16.3 W Peak-Power Pulsed All-Diode Laser Based Multi-Wavelength Master-Oscillator Power-Amplifier System at 964 nm

Round 1

Reviewer 1 Report

The resubmitted paper has been refocused on the laser source development and its characterization. This is a good choice made by the authors since the lidar and spectroscopy aspects proposed in the previous version were too shallow.
The added discussion about ASE increase respect with the power amplifier current is also of interest, in regard of the foreseen applications.
The work and developments proposed here are certainly worth being published.
However there are still some important method and coherence issues in the way the results are presented and delivered, which would require mandatory revisions.

- The discussion about the linewidth is rather clear, and it makes a convincing justification that, by construction and comparison with similar architectures, the actual linewidth of the device should be much better than the 10 pm (i.e. few GHz at the considered wavelengths) resolution limit of the optical spectrum analyzer. There is no doubt about the capacity to reach typical ~1 MHz linewidth for this type of DFB CW master oscillators. However, at the power amplifier output the linewidth can be expected to be degraded due to pulsed operation, since the Fourier transform limit for 8 ns pulses should be in the ~100 MHz range. Still suitable for lidar application, but clearly not in the “sub-MHz level” as stated line 292. These aspects should be discussed much more cautiously, in absence of an actual experimental linewidth measurement at the output of the whole laser device.

- Another point that is missing in the paper is the description on how the wavelengths are interleaved in time. Indeed one of the authors claims is fast wavelength switching (e.g. line 280), but this aspect is not illustrated in the paper. It is clear that the pulsed operation (at 25 kHz) is provided by the gate and the tapered sections in the power amplifier, but the characterizations shown in the different figures are done with the two lasers either working at the same time (Fig. 4), or working separately with one laser switched on and the other one switched off (Fig. 5). Since Wavelength (fast) switching is a critical functionality for the targeted lidar application (especially for devices relying on waveguides), the paper would gain a lot of value by clearly exposing how the switching works.

Author Response

Reply comment 1:   We are indebted to the reviewer to refer to this fact. It is correct – the Fourier transform limit must be accounted for. We modified the statement and the summary accordingly.

Reply comment 2: 

In this work, the injection currents were switched manually from one laser to the other one. We used the term “fast wavelength switching” in comparison with the laser transmitters for water vapor DIAL instruments as cited in [7-12]. By using a fast driver circuit, the currents could be turned on and off on a ns time scale. But this was not demonstrated in our work. We clarified this in the manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

The alternating dual-wavelength pulse lasers are the important laser sources for the differential absorption LIDAR (DIAL) where the absorption line of water vapor molecules can be effectively measured. The MOPA system presented in this manuscript is based on a Y-branch DFB laser operated in continuous-wave (CW) mode and a pulsed-driven two-section power amplifier (PA). Alternating dual wavelength operation is obtained between 964 nm and 968 nm with a measured spectral linewidth below 10 pm and a peak power of more than 10 W. A repetition rate of 25 kHz is used in the experiment and a pulse width of 8 ns is obtained. The manuscript can be accepted to published after some questions are answered.

1. The writing is not rigorous, with some grammatical mistakes, some spelling mistakes, inaccurate and unprofessional words, incongruent description of pictures and pictures, and some writing mistakes.
2. Laser output performance and electro-optic conversion efficiency at different repetition rates can be added in the manuscript.
3. What about the beam quality, can you further discuss the laser beam variation between the amplifying process?
4. "Beam splitter" is not described in Figure 3, nor is τ_delay on page 6.

Author Response

Reply:  We thank you very much for your useful discussion. Please see below, for a point-by-point response to the your’ comments and concerns.

1. We will read through the manuscript carefully and try our best to correct all the mistakes.
2. The repetition frequency of 25 kHz corresponds to a duty cycle of only 0.02%. An increasing repetition frequency results in an increase of the duty cycle and hence of the average thermal load which could have an impact on both performance of the driver circuit and the amplifier. However, we think repetition frequencies of several 100 kHz are still feasible because the time-averaged thermal load of the PA is of the order of only 100 mW assuming a conversion efficiency of 30%. We added a corresponding remark at the end of the manuscript.
3. We did not measure the beam properties. However, we know from CW operated tapered amplifiers that they emit a near diffraction limited beam. For example, in the newly added Reference [30] a lateral beam quality factor M² = 1.7 was measured for a 3 mm long amplifier at a CW output power of 4.5 W. We added a corresponding remark at the end of the manuscript.
4. There is a mistake in the sketch of the experimental setup. “Beam splitter” was used to check the power or the spectra of the MO during performing the experiment. However, for the work described in this manuscript we did not use it. Therefore, the beam splitter was removed in Fig. 3. t_Delay is the delay time between the pulse through the TS and the OG, which we defined now in the text (before Fig. 5).

Author Response File: Author Response.pdf

Reviewer 3 Report

1- the abstract is slightly long and can be shortened.

2- In section 3: the active region of both the DFB and the PA consist of double quantum well (QW) as shown in the figure. Yet, that is not mentioned or discussed in the text.

3- In line 198: the slope effeciency is 0.55 W/A for which case ? at higher temperature, the effeciency is slightly lower, please clarify.

4- In Fig. 4c, there is slight red-shift in both branches as the current increases, please comments.

5- In line 257: regarding the spectral width: have the authors tried an RF spectrum analyzer? or have any plan to do so to overcome the resolution limitation?

6- In line 279: for the generation of a light source or for generating a light source>

7-I suggest that the authors comment on the improvements of their results.

Author Response

Reply:  We thank you very much for your useful comments and your question. Please see below, for a point-by-point response to the your’ comments and concerns:

1. We revised the abstract.
2. We added “double quantum well (QW)” for both the DFB and the PA in the text.
3. We inserted the slope efficiency at the two temperatures for both branches.
4. Yes, it is correct that there is slight redshift in both branches, it is a common effect caused by self-heating when current increases. We added a corresponding remark.
5. The measurement of the spectral linewidth of the MOPA is subject of further work.
6. We revised the whole paragraph. It was corrupted.
7. We added some directions of further work in the Summary.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Given the response effort brought by the authors, the paper can be considered for publication. 

Carefull proofreading should eliminate the remaining imperfections.

Best regards

Reviewer 2 Report

The manuscript can be accepted to be published.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

See attached comments

Comments for author File: Comments.pdf

Reviewer 2 Report

In this manuscript, the authors investigate a semiconductor laser MOPA system for differential absorption lidar applications. The paper is very interesting, however the high spectral performance claimed in the title of the manuscript is only partially addressed. The laser linewidth is vital in differential absorption lidar systems, so deeply addressing this will improve considerably the manuscript. For this reason, I cannot accept the manuscript for publication in the present form, and I will reconsider it after major revision that I describe in the following.

1. This system or the different parts of this system, have been investigated previously in references [14-18] but, correct me if I am wrong, a better estimation of the linewidth of the lasers has never been reported. The value of less than 10 pm is given by the limited resolution of your OSA. You can use for example the delayed self-heterodyne technique to get a better estimation of the linewidth. if it has not measured before, this will improve considerably the manuscript.
2. Similar to the previous comment, the impact on the spectrum of the pulsed operation is not shown in your results. Panels c) of Figures 6 and 7 show the optical spectrum in a stationary condition, they show how the ASE grows and the SNR diminishes, however the dynamic effect over the spectrum of pulsed operation it is not shown. I do not expect that your repetition frequency of 25 kHz will have an impact on that, however the fast transitions of the nice pulses shown in panels a) may have some impact.

Reviewer 3 Report

The paper deals with the realization and characterization of an integrated semiconductor based MOPA laser set-up, which was designed to meet the requirements for water vapor observation in the atmosphere with the Differential Absorption LIDAR method. The presented work relies on a well-established technological background since similar approaches have proven their utility in the recent years, essentially through the development efforts from different research labs (mainly NCAR). The objectives and requirements for this kind of laser source development are consequently quite clear, and these are well reminded and documented by the authors in the introduction.

 

The claims of the paper are essentially three: 1/ the wavelength range proposed is slightly different around 966 nm (whereas most previous systems for water vapor work in the 800 nm range), 2/ the pulse duration attainable is shorter down to 8 ns, and 3/ the ON and OFF sources are integrated with a Y-branch coupler on the same semi-conductor chip, such that the switching functionality is monolithically available on a single device.

 

The overall paper is well written, and the characterizations shown are clear. So the work proposed here is probably well suited for publication. However it seems that these results do not fully support the third claim about the spectral performances, which is probably the most interesting one. Some clarifications would thus be very welcome, and according to me mandatory, before publication.

 

Recommendations:

1. Small remark, the second sentence in the introduction should be rephrased like for example :”The MOPA emits two alternating wavelengths, whose spectral separation can be adjusted”
2. Line 105 – 107 the sign ÷ is not clear, do the other mean 5 to 10 W, 900 to 1000 ns, etc… ?
3. Line 132, the terminology “spatial resolution” can be confusing, I would recommend to use “range resolution” or “range cell resolution” terminology. Additionally, the range resolution is limited by the pulse duration as long as it is not limited by the detection and averaging method of the LIDAR system. In this regard 1.2 m seems to be very short. Is it actually an objective or a requirement for the application to have such a high resolution?
4. The title is little bit shallow. “High Spectral Performance” is not specific enough. The main spectral characteristics that are generally looked at for DIAL applications are the wavelength stability over time, the spectral purity of the radiation, and the spectral “agility” that generally cover the wavelength switching rate capability and/or the number of wavelengths available. In the paper the main aspect that is looked at is the spectral purity through optical spectrum analysis. I would consequently be more specific in the title with “high SMSR” or “high spectral purity” instead for example.
5. I would also include in the title one important aspect that is the monolithically integrated switching functionality, as it seems to be one key point
6. Spectral purity is not limited to SMSR characterization. What is significant is the integrated power within a certain spectral width (10 pm for instance with the instrumental resolution available) respect with the total power (including the whole ASE pedestal). It would be very interesting to have an estimation of this value. In other words is there 98%, 99% or even more of the light within 10 pm?
7. The fast wavelength switching functionality is not shown or explained in the paper. Maybe this has been demonstrated in previous work. In that case this should be properly stated and referenced. Otherwise we have spectral characterizations when the two lasers are working together at the same time (figure 5), or separately in steady state with one laser switched on while the other one is switched off (figure 6 and 7). The method and sequence used to interleave the wavelengths is missing, which brings complementary questions:

• How the switching is implemented? One can assume that it is managed with the current in the Y branches of the coupler, but it should be explained and illustrated
• What is the switching rate capability? Is it for example possible to go as fast as the 25 kHz rep. rate for a pulse to pulse wavelength switch?
• Does the switching sequence affect the spectral quality of the radiation? For example what is the suppression ratio between the ON and OFF wavelengths during the switching sequence