Construction of the XLS Accelerating Structure Pre-Prototype

I.FAST Innovation Fostering in Accelerator Science and Technolog y Horizon 2020 Research Infrastructures GA n° 101004730 DELIVERABLE REPORT Construction of the XLS Accelerating Structure Pre-Prototype Deliverable D7.5 Document identifier: IFAST-Del7.5 Due date of deliverable: End of Month 44 (December 2024) Report release date: 15/12/2025 Work package: WP7 High brightness accelerators for light sources Lead beneficiary: ELETTRA Document status: Final ABSTRACT This following document (D7.5) confirms the successful construction of the CompactLight accelerating structure prototype within the I.FAST WP7. This achievement marks a key step towards Deliverable D7.6 and the realization of the full accelerating structure. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 2 / 62 I.FAST Consortium, 2024 For more information on IFAST, its partners and contributors please see https://ifast-project.eu/ This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under Grant Agreement No 101004730. IFAST began in May 2021 and will run for 4 years. Delivery Slip Name Partner Date Authored by G. D’Auria ELETTRA 12/12/2025 Reviewed by M. Vretenar, L. Celona [on behalf of Steering Committee] CERN 15/12/2025 Approved by Steering Committee 15/12/2025 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 3 / 62 TABLE OF CONTENTS 1 INTRODUCTION ......................................................................................................................................... 4 2 RF DESIGN OF THE COMPACTLIGHT ACCELERATING STRUCTURE ...................................... 4 2.1 DESIGN OF THE REGULAR CELL ........................................................................................................................ 5 2.2 ANALYTICAL STRUCTURE LENGTH OPTIMIZATION ............................................................................................. 7 2.3 IRIS TAPERING ................................................................................................................................................... 8 2.4 INPUT AND OUTPUT RF POWER COUPLERS .................................................................................................... 10 2.5 RF STRUCTURE PARAMETERS ........................................................................................................................ 12 3 MECHANICAL DESIGN AND THERMAL STABILIZATION .............................................................. 13 3.1 COOLING GEOMETRY AND THERMAL STABILITY .............................................................................................. 15 3.2 MANUFACTURING PROCESS ............................................................................................................................ 18 3.2.1 Production of starting-material: ............................................................................................................ 19 3.2.2 Pre-machining: ....................................................................................................................................... 19 3.2.3 Thermal annealing: ................................................................................................................................ 19 3.2.4 Final-machining: .................................................................................................................................... 20 3.2.5 Metrology: .............................................................................................................................................. 21 3.2.6 Cleaning: ................................................................................................................................................ 21 3.2.7 Packaging: .............................................................................................................................................. 21 4 METROLOGY CONTROL ....................................................................................................................... 22 5 BRAZING PROCEDURES ....................................................................................................................... 25 5.1 PROTOTYPE STACK ASSEMBLY PROCEDURE ..................................................................................................... 26 5.2 RF COUPLERS BRAZING TESTS .......................................................................................................................... 28 5.3 PROBLEMS EXPERIENCED IN BRAZING TESTS .................................................................................................... 31 5.4 BRAZING TESTS ON NEW RF COUPLERS ............................................................................................................ 35 6 FIRST ACCELERATING STRUCTURE STACK ASSEMBLY .......................................................... 38 6.1 BRAZING OF THE FIRST PROTOTYPE.................................................................................................................. 38 6.2 METROLOGY AND VACUUM TESTS ................................................................................................................... 41 6.3 LOW LEVEL RF MEASUREMENTS AND RF CHARACTERIZATION ....................................................................... 43 7 CONCLUSIONS ......................................................................................................................................... 45 8 REFERENCES ............................................................................................................................................ 46 9 ANNEXE ...................................................................................................................................................... 49 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 4 / 62 1 Introduction This report describes the X-band accelerating structure designed for the CompactLight project (https://www.compactlight.eu/), detailing the construction phases of the first prototype of the structure. The work has been carried out for the I.FAST work package 7. CompactLight was a a Design Study funded by the European Union under the Horizon 2020 Research and Innovation programme, completed in 2021 with the design of an innovative, compact, and costeffective hard X-ray FEL facility, using the most advanced technologies developed for accelerator cmponents, in terms of high brightness photo-injectors, compact and very high-gradient X-band accelerating structures, as well as state-of-the-art undulators. The CompactLight linac consists of approximately one hundred 0.9m long X-band accelerating structures that can operate both in a high-gradient mode, at 65 MV/m and 100 Hz pulse repetition rate, and high-repetition rate mode, at 30 MV/m and 1 kHz pulse repetition rate. The structure parameters have been optimized for both radio frequency and beam dynamics performance and have higher order transverse mode suppression for stable two-bunch operation. The cooling circuit has been designed to accommodate the high average power of the high-repetition rate mode and optimized for minimum difference between operating modes. Finally, the mechanical design of the structures has also been analyzed and optimized for industrial production. 2 RF design of the CompactLight accelerating structure The design and optimization of a linear accelerator based on traveling-wave structures involve several critical steps [1]. These steps are outlined below: 1. Selection of a Reference Accelerating Field Begin by establishing a reference value for the accelerating field, which serves as a baseline for subsequent design choices. 2. Definition of RF Power Sources and Pulse Compressor Characteristics Specify the characteristics of the RF power sources and, if applicable, the pulse compression system to ensure compatibility with the desired performance parameters. 3. Determination of the Average Iris Radius Set the average iris radius of the accelerating structure according to the requirements of beam dynamics. 4. Electromagnetic Design of the Regular Cells Conduct the electromagnetic design and optimization of regular cells, which form the core of the accelerating structure. 5. Optimization of Structure Length and Iris Tapering Explore variations in the total length of the structure and the tapering of the iris to maximize the effective shunt impedance. This step aims to enhance efficiency and reduce the total number of RF power sources needed. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 5 / 62 6. Validation Against Breakdown Rate Predictors Assess breakdown rate predictors, including the modified Poynting vector, surface electric field, and pulsed heating, to ensure they remain within acceptable limits at the nominal operating gradient. 7. Wakefield Simulations Perform wakefield simulations to evaluate beam break-up effects and confirm the stability of the particle beam under operating conditions. 8. Final Electromagnetic Design of the Structure Refine the electromagnetic design to incorporate input and output couplers, ensuring that breakdown predictors such as pulsed heating are maintained at acceptable levels. 9. Design of a Realistic RF Module Develop a detailed RF module design, including the power distribution network, to enable efficient operation of the structure. The design process typically involves several iterations of these steps to achieve an optimal balance between performance, efficiency, and reliability. The following paragraphs outline the key steps involved in the design process. 2.1 DESIGN OF THE REGULAR CELL The design of the regular cell was developed using the ANSYS Electronic desktop simulation tool, that is a Finite Element Method (FEM) solver for analyzing electromagnetic structures. Figure 1 illustrates a sketch of the cell geometry. Figure 1: Sketch of the single cell with main parametrized dimensions. In the figure, a represents the iris radius, b is the outer radius, t denotes the iris thickness, r ₀ is the rounding radius of the cell, and r ₁ = r ₂ defines the aspect ratio of the iris's elliptical profile. The cell length, d, is calculated based on an operating frequency of 11.994,2 MHz and a cell phase advance of 2π/3, resulting in 8.332 mm. The design process focused on minimizing the modified Poynting vector, normalized to the average accelerating gradient 𝑆𝑐𝑚𝑎𝑥 𝐸𝑎𝑐𝑐 2 ⁄ while maximizing RF efficiency. RF efficiency was evaluated CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 6 / 62 using the shunt impedance per unit length (𝑅) parameter. An elliptical shape for the irises was adopted to reduce the peak modified Poynting vector on the surface. It was determined that setting 𝑟1𝑟2 ⁄= 1.3 provides an optimal balance between high-gradient performance and efficiency. Additionally, the rounding radius 𝑟0 was set to 2.5 mm. After defining the iris shape, the primary cell parameters (shunt impedance per unit length (𝑅), quality factor (𝑄), group velocity (𝑣𝑔), and normalized modified Poynting vector (𝑆𝑐𝑚𝑎𝑥 𝐸𝑎𝑐𝑐 2 ⁄)), were calculated as functions of the iris radius (𝑎) and iris thickness (𝑡), as reported in Figure 2. Figure 2: Main RF and geometrical parameters as a function of the iris radius and the iris thickness. The values of 𝑎 were varied between 2 mm and 5 mm, while 𝑡 ranged from 1.5 mm to 2.25 mm. Based on these calculations, the design of the accelerating structures was finalized. The data in Fig. 2 were subsequently smoothed using MATLAB’s built-in interpolator, with a mesh spacing of 1 mm for both 𝑎 and 𝑡. The results are shown in Figure 3. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 7 / 62 Figure 3: Main RF and geometrical parameters as a function of the iris radius and the iris thickness after interpolation with MATLAB. 2.2 ANALYTICAL STRUCTURE LENGTH OPTIMIZATION With the average iris radius of the structure defined, the first step in designing the entire structure was to identify the optimal lengths for both Constant Impedance (CI) and Constant Gradient (CG) structures. This process also involved simulating the RF pulse compressor, utilizing the formulas for the SLED pulse compression system [3] applicable to both CI and CG structures [4, 5], implemented through MATLAB code [6, 7]. The effective shunt impedance, as a function of the accelerating structure attenuation, is shown in Fig. 5 for both CI and CG structures, while the optimal structure length as a function of the average iris aperture is depicted in Figure 6. For the CI structure, the optimal length has been fixed at 0.890 m, and for the CG structure, 0.818 m. These values served as the basis for a numerical optimization of iris tapering, as described in the following section. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 8 / 62 Figure 5: Effective shunt impedance as a function of the section attenuation for CI and CG structures. Figure 6: Optimal structure length as function of the average iris radius for CI and CG structures. 2.3 IRIS TAPERING Two-bunch operation in CompactLight is essential for pump-and-probe experiments. However, this operational mode introduces a critical design challenge: addressing the long-range transverse wakefield behavior of the accelerating structure. The wakefield excited by the leading bunch can influence the trajectory of the trailing bunch, leading to emittance growth. To mitigate this, longrange transverse wakefield suppression is necessary. This suppression can be achieved by varying the iris diameter along the structure's length, which detunes the synchronous frequencies of the most significant transverse modes. This detuning induces decoherence in the transverse wakefield, reducing its amplitude [11–15]. The initial constant gradient (CG) design of the X-band RF structure featured a linear tapering of the iris aperture, from 4.278 mm to 2.722 mm, with a constant iris thickness of 2 mm. The long-range transverse wakefield of this design was calculated using GdfidL [16], as shown in Fig. 7. With linear iris tapering, the long-range transverse wakefield at the 4𝑡ℎ, 6𝑡ℎ, and 10𝑡ℎ RF cycles was found to be 120, 45.83, and 18.8 V/pC/mm/m, respectively. To further suppress the wakefield, a Gaussian-like aperture tapering combined with a linear tapering of iris thickness was proposed. Through a multi- CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 9 / 62 parameter optimization of the tapering parameters, the final iris aperture and thickness profiles were determined, as illustrated in Figure 8. The blue line represents the dimensions of the linear iris design, while the red line shows the Gaussian-like design. The long-range transverse wakefield for the Gaussian-like iris design is presented in Fig. 9. The wakefield of the linear iris design (blue line) is compared to the Gaussian-like design (red line). The optimal compromise between fundamental-mode performance and minimum bunch spacing was achieved with a spacing of the 10𝑡ℎ RF cycle. In this configuration, the wakefield experienced by the second bunch is reduced to 3.65 V/pC/mm/m. Moreover, the wakefield envelope of the Gaussian-like design at the 10𝑡ℎ RF cycle is smaller than that of the linear iris design, ensuring more stable and robust operation. Figure 7: Long-range transverse wakefield of the initial linear iris design. Figure 8: The iris dimension of the RF structure. Left figure is the distribution of the iris aperture. Right figure is the distribution of the iris thickness. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 16 / 62 Figure 17. Left: temperature result of disk 2 in the high heat load case. Right: the corresponding total deformation. With these results, the parameters from the RF are recalculated and adjusted, to compensate for the difference in deformation. After that, another iteration was done to compensate for the manufacturing process, which is done at a room temperature of 20 degrees. This is just a constant, calculated with difference in the foreseen operating temperature and the manufacturing temperature times the expansion coefficient of the copper. The 3D model of the structure is shown below, Figure 18. Figure 18. 3D model of the structure From this the detailed mechanical drawings are made, see Figure 19. Detailed discs drawing (drawing number: 131-92371 ) CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 17 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 18 / 62 Figure 19. Detailed discs drawing (2 sheets) 3.2 MANUFACTURING PROCESS In this section the route towards the production of the individual accelerator parts is described. The intent is to give general overview of the typical work-flow. The typical work-flow of high precision mono-parts are reported below: 1. Production of starting material 2. Pre-machining 3. Thermal annealing 4. Final-machining 5. Metrology 6. Cleaning 7. Packaging A description of the work-flow steps are reported in the following paragraphs. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 19 / 62 3.2.1 Production of starting-material: As first step of the production work-flow is the production of the starting-material. The starting material is defined as Oxygen free copper (Cu-OFE). This step is cutting the raw material (typically bar stock) into individual blanks. 3.2.2 Pre-machining: The bulk material of the blanks is removed during the pre-machining step, by so called high-precision machines, with a high reproducibility. The machining of the material has to be done such that the amount of stress introduced in the material is low and especially reproducible. All not critical dimensions of the parts are made in this production step. All critical dimensions and low roughness surfaces, such as the "iris" of the discs, will still have additional material, which will be removed in later productions steps. All non-critical features, such as brazing grooves and thermal holes, will also be made in this production step. Figure 20 is a picture of a pre-machined disc. Figure 20. Pre-machined disc 3.2.3 Thermal annealing: After pre-machining, there will be some stress in the material that needs to be relieved before final machining. This stress can cause deformation of the parts during brazing techniques due to elevated temperature. The annealing process not only releases stress in the material, it also changes the material on a microscope level i.e. the grain-size and its distribution is also affected by the annealing process. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 20 / 62 Experiences in past learn that a single annealing step at 240 oC yields the best combined result in stress relief and grain dimensions. 3.2.4 Final-machining: The final machining consists of alternating machining and measurements steps, in order to assure product quality. Furthermore, the cleanliness of the parts is of high importance as contamination of the parts will lead to contamination of the vacuum-system where they are used in. The final machining is preformed on extremely high-precision machines. And the cutting tools are so-called single point diamond tools allow to machine the soft oxygen free copper with a high accuracy combined with a mirror finished surface. See Figure 21. Figure 21. Membrane box The machining steps are executed in two separated in two distinct groups i.e. milling and turning. In a milling operation, the machine tool rotates and removes material in an intermittent behaviour i.e. each revolution of the tool a small chip of material is removed. In a turning operation, the part rotates and the machine tool remains in contact with the part and a continuous ribbon-like stream of material is removed. Due to this difference in nature a turning surface results in a lower surface roughness compared to a milled surface. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 21 / 62 3.2.5 Metrology: During and after machining parts need to be measured to ensure that the final parts are within specifications. Depending on the stability of the process the interval at which parts are to be measured can vary from 100% for critical dimensions to 1-in-4 of even 1-in-8 depending for less critical dimensions. Form and position tolerances are measured with a touch probe of a Coordinate Measuring Machine (CMM). Measuring forces need to be much lower than normally can be expected on the CMM as no metrology marks can be left on the surface. 3.2.6 Cleaning: The Parts have to be ultra clean and this to be kept in mind during the entire manufacturing process. To avoid all possible scratches on the parts gloves need to be worn during the preand end-machining. Cutting fluids need to be free of chlorine or sulphur as this can contaminate the parts, resulting in effecting the RF function. Cutting fluids, used in the process, need to be removed with an ultrasonic cleaning method. Followed by a vapor degreasing step. Given the very fragile parts special holding tools need to be developed and made for the cleaning steps. 3.2.7 Packaging: Given the fragile nature of the parts i.e. especially the centre zone (Iris) with RF functionality, packaging is a crucial step in shipping the parts to the partner for the next fabrication step. Packaging of the individual discs is done in membrane boxes, see Figure 22. Figure 22. Membrane box CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 22 / 62 4 Metrology control One of the responsibilities form CERN is to ensure the quality of the parts delivered before the assembly process. For doing so, CERN proposed to measure some randomly selected parts to ensure the process followed by VDL ETG Precision was ok. The machine used for doing that metrology control was the LEITZ PMMC 12, CMM used for the highest accurate measurements at CERN. Following the drawings mentioned before Figure 6) and establishing a reference system as seen on the Figure 23, all selected parts were measured. Figure 23. Reference system for metrology With the aim of optimising the time and resources, a selection on the key assembly features was done. Only 13 reference dimensions were taken into account. (Figure 24) CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 23 / 62 Figure 24. Dimensions to be measured at CERN After metrology reports from discs 49, 54, 57, 59, 60, 62, 64, 65, 68, 69 and all waveguides, no significant deviations were observed. Only some deviations on the shape of the iris can be spotted on the report (Figure 25) were we can see minor deviations and after conversations with metrology experts can be explained by a mispositioning of the reference between the original drawing and the CMM machine. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 24 / 62 Figure 25. Dimensions from iris on cell 54 from CERN. Those results were discussed within the team members and we agreed that that could be a reasonable explanation. Indeed, after checking the metrology reports from VDL ETG (Figure 26) all those values were on the range of 1 micron deviation or bellow that value. Figure 26. Dimensions from iris on cell 54 from VDL ETG. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 25 / 62 Respect all the rest of variables, all are within the ranges of tolerances and meets the requirements for a good assembly process and RF performance. On the waveguide area, the focus was always on the possibility of a good assembly process. Given the fact that the RF volume on that part is not extremely crucial for the correct performance of the accelerating machine, a strong focus was directed to the interface waveguide-accelerating structure. With that focus in mind a profile of the virtual future assemble (the two parts shaping the waveguide were manually assembled during metrology) was obtained as seen on the Figure 27. The deviation is smaller than it is shown on the figure because most of the deviation is coming from the fact that the assembly is temporal and a bid step is seen just in that area. If we correct that, the possible deviation is just on the negative side, giving us more room for introducing the waveguide into the pocket. The centring of them will be ensured during the brazing procedure. Figure 27. Profile from waveguide 2 measured at CERN. 5 Brazing procedures CPI TMD are responsible for the brazing of the XLS accelerating structure. As part of their work, they provided input into the structure design to enable optimal assembly and brazing, they determined the braze type and quantity to be used, and they designed and manufactured the jig required for construction and braze of the prototype structure. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 32 / 62 In addition to this, the waveguide flanges, which were nickel plated by CERN, also did not adhere well to the braze material. CPI TMD suspects the plating layer may be too thin as there is visual evidence of the plated parts discolouring post-braze, which should not happen if sufficiently nickel plated. This issue resulted in vacuum leaks and evidence of this issue can be seen in Figure 38. Figure 37. blue oxidised cooling pipe, no braze fillet between the cooling pipe and the RF coupler disk shows lack of adheresion Figure 38. waveguide flange post-braze showing evidence of discolouration and clear lack of adheresion to braze material CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 33 / 62 Braze groove too close to RF cavity The second issue CPI TMD came across during the second braze shown in figure 15. was a number visible hole in the wall of the RF cavity following the braze. This again impacted the UHV compatibility of the assembly, as the hole connected the braze channel to the UHV cavity. The braze channel is a non-UHV section as it requires vent holes to the outside world, evidence of this issue can be seen in Figure 39, with holes appearing in two areas of just one of the RF coupler assemblies. The second RF coupler assembly did not have this issue. It appears that the braze material may have eaten into the copper edge, causing this hole. There is visual evidence of braze material on the surface of the RF cavity, showing that it has flown through the hole. CPI TMD have undertaken an investigation into this failure, they are currently unsure as to why the braze material may have eaten into the copper, however they did discover what they believe to be an issue with the design of one of the RF coupler parts. The two RF coupler assemblies have different braze groove geometries from one another. The braze groove geometry shown in Figure 40. appeared to braze very well, however the RF coupler shown in Figure 41 displayed the issue shown in Figure 39. Figure 39. two holes seen post-brazing of the RF couplers connecting the braze channels to the UHV cavity CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 34 / 62 Figure 40. successfully brazed RF coupler Figure 41. unsuccessfully brazed RF coupler CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 35 / 62 Review of the models and part drawings highlighted that there appears to be no allowable gap between the braze channel in the part shown in Figure 23. and the inner wall of the RF cavity when the parts are put together. A depiction of this issue can be seen in Figure 24. This means that a minor misalignment of the RF coupler disks, which is possible given they are only aligned using the waveguides, which themselves are not an exact fit, can result in the braze channel connecting to the RF cavity, causing a significant vacuum leak. CPI TMD attempted to fill the holes seen in Figure 21. by applying a braze paste during the third braze process shown in Figure 20. The braze paste looked as though it may have worked visually, but due to the issue of the flanges not sealing as explained earlier in the report, CPI TMD were unable to leak check the assembly. CPI TMD recommended shortening the braze channel of the RF coupler shown in Figure 23. in future designs, as the length of the channel was significantly more than needed to fill the braze gap. Out of tolerance parts or braze gap too large Another issue CPI TMD have had is related to parts being delivered out of tolerance or with a larger than acceptable braze gap. The CF flanges brazed in the first RF coupler braze, seen in Figure 15., had a braze gap larger than the recommended by the braze manufacturer which also will have impacted the ability of the parts to braze successfully. Following all of the issues described within this section, the first RF coupler brazed parts were deemed to be unusable and unrecoverable within the timeframe of the project. The project team then agreed to use the second prototype parts to re-build the first planned prototype, while re-machining the part shown in Figure 23. to shorten the braze groove length. 5.4 BRAZING TESTS ON NEW RF COUPLERS Following delivery of new RF coupler parts, CPI TMD began the process of assembling and brazing the new RF coupler assemblies. CPI TMD had agreed to nickel plate all of the stainless steel parts on delivery, to avoid any plating issues. As they did before, they initially brazed the waveguide bodies and their lids together, this time using the waveguide flange to align the two parts and then screws to hold the parts in place. The waveguides were then removed from the flanges and brazed separately. This braze went well and a good braze seal could be seen visually. Following their experience from the first RF coupler braze, CPI TMD changed the second assembly step to make the assembly and alignment easier. This time around, the second braze step brazed the RF coupler disks, the CF flanges, the cooling pipes, and additionally the waveguide assemblies all in one step. CPI TMD felt this reduced the risk of misalignment of the waveguides and waveguide flanges to the disks. An issue arose during assembly where it was discovered that parts of the RF coupler assembly were poorly fitted. CPI TMD undertook metrology on the parts to discover some parts delivered to CPI TMD were out of tolerance and some parts produced by CPI TMD were also out of tolerance. In particular the braze gap between the CF flange and the coupler disk was too large and the cooling pipes were oversized and would not fit in the cooling pipe holes. This was resolved by CPI TMD CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 36 / 62 machining down the cooling pipes to size and re-machining a new CF flange, causing a small delay to the assembly schedule. A second issue arose while CPI TMD were performing metrology on the received parts. During metrology, one of the RF coupler disks fell from its granite mount and landed on the granite table surface, falling approximately 3-5 cm. In doing so, it suffered damage to the brazing surface and, following discussion with the project team, it was determined that the part could not be used in its current condition and while re-work may be possible, it was too high risk. Evidence of the damaged part can be seen in Figure 42. VDL had a spare RF coupler disk which was originally scrapped due to a scratch in the surface of the RF cavity, however the project team agreed that it would be best to use this previously scrapped part rather than attempt to fix the damage to the disk seen in Figure 42. Once re-machining of all parts had concluded and the replacement RF coupler disk was received, the parts were assembled and brazed. Figure 43 shows the parts assembled in the furnace, ready for brazing. Figure 43. Second RF coupler assemblies in the furnace ready for braze Figure 42. Microscope photo of RF coupler disk damage CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 37 / 62 The first braze step, shown in Figure 43, went very well. The braze visually adhered well to all of the parts, evidence of this can be seen in Figure 44. It was not possible to leak check the assembly at this point, due to the fact that the waveguide flanges were missing, but CPI TMD relied on a good visual assessment of the braze flow to move on to the second braze procedure. The second braze process added the waveguide flanges to the waveguides, the assembly was turned 90 degrees to ensure the flanges were flat against the face of the waveguides. Figure 45 shows the assembly in the furnace pre-braze. Figure 44. Visual post-braze analysis of first RF coupler braze Figure 45. second RF coupler braze in the furnace CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 38 / 62 The second RF coupler braze also went very well, with the braze visually adhering well to the waveguides and waveguide flanges, evidence of which can be seen in Figure 46. Following this braze, both assemblies were leak checked and were UHV leak tight, and the RF coupler assemblies were considered complete. 6 First accelerating structure stack assembly Once the RF coupler assemblies were complete, work began on construction of the full accelerating structure. CPI TMD now have all of the parts to progress with the full build. 6.1 BRAZING OF THE FIRST PROTOTYPE The plan for the full structure was to build the stack upside down (in relation to its brazing orientation), once fully built it will be rotated securely and loaded into CPI TMD’s furnace for the third and final braze step. To do this, a braze jig has been designed and manufactured that will allow secure rotation and lifting of the fully assembled structure. Figure 47 shows the partially constructed stack (on the left) and the entire structure assembled (on the right). Figure 46. brazed waveguide flange CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 39 / 62 The whole structure, with the two new end disk assemblies, was brazed mid of February 2025. Figure 48 shows the entire structure assembled before brazing. Figure 48. Entire structure before brazing Figure 47. Partially constructed accelerating structure CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 40 / 62 Prior to running this braze, a test furnace run was performed to ensure that the braze profile chosen would reach the desired brazing temperature. A thermocouple was attached to the centre of the test mass and the results showed that the braze profile was adequate. Following the braze run, the structure was removed from the furnace and placed into CPI TMD’s cleanroom for vacuum leak testing, see Figure 49. The structure was tested using ultra-high vacuum leak checking equipment and was unfortunately shown to have a gross leak. Figure 49. RF structure after brazin CPI TMD attempted to pinpoint the source of the leak, but the structure appeared to be leaking from several sources, and the size of the leak was too large to be able to accurately identify the sources. In addition, mechanical damage was noted on the outside of the RF coupler, probably due to the thermal expansion of the structure into its braze jig during the brazing cycle, see Figure 50. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 41 / 62 Figure 50. Mechanical damage outside the RF coupler of the structure Prior to furnace, CPI TMD had put the brazing jig through a wet hydrogen greening process, which forms a green oxide layer on the stainless steel. This green oxide layer is applied to prevent the brazing parts from sticking to the braze jig. In addition to this, the braze jig was loosened once the structure was positioned in the furnace, to allow room for thermal expansion of the copper structure. Despite this, copper did stick to the stainless steel braze jig and expanded into it, causing damage to at least the outside of the stack. Following preliminary tests at CPI TMD, at the beginning of April 2025 the structure has been sent to CERN for low level RF measurements, mechanical tests (to assess any damage on the RF couplers) and for further evaluation to identify the cause of the vacuum leaks. In particular CERN Metrology will check the alignment of the structure and, using an endoscope, will also check the brazing alloy penetration on the beam axis of the structure. In addition, in order to identify the sources of vacuum leaks, the vent holes of the brazing channel will be sealed. Finally, if necessary, to fully understand the cause of the vacuum leaks, the section will be cut and the brazed areas are analysed. 6.2 METROLOGY AND VACUUM TESTS Below the metrology report carried out at CERN to verify the alignment of the structure (the full report is in Annex 1). CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 48 / 62 Linear Accelerator Conference 28, JACoW, Geneva, Switzerland, 2017, 568, ISBN: 978-3-95450169-4, DOI: https://doi.org/10.18429/JACoW-LINAC2016-TUPLR047, URL: http://jacow.org/linac2016/papers/tuplr047.pdf. [26] A. Grudiev, S. Calatroni, W. Wuensch, New local field quantity describing the high gradient limit of accelerating structures, Physical Review Special Topics - Accelerators and Beams (2009), DOI: 10.1103/physrevstab.12.102001. CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 49 / 62 9 Annexe ANNEXE 1 CERN METROLOGY REPORT CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 50 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 51 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 52 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 53 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 54 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 55 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 56 / 62 CONSTRUCTION OF THE XLS ACCELERATING STRUCTURE PRE-PROTOTYPE Deliverable: D7.5 Date: 15/12/2025 Grant Agreement 101004730 PUBLIC 57 / 62