Random Laser Based on Ytterbium-Doped Fiber with a Bragg Grating Array as the Source of Continuous-Wave 976 nm Wavelength Radiation

Round 1

Reviewer 1 Report

In this paper, authors utilized an in-line technique to fabricate Bragg grating arrays in a piece of ytterbium-doped germanophosphosilicate fiber during the fiber drawing process. Then, this fiber was used to construct a random narrow-linewidth laser with a central wavelength of 976 nm. The maximum power of the laser reached 25 mW with a slope efficiency of 33%. In general, the paper is well-organized and the results are new. I recommend publication provided the following concerns are well addressed.

1. What is the limitation of the current random laser? Can the output power and the slope efficiency be further increased? I notice that some high-power and high-efficiency random laser or 980-nm laser have been achieved recently. For random laser, a 1.6 kW power with an efficiency of 77% was reported in Optics Letters(DOI:10.1364/OL.438352), and a 5.1 kW power with an efficiency of 89% was reported in IEEE Journal of Quantum Electronics (DOI:10.1109/JQE.2020.3048166). As for 980-nm laser, a 151 W power with an efficiency of 63% was reported in Optics Express(Ref. [5] in the manuscript), and a 1.1 kW power with an efficiency of 65% was reported in Optics Letters (DOI:10.1364/OL.441789). Although these papers paid less attention to the quality of the output spectrum, they might inspire authors how to achieve higher output power with higher slope efficicency.

2. Using an in-line technique to write Bragg gratings during fiber drawing process is interesting. What are the main advantages and disadvantages of this technique? I believe it will be beneficial for readers if authors can add some discriptions and classical references to this technique. I have found two related references discussing the in-line fabrication of Bragg grating arrays on a draw tower, one was published in Optics Letters in 1994 (DOI:10.1364/OL.19.000147), while another was published in IEEE Journal Of Lightwave Technology in 2015 (10.1109/JLT.2014.2384373). I think a discussing and citing will be worthwhile.

3. Authors also uses a single FBG with a reflectivity of ~90% for experiment. Current manuscript has pointed out that this FBG was spliced to the fiber. I am curious that whether this single FBG can also be writted in fiber with an in-line technique or not. Authors can add some discussions why this “deeper” FBG can or cannot be fabricated with in-line technique.

4. Some typos need to be corrected.

(a) ‘a operation’> ‘an operation’.

(b) ‘The development a random laser’>‘The development of a random laser’.

(c) ‘an crucial condition’>‘a crucial condition’.

Author Response

Thank you for your time and comments provided. In respect for your questions we would like to give our answers to following questions:

1. What is the limitation of the current random laser? Can the output power and the slope efficiency be further increased? I notice that some high-power and high-efficiency random laser or 980-nm laser have been achieved recently. For random laser, a 1.6 kW power with an efficiency of 77% was reported in Optics Letters (DOI:10.1364/OL.438352), and a 5.1 kW power with an efficiency of 89% was reported inIEEE Journal of Quantum Electronics (DOI:10.1109/JQE.2020.3048166). As for 980-nm laser, a 151 W power with an efficiency of 63% was reported in Optics Express(Ref. [5] in the manuscript), and a 1.1 kW power with an efficiency of 65% was reported in Optics Letters (DOI:10.1364/OL.441789). Although these papers paid less attention to the quality of the output spectrum, they might inspire authors how to achieve higher output power with higher slope efficicency.

Actually, one of the most promising applications of  976nm wavelength lasers, is frequency-doubling systems (with subsequent possible conversion to UV 244nm wavelength radiation), which require high power and good initial quality of a laser beam. According to the papers you have cited, the most possible ways to improve power characteristics of 976nm wavelength random laser are implementation of the “classic” MOPA scheme (narrow-band random laser as master-oscillator + high-power ytterbium amplifier as booster) or the novel design of random-FBG laser, based on the large-mode area (LMA) fiber. However, the in-lineFBG inscription in a LMA fiber seems to us more complicated than using the MOPA scheme. Using the MOPA does not require high initial power from the ”seed” laser: ~10 mW is enough for further amplification. As a possible option, for the Yb single-mode all-fiber amplifier has been studied in this paper: L. Kotov et al, Optics Letters, 45 (15), pp. 4292-429 (DOI:10.1364/OL.398251).So, such experiments are the main topic of our future investigations.

2.  Using an in-line technique to write Bragg gratings during fiber drawing process is interesting. What are the main advantages and disadvantages of this technique? I believe it will be beneficial for readers if authors can add some discriptions and classical references to this technique. I have found two related references discussing the in-line fabrication of Bragg grating arrays on a draw tower, one was published in Optics Lettersin 1994 (DOI:10.1364/OL.19.000147), while another was published in IEEE Journal Of Lightwave Technology in 2015 (10.1109/JLT.2014.2384373). I think a discussing and citing will be worthwhile.

The key advantage of the in-line FBG fabrication technique is the ability to quickly and inexpensively create arrays of many identical FBGs without compromising the mechanical strength of a fiber. Additionally, total reflection spectrum of the fabricated FBG array can be quite narrow (with minimal distortions), which will result in the stability of the random laser emission wavelength.It should be noted that in order to inscribe the FBG array with an effective total reflection (tens of percent at lengths of several meters), a relatively high initial photosensitivity of the fiber is required. Another challenge is to control precisely the drawing speed and drawing tension to minimizespectrum distortions of inscribed FBGs.Clarifications,revealing the advantages of the in-line FBG inscription process, have been added to the text (page 3) and references ([20,[21]).

3. Authors also uses a single FBG with a reflectivity of ~90% for experiment. Current manuscript has pointed out that this FBG was spliced to the fiber. I am curious that whether this single FBG can also be writted in fiber with an in-line technique or not. Authors can add some discussions why this “deeper” FBG can or cannot be fabricated with in-line technique.

The FBG with a reflectivity of ~90%, used in this article, was inscribed in a commercially-available Corning HI1060 fiber, using a standard phase-mask technique.Clarifications about this FBG have been added to the text (page 9).Unfortunately, with a FBG length of ~10 mm, such a high reflectivity (90 %) is extremely difficult to achieve in the fiber core when using the in-line FBG inscription technique.It is also important to note that this FBG should be photoinduced (non-damaged) to avoid additional optical losses. At the very least, the experiment will require a fiber that is much more photosensitive (not less than 10 times greater).

4. Some typos need to be corrected.

(a) ‘a operation’> ‘an operation’.

(b) ‘The development a random laser’>‘The development of a random laser’.

(c) ‘an crucial condition’>‘a crucial condition’.

Thank you for your correction, misprints have been found and fixed. The paper has been modified accordingly.

Reviewer 2 Report

The authors demonstrated narrow-linewidth random fiber laser at ~976 nm based on FBG arrays. The dependance of power fluctuations/lasing efficiency on the fiber length and additional HR FBG was investigated and discussed.

I think this paper could be published if the authors could properly address the following issues:

1.      What about the lasing threshold when the length of YDF is reduced? The temporal stability is usually not good when operating around the lasing threshold. If the laser threshold decreases by decreasing the YDF, could this contribute to a better temporal stability when the same pump power is applied? What about the temporal stability by simply increasing the pump power?

2.      From Fig.3, the pump power is not sufficiently absorbed and lasing efficiency is relatively low. I wonder why decreasing the length of the YDF to 5 m did not lead to a power reduction? This is not usual in conventional fiber lasers. What is the reason?

3.      So far as I know, this random laser works in a full-open-cavity scheme before adding the 99% FBG, which makes it a half-open cavity. Please specify the condition at both ends of the laser. Was the output fiber angle cleaved or kept flat?

4.      How the power fluctuation is estimated? Did the author use standard deviation to characterize the power fluctuation? The authors should provide exact values of the fluctuations.

5.      Could the author give a plot of the Fourier transform of the time traces to see if there are characteristic frequency components corresponding to the YDF length (‘cavity’ length) to better prove it is random laser rather than conventional fiber oscillator.

6.      Typo/grammar issues:

a)       line 56: ‘≈976 nm’ should change into ‘~ 976 nm’

b)       line 48, the authors used ‘…, which… , which…’. This makes the sentence complicated and is not commonly used.

Author Response

Thank you for your time and comments provided. In respect for your questions we would like to give our answers to following questions:

1. What about the lasing threshold when the length of YDF is reduced? The temporal stability is usually not good when operating around the lasing threshold. If the laser threshold decreases by decreasing the YDF, could this contribute to a better temporal stability when the same pump power is applied? What about the temporal stability by simply increasing the pump power?

The lasing threshold did not shift significantly with a reducing of the YDF length - only the level of output power and slope efficiency changed. With increasing pump power, the amplitude of laser power fluctuations also increased (however, the standard deviation slightly decreased...). Since high temporal stability of  radiation is required for most practical applications, it makes no sense (in our opinion) to use this laser at relatively high power level - a few milliwatts in a stable operation mode will be quite enough.

2. From Fig.3, the pump power is not sufficiently absorbed and lasing efficiency is relatively low. I wonder why decreasing the length of the YDF to 5 m did not lead to a power reduction? This is not usual in conventional fiber lasers. What is the reason?

We observed the similar effect earlier in the ~1550nm wavelength random-FBG laser cavity based on the Er-doped fiber (see S.M. Popov et al, Proc. of SPIE’2020, p. 113571Q. DOI: 10.1117/12.2557818). Cavity lengths (less than 10 m) and intensity of absorption near pumping wavelength are comparable in both cases.  We assume this “non-conventional cavity length dependence” of random-FBG lasers is related to the contribution of  the dynamical population inversion grating inscribed immediately in the cavity during lasing. The contrast of  the inversion grating is not the same along the laser cavity and reaches a maximum at the area where the pump radiation is coupled.

3. So far as I know, this random laser works in a full-open-cavity scheme before adding the 99% FBG, which makes it a half-open cavity. Please specify the condition at both ends of the laser. Was the output fiber angle cleaved or kept flat?

Indeed, we carried out comparative experiments with connecting FC/UPC or FC/APC fiber single-mode pigtails to the both (input and output) ends of random cavity. We expected the APC pigtail connection to improve laser temporal stability. Unfortunately, there was no effect on pigtail type.

4. How the power fluctuation is estimated? Did the author use standard deviation to characterize the power fluctuation? The authors should provide exact values of the fluctuations.

Thank you for your notes, the information about calculated standard deviation values is added to the text (page 8).

5. Could the author give a plot of the Fourier transform of the time traces to see if there are characteristic frequency components corresponding to the YDF length (‘cavity’ length) to better prove it is random laser rather than conventional fiber oscillator.

The FFT plot corresponding to the trace of a random laser with a cavity length of 1.8 m long is shown in the DOCX file.

6. Typo/grammar issues:

a)       line 56: ‘≈976 nm’ should change into ‘~ 976 nm’

b)       line 48, the authors used ‘…, which… , which…’. This makes the sentence complicated and is not commonly used.

Thank you for your notes, we revised the text and fixed the language mistakes found. The paper was modified accordingly.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors present nice experimental results on a random laser comprising an array of weak fiber Bragg gratings inscribed directly in the Yb-doped cavity fiber. In contrast to the previous studies the laser enables single-frequency operation at 980 nm attractive for many applications. The authors show the way to enhance the efficiency of the random laser operation and improve its stability by supplying the cavity by a single FBG with a 90%-reflectivity. In general, the manuscript is clearly written and well referenced. The reported results could be interesting for people working in the field.  

In my opinion, the manuscript deserves publication in Photonics.

However,  the following question to be addressed before the publication:

1) Should 90% FBG reflectivity be considered as an optimal value for the used propose?

2) What are the parameters of the Yb-doped fiber used for FBG inscription?

3) Could you estimate the linear optical losses inside the cavity? 

Author Response

Thank you for your time and comments provided. In respect for your questions we would like to give our answers to following questions:

1) Should 90% FBG reflectivity be considered as an optimal value for the used propose?

Adding a single FBG to the random laser cavity provides two benefits: increased slope efficiency and reduced output power fluctuations. In addition to 90% reflectance (10 dB peak intensity) FBG, we also tested 50% (3 dB peak intensity) and 99.9% (20 dB peak intensity) FBGs with a 1.8m-long random  laser sample. It was found that the 50%-reflectance FBG did not significantly affect laser power stability compared to the cavity without FBG. The random laser power stability in the case of 99%-reflectance FBG turned out to be lower than in 90%-reflectance one. We also observed an increase in the laser radiation wavelength instability for the 99%-reflectance FBG (possibly due to the wider reflection spectrum of this FBG). Thus, the FBG reflectivity value of 90% is close to optimal.

2) What are the parameters of the Yb-doped fiber used for FBG inscription?

The Yb-doped fiber used for the random FBG cavity fabrication had a core diameter of ~6 μm, outer cladding diameter of 125 μm, cutoff wavelength of 0.89 μm and core-cladding refractive index difference of 0.007. The single FBG with a reflectivity of ~90% used in this work to improve the random cavity output characteristics was inscribed in a commercially-available Corning HI 1060 fiber using a standard phase-mask technique. Clarifications have been added to the text.

3) Could you estimate the linear optical losses inside the cavity? 

We measured the background optical losses in the Yb-doped fiber that was used to the laser cavity fabrication. The measurements were carried out at a wavelength of 1100 nm, where the contribution of  Yb dopant absorption band is obviously negligible. The level of background  losses was about of 2 dB/km i.e. 0.002 dB/m, respectively. So, for random laser sample with an initial cavity length of 8 meters, the linear optical losses can be estimated as 0.016 dB.

Reviewer 4 Report

The manuscript by A. Rybaltovsky et al is devoted to the development of 976 nm narrow-linewidth fiber lasers which has many applications including pump lasers, visible and UV sources, and spectroscopy. The method of fiber manufacturing proposed by the authors allows mass-production of such lasers with relatively low cost as it does not require splicing of separate FBGs to the active fiber. I believe that this paper would be very interesting to the community, so I recommend to accept it for publication in the MDPI Photonics journal. Couple of questions:

1.In the laser configuration demonstrated in Fig. 1 what was the power of the 976 nm signal at the output end of the active fiber (opposite to the end spliced to the WDM)? Or only unabsorbed pump was presented there and no signal?

2. Can the authors comment on the linewidth of the laser? Was the laser output truly single-frequency or it was just narrower than the OSA resolution and actual bandwidth is unknown?

Author Response

Thank you for your time and comments provided. In respect for your questions we would like to give our answers to following questions:

1. In the laser configuration demonstrated in Fig. 1 what was the power of the 976 nm signal at the output end of the active fiber (opposite to the end spliced to the WDM)? Or only unabsorbed pump was presented there and no signal?

Indeed, the studied random-FBG laser generated radiation (976 nm) only in the direction opposite to the pump radiation (908 nm) input. Therefore, only the unabsorbed 908nm wavelength radiation was registered at the non-spliced end of random laser cavity. By the way, we observed this effect earlier in the ~1550nm wavelength random-FBG laser cavity based on the Er-doped fiber (see S.M. Popov et al, Proc. of SPIE’2020, p. 113571Q. DOI: 10.1117/12.2557818). The actual clarification has been added to the text (page 6).

2. Can the authors comment on the linewidth of the laser? Was the laser output truly single-frequency or it was just narrower than the OSA resolution and actual bandwidth is unknown?

In this work, we did not study the laser linewidth directly (for example, by using a scanning Fabry-Perot interferometer or self-heterodyne linewidth measurement setup). Such experiments are the topic for our future research. Now we can only confirm that the investigated random laser linewidth is narrower than the resolution limit (0.02 nm) of the optical spectrum analyzer (Yokogawa AQ6370). However, in our previous work it was found that the random-FBG Erbium laser with a similar design operated in a single-frequency mode and had a linewidth of less than 4 kHz.

Reviewer 5 Report

The authors proposed a random narrow-linewidth laser lasing at the wavelength of 976 nm, obtained in the ytterbium-doped grmanophosphosilicate fiber. The usage of the language was good. However, the content of the manuscript was not well investigated. Thus a rejection was presented. Some of the issues are as follows.

1.        Some of the expressions should be improved. Such as

“The development a random laser for a wavelength range of ~980 nm requires suitable…”

2.        Different line shape should be used in Fig. 2.

3.        Without the incorporation of the FBG, the configuration of Fig.1 is a typically configuration of an amplified spontaneous emission configuration. Why it is a configuration of “random” laser?

4.        What is the real application of this kind of laser?

5.        The author mentioned the narrow-linewidth. However the linewidth was not investigated in detail in the work.

Author Response

Please check the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 5 Report

Could be accepted