A high throughput facility for the RF characterisation of planar superconducting thin films

Superconductor Science and Technology PAPER • OPEN ACCESS A high throughput facility for the RF characterisation of planar superconducting thin films To cite this article: D Seal et al 2024 Supercond. Sci. Technol. 37 115012 View the article online for updates and enhancements. You may also like Thermal quench modeling of REBCO racetrack coils under either alternating current or short-circuit voltage Arif Hussain, Anang Dadhich and Enric Pardo - Local plastic deformation effects on the critical current of high-Jc Nb3Sn strands under uniaxial strain Yunhao Liu, Yi Sun, Peng Gao et al. - Understanding vortex dynamics in CaK(Fe,Ni)4As4 and Ba(Fe,Co)2As2 single crystals under the influence of random point disorder N Haberkorn, M Xu, J Schmidt et al. - This content was downloaded from IP address 194.12.132.102 on 09/12/2025 at 10:29 Superconductor Science and Technology Supercond. Sci. Technol. 37 (2024) 115012 (19pp) https://doi.org/10.1088/1361-6668/ad7643 A high throughput facility for the RF characterisation of planar superconducting thin films D Seal1,2,3,∗, O B Malyshev2,3, P Goudket2,3,5, T Sian2,3, L Gurran1,3, R Valizadeh2,3, H Marks1,3, S Pattalwar2,3, N Pattalwar2,3, C Pira4, E Chyhyrynets4and G Burt1,3 1Engineering Department, Lancaster University, Lancaster LA1 4YR, United Kingdom 2ASTeC, UKRI/STFC Daresbury Laboratory, Daresbury, Warrington WA4 4AD, United Kingdom 3Cockcroft Institute,UKRI/STFC Daresbury Laboratory, Daresbury, Warrington WA4 4AD, United Kingdom 4Istituto Nazionale di Fisica Nucleare (INFN),Laboratori Nazionali di Legnaro (LNL), 35020 Legnaro, Italy E-mail: [email protected] Received 22 February 2024, revised 27 August 2024 Accepted for publication 2 September 2024 Published 7 October 2024 Abstract Accelerator laboratories worldwide are researching copper radio frequency (RF) cavities coated with superconducting thin films to exceed the limits of bulk niobium. The development and RF testing of thin films on small planar samples is vital before cavity depositions. A team at Daresbury Laboratory have developed a cost-effective facility using a novel 7.8 GHz Choke Cavity for the RF characterisation of planar samples. RF chokes ensure that no electrical contact is required between the sample and the cavity. The main advantages are: a simple sample design (90–130 mm diameter disk with no sample-cavity welding) and easy operation using a LHe-free cryostat. This enables high sample throughput, with up to 3 sample tests per week, making the facility suitable for quick, systematic scanning of deposition parameters. With the sample thermally and physically isolated from the test cavity, it is possible to measure the average surface resistance, Rs, directly using an RF-DC compensation method. Facility commissioning has been performed with bulk and thin film niobium samples. These tests have demonstrated the ability to measure Rsat temperatures in the range 4–20 K and sample peak magnetic fields up to 3 mT. At present, the minimum resolvable Rsis 0.5 µΩwith typical uncertainties of 9%–15%. The design, operation and commissioning of this facility is reported in this paper. Keywords: superconducting radio frequency, superconductivity, thin film, choke cavity, niobium 5Now at European Spallation Source, Lund, Sweden. ∗Author to whom any correspondence should be addressed. Original Content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 1 © 2024 IOP Publishing Ltd. All rights, including for text and data mining, AI training, and similar technologies, are reserved. Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al 1. Introduction Bulk niobium has been the main material used for superconducting radio frequency (SRF) cavities in particle accelerators for more than 50 years. Nb has been the superconductor (SC) of choice because it has the highest critical temperature (Tc) and the highest lower critical magnetic field (Bc1) of any element [1]. Furthermore, because Nb is a metallic element, it can be machined, formed, and welded to the required cavity shape using straightforward techniques. With significant improvements in the RF performance of bulk Nb 1.3 GHz cavities in recent years, these cavities have moved very close to their theoretical limit in the accelerating gradient: Eacc ⩽57 MV m−1[2]. For example, 1.3 GHz TESLA shaped single-cell cavities have demonstrated Eacc ≈ 55 MV m−1at 2 K [3]. However, regularly achieving these levels with bulk Nb comes at a significant cost for raw materials, manufacturing, and liquid helium (LHe) cooling. For all new accelerators and accelerator upgrades, the production cost of the bulk Nb cavities is a significant factor. A solution to this problem is to use Cu cavities coated with a Nb thin film (TF) of a few µm in thickness. This process has been in development since the 1980s with the use of Cu cavities coated with a Nb TF (Nb/Cu) for the LEP accelerator at CERN [4]. An advantage of using a Cu substrate is that it is significantly less expensive than Nb. In addition, Cu has a thermal conductivity much higher than that of Nb at cryogenic temperatures, allowing for more thermally stable cavities. However, TF cavities trap magnetic flux efficiently at low field and detrapping of flux is one of the possible origins of the Q slope [5]. In addition, the main difficulty in replicating the performance of bulk Nb with Nb/Cu is optimising the deposition parameters and substrate preparation. With a shift towards more sustainable accelerator technology, there is a growing interest in using more TF cavities instead of bulk Nb [6]. The main drive to explore TF cavities is to be able to reach high intrinsic quality factors (Q0) at temperatures, T>4.2 K. Bulk Nb cavities can regularly reach, Q0=1010 −1011 [7], however this is at a lower T=2 K for frequencies >1 GHz. Using TFs also opens up the possibility to explore, not only Nb, but other SC materials (e.g. NbTiN, Nb3Sn, NbN, V3Si, MgB2) with higher superheating fields (Bsh) and higher Tc. In addition, Superconductor-Insulator-Superconductor (SIS) layers on top of Nb, first theorised by Gurevich [8], are also being investigated as a way of screening the underlying bulk layer from the applied magnetic field. Ultimately, high Q0TF cavity operation at T=4.2 K is the main objective, which provides two main cost-saving advantages: •Lower operational costs as a result of a reduction in cryogenic energy consumption. •Lower capital costs by switching from bulk Nb to bulk Cu and simplifying cryogenic infrastructure. For any TF study, it would be ideal to deposit films on multiple cavities and perform SRF testing. However, this approach is too costly and time consuming. Instead, the materials should first be deposited on small, planar samples to optimise the deposition parameters and identify candidate TFs. After TF deposition parameters with the desired properties are obtained, all knowledge and experience should be applied to coat the curved surface of an SRF cavity. The development of superconducting TF planar samples typically consists of five main research components: (i) Substrate preparation (ii) TF deposition (iii) TF characterisation with surface and material analysis techniques (iv) Superconducting DC/AC measurements (v) Superconducting RF measurements To perform DC/AC and RF superconducting measurements, cryogenic test facilities are vital. The main aim of these facilities is to provide high resolution measurements with a quick sample turnaround, allowing rapid sorting between samples to identify candidate TFs that should be investigated further. At Daresbury Laboratory, it has been possible for many years to characterise samples using DC superconducting techniques in upgraded facilities [9]. A variable temperature insert allows measurements of the residual resistance ratio (RRR) and Tc[10]. In addition, a recently commissioned magnetic field penetration facility allows measurements of the field of first flux penetration (Bfp) as well as Tc[11]. These methods provide useful information on film quality; however, it is yet to be proven that these properties have a strong correlation with the behaviour of the material under RF conditions, which therefore does not currently allow one to predict it. Thus, a custom facility for the RF characterisation of planar TF-coated samples is required. Several facilities have been developed around the world to characterise planar TF-coated samples under RF conditions [12]. These facilities have been designed to measure RF losses on samples no larger than the diameter 180 mm and therefore measure their average surface resistance (Rs). For this, one of two main measurement methods is usually employed: •End-plate replacement [12–15]: based on two measurements of Q0to calculate Rs—one with a reference sample made from the same material as the test cavity and one with the sample of interest. These facilities usually only require simple cavity designs, which could allow for quick sample turnover. •RF-DC compensation [16–21]: allows the RF losses on the sample to be calculated directly from a calorimetric measurement using resistive heaters connected to the sample. This method is only possible provided the sample is thermally isolated from the cavity. Having reviewed the existing RF facilities and methods, a set of criteria was established for a new test cavity: 2 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al (i) Ability to directly measure RF losses on a simple planar sample design separately from the test cavity, that is, using the RF-DC compensation method. (ii) A low-cost sample design. (iii) A simple cavity design. (iv) A fast sample turnaround that means no welding, vacuum, or RF seals between the test cavity and sample. (v) A possibility to reuse the sample for additional coatings and RF tests or for tests in other facilities. (vi) Easy to vent, open, close, and pump vacuum system. (vii) Easy and safe cryogenic operation. In order to meet this set of criteria, a novel cavity was designed, modelled, and manufactured. The test cavity became known as the Choke Cavity, and a new RF test facility has been designed, built, and tested to house this for quick sample testing. This paper provides details of the design and commissioning of the Choke Cavity facility. 2. Choke Cavity 2.1. Cavity design The simplest sample design that could be tested is in the shape of a planar disk. Planar samples are easy to manufacture, costeffective and much simpler to deposit a TF on compared to curved structures. At the time of design, the maximum sample diameter was limited to 100 mm considering the size of the available deposition chamber at the laboratory [22] and the dimensions and cost of the deposition targets. This size also defined the maximum size of the test cavity. Eliminating the need for welding meant that a small vacuum gap would be present between the test cavity and the sample. With this gap, it is possible to separate the RF losses on the sample because of the electrical and thermal isolation of the sample from the cavity. However, because of this, a method was required to contain the RF fields within the cavity and the sample to prevent all RF from leaking through the gap. To minimise RF leakage, it was decided to use RF chokes. Chokes have been proposed in the past for use in damping RF cavities to trap the accelerating mode within the resonator [23]. They have also been used by the Surface Impedance Characterisation system at JLab for small sample testing [24]. The use of quarter-wavelength chokes acts as narrow bandpass filters to contain the fundamental mode frequency within the cavity. Their sizes are tuned to a specific frequency and gap size between the cavity and sample to minimise RF leakage whilst not having a physical RF seal. In addition, this cavity would be housed in a single vacuum, LHe-free cryostat, therefore not requiring any vacuum seals. The Choke Cavity was designed in CST Studio Suite [25] with a target resonant frequency, f0=7.8 GHz. The final design is shown in figure 1which illustrates the Choke Cavity on top of a sample with a small gap. The starting point for modelling was to set up a resonance between a central elliptical half-cavity and a single choke (choke 1). The design was then simulated with the addition of a second (choke 2) and third choke (choke 3) surrounding choke 1 and the central cavity. These additional chokes maximise the RF fields contained within the system and ensure that minimal RF leakage from the first choke is radiated away from the system. Several simulations were performed with varying choke depths, widths, positions, and gap sizes. These simulations used electrical conductivity of 7 ×1013 Sm−1for both the sample and cavity corresponding to an Rs≈20 µΩ(BCS resistance of RRR =300 bulk Nb at 4.2 K and 7.8 GHz calculated using the SRIMP code [26]). For the input of RF power into the system, some power is dissipated in the Choke Cavity, some on the sample, and the rest is radiated as leakage through the gap (Prad). One of the main optimisation goals was to minimise total losses through the system. To quantify all RF losses, a quality factor was defined for the system, Qsystem, which should be maximised. It is defined as 1 Qsystem =1 Q0 +1 Qrad ,(1) with Qrad =ω0U Prad ,(2) where Q0is the quality factor of the combined Choke Cavity and sample, Qrad is the quality factor due to radiative losses, ω0 is the angular resonant frequency, and Uis the stored energy. The Qsystem as a function of choke 1 depth comparing models with 1 choke, 2 chokes, and 3 chokes is shown in figure 2. At the resonant frequency, leakage is minimal (i.e. Qrad ≫Q0), therefore from equation (1) it can be seen that Qsystem ≈Q0≈107. This shows the clear benefits of adding two chokes around the central cavity and choke 1. The range of choke 1 depths for which Qsystem >106increases dramatically between the three designs, meaning that the 3 Choke Cavity is much less sensitive to manufacturing tolerances. Figure 3 shows Qsystem as a function of the gap for the three models. It can be seen that the 1 choke design is very sensitive to small variations in the gap, whilst the 3 choke design is capable of maintaining a Qsystem >106over a much wider range of gap sizes. For the final design, the electric (E) field at the edge of the cavity (i.e. 50 mm from the centre) is ≈50 dB lower than the maximum (Emax). The E-field induces a magnetic field (B) that is maximal on the sample surface and is responsible for sample heating. The peak magnetic field on the sample (Bs,pk) is approximately 1.07 times higher than the peak magnetic field on the Choke Cavity. The corresponding field profiles are shown in figure 4. It should be noted that without a physical joint between the sample and Choke Cavity, and despite the presence of chokes, there is still some radiation (or RF leakage), represented by Qrad. If there are changes in leakage with each sample, Qrad can never be truly separated out, meaning that the end-plate replacement method is unsuitable for the Choke Cavity. For the RF-DC compensation method, only RF heating on the sample needs to be measured. Any RF leakage still present after the addition of chokes would not limit the accuracy of measurements and the ability to measure samples. It would only affect the maximum achievable Bs,pk. 3 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Figure 1. A model of the Choke Cavity showing the position of the chokes and sample. Figure 2. Qsystem vs choke 1 depth comparing the 1 choke, 2 choke and 3 choke models. 2.2. Optimisation of the gap size between the Choke Cavity and sample After design, the ideal size of the sample-cavity gap was adjusted to maximise the power dissipated on the sample to increase Bs,pk. Figure 5shows the effect of increasing the sample-cavity gap size on power dissipation in these three areas. A smaller gap size results in a higher percentage of power dissipation in the sample, allowing for higher values of Bs,pk. Leakage dominates when the size of the gap is >4.3 mm. A gap size of 1 mm was chosen, as this was easiest to maintain with spacers between the sample and the Choke Cavity without the risk of direct contact. The leakage is higher here than the gap sizes >3.5 mm because the Choke Cavity was originally designed for a larger gap; however, a smaller gap allows for optimisation of the sample losses and hence an increase in Bs,pk. With this configuration, approximately 34% of RF heating is on the sample. Based on post-manufacture gap tuning, any future Choke Cavity redesigns will be optimised for a 1 mm gap. Figure 3. A comparison of Qsystem vs of gap for the 1 choke, 2 choke and 3 choke models. 2.3. Optimisation of the pickup coupler location During the design of the cavity, the location of a pickup coupler had to be determined. Ideally, the pickup would have been inserted into the central elliptical cavity. However, given the small diameter of this section in order to accommodate the surrounding chokes, a pickup in the cavity would be too close to the input coupler, and thus cause unwanted crosstalk between the couplers. The second option was to insert the pickup coupler from the side in the gap between the sample and the cavity. However, this was considered a more complex mechanical solution with the added risk of having the coupler touch the sample or cavity due to the difficult position in a 1 mm gap. The remaining option to investigate was to drill a hole in one of the three chokes to have the pickup coupler inserted vertically. Figure 6shows simulated measurements of the % power dissipated on the Choke Cavity, sample and leakage with the 4 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Figure 4. (a) The magnitude of the E-field and (b) the magnitude of the H-field in the cavity and on the sample. Figure 5. Simulated results showing the % of RF power dissipation on the sample, cavity and leakage as a function of sample-cavity gap. pickup placed in each of the chokes compared to the original design without a pickup. The results show that a pickup coupler in choke 1 leads to a significant decrease in power dissipated on both the sample and the cavity due to an increase in leakage through the hole in choke 1. Having the pickup coupler in chokes 2 or 3 allows the cavity to perform similarly to a system without a pickup with minimal extra leakage. One issue with having the pickup in choke 2 or 3 is the potential to couple to additional eigenmodes within the system that are close to the cavity resonance. If the frequencies of two or more modes are too close, Fano resonances can arise, which would lead to greater uncertainties in RF measurements. Frequency mode simulations with the pickup in choke 2 or choke 3 showed that the additional modes were >30 MHz Figure 6. Simulated results showing the dependence of % of RF power dissipated on the sample, cavity and % of RF leakage on the position of the pickup coupler. from the cavity mode, which is much higher than the typical cavity BW (~ few kHz). Thus, it was okay to position a pickup in either choke without fear of coupling to the wrong mode. Furthermore, the transmitted signal was ≈20 dB higher in choke 2 than in choke 3. Therefore, the chosen position for the pickup was in a small hole drilled into the top of choke 2. 2.4. Final cavity design and manufacture The final cavity design consists of two parts and operates in a TM010 mode at f0=7.8 GHz [27]: •Choke Cavity: A half-cell elliptical cavity surrounded by quarter-wavelength RF chokes milled and machined from RRR =400 bulk Nb manufactured by Niowave Inc. [28]. 5 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Holes 3 mm in diameter are drilled into both the half-cell cavity and choke 2 for positioning of the couplers. The entire structure is 104 mm in diameter and 12 mm thick. The cavity received a 600 ◦C bake for 10 h followed by a light 20 µm BCP etch at INFN and 120 ◦C bake for 48 h. •Planar sample: A bulk Nb or TF SC coated copper disk separated from the Choke Cavity by a 1 mm gap. It can be 90– 130 mm in diameter and 1–10 mm thick. 3. Facility design 3.1. Cryogenic facility Testing planar samples with the Choke Cavity requires a bespoke cryogenic facility. Initially, a LHe cryostat [27,29] was used that conduction cooled the Choke Cavity and the sample via a LHe bath. However, due to the complexities in design and LHe requirements, this restricted the testing to one sample every two weeks. Given the requirement for high throughput testing, a dry system provided by a mechanical cryocooler then replaced this. The resulting LHe-free system is the result of several years of development since the initial design [30]. The main objective was to build a facility that would be easy and safe to operate with minimal training requirements. The facility was built around a Gifford-McMahon cryocooler (Sumitomo RDK-415D2) that uses a Sumitomo F-50H water-cooled compressor. This combination produces a maximum cooling power of 35 W at 50 K on stage 1 and 1.5 W at 4.2 K on stage 2. A schematic of the current iteration of this facility is shown in figure 7. The cryocooler is mounted on an upper flange of a stainless steel cryostat vacuum chamber. The cryocooler is thermally connected via Cu heat links (L1and L2) to two large oxygenfree high conductivity (OFHC) Cu plates on stage 1 and stage 2. These plates have many tapped holes to mount the electronics (i.e. thermometers, heaters, connectors, cable cooling bobbins, etc). Mechanically, the stage 1 plate is mounted with four G10 supports on the top flange, and the stage 2 plate is mounted with another four G10 supports on the stage 1 plate. G10 is used to provide good mechanical stability and low thermal conductivity. The stage 1 plate is connected to a Cu heat shield to reduce the radiative heat load on stage 2 (including the sample). The heat shield is also covered with multilayer thermal insulation (MLI) to further attenuate the incoming radiation from the vacuum chamber. The temperature of the stage 1 cold head (Ts1) and the heat shield are monitored with four platinum resistive (Pt 100) temperature sensors (T1a to T1d) connected to a Lake Shore 218 temperature monitor. The Choke Cavity is mounted directly to the underside of the stage 2 plate with two Cu supports that allow uniform cooling from the top and the edge of the cavity. Thermal simulations in CST show that the temperature gradient between stage 2 and the cavity is <0.1 K with this configuration. Two Lake Shore Cernox (CX-1050-CU-HT-1.4 L) temperature sensors (TCA and TCB) are used to monitor the temperature of stage 2, while a third Cernox (T2) is used to monitor the temperature of the stage 2 cold head (Ts2). Thermometers TCA and TCB, used in conjunction with three 10 Ωheaters mounted to stage 2, allow for control of the Choke Cavity temperature (Tcav) using a Lake Shore 331 temperature controller. The sample is mounted on top of a Cu sample holder that sits underneath the Choke cavity and is supported by three G10 studdings connected to an aluminium sample plate. In addition, 1 mm thick G10 spacers mounted between the sample and the Choke Cavity outside of choke 3 ensure that they are thermally and physically isolated while maintaining a constant sample-cavity gap of 1 mm. In this position, dielectric losses due to the spacers are negligible. An additional aluminium heat shield is mounted to the stage 2 plate around the sample and Choke Cavity to minimise radiation from stage 1. The cooling of the sample holder is provided by a separate heat link (L3) to the cold head to increase the thermal path between the sample and the cavity and thermally isolate them from each other. The sample holder is also equipped with two Cernox thermometers (TSA and TSB) and two 10 Ωheaters to control the sample temperature (Ts) using a second Lake Shore 331 temperature controller. The power dissipated by the sample and cavity heaters is measured by a 4-wire measurement. Measurements of heater current (Ih) and voltage (Vh) are made using two Multicomp Pro (MP730424) digital multimeters connected to each of the heater sets. Thus, the dissipated power, PDC, is given by PDC =Ih·Vh.(3) A sufficient amount of cooling power is needed throughout L3in order for the sample to reach Ts,min <4 K and also produce a detectable change in temperature with the applied heater power. A 1 K temperature increase was required for 1 W of power. In general, the cryogenic facility takes approximately 9 h to cool to a base temperature of Ts,min =3.6 K. 3.2. Vacuum system The cryostat vacuum system consists of a single vacuum vessel equipped with a 67 L s−1Pfeiffer (HiPace80) turbo-molecular pump (TMP) backed with a 6 m3s−1Edwards (nXDS6i) scroll pump (SP). Pressure control is achieved by vacuum valves and two Agilent (FRG-700) full-range Pirani inverted magnetron gauges. It takes approximately 4 h to pump the system down to ≈10−4mbar at which point the cryocooler is turned on (which can be controlled remotely). At base temperature, the pressure in the system is ≈10−7mbar due to cryopumping and remains stable throughout the duration of the experiment. At the end of the RF test, the facility is kept under vacuum and pumping until all parts are warmed to room temperature. A N2gas injection system allows venting to atmosphere. The cryostat vacuum chamber is equipped with a pressure relief valve to protect the system from overpressurisation during warming up and venting. In total, it takes around 30 mins to open the vacuum chamber and remove the heat shields to access the sample. It takes around the same time to re-mount the heat shields and close the vacuum chamber for pumping. 6 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Figure 7. A schematic of the LHe-free cryostat housing the Choke Cavity for the RF testing of planar samples. 3.3. RF system RF is input into the centre of the cavity via a short Cu-coated semi-rigid coax that acts as the input coupler. This is attached by a long stainless steel rod to a linear drive micrometer shown in figure 7. This allows for controlled vertical motion of the input coupler, and hence variable coupling to the cavity. A second, short, Cu-coated semi-rigid coax is clamped in a fixed position and inserted into choke 2. Coaxial cables from an SMA feedthrough at room temperature to an SMA feedthrough mounted at stage 1, as well as from stage 1 to stage 2, are made from stainless steel to minimise thermal loads between stages. Having a pickup coupler allows transmission (S21) measurements to be made alongside reflection measurements through the input coupler (S11). Given the wide bandwidth (BW) of this cavity (typically ~1 kHz at 4.2 K with a Nb sample), RF measurements can be made without the need for a phase-locked loop (PLL) or self-excited loop. A PLL/SEL is in development for future testing of alternative SC TFs where the bandwidth might be <1 kHz with the addition of an alternative SC TF coated Choke Cavity (e.g. Nb3Sn/Cu). As a result, a simple RF system is used, as shown in figure 8. In this system, the RF power is provided by a vector network analyser (VNA, Keysight P5024A) that provides a maximum output power of 12 dBm at 7.8 GHz. Power is input into an RF amplifier (Microwave Amps AM43-7.8S-35-43) through some fixed attenuators. Fixed attenuators are used to restrict the maximum input power at the cryostat to 1 W due to the RF health and safety limits—although this will be higher after moving to a radiation test bunker. A circulator is used at the 7 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Figure 8. A schematic of the simple RF system used with this facility. output of the amplifier to prevent amplifier reflections and damage. RF power is input into the system through the SMA feedthrough. The power transmitted from the cavity is measured through a 20 dB directional coupler using a power sensor (Keysight E9326A) connected to a power meter (Keysight E4416A). 3.4. Microphonics A common issue with SRF cavity testing is the effect of microphonics, typically due to pressure fluctuations within the LHe bath. For the 7.8 GHz bulk Nb Choke cavity with Rsclose to the theoretical resistance described by the Bardeen–Cooper– Schrieffer (BCS) theory [31], the minimum cavity BW is typically ~1 kHz therefore, any vibrations that cause frequency shifts greater than this will lead to greater uncertainties when measuring stored energy. This is ≈3 orders of magnitude higher than a typical SRF cavity; therefore, it was assumed that microphonics would not be an issue, allowing the simple RF system to be used without a PLL or SEL. However, it was soon shown that the level of microphonics is higher than that of LHe systems. This is thought to be mainly due to the long, movable stainless steel rod between the input coupler and the upper cryostat flange, which also couples to cryocooler noise generated by the periodic motion of components within the cryocooler. The level of microphonics was measured by recording the phase fluctuations of the |S21|trace (∠S21) over a sweep time of 10 s with the VNA centred at the resonant frequency, f0, with zero span. This is given by: ∠S21 =tan−1(2QL∆fshift f0).(4) A typical level of frequency shift (∆fshift) due to microphonics for the Choke Cavity is shown in figure 9. The maximum ∆fshift ≈1.5 kHz. This level of detuning occurs with a period of 1 s, matching the GM cryocooler operating frequency of 1 Hz. Therefore, while using the simple VNAbased RF system, care is taken to ensure that the input coupler depth is adjusted to give a cavity BW ⩾3 kHz to mitigate the effects of microphonics and ensure accurate measurements of the transmitted quality factor (Qt). For measurements of Qt, as discussed in appendix B, it should be noted that it is important that the intermediate frequency bandwidth (IF BW) on the VNA is set correctly to reduce noise and increase the accuracy of measurements. The IF BW, ultimately controls the measurement time and resolution level of a measurement. A lower 8 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Figure 18. Rs(Tcav)for Nb/Cu 1 with Ts=4.2 K. 6. Discussion The sample measurements made with the Choke Cavity during the commissioning phase demonstrate that this facility can be used to characterise and compare the performance of SC TFs under RF conditions. Measurements with multiple samples, in particular the Nb/Cu samples deposited at different temperatures, have shown that it is possible to differentiate between samples deposited under different conditions. This means that the Choke Cavity facility is suitable for mass parameter deposition studies and will be the main tool for optimising deposition processes in future studies. Results have shown that measurements of Rs(Ts) and Bs,pk(Ts) are possible and repeated tests of Nb/Cu samples have shown the ability to achieve consistent measurements between cooldowns. In addition, measurements of ∆f(Ts) have demonstrated the ability to estimate Tc. These measurements have also shown to be a useful indication of material properties such as the shift in penetration depth as well as the level of sample impurities and performance. The main objective in the near future will be to show whether there is any correlation between RF measurements and results from DC tests on the same samples with other facilities [9]. These include a DC magnetic field penetration facility [11] and a variable temperature insert to measure Tcand RRR [10]. A key advantage of using the Choke Cavity is that samples can still be measured independently of the quality of the cavity, i.e. the cavity surface resistance Rs,cav does not affect the measurements of sample Rs. This is because the RF losses on the sample are always measured independently of the Choke Cavity. However, a reduction in cavity quality affects the maximum achievable Bs,pk given that the ratio of RF power deposited on the sample to the cavity changes. This would then limit Rs,min as given in equation (16). Improvements in both Bs,pk and Rs,min are possible by increasing the input of RF power into the cavity or improving the quality of the choke Figure 19. Rs,min vs Rs,cav and Bs,pk vs Rs,cav for two levels of cavity RF power. cavity to lower Rs,cav. Figure 19 illustrates how these improvements can result in a decrease in Rs,min. One limitation of using a LHe-free system is that the maximum RF power is dictated by the cryocooler capacity. For this system, the limit is 1.5 W (though this is slightly reduced due to static heat loads). Powers >1 W will be achievable once this facility has been moved to a radiation-safe test bunker (scheduled by late 2024) and pulsed RF power will be investigated. Assuming a 1.5 W upper limit, with a bulk Nb cavity operating close to BCS resistance (i.e. from a full chemical and heat treatment), it is theoretically possible to increase the peak field to Bs,pk ≈11 mT, resulting in Rs,min ≈0.03 µΩ. Investigation into the use of TF coated Choke Cavities is ongoing. This will allow for slight gains at maximum power: a TF Nb coated Cu choke cavity could reach Bs,pk ≈14 mT and Rs,min ≈0.02 µΩ while a TF Nb3Sn/Cu cavity could reach Bs,pk ≈20 mT with Rs,min ≈0.01 µΩ. This shows the gains that can be achieved by increasing the RF power and improving the quality of the Choke Cavity. A comparison of the existing RF characterisation facilities at laboratories worldwide is shown in table 4. The facilities that use the end-plate replacement method (SLAC [13], Cornell [14] and IMP [15]) have simple mushroom-shaped cavity designs allowing for fairly quick sample turnovers. However, the sample and cavity need to be in direct contact (often with use of an indium seal) which risks cavity contamination during sample changeover, potentially changing the cavity Q0and affecting sample measurement accuracy. The RF-DC compensation method, compared to the endplate replacement method, allows a higher resolution down to sub-nΩ, mainly because a reference sample is not required. This technique is used by the commonly employed quadrupole resonator (QPR), at CERN [17,18], HZB [19] and DESY [20]. These facilities allow measurements at three quadrupole mode frequencies. The main disadvantage is the complex design of the sample holder. A similar measurement procedure can also be performed the SIC system at JLab [24]. 15 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Table 4. A comparison of RF test facilities for small samples. End-plate replacement RF-DC compensation SLAC Cornell IMP QPR SIC [13] [14] [15] [17–20] [24] Choke Cavity f0(GHz) 11.4 4.78/6.16 3.9 0.4/0.8/1.3 7.4 7.8 ∆Rs(µΩ) 100/10b— — <0.001 <1 0.5/0.01a Bs,pk,max (mT) 400 (4 K) 100 (2 K) — 120 (2 K) 14 (2 K) 3 (4.2 K) 20a(4.2 K) Cooling LHe LHe LHe LHe LHe Cryocooler Ts(K) >4 1.6–4.2 1.5–4.2 2–20 2–40 4–20 Sample shape Disk Disk Disk Cylinder Disk Disk Sample diam. (mm) 50–80 100 110 75 50 90–130 aWith upgrades. bCu cavity/Nb cavity. All existing facilities use LHe. This has advantages and disadvantages in comparison to a system based on a cryocooler. LHe based systems allow for the cooling of samples to temperatures down to at least 2 K and can operate at higher cooling capacities in the cryogenic system. However, this restricts the rate of sample turnover. In contrast, a cryocooler-based system is not dependent on LHe supply. It has a simpler design and fewer cryogenic health and safety risks. However, it is limited by the cooling power and restricted to Ts>4 K. The commissioning of the Choke Cavity facility has demonstrated the ability to accurately control the temperature to ±10K by using a cryocooler, which means that there are no fluctuations in pressure and temperature due to LHe. Stable temperatures from 4–20 K allow for accurate measurements of Rs(Ts) and ∆f(Ts) for alternative superconductors, e.g. Nb3Sn. It should be noted that the ultimate purpose of the Choke Cavity facility is to perform rapid sample comparison and TF optimisation under RF conditions. Given that 3 samples can be measured per week, this facility opens up the ability to perform mass parameter studies allowing for the understanding of quantities such as deposition temperature, sample thickness, magnetron power etc in a short space of time. Therefore, not being able to reach very high peak fields is acceptable. Once a TF deposition has been optimised, the process will be repeated on a QPR sample allowing low frequencies up to 1.3 GHz and high field analysis up to 120 mT under RF conditions [19]. A QPR sample typically takes longer to prepare and measure (1– 2 per month), which is why TF optimisation with the Choke Cavity facility with high turnover is such a useful and timesaving precursor step. Choke Cavity samples and QPR samples are planar geometries. The ultimate aim is to measure the RF properties of the TF 1.3 GHz cavities. To bridge the gap between planar and curved geometries, the Choke Cavity tests can be followed by split 6 GHz cavity tests. These cavities have the advantage of being able to visually inspect film quality as well as perform the RF test [41–43]. In this case, the TF depositions can also be performed using the same planar magnetron as Choke Cavity samples. Sample measurements with all mentioned RF and DC facilities will provide useful information for future 1.3 GHz cavity depositions. 7. Conclusions A facility based on a 7.8 GHz Choke Cavity has been designed and optimised for shape, number of chokes and gap size. It was then built and commissioned at Daresbury Laboratory. This facility is able to test planar thin film coated samples with a diameter of 90–130 mm and thickness 1–10 mm under RF conditions. The testing of bulk Nb and thin film Nb coated Cu samples has shown that this facility is able to measure the average surface resistance, Rs⩾0.5µΩ, with sample temperatures, Ts>4 K. In addition, at fixed Ts, samples in this facility can be measured with a range of peak sample magnetic fields, Bs,pk ⩽3.0 mT with potential for future improvements up to a maximum of 20 mT. The main advantage of this facility over others is its simplicity in design and operation. Given the simple sample design, mounting procedure, and use of a LHe-free cryostat, the facility can achieve fast sample changeovers allowing for a very high sample throughput of up to 3 per week. Thus, this is a very effective facility for quickly providing low power RF evaluation during mass parameter optimisation of superconducting thin films prior to tests in other facilities with lower RF frequencies, lower operating temperature, and higher RF power. Data availability statement All data that support the findings of this study are included within the article (and any supplementary files). Acknowledgments This work has been supported by: the IFAST collaboration which has received funding from the European Union’s Horizon 2020 Research and Innovation programme under Grant Agreement No. 101004730. D Seal would like to thank UKRI for his PhD funding and STFC for the Cockcroft Institute core grant. The authors wish to acknowledge the support received by members of ASTeC, namely J Conlon, A Vick, S Hitchen, A Blackett-May, A Palmer, A Wootten, K Dumbell, S Wilde, C Benjamin, L Smith, A Hannah. 16 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al The authors also acknowledge the work carried out by S Bibby-Trevor, R McAllister and the team at the Engineering Technology Centre at Daresbury Laboratory and the polishing of Cu disks carried out by the team at STFC RAL Space. The authors thank G Ciovati for sharing the WinSuperFit code. Appendix A. Rsderivation The RF power dissipated on the sample (PRF) is defined as PRF =1 2µ2 0ˆsample Rs|B|2dS=1 2µ2 0 Rsˆsample |B|2dS,(A.1) where µ0is the vacuum permeability, Bis the magnetic field strength, and it is assumed that Rsis constant over the sample surface. A rearrangement of equation (A.1) gives: Rs=2µ2 0PRF ´sample |B|2dS.=c1µ2 0PRF B2 s,pk .(A.2) The surface integral in equation (A.2) cannot be determined experimentally, therefore a constant has to be defined to calculate Rs. This constant, c1, is defined as: c1=2B2 s,pk ´Sample |B|2dS=B2 s,pk,CSTRS,CST µ2 0Ps,CST =2.63 ×103m−2, (A.3) where equation (A.2) and calculated values of the magnetic field (Bs,pk,CST), power dissipated on the sample (Ps,CST) and surface resistance (Rs,CST), normalised for a stored energy of 1 J, have been used to determine a value of c1. These values can easily be calculated using eigenmode simulations in CST [25]. Thus a substituting the integral in equation (A.3) into equation (A.2) gives Rs=c1µ2 0PRF B2 s,pk .(A.4) The peak magnetic field on the sample surface (Bs,pk) is defined as Bs,pk =Bs,pk,CST ×√U,(A.5) where Uis the stored energy. Using the definition for the quality factor of the pickup coupler (Qt), Qt≡ω0U Pt ,(A.6) Equation (A.5) can be rearranged to give Bs,pk =Bs,pk,CST (QtPt ω0)1 2 .(A.7) Appendix B. Qtcalibration and uncertainty Measurements of stored energy (U) in a 2-port system can be made directly from measurements of Pt. This requires a calibration of Qtdefined by: Qt=ω0U Pt (B.1) where ω0is the resonant angular frequency. It can be assumed that Qtis constant throughout a sample measurement cycle (detailed in section 4.3). Only Ptvaries between measurements, resulting in a change in U. For a 2-port system, transmission measurements are possible with the VNA. Given the wide bandwidth of the cavity, it is possible to calculate Qtdirectly from the VNA. In this case, the input coupling factor is given by β1with external quality factor Qe. The pickup coupling factor is given by β2with the external quality factor Qt. The coupling factors are defined as β1≡Qsys Qe =Pr Pc ,(B.2) and β2≡Qsys Qt =Pt Pc ,(B.3) where Pris the reflected power through the input, Ptis the transmitted power through the pickup coupler, Ptand Pcis the total power dissipated on the choke cavity, sample and leakage. Qsys is the quality factor accounting for the Pclosses. Qtcan be calculated from equation (B.3) provided that Qsys is calculated. It is defined as 1 QL =1 Qsys +1 Qe +1 Qt ,(B.4) where QLis the loaded quality factor. Using equations (B.2)–(B.4) can be rewritten as Qsys =QL(1+β1+β2).(B.5) The three variables in equation (B.5) can be calculated from the measurements of |S21|and |S11|which are defined by relating the measurements Prand Ptto the forward power into the cavity (Pf): The input reflection parameter (|S11|) is defined as |S11|2≡Pr Pf (B.6) and the forward transmission parameter (|S21|) as |S21|2≡Pt Pf .(B.7) 17 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al Firstly, a measurement of QLcan be made from the 3 dB bandwidth (∆f3dB) of the |S21|trace: QL=f0 ∆f3dB .(B.8) Next, to find an equation in terms of S-parameters for β2 one starts by noting that the total power is conserved, that is. Pf=Pc+Pr+Pt.(B.9) Substituting equation (B.9) into the inverse of (B.7) gives 1 |S21|2=Pc Pt +Pr Pt +Pt Pt =1 β2 +|S11|2 |S21|2+1,(B.10) where equations (B.3), (B.6) and (B.7) have been used. This can be rearranged to give an equation for β2in terms of S-parameters only: β2=|S21,cav|2 1−|S11,min|2−|S21,cav|2,(B.11) where |S11,min|denotes the value of |S11|measured at f0and |S21,cav|is the transmission loss measured at the cavity. This is to account for cable losses. In order to derive an equation for β1, one starts by noting that the input coupler cannot distinguish Ptfrom Pc. A new coupling coefficient, β∗is defined: β∗=Pr Pc+Pt =1/Qe 1/Qsys +1/Qt =β1 1+β2 ,(B.12) β∗is calculated from the defined reflection parameter (Γ), where Γ≡β∗−1 β∗+1.(B.13) Using this, it is possible to relate Prto the forward power (Pf) Pr= Γ2Pf.(B.14) Substituting equation (B.13) in (B.13), rearranging and taking the square root gives an equation for β∗ β∗= 1±√Pr Pf 1∓√Pr Pf =1±|S11| 1∓|S11|.(B.15) Equating equations (B.13) and (B.15) and accounting for cable losses give an equation for β1in terms of S-parameters only: β1=|S11,max|±|S11,min| |S11,max|∓|S11,min|·(1+|S21,cav|2 1−|S11,min|2−|S21,cav|2), (B.16) where |S11,max|which is the maximum off-band measurement of the |S11|trace to account for cable loss. Furthermore, the sign changes in equations (B.15) and (B.16) depend on whether the coupler is overcoupled (+/-) or undercoupled (- /+), which can be inferred from the polar plot of the S11 trace. Therefore, from equations (B.3), (B.5), (B.11) and (B.16), it is possible to calibrate Qtusing the VNA. ORCID iDs D Seal https://orcid.org/0000-0003-2871-8690 O B Malyshev https://orcid.org/0000-0001-9345-3225 H Marks https://orcid.org/0000-0001-6550-3854 C Pira https://orcid.org/0000-0002-5893-1567 E Chyhyrynets https://orcid.org/0000-0002-8167-3957 G Burt https://orcid.org/0000-0001-8288-1216 References [1] Poole C 2000 Handbook of Superconductivity (Academic) [2] Padamsee H, Knobloch J and Hays T 1998 RF Superconductivity for Accelerators (Wiley) [3] Bafia D, Grassellino A, Sung Z, Romanenko A, Melnychuk O S and Zasadzinski J F 2019 Gradients of 50 MV/m in TESLA shaped cavities via modified low temperature bake Proc. 19th Int. Conf. on RF Superconductivity (SRF’19) pp 586–91 [4] Benvenuti C, Circelli N and Hauer M 1984 Niobium films for superconducting accelerating cavities Appl. Phys. Lett. 45 583–4 [5] Miyazaki A and Delsolaro W V 2019 Two different origins of the Q-slope problem in superconducting niobium film cavities for a heavy ion accelerator at CERN Phys. Rev. Accel. Beams 22 073101 [6] Valente-Feliciano A M et al 2022 Next-generation superconducting RF technology based on advanced thin film technologies and innovative materials for accelerator enhanced performance and energy reach (arXiv:2204.02536) [7] Dhakal P 2020 Nitrogen doping and infusion in SRF cavities: a review Phys. Open 5100034 [8] Gurevich A 2006 Enhancement of RF breakdown field of superconductors by multilayer coating Appl. Phys. Lett. 88 012511 [9] Seal D et al 2023 Characterisation facilities for evaluating superconducting thin films for SRF cavities Proc. 14th Int. Particle Accelerator Conf. (IPAC’23) pp 2983–6 [10] Malyshev O B, Bizel-Bizellot L, Dumbell K, Goudket P, Pattalwar N, Pattalwar S, Pizzol P, Smith P A, Valizadeh R and Wilde S 2018 Design, assembly and commissioning of a new cryogenic facility for complex superconducting thin film testing Proc.9th Int. Particle Accelerator Conf. (IPAC’18) pp 3859–61 [11] Turner D A, Malyshev O B, Burt G, Junginger T, Valizadeh R and Gurran L 2022 A facility for the characterisation of 18 Supercond. Sci. Technol. 37 (2024) 115012 D Seal et al planar multilayer structures with preliminary niobium results Supercond. Sci. Technol. 35 095004 [12] Goudket P, Junginger T and Xiao B P 2016 Devices for SRF material characterization Supercond. Sci. Technol. 30 013001 [13] Nantista C, Tantawi S, Weisend J, Siemann R, Dolgashev V and Campisi I 2005 Test bed for superconducting materials Proc. 2005 Particle Accelerator Conf. (PAC’05) pp 4227–9 [14] Xie Y, Hinnefield J and Liepe M 2009 Design of a TE-type cavity for testing superconducting material samples Proc. 14th Int. Conf. on RF Superconductivity (SRF’17) pp 281–5 [15] Huang S, He Y, Tan T and Zhang S 2018 The development of TE-sample host cavity at IMP Proc. 29th Int. Linear Accelerator Conf. (LINAC’18) pp 852–4 [16] Junginger T 2012 Investigations of the surface resistance of superconducting materials PhD Thesis University of Heidelberg [17] Chiaveri E, Haebel E, Mahner E and Tessier J M 1998 The quadrupole resonator, construction, RF system, field calculations and first applications Proc. 6th European Particle Accelerator Conf. (EPAC’98) pp 1852–4 [18] del Pozo Romano V, Gerigk F, Betemps R, Illan Fiastre R and Mikkola T P 2018 Redesign of CERN’s quadrupole resonator for testing of superconducting samples Proc. 18th Int. Conf. on RF Superconductivity (SRF’17) pp 420–2 [19] Keckert S, Kleindienst R, Kugeler O, Tikhonov D and Knobloch J 2021 Characterizing materials for superconducting radiofrequency applications - a comprehensive overview of the quadrupole resonator design and measurement capabilities Rev. Sci. Instrum. 92 064710 [20] Monroy-Villa R, Lemke M, Reschke D, Röhling M, Thie J H, Hillert W, Wenskat M, Gorgi Zadeh S and Putek P 2022 Status of the new quadrupole resonator for SRF R&D Proc. 20th Int. Conf. on RF Superconductivity (SRF’21) pp 751–5 [21] Monroy-Villa R, Gossel A, Reschke D, Rohling M, Schmokel M, Thie J-H, Wiencek M, Hillert W, Martens C and Wenskat M 2023 Commissioning of the UHH quadrupole resonator at DESY Proc. 21th Int. Conf. on RF Superconductivity (SRF’23) pp 952–9 [22] Wilde S 2018 Development of superconducting thin films for use in SRF cavity applications PhD Thesis Loughborough University [23] Shintake T 1992 The choke mode cavity Jpn. J. Appl. Phys. 31 1567–70 [24] Xiao B, Reece C E, Phillips H L, Geng R L, Wang H, Marhauser F and Kelley M J 2011 Note: radio frequency surface impedance characterization system for supercon-ducting samples at 7.5 GHz Rev. Sci. Instrum. 82 056104 [25] CST Studio Suite (available at: www.3ds.com/productsservices/simulia/products/cst-studio-suite/) [26] Liepe M (Cornell University: Liepe group research) 2012 SRIMP (available at: https://www.classe.cornell.edu/~liepe/ webpage/researchsrimp.html) (Accessed 27 September 2024) [27] Goudket P, Pattalwar S, Malyshev O B, Gurran L, Burt G, Jordan E S, Jones T J, Malyshev D O and Valizadeh R 2015 Test cavity and cryostat for SRF thin film evaluation Proc. 6th Int. Particle Accelerator Conf. (IPAC’15) pp 3232–4 [28] Niowave Inc (available at: www.niowaveinc.com/) [29] Goudket P, Malyshev O B, Pattalwar S, Pattalwar N, Gurran L, Valizadeh R and Bizellot L B 2018 First full cryogenic test of the SRF thin film test cavity Proc. 18th Int. Conf. on RF Superconductivity (SRF ‘17) pp 644–6 [30] Malyshev O B, Goudket P, Pattalwar N, Pattalwar S, Conlon J and Burt G 2019 The SRF thin film test facility in LHe-free cryostat Proc. 19th Int. Conf. on RF Superconductivity (SRF ‘19) pp 610–2 [31] Bardeen J, Cooper L N and Schrieffer J R 1957 Theory of superconductivity Phys. Rev. 108 1175–204 [32] Turneaure J P, Halbritter J and Schwettman H A 1991 The surface impedance of superconductors and normal conductors: the Mattis-Bardeen theory J. Supercond. 4341–55 [33] Ciovati G 2005 Investigation of the superconducting properties of niobium radio-frequency cavities PhD Thesis Old Dominion University [34] Langley B W, Anlage S M, Pease R F W and Beasley M R 1991 Magnetic penetration depth measurements of superconducting thin films by a microstrip resonator technique Rev. Sci. Instrum. 62 1801–12 [35] Miyazaki A and Venturini Delsolaro W 2019 Determination of the bardeen-cooper-schrieffer material parameters of the hie-isolde superconducting resonator Supercond. Sci. Technol. 32 025002 [36] Ciovati G and Xiao B 2010 Winsuperfit version 2.0 [37] Halbritter J 1970 Fortran-program for the computation of the surface impedance of superconductors [38] Posen S, Romanenko A, Grassellino A, Melnychuk O S and Sergatskov D A 2020 Ultralow surface resistance via vacuum heat treatment of superconducting radio-frequency cavities Phys. Rev. Appl. 13 014024 [39] Singer W, Ermakov A and Singer X 2010 DESY RRR-measurement techniques on high purity niobium Technical Report, TTC-Report 2010-02 [40] Malyshev O B, Gurran L, Goudket P, Marinov K, Wilde S, Valizadeh R and Burt G 2017 RF surface resistance study of non-evaporable getter coatings Nucl. Instrum. Methods Phys. Res. A844 99–107 [41] Sian T, Burt G, Seal D, Marks H, Valizadeh R and Malyshev O B 2022 Split thin film SRF 6 GHz cavities Proc. 31st Int. Linear Accelerator Conf. (LINAC’22) pp 814–7 [42] Sian T, Conlon J, Malyshev O B, Valizadeh R, Burt G, Leicester N, Marks H, Seal D, Chyhyrynets E and Pira C 2023 Split 6 GHz SRF thin film cavities Proc. 14th Int. Particle Accelerator Conf. (IPAC’23) pp 3074–7 [43] Leicester N, Burt G, Marks H, Seal D, Conlon J, Malyshev O B, Sian T, Valizadeh R, Chyhyrynets E and Pira C 2023 Development and testing of split 6 GHz cavities with niobium coatings Proc. 21st Int. Conf. on RF Superconductivity (SRF’23) pp 51–55 19