfer directly transmits a continuous-wave (cw) of a fixed laser frequency through optical fibers, which is phase-coherently linked to an optical frequency standard used for a time-keeping of the optical clock in a local site. The single frequency of the optical carrier transferred is received at the remote site to be compared with another optical standard operating at a nearby frequency. Because optical phase noise induced by fiber transmission degenerates the frequency stability of the optical carrier, optical carrier transfer with active fiber-induced phase noise cancellation (FNC) has been employed in comparing optical clocks located at distance from over one to thousands kilometers away [2–4]. Optical frequency comb transfer sends periodic optical pulse trains emitted from a femtosecond (fs)-pulse mode-locked laser to remote sites. This manner supplies the remote end users with both microwave and optical frequency references simultaneously, which are phase-coherent with the optical frequency standard at the local site. The spectral components of the fs-laser pulse form an optical comb, which are regularly spaced by pulse repetition rate (frep) and shifted by a carrier-envelope offset frequency (fceo). The frequency of the n-th component f (n) of the fs-laser frequency comb (FLFC) is determined by two frequencies: f (n) = n frep + fceo [5]. The stabilization of both frep and fceo makes all comb components stable. The fceo can be observed by the self-referencing technique. It is usually phase-locked to the microwave standard by controlling the difference between the group and optical-phase velocities. Likewise, frep can be stabilized to the microwave standard directly (microwave locking) or optical frequency standard by stabilizing the heterodyne beat between the m-th component of the FLFC and the optical standard (optical locking): frep = (νL − fceo − fbeat)/m, where νL is the frequency of the optical standard and fbeat is an arbitrary frequency of the heterodyne beat. The frep stabilization is achieved by controlling the group delay of pulses in the laser cavity. Although optical comb transfer has upward compatibility with optical carrier transfer, its study has been advanced steady because of the complex control topology required in principle for stabilization with two degrees of freedom. Much research has been performed on active and passive optical comb transfer with the microwave locking of frep to investigate the characteristics of either or both of the delivered microwave and optical references [6–11]. Among them, the optical comb transfer was accomplished with a 10−18-level accuracy over a 7.7 km spooled fiber by G. Marra et al [12]. To attain the accuracy in transmitting the optical reference, an optical phase detection technique played an important role to measure the fiber-induced noise with a high signal-to-noise ratio (SNR). The group delay of the fs pulse train was controlled only to compensate the Doppler frequency shift on the optical components of the FLFC, which arise from environmental perturbations along the fiber. The accuracy of these two kinds of optical frequency transfer has reached a suitable level enabling comparison of the best optical clocks. However, the end users would need to equipped with a costly and complicated FLFC at their remote sites if the optical frequencies planned for comparison or measurement either were over ~ 100 GHz apart from the frequencies of the delivered reference of around 1550 nm or were not sufficiently close to the frequency of its second or third harmonic. This is because optical frequency transfer exploits minimizing losses in standard silica-glass fiber at around 1550 nm and hence inherently restricts the bandwidth of delivered frequency references in transmissions over a kilometer-long fiber link. The actual comparison of optical clocks is generally in such instance, because their clock frequencies are widely disparate in the visible and near-infrared regions [13]. Hence, the optical comb transfers previously reported with broad bandwidth of at most 100 nm still need an auxiliary FLFC at the remote site to bridge the 1.5 μm band and targeted clock frequencies [12].
An optical-frequency reference that is simple, ultra-broadband and cost-effective is needed not only for optical-clock comparisons but also in many fields of research requiring laser frequency stabilization and measurement. In high-speed optical communications, the dramatic growth of internet traffic supports the rapid development of a large data transmission bandwidth. While conventional optical communications have been implemented around the 1.5 μm wavelength range, where silica-fiber losses are low and reliable amplification technologies exists, recent data transmission studies are looking at shorter [14] and longer wavelength bands [15]. Multilevel modulation formats with coherent heterodyne or homodyne detection have demonstrated high spectral efficiency and dramatically improved capacity for optical transmission systems. In applying these formats with large transmission bandwidths, numerous costly narrow-linewidth lasers must be used as local oscillators at the user end [16]. An ultra-broadband optical frequency reference (UBOFR) enables these local oscillators to be stabilized and thereby establish an option for future high-speed large-capacity data transmission networks. In environmental science, the impact of global warming and climate change on Earth’s environment has been a great interest [17]. A coherent differential-absorption (DA) light detection and ranging (LIDAR) is expected to improve the absolute error of measurements of greenhouse-gas concentrations in the atmosphere [18]. The absorption spectra of gas molecules such as carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O) lie in the wide infrared region. Therefore, a high-energy single-axial-mode laser for the coherent LIDAR system must oscillate at the specific wavelength for gas spectroscopy. An UBOFR is prerequisite for determining the absolute frequency and reducing the frequency drift of the laser. Although its dissemination have been required by diverse applications including the above, the development of such optical frequency transfer has not been addressed.
In this paper, we present the dissemination of an optical-comb-based UBOFR through a fiber network over distances of kilometers. Its optical spectrum covers an unprecedented wide range sufficient for actual comparisons of optical clocks without costly and bulky FLFCs at the remote end of the network. The control of the Doppler frequency shift imposed during the fiber transmission was performed against the optical-comb modes centered around the 1.5 μm wavelength, and its effect was generated across a wide spectral range as the suppression of excess phase noise for all comb modes. As a practical demonstration, the UBOFR, which was optical-locked to a narrow-linewidth laser having an absolute frequency traceable to Japan Standard Time (JST), was provided via the fiber network to a CO2 DA/Doppler wind (DW) LIDAR system to stabilize the frequency of a Q-switched 2 μm pulse laser.
2. Experimental setup
Figure 1 illustrates the experimental setup for the dissemination of an optical-comb-based UBOFR with cancellation of excess optical phase noise induced during transmission in the fiber link. The fiber link comprises two cascaded polarization-independent single-mode (SM) fibers with a length of about 900 m, which was revealed by optical time-domain reflectometry. Situated in a utility corridor, these SM fibers connect two laboratories located in different buildings in adjacent campuses at the National Institute of Information and Communications Technology (NICT). Both local and remote sites were collocated in a single laboratory to facilitate the assessment of the system.
Fig. 1 Experimental setup of an ultra-broadband optical frequency reference transfer via a fiber network. AOM: acousto-optic modulator, CIR: optical circulator, DCF: dispersion compensating fiber, EDFA: erbium-doped fiber amplifier, FS: piezo-actuated fiber stretcher, HNLF: highly nonlinear fiber, SCL: single-mode fiber coupler, SMF: polarization-independent single-mode fiber, VCM: voice-coil-motorized delay-line stage, fbeat: frequency of heterodyne beat signal.
A narrow-linewidth laser with a wavelength of 1538 nm served as an optical frequency standard. The laser frequency was stabilized to the resonance frequency of a high-finesse Fabry-Perot cavity by means of the Pound-Drever-Hall technique [19]. The Fabry-Perot cavity was composed of two high-reflective mirrors and an ultra-low expansion glass spacer of length 10 cm. This cavity was put in a vacuum chamber, which was kept below 10−6 Pa, and its temperature was actively controlled to reduce the laser frequency drift. The vacuum chamber was placed on a passive vibration isolation platform to decrease the laser frequency noise. The control topology for the frequency stabilization was designed to narrow the laser linewidth to less than 3 Hz and to maintain long-term continuous-operation for over half a year.
The optical frequency reference delivered to the remote site was a FLFC stabilized to the 1538 nm narrow-linewidth laser by the optical locking. The FLFC was emitted from a self-referenced mode-locked erbium (Er)-doped fiber laser (Menlo Systems, FC 1500). Its frep and pulse width were 250 MHz and 85 fs, respectively. The optical spectrum of the FLFC has a central wavelength of 1570 nm and a spectral width of about 60 nm as shown in Fig. 2(a). The fs-pulse mode-locked laser has three equivalent output ports with an average optical power of 28 mW (Port 1,2,3). The output from Port 1 was mixed with the narrow-linewidth laser to obtain a heterodyne beat signal. The frequency of the beat signal fbeat 1 was compared with a local oscillator of 60 MHz to extract a control signal for the optical locking. The low-frequency components of the control signal were fed back to a piezo–actuator to modulate the cavity length of the fs-pulse laser. The high-frequency components were fed back to an electro-optical modulator inside the laser cavity to expand the control bandwidth. The bandwidth achieved was approximately 300 kHz.
Fig. 2 Optical spectra of the original comb emitted from a fs-pulse mode-locked laser (a), ultra-broadband optical reference disseminated after passing a 1.8 km fiber link (b), and broadband continuum related to the original comb at the local site (c). The spectrum of the ultrabroad optical reference ranges from 1160 to 2180 nm at −20 dB below the maximum. The wavelengths of the three auxiliary lasers used in the fiber-induced phase noise measurement are indicated by dotted lines.
The dissemination of the UBOFR was accomplished by spectral broadening of the FLFC received at the remote site. This novel approach circumvents the need for an additional fs-pulse mode-locked laser at the remote site for spanning the gap between the delivered reference and measured optical frequencies. In addition, it overcomes limitation in the optical amplification technology for compensation of intrinsic fiber loss, which originates from Rayleigh scattering at the shorter wavelengths and infrared absorption at longer ones. Although chalcogenide glass fibers and hollow-core photonic band-gap fibers [15] offer the potential for lower fiber loss around 2 μm, they are still rather expensive or not mature technologically to be available for practical fiber networks. The output from Port 2 was introduced into a SM fiber coupler 1 (SCL 1) with a coupling ratio of 95: 5. The 95 % output of the SCL 1 was sent to the remote site after entering an optical circulator 1 (CIR 1) and actuators for FNC explained later. The average optical power of 9 mW was launched into the fiber link, whereas the transmitted power through the fiber was eventually decreased to 2 mW by mainly optical loss associated with the fiber connectors. Moreover, the group velocity dispersion (GVD) of the fiber link stretched the pulse width out to nanoseconds, consequently the lower peak power of laser pulses. It was insufficient to draw on the potential of a highly nonlinear fiber (HNLF). The HNLF has been employed to broaden the optical spectrum of fs-pulse Er-fiber lasers. Thus, precise GVD control against the distribution fiber is crucial in recovering the ultrashort pulse and disseminating the UBOFR; fiber-length management must attain levels of 10 cm relative to the km-long distribution fiber. We adopted a three-step process for GVD control: installation of a dispersion-compensating fiber (DCF), cancellation of the overcompensated GVD, and optimization of spectral broadening. A DCF with dispersion-slope compensation was inserted between the CIR 1 and the actuators for the FNC at the local site. Its dispersion and relative dispersion slope have opposite signs to those of the standard fiber and they are −34 ps/nm and 3.57 × 10−3/nm at 1545 nm, respectively (Fujikura, DC-C+L-N017-UW). Its insertion loss was measured to be 2.4 dB. The length of the DCF was designed to overcompensate the magnitude of the GVD concerning the fiber link. This intentional overcompensation toward normal dispersion regime facilitated the minimization of a net dispersion and relative dispersion slope in the vicinity of 1570 nm by simply adding an extra standard fiber with anomalous dispersion. By following the strategy, the DCF length was set to attain near-perfect chirp compensation of the fs pulse transmitted through a 2 km SM fiber. After passing CIR 2 and SCL 2 at the remote site, the pulse shape was measured by an autocorrelator to minimize the overcompensated GVD after adding the SM fiber for dispersion management. To decide the fiber length, the pulse width was monitored simultaneously with changing the magnitude of the anomalous dispersion, that was assisted by a low-loss variable dispersion compensator (Optohub, DE-G 019). The length of the extra SM fiber was calculated to be 380 m from this measurement. A fiber of this length was arranged in front of the CIR 2 to re-compress the pulse width to sub-picosecond (ps) level. The sub-ps pulse was amplified by an Er-doped fiber amplifier (designated as EDFA 1) and subsequently introduced into HNLF 1 with a length of 40 cm. EDFA 1 consists of two fiber-based wavelength division multiplexers and 4 m Er-doped fiber fusion-spliced between them. It was pumped by two 980 nm laser diodes from the input and output sides, both of power 400 mW. The dispersion management of the incident pulse light into EDFA 1 was provided by adjusting the length of the extra SM fiber, since it could enhance the output power from EDFA 1 [20]. Finally, the length of the SM fiber after EDFA 1 was also optimized to fully broaden the output spectrum from HNLF 1.
The cancellation of the Doppler frequency shift imposed during fiber transmission was conducted with regard to only modes of the FLFC around 1570 nm. The design of this FNC is borrowed from one previously reported, with a slight modification [12]. The optical pulses delivered to the remote site were split using SCL 2 with a coupling ratio of 95: 5. The 5 % output was amplified by EDFA 2 and sent back to the local site through CIR 2 and the same fiber link. The round-trip pulses were separated from the pulses launched into the fiber link by CIR 1. Its average optical power was about 200 μW. The optical comb of the round-trip pulses must be combined with an original FLFC for the optical phase detection technique. The original FLFC was provided from the 5 % output of the SCL 1. For the signal detection of heterodyne beat frequency fbeat 2, it was frequency-shifted by 55 MHz using an acousto-optic modulator. A manually adjustable optical delay line was used for a temporal overlap between the two pulse electric fields. The fbeat 2 signal had a SNR of more than 25 dB in a 300 kHz resolution bandwidth (RBW). The signal was bandpass-filtered and then down-converted with a local oscillator at 55 MHz to extract a control signal for the FNC. The control signal was fed back to two actuators for stabilizing the effective transmission length and thereby group delay of the distribution fiber: a voice-coil-motorized (VCM) delay-line stage and piezo-actuated fiber stretcher. The low-frequency components of the signal were directed toward the VCM stage and the high frequency ones were fed back to the fiber stretcher. The unity gain frequency of the feedback loop was about 1 kHz. The continuous operation of the FNC exceeded 3 hours, which was limited by the 6 mm stroke of the VCM stage. We note that the present design of the FNC does not compensate the phase noise originated from the EDFA 1 and HNLF 1. However, they have been determined to be a negligible level by the result of a preceding measurement, as explained in the following section.
The UBOFR disseminated was compared with a reference related to the original FLFC to characterize the fiber-induced noise. The FLFC emitted from Port 3 was spectral-broadened to yield the reference. A comparison of the disseminated UBOFR and the broadened original FLFC was implemented with three auxiliary flywheel lasers oscillating at different wavelengths: 1310 nm, 1565 nm and 2051 nm. The two flywheel lasers with the shorter wavelengths were extended cavity diode lasers and the one with the longest wavelength was a Tm, Ho: YLF solid-state laser. They were phase-locked to a nearby mode of the UBOFR by forcing their heterodyne beatnote frequency fbeat 3 to oscillate at 30 MHz. The control bandwidth of the phase-locked loop for the two diode lasers and the solid-state laser were about 700 kHz and 2 kHz, respectively. The relative linewidth and fractional frequency instability of these lasers to the mode of the UBOFR were below 1 Hz and 2 × 10−15/τ in the Allan standard deviation, where τ is the averaging time. The performance limitation of the UBOFR was determined by measuring of the heterodyne beat fbeat 4 between the output of the flywheel lasers and broadened original FLFC. The influence of the fiber-induced noise was determined in terms of the linewidth, frequency instability and single-sideband (SSB) phase noise. Optical bandpass filters transmitting the aimed wavelength bands were used to detect the beat signals at fbeat 3 and fbeat 4 with high SNR.
3. Performance of an optical-comb-based ultra-broadband frequency reference transfer
The output spectrum from the HNLF 1 was measured by an optical spectrum analyzer. Figure 2(b) shows a typical spectrum of the UBOFR disseminated via the 1.8 km long fiber link. Formed by approximately 5 × 105 optical-comb modes, the spectrum ranges from 1160 to 2180 nm at −20 dB below its maximum at 1570 nm. To the best of our knowledge, such spectral coverage is the first report in optical frequency reference transfer via a km-long fiber link without using additional fs-pulse lasers at the remote site. For reference, the spectrally broadened original FLFC is depicted in Fig. 2(c).
The fbeat 4 signals were produced by photo-mixing of each flywheel laser and the original FLFC broadened at the local site. Their radio-frequency (rf) spectra were measured by a spectrum analyzer as shown in Fig. 3. They have a SNR of better than 35 dB with a 300 kHz RBW. As the Doppler-frequency modulation was added to the disseminated UBOFR during fiber transmission, these linewidths without the FNC became considerably broader than the narrow-linewidth laser serving as the optical frequency standard. The linewidths observed were about 100 Hz as plotted by dotted lines in Fig. 3. They are, however, still useful for many practical applications except for the comparison of optical clocks. The linewidths were reduced to 1 Hz with the activation of the FNC, which were restricted by the measurement resolution of the rf spectrum analyzer (bold lines in Fig. 3). The line noise at 100 Hz and its harmonics in Figs. 3(b), 3(d), and 3(f) are caused by the phase-locking electronics for flywheel lasers, although they disappeared in Figs. 3(a), 3(c), and 3(e) due to the averaging process of the noise power in the wider RBW of 1 kHz.
Fig. 3 Heterodyne beat spectra between the original FLFC broadened at the local site and flywheel lasers phase-locked to a nearby mode of the ultra-broadband optical reference disseminated through the fiber link. They were measured at wavelengths of 1310 nm (a, b), 1565 nm (c, d), and 2051 nm (e, f). The beat spectra with and without the fiber-induced noise cancellation are plotted by bold and dotted lines, respectively. RBW: resolution bandwidth, VBW: video bandwidth.
The fbeat 4 associated with each flywheel laser was recorded by a zero-dead time frequency counter having a measurement bandwidth of more than 1 kHz (K+K Messtechnik, FXE). The fractional frequency instabilities of fbeat 4 are shown in Fig. 4(a). The instabilities without the FNC were below 2 × 10−14 at a 1 s averaging time, and remained at the 10−15 level until 2000 s (open diamonds). With the activation of FNC, they were stabilized to below 3 × 10−15 at 1 s and dropped to 10−18 with a dependence of 3×10−14/τ (filled circles). The measurement noise floor, indicated by a blue line in Fig. 4(a), was measured at 1565 nm by replacing the 1.8 km-long fiber link with a 2 m-long fiber and adjusting the detected pulse power to the same operational power. The instabilities with the FNC were dominated by measurement noise over 100 s. Since these sampled wavelengths span nearly the full spectral region of the disseminated UBOFR, their measured linewidths and frequency instabilities are deduced as reflecting those of all its modes. There exists some discrepancies between the instabilities of each wavelength below 100 s. The discrepancy below 1 s resulted from the carrier-to-noise ratio of each fbeat 4 signal, whereas that between 1 s and 100 s was caused by the change in laboratory environmental conditions at the each measurement. Note that the lower instability at 1310 nm and 2051 nm than the measurement noise floor at 1565 nm as well as the higher instability in freerun at 1565 nm than those at the other wavelengths arose from this laboratory condition change. The instability of a combination of EDFA 1 and HNLF 1 was measured to be below 4 × 10−16 at 1 s and decreased to 10−18-level within 1000 s as plotted by a red-dashed line in Fig. 4 (a), and thus it is not significant in the present instability of the UBOFR. Figure 4(b) shows the SSB phase noise spectra for the 1565 nm mode of the UBOFR disseminated. They were measured by a phase comparison of fbeat 4 and the output from an rf synthesizer externally-referenced to a microwave standard. To overcome the large frequency excursions, the fbeat 4 at 30 MHz was divided by 200 using a direct digital synthesizer, which can generate an arbitrary-frequency output from the fbeat 4 as the input clock signal. The output of a double-balanced mixer for the phase comparison was sent to a fast Fourier transform analyzer. Whereas the SSB phase noise without FNC was more than −10 dBc/Hz at the 10 Hz from the carrier (green dotted line), it decreased to −50 dBc/Hz using active FNC (green bold line). Below 10 Hz, the phase noise was dominated by the measurement noise floor (plotted blue line). This feature of the SSB phase noise was consistent with the measurement of frequency instabilities. The electronic noise level of the system was not significant for these measurements (black dotted line).
Fig. 4 Allan standard deviations (a) and single-sideband phase noise spectra (b) of sampled optical modes comprising the ultra-broadband optical reference. (a) filled circles: fiber-induced noise cancellation (FNC) activated, open diamonds: FNC inactivated. The blue and red-dashed lines indicate the measurement noise floor at 1565 nm and the instability of a combination of EDFA 1 and HNLF 1, respectively. (b) green bold line: FNC activated, green dashed line: FNC inactivated, blue line: measurement noise floor and black line: system electric noise level. These were determined at a wavelength of 1565 nm.
4. Application to Q-switched 2 μm pulse laser for coherent LIDAR system
At NICT, a coherent CO2 DA/DW LIDAR system has been operating to observe the time variation of the CO2 concentration in air as well as the temporal and spatial distribution of wind [21]. Its target is the R (30) absorption line of the (20°1)III ← (00°0) band of CO2 at 2050.967 nm. The R (30) line meets the criteria for LIDAR measurements because of its absorption depth and low temperature dependence. This coherent LIDAR system employs a single-axial-mode pulse laser with high energy and low frequency noise; atmospheric backscattered light produced from the high-energy pulse laser must be photo-mixed with a stable local oscillator. An injection seeding technique is employed to satisfy the requirements; while a Q-switched slave laser provides the high-energy pulses under the multi-axial-mode oscillation, a maser laser with low cw power, called a seed laser, is easily frequency-stabilized for spectroscopy of the target absorption line. The seed laser is also used as the local oscillator. The DA spectroscopy requires three optical frequency references around 2051 nm for stabilizing the frequency of the seed lasers at the center, online, and offline of the absorption spectrum. Toward improving measurement accuracy of the coherent LIDAR, an injection-seeded Q-switched 2 μm pulse laser must have an absolute frequency stabilized to within 100 kHz accuracy, a frequency drift rate of less than 10 kHz/hour, and an uninterrupted operation of over 10 hours. Hence, the frequency stabilization of the Q-switched 2 μm pulse laser is regarded as a suitable demonstration platform for the dissemination of UBOFR.
Figure 5 illustrates the experimental setup for the frequency stabilization of the injection-seeded Q-switched 2 μm pulse laser by means of the disseminated UBOFR in NICT. The local site and user end are connected by a polarization-independent SM fiber of about 900 m in length, situated in the utility corridor. In the local site, the FLFC was optical-locked to the 1538 nm narrow-linewidth laser using their heterodyne beat at fbeat 1. The absolute frequency of the narrow-linewidth laser was traceable to JST with an uncertainty of 10−15, accordingly every mode-component of the FLFC. The fs-pulse trains from the mode-locked laser were sent to the user end over the fiber link. Its average optical power was decreased to approximately 4 mW at an output of the link. The optical spectrum of the pulse trains was broadened after the dispersion management and amplification. The UBOFR has sufficient spectral coverage for providing three frequency references around 2051 nm. Since the robust operation of the UBOFR without interruption for over 10 hours was considered to be a higher-priority requirement for practical environmental observations in comparison to the others related to the frequency, FNC explained in earlier sections was not active during the demonstration taking into account the present continuous operation period with several hours. Nevertheless, this is the first convincing demonstration of UBOFR transfer, because its absolute frequency stability and drift rate without the FNC have reached the sufficiently high-precision satisfying the requirements for a Q-switched 2 μm pulse laser.
Fig. 5 Dissemination of ultra-broadband optical frequency reference (UBOFR) for the stabilization of an injection-seeded Q-switched 2 μm pulse laser. A distribution fiber with a 900 m length connects the Space-Time Standard Laboratory (local site) and Remote Sensing Fundamentals Laboratory (user end) in NICT. They are located in adjacent campuses. The UBOFR was phase-locked to the 1538 nm narrow-linewidth laser, with an absolute frequency that is traceable to Japan Standard Time. At the user end, a seed laser with a 2051 nm wavelength was stabilized to the disseminated UBOFR and then injected into a Q-switched slave laser. fhet: heterodyne beat frequency, DAQ: data acquisition system, LO: local oscillator.
The UBOFR was photo-mixed with the 2051 nm seed laser 1 to obtain their beat frequency 2 (fbeat 2). The seed laser 1 is a Tm, Ho: YLF solid-state laser. The SNR of fbeat 2 was about 40 dB at 300 kHz RBW. It was down-converted by an rf signal at 60 MHz from a direct digital synthesizer to obtain a control signal for the seed laser 1. The control signal was fed back to the frequency-modulation actuator of the seed laser 1 to be phase-locked to the UBOFR. A part of the output from the seed laser 1 was injected into a Q-switched Tm, Ho: YLF slave laser after frequency-shifting by 108 MHz. This injection-seeded Q-switched 2 μm pulse laser generated a 25 mJ output energy with a 150 ns pulse width at a 30 Hz repetition rate. Its pulse light and a fraction of the output from the seed laser 1 were superimposed to detect the heterodyne beat signal (fhet). The fhet was recorded using a data acquisition system with a high-speed sampling frequency of 400 MHz and an analog-to-digital converter resolution of 14 bit. The sample length of each data was 4096 points. A typical fhet signal recorded is shown in Fig. 6(a). The single-axial-mode operation of the Q-switched 2 μm pulse laser was confirmed from the heterodyne beat signal at 108 MHz inside the envelope associated with the pulse electric field. Its rf spectrum is also shown in Fig. 6(b), which has a Fourier transform limited linewidth.
Fig. 6 Characterization of an injection-seeded Q-switched 2 μm pulse laser phase-locked to the ultra-broadband optical frequency reference disseminated via a fiber network. (a) Typical heterodyne beat signal (fhet) between the Q-switched 2 μm pulse laser and stabilized seed laser 1 as a local oscillator. (b) Rf spectrum of the fhet. (c) Measured frequencies of Q-switched 2 μm pulse laser injection-seeded by the seed laser 1 (red line) and free-running seed laser 2 (blue line), respectively. (d) Fractional frequency instabilities as given by the Allan standard deviation. Instabilities of the Q-switched 2 μm laser injection-seeded by seed laser 1 (filled circles) and seed laser 2 (open circles), respectively. The instability of the returned comb mode at 2051 nm after passing through the 1.8 km fiber link (open diamonds). The instability of the 1538 nm narrow-linewidth laser against a cryogenic sapphire oscillator [24] (filled squares).
The frequency fluctuation of the Q-switched 2 μm pulse laser relative to the local oscillator is plotted by a red line in Fig. 6(c). This plots the frequencies of the maximum of the rf spectrum computed from the fhet recorded. Whereas a frequency drift of 17 kHz/s remained using an equivalent seed laser 2 in the free-running state (blue line), it was suppressed by employing the seed laser 1. It should be noted that this result indicates the existence of no excess frequency drift relative to the UBOFR within the measurement accuracy. There is a non-stationary frequency fluctuation with amplitude of order 1 MHz. These were caused by the short-term frequency fluctuation of the Q-switched laser, which was unable to be stabilized by the injection-seeding technique under the present conditions. According to injection-seeding theory, an mode-selection process of the slave laser takes place instead of the frequency locking process, and the slave laser oscillates eventually at the single mode nearest the frequency of the seed laser and not at the injected frequency [22, 23]. Therefore, the short-term frequency fluctuation of the slave laser still remains as that of the Q-switched laser. Figure 6(d) shows the Allan standard deviations of the Q-switched 2 μm pulse laser. The instability of the Q-switched pulse laser with seed laser 2 has 8 × 10−11 × τ dependence above 10 s, which corresponds to a frequency drift of about 17 kHz/s (open circles). This instability was decreased using seed laser 1 (filled circles), although it was not improved below 10 s because of the above-mentioned property of injection seeding. The absolute frequency measurement of the Q-switched laser was traceable to JST within an accuracy of 50 kHz. The Q-switched 2 μm pulse laser with the UBOFR disseminated was found to be available for the coherent CO2 DA/DW LIDAR system with the projected measurement accuracy.
5. Conclusion
We demonstrated the dissemination of an optical-comb-based UBOFR via a fiber network over km-long distances. Its spectrum ranged from 1160 nm to 2180 nm without additional fs-pulse mode-locked lasers at the user end. A difficulty in direct transmission of such UBOFR through the standard fibers was circumvented by the precise dispersion management of the delivered optical pulse and subsequent spectral-broadening of the retrieved ultrashort pulse at the end of the network. The linewidth and frequency instability for all components of the UBOFR were convinced from measurement results at sampled wavelengths spanning nearly the full spectrum region, and they were of order 1 Hz and 10−18 in a 5000 s averaging time with FNC technique. Toward the improvements in measurement accuracy of atmospheric greenhouse-gas dry mixing ratio, the UBOFR phase-locked to the narrow-linewidth laser, with an absolute frequency traceable to JST, was applied for frequency stabilization of the injection-seeded Q-switched 2 μm pulse laser in the coherent CO2 DA/DW LIDAR system at NICT. This new approach for dissemination of the UBOFR via fiber network has a potential to directly compare optical clocks with widely disparate clock frequencies, and enables to link diverse frequency-related equipment and devices while ensuring traceability to national standards.
References and links
1. S. M. Foreman, K. W. Holman, D. D. Hudson, D. J. Jones, and J. Ye, “Remote transfer of ultrastable frequency references via fiber networks,” Rev. Sci. Instrum. 78(2), 021101 (2007). [CrossRef] [PubMed]
2. A. D. Ludlow, T. Zelevinsky, G. K. Campbell, S. Blatt, M. M. Boyd, M. H. G. de Miranda, M. J. Martin, J. W. Thomsen, S. M. Foreman, Jun Ye, T. M. Fortier, J. E. Stalnaker, S. A. Diddams, Y. Le Coq, Z. W. Barber, N. Poli, N. D. Lemke, K. M. Beck, and C. W. Oates, “Sr lattice clock at 1 × 10−16 fractional uncertainty by remote optical evaluation with a Ca clock,” Science 319(5871), 1805–1808 (2008). [CrossRef] [PubMed]
3. A. Yamaguchi, M. Fujieda, M. Kumagai, H. Hachisu, S. Nagano, Y. Li, T. Ido, T. Takano, M. Takamoto, and H. Katori, “Direct comparison of distant optical lattice clocks at the 10−16 uncertainty,” Appl. Phys. Express 4, 082203 (2011). [CrossRef]
4. C. Lisdat, G. Grosche, N. Quintin, C. Shi, S.M.F. Raupach, C. Grebing, D. Nicolodi, F. Stefani, A. Al-Masoudi, S. Dörscher, S. Häfner, J. L. Robyr, N. Chiodo, S. Bilicki, E. Bookjans, A. Koczwara, S. Koke, A. Kuhl, F. Wiotte, F. Meynadier, E. Camisard, M. Abgrall, M. Lours, T. Legero, H. Schnatz, U. Sterr, H. Denker, C. Chardonnet, Y. Le Coq, G. Santarelli, A. Amy-Klein, R. Le Targat, J. Lodewyck, O. Lopez, and P. E. Pottie, “A clock network for geodesy and fundamental science,” arXiv:1511.07735 (2015).
5. S. T. Cundiff, J. Ye, and J. L. Hall, “Femtosecond optical frequency combs,” Rev. Sci. Instrum. 72(10), 3749–3771 (2001). [CrossRef]
6. K. W. Holman, D. J. Jones, D. D. Hudson, and J. Ye, “Precise frequency transfer through a fiber network by use of 1.5-µ m mode-locked sources,” Opt. Lett. 29(13), 1554–1556 (2004). [CrossRef] [PubMed]
7. K. W. Holman, D. D. Hudson, and J. Ye, “Remote transfer of a high-stability and ultralow-jitter timing signal,” Opt. Lett. 31(10), 1225–1227 (2005). [CrossRef]
8. G. Marra, R. Slavík, H. S. Margolis, S. N. Lea, P. Petropoulos, D. J. Richardson, and P. Gill, “High-resolution microwave frequency transfer over an 86-km-long optical fiber network using a mode-locked laser,” Opt. Lett. 36(4), 511–513 (2011). [CrossRef] [PubMed]
9. K. Jung, J. Shin, J. Kang, S. Hunziker, C.-K. Min, and J. Kim, “Frequency comb-based microwave transfer over fiber with 7 × 10−19 instability using fiber-loop optical-microwave phase detectors,” Opt. Lett. 39(6), 1577–1580 (2014). [CrossRef] [PubMed]
10. B. Ning, S.-Y. Zhang, D. Hou, J.-T. Wu, Z.-B. Li, and J.-Y. Zhao, “High-precision distribution of highly stable optical pulse trains with 8.8 × 10−19 instability,” Sci. Rep. 4, 5109 (2014).
11. C.-W. Kang, T.-A. Liu, R.-H. Shu, C.-L. Pan, and J.-L. Peng, “Simultaneously transfer microwave and optical frequency through fiber using mode-locked fiber laser,” in Conference on Lasers and Electro-Optics International Quantum Electronics Conference, OSA Technical Digest, (Optical Society of America, 2009), p. JWA70.
12. G. Marra, H. S. Margolis, and D. J. Richardson, “Dissemination of an optical frequency comb over fiber with 3 × 10−18 fractional accuracy,” Opt. Express 20(2), 1775–1782 (2012). [CrossRef] [PubMed]
13. A. D. Ludlow, M. M. Boyd, J. Ye, E. Peik, and P. O. Schmidt, “Optical atomic clocks,” Rev. Mod. Phys. 87(2), 637–701 (2015). [CrossRef]
14. N. Yamamoto and H. Sotobayashi, “All-band photonic transport system and its device,” Proc. SPIE 7235, 72350C (2009). [CrossRef]
15. M. N. Petrovich, F. Poletti, J. P. Wooler, A. M. Heidt, N. K. Baddela, Z. Li, D. R. Gray, R. Slavík, F. Parmigiani, N. V. Wheeler, J. R. Hayes, E. Numkam, L. Grüner-Nielsen, B. Pálsdóttir, R. Phelan, B. Kelly, John O’Carroll, M. Becker, N. MacSuibhne, J. Zhao, F. C. GarciaGunning, A. D. Ellis, P. Petropoulos, S. U. Alam, and D. J. Richardson, ‘Demonstration of amplified data transmission at 2 µ m in a low-loss wide bandwidth hollow core photonic bandgap fiber,” Opt. Express21(23), 28559–28569 (2013).
16. M. Seimetz, “Laser linewidth limitations for optical systems with high-order modulation employing feed forward digital carrier phase estimation,” in Proceedings of Optical Fiber Communication Conference (OFC2008), paper OTuM2.
17. J. R. Petit, J. Jouzel, D. Raynaud, N. I. Barkov, J.-M. Barnola, I. Basile, M. Bender, J. Chappellaz, M. Davis, G. Delaygue, M. Delmotte, V. M. Kotlyakov, M. Legrand, V. Y. Lipenkov, C. Lorius, L. Pépin, C. Ritz, E. Saltzman, and M. Stievenard, “Climate and atmospheric history of the past 420, 000 years from the Vostok ice core, Antarctica,” Nature 399, 429–436 (1999). [CrossRef]
18. B. I. Vasil’ev and O. M. Mannoun, “IR differential-absorption lidars for ecological monitoring of the environment,” Quantum Electron. 36(9), 801–820 (2006). [CrossRef]
19. R. W. P. Drever, J. L. Hall, F. V. Kowalski, J. Hough, G. M. Ford, A. J. Munley, and H. Ward, “Laser phase and frequency stabilization using an optical resonator,” Appl. Phys. B 31(2), 97–105 (1983). [CrossRef]
20. Y. Nakajima, H. Inaba, F.-L. Hong, A. Onae, K. Minoshima, T. Kobayashi, M. Nakazawa, and H. Matsumoto, “Optimized amplification of femtosecond optical pulses by dispersion management for octave-spanning optical frequency comb generation,” Opt. Commun. 281(17), 4484–4487 (2008). [CrossRef]
21. S. Ishii, K. Mizutani, P. Baron, H Iwai, R. Oda, T. Itabe, H. Fukuoka, T. Ishikawa, M. Koyama, T. Tanaka, I. Morino, O. Uchino, A. Sato, and K. Asai, “Partial CO2 column-averaged dry-air mixing ratio from measurements by coherent 2-µ m differential absorption and wind lidar with laser frequency offset locking,” J. Atmos. Oceanic Technol. 29, 1169–1181 (2012). [CrossRef]
22. Y. K. Park, G. Giuliani, and R. L. Byer, “Single axial mode operation of a Q-switched Nd:YAG oscillator by injection seeding,” IEEE J. Quantum Electron. QE-20, 117–125 (1984). [CrossRef]
23. J.-L. Lachambre, P. Lavigne, G. Otis, and M. Noël, “Injection locking and mode selection in TEA − CO2 laser oscillators,” IEEE J. Quantum Electron. QE-12, 756–764 (1976). [CrossRef]
24. J. G. Hartnett, C. R. Locke, E. N. Ivanov, M. E. Tobar, and P. L. Stanwix, “Cryogenic sapphire oscillator with exceptionally high longterm frequency stability,” Appl. Phys. Lett. 89(20), 203513 (2006). [CrossRef]
