- Email: [email protected]
Dynamic power and chirp measurements of electroabsorption and Mach-Zehnder modulator pulse generators and chirp factor extraction using a linear pulse characterization technique
Optics and Laser Technology 121 (2020) 105783
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Full length article
Dynamic power and chirp measurements of electroabsorption and MachZehnder modulator pulse generators and chirp factor extraction using a linear pulse characterization technique
T
⁎
Michael J. Connelly , Javier Romero-Vivas, Aidan Meehan, Lukasz Krzczanowicz Optical Communications Research Group, Dept. Electronic and Computer Engineering, University of Limerick, Limerick, Ireland
H I GH L IG H T S
of a linear measurement technique for modulator characterization. • Application of pulse dynamic power and chirp. • Measurement • Use of simple models to determine modulator chirp factor.
A R T I C LE I N FO
A B S T R A C T
Keywords: Optical modulator Electroabsorption modulator Mach-Zehnder modulator Pulse characterization
Electroabsorption and Mach-Zehnder modulators are routinely used to generate optical pulses for use in optical transmission systems. Both types of modulator impose dynamic frequency chirp on the generated pulses, which can lead to significant pulse broadening when transmitted in optical fiber. This paper describes the use of a modified linear pulse characterization method to measure dynamic pulse power, phase and chirp. Modulator chirp factors are determined by the use of simple models in conjunction with experimental pulse temporal profiles for 40 and 17.7 ps pulsewidth pulses at 10 and 20 GHz repetition rates, generated by the Electroabsorption and Mach-Zehnder modulators respectively.
1. Introduction Niobate based Electroabsorption optical modulators (EAMs) and Mach-Zehnder Modulators (MZMs) and are routinely used to generate optical pulses for use in optical transmission systems [1,2]. Both types of modulator impose chirp on the generated pulses [3]. Chirp has the effect of broadening the pulse spectrum and consequently when such pulses are used for data transmission in optical fiber they will experience more dispersion than would be the case if there were no chirp. Coherent optical communication systems use external modulators for pulse generation but also for amplitude and phase modulation of laser sources. Therefor it is desirable to have a relatively simple technique to measure the dynamic power and phase of such pulses. Among available pulse characterization techniques, Frequency Resolved Optical Gating (FROG) has become the de-facto technique [4]. Because of its use of a nonlinear interaction in a crystal, FROG requires high optical powers at the input and has more polarization sensitivity than linear techniques [5]. A limit is effectively imposed on the maximum pulsewidth that can
⁎
Corresponding author. E-mail address: [email protected] (M.J. Connelly).
https://doi.org/10.1016/j.optlastec.2019.105783 Received 4 July 2019; Accepted 25 August 2019 0030-3992/ © 2019 Elsevier Ltd. All rights reserved.
be measured. In this paper, the post-processing of the experimental data to obtain the pulse chirp is carried out using a deterministic algorithm unlike FROG, which uses blind deconvolution techniques. Other techniques impose restrictions on the temporal support of the pulse. For example, chronocyclic tomography, another linear technique, can work only when a certain condition on the quadratic phase modulation is fulfilled [6]. In contrast, the modified linear pulse characterization technique described in this paper enables the measurement of relatively low power pulse trains. We focus on the measurement of the dynamic power and phase of 40 ps and 17.7 ps pulsewidth pulses generated by sinusoidal modulation of an EAM and MZM at 10 and 20 GHz repetition rates respectively [7], which in prior work we used to characterize an integrated reflective semiconductor optical amplifier-EAM pulse generator [8]. It is then simple to calculate the pulse dynamic chirp and power spectrum. The experimental results are used as an input to simple modulator models to determine their chirp factors, which show good agreement with manufacturer specifications. As expected the EAM chirp factor is much larger that of the MZM.
Optics and Laser Technology 121 (2020) 105783
M.J. Connelly, et al.
Fig. 1. Experimental setup and modulators.
Fig. 2. EAM insertion loss versus reverse voltage.
Fig. 3. Typical experimental MZM generated 20 GHz repetition rate pulse train input SLs and output ISLs. Line broadening is caused by the finite resolution bandwidth of the OSA.
2. Pulse generation and characterisation system
for the EAM and MZM respectively at corresponding repetition rates of 10 and 20 GHz. The tunable laser wavelength in both case is 1550 nm. The EAM output power depends on its drive voltage and input light polarization dependent insertion loss, as shown in experimental data of Fig. 2. The MZM normalized optical power transfer function depends on the modulator drive voltage V as [2],
The experimental setup for optical pulse generation and characterization is shown in Fig. 1. To generate an optical pulse stream, an EAM (CIP Photonics 10G-PS-EAM-1550 - optical sampling window generator) or MZM (low chirp OptiLab IM-1550-20) is driven by a sinusoidal clock voltage, superimposed on a bias voltage, to generate a pulse stream at a repetition rate, equal to the clock frequency F. The Full Width at Half Maximum (FWHM) pulsewidths are 40 and 17.7 ps
πV⎞ P (V ) = cos2 ⎛ ⎝ 2Vπ ⎠ ⎜
2
⎟
(1)
Optics and Laser Technology 121 (2020) 105783
M.J. Connelly, et al.
Fig. 4. EAM (a) experimental and (b) simulated pulse power and chirp.
Fig. 5. MZM (a) experimental and (b) simulated pulse power and chirp.
Fig. 6. (a) EAM and (b) MZM experimental and simulated pulse power spectrums.
where Vπ = 5 V is the voltage required to induce a π phase shift in the phase modulator arm of the MZM. Fig. 2 and (1) can be used to determine the instantaneous output power of either modulator for a given applied time-varying voltage centered on a given bias voltage. A commonly used relationship between the dynamic chirp Δν (t ) and output power p (t ) of both types of modulator is,
Δν (t ) =
α dp (t ) 4πp (t ) dt
(2)
A common figure of merit for modulator chirp inducing properties is the chirp factor α defined as the derivative of the real with respect to the imaginary part of the modulator waveguide modal index [3]. If p (t ) and Δν (t ) are known, then α can be determined using (2). A pulse train spectrum consists of discrete Spectral Lines (SLs), 3
Optics and Laser Technology 121 (2020) 105783
M.J. Connelly, et al.
located at integer multiples of F, centered around the laser frequency ν0. In the pulse characterization system, the SL powers are measured by directly connecting the pulse train to a 0.06 nm resolution bandwidth Optical Spectrum Analyzer (OSA), as shown in Fig. 3. When MZMc is driven by a low-amplitude sinusoidal signal of frequencyF/2 and biased at the minimum transmission point, it results in carrier-suppressed double-sideband modulation of the input pulse train. The delay line (0–18 GHz bandwidth) is used to provide a controllable time delay τ (0– 360°) of the MZMc modulation with respect to the pulse train clock [9,10]. The up-converted sideband of each SL of the input pulse train interferes with the down-converted sideband of the next SL. This produces an interference signal, whose spectrum is a series of Interference Spectral Lines (ISLs) interlaced between the SLs of the original pulse spectrum as illustrated in Fig. 3 for four values of τ. The dynamic pulse power and phase is reconstructed by post processing of set of acquired MZMc’s output spectrums for twelve values of τ and the powers of the input pulse train SLs as described in [9], with some modifications to improve measurement accuracy [10]. The chirp (Hz) is calculated as the time derivative of the dynamic phase ϕ(t) divided by 2π.
confirmed by our results. The experimental and simulated pulse power spectrums shown in Fig. 6(b), have FWHM bandwidths of 15.6 and 17.1 GHz respectively, with corresponding TBPs of 0.56 and 0.52.
3. Experimental results and simulations
[1] Y. Kim, H. Lee, J. Lee, J. Han, T.W. Oh, J. Jeong, Chirp characteristics of 10-Gb/s electroabsorption modulator integrated DFB lasers, IEEE J. Quantum Electron. 36 (8) (2000) 900–908, https://doi.org/10.1109/3.853536. [2] E.L. Wooten, K.M. Kissa, A. Yi-Yan, E.J. Murphy, D.A. Lafaw, P.F. Hallemeier, D. Maack, D.V. Attanasio, D.J. Fritz, G.J. McBrien, D.E. Bossi, A review of Lithium Niobate modulators for fiber-optic communications systems, IEEE J. Select. Top. Quantum. Electron. 6 (1) (2000) 69–82, https://doi.org/10.1109/2944.826874. [3] F. Koyama, K. Iga, Frequency chirping in external modulators, J. Lightwave Technol. 6 (1) (1988) 87–93, https://doi.org/10.1109/50.3969. [4] R. Trebino, K.W. DeLong, D.N. Fittinghoff, J.N. Sweetser, M.A. Krumbugel, B.A. Richman, D.J. Kane, Measuring ultrashort laser pulses in the time-frequency domain using frequency-resolved optical gating, Rev. Sci. Instrum. 68 (1997) 3277–3295, https://doi.org/10.1063/1.1148286. [5] B.C. Thomsen, M.A.F. Roelens, R.T. Watts, D.J. Richardson, Comparison between nonlinear and linear spectrographic techniques for the complete characterization of high bit-rate pulses used in optical communications, IEEE Phot. Technol. Lett. 17 (9) (2005) 1914–1916, https://doi.org/10.1109/LPT.2005.853261. [6] C. Dorrer, K. Inuk, Complete temporal characterization of short optical pulses by simplified chronocyclic tomography, Opt. Lett. 28 (16) (2003) 1481–1483, https:// doi.org/10.1364/OL.28.001481. [7] M.J. Connelly, J. Romero-Vivas, A. Meehan, L. Krzczanowicz, Modeling of MachZehnder and electroabsorption modulator pulse generators and extraction of the chirp factor, International Conference on Numerical Simulation of Optoelectronic Devices, (2015). [8] V. Moreno, Integrated 16-ps pulse generator based on a reflective SOA-EAM for UWB schemes, IEEE Photonic Tech. Lett. 28 (20) (2016) 2180–2182, https://doi. org/10.1109/LPT.2016.2586893. [9] J. Debeau, B. Kowalski, R. Boittin, Simple method for the complete characterizationof an optical pulse, Opt. Lett. 23 (22) (1998) 1784–1786, https://doi.org/10. 1364/OL.23.001784. [10] M.J. Connelly, J. Romero-Vivas, A. Meehan, K. Lukasz, Dynamic power and chirp measurements of amplified 19 ps pulses in traveling-wave and reflective semiconductor optical amplifiers using a linear pulse characterization technique, Opt. Quan. Electron. 51 (7) (2019) 248–1–248–10, https://doi.org/10.1007/s11082019-1948-z.
4. Conclusion A modified linear pulse characterization technique was used to measure dynamic power and chirp of EAM and MZM generated pulses. Because of its linearity, the technique is particularly suited to the characterization of relatively wide low power pulses. Simple models were used in conjunction with the experimental results to obtain modulator chirp factors. Acknowledgement This research was supported by Science Foundation Ireland Investigator Grant 09/IN.1/I2641. References
Fig. 4 shows the EAM experimental and modeled TE polarized pulse power and chirp for a 0.6 V amplitude sinusoidal drive voltage with a bias of 1.0 V. These voltage values were obtained by fitting the modulated EAM normalized response to the measured normalized power. The experimental and simulated pulsewidths are 40.0 and 39.4 ps respectively. The peaks in the chirp profile are caused by a phase reversal effect that occurs when the modulation voltage approaches and then departs from the transmission response minima. The chirp factor was determined by carrying out a fit between the measured and simulated chirp in the region where the normalized pulse power is greater than 0.5. The extracted chirp factor is 2.2. The experimental and simulated pulse power spectrums shown in Fig. 6(a), calculated by applying a fast Fourier transform to the complex pulse field envelope p (t ) exp[jϕ (t )], have FWHM bandwidths of 15.6 and 17.1 GHz respectively. The resulting measured and simulated pulse Time Bandwidth Products (TBPs) are 0.63 and 0.68 respectively. Fig. 5 shows the MZM experimental and modeled pulse power and chirp for a 2.4 V amplitude sinusoidal drive voltage with a bias of 6.6 V. The amplitude and bias parameters were obtained by fitting (1) with a multiplicative attenuation factor to take into account other insertion losses, with the measured normalized power. These values were close to the measured values of 2.5 V and 7.2 V obtained at the modulator input. The experimental and theoretical pulsewidths are 17.7 and 18.3 ps respectively. As is the case for the EAM, the peaks in the chirp plot are caused by a similar phase reversal effect. The extracted chirp factor is 0.065. The manufacture specified chirp factor is less than 0.1, which is
4
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Full length article
Dynamic power and chirp measurements of electroabsorption and MachZehnder modulator pulse generators and chirp factor extraction using a linear pulse characterization technique
T
⁎
Michael J. Connelly , Javier Romero-Vivas, Aidan Meehan, Lukasz Krzczanowicz Optical Communications Research Group, Dept. Electronic and Computer Engineering, University of Limerick, Limerick, Ireland
H I GH L IG H T S
of a linear measurement technique for modulator characterization. • Application of pulse dynamic power and chirp. • Measurement • Use of simple models to determine modulator chirp factor.
A R T I C LE I N FO
A B S T R A C T
Keywords: Optical modulator Electroabsorption modulator Mach-Zehnder modulator Pulse characterization
Electroabsorption and Mach-Zehnder modulators are routinely used to generate optical pulses for use in optical transmission systems. Both types of modulator impose dynamic frequency chirp on the generated pulses, which can lead to significant pulse broadening when transmitted in optical fiber. This paper describes the use of a modified linear pulse characterization method to measure dynamic pulse power, phase and chirp. Modulator chirp factors are determined by the use of simple models in conjunction with experimental pulse temporal profiles for 40 and 17.7 ps pulsewidth pulses at 10 and 20 GHz repetition rates, generated by the Electroabsorption and Mach-Zehnder modulators respectively.
1. Introduction Niobate based Electroabsorption optical modulators (EAMs) and Mach-Zehnder Modulators (MZMs) and are routinely used to generate optical pulses for use in optical transmission systems [1,2]. Both types of modulator impose chirp on the generated pulses [3]. Chirp has the effect of broadening the pulse spectrum and consequently when such pulses are used for data transmission in optical fiber they will experience more dispersion than would be the case if there were no chirp. Coherent optical communication systems use external modulators for pulse generation but also for amplitude and phase modulation of laser sources. Therefor it is desirable to have a relatively simple technique to measure the dynamic power and phase of such pulses. Among available pulse characterization techniques, Frequency Resolved Optical Gating (FROG) has become the de-facto technique [4]. Because of its use of a nonlinear interaction in a crystal, FROG requires high optical powers at the input and has more polarization sensitivity than linear techniques [5]. A limit is effectively imposed on the maximum pulsewidth that can
⁎
Corresponding author. E-mail address: [email protected] (M.J. Connelly).
https://doi.org/10.1016/j.optlastec.2019.105783 Received 4 July 2019; Accepted 25 August 2019 0030-3992/ © 2019 Elsevier Ltd. All rights reserved.
be measured. In this paper, the post-processing of the experimental data to obtain the pulse chirp is carried out using a deterministic algorithm unlike FROG, which uses blind deconvolution techniques. Other techniques impose restrictions on the temporal support of the pulse. For example, chronocyclic tomography, another linear technique, can work only when a certain condition on the quadratic phase modulation is fulfilled [6]. In contrast, the modified linear pulse characterization technique described in this paper enables the measurement of relatively low power pulse trains. We focus on the measurement of the dynamic power and phase of 40 ps and 17.7 ps pulsewidth pulses generated by sinusoidal modulation of an EAM and MZM at 10 and 20 GHz repetition rates respectively [7], which in prior work we used to characterize an integrated reflective semiconductor optical amplifier-EAM pulse generator [8]. It is then simple to calculate the pulse dynamic chirp and power spectrum. The experimental results are used as an input to simple modulator models to determine their chirp factors, which show good agreement with manufacturer specifications. As expected the EAM chirp factor is much larger that of the MZM.
Optics and Laser Technology 121 (2020) 105783
M.J. Connelly, et al.
Fig. 1. Experimental setup and modulators.
Fig. 2. EAM insertion loss versus reverse voltage.
Fig. 3. Typical experimental MZM generated 20 GHz repetition rate pulse train input SLs and output ISLs. Line broadening is caused by the finite resolution bandwidth of the OSA.
2. Pulse generation and characterisation system
for the EAM and MZM respectively at corresponding repetition rates of 10 and 20 GHz. The tunable laser wavelength in both case is 1550 nm. The EAM output power depends on its drive voltage and input light polarization dependent insertion loss, as shown in experimental data of Fig. 2. The MZM normalized optical power transfer function depends on the modulator drive voltage V as [2],
The experimental setup for optical pulse generation and characterization is shown in Fig. 1. To generate an optical pulse stream, an EAM (CIP Photonics 10G-PS-EAM-1550 - optical sampling window generator) or MZM (low chirp OptiLab IM-1550-20) is driven by a sinusoidal clock voltage, superimposed on a bias voltage, to generate a pulse stream at a repetition rate, equal to the clock frequency F. The Full Width at Half Maximum (FWHM) pulsewidths are 40 and 17.7 ps
πV⎞ P (V ) = cos2 ⎛ ⎝ 2Vπ ⎠ ⎜
2
⎟
(1)
Optics and Laser Technology 121 (2020) 105783
M.J. Connelly, et al.
Fig. 4. EAM (a) experimental and (b) simulated pulse power and chirp.
Fig. 5. MZM (a) experimental and (b) simulated pulse power and chirp.
Fig. 6. (a) EAM and (b) MZM experimental and simulated pulse power spectrums.
where Vπ = 5 V is the voltage required to induce a π phase shift in the phase modulator arm of the MZM. Fig. 2 and (1) can be used to determine the instantaneous output power of either modulator for a given applied time-varying voltage centered on a given bias voltage. A commonly used relationship between the dynamic chirp Δν (t ) and output power p (t ) of both types of modulator is,
Δν (t ) =
α dp (t ) 4πp (t ) dt
(2)
A common figure of merit for modulator chirp inducing properties is the chirp factor α defined as the derivative of the real with respect to the imaginary part of the modulator waveguide modal index [3]. If p (t ) and Δν (t ) are known, then α can be determined using (2). A pulse train spectrum consists of discrete Spectral Lines (SLs), 3
Optics and Laser Technology 121 (2020) 105783
M.J. Connelly, et al.
located at integer multiples of F, centered around the laser frequency ν0. In the pulse characterization system, the SL powers are measured by directly connecting the pulse train to a 0.06 nm resolution bandwidth Optical Spectrum Analyzer (OSA), as shown in Fig. 3. When MZMc is driven by a low-amplitude sinusoidal signal of frequencyF/2 and biased at the minimum transmission point, it results in carrier-suppressed double-sideband modulation of the input pulse train. The delay line (0–18 GHz bandwidth) is used to provide a controllable time delay τ (0– 360°) of the MZMc modulation with respect to the pulse train clock [9,10]. The up-converted sideband of each SL of the input pulse train interferes with the down-converted sideband of the next SL. This produces an interference signal, whose spectrum is a series of Interference Spectral Lines (ISLs) interlaced between the SLs of the original pulse spectrum as illustrated in Fig. 3 for four values of τ. The dynamic pulse power and phase is reconstructed by post processing of set of acquired MZMc’s output spectrums for twelve values of τ and the powers of the input pulse train SLs as described in [9], with some modifications to improve measurement accuracy [10]. The chirp (Hz) is calculated as the time derivative of the dynamic phase ϕ(t) divided by 2π.
confirmed by our results. The experimental and simulated pulse power spectrums shown in Fig. 6(b), have FWHM bandwidths of 15.6 and 17.1 GHz respectively, with corresponding TBPs of 0.56 and 0.52.
3. Experimental results and simulations
[1] Y. Kim, H. Lee, J. Lee, J. Han, T.W. Oh, J. Jeong, Chirp characteristics of 10-Gb/s electroabsorption modulator integrated DFB lasers, IEEE J. Quantum Electron. 36 (8) (2000) 900–908, https://doi.org/10.1109/3.853536. [2] E.L. Wooten, K.M. Kissa, A. Yi-Yan, E.J. Murphy, D.A. Lafaw, P.F. Hallemeier, D. Maack, D.V. Attanasio, D.J. Fritz, G.J. McBrien, D.E. Bossi, A review of Lithium Niobate modulators for fiber-optic communications systems, IEEE J. Select. Top. Quantum. Electron. 6 (1) (2000) 69–82, https://doi.org/10.1109/2944.826874. [3] F. Koyama, K. Iga, Frequency chirping in external modulators, J. Lightwave Technol. 6 (1) (1988) 87–93, https://doi.org/10.1109/50.3969. [4] R. Trebino, K.W. DeLong, D.N. Fittinghoff, J.N. Sweetser, M.A. Krumbugel, B.A. Richman, D.J. Kane, Measuring ultrashort laser pulses in the time-frequency domain using frequency-resolved optical gating, Rev. Sci. Instrum. 68 (1997) 3277–3295, https://doi.org/10.1063/1.1148286. [5] B.C. Thomsen, M.A.F. Roelens, R.T. Watts, D.J. Richardson, Comparison between nonlinear and linear spectrographic techniques for the complete characterization of high bit-rate pulses used in optical communications, IEEE Phot. Technol. Lett. 17 (9) (2005) 1914–1916, https://doi.org/10.1109/LPT.2005.853261. [6] C. Dorrer, K. Inuk, Complete temporal characterization of short optical pulses by simplified chronocyclic tomography, Opt. Lett. 28 (16) (2003) 1481–1483, https:// doi.org/10.1364/OL.28.001481. [7] M.J. Connelly, J. Romero-Vivas, A. Meehan, L. Krzczanowicz, Modeling of MachZehnder and electroabsorption modulator pulse generators and extraction of the chirp factor, International Conference on Numerical Simulation of Optoelectronic Devices, (2015). [8] V. Moreno, Integrated 16-ps pulse generator based on a reflective SOA-EAM for UWB schemes, IEEE Photonic Tech. Lett. 28 (20) (2016) 2180–2182, https://doi. org/10.1109/LPT.2016.2586893. [9] J. Debeau, B. Kowalski, R. Boittin, Simple method for the complete characterizationof an optical pulse, Opt. Lett. 23 (22) (1998) 1784–1786, https://doi.org/10. 1364/OL.23.001784. [10] M.J. Connelly, J. Romero-Vivas, A. Meehan, K. Lukasz, Dynamic power and chirp measurements of amplified 19 ps pulses in traveling-wave and reflective semiconductor optical amplifiers using a linear pulse characterization technique, Opt. Quan. Electron. 51 (7) (2019) 248–1–248–10, https://doi.org/10.1007/s11082019-1948-z.
4. Conclusion A modified linear pulse characterization technique was used to measure dynamic power and chirp of EAM and MZM generated pulses. Because of its linearity, the technique is particularly suited to the characterization of relatively wide low power pulses. Simple models were used in conjunction with the experimental results to obtain modulator chirp factors. Acknowledgement This research was supported by Science Foundation Ireland Investigator Grant 09/IN.1/I2641. References
Fig. 4 shows the EAM experimental and modeled TE polarized pulse power and chirp for a 0.6 V amplitude sinusoidal drive voltage with a bias of 1.0 V. These voltage values were obtained by fitting the modulated EAM normalized response to the measured normalized power. The experimental and simulated pulsewidths are 40.0 and 39.4 ps respectively. The peaks in the chirp profile are caused by a phase reversal effect that occurs when the modulation voltage approaches and then departs from the transmission response minima. The chirp factor was determined by carrying out a fit between the measured and simulated chirp in the region where the normalized pulse power is greater than 0.5. The extracted chirp factor is 2.2. The experimental and simulated pulse power spectrums shown in Fig. 6(a), calculated by applying a fast Fourier transform to the complex pulse field envelope p (t ) exp[jϕ (t )], have FWHM bandwidths of 15.6 and 17.1 GHz respectively. The resulting measured and simulated pulse Time Bandwidth Products (TBPs) are 0.63 and 0.68 respectively. Fig. 5 shows the MZM experimental and modeled pulse power and chirp for a 2.4 V amplitude sinusoidal drive voltage with a bias of 6.6 V. The amplitude and bias parameters were obtained by fitting (1) with a multiplicative attenuation factor to take into account other insertion losses, with the measured normalized power. These values were close to the measured values of 2.5 V and 7.2 V obtained at the modulator input. The experimental and theoretical pulsewidths are 17.7 and 18.3 ps respectively. As is the case for the EAM, the peaks in the chirp plot are caused by a similar phase reversal effect. The extracted chirp factor is 0.065. The manufacture specified chirp factor is less than 0.1, which is
4