We propose and demonstrate a distributed Brillouin fiber sensor using Stokes and anti-Stokes differential pulse pair based on double- sideband probe wave, in which the two sidebands of probe wave are used to balance the power of two pump pulses. The spatial resolution is determined by the slightly width difference of the two balanced pulses, without Brillouin gain spectrum broadening. The pulses perform gain-loss process in optical field before the probe signal being detected, without any post-processing or extra measurement time. The proposed technique can achieve high spatial resolution, natural Brillouin gain spectrum linewidth, normal measurement time and long sensing range simultaneously.
© 2014 Optical Society of America
Distributed optical fiber sensors based on Brillouin optical time-domain analysis (BOTDA) provide applicable solutions for the monitoring of temperature and strain in large structures like pipelines and electrical power cables. In stimulated Brillouin scattering (SBS) process, the signal gain spectrum corresponds to the convolution between pump pulse spectrum and the natural Brillouin spectrum of the fiber. The observed Brillouin spectrum broadens as the pulse width reduced to the value comparable with the acoustic relaxation time (about 10 ns), which leads to a reduced peak gain and uncertainties in the determination of the Brillouin frequency shift (BFS) . As a result, the spatial resolution of conventional BOTDA configuration was not better than 1 m. Pre-pumping techniques can achieve sub-meter spatial resolution by combining a long CW base and a short pulsed light as pump wave , in which the CW base interacts with the probe wave to generate a pre-existing acoustic wave over the entire sensing length, so that the so sustained interaction of the pulsed portion and the probe wave provides spatial information for temperature or strain measurements. However, in case of a long sensing range, the contribution of the CW base massively dominates the measured Brillouin gain spectrum (BGS), so that the minute contribution from a small stressed fiber can be hardly detected. After that, several configurations employing dark pulse  or -phase-shifted pulse  are proposed, but suffers from a problem called second echo phenomenon. This echo extends over a duration given by the acoustic lifetime and thus creates a ghost response over the subsequent measured points after dark pulse passes, which is detrimental for the measurement results. Approaches using Brillouin gain-profile tracing , deconvolution-based method  and four-section dark pulse  can mitigate the second echo but the sensing range is limited by pump depletion. The differential pulse-width pair Brillouin optical time-domain analysis (DPP-BOTDA)  provides long distance, high spatial resolution and high frequency resolution simultaneously, whereas sacrifices the measurement speed because it requires two consecutive measurements, as well as some post-processing. Additionally, based on Brillouin optical correlation domain analysis (B-OCDA), approaches can achieve 1-2 cm resolution and range of 200 m  and even 3 km  are recently reported, but spending relative long measurement time. More recently, the technique called optical differential parametric amplification (ODPA) Brillouin sensing, which is based on the simultaneously interaction of Brillouin gain and loss process, achieved 50 cm spatial resolution over a 15m polarization maintaining fiber . After that, a similar technique that uses slightly modified experiment setup, realized 10cm spatial resolution over a 12m standard single mode fiber . No matter whether the working process of those two approaches can be defined as differential parametric amplification  or just the combination of Brillouin gain and loss , the key ideas of them are the same. Both of them employ two simultaneously launched Stokes and anti-Stokes long pump pulses with same power but slightly different widths to realize the subtraction process in optical domain before the probe wave been detected, thus can achieve high spatial resolution without penalizing the measurement time. However, both of them have a limit of sensing range since the anti-Stokes and Stokes pump pulses experience Brillouin depletion and amplification all the time as they propagate along the fiber, respectively, leading to power unbalance of them. Consequently, the contribution of the Stokes pump pulse will dominate the measurement BGS at the far end of the sensing fiber, resulting in measurement error. The longer of the sensing fiber, the more Brillouin depletion and amplification will be accumulated.
In this paper, a novel distributed Brillouin sensor based on Stokes and anti-Stokes pump pulses with double-sideband probe wave is proposed, which has the same advantages with ODPA sensing while long range measurement can be realized. In our proposal, the probe wave consists of three frequency components with equal intensities, in which the carrier still acts as probe signal, while the upper sideband with twice of anti-Stokes frequency and lower sideband with twice of Stokes frequency are implemented to balance the power of pump pulses. The Brillouin depletion and amplification can be greatly mitigated [13, 14], thus the change of pulse power only depends on fiber loss of 0.2 dB/km. On this condition, the limitation of probe wave’s power and nonlocal effects in conventional BOTDA will not be concerned because they only appear when Brillouin depletion or amplification occurs. By using this technique, high spatial resolution, natural linewidth of BGS, normal measurement time and long sensing range can be achieved simultaneously. The proposal is theoretical derivated and experimental validated in this paper.
2. Working principle
The schematic model of the proposed technique is shown in Fig. 1. The anti-Stokes pump pulse at frequency and Stokes pump pulse at frequency having same power and polarization with different widths of and are launched into the sensing fiber simultaneously. is the modulation frequency, which is approximately equal to Brillouin frequency shift . The counter-propagating probe wave consists of three frequency components (probe signal spectrally positioned at frequency , the upper sideband positioned at frequency and the lower sideband positioned at frequency ) with same power, which interact with the two pump pulses. Due to the conversation laws of energy and momentum, the Brillouin interaction energy will transfer from the wave with higher frequency to the wave of lower frequency, as shown in Fig. 1, where the orange arrows represent the direction of energy transfer.
The key point of our proposal is using the two sidebands of probe wave to balance the power of the two pump pulses, maintaining high spatial resolution in case of long range sensing. Note that although the slightly unbalanced double sidebands have the best performance to balance the pump pulse power , the sidebands with equal power can also mitigate the pump depletion and nonlocal effects greatly [13, 14]. Considering the complexity and efficiency of the system, the double sidebands with equal power are employed. As can be seen in Fig. 1, the anti-Stokes pump pulse is balanced by the carrier signal and upper sideband of probe wave, and Stokes pump is balanced by carrier signal and lower sideband. During the pump and probe interaction process, the carrier signal is firstly balanced by the overlapping regions of the two pump pulses, and then amplified by the width difference of those two pulses. Thus by filtering the carrier signal, the spatial information of temperature or strain can be obtained. The overlapping regions of the two pulses are used to pre-excite steady state acoustic waves. In this technique, the spatial resolution is supposed to only depend on the section, which means the contribution of overlapping regions of the two pulses on carrier signal should be always canceled. Only the same pump pulses power can ensure this condition.
The working process can be described by using the steady-state five waves coupled intensity SBS equations because the widths of two pump pulses are both sufficiently longer than the phonon life time (about 10ns) :
Where is the fiber loss coefficient, and is the Brillouin gain coefficient. The Brillouin gain coefficient at each section depends on the mismatch between the local Brillouin frequency shift and the frequency offset between the interacting light waves.
Equations (1-2) can be solved by employing the perturbation method . The variety of the signal wave can be considered as a perturbation, so assuming the signal wave only experiences natural fiber loss. Setting the boundary conditions that , , , and , the intensities of two pump pulses can be described as:
Where the factors of and are given by:
The factors of and represent the depletion or amplification experienced by the pump pulses due to Brillouin interaction. It is noted that the size of and greatly depends on the size of and , respectively.
Assuming that the temperature or strain is uniform in each interacting distance , which means maintains a same value in each . Equation (10) could be simplified as:
Where represents the light velocity in the fiber. The detected power of carrier signal at the location z = 0 at time t is given by:
As a result, the intensity of Brillouin signal not only depends on the section , but also the section T. the longer the fiber, the large intensity difference between and . Predictably, when the sensing range is sufficient long, the contribution of will dominate the measured BGS compare with the contribution of because the width of section is constant. The following expression can be obtained:
It is clear that the measurement error occurs by using previous method in case of long sensing range.
Finally we can obtain:
The intensity of Brillouin signal only depends on section as expected.
3. Experiment results and discussion
The experiment setup was employed as shown in Fig. 2. A narrow linewidth (15KHz) tunable laser source at 1550 nm with 20dBm optical power was split into Brillouin pump and probe waves after a 50:50 coupler. The pump wave was injected into a high extinction ratio (>25dB) electro-optic modulator (EOM1). A polarization controller (PC) was positioned in front of the EOM to insure optimum polarization state of the input light. The CW pump wave is modulated by EOM1 operating at null transmission point to generate carrier-suppressed double-sideband (SC-DSB) wave. A commercial programmable optical filter (waveshaper 4000s) with 3dB filtering bandwidth of 10 GHz was employed to separate the anti-Stokes and Stokes sidebands. The two sidebands were pulsed by EOM2 and EOM3 with widths of 30 ns and 25 ns, respectively. The two pulses were amplified to 100mW peak power with aligned rising edges by properly tuning the tunable optical delay line (ODL). Then they combined with the same polarization using the two PCs and a polarizer before being launched into the start end of the fiber. In the other branch, the probe wave was first modulated by EOM4 operating at quadrature transmission point to generate the double-sideband (DSB) signal with 2mW power of each frequency component, then was scrambled by a polarization scrambler (PS) to average the effect of state of polarization (SOP), and launched into the far end of the sensing fiber through an isolator. The carrier signal component of probe wave was extracted at the fiber end using an optical circulator and another optical filter with 3dB bandwidth of 10GHz, then was detected by an AC-coupled photodiode (PD), which was connected to an oscilloscope with 1GSamples/s.
A 524m long sensing fiber was used to preliminary demonstrate our technique, which was at room temperature of 25°C corresponding to Brillouin frequency shift of 10.873GHz. A 2m long heated segment was located approximately between 509.5m and 511.5m with temperature of 45°C corresponding to 20MHz frequency shift. The position of heated segment was deliberate chosen at the far end of the sensing fiber for demonstrating the performance of spatial resolution in case of long range. The frequency launched into EOM4 from microwave generator was step-swept by 2MHz across the spectral range from 21.66GHz to 21.888GHz, and that frequency was divided into half (i.e. 1MHz across the spectral range from 10.83GHz to 10.944GHz) by a frequency divider, which was launched into EOM1 synchronously. The time traces were obtained with 4096 times averaging per trace.
The ODPA method and proposed method were both employed for comparison. The 3D-mappings of Brillouin gain versus both location and frequency shift are illustrated in Fig. 3. By using ODPA method, as shown in Fig. 3(a), the curve decreases rapidly, resulting in the deterioration of measured Brillouin gain trace, even appearing “negative gain” over the region from 300m to the end of the fiber. The reason can be explained as the Stokes and anti-Stokes pump pulses continuously experience Brillouin amplification and depletion, respectively, leading to unbalanced pulse peak power which further gives rise to unbalanced gain-loss process. As time goes on, the Brillouin loss of probe signal dominates the gain-loss process, resulting in signal distortion. In other words, the probe signal depends on not only the differential pulse width (i.e. 5ns section of anti-Stokes pulse) but also the overlapping section of the two pulses (i.e. 25ns sections of both pulses). In order to certify that the rapid decrease of the measured curve is caused by the unbalance of pump pulses, we insert the experiment parameters (i.e. intensity of probe signal, widths of two pump pulses) of this experiment into Eq. (6-9) to calculate the depletion value. The result shows that the gain becomes negative (i.e. condition of Eq. (16) occurs) at the position of about 360m distance, which well accords with the experimental results showed in Fig. 3(a). The result of our technique is ideal, as shown in Fig. 3(b).
To observe clearly, the top view of the two 3D-mappings are also given, as shown in Fig. 4(a) and Fig. 4(b), respectively. We can observe the changes of Brillouin gain around heated section in both figures but the change in Fig. 4(a) manifests negative.
For further investigating the performance on spatial resolution and BGS linewidth of our technique, the measured signal time domain traces at the frequency shift tuned to the Brillouin frequency corresponding to the temperature of heated segments are shown in Fig. 5(a), in which the blue, red and black traces represent the results using conventional BOTDA technique with 30 ns pump pulse, our proposed technique with 30/25 ns pump pulse pair and ODPA technique with 30/25 ns pump pulse pair, respectively. Observing the 10% to 90% response of temperature step near the heated segment in the trace is the general way to evaluate the spatial resolution . When the length of heated segment is shorter than the theoretical spatial resolution calculated using pulse width, the 10% to 90% response represents the length of heated segment directly. As can be seen in Fig. 5(a), the 10% to 90% response of our proposed technique is 0.5m, which accords with the theory. While the response of conventional BOTDA curve is 2m because the 2m heated segment is shorter than the corresponding 3m spatial resolution. The “negative” 10% to 90% response of ODPA technique is also 2m because the Stokes pump pulse dominates the gain-loss process, the change of probe signal is almost determined by the width of Stokes pump pulse, corresponding to 2.5m spatial resolution, which is also longer than 2m heated segment. Note that in our proposed technique, the measured heated section is shifted by 2.5m that corresponds to the 25ns Stokes pump pulse of the pulse pair. In order to achieve correct distributed information, we should subtract an amount that corresponds to the smaller pulse width of the used pulse pair from all the location information. Figure 5(b) shows the measured BGS and fitting BGS curves at 500m (at room temperature) and 513m (heated segment), which correspond to real locations at the sensing fiber of 497.5m and 510.5m by subtracting 2.5m length as mentioned above. The two BGSs are both with natural FWHM linewidth of about 35MHz, demonstrating that our technique can achieve high spatial resolution and natural BGS linewidth simultaneously at the far end of the fiber.
In conclusion, a 524m sensing range distributed Brillouin sensor with 0.5m spatial resolution based on anti-Stokes and Stokes pump pulse pair and double-sideband prove wave was theoretically derivated and experimentally demonstrated. In this technique, the impact of the overlapping sections of two pulses on the prove signal was canceled in optical field without post processing or doubled measurement time, and the power of two pump pulses were automatically balanced by the upper and lower sidebands of probe wave. We experimentally confirmed that this technique can achieve high spatial resolution, natural BGS linewidth, normal measurement time and long sensing range simultaneously. Due to the good performance on those aspects, we expect this technique can help the development of future practical distributed sensing systems.
This work is supported by 863 Program under Contract 2013AA014202 and the 111 Project (B07005) of State Key Lab of Information Photonics and Optical Communications (Beijing University of Posts and Telecommunications).
References and links
1. A. Fellay, L. Thévenaz, and M. Facchini, “Distributed sensing using stimulated Brillouin scattering: towards ultimate resolution,” in Proceedings of 12th International Conference on Optical Fiber Sensors, 324–327(1997).
2. X. Bao, A. Brown, M. Demerchant, and J. Smith, “Characterization of the Brillouin-loss spectrum of single-mode fibers by use of very short (<10-ns) pulses,” Opt. Lett. 24(8), 510–512 (1999). [CrossRef] [PubMed]
3. A. W. Brown, B. G. Colpitts, and K. Brown, “Dark-pulse Brillouin optical time-domain sensor with 20-mm spatial resolution,” J. Lightwave Technol. 25(1), 381–386 (2007). [CrossRef]
4. S. F. Mafang, M. Tur, J. C. Beugnot, and L. Thévenaz, “High spatial and spectral resolution long-range sensing using Brillouin echoes,” J. Lightwave Technol. 28(20), 2993–3003 (2010). [CrossRef]
5. T. Sperber, A. Eyal, M. Tur, and L. Thévenaz, “High spatial resolution distributed sensing in optical fibers by Brillouin gain-profile tracing,” Opt. Express 18(8), 8671–8679 (2010). [CrossRef] [PubMed]
6. J. C. Beugnot, M. Tur, S. F. Mafang, and L. Thévenaz, “Distributed Brillouin sensing with sub-meter spatial resolution: modeling and processing,” Opt. Express 19(8), 7381–7397 (2011). [CrossRef] [PubMed]
7. Z. Yang, X. Hong, J. Wu, H. Guo, and J. Lin, “Distributed Brillouin sensing with sub-meter spatial resolution based on four-section pulse,” in Optical Fiber Communication Conference, Technical Digest (CD) (Optical Society of America, 2013), paper OM3G.3. http://www.opticsinfobase.org/abstract.cfm?URI=OFC-2013-OM3G.3 [CrossRef]
9. A. Zadok, Y. Antman, N. Primerov, A. Denisov, J. Sancho, and L. Thévenaz, “Random-access distributed fiber sensing,” Laser Photon. Rev. 6(5), L1–L5 (2012). [CrossRef]
10. A. Denisov, M. A. Soto, and L. Thévenaz, “Time gated phase-correlation distributed Brillouin fiber sensor,” Proc. SPIE 8794, Fifth European Workshop on Optical Fibre Sensors, 87943I (2013). [CrossRef]
11. Y. Li, X. Bao, Y. Dong, and L. Chen, “A novel distributed Brillouin sensor based on optical differential parametric amplification,” J. Lightwave Technol. 28(18), 2621–2626 (2010). [CrossRef]
12. A. Motil, O. Danon, Y. Peled, and M. Tur, “High spatial resolution BOTDA using simultaneously launched gain and loss pump pulses,” Proc. SPIE 8794, 87943L (2013). [CrossRef]
13. R. Bernini, A. Minardo, and L. Zeni, “Pump depletion reduction technique for extended-range distributed Brillouin fiber sensors,” Proc. SPIE 7356, 73560L (2009). [CrossRef]
16. R. W. Boyd, Nonlinear Optics (Academic, 2003) 3th ed.
17. R. Bernini, A. Minardo, and L. Zeni, “Reconstruction technique for stimulated Brillouin scattering distributed fiber-optic sensors,” Opt. Eng. 41(9), 2186–2194 (2002). [CrossRef]
18. M. A. Soto, M. Taki, G. Bolognini, and F. Di Pasquale, “Optimization of a DPP-BOTDA sensor with 25 cm spatial resolution over 60 km standard single-mode fiber using Simplex codes and optical pre-amplification,” Opt. Express 20(7), 6860–6869 (2012). [CrossRef] [PubMed]