High gradient performance and quench behavior of a verification cryomodule for a high energy continuous wave linear accelerator

An eight-cavity, 1.3 GHz, SLAC linac coherent light source II High Energy cryomodule was assembled and tested at Fermilab to verify performance before the start of production. Its cavities were processed with a novel nitrogen doping treatment to improve gradient performance. The cryomodule was tested with a modified protocol to process sporadic quenches, which were observed in Linac Coherent Light Source II production cryomodules and are attributed to multipacting. Dedicated vertical test experiments support the attribution to multipacting. The verification cryomodule achieved an acceleration voltage of 200 MV in continuous wave mode, corresponding to an average accelerating gradient of $24.1\mathrm{MV}/\mathrm{m}$, significantly exceeding the specification of 173 MV. The average ${\mathrm{Q}}_0$ ($3.0\times {10}^{10}$) also exceeded its specification ($2.7\times {10}^{10}$). After processing, no field emission was observed up to the maximum gradient of each cavity. This paper reviews the cryomodule performance and discusses operational issues and mitigations implemented during the several month program.

Figure 1

Cross section of LCLS-II-HE cryomodule, which contains eight 1.3 GHz cavities. Each module is 12.2 m long.

Figure 2

Statistics for vertical testing of production L2 cavities, showing that the gradient reach was not sufficient for HE. For L2, the minimally accepted gradient in VT was $19\mathrm{MV}/\mathrm{m}$; for HE it is $23\mathrm{MV}/\mathrm{m}$.

Figure 3

Vertical test qualification of ten cavities tested for the vCM. All ten qualified and eight were selected (open symbols indicate the two not selected). Measurements were performed at 2 K. Uncertainty in ${\mathrm{Q}}_0$ and $E_{\mathrm{acc}}$ measurements is approximately 10%.

Figure 4

Histogram of the maximum gradient minus the usable gradient for L2 cavities without excessive field emission, HOM port heating, or similar limiting extrinsic factors.

Figure 5

A strong correlation was observed between quenching (appearing in the top plot as a sudden drop in gradient) and bursts of x rays as observed with radiation detectors located alongside the cryomodule (bottom plot). The data are from Fermilab cryomodule 16 in the L2 production series.

Figure 6

Single cell cavity setup for evaluating sporadic quenching with a temperature map (left) and second sound sensors (right).

Figure 7

Data taken during a vertical test of a single cell 1.3 GHz cavity. The rapid drops in gradient indicate when quenching occurred. ${\mathrm{P}}_{\mathrm{i}}$ is the cavity input rf power.

Figure 8

Second sound (top) and temperature map data for the 1.3 GHz single cell cavity. The middle plot shows heating in steady state at $21\mathrm{MV}/\mathrm{m}$ after a sporadic quench, and the bottom plot shows heating immediately following a quench at $32\mathrm{MV}/\mathrm{m}$.

Figure 9

Quench processing of the eight vCM cavities after the initial ramp to $16\mathrm{MV}/\mathrm{m}$. The solid lines are cavity gradients and the dots are readings from the radiation detectors located along the side of the module, which have an integration time of $\sim 1\mathrm{minute}$. Detector colors correspond to those in Fig. 5. Some cavities have pauses in testing removed from the plots for ease of visualization.

Figure 10

Example of sporadic quenching over a long duration of operation at elevated gradient. A $\sim 1\mathrm{hr}$ pause in testing cavity 2 (during which other cavities were tested) is removed from the plot for ease of visualization.

Figure 11

VT maximum gradient, cryomodule maximum gradient, and cryomodule usable gradient for each cavity in the vCM. The uncertainty in the gradient measurements is approximately 10%.

Figure 12

${\mathrm{Q}}_0$ in the cryomodule compared to ${\mathrm{Q}}_0$ in VT. Measurements were made at the nominal gradient ($21\mathrm{MV}/\mathrm{m}$).

Figure 13

Data recorded during the ${\mathrm{Q}}_0$ measurement of cavity 8. Multiple quenches occurred during the process, but there were stable intervals long enough that allowed multiple ${\mathrm{Q}}_0$ measurements to be made. This plot also shows the heater and static load calibration steps. The regions where the mass flow was evaluated are highlighted with circles and dotted blue lines. The helium inlet conditions (Joule Thomson valve setting and supply temperature) were held as constant as possible during this period.

Figure 14

${\mathrm{Q}}_0$ versus quench number for three cavities (top) and the corresponding change in surface resistance relative to zero quenches (bottom). The degradation was recovered after thermal cycling.

Figure 15

${\mathrm{Q}}_0$ comparison before and after the unit test (top) and the corresponding changes in surface resistance versus the number of quenches that occurred since the most recent thermal cycle (bottom).

Figure 16

Top to bottom cavity temperature difference at the critical temperature during cooldown. Having a high ΔT is important to achieve strong flux expulsion.

Figure 17

Data showing that lowering the gradient in cavities 1 and 2 recovered the liquid level drop in the upstream reservoir.

Figure 18

Cryomodule voltage (top) and coupler vacuum level (bottom) verses time. Vacuum spikes that exceed the trip level (dashed line) shut off rf power to all cavities.

Figure 19

Cryomodule voltage, upstream liquid level and coupler pressure over four days during the unit test. The decrease in the coupler vacuum pressure over time was due in part to $Q_{\mathrm{ext}}$ adjustments that lowered the rf input power.

Figure 20

Data taken during the high gradient test in which a total voltage of 200 MV was maintained for more than one hour.

Figure 21

Top: histogram of the detuning of the eight cavities over a 4.5 hour period. The goal is not to exceed $\pm 10\mathrm{Hz}$ (red lines) at six sigma. Bottom: integrated rms detuning for the eight cavities from the frequency computed to 500 Hz.

Figure 22

Spectrogram of the cavity 6 detuning near 20 Hz and the variation in the 2.2 K helium supply line pressure.