Collimator irradiation studies in the Argonne Advanced Photon Source at energy densities expected in next-generation storage ring light sources

The Advanced Photon Source (APS) team is building a fourth-generation storage ring (4GSR), replacing the present double-bend achromat lattice with a multibend achromat system thereby allowing the production of ultrabright x-ray beams. The new lattice enables a 2-order-of-magnitude reduction in horizontal beam emittance and a factor of two increase in beam current. The result is an electron beam of very high energy and power densities. Initial predictions suggest many common ultrahigh-vacuum-compatible materials struck by the full-intensity electron beam will be damaged. Two experimental beam abort studies have been conducted on collimator test pieces in the present APS SR to inform the design of a fully-functional machine protection system for APS 4GSR operations at 200 mA. A comprehensive suite of diagnostics was utilized during the studies. The diagnostics used in these experiments are not new, but employed in different ways to obtain unique data sets. With these sets now in hand, we are developing new numerical tools to guide collimator design using pelegant [M. Borland, elegant: A flexible SDDS-compliant code for accelerator simulation, Technical Report No. LS-287, Advanced Photon Source, 2000; Y. Wang and M. Borland, Implementation and performance of parallelized elegant, in Proceedings of the 2007 Particle Accelerator Conference, http://cern.ch/AccelConf/p07/PAPERS/THPAN095.PDF, pp. 3444–3446], mars [N. V. Mokhov and S. I. Striganov, Fermilab-conf-07/008-ad, AIP Conf. Proc. 896, 50 (2007)], and flash [B. Fryxell et al., flash: An adaptive mesh hydrodynamics code for modeling astrophysical thermonuclear flashes, Astrophys. J. Suppl. Ser. 131, 273 (2000); P. Tzeferacos et al., Laboratory evidence of dynamo amplification of magnetic fields in a turbulent plasma, Nat. Commun. 9, 591 (2018)].

Figure 1

Damage in the beam-facing surface of a W collimator after irradiation by separate beam aborts at 7-GeV, 100-mA stored current, yielding energy deposition of 2–3 MGy.

Figure 2

Simplified horizontal collimator geometry implemented with mars. Shown in (a) is a cross section through the collimator apex at $z=10\mathrm{cm}$ with the e-beam into the page, and in (b) is an anamorphic view of a lateral slice at beam elevation ($y=0$).

Figure 3

Static mars simulation of combined electron-positron flux in collimators fabricated from (a) aluminum and (b) tungsten. The $x\text{−}y\text{−}z$ voxel dimensions are $0.02\mathrm{cm}\times 0.1\mathrm{cm}\times 0.1\mathrm{cm}$ centered at beam elevation ($y=0$).

Figure 4

Simulated equilibrium beam parameters for the second collimator experiment, from particle tracking simulations with pelegant. The horizontal axis label, “Abort” refers to the sequence number of the beam dumps (see Table 6).

Figure 5

Simulated peak dose in aluminum for the second collimator experiment. Shown are the raw results, mars results, and those including thermal diffusion.

Figure 6

Vertical extent of the liquified-aluminum region for the second collimator experiment, based on the thermally diffused simulated dose map.

Figure 7

(a) flash 2D simulation domain (initial condition) and (b) the mars dose map overlaid on the same region.

Figure 8

flash 2D-density simulation of a 48-bunch, 200-mA beam dump in aluminum spread over five turns. The domain is the same shown in Fig. 7. The results are shown at (a) $1.28\mu \mathrm{s}$ [0.35 turn] and (b) $4.37\mu \mathrm{s}$ [1.19 turns].

Figure 9

Dose map at the collimator apex for a simulated APS-U 48-bunch, 200-mA whole-beam dump fill pattern using a $128\times 128$ grid over an x-y domain of 0.3 mm by 0.3 mm (voxel size: $2.34\mu \mathrm{m}$). The dose is scaled to that for a single bunch. The insets plot the x- and y-dose profiles at the location of maximum integrated dose, yielding ${\mathrm{FWHM}}_x=31.3\mu \mathrm{m}$ and ${\mathrm{FWHM}}_y=26.3\mu \mathrm{m}$.

Figure 10

Dose map at the collimator apex for a 200-mA, whole-beam dump from the January 2020 experiment simulating the eighth loss event (SN08). Note the increased size of the footprint relative to the APS-U case presented in Fig. 9. The x- and y-dose profiles plotted at the location of maximum integrated dose yield ${\mathrm{FWHM}}_x=132.1\mu \mathrm{m}$ and ${\mathrm{FWHM}}_y=43.9\mu \mathrm{m}$.

Figure 11

Horizontal collimator assembly plan view. The collimator test pieces are stacked in the y direction so that only one is visible in this view; in addition, the collimator pieces are shown fully inserted, past the centerline. In the experiments, the collimators are withdrawn to a parking position allowing the beam to circulate. Dimensions are given in mm.

Figure 12

Collimator test piece. (a) Oblique view. (b) The manufactured article showing 1-mm-spaced rulings along one of the lateral surfaces. Distance along the chord on the beam-facing surface is 41.5 mm.

Figure 13

An anamorphic view of the collimator surface showing curvature and roughness for the aluminum test piece used in May 2019.

Figure 14

The reduced-horizontal-beam-size lattice functions in S37 and S38. The collimators were positioned at $s=1023.78\mathrm{m}$. Horizontal and vertical tunes were 36.169 and 19.241.

Figure 15

Beam emittances, scraper position, and beam current for a 200-mA beam with a 2.6 mm vertical bump. The plot ends with the beam dump.

Figure 16

Loss intensity profiles from (a) ID1 fast BLM and (b) pelegant modeling. Both ordinate axes are given in arbitrary units. The leading two-digit numbers in the legend correspond to the beam dump sequence numbers in Table 6.

Figure 17

Comparison of loss signal arrival times determined from (a) the fast BLMs and TBT BPMs and (b) pelegant simulations. The fast BLMs are based on high-purity, fused-silica fiber optic bundles; in sector 1 these are labeled ID1_ChM, where $\mathrm{M}=1,2,3$ and in sector 6 labeled ID6_ChN, where $\mathrm{N}=1,2,3,4$. pelegant results are presented for 10, 50, and 90 percent particle loss.

Figure 18

Visible light emission from beam strikes during experiment 2 at (a) 69.4 mA and (b) 201.2 mA corresponding to SN02 and SN09 in Table 6. Beam is moving from right to left. The yellow dashed circles in (a) are locations where ejecta tracks appear to bend. The horizontal scale is centered on the field of view which is near the upstream end of the collimator. The apex is out of the field of view to the left of the image. The vertical scale is centered at nominal beam elevation ($y=0$).

Figure 19

Emission images at 200 mA from two beam dumps at the same y-offset location; referring to Table 6 these strikes are (a) SN11 and (b) SN12. A reduction in light intensity is evident in the second strike.

Figure 20

Emission images at 200 mA from the five-strike case at the same $y$ location. Again referring to Table 6, light intensities from beam strikes (a) SN14 (first) and (b) SN18 (fifth) are shown.

Figure 21

Last image of the collimator surfaces after the final beam dump of the May 2019 study, shown in false color. The titanium alloy collimator test piece is above the gap and the aluminum piece is below. The field of view is 11 mm. Color indicates reflected light intensity.

Figure 22

Photograph of the January 2020 collimator test pieces still attached to the scraper body. Sequence numbers are shown on the left; targeted currents are listed on the right. “Targeted” refers to nominal current we were intending to store before the beam was dumped. The actual currents are given in Table 6. The gap between upper and lower collimator pieces is 0.83 mm. The pieces were positioned at $x=-2\mathrm{mm}$ prior to dumping the beam by rf muting.

Figure 23

An enlargement of the boxed region shown in Fig. 22. Sequence numbers, shown on the left, correspond to those given in Table 6; thus, starting from the first visible mark at the bottom are strikes of 34.6, 69.4, 99.1, 202.0, 100.0, 201.2, and 202.1 mA corresponding to sequence numbers 01, 02, 03, 08, 06, 09, and 10, respectively. The vertical separation between each strike position is 0.4 mm. Illumination is from the bottom of the image.

Figure 24

Overview of the top and bottom collimator test pieces struck by beam dumps recorded at $25\times$ magnification. The top piece was subjected to single-strike beam dumps; the bottom piece received both single and multiple beam dumps at the same vertical offsets. Beam was moving from right to left. Three $z$ locations of interest are indicated on the top image; the apex location is just upstream of the $z_2$ position. All dimensions are in mm.

Figure 25

Adjacent, vertically merged microscopy images of beam strikes SN 06 and 08 (100 mA, upper, and 202 mA, lower), from a $700\text{−}\mu \mathrm{m}$ section of the top collimator at location $z_0$. The flattened images are recorded at a magnification of 200x.

Figure 26

Anamorphic topological map of beam strike SN 08, 202 mA from the same $700\mu \mathrm{m}$-length section of the top collimator shown in Fig. 25. The increment in the $x$ direction is $2\mu \mathrm{m}$.

Figure 27

Surface profiles extracted from the SN08 stack data of Fig. 26 ($z_0$) in waterfall format. The $z$ separation between profiles is $100\mu \mathrm{m}$.

Figure 28

Flattened image of the SN08, 200 mA beam dump at location $z_1$. A large variation of trench depth occurs at this position.

Figure 29

Surface profiles from the SN08 strike zone of Fig. 28 showing relatively large variation in depth along the trench. Separation between the profiles is $300\mu \mathrm{m}$ starting from $z=0$.

Figure 30

Flattened image of the SN08, 200-mA, single-strike beam dump at location $z_2$.

Figure 31

Flattened image of multiple-strike zone at $y_{\mathrm{off}}=-2.6\mathrm{mm}$; SN 14-18 (Table 6).

Figure 32

Five-strike FWHM loss durations taken from ID1 and ID6 fast BLMs compared with pelegant 90% loss increment. The BLMs in ID1 are labeled ID1_ChM where $M=2,3$; and the BLMs in ID6 are labeled ID6_ChN where $N=1,2,3,4$.

Figure 33

Polished and etched cross sections of the aluminum test piece from the May 2019 study showing regions struck at (a) 33.1 mA and (b) 67.4 mA.

Figure 34

Comparison of FWHM loss durations for all 20 beam dumps of the second experiment determined from (a) the fast BLMs and TBT BPMs and (b) pelegant simulations. pelegant results are presented for time required to lose 50, 80, and 90 percent of beam particles. The fast BLMs in ID1 are labeled ID1_ChM where $M=1,2,3$ and those in ID6 are labeled ID6_ChN where $N=1,2,3,4$. The fast BLM in S37 is simply labeled S37.

Figure 35

Contour plots from the ID1 external FO bundle BLM during the 18 mA beam dump. The fill pattern held 27 equally spaced 2.2 nC bunches. The horizontal (fast) axis shows loss signals within a turn, and the vertical (slow) axis shows behavior turn by turn. Part (a) shows loss signals over a full turn on the fast axis; (b) zooms in on losses from three adjacent bunches.

Figure 36

Average and peak circulating power densities in the 6-GeV electron beam.