Multi-wavelength in-band OSNR monitor based on Lyot-Sagnac interferometer

Jifang Qiu; Zhuili Huang; Bo Yuan; Na An; Deming Kong; Jian Wu

doi:10.1364/OE.23.020257

1. Introduction

Optical performance monitoring (OPM) is a promising technology to improve control of transmission- and physical-layer fault management, which is essential for the operation of complex and transparent wavelength-division-multiplexing (WDM) transmission and switching systems [1,2 ]. Within various OPM techniques, monitoring of the optical signal-to-noise ratio (OSNR) is very important as OSNR is one of the most fundamental parameters indicating the signal transmission performance. Since traditional out-band OSNR monitoring is not accurate in a dynamic WDM network with optical add-drop modules (OADM) due to the intermediate filtering, a number of techniques have been reported to monitor the in-band OSNR, such as polarization nulling [3], waveform sampling [4], interferometry [5–7 ] and the approaches based on nonlinear effects [8–10 ]. Among them, OSNR monitor based on interferometry is very promising because that it exploits the coherence difference between signal and amplified spontaneous emission (ASE) to calculate the OSNR, making it intrinsically independent to chromatic dispersion (CD), polarization mode dispersion (PMD) and polarized noise [7]. Plus, it has advantages of simple setup, convenient operation and easy maintenance. However most interferometry-based methods, using single interferometer, require the prior knowledge of the noise-free coherence properties of the signal and therefore basically require system calibration by turning off the noise and scanning signal interference, which is impossible in practical networks. To overcome this problem, a modified scheme based on a pair of Michelson interferometers with different optical delays [6] has been reported.

On the other hand, monitoring WDM signal in a single operation is of interest for reducing measurement latency, providing a more scalable solution. Achieving this without compromising the measurement sensitivity and inter-channel cross talk is also important. There are few papers reporting OSNR monitoring for WDM signals [7,8 ]. However [7], using a delay-line interferometer (DLI) requires turning off the noise to measure the signal distribution ratio [8]; exploiting Stimulated Brillouin Scattering requires high input optical power and only monitors the OSNR by tracking the optical power of the back-scattered stokes wave rather than giving the OSNR value.

Therefore, in this paper, we propose and experimentally demonstrate a novel in-band OSNR monitoring scheme for 4 × 40 GBaud non-return-to-zero quadrature-phase-shift-keying (NRZ-QPSK) WDM signal, the preliminary results of which was report in [11]. By using a phase modulator-embedded Lyot-Sagnac interferometer, the scheme requires no prior knowledge of the noise-free coherence properties of the signal. Experimental results show that OSNR measurements from 7.5 to 25 dB within error of ± 0.5 dB were achieved for all four channels. Results also show that the monitor was insensitive to CD, PMD, and polarized noise. Moreover, the monitor worked very well even with input optical power as low as −8.24 dBm, and also worked in C-band. It is interesting to know that the scheme is able to measure OSNR for polarization-division-multiplexing (PDM) signals.

2. Principle of operation

The schematic of the proposed OSNR monitor is shown in Fig. 1(a) , which is composed of a 3dB coupler, a length of polarization maintaining fiber (PMF, with a length of $L_{1}$ , and birefringence of $Δ n_{1}$ ), a PMF-pigtailed phased modulator ( $L_{2}$ , $Δ n_{2}$ ), and two polarization controllers (PCs, PC1 and PC2) acting as half-wave plates (can be replaced by polarization switches). $θ_{1}$ and $θ_{2}$ are the relative angle directions of the phase modulator (including its pigtails) and the PMF respectively. With different polarization alignment between the PMF and the phase modulator, optical signals propagating in two directions experience different time delays, due to the intrinsic birefringence inside the loop induced by the PMF and the phase modulator. For example, when the two polarizations in the PMF and the phase modulator are parallel to each other ( $θ_{1} = θ_{2}$ ), the time delay is $Δ τ_{1}$ , as shown in Fig. 1(b). While, if the two polarizations are orthogonal to each other ( $θ_{1} - θ_{2} = 90^{\circ}$ ), the time delay is $Δ τ_{2}$ , as illustrated in Fig. 1(c). Note that the different delay times result in different free spectral ranges (FSRs) of the transmission spectra of the Lyot-Sagnac loop. In each case ( $Δ τ_{1}$ or $Δ τ_{2}$ ), periodically altering the voltage applied to the phase modulator changes the phase difference between the counter-propagating signals, thus leads to periodic shift in the transfer functions. $Δ τ_{1}$ and $Δ τ_{2}$ can be calculated as $Δ τ_{1, 2} = \frac{Δ n_{1} L_{1} \pm Δ n_{2} L_{2}}{c}$ [12].

Fig. 1 (a) Schematic of the proposed OSNR monitor. Transmission spectra when the axes between the PMF and PMF-pigtailed phased modulator are (b) parallel, (c) orthogonal. PC: polarization controller, PMF: polarization maintaining fiber, Mod.: modulator.

Download Full Size | PDF

With each time delay $Δ τ$ , the OSNR can be calculated from [6]

O S N R = 10 \log_{10} (\frac{1}{r} \frac{N E B}{0.1 (n m)}) = - 10 \log_{10} (r) + 10 \log_{10} (\frac{N E B}{0.1 (n m)})

r = P_{n} / P_{s} = (γ_{s} (Δ τ) - μ) / (μ - γ_{n} (Δ τ))

where

P_{s}

and

P_{n}

are the signal and noise powers respectively,

N E B

is the noise equivalent bandwidth of the channel filter.

μ = (P_{\max} - P_{\min}) / (P_{\max} + P_{\min})

,

P_{\max}

and

P_{\min}

are the maximum and the minimum output powers tracked by an optical power meter when the signal experiences constructive or destructive interference, respectively.

γ_{n} (Δ τ)

is the normalized autocorrelation function of the noise,

γ_{s} (Δ τ)

is the normalized autocorrelation function of the signal which has to be measured by turning off the noise.

γ_{s} (Δ τ)

can be expended as

γ_{s} (Δ τ) = 1 - \sum_{i = 1}^{I} c_{i} Δ τ_{}^{2 i}

. In the case of first-order approximation,

γ_{s} (Δ τ) \approx 1 - c_{1} Δ τ_{}^{2}

, in which

Δ τ = Δ τ_{j}

(

j = 1, 2

). Substituting

γ_{s} (Δ τ) = γ_{s} (Δ τ_{j}) \approx 1 - c_{1} Δ τ_{j}^{2}

(

j = 1, 2

) into the Eq. (2), we can get the two equations,

c_{1} Δ τ_{1}^{2} + r (μ_{1} - γ_{n} (Δ τ_{1})) = 1 - μ_{1}

c_{1} Δ τ_{2}^{2} + r (μ_{2} - γ_{n} (Δ τ_{2})) = 1 - μ_{2}

where

μ_{j} = (P_{\max_j} - P_{\min_j}) / (P_{\max_j} + P_{\min_j})

(

j = 1, 2

),

P_{\max_j}

and

P_{\min_j}

are the maximum and the minimum output powers tracked by an optical power meter when the signal experiences constructive or destructive interference when

Δ τ = Δ τ_{j}

(

j = 1, 2

), respectively.

From the two equations, it is possible to solve the unknown values of $r$ and $c$ . Therefore, the measured OSNR is calculated from Eq. (1) without prior knowledge of the noise-free coherence properties of the signal [6].

If a WDM demultiplexer is added at the output of the Lyot-Sagnac loop, then the OSNR of the multi channels can be measured. Note that multi power meters have to be connected to the WDM demultiplexer to track the maximum and minimum optical powers of each channel.

3. Experimental setup

The proposed scheme is demonstrated in an experimental setup shown in Fig. 2 . Four-channel WDM signal was generated by using four continuous-wave lasers at wavelengths, $λ_{1 - 4}$ , of 1547.72 nm, 1549.32 nm, 1550.92 nm and 1552.52 nm, separated by 200 GHz on the ITU grid. The four continuous-wave lights were combined by polarization maintaining couplers and modulated by an IQ modulator driven by 40 Gbit/s pseudo-random binary sequence (PRBS) with a length of 2⁷-1 to generate 4 × 40 GBaud NRZ-QPSK signals. The erbium-doped fiber amplifiers (EDFAs) were used to produce ASE noise, which was then combined with the modulated signals via a 3 dB fiber coupler. The noise from EDFA2 polarized by a polarizer before combining with noise from EDFA1 was to generate different degree of polarization of noise, in order to investigate the effect of polarized noise on the performance of the OSNR monitor. Attenuators after the EDFAs controlled the amount of ASE to form different values of OSNR. The noisy signal was then sent into both the OSNR monitor and an optical spectrum analyzer (OSA), for comparison. The OSA was used to measure the reference OSNR. In the experiment, we used only one power meter after the WDM demultiplexer (with 3-dB bandwidth of 2 nm) and shifted it alternately to four channels to track the maximum and the minimum output powers of each channel when periodically altering the voltage applied to the phase modulator. In order to investigate the effect of PMD and CD on the OSNR monitor, we used a PMD emulator and different lengths of single mode fiber (SMF) to add different amount of first-order PMD and CD to the noisy signal separately. The lengths of the PMF and the phase modulator were chosen to be ~1 m and ~4.2 m respectively, considering the OSNR measurement accuracy; the total length of the loop was ~16.5 m. Note that we added different lengths of SMF into the loop, and found that the length of the fiber loop did not affect the transmission spectra at all.

Fig. 2 Experimental setup. Att.: optical attenuator, OSA: optical spectrum analyzer, PMD: polarization mode dispersion, SMF: single mode fiber, Pol.: polarizer, MZM: Mach-Zehnder modulator.

Download Full Size | PDF

4. Experimental results and discussions

Figures 3(a) and 3(b) show the experimental measurement of the two distinct periods of transmission spectra of the Lyot-Sagnac interferometer in the case of $θ_{1} - θ_{2} = 90^{\circ}$ , and $θ_{1} = θ_{2}$ , respectively. The solid and dotted lines stand for the transmission spectra when destructive and constructive interferences occur respectively. We can see that the FSRs of the transfer function are $Δ λ_{1}$ = 1.91 nm and $Δ λ_{2}$ = 1.19 nm, corresponding to time delays of $Δ τ_{1}$ = 4.19 ps and $Δ τ_{2}$ = 6.72 ps, respectively. $γ_{n} (Δ τ_{1})$ in Eq. (3), and $γ_{n} (Δ τ_{2})$ in Eq. (4) are calculated from the Fourier transform of the filter function as shown in Fig. 3(c), as $γ_{n} (Δ τ_{1})$ = 0.06358 and $γ_{n} (Δ τ_{2})$ = −0.1445 respectively.

Fig. 3 Experimental measurement of the two distinct periods of transmission spectra of the Lyot-Sagnac interferometer when (a) $θ_{1} - θ_{2} = 90^{\circ}$ , (b) $θ_{1} = θ_{2}$ ; solid and dotted lines stand for the transmission spectra when destructive and constructive interferences occur. (c) Fourier transform of the filter function showing $γ_{n} (Δ τ)$ and the filter function (inset).

Download Full Size | PDF

Figure 4(a) shows the measured OSNR versus the reference OSNR for four channels, the errors are given as well. We can find that OSNR measurement from 7.5 to 25 dB within error of $\pm$ 0.5 dB is achieved for all channels in the presence of neighboring channels. As the most concerned OSNR range is from 10 to 25 dB, thus our scheme is effective enough for monitoring OSNR. Figure 4(b) shows the measured OSNR versus the reference OSNR of one channel (1550.92 nm) when the OSNRs of the other three channels were fixed at 5.4 dB (1547.72 nm), 15.6 dB (1549.32 nm) and 9.9 dB (1552.52 nm), respectively, indicating that the OSNR of adjacent channels does not affect the OSNR measurement at all.

Fig. 4 (a) Measured OSNR and error versus reference OSNR of four channels; (b) Measured OSNR and error versus reference OSNR of one channel while the OSNRs of the other channel were fixed.

Download Full Size | PDF

Then, we investigated the performance of our OSNR monitor by changing the input optical power, tuning the wavelength, adding different amounts of CD or PMD, and adding partially polarized ASE noise. Note that only channel of 1550.92 nm was investigated, the other three channels were blocked during this investigation. The results are shown in Fig. 5 and Fig. 6 . We can see from Fig. 5(a), the error as a function of the input optical power at OSNR of 20 dB, that the error changed little when the input optical power changing from −8.24 dBm to −0.81 dBm. The measured monitoring error versus wavelength is shown in Fig. 5(b). The error kept within $\pm$ 0.5 dB during the whole C-band, indicating that this scheme can be used to measure even more WDM channels. Note that since the basic principle of the OSNR monitor is based on the coherence difference between signal and ASE to calculate the OSNR, it is only valid for monitoring WDM channels without inter-channel crosstalk, because the unwanted crosstalk signal still has relatively strong coherence, as a result would be falsely treated as signal by the monitor. As shown in Figs. 6(a) and 6(b), the changes in error are negligible even when the signal is added with CD as large as 1152 ps/nm, or with DGD as large as 100 ps, indicating that the influence of CD and PMD on the measured OSNR can be neglected.

Fig. 5 Measured OSNR and the monitoring errors when (a) the OSNR was fixed at 20 dB, the input power was changed from −8.24 dBm to −0.81 dBm; (b) signal wavelength was tuned from 1530 nm to 1565 nm, when the OSNR was fixed at 15 dB and 20 dB.

Download Full Size | PDF

Fig. 6 Measured OSNR and the monitoring errors when (a) the input signal was added with CD from 0 to 1152 ps/nm, the OSNR was fixed at 5 dB and 15 dB respectively; (b) with DGD from 0 to 100 ps, the OSNR was fixed at 10 dB and 20 dB respectively; (c) in the cases of adding unpolarized, partially polarized and or polarized noises.

Download Full Size | PDF

The influence of polarized ASE to our scheme was investigated as well. Partially polarized ASE was generated by combining the output of a polarized ASE source (EDFA2) with the output of an unpolarized ASE source (EDFA1) using a 3 dB coupler. Figure 6(c) shows the measured OSNR versus reference OSNR when the noise is fully polarized, partially polarized or unpolarized. The three curves are in superposition, indicating that the state of polarization of the noise did not affect the results at all.

It is interesting to know that when we changed the polarization of the noisy signal in front of the monitor, the measured OSNR did not change at all, which can be theoretically explained as follow.

We expand the noisy signal with arbitrary polarization as a superposition of two linearly orthogonal polarizations, one of which is horizontal to the fast axis of the PMF-pigtailed phase modulator, named as $P o l . x$ , another is vertical to the fast axis of the PMF-pigtailed phase modulator, named as $P o l . y$ , respectively. The noisy signal input into the Lyot-Sagnac loop propagates in two directions, as clockwise (CW) and counter-clockwise (CCW). As mentioned in above operation principle, optical signals propagating in two directions (CW and CCW) experience different time delays, we can calculate the time delays for two orthogonal polarizations ( $P o l . x$ and $P o l . y$ ) separately as follows.

When $θ_{1} = θ_{2}$ , for $P o l . x$ , CW and CCW light experience time delay as, (note that CCW light is aligned to the slow axis of the PMF by adjusting PC1 in Fig. 1)

\begin{matrix} Δ τ_{1_P o l . x} = | τ_{P o l . x}^{c w} - τ_{P o l . x}^{c c w} | \\ = | \frac{L_{2} n_{21} + L_{1} n_{11}}{c} - \frac{L_{1} n_{12} + L_{2} n_{22}}{c} | \\ = | \frac{L_{1} (n_{11} - n_{12}) + L_{2} (n_{21} - n_{22})}{c} | \\ = \frac{L_{1} Δ n_{1} + L_{2} Δ n_{2}}{c} \end{matrix}

where $τ_{P o l . x}^{c w}$ and $τ_{P o l . x}^{c c w}$ are the times of $P o l . x$ propagating in CW and CCW directions in the loop respectively, $n_{11}$ and $n_{12}$ are the refractive indexes of the fast and slow axis of the PMF respectively, $n_{21}$ and $n_{22}$ are the refractive indexes of the fast and slow axis of the PMF-pigtailed phase modulator respectively.

For $P o l . y$ , CW and CCW light experience time delay as,

\begin{matrix} Δ τ_{1_P o l . y} = | τ_{P o l . y}^{c w} - τ_{P o l . y}^{c c w} | \\ = | \frac{L_{2} n_{22} + L_{1} n_{12}}{c} - \frac{L_{1} n_{11} + L_{2} n_{21}}{c} | \\ = | - \frac{L_{1} (n_{11} - n_{12}) + L_{2} (n_{21} - n_{22})}{c} | \\ = \frac{L_{1} Δ n_{1} + L_{2} Δ n_{2}}{c} \end{matrix}

Clearly, $Δ τ_{1_P o l . x} = Δ τ_{1_P o l . y} = Δ τ_{1}$ , which indicates that the two polarizations experience the same time delay, have the same transmission spectra, experience the same interferometry. The calculation of time delays for the case of $θ_{1} - θ_{2} = 90^{\circ}$ is likewise.

For $P o l . x$ , CW and CCW light experience time delay as,

\begin{matrix} Δ τ_{2_P o l . x} = | τ_{P o l . x .2}^{c w} - τ_{P o l . x .2}^{c c w} | \\ = | \frac{L_{2} n_{21} + L_{1} n_{12}}{c} - \frac{L_{1} n_{11} + L_{2} n_{22}}{c} | \\ = | - \frac{L_{1} (n_{11} - n_{12}) - L_{2} (n_{21} - n_{22})}{c} | \\ = | \frac{L_{1} Δ n_{1} - L_{2} Δ n_{2}}{c} | \end{matrix}

For $P o l . y$ , CW and CCW light experience time delay as,

\begin{matrix} Δ τ_{2_P o l . y} = | τ_{P o l . y .2}^{c w} - τ_{P o l . y .2}^{c c w} | \\ = | \frac{L_{2} n_{22} + L_{1} n_{11}}{c} - \frac{L_{1} n_{12} + L_{2} n_{21}}{c} | \\ = | \frac{L_{1} (n_{11} - n_{12}) - L_{2} (n_{21} - n_{22})}{c} | \\ = | \frac{L_{1} Δ n_{1} - L_{2} Δ n_{2}}{c} | \end{matrix}

Likewise, we get $Δ τ_{2_P o l . x} = Δ τ_{2_P o l . y} = Δ τ_{2}$ from Eqs. (7) and (8) , which indicates that the two polarizations experience the same time delay, have the same transmission spectra, experience the same interferometry.

In order to prove above theoretical analysis, we investigated the effect of input polarization on the OSNR measurement by using a modified experimental setup shown in Fig. 7 , only one channel with wavelength of 1550.92 nm was used for simplicity. The OSNR-degraded noisy signal was generated by combining a 40 Gbaud NRZ-QPSK signal with ASE and then filtered by a filter with 3-dB bandwidth of 2 nm. An OSA was used to measure the actual OSNR of the noisy signal for reference. After passing through a polarization controller (PC), half of the noisy signal was injected into our OSNR monitor for OSNR measurement, another half was used for recording the change of input polarization by using a polarizer and a power meter. The change of the polarized optical power $P_{P o l .}$ means the change of the input polarization of the noisy signals into our OSNR monitor scheme. By adjusting the PC to change the input polarization, the $P_{P o l .}$ can reach its maximum or minimum, which means the signal passes the polarizer or is blocked thoroughly. Figure 8 shows the measured OSNR error as a function of $P_{P o l .}$ for different OSNR levels. It can be seen that all the errors are less than 0.5 dB for OSNR of 10.5 dB, 15.6 dB and 20.5 dB, indicating that our OSNR monitor is insensitive to the input polarization. Note that in Fig. 8 the maximum (minimum) values of $P_{P o l .}$ are different for different OSNR levels, because $P_{P o l .}$ includes not only the signal power but also the noise power, which is different for different OSNR levels. In conclusion, this OSNR monitor is capable of measuring OSNR of PDM signals.

Fig. 7 Experimental setup for investiging the effect of input polarization on OSNR measurement. Att.: optical attenuator, OSA: optical spectrum analyzer, MZM: Mach-Zehnder modulator.

Download Full Size | PDF

Fig. 8 OSNR measurement error as a function of polarized input optical power.

Download Full Size | PDF

5. Conclusion

We propose a novel in-band OSNR monitoring scheme based on a phase modulator-embedded Lyot-Sagnac interferometer. By introducing two different time delays in the Lyot-Sagnac loop, the scheme requires no prior knowledge of the noise-free coherence properties of the signal. Experimental results show that this monitor allows measurement of OSNR ranging from 7.5 to 25 dB within error of ± 0.5 dB for 4 × 40 GBaud NRZ-QPSK signals separated at 200 GHz ITU grid. The monitor was experimentally proved to be robust to CD, PMD, polarized noise and input optical power. As the OSNR monitoring scheme is wavelength insensitive, it can be used for more channels in C-band (more PDs/power meters need to be added too). It is also interesting to know that the scheme is able to measure OSNR for polarization multiplexed signals, based on the theoretical analysis of the Lyot-Sagnac loop and experimental results.

Acknowledgments

This work is partly supported by NSFC program (61205031, 61335009, 61307055, 61475022), 863 program (2013AA014202), 973 program (2014CB340100), the Fundamental Research Fund of State Key Laboratory of IPOC (BUPT), China.

References and links

1. Z. Pan, C. Yu, and A. E. Willner, “Optical performance monitoring for the next generation optical communication networks,” Opt. Fiber Technol. 16(1), 20–45 (2010). [CrossRef]

2. D. C. Kilper, R. Bach, D. J. Blumenthal, D. Einstein, T. Landolsi, L. Ostar, M. Preiss, and A. E. Willner, “Optical Performance Monitoring,” J. Lightwave Technol. 22(1), 294–304 (2004). [CrossRef]

3. J. H. Lee, H. Y. Choi, S. K. Shin, and Y. C. Chung, “A review of the polarization-nulling technique for monitoring optical-signal-to-noise ratio in dynamic WDM networks,” J. Lightwave Technol. 24(11), 4162–4171 (2006). [CrossRef]

4. I. Shake and H. Takara, “Averaged Q-factor method using amplitude histogram evaluation for transparent monitoring of optical signal-to-noise ratio degradation in optical transmission system,” J. Lightwave Technol. 20(8), 1367–1373 (2002). [CrossRef]

5. Z. N. Tao, Z. Y. Chen, L. B. Fu, D. M. Wu, and A. S. Xu, “Monitoring of OSNR by using a Mach-Zehnder interferometer,” Microw. Opt. Technol. Lett. 30(1), 63–65 (2001). [CrossRef]

6. E. Flood, W. H. Guo, D. Reid, M. Lynch, A. L. Bradley, L. P. Barry, and J. F. Donegan, “In-band OSNR monitoring using a pair of Michelson fiber interferometers,” Opt. Express 18(4), 3618–3625 (2010). [CrossRef] [PubMed]

7. M. R. Chitgarha, S. Khaleghi, W. Daab, A. Almaiman, M. Ziyadi, A. Mohajerin-Ariaei, D. Rogawski, M. Tur, J. D. Touch, V. Vusirikala, W. Zhao, and A. E. Willner, “Demonstration of in-service wavelength division multiplexing optical-signal-to-noise ratio performance monitoring and operating guidelines for coherent data channels with different modulation formats and various baud rates,” Opt. Lett. 39(6), 1605–1608 (2014). [PubMed]

8. M. D. Pelusi, A. Fu, and B. J. Eggleton, “Multi-channel in-band OSNR monitoring using Stimulated Brillouin Scattering,” Opt. Express 18(9), 9435–9446 (2010). [CrossRef] [PubMed]

9. J. Y. Huh and Y. C. Chung, “Simultaneous monitoring technique for OSNR and PMD based on four-wave mixing in SOA,” in Optical Fiber Communication Conference2008(OFC/NFOEC 2008), paper OThW1. [CrossRef]

10. K. Xu, H. K. Tsang, G. K. P. Lei, Y. M. Chen, L. Wang, Z. Cheng, X. Chen, and C. Shu, “OSNR monitoring for NRZ-PSK signals using silicon waveguide two-photon absorption,” IEEE Photonics J. 3(5), 968–974 (2011). [CrossRef]

11. N. An, J. Qiu, Z. Huang, B. Yuan, D. Kong, and J. Wu, “Multi-wavelength in-band OSNR monitoring based on Lyot-Sagnac interferometer,” in Optical Fiber Communication Conference2015(OFC 2015), paper Th2A.3. [CrossRef]

12. C. S. Kim, Y. G. Han, R. M. Sova, U. C. Paek, Y. Chung, and J. U. Kang, “Optical fiber modal birefringence measurement based on Lyot-Sagnac interferometer,” IEEE Photonics Technol. Lett. 15(2), 269–271 (2003). [CrossRef]

Multi-wavelength in-band OSNR monitor based on Lyot-Sagnac interferometer

Abstract

1. Introduction

2. Principle of operation

3. Experimental setup

4. Experimental results and discussions

5. Conclusion

Acknowledgments

References and links

Cited By

Figures (8)

Equations (8)

Optics Express