A one dimensional, absolutely calibrated pinhole camera system was installed on the DIII-D tokamak to measure edge Lyman-alpha (Ly-α) emission from hydrogen isotopes, which can be used to infer neutral density and ionization rate profiles. The system is composed of two cameras, each providing a toroidal fan of 20 lines of sight, viewing the plasma edge on the inboard and outboard side of DIII-D. The cameras’ views lie in a horizontal plane 77 cm below the midplane. At its tangency radius, each channel provides a radial resolution of ∼2 cm full width at half maximum (FWHM) with a total coverage of 22 cm. Each camera consists of a rectangular pinhole, Ly-α reflective mirror, narrow-band Ly-α transmission filter, and a 20 channel AXUV photodetector. The combined mirror and transmission filter have a FWHM of 5 nm, centered near the Ly-α wavelength of 121.6 nm and is capable of rejecting significant, parasitic carbon-III (C-III) emission from intrinsic plasma impurities. To provide a high spatial resolution measurement in a compact footprint, the camera utilizes advanced engineering and manufacturing techniques including 3D printing, high stability mirror mounts, and a novel alignment procedure. Absolutely calibrated, spatially resolved Ly-α brightness measurements utilize a bright, isolated line with low parasitic surface reflections and enable quantitative comparison to modeling to study divertor neutral leakage, main chamber fueling, and radial particle transport.
At the boundary of magnetically confined plasmas, neutral particles ionize as they move toward the center of the plasma, acting as a source of ionized particles for the plasma density. The neutral source of ionized particles, typically neutral hydrogen or deuterium, is a key boundary condition for the core performance of current devices and extrapolation to future fusion energy devices.1 In particular, for tokamaks, a high performance toroidal magnetic confinement fusion device, neutrals are expected to play a key role in a range of physical processes: setting global plasma pressure, the height of the high confinement mode (H-mode) pedestal, divertor detachment, compatibility with advanced divertors, and divertor closure.2–5
Measurements of neutral hydrogen in fusion devices are generally made through either active spectroscopy, such as laser induced fluorescence (LIF), or passive measurement of hydrogen brightness. LIF techniques allow a direct spatial localization of the measurement of neutral density using stimulated emission from laser systems.6–9 However, LIF requires high power laser systems and currently have limited applicability in fusion plasmas.6–8 In contrast to LIF, passive hydrogenic emission measurements, particularly in the visible range of wavelengths, rely solely on collection optics and, therefore, are more widely adopted. Passive spectroscopy measurements are line integrated and rely on the filtering of one or more lines of hydrogen emission. A measurement of the line brightness is required to calculate the local neutral density, and the brightness must then be inverted to determine the local emissivity and combined with measurements of the local electron temperature and density to evaluate standard transition rate coefficients.10 In fusion plasmas, passive emission studies of neutral particles have utilized Balmer (ni>2 → n = 2) and Lyman (ni>1 → n = 1) series emissivity measurements in conjunction with density and temperature measurements, for instance, from Thomson scattering.11 Since Balmer-alpha (B-α) (λ = 656.3 nm) is routinely measured in tokamaks, it has been adopted for neutral studies on the DIII-D, Alcator C-Mod, NSTX-U, LHD, and TEXT tokamaks.12–17 Lyman-alpha (Ly-α) (λ = 121.6 nm) diagnostics for neutral studies have been implemented on C-Mod18,19 and TCV20 and used for studies on the role of neutrals in the formation of the edge profiles, through particle and heat balance.21–25
B-α measurements are more frequently performed, but Ly-α measurements offer key advantages for neutral studies. One challenge in using B-α measurements to measure hydrogen neutral density is the increased contribution to the line brightness of excited hydrogen neutrals from molecular reactions. Recent reports on NSTX-U have utilized B-α to estimate neutral density; however, due to the large molecular interactions, advanced modeling is required to determine the relative molecular and atomic contributions to the line brightness, which alters the inferred neutral densities.14,26 On DIII-D, molecules are suspected to be a significant fraction of main chamber recycling neutrals and can lead to an overestimation of ionization and density.27–29 In contrast, EDGE2D-EIRENE modeling on DIII-D suggests that molecular contributions to emission are negligible for measurements of the Ly-α line,30 suggesting that atomic neutral densities can be directly evaluated. Unlike B-α, Ly-α does not readily reflect off of surfaces. Also, the intensity of Ly-α is at least an order of magnitude larger for typical plasma parameters found in the tokamak edge.10 These advantages come with additional engineering challenges. Optical filtering in the ultraviolet spectrum is more challenging than the visible spectrum due to a lack of high transmission narrowband filters. In particular, rejection of a nearby carbon-III (C-III) line (λ = 117 nm) becomes a concern. Since Ly-α is readily absorbed in air and most materials, optical components and detectors need to be placed in vacuum. Placing a diagnostic inside the primary vacuum of a tokamak further introduces engineering challenges in terms of limited space availability, limited access to the vessel, restrictions on materials, and resilience to heat loads from plasma operation and cleaning procedures.
In this paper, we detail a new, compact, absolutely calibrated diagnostic, LLAMA (Lyman Alpha Measurement Apparatus) installed on the DIII-D tokamak to measure the Ly-α hydrogen neutral line brightness. The LLAMA diagnostic advances both the engineering and measurement solutions compared to previous Ly-α diagnostics. The new Ly-α camera on DIII-D introduces a second filtering component, a Ly-α notch reflective mirror, which narrows the bandpass and significantly reduces the neighboring C-III line intensity. A pinhole that aligns with flux surfaces allows maximization of the resolution in magnetic flux space. The camera utilizes advanced engineering, including vacuum compatible three-dimensional (3D) printed Inconel parts, thermally and shock resistant adjustable mirror mounts, water-cooling, and a compact footprint providing two views. The custom amplifier system mounts directly to the flange and provides a high gain adjustable amplification system in a small footprint with pseudo-differential outputs. Finally, a novel alignment procedure relying on a coordinate measuring machine (CMM) allows alignment on the bench, minimizing the in-vessel time required for installation and alignment.
II. DIAGNOSTIC DESIGN AND OVERVIEW OF IN VACUUM COMPONENTS
LLAMA consists of two one-dimensional pinhole cameras placed in the primary vacuum of the DIII-D tokamak, viewing the high field side (HFS) and low field side (LFS) of the plasma. The optical components of each camera are highlighted in Fig. 1. Plasma emission enters the camera through a 2 × 8 mm2 rectangular pinhole cut by a waterjet and held in a rotatable mount for optimal alignment with plasma flux surface geometry. Next, the incoming spectrum is narrowed by a Ly-α notch reflective filtering mirror31 mounted in a customized THORLABS mount.32 The reflected light is then directed to a light tight box, the detector box, where it passes through a transmission filter toward33 an AXUV photo-diode detector. The detector consists of 20 channels providing 20 lines of sight.34 Photocurrent is carried out of vacuum through Kapton insulated cable to a 25 D-sub connection. Cable lengths are minimized to reduce electrical pickup. The two cameras are separated by a 3D printed Inconel septum,35 which can be seen in Fig. 1. The internal components are protected from tokamak bakes and glow discharges by a linearly actuated shutter. Enclosing the system is a light tight tube, which can be seen in Fig. 5(b). Typical solid angles are 5 × 10−4 for the HFS and 2 × 10−3 for the LFS. The etendue for the LFS and HFS is ∼5 × 10−9 m2 and 1 × 10−9 m2.
The filtering of the Ly-α line is achieved through a combination of notch reflective and transmission thin-film filters both produced by Acton Optics. The reflective filtering mirror, thin-film transmission filter, and combined transmission curves can be seen in Fig. 2. Previous Ly-α cameras isolated the Ly-α line utilizing only the narrow-band transmission filter.18,20 However, due to the carbon wall on DIII-D, there is significant intensity from a neighboring C-III line at 117.5 nm. The intensity of the C-III line, as measured by the divertor spectrometer, is comparable to that of Ly-α on many shots. The addition of the reflective filter provides a greater reduction of the C-III signal relative to Ly-α by narrowing the bandpass, as highlighted in Fig. 2(c), showing the normalized transmission curves. The combined mirror and transmission filter system is capable of reducing the carbon intensity relative to Ly-α by a factor of 10 with a full-width half maximum of 5 nm centered around 123 nm. In addition, the inclusion of the mirror facilitates a more compact optical design. The angle of incidence on the mirror varies from 35° to 45° for the HFS and 45° to 55° for the LFS. This introduces some dependence of the Ly-α transmission on the incident angle of light. However, the transmission dependence is accounted for in the absolute calibration procedure.36 The bandpass of the system allows measurement of both the hydrogen and deuterium Ly-α line. DIII-D primarily operates with deuterium but has dedicated hydrogen campaigns; the LLAMA is capable of providing measurements for both hydrogen isotopes.
The reflective filter is held in a THORLABS mirror mount shown in Fig. 1(d) selected and customized to survive the challenging in-vessel environment of a tokamak while facilitating alignment on the optical bench. For the stability of the views, the mirror mount requires stability to 100 µrad through thermal and vibrational shock from the tokamak environment. A THORLABS Polaris K05 manual mirror mount was selected as the base design due to its microradian level stability over 15 °C thermal and two times gravity shock cycling.37 The mount was further customized with non-magnetic parts to avoid displacement from the tokamak’s magnetic field and error fields. Components were selected to comply with material restrictions inside the primary vacuum of DIII-D, which included removing lubricants from all screws by using Nitronic-60 screws and replacing ball bearing contacts with ceramic pieces. This was all achieved in a compact footprint while still providing three-axes manual manipulation for easy alignment on the bench.
The primary engineering challenges constraining the mechanical design were the geometry of the port and thermal loads during the 350 °C bakes performed on DIII-D before operation.38 The camera body is designed around a 3D printed Inconel septum, highlighted in Figs. 1(b) and 1(c), which provides support to the optical components, thermal stability with an embedded 6 mm diameter cooling loop, and a channel for the shutter actuation. Cooling tubes from the 3D printed Inconel septum pass through the stainless steel vacuum flange and are welded in place providing structural support and removing all water connections from inside the vessel. The 3D printed Inconel septum is deployed in-vessel and pressurized with water up to 850 kPa. Only during bakes, thermal loads are large enough to require water cooling, with 15 °C thermal excursion measured in the diagnostic by a thermocouple. However, the flow is left on during plasma to ensure thermal stability of the detector response. Embedding the cooling tubes in the mounting plate minimized the size of in-vessel components. The shutter, which protects the optics during bakes and plasma glow discharges, is also made of 3D printed Inconel to conform to the outer tube and has linear pneumatic actuation. The cameras are enclosed in an outer tube nearly matching the diameter of the port and mounted to the Inconel septum. The flange, which supports the entire diagnostic, is highlighted in Fig. 3. It consists of two 25 pin D-sub connectors, water cooling feedthroughs, and a pass-through for the shutter actuator. The compact size and flange-supported design allow the LLAMA to easily be exchanged with a duplicate from outside the vessel. This allows for regular calibration and maintenance to be performed away from the tokamak.
III. LLAMA AMPLIFIER SYSTEM
Due to previous experience with AXUV detectors on DIII-D, noise was a primary concern in the design of the wiring and electronics for the LLAMA. The HFS and LFS detectors and signals are electrically isolated against the DIII-D tokamak (DIII-D) vacuum vessel potential and each other. The detectors are isolated from the machine by a custom-made Macor mount and insulating film, which can be seen in Fig. 1(d). This setup provides a 5 kV standoff. Wire paths are kept short to minimize common mode pickup, which was a concern since AXUV photodiode arrays consist of 20 photodiode detectors connected to one common ground. To further minimize pickup, the signals are immediately amplified by a variable gain 107–108 V/A custom-built amplification board mounted directly to the air side of the D-sub connectors on the flange, as shown in Fig. 3. The custom amplifier boards shown in green mount directly to the vacuum feedthroughs and cover the air side of the DB-25 connectors in Fig. 3. There are separate amplifier boards for the LFS and HFS systems that are enclosed by a metal housing for shielding (shown for the HFS board). The amplifier boards have smaller, removable circuit boards, which contain the actual electronic elements of the amplifier and output circuits. The amplifier boards are mounted to a water cooled base plate whose cooling fluent copper leads exit to the right. Thermally conductive 5584 silicone interface pads39 enable good thermal exchange between the amplifier boards, boxes, and base plate and fulfill the safety requirement of a 5 kV stand-off for DIII-D.
The AXUV photodiode detector arrays with 20 channels require a transimpedance amplifier system to measure the generated photocurrents. The photocurrents were modeled from the LLAMA responsivity and estimated Ly-α brightness in typical DIII-D scenarios to be on the order of 1 nA–1000 nA. The expected wide range of photocurrents motivated variable gains of 107 and 108 V A−1 to create measurable output voltages.
A schematic drawing of the implemented transimpedance amplifier is presented in Fig. 4(a). Typically, a reversed bias voltage (Vbias) of 3 V is applied to the photodiode detector. The generated photocurrent is amplified and converted by an operational amplifier and a parallel resistor (107 or 108 Ω), which determines the gain. The transimpedance amplifiers of the photodiode detector channels are grouped on amplifier stage boards. One amplifier stage board accommodates four photodiode detector channels, using a four channel operational amplifier in combination with the corresponding circuit resistors and capacitors. Each exchangeable amplifier stage board is equipped with one of the design gains; therefore, changing the amplifier stage board alters the gains for four photodiode detector channels at a time. Each amplifier stage board can be easily replaced if a single component is damaged or fails due to exposure to high energy neutrons and ionizing radiation from the tokamak. Five amplifier stage boards are shown on the LFS amplifier board presented in Fig. 3 next to the DB-25 feedthrough.
The output of the amplifier boards is converted to a pseudo-differential signal that allows for a transmission less susceptible to noise. The generated differential signals are then passed through a DB-50 connector, which transfers them in pairs of twisted and shielded wires to the driver stage (second stage) located roughly 10 m of the wire away from the LLAMA flange. The second stage supports the signal transmission along 55 m of wire to the commercially available ACQ424ELF digitizer40 outside of the DIII-D machine hall.
Data are digitized and stored at a rate of 500 kHz. Although bright transient events such as edge localized modes (ELMs) can be analyzed at higher frequencies, data analysis for steady state behavior of the Ly-α brightness is presently limited to about 1 kHz by noise from induced currents due to high frequency oscillations of the tokamak’s poloidal magnetic field coil’s power supplies. The signal amplitude to noise amplitude ratio at these frequencies ranges from 5% to 20%, depending on the configuration of power supplies and coils, plasma conditions, and the amplifier gain. The inferred neutral density from the brightness measurements is also dependent on the frequency of electron temperature and density measurements performed. For example, for measurements of electron temperature and density by Thomson scattering, the neutral density is determined at the interleaved nominal laser pulse repetition rate of 250 Hz.41
To be able to resolve fast transient events in the plasma edge, such as edge localized modes (ELMs),42,43 which occur on a sub-ms timescale, a constant gain over a wide range of frequencies is desired. Figure 4(b) presents the modeled frequency and phase dependence of the circuits. As desired, both gains are constant up to almost 103 Hz providing access to inter-ELM dynamics.13,44,45 The lower, 107, gain exhibits a constant gain up to roughly 104 Hz, allowing Ly-α brightness measurements of low frequency plasma fluctuations in the plasma edge like the fundamental frequency of the edge harmonic oscillation (EHO).46,47
IV. LLAMA VIEWS AND ALIGNMENT
Due to the geometry of the port, it was not possible to align the diagnostic in-vessel. Therefore, a novel alignment scheme was developed in which the majority of alignment can be performed on an optical bench. The alignment method relies on the use of a CMM, fiducials placed on the front enclosure of the diagnostic, and an alignment mount recreating the port geometry on the optical bench. The views are determined on the bench with the diagnostic installed in the alignment mount. Then, the views are measured with the CMM along with the fiducials. After the installation in DIII-D, the fiducials are measured with a CMM in the tokamak geometry. By overlaying the fiducial measurements from the bench onto the tokamak fiducial measurements, the orientation of the views is determined in the tokamak coordinate system. This alignment procedure reduces the scarce time for in-vessel alignment available at DIII-D and allows the system to be operated immediately upon installation while still providing accurate positioning of the views.
An overview of the alignment procedure is shown in Fig. 5. The diagnostic is installed in the alignment mount, which is designed to recreate the port geometry on DIII-D. Once installed in the alignment mount, from the perspective of the diagnostic, the plane of the bench is parallel to a horizontal plane inside DIII-D. A He–Ne laser is then used to create a fan of rays to which the internal components are aligned. The center channel is illuminated by aligning the laser parallel to the optical surface and perpendicular to the pinhole, as seen in Fig. 5(a) The outer channels, seen in Fig. 5(a), can then be determined by maximizing the signal on the corresponding detector channel. The He–Ne rays, detector, and pinhole location are then measured using the CMM. All other views can then be interpolated from these quantities since the dimensions of the detector are known to have 0.05 mm accuracy. The outer tube, detector box lid, and transmission filter are removed during the He–Ne laser alignment, to allow significant light transmission to the detector.
The views determined on the benchtop are modeled in the tokamak coordinate system using a CMM and the fiducials marked on the front cap of the diagnostic. The three fiducials shown in Fig. 5(b) are measured on the bench along with the three He–Ne laser lines and fit using an error reducing method.48 After the diagnostic is installed, the CMM is used in the tokamak to re-measure the fiducials. As seen in Fig. 5(c), by overlaying the two fiducial measurements, the views are reconstructed in the tokamak coordinate system.
The resulting views in the toroidal and poloidal cross-section can be seen in Figs. 6 and 7. The views are centered near the separatrix below the midplane for a variety of common plasma shapes on DIII-D. The central channel of each view has its tangency radius at 121.0 cm (HFS) and 192.9 cm (LFS) with an error of 1 mm in the radial direction resulting from the CMM measurement. The views are tilted from the horizontal by 1 ± 1° with HFS and LFS views pointing slightly downwards. Since there is no dependent coordinate of the 3D points measured by the CMM, the uncertainty of the views from the alignment procedure was determined using a Monte Carlo simulation. The individual 3D coordinates measured by the CMM were varied by the experimentally determined uncertainty of a point, 0.5 mm. The fiducial alignment procedure was then performed with the varied coordinates. The resulting uncertainty in the position of the tangency radii and angles from the horizontal is reported to contain two standard deviations of the simulations.
The projection of individual channels onto their tangency radii plane can be seen in Figs. 7(b) and 7(c). The projected image of each channel differs in size due to the varying distance to the tangency radii plane. Tangency radii for all channels are shown in magenta; the projected channels’ radii are highlighted with a white border. For the LFS, the full-width half maximum for each channel measurement is 2.6 ×8.4 cm2 for the radial and vertical directions, respectively. For the HFS, the full-width half maximum is 2.1 × 9.2 cm2. The finite size of the pinhole and detector results in a spot size, which is much larger than the uncertainty in the positioning of the center of each channel determined from the CMM Monte Carlo simulation. The spatial sensitivity distribution is emphasized by the higher density color near the center of the channel projection. Due to the finite size pinhole, there is overlap between channels, about half of the full width at half maximum (FWHM) overlaps with the neighboring channel.
The alignment and size of the pinhole were chosen to maximize signal while maintaining the lines of sight along a flux surface. The Ly-α emissivity depends on electron temperature, electron density, and neutral density. Electron temperature and density are flux functions; therefore, for small gradients in neutral density, the emissivity is expected to be a flux function. While modeling suggests poloidal asymmetries in neutral density, their scale is larger than a LLAMA channel view.2 Therefore, by aligning the detector projection with the flux surface, it is possible to minimize averaging of emissivity across flux surfaces improving the resolution with a relatively large detector projection. The effect of rotation of the pinhole can be seen by comparing the HFS and LFS detector projections in Figs. 7(b) and 7(c). On the LFS, the pinhole is rotated; the projection of the detector more closely matches the flux surface geometry, reducing the averaging across flux surfaces. By rotating the LFS pinhole, both the HFS and LFS can resolve features on a 50 mm radial scale. The initial estimate from the ionization cross-section evaluated at the pedestal top suggests neutral mean free paths of ∼10–100 mm, while previous models suggest neutral decay lengths comparable to the electron density pedestal width of 20 mm–50 mm.49 These scale lengths should be resolvable by the LLAMA system. Furthermore, shifts smaller than 50 mm in the emissivity should be measurable due to the overlapping channel views.
The LLAMA diagnostic provides absolutely calibrated brightness measurements through a benchtop calibration, which is regularly performed before and after campaigns to monitor the stability of the diagnostics’s spectral response. The calibration procedure will be described briefly here and in detail in a dedicated article.36 Due to its compact size, the calibration is performed on a dedicated ex-vessel vacuum enclosure, which reaches pressure below 2 × 10−5 Pa to avoid the loss of Ly-α transmission. The procedure uses an absolutely calibrated Ly-α electrodeless gas discharge powered by a radio-frequency source, supplied by Resonance LTD,50 to illuminate the pinhole of the diagnostic with Ly-α light. The calibration source spectra are characterized by a vacuum ultraviolet (VUV) spectrometer,51 and its intensity is measured by a NIST-absolutely calibrated photodiode52 to determine the intensity of Ly-α illuminating the diagnostic. Initial calibration of the LLAMA shows brightness measurements on the tokamak of 1 × 1021 Ph sr−1 m−2 s−1 and absolutely calibrated profiles can be seen in Figs. 9 and 10.
V. INITIAL Ly-α BRIGHTNESS MEASUREMENTS IN DIII-D
The diagnostic regularly observes variations due to L-mode to H-mode transitions, ELMs, gas puffs, shape changes, and density changes. An example of observed brightness from channel four, highlighted in yellow in Fig. 7(b), on the LFS camera from a 2019 DIII-D experiment can be seen in Fig. 8. As observed on other tokamaks, when the plasma transitions from the low to high confinement mode, there is a drop in emissivity in the near scrape off layer.25 ELMs are also observed in the Ly-α emissivity due to the increase of particles outside of the core due to the collapse of the pressure pedestal.53,54 As a passive diagnostic, LLAMA can obtain brightness data on most shots, except for scenarios that could damage the system’s optics, such as counter neutral beam injection, which populate the scrape off layer with suprathermal beam ions.
The diagnostic was recently commissioned, which involved a rigid shift of the plasma boundary and increasing outboard gas puff. The commissioning discharges were performed in a lower single null configuration, which is shown in solid black in Fig. 7(a), with a toroidal field of 2 T, 1.8 MA plasma current, 6 MW injected power, and a line averaged density of 7 × 1019 m−3. Brightness profiles from a rigid outward shift of the plasma boundary by 5 cm are shown in Fig. 9. The resulting movement of the separatrix in real space is indicated by the vertical line in Fig. 9(a). LLAMA’s measured brightness profile tracks the overall movement across channels in absolute space but maintains a similar profile when plotted in normalized flux space. The tracking with plasma movement demonstrates that the measured profiles are not due to diagnostic or geometric effects. The profiles in Fig. 9 are from ELM-synced analysis examining the last 65%–98% of the ELM cycle. Figure 10 demonstrates the response of the brightness profiles to an increasing outboard gas puff across three shots: a base case with no additional gas (180 910), 40 Torr-L/s gas puff (180 914), and 80 Torr-L/s (180 916). The displayed data examine time windows with steady state behavior and are ELM-synchronized, including only the last 65%–98% of the ELM cycle. As the gas puff rate increases, increasing the neutrals in the scrape of layer, the brightness profile increases as anticipated. The brightness gradient also increases as the gas puff rate increases. The resulting neutral density and ionization rate evaluated at the peak of the LFS brightness can be seen in Table I.
|.||Additional gas rate .||Neutral density .||Ionization rate .|
|Shot .||(Torr L s−1) .||(1014 m−3) .||(1021 m−3 s−1) .|
|.||Additional gas rate .||Neutral density .||Ionization rate .|
|Shot .||(Torr L s−1) .||(1014 m−3) .||(1021 m−3 s−1) .|
Overall, the profile trends for both views agree with expectation and track with plasma shifts. The LFS profile shows a brightness, which peaks around the separatrix, similar to observations on C-Mod.18 The signal on the HFS is an order of magnitude larger than the LFS. Similar inboard-outboard asymmetries were observed for Ly-α measurements on C-Mod,18 B-α measurements on DIII-D,27 and neutral density temperature on Aditya-U.55 Poloidal asymmetries have also been suggested by modeling on DIII-D using the SOLPS code2,56 and several other machines.57 The HFS measurements also show significant brightness from well inside the separatrix, where the temperature would not suggest a significant neutral population for emitting Ly-α radiation. The source of this radiation is under investigation. The core radiation peak could be the result of neutrals penetrating deeper than expected through charge exchange interactions in a local colder region of the plasma. Another possibility is a source of radiation either from the core or where the HFS views terminate on the LFS wall, which could result in signal miss-attributed to the tangency radius in the HFS brightness.
VI. CONCLUSION AND OUTLOOK
A one-dimensional Ly-α camera providing radial profiles of the HFS and LFS Ly-α brightness profiles has been developed and implemented on DIII-D. A measurement of Ly-α is achieved with VUV reflective and transmission filters, which can be utilized to determine the hydrogenic neutral density. The optical filtering is able to significantly reject the neighboring C-III line emission, isolating the Ly-α line. The optical system is successfully deployed in the primary vacuum of DIII-D by utilizing advanced manufacturing techniques such as embedding a cooling loop in 3D printed Inconel septum plate and non-magnetic mirror mounts. Utilizing a custom amplifier system, a rotatable pinhole mount, and a novel alignment scheme, the system is able to measure the Ly-α brightness on spatial and temporal scales relevant to pedestal physics. The results from a recent commissioning experiment demonstrate anticipated brightness magnitudes and trends with gas puffing showing that the diagnostic is ready for physics studies of main chamber neutral behavior.
As a passive diagnostic, LLAMA provides brightness profiles on most plasma discharges and is well suited for quantitative studies on neutral transport with relevance to pedestal structure, divertor studies, and model validation. The two views provided by the system offer a unique opportunity for investigating in–out asymmetries of neutral particle behavior. Further improvements to the diagnostic are expected including improvement of the noise characteristics to fully exploit the time resolution for brightness measurements. A second system is also under consideration for installation above the midplane, which would allow the investigation of up–down as well as inboard–outboard asymmetries.
A. M. Rosenthal would like to thank Theodore Golfinopoulos and Matthew Reinke for the help and mentorship on the project.
This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Fusion Energy Sciences, using the DIII-D National Fusion Facility, a DOE Office of Science user facility, under Award Nos. DE-AC02-09CH11466, DE-AC05-00OR22725, DE-FC02-04ER54698, and DE-SC0014264.
This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
The data that support the findings of this study are openly available in DATAVERSE at https://doi.org/10.7910/DVN/TWNBPM (Reference No. PSFC/JA-20-17).