High resolution imaging with anomalous saturated excitation
Bo Du, Xiang-Dong Chen, Ze-Hao Wang, Shao-Chun Zhang, En-Hui Wang, Guang-Can Guo, Fang-Wen Sun
aa r X i v : . [ phy s i c s . op ti c s ] F e b High resolution imaging with anomalous saturated excitation
Bo Du,
1, 2
Xiang-Dong Chen,
1, 2
Shao-Chun Zhang,
1, 2
Ze-Hao Wang,
1, 2
En-Hui Wang,
1, 2
Guang-Can Guo,
1, 2 and Fang-Wen Sun
1, 2, ∗ CAS Key Lab of Quantum Information, University of Science and Technology of China, Hefei, 230026, P.R. China CAS Center For Excellence in Quantum Information and Quantum Physics,University of Science and Technology of China, Hefei, 230026, P.R. China
The nonlinear fluorescence emission has been widely applied for the high spatial resolution opticalimaging. Here, we studied the fluorescence anomalous saturating effect of the nitrogen vacancy defectin diamond. The fluorescence reduction was observed with high power laser excitation. It increasedthe nonlinearity of the fluorescence emission, and changed the spatial frequency distribution of thefluorescence image. We used a differential excitation protocol to extract the high spatial frequencyinformation. By modulating the excitation laser’s power, the spatial resolution of imaging wasimproved approximate 1.6 times in comparison with the confocal microscopy. Due to the simplicityof the experimental setup and data processing, we expect this method can be used for improvingthe spatial resolution of sensing and biological labeling with the defects in solids.
I. INTRODUCTION
With stable photon emission and optically controlledspin, the fluorescent defects in solids have been widelystudied for the quantum information processing, sens-ing and biological imaging [1–7]. One of the mostpromising candidates is the nitrogen vacancy (NV) cen-ter in diamond. Due to its atomic size, the sensingand imaging with NV center show the advantage of highspatial resolution. Several optical super-resolution mi-croscopy techniques have been demonstrated for the sub-diffraction resolution imaging of NV center[8–14]. Shape-modulation of the laser beam or selective excitation ofsingle NV centers is usually required for these imagingmethods.Meanwhile, the nonlinear optical response of the flu-orescence is also explored to improve the spatial res-olution of the microscopy [15–20]. The key is to in-crease the high spatial frequency component of the im-age, and suppress the low spatial frequency component.For example, saturated excitation microscopy (SAX) hasbeen demonstrated by utilizing the nonlinearity of thesaturation[21, 22] . The high spatial frequency compo-nent of a fluorescence image can be extracted with theharmonic demodulation or the differential excitation[23].The nonlinearity of the fluorescence determines the res-olution and signal-to-noise ratio of the images. There-fore, emitters with a high nonlinearity of the fluores-cence emission are required for the high resolution imag-ing [16, 24, 25]. Combined with STED microscopy,the nonlinear fluorescence emission has also been uti-lized for the super-resolution imaging with upconversionnanoparticles[26–28].Recent experiments reveal that the excitation depen-dence of the NV center fluorescence is different from atraditional saturating[29–31]. Fluorescence reduction is ∗ [email protected] observed with a high-intensity laser excitation. This ef-fect could enhance the nonlinearity of fluorescence emis-sion. Here, by utilizing this anomalous saturation effect,we demonstrated a sub-diffraction microscopy for the NVcenter imaging. A single Gaussian-shaped laser beampumped the fluorescence of NV center through a scanningconfocal microscope. Due to the fluorescence reductionwith a high intensity laser excitation, a doughnut-shapedfluorescence image with high spatial frequency compo-nent was obtained. To extract the high spatial frequencyinformation of the images, we applied a differential exci-tation protocol, where the fluorescence intensities of NVcenter with high and low power lasers were simultane-ously recorded and compared. The results showed thatthe resolution of NV center imaging can be improved 1.6times by utilizing the anomalous saturating. With thesimple experimental setup and data processing, we ex-pected this anomalous saturated excitation (ASAX) mi-croscopy can be used to improve the spatial resolution ofNV center’s sensing and imaging. II. PRINCIPLE AND EXPERIMENTAL SETUP
The NV center in diamond usually shows two chargestates: the negatively charged NV − and the neutrallycharged NV . The ground state and excited state of NV − are spin triplet. The excited state of NV − can decay tothe ground state through the spontaneous emission orthe non-radiative inter-system crossing (ISC). The ISCincludes the decay from the excited state to a metastablestate, and the subsequent decay from the metastablestate to the ground state. The transition between the ex-cited state and the metastable state mainly occurs with m s = ± m s = 0. As a result, the fluorescenceintensity of NV − with electron spin m s = 0 is higher thanthat with m s = ±
1. The ISC process also polarizes thespin state to m s = 0. The property of NV − electron spin -2 -1 Amplitude (a.u.)
Intensity (a.u.) A m p li t ud e ( a . u . ) Frequency ( mm -1 ) (d) E m i ss i on ( c t s / . ) Laser intensity ( MW/cm ) Spatial-domain Frequency-domain (a) (c)(b) P P AFG
532 nm
Mirror MirrorDMPinholeFilterAPD ObjectiveDiamondPulsed laser LensAOM R a t i o ( % ) Power (mW)
FIG. 1. (a) Experimental setup. The pulsed 532 nm laser is circularly polarized, with a repetition rate of 5 MHz. DM: dichroicmirror; APD: avalanche photodiode. (b) The fluorescence emission of a single NV center with different laser intensities. Theblue line is the fitting with Eq.1. (c) Simulated point spread function and the spatial frequency distribution of the confocalmicroscopy with different powers. (d) The normalized spatial frequency with the excitation power of 0.06 and 1.2 mW. Forcomparison, we also show the results with a traditional saturating (b = 0 in Eq.1). The insert is the ratio of the high spatialfrequency signal with different excitation laser powers. enables it to be widely used as a qubit, while the spinof NV is rarely studied. Therefore, ’NV center’ in thiswork, unless otherwise specified, will refer to the NV − charge state. The zero phonon line of NV − spontaneousemission is at 637 nm, while the phonon sideband with apeak around 700 nm is observed at room temperature.The charge state conversion between NV − and NV can be pumped by laser with ultra-violet and visiblewavelength. It has been demonstrated that the chargestate conversion is a spin depolarization process[29],which will decrease the fluorescence intensity of NVcenter. The fluorescence reduction during the photon-induced spin depolarization process can be written as1 + b · e − γt . γ is the charge state conversion rate, and t is the laser duration. b presents the amplitude of flu-orescence reduction, which is determined by the fluores-cence difference between m s = 0 and m s = ± γ ∝ P [32, 33].Therefore, the fluorescence reduction that is induced bythe spin depolarization can be written as 1 + b · e − P/P ,where P is the saturation power of this spin depolariza-tion effect. Here, we assume that the width of laser pulsedoes not change with the laser intensity.Based on the theoretical analyzing, we can write theexcitation power dependence of NV center fluorescence by including the traditional saturating and the spin de-polarization: I ( P ) = I sat PP + P (1 + b · e − P/P ) . (1)Here, PP + P presents the saturation of the ground state toexcited state transition, where P is the saturation powerof this transition. I sat is the fluorescence intensity withan infinity excitation intensity.For sub-diffraction NV center imaging, the experi-ments were carried out with a home-built confocal micro-scope, as shown in Fig. 1(a). NV centers in a chemicalvapor deposition diamond plate were produced by nitro-gen ion implanting and subsequent annealing. A 0.9 N.Aobjective was used to focus the laser beam and collectthe fluorescence of NV center. The laser beam was a 532nm picosecond pulsed laser with a Gaussian shape. Thelaser passed through an acousto-optic modulator (AOM),the diffraction efficiency of which was modulated by anexternal analog signal (0 - 1 V) from an arbitrary func-tion generator (AFG). The first order of the diffractionbeam was used to excite the NV center. In this way, wecan change the power of the excitation laser at a timescale of nanoseconds. A long pass dichroic mirror withan edge wavelength of 640 nm separated the excitationlaser and the fluorescence. A long pass optical filter withan edge wavelength of 659 nm was used to further block I low g (cid:215) I high I r =- g S NR g = 0 g = 0.2 g = 0.6 g = 1 A m p li t ud e ( a . u . ) Frequency (mm -1 ) (a) (b)(c) (d)(e) (f) Exc.Fluor. I high I low R es o l u t i on ( n m ) g g = 0 g = 0.6 g = 1.0 I n t e n s i t y ( a . u . ) Lateral Position ( mm ) FIG. 2. (a) The sequences of the laser’s modulation and fluorescence detection. (b) The images of a single NV center. Theconfocal microscopy images of I low and I high are recorded to calculate the final image I r . Cross-section profiles correspondingto the dashed lines in the images of single NV center are shown on the upsides. (c) The spatial resolution of ASAX microscopywith different γ . The error bars are the deviation of the results. The insets are images of single NV center with γ =0, 0.6 and1.0. The scale bars are 500 nm in length. (d) The lateral cross-section profiles of ASAX images with γ =0, 0.6 and 1.0. (e) Thespatial frequency distribution of ASAX images with different γ . (f) The SNR of the I r as a function of γ . Confocal (a) (b) (c)(d)
ASAX I n t e n s i t y MinMax
Confocal ASAX I n t e n s i t y MinMax (e) (f)
Confocal ASAX I n t e n s i t y ( a . u . ) Position (mm)280 nm
Confocal ASAX I n t e n s i t y ( a . u . ) Position (mm)350 nm
FIG. 3. Multiple NV centers imaged with the confocal (a)(d) and ASAX microscopies (b)(e). The scale bars are 1 µm inlength. (c) and (f) are the cross-section profiles indicated by the arrows. the excitation laser and the spontaneous emission fromNV . Finally, the phonon sideband of NV − center fluo-rescence was detected by an avalanche photodiode. Andthe numbers of photon counts were recorded by a dataacquisition card for further analysis. III. RESULTS
The fluorescence emission of a single NV center underthe excitation of a pulsed laser with a repetition rate of 5MHz was measured in Fig. 1(b). As expected, it showedthat the fluorescence intensity increased with the excita-tion power in the weak pumping region, but decreasedwith the laser power in the strong pumping region. Thethermal effect was excluded by measuring the spin tran-sition signal of NV center[29]. The results in Fig. 1(b)can be well fitted by Eq. 1, with the coefficients of P =0.048 MW/cm and P = 0.32 MW/cm . Here, with a5 MHz repetition rate of the excitation laser, the highestfluorescence intensity was observed with an average laserintensity of 0.128 MW/cm . The contrast between thehighest fluorescence intensity and the saturated fluores-cence intensity was I max − I sat I max = 0 .
61, corresponding to afluorescence reduction amplitude of b ≈ . and 0.256 MW/cm at thebeam center, respectively. It showed that, with a highlaser power, the image of a scanning confocal microscopechanged to a doughnut shape. In Fig. 1(d), we showedthe spatial frequency distribution with different powers.To quantitatively present the relation between spatial fre-quency distribution and the excitation laser power, wecalculated the ratio of spectral component with the spa-tial frequency higher than that corresponded to the reso-lution of the confocal microscopy. As shown in the insertof Fig. 1(d), the component of the high spatial frequencywas improved by increasing the laser power. The spatial frequency distribution of the traditional saturating wasalso simulated by setting b = 0 in Eq. 1. It showedthat the high spatial frequency component with anoma-lous saturation was higher than that with the traditionalsaturation.To experimentally obtain a sub-diffraction resolutionimage, a differential excitation protocol was applied.The key is to simultaneously record two or more imageswith different spatial frequency distributions. It canbe realized by changing the shape of excitation laserbeam[17, 34–36]. Here, we modulated the power of theexcitation laser to obtain images with different spatialfrequency distributions. Two excitation laser powerswere used. To ensure a large spatial frequency differencebetween two images, the peak intensity of the laser beamfor the weak excitation was lower than P . And for thestrong excitation, the peak intensity of the laser beamwas at the scale of P . A square wave pulse sequencewith a period of 100 µs was generated by the AFG tocontrol the diffraction efficiency of AOM. As shown inFig. 2(a), The high and low levels of the square wavesequence were 1 and 0.2 V, respectively. And the dutyratio of the pulse sequence was 50%. Then the excitationlaser power would switch between P and P with aperiod of 100 µs . The fluorescence intensities with thehigh and low power excitation were separately recordedby two channels as I low and I high .In Fig. 2(b), we showed the image of a single NVcenter. The high and low levels of laser power were set to1.2 mW and 0.06 mW. The scanning confocal image wasobtained with a pixel dwell time of 50 ms. As expectedthe fluorescence image of I high showed a doughnut shape,while the image of I low was Gaussian shape. To obtainthe image with a sub-diffraction spatial resolution, wesimply calculated a signal as I r = I low − γ × I high . (2)Here, we have normalized I low and I high by dividing themaximum of each channel. An intuitive understandingof this calculation is that the off center signal of I low issuppressed because of the doughnut shape of the imagewith I high . It would subsequently improve the spatialresolution of imaging with I r . With a small γ , the imageof I r is almost the same as I low . And with a large γ , I r would be mainly determined by the distribution of I high .By adjusting the factor γ , the suppression of low spatialfrequency signals will be optimized.In Fig. 2(c), we presented the single NV center imag-ing with different γ . For γ < .
6, the image of I r stillshowed a single peak. The full width at half maximum(FWHM) of the image decreased with the factor γ . For γ > .
6, negative values would emerge for I r . The imagecan not be simply fitted by a single peak function. Andthe width of the image did not show a further decreaseby increasing the factor γ . Specifically, in Fig. 2(c), theFWHM of the image I r with γ = 0 . γ = 0) was 445 nm,as shown in Fig. 2(d). It demonstrated that the reso-lution of imaging can be improved 1.6 times by utilizingthe ASAX microscopy. The frequency distributions withdifferent γ were also shown in Fig. 2(e). The results in-dicated that the ratio of the high-frequency componentreached its maximum at γ = 0 .
6. It agreed with thechange of spatial resolution with γ . On the other hand,the signal-to-noise ratio (SNR) determines the visibilityof the image. By increasing γ , the signal (the peak-to-peak value of fluorescence image) will decrease, while thenoise is increased. For the single NV center imaging inFig. 2(f), it showed that the SNR decreased from 23 to 15by increasing γ from 0 to 0.6. And it further decreased toapproximate 7 with γ = 1. Therefore, considering boththe spatial resolution and SNR, the subtractive factor γ was chosen to be 0.6 for the best imaging quality in oursystem. The noise mainly originated from the shot noise.It caused the fluctuation of photon counts at each pixel.For the imaging in this work, we chose the pixel size muchsmaller than the FWHM of the point spreading function.Therefore, the shot noise would mainly affect the highspatial frequency distribution. It would not significantlyaffect the estimation of spatial resolution. But the noisemight decrease the accuracy of NV center localizationand the contrast of the image.To further confirm the improvement of the spatial reso-lution with ASAX microscopy, we applied this techniquefor the imaging of multi NV centers. In Fig.3, the high(low) level of the excitation laser power was set to 1.4(0.07) mW. To extract the high spatial frequency signal,the factor γ was set to 0.6. We compared the images ofconfocal microscopy ( γ = 0) and ASAX microscopy. Itconfirmed that the resolution for high density NV cen-ter imaging can be improved with ASAX microscopy. Asindicated in the cross-section profiles (Fig.3(c) and (f)),the NV centers with distances of 280 and 350 nm canbe hardly distinguished with confocal microscopy, as thedistances were shorter than the resolution of confocal mi-croscopy. In contrast, the image of ASAX microscopyseparated the signal from multi NV centers. However,for the imaging with high density NV center, the signalfrom adjacent emitters contributed to the background ofimage with high excitation power. It might decrease thesignal to noise ratio of the final image with differentialexcitation[28]. IV. DISCUSSION AND CONCLUSION
Solid state defect with spin-dependent fluorescenceemission is of interest to researches of quantum informa- tion processing, biological imaging and nanophotonics.And the charge state conversion has been observed invarious solid state defects[37, 38]. We expected that thefluorescence modulation with charge state manipulationin these defects could be utilized to develop the superresolution imaging technique. It could help to realize thehigh resolution spin state manipulation and detection.The ASAX microscopy provided one of the most simpletechniques for detecting the defect with a resolution be-low the diffraction limit. It can be applied for the highresolution imaging of fluorescent defects and other parti-cles with similar nonlinear optical response[39]. Accord-ing to the results, the amplitude of fluorescence reductionwith the high power excitation determines the final imag-ing quality. Since the spin depolarization of NV centerchanges the fluorescence lifetime of NV center, the time-gated fluorescence detection technique [40, 41] could beused to improve the amplitude of fluorescence reduction.Other methods, such as that using laser beams with dif-ferent shapes can also be applied to further improve thespatial resolution of the microscopy with anomalous sat-urated excitation[42].In summary, we studied the anomalous saturation ofNV center fluorescence emission. The fluorescence reduc-tion with high power excitation increased the nonlinear-ity of fluorescence emission. The sub-diffraction ASAXmicroscopy was demonstrated by extracting the high spa-tial frequency information with the high power laser exci-tation. The spatial resolution of NV center imaging wasimproved approximately 1.6 times in comparison with theconfocal microscopy. Considering the simple experimen-tal setup and data processing of this microscopy method,we expect that the technique can be applied for the quan-tum sensing and biological imaging with NV center andother defects.
FUNDING
This work was supported by National Key Re-search and Development Program of China (No.2017YFA0304504); Anhui Initiative in Quantum Infor-mation Technologies (AHY130100); National NaturalScience Foundation of China (Nos. 91536219, 91850102).
DISCLOSURES
The authors declare no conflicts of interest. [1] F. Jelezko and J. Wrachtrup, Phys. Status Solidi A ,3207 (2006). [2] M. W. Doherty, N. B. Manson, P. Delaney, F. Jelezko,J. Wrachtrup, and L. C. Hollenberg, Phys. Rep , 1(2013). [3] R. Schirhagl, K. Chang, M. Loretz, and C. L. Degen,Annu. Rev. Phys. Chem , 83 (2014).[4] L. J. Rogers, K. D. Jahnke, M. H. Metsch,A. Sipahigil, J. M. Binder, T. Teraji, H. Sumiya,J. Isoya, M. D. Lukin, P. Hemmer, and F. Jelezko,Phys. Rev. Lett. , 263602 (2014).[5] A. Gottscholl, M. Kianinia, V. Soltamov, S. Orlin-skii, G. Mamin, C. Bradac, C. Kasper, K. Krambrock,A. Sperlich, M. Toth, I. Aharonovich, and V. Dyakonov,Nat. Mater. , 540 (2020).[6] J.-W. Fan, I. Cojocaru, J. Becker, I. V. Fedotov, M. H. A.Alkahtani, A. Alajlan, S. Blakley, M. Rezaee, A. Lyamk-ina, Y. N. Palyanov, Y. M. Borzdov, Y.-P. Yang,A. Zheltikov, P. Hemmer, and A. V. Akimov, ACS Pho-tonics , 765 (2018).[7] G. Wolfowicz, C. P. Anderson, B. Diler, O. G. Poluek-tov, F. J. Heremans, and D. D. Awschalom, Sci. Adv. (2020).[8] E. Rittweger, K. Y. Han, S. E.Irvine, C. Eggeling, andS. W.Hell, Nat. Photon. , 144 (2009).[9] X.-S. Yang, Y.-K. Tzeng, Z. Zhu, Z.-H. Huang, X.-Z.Chen, Y.-J. Liu, H.-C. Chang, L. Huang, W.-D. Li, andP. Xi, RSC Adv. , 11305 (2014).[10] M. Pfender, N. Aslam, G. Waldherr, P. Neumann, andJ. Wrachtrup, Proc. Natl. Acad. Sci. , 14669 (2014).[11] X.-D. Chen, C.-L. Zou, Z.-J. Gong, C.-H. Dong, G.-C.Guo, and F.-W. Sun, Light Sci. Appl. , e230 (2015).[12] E. Bersin, M. Walsh, S. L. Mouradian, M. E. Trusheim,T. Schr¨oder, and D. Englund, npj Quantum Information , 38 (2019).[13] J.-C. Jaskula, E. Bauch, S. Arroyo-Camejo, M. D. Lukin,S. W. Hell, A. S. Trifonov, and R. L. Walsworth, Opt.Express , 11048 (2017).[14] M. Barbiero, S. Castelletto, X. Gan, and M. Gu, LightSci. Appl. , e17085 (2017).[15] E. H. Rego, L. Shao, J. J. Macklin, L. Winoto, G. A.Johansson, N. Kamps-Hughes, M. W. Davidson, andM. G. L. Gustafsson, Proceedings of the NationalAcademy of Sciences , E135 (2012).[16] D. Denkova, M. Pl¨oschner, M. Das, L. M. Parker,X. Zheng, Y. Lu, A. Orth, N. H. Packer, and J. A.Piper, Nat. Commun , 3695 (2019).[17] G.-Y. Zhao, C. Zheng, C.-F. Kuang, R.-J. Zhou, M. M.Kabir, K. C. Toussaint, W.-S. Wang, L. Xu, H.-F. Li,P. Xiu, and X. Liu, Phys. Rev. lett. , 193901 (2018).[18] R. Heintzmann, T. M. Jovin, and C. Cremer,J. Opt. Soc. Am. A , 1599 (2002).[19] X.-D. Chen, S. Li, B. Du, Y. Dong, Z.-H. Wang, G.-C.Guo, and F.-W. Sun, Opt. Lett. , 699 (2018).[20] I. Gregor, M. Spiecker, R. Petrovsky,J. Großhans, R. Ros, and J. Enderlein,Nat. Methods. , 1087 (2017).[21] K. Fujita, M. Kobayashi, S. Kawano, M. Yamanaka, andS. Kawata, Phys. Rev. Lett. , 228105 (2007). [22] Y. Nawa, Y. Yonemaru, A. Kasai, R. Oketani,H. Hashimoto, N. I. Smith, and K. Fujita, APL Pho-tonics , 080805 (2018).[23] C. Li, V. N. Le, X. Wang, X. Hao, X. Liu, and C. Kuang,Laser & Photonics Reviews , 1900084 (2020).[24] J. Vogelsang, T. Cordes, C. Forthmann, C. Steinhauer,and P. Tinnefeld, Nano Letters , 672 (2010).[25] S. Hennig, S. van de Linde, M. Heilemann, and M. Sauer,Nano Letters , 2466 (2009).[26] M. Pl¨oschner, D. Denkova, S. D. Camillis, M. Das, L. M.Parker, X. Zheng, Y. Lu, S. Ojosnegros, and J. A. Piper,Opt. Express , 24308 (2020).[27] S. De Camillis, P. Ren, Y. Cao, M. Pl¨oschner,D. Denkova, X. Zheng, Y. Lu, and J. A. Piper, Nanoscale , 20347 (2020).[28] C. Chen, F. Wang, S. Wen, Q. P. Su, M. C. L. Wu, Y. Liu,B. Wang, D. Li, X. Shan, M. Kianinia, I. Aharonovich,M. Toth, S. P. Jackson, P. Xi, and D. Jin, Nat. Commun , 3290 (2018).[29] X.-D. Chen, L.-M. Zhou, C.-L. Zou, C.-C. Li, Y. Dong,F.-W. Sun, and G.-C. Guo, Phys. Rev. B , 104301(2015).[30] R. Chapman and T. Plakhotnik,Phys. Rev. B. , 045204 (2012).[31] K. Y. Han, D. Wildanger, E. Rittweger, J. Meijer, S. Pez-zagna, S. W. Hell, and C. Eggeling, New J. Phys. ,123002 (2012).[32] X.-D. Chen, S. Li, A. Shen, Y. Dong, C.-H. Dong, G.-C. Guo, and F.-W. Sun, Phys. Rev. Applied. , 014008(2017).[33] N. Aslam, G. Waldherr, P. Neumann, F. Jelezko, andJ. Wrachtrup, New J. Phys. , 013064 (2013).[34] G. Zhao, M. M. Kabir, K. C. Toussaint, C. Kuang,C. Zheng, Z. Yu, and X. Liu, Optica , 633 (2017).[35] K. Korobchevskaya, C. Peres, Z.-B. Li, A. Antipov,C. J. R. Sheppard, A. Diaspro, and P. Bianchini, Sci.Rep. , 25816 (2016).[36] S. J. Hewlett and T. Wilson, Mach Vision. Appl. , 233(1991).[37] S. Dhomkar, P. R. Zangara, J. Henshaw, and C. A. Mer-iles, Phys. Rev. Lett. , 117401 (2018).[38] G. Wolfowicz, C. P. Anderson, A. L. Yeats, S. J. White-ley, J. Niklas, O. G. Poluektov, F. J. Heremans, andD. D. Awschalom, Nat. Commun. , 1876 (2017).[39] X. Ouyang, F. Qin, Z. Ji, T. Zhang, J. Xu, Z. Feng,S. Yang, Y. Cao, K. Shi, L. Jiang, and X. Li, APLPhotonics , 110801 (2018).[40] X.-D. Chen, Y. Zheng, B. Du, D.-F. Li, S. Li, Y. Dong,G.-C. Guo, and F.-W. Sun, Phys. Rev. Applied ,064024 (2019).[41] B.-W. Zhao, X.-D. Chen, E.-H. Wang, Y. Zheng,B. Du, S. Li, Y. Dong, G.-C. Guo, and F.-W. Sun,Appl. Opt. , 6291 (2020).[42] G. Zhao, C. Kuang, Z. Ding, and X. Liu, Opt. Express24