Open Access
Add to BibSonomyAbstract
High resolution X-ray diagnosis is a significant method for obtaining ablation-front and trajectory measurements targeting Rayleigh-Taylor (RT)-instability growth in initial confinement fusion (ICF) experiments. In this paper, a novel Kirkpatrick-Baez-type structure, as a kind of essential X-ray micro-imaging apparatus, has been developed that realizes a large field of view (FOV) and images with high resolution and energy response. Zoned multilayer coating technology is applied to the Kirkpatrick-Baez mirrors to transmit two specific quasi-monochromatic light through the same mirror and enables a compact dual-channel structure. This microscope has been assembled in the laboratory and later implemented at the Chinese SG-III laser facility. The characterization results show that this imaging system can achieve a good spatial resolution of 5 µm in a large FOV of 500 µm, while maintaining a strong monochromatic performance with bandwidth of 0.5 keV at 2.5 keV and 4.3 keV respectively.
© 2017 Optical Society of America
1. Introduction
Obtaining more precise experimental data with diagnostic systems for simulating target implosion processes precisely is significant when designing ignition methods for initial confinement fusion (ICF) experiments [1]. In the acceleration phase of the target compression, a gated, time-resolved X-ray imager is an essential device for the diagnosis of the temporal evolution of the target including the target trajectory, and the Rayleigh-Taylor (RT) instability growth of the ablator front, which seriously impacts the target asymmetry and laser illumination nonuniformity of the system [2]. Currently various experimental studies of RT instability focus on the simplified model of a sinusoidal perturbation target, especially a 1-D sinusoidal planar target, which has 50 µm period and different perturbation amplitudes. The size of target working area is almost 500 µm [3]. The aim of the X-ray diagnosis is to record the change in width of the fringe with high spatial resolution because the information modulated by RT growth microstructure is in the high spatial frequency range.
X-ray pinhole frame camera is the basic device used for X-ray diagnosis in ICF because of its simple structure, large range of magnification and large field of view (FOV). Almost all physical experimental results for RT instability obtained from ICF experiments have been diagnosed by pinhole cameras. For example, experiments at the NOVA [4], the OMEGA [5], the National Ignition Facility (NIF) [6], and the Chinese SG-II laser facility have used this approach. While, the precision of this technique is limited mainly by three factors. First, pinhole cameras have a spatial resolution limit of ~10 µm. Once this resolution is reached it is difficult to ensure enough power for light gathering. The X-ray imager loses high-frequency information at this resolution level. Second, pinhole camera has no ability for spectral resolution, which is limited by its imaging principle. In the high-energy X-ray imaging experiments with an energy of greater than 2 keV especially, the source is linear emission. Different energy points for synchronous monochromatic X-ray imaging become necessary. The third limitation is potential device damage when working at close range (of less than 60 mm from the pinhole camera) to high laser power implosion experiments [7]. Extending working distance is not appropriated as the gathering light power will seriously degrade.
An alternative device which may diminish these limitations is the reflective Kirkpatrick-Baez (KB) mirror system [8], which is also widely used in ICF experiments as an X-ray imager. It has a high spatial resolution, high power gathering efficiency, large working distance of about 100 mm∼200 mm and good monochromatic energy response when using multilayer technology [9]. At the OMEGA facility, KB mirrors were used for physical experiments in diagnosis because of their spatial resolution of 5 µm and 200 µm FOV [10]. A new KB microscope for the NIF has been in development since 1998 [11]. The results of core implosion imaging were obtained in 2013 [12] and the first KB microscope was established in 2016 [13]. However, a traditional KB mirror is only applicable for core target implosion experiments because its target size of 100 µm fits the high resolution and high reflectivity area of the FOV. Beyond the area however, both performances will degrade significantly due to the off-axis aberration and narrow angular bandwidth of the periodic multilayer, which cannot match the specific requirements for RT instability experiments effectively.
Influenced by the work of Bennett (1999) [14], a primary compound KB-advanced (KBA) system was designed by our team, which can maintain a high level of imaging performance for a large FOV [15,16]. In this paper, the development of this novel compound KBA system with two channels which can image at 2.5 keV and 4.3 keV simultaneously is discussed. The dual-energy imaging of RT instability and shell trajectory is beneficial to the analysis of data [17]. The optical, specific multilayer design is shown and various adjustment error allowances are discussed. This system has already been implemented at the new generation SG-III laser facility and has already obtained two static images of a gold mesh target in implosion experiments. Characterization results show that the system can achieve a high spatial resolution of 5 µm in one dimension with a 500 µm FOV and an energy resolution of 0.5 keV at the two designed energy points.
2. Design of microscope system
2.1 Optical design
A traditional KB mirror has the form of two spherical mirrors set perpendicular to each other to focus light in the tangential and sagittal direction respectively. Due to the dissatisfaction of the Abbe sine condition, the KB system has a serious off-axis coma aberration and the resolution degrades rapidly with an extension of the FOV, as shown in Fig. 1(a). This limits the effective FOV down to ∼200 µm. Alternatively, to maintain a high energy response, a strong monochromatic performance, and a high reflectivity at grazing angles of greater than 1°, depositing a specific periodic multilayer is a commonly used method. However, a periodic multilayer has a narrow angular bandwidth, which means that the multilayer can only transmit light from a small area of object plane with high reflectivity. Therefore, there are two factors that limit the size of the effective FOV for a KB system.
To reduce the off-axis aberration, we use KBA structure [15,16]. The KBA structure has two spherical mirrors arranged in the same direction with the second mirror having a deflection angle of 2θ to realize a dual-reflection, which can satisfy the Abbe sine law and eliminate the obliquity of the image plane as shown in Fig. 1(b). The focal length and object distance of KBA structure needs to satisfy Eq. (1).
Where f is the focal length of the KBA system, u is the object distance, R is the curve radius of both mirrors, M is the magnification of the whole system, θ is the central grazing angle on each mirror, and d is the distance between the two centers of mirrors. f1 and f2 are the focal lengths of two spherical mirrors respectively. Because both of the mirrors have the same R and θ, the focal length f of the whole system can be easily derived from the second formula in Eq. (1). The angles of incidence light and outgoing light of the system are nearly equal to 4θ, which lets light from the off-axis object focus close to the ideal image plane. The simulated geometrical resolution results of KBA and KB mirrors with the same numerical aperture are calculated by MATLAB as shown in Fig. 2. In ideal conditions, the FOV of the KB system at a spatial resolution level greater than 5 µm is 300 µm, while for the KBA configuration, this can be extended up to 1400 µm.
From the aspect of energy transmission, the periodic multilayer has a narrow angular bandwidth not beyond 0.1° and this factor limits the size of FOV less than 200 µm. The Pt single layer is used instead of multilayer, because when light is transmitted by the mirrors over the total reflective angle range, the single metal layer can maintain a high reflectivity of approximately 60% for a large angle bandwidth and the FOV can be extended to 1∼2 mm. However, the Pt layer has no ability for spectra resolution, so a single spherical mirror deposited with a periodic multilayer is set vertically after the KBA configuration as a focusing mirror and a monochromator. Because the reflective angles of two directions are independent, the FOV will be restricted to 200 µm just in the KB direction, while the KBA direction will be maintained as large as ∼1 mm.
2.2. Multilayer design
To record two images at 2.5 keV and 4.3 keV simultaneously, two independent channels are necessary. We used four mirrors in the form of two KBA channels in the tangential direction, as shown in Fig. 3. And in the sagittal direction, we used one spherical KB mirror with enough width to receive the light emitted from both KBA channels at the same time.
Zoned multilayer coating technology was applied on the KB mirror. The working surface of the mirror is divided into two parts and two kinds of multilayers which are designed to response at 2.5 keV and 4.3 keV relatively are zoned deposited. The size of each area is 15 mm × 20 mm, which is large enough for reflecting light from the KBA channel. By this method, the system becomes more compact and benefits associated with the mechanical design and assembling of the structure. The optical design parameters of the system are shown in Table 1.
Table 1. Optical parameters of KBA-KB system
All of the KBA mirrors are deposited with Pt single layer and the grazing angle is designed to be 0.77°, so that X-ray has a broad energy bandwidth from 2 keV to 4.5 keV that can be reflected with high reflectivity of almost 60%. The energy response of Pt layer at designed grazing angle is shown in Fig. 4(a). The spectral resolution results of these two multilayers are shown in Fig. 4(b). Both of two multilayers have a 0.5 keV resolved capability. Figure 4(c) shows the reflectivity of two multilayers versus the grazing angle at each designed energy point. The Cr/C multilayers have a high energy response beyond 70% at 2.5 keV and the W/B4C multilayers at 4.3 keV. Both of those response peaks are overlapped near 1.2427°. The bandwidth angle of the W/B4C multilayer is 0.1°, which can transmit light from the area of object plane as large as 400 µm. The size of that for Cr/C multilayer is obviously larger. We simulate the 2-D energy response of the system versus the FOV at two energy points and the results are shown in Figs. 4(d) and 4(e). The energy centers are slightly deviated from the geometrical FOV centers, because the optimal grazing angles of two kinds of multilayers do not coincide well in the actual deposition process. However, the system still ensures that a 600 µm FOV can be realized in vertical direction and has the same level of capability to transmit power at two energy points. Additionally, this system has a better light gathering performance which is proportional to the square of numerical aperture. Especially compared with the common pinhole camera used in ICF diagnostics which has a working distance of tens of millimeters, a pinhole diameter of ~10 µm, and thus a solid angle of ~10−8 sr, this KBA-KB microscope can improve to 2.3 × 10−6 sr.
Fig. 4 Energy response simulation at 2.5 keV and 4.3 keV. (a) is the Pt monolayer energy response curve; (b) is spectral resolution of two multilayers; (c) is two multilayer reflectivity curve versus grazing angle at 2.5 keV and 4.3 keV; (d) and (e) are two intensity simulation results versus view field at two energy points, white cross dashes point out the energy center in the FOV.
3. Mechanical adjustment and characterization experiments in the lab
The adjustment errors impact the ultimate imaging performance. Before the process of system assembling, various mechanical adjustment error allowances need to be discussed firstly. The degradations of the spatial resolution by different assembling errors are calculated using MATLAB. We defined the error allowance as the error value which lets the resolution become twice as bad of the ideal condition. Table 2 shows the calculated results of major error allowances. The deflection angle of KBA double-mirrors is only 1.5°, the error of angle is limited to 0.1°. Object distance has only a slight effect on resolution. Meanwhile, two KBA channels need to have a coincident FOV and two image spots need to be separated by a distance of 14 mm, which means the vertical distance and relative angle of two sets of KBA configurations should be strictly restricted.
Table 2. Adjustment error allowance of mechanical assembling
Various unpredictable errors may be caused when adjusting each mirror individually by translation stages and then fixing them together by traditional mechanical methods. Meanwhile, the strength of the entire KBA structures should be strong enough when facing the shocks from target implosion. To solve these problems, we use a finely polished SiO2 prism as a support to hold four mirrors, as shown in Fig. 5. The height and the angle of four facets of the prism can restrict the position of each mirror effectively. The rigorous precision requirements for mirrors’ positioning have been attributed to the accuracies of prism manufacture. There are two channels carved on each working side of the prism with 3 mm depth to let the light which can reflected on mirror pass through, while the direct light is stopped. The mirrors are fixed on the working surfaces by ball screws and epoxy glue. The KBA structures are connected into a whole with high stability and easy to be adjusted by adjustment mounts.
The system was assembled in the lab and characterized with an X-ray source at 8 keV. All the mirrors were deposited with a periodic multilayer designed to provide a good monochromatic response for specific grazing angles. By this method, we can obtain an X-ray image directly with high intensity. A narrow angular bandwidth of approximately 0.05° can help us to adjust the mirrors into right grazing angles with high precision. In the object plane, a gold mesh with a period of 40 µm and a wire width of 5 µm was set as an imaging sample. An X-ray charged couple device (CCD) camera was set on the image plane. The pixel size of CCD is 6.45 µm and the size of working area is 9.0 mm × 6.7 mm.
Figure 6 shows the results of imaging with the X-ray source. The exposure time of CCD is 30 min with a gain of 30. We have obtained 1-D high resolution mesh images. From the intensity profile, the resolution is estimated with a width of 10% to 90% intensity and the results can achieve 5 µm. The mesh image is shown in Fig. 6(a) and the intensity profile is shown in Fig. 6(d) which corresponds to the yellow line in Fig. 6(a). A small hole with a diameter of 120 µm is used as the mark of FOV center. Figures 6(b) and 6(c) show the central-hole’s images through different channels. Due to the long time exposure, the thermodynamic variations may degrade the system resolution slightly.
Fig. 6 Results of gold mesh imaging experiment with X-ray source. (a) is intensity image of gold mesh with X-ray source, (b) and (c) are intensity images of central hole on the mesh by two KBA channels, (d) is estimated resolution from the intensity profile of selected column which illustrated with yellow line in (a).
4. Results of implosion experiments and discussion
4.1. Binocular target aiming method in the SG-III facility
At the terminal of the SG-III laser facility there is a huge spherical chamber 6.4 m in diameter and the internal environment keeps high vacuum. This implies all the diagnostic devices need to be equipped on a Diagnostic Instrument Manipulator (DIM) to move the designing position without a manual adjustment. For this process, a visible target aiming device is required because before the implosion experiments the chamber cannot be entered for adjusting the microscope position. Meanwhile, the target is a 2 mm capsule which positioned at the center of spherical chamber. In the former section, we have simulated and tested that the efficient FOV of KBA system is nearly 500 µm, which seriously restricts the position and direction of the entire microscope system in the chamber. Therefore, a precise and reliable method is needed for aiming at the target and adjusting the microscope.
The key point of target aiming is that the microscope has an ability of indicating the position of the object and the image reliably. An auxiliary device is designed to assist this process for the whole system as shown in Fig. 7. There are two steps for target aiming. The first is off-line calibration of the object and the image position. During the process of mechanical adjustment in the lab, the best position of mesh center hole which satisfies the coincident center of both channels can be precisely obtained by imaging experiments, and then use an analog needle to indicate the center. The width of analog needle is nearly 100 µm, two visible magnifying CCDs are used as monitors and record the position of the mesh when it is replaced by the needle tip. The accuracy of positioning can reach 10 µm.
The second step is on-line target aiming. A double light path visible aiming monitor assists the microscope to aim at the target. The monitor is composed of three lenses and the working distance is almost 350 mm. As shown in Fig. 8, two upside lenses are used to position the target precisely and they can work as a binocular system to position a point in 3-D space effectively. The optical axis of two lenses has an included angle of 9.6°. The object planes of these two lenses are intersected at the FOV center of the microscope. The included angle between monitor axis and optical axis of the microscope is 21.5°. A small offset of target aiming will cause different scales of deviation for each image of the two lenses. Based on these offset values the target position can be calculated quantitatively. The resolution of each lens is 24 µm. And the lower lens has a wide view with small magnification. The FOV of this lens can reach 170 mm for searching the target position roughly. Each lens has an electronically controlled shield to protect them from the damages of the target pieces sputtering during the implosion experiments. The position of analog needle tip needs to be recorded precisely by the detectors of three lenses before inserting the chamber. After rechecking, the analog needle is dismounted carefully and the object position has been determined uniquely. Depending on the accuracy and stability of the positioning of the analog elements, this process may be repeated many times after the microscope is mounted on the DIM for checking the aim at the target.
The indication of the image plane also plays an essential role for adjusting the detector. In the laboratory calibration experiments, a 632.8 nm cross-shape laser is used to indicate the center of two images. In addition, this laser is also connected together with the microscope body by a dismountable slider. The light path of system is symmetric about the KBA optical axis. This means the accuracy of imaging indication can maintain a high level of 0.5 mm even when the real image plane has a displacement of several tens of millimeters. However, the laser emitter blocks the out light from the mirrors and needs to be dismounted before inserting it. Using two lasers positioned off-axis and it is possible to point out the centers of the two image spots referred by crossing shape laser before the whole system is mounted on the DIM. After inserting it into the chamber, the detector on the image plane can receive the two spots from these lasers and then be corrected.
4.2. Results of implosion experiment
The implosion target is a gold mesh with the period of 42.3 µm and the wire width of 5 µm. The backlight producer is a metal plate mixed with Mo and Sc, which can emit 2.5 keV and 4.3 keV X-rays simultaneously when stimulated by lasers. The exposure time is 3ns and the power of the backlight lasers is 3 kJ. An X-ray image plate is used as a detector which can receive two static image spots. This result is shown in Fig. 9.
Fig. 9 (a) is static image of gold mesh target at 2.5keV and 4.3keV in implosion experiments; (b) and (c) are intensity profiles of the center column from (a).
The laser is focused on the back-lighter plate and forms a circular area with a diameter of 500 µm to illuminate the mesh target. From these results, the whole illuminating area is transmitted completely by the microscope. The pixel size of image plane is 25 µm. Figure 9(a) shows the static image from implosion experiment. The average resolution of the image spot for the whole FOV at 4.3 keV is 8.75 µm, as demonstrated by the intensity profile in Fig. 9(b). While, the alternative channel for 2.5 keV has a better resolution performance. As shown in Fig. 9(c), the measured result of the average resolution is 5.2 µm. From the intensity profile, the signal to noise ratio (SNR) of the 4.3 keV image is lower than the 2.5 keV image. The main reason for the conflicting resolution is the low frequency figure error of the spherical mirrors, in particular the deviation of designed curve radius. This would change the image distance of single channel and causes a ‘defocusing’ effect at the ideal image plane.
5. Conclusion
KB microscopy has been widely used for X-ray diagnostic implementation of target implosion experiments. We designed a new type of KBA-KB microscope which is applicable for planar target trajectory and RT instability growth of target recording. The system has been already assembled and characterized regarding its spatial resolution performance in the laboratory and applied on the Chinese SG-III laser facility for implosion experiments. This system has an excellent performance for high resolution imaging. The whole field resolution can achieve 5 µm with an efficient field of view beyond 500 µm. Meanwhile, the system has two channels and through the zoned coating technology on the mirror, it can realize two images at 2.5 keV and 4.3 keV simultaneously. This imaging diagnostic system has been verified for its optical and mechanical performance at the SG-III laser facility and will be used in several rounds of physical experiments in the future. Combined with a streak camera, this microscope can record high resolution images of target compression in RT instability growth experiments.
Funding
Key Projects of the National Science and Technology Pillar Program, (2013BAK14B02).
Acknowledgments
We would like to thank for the supporting from the staffs of SG-III laser facility.
References and links
1. J. Lindl, O. Landen, J. Edwards, E. Moses, and N. I. C. Team, “Review of the National Ignition Campaign 2009-2012,” Phys. Plasmas 21(2), 020501 (2014). [CrossRef]
2. V. A. Smalyuk, M. Barrios, J. A. Caggiano, D. T. Casey, C. J. Cerjan, D. S. Clark, M. J. Edwards, J. A. Frenje, M. Gatu-Johnson, V. Y. Glebov, G. Grim, S. W. Haan, B. A. Hammel, A. Hamza, D. E. Hoover, W. W. Hsing, O. Hurricane, J. D. Kilkenny, J. L. Kline, J. P. Knauer, J. Kroll, O. L. Landen, J. D. Lindl, T. Ma, J. M. McNaney, M. Mintz, A. Moore, A. Nikroo, T. Parham, J. L. Peterson, R. Petrasso, L. Pickworth, J. E. Pino, K. Raman, S. P. Regan, B. A. Remington, H. F. Robey, D. P. Rowley, D. B. Sayre, R. E. Tipton, S. V. Weber, K. Widmann, D. C. Wilson, and C. B. Yeamans, “Hydrodynamic instability growth and mix experiments at the National Ignition Facility,” Phys. Plasmas 21(5), 056301 (2014). [CrossRef]
3. S. G. Glendinning, S. V. Weber, P. Bell, L. B. DaSilva, S. N. Dixit, M. A. Henesian, D. R. Kania, J. D. Kilkenny, H. T. Powell, R. J. Wallace, P. J. Wegner, J. P. Knauer, and C. P. Verdon, “Laser-driven planar Rayleigh-Taylor instability experiments,” Phys. Rev. Lett. 69(8), 1201–1204 (1992). [CrossRef]
4. M. M. Marinak, S. G. Glendinning, R. J. Wallace, B. A. Remington, S. V. Weber, S. W. Haan, and G. W. Collins, “Nova indirect drive Rayleigh-Taylor experiments with beryllium,” Phys. Plasmas 9(8), 3567–3572 (2002). [CrossRef]
5. V. A. Smalyuk, S. X. Hu, J. D. Hager, J. A. Delettrez, D. D. Meyerhofer, T. C. Sangster, and D. Shvarts, “Rayleigh-Taylor growth measurements in the acceleration phase of spherical implosions on OMEGA,” Phys. Rev. Lett. 103(10), 105001 (2009). [CrossRef]
6. D. A. Martinez, V. A. Smalyuk, J. O. Kane, A. Casner, S. Liberatore, and L. P. Masse, “Evidence for a bubble-competition regime in indirectly driven ablative Rayleigh-Taylor instability experiments on the NIF,” Phys. Rev. Lett. 114(21), 215004 (2015). [CrossRef]
7. F. J. Marshall, J. A. Oertel, and P. J. Walsh, “Framed, 16-image, Kirkpatrick–Baez microscope for laser-plasma X-ray emission,” Rev. Sci. Instrum. 75(10), 4045–4047 (2004). [CrossRef]
8. P. Kirkpatrick and A. V. Baez, “Formation of optical images by x-rays,” J. Opt. Soc. Am. 69(8), 1201–1204 (1948).
9. T. Pardini, T. J. McCarville, C. C. Walton, T. A. Decker, J. K. Vogel, P. B. Mirkarimi, J. B. Alameda, R. M. Hill, L. A. Pickworth, V. A. Smalyuk, J. M. Ayers, P. M. Bell, D. K. Bradley, J. D. Kilkenny, and M. J. Pivovaroff, “Optical and multilayer design for the first Kirkpatrick-Baez optics for X-ray diagnostic at NIF,” Proc. SPIE 8850, 88500E (2013). [CrossRef]
10. O. V. Gotchev, L. J. Hayes, P. A. Jaanimagi, J. P. Knauer, F. J. Marshall, and D. D. Meyerhofer, “Large-grazing angle, multi-image Kirkpatrick–Baez microscope as the front end to a high-resolution streak camera for OMEGA,” Rev. Sci. Instrum. 74(12), 5065–5069 (2003). [CrossRef]
11. J. A. Koch, O. L. Landen, T. W. Barbee, P. Celliers, L. B. Da Silva, S. G. Glendinning, B. A. Hammel, D. H. Kalantar, C. Brown, J. Seely, G. R. Bennett, and W. Hsing, “High-energy X-ray microscopy techniques for laser-fusion plasma research at the National Ignition Facility,” Appl. Opt. 37(10), 1784–1795 (1998). [CrossRef]
12. L. Pickworth, D. Bradley, T. Pardini, V. Smalyuk, N. Izumi, M. Pivovaroff, J. Vogel, C. Walton, P. Mirkarimi, P. Bell, T. Decker, T. McCarville, M. Ayers, and J. Kilkenny, “A Kirkpatrick-Baez microscope for core implosion imaging at NIF,” APS Meeting Abstracts 1, 5013 (2013).
13. L. A. Pickworth, J. Ayers, P. Bell, N. F. Brejnholt, J. G. Buscho, D. Bradley, T. Decker, S. Hau-Riege, J. Kilkenny, T. McCarville, T. Pardini, J. Vogel, and C. Walton, “The National Ignition Facility modular Kirkpatrick-Baez microscope,” Rev. Sci. Instrum. 87(11), 11E316 (2016). [CrossRef]
14. G. R. Bennett, “Advanced one-dimensional X-ray microscope for the Omega Laser Facility,” Rev. Sci. Instrum. 70(1), 608–612 (1999). [CrossRef]
15. Y. Zhang, B. Mu, S. Yi, X. Wang, Z. Wang, Z. Cao, S. Liu, and Y. Ding, “Design of mixed KB-KBA microscope for multi-keV dynamic imaging diagnostics,” High Power Laser Particle Beams 26(5), 052002 (2012). [CrossRef]
16. S. Yi, B. Mu, X. Wang, Z. Zhang, J. Zhu, Z. Wang, P. He, Z. Cao, J. Dong, S. Liu, and Y. Ding, “Large-field high-resolution Kirkpatrick–Baez amélioré-Kirkpatrick–Baez mixed microscope for multi-keV time-resolved X-ray imaging diagnostics of laser plasma,” Opt. Eng. 53(5), 053114 (2014). [CrossRef]
17. D. B. Sinars, K. J. Peterson, S. A. Slutz, M. C. Herrmann, E. P. Yu, M. E. Cuneo, I. C. Smith, G. R. Bennett, B. W. Atherton, J. L. Porter, and D. F. Wenger, “Observation of instability growth in a Copper-Pinch target using two-color monochromatic X-ray backlighting,” IEEE Trans. Plasma Sci. 39(11), 2408–2409 (2011). [CrossRef]
