High output mode-locked laser empowered by defect regulation in 2D Bi2O2Se saturable absorber

Atomically thin Bi2O2Se has emerged as a novel two-dimensional (2D) material with an ultrabroadband nonlinear optical response, high carrier mobility and excellent air stability, showing great potential for the realization of optical modulators. Here, we demonstrate a femtosecond solid-state laser at 1.0 µm with Bi2O2Se nanoplates as a saturable absorber (SA). Upon further defect regulation in 2D Bi2O2Se, the average power of the mode-locked laser is improved from 421 mW to 665 mW, while the pulse width is decreased from 587 fs to 266 fs. Moderate Ar+ plasma treatments are employed to precisely regulate the O and Se defect states in Bi2O2Se nanoplates. Nondegenerate pump-probe measurements show that defect engineering effectively accelerates the trapping rate and defect-assisted Auger recombination rate of photocarriers. The saturation intensity is improved from 3.6 ± 0.2 to 12.8 ± 0.6 MW cm−2 after the optimized defect regulation. The enhanced saturable absorption and ultrafast carrier lifetime endow the high-performance mode-locked laser with both large output power and short pulse duration.

1) 2D-material-based saturable absorber (graphene) has been discovered and used to passively mode-lock fiber laser as early as in 2009[T.Hasan et al., Adv. Mater. 21, 3874, 2009//Q. Bao et al. Adv. Mater. Funct. 19, 3077, 2009, and the 2D Bi2O2Se has been previously found as an ultrabroadband saturable absorber, so it is not surprising to obtain the mode-locked laser using the 2D Bi2O2Se SA.
2) Defect regulation of mediation is a common way to improve the saturable absorption performance of 2D material, so the use of defect regulated 2D Bi2O2Se is also not novel.
3) As seen in Fig.4c, the output power (665 mW) and the pulse duration (266 fs) in this work are also not highest level in the 2D-material mode-lock laser. In Refs. [41,11], the mode-locked pulse duration can be even as short as 30 fs, and in Refs. [44,48] the output power is as high as 800 mW.
In summary, this work is not novel in principle and the output laser performance is not surprising. I can not recommend it for publication in the high-level journal Nat. Commun.

Point-by-point responses to the reviewers' comments of the manuscript "High output mode-locked laser empowered by defect regulation in 2D Bi2O2Se saturable absorber".
Thank you very much for the reviewers' keen interest and constructive comments on our manuscript entitled "High output mode-locked laser empowered by defect regulation in 2D Bi2O2Se saturable absorber". These comments are all valuable and very helpful for revising and improving our manuscript, as well as making important guiding significance to our research. We have carefully studied the comments and thoroughly made the revisions. The point-by-point responses to the reviewers' comments are listed as follows:

Reviewer #1:
This work demonstrates the availability of mode-locking performance control using Bi2O2Se nanoplates as a saturable absorber (SA) via defect engineering (i.e., plasma irradiation), which enables tailoring of the saturation intensity and carrier lifetime. An improved output power and narrowed pulse width of mode locking have been achieved based on the optimized SA. Detailed characterizations of both the material and laser have also been performed for intuitive comparisons. The approach proposed in this paper is new and should be quite valuable for the development of ultrafast laser fields.
In my option, this work can be considered to be published, after addressing the following issues: Response: We thank the reviewer for his/her positive comments.
1. The achieved mode-locked pulses are not operated at soliton state, why sech 2 instead of Gauss profile was selected to fit the autocorrelation trace?
Response: We thank the reviewer very much for this kind suggestion. The achieved mode-locking pulses are not operated at the soliton state and are slightly chirped, so the autocorrelation trace should be fitted by Gauss rather than the sech 2 model. As shown in Fig. R1, under different plasma irradiation times (0 min, 2 min, 3 min, 5 min) treated on the Bi2O2Se nanoplates, the pulse durations were measured to be 661 fs, 591 fs, 400 fs and 312 fs by Gaussian pulse shape fitting. We have revised and updated it in the manuscript (highlighted in yellow, Fig. 1b and Fig. 4a).

How about the long-term stability of the mode-locked laser?
Response: We thank the reviewer very much for the good suggestion. By using the same Bi2O2Se nanoplates under a plasma irradiation time of 5 min, we reperformed the mode-locked experiment and measured the corresponding output power instability and pulse width. As shown in Fig. R2a, the maximum output power was 654 mW obtained three months later, and the instabilities (average output power, rms) of mode-locked operation were measured to be 1.69% at 12 hours. This not only indicates the long-term stability of the mode-locked laser operation but also illustrates the long-term stability of the Bi2O2Se nanoplates. In addition, the characterization of CWML operation  for the same Bi2O2Se SA before and after exposure to air for three months demonstrates that the Bi2O2Se nanoplate-based SA has excellent long-term stability. We have revised and updated it in the manuscript (highlighted in yellow, page 11, line 218 and Supplementary Fig. 9). 3. What limit further power scaling of the mode-locked pulses? Damage of the SA, disappearance of mode-locking, or others? The relative comments should be provided.

Response:
We thank the reviewer very much for pointing out this. In our experiment, no optical damage was observed in the Bi2O2Se SA during the mode-locked operation.
Upon further increasing the pump power, the Bi2O2Se SA is completely bleached, and there is insufficient further modulation to drive the pulse forming process. Therefore, the CWML operation would become unstable and eventually disappear under high pump power. We have added the corresponding comments to the revised manuscript (highlighted in yellow, page 10, line 215).

Following the above question, whether the mode-locking pulse width can be further narrowed? If yes, how to realize it.
Response: We thank the reviewer very much for pointing out this. The saturable absorption parameters (saturation intensity, recovery time, etc.) of SA are one of the factors that determine the mode-locked pulse width. Therefore, defect engineering is employed in Bi2O2Se SA by moderating Ar + plasma treatments, which significantly accelerates the trapping rate and defect-assisted Auger recombination rate of photocarriers and empowers strong saturable absorption and self-defocusing properties in Bi2O2Se. The improved saturation intensity and ultrafast carrier lifetime make synergetic contributions to the high-performance mode-locked laser with both high output and ultrashort pulse width. The CWML operation was not achieved when further increasing the irradiation time on the Bi2O2Se SA to 6 minutes (The detailed explanation can be found in the response to Question 5). In addition to the saturable absorption parameters of SA, the mode-locked pulse width is also determined by the fluorescence linewidth (emission gain linewidth) of the laser crystal, the intracavity dispersion and the energy. Therefore, by applying a high-quality low-loss cavity mirror (larger coating wavelength band with higher reflection/transmission) and optimizing the dispersion compensation and cavity designation, the mode-locking pulse width can be further narrowed.

If further increasing the irradiation time, whether the mode-locking performance can be improved?
Response: We thank the reviewer very much for this kind suggestion. As shown in Fig.   R3, when increasing the irradiation time on the Bi2O2Se nanoplates to 6 minutes, the saturation recovery time of Bi2O2Se nanoplates shows an accelerated trend, and the saturation intensity is increased, which is consistent with the phenomenon observed in the manuscript. It seems that the mode-locking performance should be improved by further increasing the irradiation time. However, we did not achieve CWML operation based on the Bi2O2Se nanoplate SA treated by 6 min plasma irradiation. First, the larger saturation intensity increases the difficulty of the mode-locking operation. Second, the introduction of a large number of defect states reduces the material quality, as well as the damage threshold of the Bi2O2Se nanoplate. Nonlinear transmittance of Bi2O2Se nanoplates under 6 min plasma irradiation.

Whether this defect engineering method is accessible for other low-dimensional materials?
Response: We thank the reviewer very much for pointing out this. Recently, plasma treatment has become an effective method to create atomics defects in 2D materials 1, 2, 3, 4 , including oxygen plasma 4 and argon plasma 5 . For example, oxygen plasma treatment of ReS2 introduced S vacancies and improved its electrical property and photodetection performance 1 . Argon plasma generated Se vacancies PdSe2 to mediate the phase transition of Pd17Se15 2 and Se vacancies in PtSe2 to result in thicknessindependent semiconducting to metallic conversion 3 . In our manuscript, we prove that the defect regulation of Bi2O2Se nanoplates by oxygen plasma can effectively accelerate their carrier recombination and greatly improve their saturation intensity, thus improving their mode-locked laser performance. Therefore, we believe that this defect engineering method can control the ultrafast carrier dynamics and nonlinear absorption properties of other two-dimensional materials to enhance their pulsed laser performance. R1

Reviewer #2:
This manuscript reports interesting novel experimental results of high output modelocked laser empowered by defect regulation in 2D Bi2O2Se saturable absorber, however several points must be revised before being suitable for publication; Response: Many thanks for the reviewer's positive comments.
With the Bi2O2Se SA used in the cavity, after careful adjustment, the laser runs into a continuous wave mode-locked (CWML) regime when the absorbed pump power exceeds 5.53 W, as shown in Fig. R4a. Under an absorbed pump power of 6.22 W, a maximum average output power of 51 mW is obtained. As shown in Fig. R4b, the output mode-locked laser spectrum is centered at 1941 nm with a full-width at halfmaximum (FWHM) of 6.7 nm. Fig. R4c shows the pulse trains on the 200 ns and 500 μs time scales at the maximum output power, which indicates the realization of modelocked laser operation. In addition, the recorded radio frequency spectrum is shown in     Supplementary Fig. 2b, 8a-c).

The authors claim that 2D Bi2O2Se has excellent air stability, but no data for longterm operation in open air condition can be found. The authors are recommended to make long term operation of their 2D Bi2O2Se saturable absorber and show that the deterioration is negligible under photo-degradation and/or air environment
Response: Thank you very much for the kind suggestion. As shown in Fig. R6, the characterization of the Raman spectra (Fig. R6a) and CWML operation (Fig. R6b-c) for the same Bi2O2Se SA before and after exposure to air for three months demonstrate that the device has excellent long-term stability. Moreover, the instabilities (average output power, rms) of the mode-locked operation are measured to be 1.69% at 12 h.
The corresponding information has been revised and updated in the manuscript (highlighted in yellow, page 11, line 218 and Supplementary Fig. 1d, 9a-c).  Response: Many thanks to the reviewer for reviewing our manuscript and giving his/her constructive comments.

2D-material-based saturable absorber (graphene) has been discovered and used
to passively mode-lock fiber laser as early as in 2009[T. Hasan et al., Adv. Mater. 21, 3874, 2009//Q. Bao et al. Adv. Mater. Funct. 19, 3077, 2009, and the 2D Bi2O2Se has been previously found as an ultrabroadband saturable absorber, so it is not surprising to obtain the mode-locked laser using the 2D Bi2O2Se SA.
Response: Graphene was proven to be an excellent saturable absorber as early as air stability is the largest obstacle for its real applications; for TIs, poor air compatibility and relatively low modulation depth limit its application in high-power high-energy pulsed lasers. Therefore, it is urgent to continue exploring novel 2D materials as well as the properties and performance modulation techniques. Moreover, due to the limited density of states and extremely high saturation intensity of common 2D materials (e.g., graphene and TMDs), it is difficult to generate ultrafast lasers with both high output power and ultrashort pulse width simultaneously.
Recently, 2D bismuth oxyselenide (Bi2O2Se) has been demonstrated to have ultrabroadband nonlinear modulation (from 1.55 µm to 5.0 µm) and a larger nonlinear absorption coefficient β (-2.91×10 -6 cm/W @800 nm. Although it has been demonstrated to have saturable absorption 15 (this work was based on Bi2O2Se nanoflakes produced by a solution method wherein many chemicals were used, which may have contamination effects on the optical property investigations), due to the low saturation intensity of Bi2O2Se (~ MW/cm 2 ), ultrafast mode-locked lasers based on Bi2O2Se SA have not been devised. In this work, for the first time, to the best of our knowledge, a femtosecond solid-state laser with 2D Bi2O2Se nanoplates as SA is Overall, our main innovation is illustrating that Bi2O2Se is a superior 2D SA, defect modulation is an effective way to modulate the SA performance, and a high output power and ultrashort pulse mode-locked bulk laser can be realized and improved by plasma-treated Bi2O2Se SA. 2. Defect regulation of mediation is a common way to improve the saturable absorption performance of 2D material, so the use of defect-regulated 2D Bi2O2Se is also not novel.

Response:
We agree with the reviewer that atomic defects, being the most prevalent zero-dimensional topological defects, are ubiquitous in a wide range of 2D materials.
Hence, controlling atomic defects provides an alternative avenue for engineering a wide range of physical and chemical properties of 2D materials to enhance the performance of their corresponding devices. However, there are few reports on enhancing the performance of 2D material-based mode-locked lasers by defect regulation. In this work, defect regulation effectively accelerates carrier recombination and greatly improves the saturation intensity of Bi2O2Se SA. The improved saturation intensity and ultrafast carrier lifetime endow the high-performance mode-locked laser with both high output and short pulse duration. In our opinion, the emergence of this work effectively fills the gap in the current field. Fig. 4c, the output power (665 mW) and the pulse duration (266 fs) in this work are also not highest level in the 11], the mode-locked pulse duration can be even as short as 30 fs,and in Refs. [44,48] the output power is as high as 800 mW.

As seen in
Response: It is certainly that the output power is not the highest, and the pulse width is not the shortest result among 2D material-based mode-locked lasers. In Refs. [Opt. Lett. 41, 890-893 (2016)], although the mode-locked pulse duration can be as short as 30 fs, its average power is only 26 mW. In Refs. [Opt. Lett. 38, 4189-4192 (2013)], the output power is as high as 800 mW, while the mode-locked pulse duration has reached 643 fs.
Therefore, it is a great challenge to realize mode-locked lasers with high average output power and narrow pulse width at the same time. The main novelty of our work is to explore a new material and a novel performance modulation technique to realize ultrafast lasers with both high output power (665 mW) and an ultrashort pulse width (312 fs) simultaneously.
constructive comment, which can significantly improve our manuscript.
In real mode-locked lasers, the pulse shape is complicated since dispersion management can produce many different pulse shapes: (i) sech 2 pulses with a small dispersion swing; (ii) Gaussian pulses with moderate dispersion; and (iii) the pulse shape depends more sensitively on gain filtering with large dispersion [1]. In addition, for passive mode locking based on a slow saturable absorber with self-phase modulation (SPM) and group delay dispersion (GDD), the pulse shape should be sech 2 if there is soliton formation, while it should be a Gaussian shape if there is no soliton We apologize for the confusing description of the obtained pulse shapes in the previous response letter. We agree with the reviewer that mode-locking soliton pulses have a nearly sech 2 shape even though they are chirped and must be fitted as such.
According to the reviewer's suggestion, we have carefully analyzed our results (especially the autocorrelation curve and spectrum of the mode-locking pulses), reconsidered the dispersion of the laser cavity, and finally concluded that the achieved mode-locked pulses in our experiment should be sech 2 shaped.
First, it is a passive mode-locking operation with a slow saturable absorber since the obtained mode-locking pulse duration is 10 times shorter than the recovery time of the 2D Bi2O2Se saturable absorber.
Second, two Gires-Tournois interferometer (GTI) mirrors with a total negative GDD of ~-750 fs 2 per round were used to compensate for the mirror dispersion (lowloss chirped mirrors with near-zero GDD) and material dispersion (mainly derived from the laser crystal with positive GDD ~700 fs 2 ) inside the resonator and balance the SPM induced by Kerr nonlinearity of the crystal. The total dispersion of the laser cavity was small and negative.
Third, as shown in Fig. R1-R2, the autocorrelation trace and spectrum of modelocked pulses are both almost perfectly sech 2 fitted, which is evidence for soliton-like pulse generation (although in theory, in both cases, this is an approximation) [3]. As shown in Fig. R1 In conclusion, (i) the dispersion of our laser cavity is small and negative; (ii) the autocorrelation trace and spectrum of mode-locked pulses both are well described by sech 2 fitting; (iii) the TBP of obtained mode-locking pulses with sech 2 shapes is slightly higher than that of the transformed limitation and indicates slight chirping; and (iv) limited by the elements in hand, the dispersion of the cavity has been optimized for proper mode locking with long-term stability (rms=1.69%@12 hrs).
Based on the above experimental evidence and analysis, we are confident that the mode-locking pulses obtained in our experiment are sech 2 shapes.
We have carefully revised and updated it in the revised manuscript (highlighted in yellow, page 4, line 77, page 10, line 206, Fig. 1b, Fig. 4a, Fig. 4c, Supplementary Fig.   2b and Supplementary Fig. 8).   B, 16, 1999-2004 (1999 In a word, they give the convincing explanation to my questions, and this work may be an important step for 2D-material mode-locked high-power laser into practical application. Therefore, I agree to accept the revised manuscript for publication in Nature Communication now.
Response: Many thanks for the reviewer's positive remarks, which certainly confirm the novelty and importance of our work.