Plasma cleaning of the SLAC Linac Coherent Light Source II high energy verification cryomodule cavities

Plasma cleaning is a technique that can be applied in superconducting radio-frequency cavities in situ in cryomodules to decrease their level of field emission (FE). We developed the technique for Linac Coherent Light Source II (LCLS-II) cavities and we present in this paper the full development and application of plasma processing to the LCLS-II High Energy verification cryomodule (vCM). We validated our plasma processing procedure on the vCM, fully processing four out of eight cavities of this CM, demonstrating that cavity performance was preserved in terms of both accelerating field and quality factor. Applying plasma processing to this clean, record breaking cryomodule also showed that no contaminants were introduced in the string, maintaining the vCM FE-free up to the maximum field reached by each cavity. We also found that plasma processing eliminates multipacting (MP)-induced quenches that are frequently observed within the MP band field range. This suggests that plasma processing could be employed in situ in CMs to mitigate both FE and MP, significantly decreasing the testing time of cryomodules, the linac commissioning time and cost, and increasing the accelerator reliability.

Figure 1

Photo of TESLA shape nine-cell 1.3 GHz cavity, with highlighted ports for the higher order modes (HOM) couplers, the fundamental power coupler (FPC), and the pickup (PU) coupler.

Figure 2

Temperature sensors installed on the cable end (left, 90° connector) and cavity HOM feedthroughs (right).

Figure 3

Power versus time during cable testing for LCLS-II plasma processing: input power of 100 W for 30 min. The input and reflected power, $P_{\mathrm{IN}}$, $P_{\mathrm{R}}$, respectively, follow the left y axis, while the right y axis refers to the transmitted power $P_{\mathrm{T}}$.

Figure 4

Temperature of HOM1 and HOM2 cables and feedthroughs versus time during cable testing for plasma processing: input power of 100 W for 30 min. HOM1 denotes the cable connected to the HOM FPC side and used as an input line, HOM2 denotes the cable connected to the HOM pickup side and used as the transmitted line. For both HOM1 and HOM2 cables, the 90° connector (shown in Fig. 2, left) is the side connected to the HOM coupler, while the straight connector is the opposite end of the cable. The HOM-FPC and HOM-PU copper connectors denote the sensors placed on the HOM antenna feedthroughs, shown in Fig. 2, right.

Figure 5

Experimental setup used to apply plasma processing to the LCLS-II-HE vCM. The gas station is connected to one end of the CM (a) and the vacuum and RGA station to the other end (b). The rf rack and computers are located in between the gas and vacuum stations (c).

Figure 6

Plot of RGA signals measured on CAV1, during the first day of plasma processing from cell no. 1 to no. 5 and cell no. 9. As the gas flow is started and the pressure in the vCM increases, there is an initial peak in the oxygen signals (at 16 and 32 amu), which then decreases and stabilizes to $\approx 1.6\%$ of the neon content. The partial pressures measured by the RGA differ from the total pressure in the cavity string since the RGA is positioned on a bypass line near a turbo pump on the vacuum cart to allow the RGA filament to sample the gas composition. A leak valve is used to regulate the gas flow through the RGA.

Figure 7

Pressure measured at the two ends of the vCM during initial neon-oxygen mixture injection and during the flow maintained for the entire duration of plasma processing. The pressure drop across the string is approximately 24 mTorr. The pressure values are scaled to account for the gauge sensitivity to neon. These data were acquired during plasma processing of CAV4.

Figure 8

Experimental data collected during plasma processing of CAV4. (a) contains a plot of the RGA signals measured during plasma processing of cells no. 9 to no. 4. The peak at approximately $9:30$ measured for CO and ${\mathrm{CO}}_2$ indicates the accidental ignition of the plasma in the HOM coupler. The following peak at approximately $11:00$ is measured when the plasma is successfully ignited inside the cavity. (b) shows the input and reflected power measured at the CAV4 HOM1 coupler and transmitted by the HOM2 during plasma processing. Note the different scales on the y axis between the top and bottom plots. Additionally, in this case, the $9:30$ peak corresponds to accidental HOM ignition, while the next peak corresponds to the ignition of the plasma in the cavity, followed by a rapid sequence of smaller intensity peaks that identify the plasma being moved from cell no. 5 to no. 9 and then tuned in cell no. 9. After approximately 1 h, the plasma density is decreased, and the glow discharge is moved to the next cell. This procedure is repeated for each cell. The temperature increase measured on the HOM1 and HOM2 connectors of all four plasma-processed cavities is shown in (c). In the period plotted here, the glow discharge was ignited in CAV4, as confirmed by the temperature increase registered by its HOM1 and HOM2 sensors. (d) shows the temperature increase measured on the outside of cells no. 1 and no. 9. The four sensors are placed inside the He vessel at the top and bottom of the end cells.

Figure 9

Plots of the $S_{21}$ (green), $S_{11}$ (red), and $S_{22}$ (black) parameters for some of the dipole modes used for plasma processing. The parameters were measured on CAV1-4-5-8 from top to bottom using a network analyzer, the y axis is in the unit of dB. (a), (d), (g), and (j) contain the 2D-1 peaks, (b), (e), (h), and (k) show the 1D-5 peaks, and (c), (f), (i), and (l) show the 1D-1, 1D-2, and 1D-3 peaks. The plots present an example of the different couplings to the HOM and the dipole splitting characteristics of each cavity. Exact measurements of the $S_{21}$, $S_{11}$, and $S_{22}$ are critical, particularly for mode 2D-1, to correctly identify the rf frequency to use during plasma ignition in order to minimize the probability of causing plasma ignition at the HOM coupler.

Figure 10

The three plots show the number of MP quenches found for each cavity as a function of the integrated time that the cavity spent powered on. Plot (a) shows these results for the first cooldown, (b) for the second cooldown measured before plasma processing, and (c) for the after plasma processing cooldown.