2 Design

The design for the COAX targets is shown in Figure 2. Figure 2(a) shows the foam package design with the doped aerogel foam region signified in gray hash. The tolerance of the doped foam is $+0.0$ and $-5.0~\unicode[STIX]{x03BC}\text{m}$ to ensure the foam is smaller than the specification. Similarly, the inner dimensions of the undoped foam are between $+0.0$ and $-5.0~\unicode[STIX]{x03BC}\text{m}$ , smaller as well. The tolerance of the foams is the most important specification, set so the foams are not larger than the specified size to ensure the foams can be assembled without damage and the foam stick together without glue. The tight tolerance keeps the gap between the two foams small enough to ensure the radiation is transported through the foams and not along the gap. The foam density can be varied between ${\sim}50$ and $80~\text{mg}/\text{mL}$ with ${\sim}$ 1 : 5 ratio of dopant material to aerogel. Machinability of the foam sets the lower density limit while laser energy sets the upper density limit. Foam with both Ti and Sc dopants have been manufactured. The surrounding foam is undoped aerogel foam usually with a slightly higher density than the doped foam between $50$ and $80~\text{mg}/\text{mL}$ to maintain similar electron densities and temperatures between the two foam regions. Figure 2(b) shows the drawing of the assembled target with the hohlraum and the foam package. To increase robustness, the foam package is mounted in a beryllium tube that is attached to the hohlraum. Figure 2(c) shows the final target assembly including the stalk. Another feature is a collar on the outside of the hohlraum to support the beryllium tube for easy and reproducible assembly of the target. The hohlraums are manufactured by General Atomics^[9].

Figure 2. (a) Foam target designs with aperture, (b) hohlraum foam assembly, (c) final target assembly with stalk mounting.

On a historical note, the target design has evolved. The original design of these targets deposited near solid density Ti on one half of the foam that had been split into two parts^{[Reference Taccetti, Keiter, Lanier, Mussack, Belle and Magelssen10]}. The foams were reunited and mounted. However, the dense titanium absorbs more x-rays from the hohlraum than the surrounding foam leading to a pressure imbalance not acceptable for the experiments. The second iteration used a rectangular block of doped foam with a slot in the undoped foam to simplify machining (Figure 3). While this configuration succeeded in providing absorption spectra, the orthogonal views of the two instruments measuring spectra and subsonic foam motion are not the same. In addition, it is more difficult to keep the slotted foam in the undoped foam while preventing glue from wicking into the experimental region of the foam package. The current design is the third iteration. The beryllium tube has been added to make the assembly more robust. The initial designs mounted the foam directly to the aperture and on occasion, the foams would get damaged, separate from the hohlraum or glue would wick into the foam rendering the target unacceptable for the experiments. The beryllium sleeve protects the foam and holds it in place and a larger amount of glue can be used between the hohlraum and beryllium sleeve away from the experimental region. The beryllium sleeve was never used for the slotted foam design, but would have prevented glue from spreading into the experimental line of sight, thus making the target more robust.

Figure 3. Schematic of foam with slot insert.

3 Ti- and Sc-doped foams

Creating the doped aerogel foams with high- $z$ particles uses previously developed technology^{[Reference Hamilton, Honnell, Patterson, Schmidt and Obrey11]}. All chemicals are off the shelf with no further purification required. The solution starts with 12 g of methyl silicate 51 precursor from BOC sciences^[12] (50% solids by mass) yielding 6 g (0.1 mol). This is diluted to 100 mL with methanol (Fisher Scientific^[13]). 1.6 g of titania (TiO $_{2}$ ) nanoparticle powder from Sigma-Aldrich^[14] (22 nm nominal particle size) is added containing 0.02 mol TiO $_{2}$ . Finally, 3 mL of 5 mol/L aqueous ammonium hydroxide (Fisher Scientific) is added to catalyze gelation (typical gel times were 20–30 min). The mixture is placed in a bath sonicator and mixed until viscosity begins to increase enough to prevent TiO $_{2}$ from sedimenting. The mixture is quickly sealed in stainless steel molds (before reaching the gel point) which are aged 24 h at room temperature before supercritical drying.

Supercritical drying is achieved by placing the molds in a methanol filled autoclave. The system is pressurized to 140 atm and heated to $300\,^{\circ }\text{C}$ over 4 h, maintaining constant pressure by slowly releasing methanol. The vessel is then held at constant temperature and pressure (140 atm, $300\,^{\circ }\text{C}$ ) for 1 h. Pressure is slowly released over 2 h while maintaining temperature. Finally, the reactor is placed under vacuum and allowed to cool to room temperature overnight. Producing the desired foam density is challenging due to volumetric shrinkage that can increase the density by 15%–20%.

Sc-doped SiO $_{2}$ is prepared in exactly the same way except that the amount of methanol used is 120 mL and 2.76 g of Sc $_{2}$ O $_{3}$ powder ( $7~\unicode[STIX]{x03BC}\text{m}$ nominal particle size) from Goodfellow inc.^[15]. 0.02 mol is added to the 12 g of methyl silicate. Ideally, the Sc particles would be much smaller ${<}0.5~\unicode[STIX]{x03BC}\text{m}$ . Efforts to reduce their size via mortar and pestle did not work. Currently, an effort is underway to use Sc nanoparticles to form more uniform distribution of Sc in the foam. This will allow the Sc dopant to ionize and to equilibrate with the surrounding foam faster.

Once the foams are completed, the foams are characterized using both gravimetric measurements and x-ray transmission. First, the difference in mass between the pure SiO $_{2}$ bulk sample (30 mm long $\times$ 2 mm diameter cylinders) and the Ti-doped SiO $_{2}$ bulk sample is used to infer the relative concentration of Ti in the doped sample. The machined Ti-doped foams are not massive enough to measure the density gravimetrically. After the individual samples are machined the density is characterized again using the density characterization station (DCS)^{[Reference Lanier, Hamilton and Taccetti16]}, a mono-energetic chromium x-ray imaging system at 5.6 keV. A CCD camera images the x-rays transmitted through the samples as shown in Figure 4. From the path length through the foam, the transmission and cold opacity, the density can be determined. The cold opacity is derived from the Henke tables based on the foam stoichiometry. The largest uncertainty in the density measurement is the stoichiometry. The uniformity of the doped and the undoped foams is also determined from the DCS images. If the images show non-uniformities, the foams are not used for experiments. Figure 4 shows end-on images of two sample foams. The images show both a uniformly doped foam and a chunky foam where the Ti particles coagulate to the desired size.

Figure 4. End-on images of two Ti-doped aerogel foams with (a) uniform and (b) chunky dopant profiles.

4 Manufacturing and assembly

Micro-machining processes have been developed to produce exacting fits between (SiO $_{2}$ ) aerogel cylinders and the doped aerogel inserts. Outer aerogel cylinders are machined to fit the final insertion diameter of the very thin-walled beryllium tubes. An intensive inspection process is required before machining, throughout machining and into the finished final assembly processes in order to produce these very delicate target assemblies.

Machining of the aerogel is done with Nakanishi Shika Kikai (NSK) live tooling spindles using a process called ‘turn milling’. This process is useful because it eliminates any tool pressure, allowing us to take heavier roughing cuts without having the aerogel detach from the mandrel. This method also saves crucial machining time and allows more target hardware to be machined during a production run. The live tooling is set for 20,000 RPM, with the Hardinge spindle running at 12,000 RPM in the opposite direction. In the final procedure, a $12~\unicode[STIX]{x03BC}\text{m}$ radius single crystal diamond tool is used for the final finishing of the O.D and the parting-off to finish length of the completed hardware. One live tooling mounted along the $z$ -axis on the top plate is used for facing and roughing purposes. A second live tool spindle is set at $90^{\circ }$ along the $z$ -axis for roughing out the backside. When finished, the part is done in one operation; there is no need for a second operation fixture, and hence, a time-saving process.

The first step in the assembly process is attaching the beryllium tube to the gold hohlraum. Manufacturing of the beryllium tubes uses the same technique as in Ref. [Reference Capelli, Schmidt, Cardenas, Rivera, Randolph, Fierro, Merritt, Flippo, Doss and Kline17]. The beryllium tube has three fiducial grooves machined on the outer surface with different spacing between the grooves meaning the tube has a direction. Before attachment of the Be cylinder the orientation is checked by examining the distance between the fiducials on the ends of the tube. Then the Be tube is attached to the hohlraum using the triple theta assembly station^{[Reference Schmidt18]}. Before gluing, the orientation of the cylinder and the alignment of the tube with respect to the hohlraum is checked. In addition, the seating of the cylinder on the hohlraum holding ring is verified. The two pieces are glued using an ultraviolet adhesive.

Figure 5. (a) Image of assembled target with CAD drawing showing the same view.

The foam packages are prepared next. The larger undoped aerogel foam is placed on a vacuum tip with the bored out region facing upward using the triple theta assembly station^{[Reference Schmidt18]}. Special care must be taken when handling these foams as they are very delicate and difficult to see. The foam is leveled and centered. The Sc- or Ti-doped foam is picked up with a small vacuum tip using the 5th-axis arm. The foam is also leveled and centered. Leveling and centering the two foams is a critical step since damage to both foams could occur if they are not aligned during insertion. Once both foams are centered and leveled, the Sc/Ti foam is slowly inserted into the SiO $_{2}$ foam. While inserting the foam, a side camera and a front camera are used to watch the insertion process to ensure the smaller foam is completely pushed in and to verify there are no gaps before turning off the vacuum to release the doped aerogel.

Once the foams and the hohlraum with the Be tube are assembled, the foams are inserted into the beryllium cylinder. Again, this is done using the triple theta assembly station^{[Reference Schmidt18]}. The two parts are centered and leveled. The Be tube and hohlraum are lowered over the foam. After the foam is inserted, care is taken to ensure that there is no gap between the washer at the end of the half-hohlraum. Then small glue beads are placed at the top foam/Be cylinder interface on opposite sides. A gold cone shield is added around the laser entrance hole of the hohlraum to complete the main target build.

The last step is mounting the main target to the stalk (Figure 5). The stalk consists of a rod with a rapid prototype 3D printed holder^{[Reference Cardenas, Schmidt and Peterson19]}. The rapid prototype holders are ordered from Fineline Proto Labs^[20]. The microfine green material is an approved material at the Omega laser facilities and the NIF. Using the 3D printed holder simplifies attaching the main target to the stalk since the design allows the target to rest in the holder. The half-hohlraums are mounted to the prototype holders using the hohlraum mounting station (HMS). After mounting, the position on the holder is verified and recorded on the shadow graph before finally taking Xradia^[8] images to check for gaps, defects and additional measurements.