Operational experience of a low beam coupling impedance injection kicker magnet for the CERN SPS ring

The CERN SPS injection kicker magnets (MKP) were developed in the 1970’s, before beam power deposition was considered an issue. There are two types of these magnets in the SPS: MKP-S (small aperture) and MKP-L (large aperture) versions. The MKP-L magnets are very lossy from a beam impedance perspective: this would be an issue during SPS operation with the higher intensity beams needed in the future for HL-LHC. Hence, a beam screen has been developed, which is inserted in the aperture of each MKP-L module. The screen consists of silver fingers applied to alumina U-shaped chambers: the fingers have been optimized to achieve both adequately low beam induced power deposition and good high voltage (HV) behaviour. A surface coating, with a low secondary electron yield, is applied to the inner surface of the alumina chambers to reduce dynamic vacuum. The low-impedance MKP-L has been extensively HV tested in the lab before installation in the SPS. This paper briefly presents the design and focuses on the operational experience in the SPS, including heating and vacuum.


Introduction
In CERN's Super Proton Synchrotron (SPS), a fast kicker system (MKP) is used for injection of the beam into the accelerator [1].Two different types of the MKP magnets are used: the MKP-S and MKP-L.The MKP-S has an aperture of 100 mm wide by 61 mm high and the MKP-L aperture is 141.5 mm wide by 54 mm high: the width is the distance between the high voltage (HV) and return conductors.The two apertures are used to both meet beam optics requirements and provide the required deflection, within the constraints of available length, voltage and current demands on the pulse generators.There are four MKP-L modules in the MKP-L vacuum tank.The MKP modules are transmission line type [2], constructed of multiple cells and operated in machine vacuum: the MKP-L modules have 22 cells each.
As a result of the difference in the aperture dimensions, between the MKP-S and MKP-L modules, the MKP-L is intrinsically more susceptible to the electromagnetic wakefields of the circulating beams, and therefore, the original design had a high broadband beam coupling impedance [3].
Figure 1 shows temperature and pressure measurements, during March 2022, for an MKP-S and an MKP-L: the measured maximum temperature rise of the MKP-L module is a factor of ∼4 times higher than for the MKP-S.The relatively large temperature rise of the MKP-L is attributable to its high beam coupling impedance.The temperature probes are mounted on MKP side plates, which are at ground potential.Hence, the ferrite temperature will be higher than measured during beam induced heating [4].The measured pressure in the MKP-L tank is considerably higher than for the MKP-S tank.The pressure consists of both a static component (without beam) and a dynamic component (i.e.due to Electron cloud).The static pressure (the envelope of the base of the pressure) reaches a maximum of ∼5×10 −7 mbar for the MKP-Lthis is mainly due to the heating of the ferrite: during the scrubbing, to avoid damage to the MKP-L modules, the measured temperature was purposefully limited to 70 • C, and this limited scrubbing [5].The dynamic pressure, which is the difference between the envelope of the peak pressure and the static pressure, is also significant for the MKP-L.Figure 2 shows a plot of the intensity and pressure data during the first two days of the March 2022 scrubbing run, for an MKP-L module.During the last 12 hours of this period the maximum total intensity of beam is 4×10 13 protons (p): due to heating of the MKP-L both the static and dynamic pressure increase.The dynamic pressure, normalized to the total number of protons, increases from 2.2×10 −21 mbar/p to 3.5×10 −21 mbar/p.
With the advent of higher bunch intensities for High Luminosity (HL) LHC, larger beam induced heating will occur in the ferrite yoke of the MKP-L.If this magnet were not upgraded, the Curie point of the ferrite could be reached, temporarily preventing injection while the magnet cools-down.In addition, there is a danger of mechanical damage to the MKP-L as the original design did not envisage such high temperatures.The new design has a beam screen to shield the ferrite from the wakefields of the beam and, hence, mitigates the high heat load [3,6,7,8].

Pulse output end
Alumina chamber (grey) Fingers connected to HV input plate Silver fingers (black), numbered

Beam screen design
Silver fingers serve as a conductive path for the image current of the beam, which shields the ferrite yoke from the electromagnetic (EM) wakefields of the beam, and therefore reduces the  [8].The fingers cannot be the full length of the aperture of a module, as they would conduct current during the magnetic field rise and fall times, greatly increasing the overall field rise and fall times.Similarly, the cells of the MKP-L module are too short to apply the fingers directly to the ferrite yoke [3,8].Hence, a carrier is required to mechanically support the fingers in the aperture: alumina is used for this purpose.

Alumina Carrier
The alumina will reduce the aperture available to the beam the MKP-L modules.Hence, a detailed model of the injection region was constructed and subsequent aperture studies carried out [9].These studies defined both the required beam aperture and good field regions at the entrance and exit of the MKP-L modules [9], and hence defined the maximum allowable thickness of the alumina.
In the initial design of the low-impedance MKP-L the carrier for the silver fingers was envisaged to be two flat plates, one at each of the top and bottom of the aperture.However, a pressure rise, due to electron cloud in the aperture of this low-impedance MKP-L, could result in an HV electrical breakdown between the silver fingers and the module busbars.Hence, to prevent this, the alumina chamber is closed on its sides (Fig. 3): to be able to apply the silver fingers to the chamber, it is constructed from two U-chambers.The MKP-L modules installed in the SPS prior to the Year End technical Stop (YETS) 2022-23 exhibited high pressure rise, with circulating beam, due to electron-cloud (Fig. 1).Measurements of the secondary electron yield (SEY) of the ferrite typically used at CERN for kicker magnets gave a maximum value of ∼2.1 [10].However, alumina has a significantly higher SEY (∼9) [11].To mitigate electron-cloud in the low-impedance MKP-L, each set of U-chambers was coated on their ends with amorphous Carbon [12], at CERN, and on their interior surface with Cr 2 O 3 , by Polyteknik1 .To verify the SEY of the Cr 2 O 3 , two witness samples were coated together with each chamber.The SEY of each witness sample was measured: the maximum values were in the range 1.5 to 2.0 [13].Even with the Cr 2 O 3 coating electron cloud will occur until the Cr 2 O 3 is conditioned with beam [14] and the coatings maximum SEY is reduced to ∼1.4 or less.

Silver Finger Design
The final design of the low-impedance MKP-L has eight silver fingers in total, with four on each of the top and bottom alumina U-chambers (Fig. 3).Each set of four has two fingers connected to the HV plate at the input end of the module and two to the HV plate at the output end (Fig. 3).Each finger is only connected at one end, and capacitively connected to the opposite pair of fingers, via the permittivity of the alumina.This capacitive path allows high frequency components of the image current of the beam to flow, but blocks the lower frequencies associated with the rise and fall times of the magnetic field.The resulting reduction in beam impedance is confirmed by both predictions and measurements [8].This final design of the screen is the result of several iterations to optimize the beam-coupling impedance [8] and the HV behaviour [6,7,12,15].

Operation in the SPS with Beam
The low-impedance MKP-L was installed in the SPS during the YETS 2022-23.As expected from predictions and measurements [7], the silver fingers did not influence the trajectory of the beam during injection.
Figure 4 shows temperature and pressure measurements, from 24/3/2023 to 4/4/2023, for an MKP-S module and a low-impedance MKP-L module: the plots in Fig. 1  MKP-S and MKP-L modules in Fig. 4 are 22.8 • C and 7.9 • C, respectively.Thus, the maximum temperature rise of the low-impedance MKP-L is ∼35% of the MKP-S, compared to a factor of ∼4 higher before the upgrade (Fig. 1).The smaller temperature rise in the low-impedance MKP-L, with respect to the MKP-S, is expected from predictions [8].The data shown in Fig. 1 is for beam at injection energy (prior to 12:00 hrs on 23/03/2022) and accelerated beam (after 12:00 hrs on 23/03/2022): the data shown in Fig. 4 is for beam at injection energy until 30/3/2023.The low static pressure rises in both the MKP-S (Fig. 1) and the low-impedance MKP-L (Fig. 4) vacuum tanks is consistent with the relatively small temperature rises.
The top plot of Fig. 5 shows total beam intensity and measured pressure, from 24/3/2023 to 4/4/2023, for the low-impedance MKP-L's.To ensure reliable operation of the low-impedance MKP-L, since no spare low-impedance magnet is available yet, the goal was to limit the tank pressure to 2×10 −7 mbar when the magnet is pulsed for injecting beam: thus, injection was inhibited while this pressure was exceeded.By 29/03/2023, the total beam intensity was ∼6×10 13 protons, consisting of four batches each of 72 bunches and an intensity of 2.2×10 11 protons per bunch.
The bottom plot of Fig. 5 shows a more detailed view of the intensity and pressure on 26/03/2023 and 27/03/2023.During this period the maximum total intensity of beam in the SPS is reasonably constant (∼4×10 13 p), whereas the MKP-L pressure reduces from 1.7×10 −7 mbar to 1.0×10 −7 mbar.Hence, the normalized pressure decreases from 4.3×10 −21 mbar/p to 2.5×10 −21 mbar/p: this clearly demonstrates conditioning of the surfaces facing the beam.Comparing Fig. 2 and the bottom plot in Fig. 5, at a beam intensity of ∼4×10 13 p, it is clear that the dynamic pressure in the low-impedance MKP-L, at the end of these time periods, is smaller than in the originally installed MKP-L: the lower pressure rise is despite the original MKP-L having been in the SPS since 2017 and therefore would be expected to be well conditioned with beam.
In 2017 an aluminium liner coated with 50 nm thick Cr 2 O 3 was installed in the SPS, to validate the performance of the coating [16].During the YETS 2022-23 this liner was removed and SEY measurements have been carried out on it.The measurements show that the maximum SEY of the coating, in regions closest to the core of the beam, have conditioned to a maximum SEY of ∼1.3.In regions further from the core of the beam the maximum SEY is ∼1.5 [17].These measurements indicate that, after six years of operation with beam, the Cr 2 O 3 coating is still very effective.

Conclusion
Silver fingers, for reducing beam induced heating, are applied to an alumina chamber which is inserted into the aperture of each of the four MKP-L modules.The final design of the fingers is a result of an optimization of the beam coupling impedance and the HV behaviour.The interior of the alumina chamber is coated with Cr 2 O 3 to reduce its SEY and thus help to mitigate electron cloud.The 4-module MKP-L tank was installed in the SPS during the YETS 2022-23.Operation with high intensity beam shows that the low-impedance design has significantly reduced beam induced heating compared with the original MKP-L, which had high broadband beam coupling impedance.In addition, the dynamic pressure in the vacuum tank of the low-impedance MKP-L demonstrates relatively rapid conditioning of the surfaces facing beam.These observations validate the low-impedance design of the MKP-L: the low-impedance design improves machine availability, and allows for the full HL-LHC potential for the SPS injection system.

Figure 1 .
Figure 1.Temperature and pressure measurements, during March 2022, for an MKP-S module and an original (high broadband beam coupling impedance) MKP-L module.Until 12:00 hrs on 23/03/2022, scrubbing beam was kept at injection energy; after that date, it was accelerated.

Figure 2 .
Figure 2. Total beam intensity and pressure measurements, during the first two days of the March 2022 scrubbing run, for an MKP-L module (scrubbing beam at injection energy).

Figure 3 .
Figure 3. Cross-section of the upgraded MKP-L.
and Fig. 4 are for modules with the highest measured temperature.The maximum temperature rise of the 14th International Particle Accelerator Conference Journal of Physics: Conference Series 2687 (2024) 082034