Preamble

Experimental Study on Online Whole Machine Demagnetization of High-Q SHINE Modules Yong-zhou He, Jian Dong, Lijun Lu, Zheng Li, Chen Luo, Hongtao Hou, Xiaowei Wu, Sen sun, Yiyong Liu, and Lixin Yin 1 Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai, 201204, China Zhangjiang Laboratory, Shanghai, 201204, China Cryomodules are key systems in SHINE. To achieve a high quality factor for the superconducting cav- ities in cryomodules, it is essential to provide a weak magnetic field operating environment. Efficient mag- netic shielding for superconducting cavities effectively shields against background magnetic fields; excessive background magnetic fields significantly reduce the effectiveness of cryogenic Permalloy magnetic shielding.

During manufacturing, installation, integration, testing, and transportation, cryomodules acquire residual mag- netic fields of varying degrees, necessitating online whole machine precision demagnetization of the entire cryomodule to reduce the background magnetic field for magnetic shielding. Taking the SHINE cryomodule as an example, this paper presents the procedure and experimental results. The study shows that after online demagnetization of the entire SHINE module, the magnetic field inside the superconducting cavity is less than 1.0 mGs when the superconducting cavity reaches . These results provide a reference for the demagnetization of cryomodules in similar free-electron laser facilities worldwide.

Keywords

cryomodule, whole machine demagnetization, weak magnetic field, quality factor

INTRODUCTION

The Shanghai high-repetition-rate XFEL and extreme light facility (SHINE) is designed to operate 75 large-scale cryo-

genic superconducting accelerator modules, incorporating 4

600 standard 1.3 GHz nine-cell elliptical superconducting 5

cavities [ Achieving the design quality factor at 16 MV/m for these cavities, reducing the ther-

mal load on the SHINE cryogenic system, and lowering costs 8

in construction, equipment development, integration testing, and operational maintenance require a weak magnetic field operating environment for the cavities [ ]. A lower resid- ual magnetic flux density (hereafter referred to as the residual magnetic field ) inside the superconducting cavi- ties leads to a higher ]. Previous studies indicate that when the residual magnetic field inside the cavity is reduced to about 3.00 mGs, the cavity reaches further reducing to 0.50 mGs allows the to approach According to the SHINE high- module full-machine on-

line demagnetization technical standard, after online demag- 20

netization of a standard 1.3 GHz module, the magnetic field readings from the two fluxgate probes located between the magnetic shielding layers shall satisfy the following require- ments:

1. At the ground-level horizontal test position, the average

absolute value of the two fluxgate probes shall be below 0.60 mGs.

2. At the underground tunnel operation position, the aver-

age absolute value shall be below 1.00 mGs. The authors would like to express their gratitude to the SHINE module development team for their support in experimental test.

The weak magnetic field environment required for super- conducting cavities is critically dependent on the effective- ness of magnetic shielding. Such shielding is typically con- structed from high-permeability magnetic materials—often formed into geometric shapes such as spherical rings, frames, or cylinders—to divert ambient magnetic field lines and es- tablish a magnetically clean background, as illustrated in . The underlying principle of low-frequency static magnetic shielding is to employ high-permeability materials such as pure iron, silicon steel, and Permalloy to provide a low-reluctance path for interfering magnetic fields, thereby substantially reducing the residual field in the equipment’s working volume [ ]. Based on the principles of static magnetic shielding, two key conclusions can be drawn: first, the greater the permeability of the ferromagnetic mate- rial and the larger the cross-sectional area of the magnetic shield, the lower the magnetic reluctance of the mag- netic circuit, which results in a greater fraction of the mag- netic flux being confined within the shielding material and

significantly reduces leakage flux in the air; second, in mag- 49

netic shields made of soft magnetic materials, openings or 50

gaps oriented perpendicular to the magnetic field lines should be avoided since such gaps force the field lines to traverse air, increasing the magnetic reluctance and thus degrading the shielding effectiveness.

The permeability µ of cryogenic Permalloy used in super- 55

conducting cavity magnetic shielding exhibits strong depen- dence on the background magnetic field . Experimental characterization demonstrates that optimal shielding perfor- mance is achieved when the background field is maintained in the range of 0–50 mGs (see Fig. ), where the ma- terial maintains high permeability. Consequently, establish- ing a weak magnetic environment (low ) that matches the permeability characteristics ( ) becomes essential for maxi- mizing shielding effectiveness. This relationship reveals that the magnetic shield surrounding the superconducting cavity alone provides insufficient protection against stronger ambi- ent fields.

Therefore, creating a further reduced magnetic

field environment external to the shield itself is necessary to establish the fundamental conditions required for achieving high shielding performance [

(b) Relationship between the permeability of cryogenic permalloy and the background magnetic field H .

The background magnetic field environment surrounding the superconducting cavity magnetic shielding is influenced by several key factors:

1. Inherent geomagnetic field orientation: The overall

module orientation creates a directional dependence in the background field, with the east–west orientation exhibiting

significantly lower field strength compared to the north–south 77

direction. SHINE cryomodules are aligned east–west at the ground-level test station but oriented north–south during un- derground tunnel installation and operation.

2. Cryomodule magnetic shielding design: The SHINE

1.3 GHz cryomodule employs a double-layer magnetic shield exterior to the helium vessel, constructed from Chinese-

developed cryogenic permalloy 1JL0. This material achieves 84

a relative permeability µ r > 30 , 000 at temperatures ≤ 77 K, 85

with a sheet thickness of 1.3 mm.

3. Building magnetic shielding effectiveness: The mag-

netic shielding performance of test facilities varies signif- 88

icantly. In Hall 1 of the SHINE ground test station, two

test stations provide moderate shielding, maintaining an ax- 90

ial background field of 150–250 mGs. Hall 2 test stations offer poorer shielding, with axial fields of 250–350 mGs.

The underground north–south tunnel, constructed with rein- forced concrete, exhibits the highest background field of 400– 550 mGs at module installation positions.

4. Magnetic hygiene control: Strict magnetic hygiene pro-

tocols were implemented throughout manufacturing, assem- bly, and testing. All components within 100 mm of super- conducting cavities (including couplers, tuners, flanges, and bellows) were fabricated from non-magnetic materials (316

stainless steel, titanium alloy, or aluminum alloy) with per- 101

meability µ r < 1 . 1 and remanent field < 1 Gs. Assembly 102

tools and fixtures were controlled to < 2 . 5 Gs remanent field, 103

with mandatory demagnetization when exceeded. The verti- cal test background for bare cavities was maintained below 2.0 mGs [

5. Whole machine demagnetization precision: Despite

strict magnetic control during cryostat installation, subse- quent processes (integration, debugging, welding, and con-

ditioning) introduce magnetization. Online whole machine 110

demagnetization is therefore essential to reduce the cryostat’s background field environment [ The alternating current (AC) demagnetization method em- ploys a gradually decaying AC current to eliminate residual magnetism in ferromagnetic materials. As the current ampli- tude decreases cyclically, the material’s hysteretic behavior drives the magnetic domains through progressively smaller loops, effectively reducing the remanent field to near zero when the current approaches zero [ ]. This well-

established technique has found extensive application in en- 120

gineering domains including ship degaussing and pipeline de-

magnetization [ 26 , 27 ]. Significant advancements in super- 122

conducting cryomodule demagnetization have been achieved internationally. For the IFMIF project, CEA-Saclay imple- mented a comprehensive magnetic hygiene strategy involv- ing replacement of magnetic components with non-magnetic alternatives and offline demagnetization of susceptible com-

ponents to minimize internal magnetic fields [ 28 ]. At FRIB, 128

validation of the local magnetic shielding scheme demon- strated attenuation of the cavity wall field below 15 mGs while achieving cost reduction through material optimization.

Their research further established that controlled cooldown conditions eliminate the need for additional demagnetization after solenoid excitation cycles [ ]. Fermilab researchers have developed a systematic theoretical framework for whole machine demagnetization and successfully implemented it on LCLS-II cryomodules, achieving residual fields below

2 mGs [ 30 , 31 ]. In China, research at Beihang University has 138

focused on low-frequency demagnetization techniques for 139

magnetic shielding rooms, resulting in proposed methods for estimating shielding coefficients and residual magnetism [ However, systematic investigation of whole machine demagnetization for large-scale accelerator superconducting modules remains underdeveloped domestically. Bridging this technological gap is essential for achieving world-leading performance in superconducting accelerator module technol- ogy. This paper presents experimental results from online whole machine demagnetization of high- modules and mul- tiple engineering cryomodules, demonstrating the creation of ultra-low magnetic field environments necessary for optimal superconducting cavity performance.

EXPERIMENTAL METHOD The demagnetization of the SHINE cryomodule essentially involves demagnetizing the vacuum cryostat [ ]. Com- ponents such as the tuner, coupler, flange, and support pieces around the superconducting cavities inside the module cannot have their residual magnetic fields reduced through whole- module demagnetization [ ]. As previously described, these components can only be managed through material se- lection and individual component demagnetization. There- fore, the demagnetization system and process for the entire module are almost identical to those for the cryostat. The SHINE module whole machine demagnetization uses the de- magnetization system IPS-403E jointly developed with Hu-

nan United Information Security Technology Co., Ltd. The 165

demagnetization system for the whole module consists of two parts: the demagnetization power supply system and the load coil system. The demagnetization power supply system has a total power of about 33 kW, with a demagnetization output current ranging from 6.5 mA to 130 A and a demagnetization frequency of 0.1–1.0 Hz. The switching frequency is switch- able between 5 kHz and 10 kHz. It includes an industrial computer, a low-frequency power supply, and a resistor box.

The load coil system has a total length of about 1300 m and a total resistance of about 2.0 when connected in series.

It uses 12 mm lightweight, high-temperature, and radiation- resistant cables. It includes a 100-turn coil at the module end, 20 groups of coils in the middle (10 turns per group), and output/input leads.

During the SHINE module whole-machine demagnetiza- tion process, operations such as vacuum pumping, heat-

ing, RF conditioning, and welding—all of which can in- 182

terfere with demagnetization to varying degrees—must be avoided [ ]. The values of individual or groups of eight superconducting cavities are measured before and af- ter demagnetization using the flowmeter-small range method, and the results are compared with vertical test data. The phys- ical setup of the SHINE cryomodule prepared for demagne-

tization is shown in Fig. 2 FIGURE:2 .To monitor magnetic field tran- 189

sients and final demagnetization results during online demag- netization of the SHINE high- module, 13 Bartington In- struments fluxgate probes were installed at strategic positions inside the module [ ], as illustrated in Fig. . The fluxgate probes were distributed as follows: CH1, CH12, and CH2 were placed between the two magnetic shielding layers

outside the titanium jackets of the first, fourth, and fifth cavi- 196

ties, respectively; CH4 was mounted horizontally atop Cell 5 of Cavity 1; in Cavity 4, CH6 and CH7 were installed at a angle on top of Cell 9 and Cell 1, respectively, while CH13 and CH14 were positioned vertically along the axis at the bot- tom of Cell 9 and Cell 1; for Cavity 5, CH8 was placed at a angle on Cell 9, with CH9 arranged vertically at the bot- tom of the same cell; finally, in Cavity 8, CH10 and CH11 were both oriented at on Cell 9 and Cell 1, respectively, ture more extreme field conditions in this experiment, all in- ternal fluxgate probes except CH4 were located at both ends of the magnetic shield—regions identified by simulation as having the strongest magnetic field. Simulation results indi- cate that if the field at the shield ends meets expectations, the

field uniformity in the central region will be even more favor- 211

ton fluxgate probes were installed between the two magnetic

shielding layers outside the titanium jackets of the first and 214

fifth cavities to monitor field variations and demagnetization 215

effectiveness during online operation. Unlike the demagnetization of cryomodules under well- insulated and isolated conditions, the online whole-machine demagnetization of the entire SHINE module–in both the horizontal test station and underground installation states– requires numerous signal cables for data acquisition. Addi- tionally, the conductive and magnetic support platform lacks insulation design, resulting in extensive leakage currents at test and installation sites. These leakage currents cause un- predictable interference during the demagnetization process.

Theoretically, disconnecting the leakage current grounding path before demagnetization could prevent demagnetization interruption. However, this approach is impractical in opera- tion, as it would compromise the safety of the module, equip- ment, and personnel. In early 2024, online demagnetization of the SHINE high- prototype module used a non-timed balance mode, which frequently triggered the leakage cur- rent protection, causing unexpected interruptions. This mode proved unsuitable for general cryomodule demagnetization.

To address this, subsequent demagnetization adopted two al- ternative modes–timed exponential decay (TED) and timed linear decay (TLD)–that are less likely to trigger protection.

The low-frequency power supply switching frequency was

also reduced to minimize interruptions. The TED mode cur- 239

rent waveform follows:

I t = I a � K 2 f � ( t − 1 4 f ) sin(2 πft ) + I dc (1) 241

where is the real-time current, is the amplitude of the

initial peak current, K is the decay coefficient (fixed at 0.95 in 243

this mode), and is the DC offset. The output termination condition is met when the total operating time reaches the preset duration . The current waveform of TLD mode is

determined by the following equation: sin(2

I t = I a

The current waveforms for the TED and TLD modes dur- ing SHINE high- module demagnetization are shown in , respectively. The corresponding mag- netic field waveforms measured by two fluxgate probes lo- cated outside the superconducting cavity are presented in . Analysis of these waveforms reveals distinct de- cay characteristics: in the TED mode, both current and mag- netic field exhibit a sharp exponential decay after reaching

the initial peak, whereas in the TLD mode, they demonstrate 257

a gradual linear decay following the initial maximum. 258

During the testing of the cryomodule, the cooldown procedure was executed as follows.

The system was first cooled from 300 K to 160 K at a controlled rate of 4–7 K/h.

Subsequently, the temperature was further reduced to 45 K at a rate of 10–20 K/h. A 12-hour holding period was imple- mented at 45 K to ensure temperature stabilization. Follow-

ing this, a rapid cooldown phase with a rate ≥

10 K/min was

265

initiated to reach 4.5 K. Finally, the saturated helium vapor 266

pressure was reduced to 31 mbar to establish the operational state of superfluid helium at 2 K. The accelerating gradient of the superconducting cavity is determined by the fol- lowing equation:

E acc = 1 L eff · � r Q · Q t · P t (3) 271

where refers to the effective accelerating length of the cavity, denotes the shunt impedance, is the quality factor derived from the pick-up probe, and is the power measured at the pick-up port. The accelerating gradient can be calculated following the calibration of and the mea- surement of . A direct comparison of the intrinsic quality factor between vertical and horizontal tests is ensured by evaluating them at an identical EXPERIMENTAL RESULT summarizes the demagnetization status of the SHINE high- prototype module across different worksta- tions at the installation and test site, including position, de- magnetization current, temperature, and orientation. Fig. shows the magnetic field measurements inside and outside the superconducting cavities before and after demagnetiza- tion of the prototype module. For the three external flux- gate probes (CH1, CH12, CH2) at the Hall 1 test station, the magnetic fields before 60 A demagnetization were 0.51 mGs, 2.76 mGs, and 7.59 mGs, respectively.

After demagneti- zation, these values decreased to 0.14 mGs, 0.43 mGs, 1.09 mGs.

After cooling the module to 2.0 K, the fields increased slightly but remained lower than the pre- demagnetization levels.

When the module was moved to a north–south orientation in Hall 1, the magnetic fields at TLD mode and (c) magnetic field changes during the demagneti- zation process (from left to right in the figure, the four magnetic field waveforms correspond to: 30A/TED mode, 75A/TED mode, 30A/TLD mode, 75A/TLD mode).

these probes increased significantly, except for CH1 which 296

provided faulty data. Relocating the module to an east– west position away from the Hall 1 test station reduced the fields compared to the north–south orientation; however, due to environmental field differences from the position change, they remained higher than immediately after demagnetiza- tion. Moving the module from ground to the north–south oriented underground tunnel again increased the fields sig-

nificantly, though demagnetization again reduced them. Fi- 304

nally, after moving the module from the underground tunnel to a ground-level east–west position, the magnetic fields de- creased dramatically following demagnetization, showing a substantial difference compared to pre-demagnetization val- ues. For the ten internal fluxgate probes at the Hall 1 test sta- tion, the magnetic fields measured at CH6, CH7, and CH15 exceeded 5.0 mGs prior to demagnetization at 60 A. Af- ter demagnetization, these values decreased to 1.16 mGs, 0.22 mGs, and 0.14 mGs, respectively, with the remain- ing probes also exhibiting reduced field levels. Following cooldown to 2 K, the magnetic fields experienced a slight overall increase but remained below pre-demagnetization val- ues. When the module was reoriented to a north–south align-

ment within Hall 1, nearly all probes showed a significant 318

increase in magnetic field strength. Subsequent relocation of the module to an east–west position away from the Hall 1 test station resulted in lower field readings relative to the north– south orientation; however, due to differences in the ambient magnetic environment, field levels remained elevated com- pared to the immediately post-demagnetized state. Transfer- ring the module from ground level to the north–south ori- ented underground tunnel again produced a substantial field increase across most probes. Although demagnetization at 75 A again reduced field strengths, certain probes continued to register fields exceeding 5.0 mGs. Finally, after moving the module from the underground tunnel to an east–west ground- level position and performing 75 A demagnetization, inter- nal magnetic fields decreased dramatically, with all cavity- internal fluxgate probes measuring below 1.0 mGs. These results confirm the efficacy of the demagnetization process in reducing magnetic fields both inside and outside the su- perconducting cavities, successfully establishing the required weak-field environment for the SHINE high- module.

From the results presented above, the following observa- tions can be made: 1. The full-module demagnetization sig-

nificantly influences the magnetic fields both inside and out- 340

side the cavities. Importantly, the module should not be re- located after demagnetization, as substantial variations in the ambient background field can compromise the demagnetiza- tion effect. 2. The temperature reduction following demagne- tization induces non-equilibrium thermal currents and related effects, resulting in a slight increase in the magnetic field.

3. The overall demagnetization efficacy is strongly correlated

with the module orientation. For identical demagnetization currents and process parameters, the north–south orientation– characterized by a higher background field–exhibits less ef- fective demagnetization compared to the east–west orienta- probes inside and outside the cavities show general consis- tency before and after demagnetization.

As in-situ probe placement within the cavity is infeasible during operation,

external probes can serve to monitor and track internal field 356

changes post-demagnetization, offering valuable reference for operational maintenance.

Prior to demagnetization, the average magnetic field values measured by the two fluxgate probes (CH12 and CH2) be- tween magnetic shielding layers were 5.56 mGs in the east– west direction at the Hall 1 test station, 16.32 mGs in the east–west direction at ground level at Well 2, and 9.42 mGs in the north–south direction at the Well 2 tunnel entrance.

The ten fluxgate probes (CH4, CH6, etc.) inside the mag- netic shielding cavity recorded averages of 4.03 mGs (east– west) at Hall 1, 4.12 mGs (east–west) at Well 2 ground level, and 4.35 mGs (north–south) at the Well 2 tunnel entrance.

After demagnetization, the corresponding averages for CH12 and CH2 were 0.76 mGs (60 A demagnetization, east–west, Hall 1; 1.81 mGs at 2.0 K), 0.94 mGs (75 A, east–west, Well 2 ground), and 1.90 mGs (75 A, north–south, Well 2 tunnel entrance). The internal probes (CH4, CH6, etc.) regis- tered 0.34 mGs (60 A, east–west, Hall 1; 2.06 mGs at 2.0 K), 0.32 mGs (75 A, east–west, Well 2 ground), and 1.53 mGs (75 A, north–south, Well 2 tunnel entrance). Furthermore, probe CH4–positioned in the central region of the magnetic shield–exhibited relatively low field amplitude throughout module transfers and demagnetization cycles, never exceed-

ing 2.0 mGs. In contrast, the nine probes (CH6, etc.) located 380

near the shield ends, where fields are strongest, showed larger variations in amplitude, with certain locations approaching

or exceeding 5.0 mGs. Following initial demagnetization, 383

the SHINE high- module achieved an overall field below

1.0 mGs, outperforming comparable demagnetization results

from the LCLS-II project in the United States, where the east– 386

west oriented field remained below 2.3 mGs [ test results of the SHINE high- prototype module before and after online demagnetization.

The key observations are summarized as follows: 1. For the

full module, the initial Q 0 values of CAV1, CAV6, and CAV7 391

, and , respectively.

After demagnetization, these values increased significantly to 393

, and . When demagnetiza- tion was interrupted under the non-timed balance mode (sim- ulating re-magnetization), the values dropped dramati- cally to , and , respectively.

This confirms the strong dependence of on the magnetic field inside the cavity shielding structure. 2. Other cavities, including CAV5 and CAV8, also exhibited substantial provement after demagnetization. CAV2, CAV3, and CAV4 reached the design target of , with CAV2 achieving . This enhancement is attributed to the reduction of trapped magnetic flux and the intrinsic surface resistance of each cavity. 3. The overall across all eight cavities at 134.5 MV increased from after demagnetization, meeting the key acceptance criterion for the module.

These results underscore the critical role of demagnetiza- tion in enhancing cavity and the necessity of effective magnetic field management for optimal performance. engineering module CM01, the variation before and af- ter demagnetization is shown in Fig. . Nearly all cav- ities showed improved , though quenching during post-

demagnetization testing significantly influenced the results. 415

Cavities without quenching generally exhibited higher compared to vertical test results.

Module CM10 underwent online demagnetization at 41 K.

As shown in Fig. , the increased from at 112 MV. The overall performance of low-temperature demagnetization is consistent with room- temperature results. No quenching occurred during the test of CM10 after demagnetization; however, the horizontal- to-vertical ratio was 84%, below the expected 90%. The

thermal gradients present during cryogenic demagnetization 425

generate localized thermal currents in the module, whose magnetic fields complicate the internal magnetic environment and may influence demagnetization outcome. This effect re-

quires comprehensive monitoring, as it may not be fully cap- 429

tured by only two fluxgate probes. Whether the observed ratio discrepancy is linked to such thermal currents remains an open question. Further data collection and analysis are needed to fully understand these phenomena. modules CM01–CM12 following online whole-machine de- magnetization. Across all twelve 1.3 GHz eight-cavity mod- ules, the magnetic field between magnetic shielding lay- ers was reduced to 0.15–0.50 mGs after demagnetization, suggesting even lower field levels in the central region of the superconducting cavities. With the exception of CM02, all modules satisfied the design specification. The un- derperformance of CM02 is likely attributable to the rela- for superconducting cavities in the high- Q modules before and after demagnetization: (a) Prototype module, (b) CM01 module, (c) CM10 module. tively low demagnetization current applied, the intrinsic per- formance limitations of its cavities, and quenching events

during testing. Of the twelve standard 1.3 GHz engi- neering modules, CM01–CM08 were demagnetized using the timed exponential decay (TED) mode, achieving post- demagnetization inter-layer fields of 0.20–0.50 mGs. Mod- ules CM09–CM12 employed the timed linear decay (TLD) mode, which yielded slightly better performance, with resid- ual fields below 0.20 mGs.

In general, horizontal measurements exceeded verti- cal test results for cavities that did not experience quench-

ing. Notably, CM01 and CM08 exhibited significantly higher 454

values than their vertical test baselines. Excluding CM05, all other cavities achieved horizontal-to-vertical ratios ex- ceeding 90%. In contrast, CM02, CM07, and CM09 under- went partial or full quenching during testing after demag- netization, resulting in substantially degraded performance.

DISCUSSION AND CONCLUSION This paper presents the results of online whole-machine de- magnetization of SHINE high- modules and demonstrates that the background magnetic field of the entire module sig-

nificantly influences the demagnetization outcome. Follow- 464

ing online demagnetization, the magnetic field inside the superconducting cavities is reduced to 0.15–0.50 mGs, with values reaching These results satisfy the key performance requirements for SHINE modules and exceed reported online demagnetiza-

tion performance from international counterparts. Significant 470

variation in among different superconducting cavities af- ter demagnetization is observed, which may be attributed not only to magnetic field conditions but also to intrinsic cav- ity performance and quenching events during testing. The magnetic field variations measured by fluxgate probes in- side and outside the cavities show strong correlation before and after demagnetization. This supports the use of exter-

nal fluxgate probes for monitoring the demagnetization pro- 478

cess during module installation and operation. To optimize demagnetization efficiency and results, the background mag- netic field should be maintained stable after online whole- machine demagnetization.

Demagnetization in the north– south orientation requires higher currents to achieve equiv- alent performance to east–west orientation, indicating greater Hard X-ray Free Electron Laser Facility. Project pro- file, [accessed January 2025] Shanghai Zhangjiang Advanced Light Source Large Scientific Device Cluster.

Cluster project, [accessed April 2025] Y. Zhang, J. Chen, D. Wang, RF design optimization for the SHINE 3.9 GHz cavity. Nucl. Sci. Tech. , 73 (2020). difficulty in field configurations with higher ambient back- ground.

Post-demagnetization activities that may compro-

mise results–including transport, welding, conditioning, and 487

mechanical vibration–should be minimized, as these opera- 488

tions can induce partial re-magnetization of the module struc- ture.

Further investigation is warranted to understand the under-

lying mechanism of occasional interruptions during online 492

demagnetization in the non-timed balance mode, and to verify whether the timed exponential decay (TED) and timed linear decay (TLD) modes maintain interruption-free operation un- der more complex leakage field conditions. Compared to the TED mode, the TLD mode maintains higher current levels for a longer duration, facilitating more complete randomiza- tion of stubborn magnetic domains with high coercivity ( in the cryostat material. Theoretically, this leads to more thor- ough demagnetization, a conclusion preliminarily supported by the results from modules CM09–CM12. Additional data from engineering modules must be accumulated and analyzed to fully elucidate the performance differences between these two timed modes. The demagnetization effectiveness shows strong dependence on the background magnetic field charac- teristics of the test station or operational environment. With the exception of CM10, all high- and engineering mod- ule demagnetization tests were conducted at two locations in Hall 1, where background fields are lower and leakage cur- rents are relatively straightforward. By contrast, demagneti- zation under the stronger background fields and more com- plex leakage conditions in Hall 2, or in the north–south ori-

ented underground tunnel with its significantly higher am- 514

bient field, may yield different results. Process parameters likely require optimization under these more challenging con- ditions to maintain performance. Prior to demagnetization, the magnetic field in the central region of the SHINE super- conducting cavities typically meets the specification of be- ing below 3.0 mGs. However, actual measurements in- dicate that this level is insufficient for optimal performance.

To achieve superior , the post-demagnetization field in- side the cavities must be reduced below 1.0 mGs, or even

to 0.50 mGs. Analysis of the cryogenic demagnetization re- 524

sults from CM10 suggests minimal performance difference 525

between low-temperature and room-temperature processes, though neither has yet achieved optimal results. Further cryo-

genic demagnetization data collection and detailed analysis 528

are essential to clarify these findings. C. Feng, H. Deng, Review of fully coherent free-electron lasers. Nucl. Sci. Tech. , 160 (2018). s41365-018-0490-1 Z. Zhao, H. Ding, B. Liu, Shanghai hard X-ray FEL facility progress status. In: Proceed. IPAC’24, Nashville, TN, USA, S. Posen, G. Wu, A. Grassellino, et al., Role of magnetic

flux expulsion to reach Q 0 > 3 × 10 10 in superconducting 547

rf cryomodules. Phys. Rev. Accel. Beams , 032001(2019).

W. Pan, J. Zhai, F. He et al., High Q and high gradient perfor- mance of first medium-temperature baking 1.3 GHz cryomod- ule. Phys. Rev. Accel. Beams , 092003 (2024). org/ 10. 1103/PhysRevAccelBeams.27.092003

[8] W. Pathirana, A. Gurevich, Effect of random pinning 554

nonlinear dynamics dissipation vortex driven by a strong microwave current. Phys. Rev. B 184518(2021).

A. Romanenko, A. Grassellino, A. Crawford, et al., Ultra-high

quality factors in superconducting niobium cavities in ambient 559

magnetic fields up to 190 mG. Appl. Phys. Lett. , 234103 (2014).

J. Chen, H. Hou, Z. Ma, et al., Cryomodules development for SHINE project. In: Proceed. IPAC’24, Nashville, TN, USA (2024) F. He, W. Pan, P. Sha, Medium-temperature furnace baking of 1.3 GHz 9-cell superconducting cavities at IHEP. Supercond.

Sci. Technol. , 095005(2021). 6668/ac1657 T. Okada, E. Kako, T. Konomi, et al., Systematic eval-

uation of magnetic sensitivities of anisotropic magne- 570

toresistive sensors at liquid helium temperature for su- perconducting cavities.

Instrum. (2021). F. Kramer, S. Keckert, O. Kugeler, et al., Study of the dynam- ics of flux trapping in different SRF materials. In: Proceed.

SRF’23, JACoW, Grand Rapids, MI(2023). D. Turner, A. Terricabras, I. Díaz-Palacio, et al., Flux expul- sion lens: Concept and measurements, In: Proceed. SRF’23, JACoW, Grand Rapids, USA(2023) P. Arpaia, M. Buzio, O. Capatina, et al., Effects of tem-

perature and mechanical strain on Ni-Fe alloy CRYOPHY 581

for magnetic shields. J. Magn. Magn. Mater. , 514- 523(2019).

[16] S. Sah, G. Myneni, J. Atulasimha, et al., Experimental Char- 584

acterization of Magnetic Materials for the Magnetic Shielding of Cryomodules in Particle Accelerators. IEEE T. Magn. 1-6(2016).

S. Chandrasekaran, K. Saito, S. Shanab, et al., Magnetic Shield Material Characterization for the Facility for Rare Iso- tope Beams’ Cryomodules. IEEE T. Appl. Superc. (3), 1-5, (2015).

R. Augusto, A. Trudel, Z. Liu, et al., An overview of shielding optimization studies TRIUMF- ARIEL facility.

Nucl. Instrum. Meth. 165401(2021). F. Kramer, S. Keckert, O. Kugeler, et al., A new exper- iment to enable rapid systematic investigations of flux trapping dynamics superconducting radio-frequency cavity applications.

Instrum. (2024). Y. Iwashita, Y. Kuriyama, H. Tongu, et al., Magnetic Field Shielding With Superconductors. IEEE T. Appl. Supercond. (6), 1-4(2022).

N. Huang, Z. Liu, B. Deng, et al., The MING proposal at SHINE: megahertz cavity enhanced X-ray generation. Nucl.

Sci. Tech. , 6(2023). Y. Zong, Q. Chen, X. Huang, et al., Realization of accelerating gradient larger than 25 MVm on high-Q 1.3 GHz 9-cell cavities for SHINE. In: Proceed. SRF2023, Grand Rapids, MI, USA (2023).

J. Chen, Y. Zong, X. Pu, et al., Ultra-high quality factor and ultra-high accelerating gradient achievements in a 1.3 GHz continuous wave cryomodule. Nucl. Sci. Tech. , 25 (2025).

Fierlinger, Limits Magnetic Field Environments Magnetic Shields. IEEE T. Indus. Elec. 5385-5395(2021).

J. Yang, M. Shi, X. Zhang, et. al. , Demagnetization Pa- rameters Evaluation of Magnetic Shields Based on Anhys- teretic Magnetization Curve. Materials (15), 5238(2023).

C. Zuo, Z. Wang, W. Yang, et. Al. , Research on Induc- tance Calculations for Degaussing Windings of a Double- Layer Steel Cylinder. IEEE Access , 55205-552139(2021).

Y. Lin, J. Yang, F Cao, et. al. , Enhancement of Magnetic Shielding Based on Low-Noise Materials, Magnetization Con- trol, and Active Compensation: A Review. materials (22), 5469 (2024).

J. Plouin, N. Bazin, P. Charon et. al., CEA Experience and Ef- fort to Limit Magnetic Flux Trapping in Superconducting Cav- ities. In Proceedings of SRF2015, Whistler, BC, Canada.

S. Chandrasekaran, K. Saito, Validation of Local Magnetic Shielding for FRIB Using A Prototype Cryomodule. In Pro- ceedings of SRF2015, Whistler, BC, Canada.

S. Chandrasekaran, A. Crawford, Demagnetization of a Com- plete Superconducting Radiofrequency Cryomodule: ory and Practice, IEEE T. Appl. Superc. (1), 1-6(2017).

Checchin, Martinello, Grassellino, Frequency dependence trapped sensitivity cavities.

Appl. Phys. Lett. (2018) M. Shi, X. Zhang, J. Yang, et. al. , The SF and remanence evaluation of magnetic shields based on SST for low-frequency and degaussing situation. J. Phys. D: Appl. Phys. , 365003 (2024).

J. Yang, X. Zhang, M. Shi, et. al. , Experimental studies on the performance of magnetic shields under different magneti- zation conditions. J. Phys. D: Appl. Phys. , 215001 (2023).

Z. Wang, H. Li, S. Yang, et. al. , Modeling and exper- imental studies of degaussing hysteresis in near-zero mag- netic shielding systems. J. Appl. Phys. , 084501 (2024).

Y. Zong, J. Chen, D. Wang, et al., Accelerating gradi-

ent improvement in nitrogen-doped superconducting radio- 659

frequency cavities for SHINE. Nucl. Instrum. Meth. A. 168724(2023).

X. Wang, S. Barbanotti, J. Eschke et al., Thermal performance analysis and measurements of the prototype cryomodules of Euro pean XFEL accelerator-Part II. Nucl. Instrum. Methods Phys. Res. Sec. A , 688–700(2014).

1016/j. nima. 2014. 07.045 666

S. Posen, M. Checchin, A. Crawford, et al., Efficient expulsion of magnetic flux in superconducting radio fre- quency cavities for high applications. J. Appl. Phys. 213903(2016).

[38] J. Gao, CEPC technical design report: accelerator. Radiat. De- 671

tect. Technol. Methods , 1–1105 (2024). 1007/ s41605-024-00463-y Okada, Kako, Masuzawa, Obser- vation quenching-induced magnetic trap-

using magnetic field temperature mapping system, Phys.

Accel. Beams

[40] S. Huang, Y. He, L. Peng, Exploration of the copper–niobium 679

composite superconducting cavities for pursuing extremely high operational stability at IMP. Nucl. Sci. Tech. 81(2025).

M. Martinello, M. Checchin, A. Grassellino, et al., Mag- netic flux studies in horizontally cooled elliptical super- conducting cavities, J. Appl. Phys. , 044505 (2015).

Chitarin, Marconato, Mayer, Realization Tests Prototype Fluxgate Magnetic Sensors Neutral Injectors.

Sensors 1492(2023). X. Xu, Y. Liu, X. Sun, et al., Analysis and Design of Biplanar Coils Within Magnetic Shielding Room Considering Actual Ferromagnetic Boundaries. IEEE Sensors J. (24), 40623- 40633(2023). 082002(2022).

Status number Module position Demagnetization current Measure status Average reading of external fluxgate Average reading of internal fluxgate Hall 1 test station

60 A

293K before demagnetization 6.56 mGs 4.03 mGs 293K after demagnetization 0.76 mGs 0.34 mGs 2K after demagnetization 1.81 mGs 2.06 mGs Hall 1 north-south direction 293K after demagnetization 37.9 mGs 12.36 mGs Hall 1 east-west direction 293K after demagnetization 6.78 mGs 1.69 mGs Well 2 underground north-south direction

75 A

293K before demagnetization 9.42 mGs 4.35 mGs 293K after demagnetization 1.89 mGs 1.53 mGs Well 2 ground north-south direction

75 A

293K before demagnetization 16.32 mGs 4.12 mGs 293K after demagnetization 0.94 mGs 0.41 mGs test results after demagnetization of SHINE engineering modules.

Module number Position Demagnetization mode/current Magnetic field Whether quenching before/after demagnetization Hall 1 TED mode/75 A 2.00 mGs/0.20 mGs @133 MV 120% Hall 1 TED mode/60 A 1.00 mGs/0.30 mGs @166 MV Hall 1 TED mode/75 A 1.77 mGs/0.50 mGs @166 MV Hall 1 TED mode/75 A 0.80 mGs/0.40 mGs @166 MV Hall 1 TED mode/75 A 2.00 mGs/0.25 mGs @166 MV Hall 1 TED mode/75 A 1.40 mGs/0.30 mGs @166 MV Hall 1 TED mode/75 A 2.40 mGs/0.20 mGs @166 MV Hall 1 TED mode/75 A 0.35 mGs/0.29 mGs @166 MV 108% Hall 1 TLD mode/75 A 0.34 mGs/0.18 mGs @165 MV Hall 2 TLD mode/75 A 1.40 mGs/0.19 mGs @112 MV Hall 1 TLD mode/75 A 0.35 mGs/0.15 mGs @166 MV Hall 1 TLD mode/75 A 0.25 mGs/0.16 mGs @163 MV