A large area 3He tube array detector with vacuum operation capacity for the SANS instrument at the CSNS

Xing-Fen Jiang• Jian-Rong Zhou,3 • Hong Luo • Liang Xiao •Xiao-Juan Zhou • Hong Xu • Yuan-Guang Xia • Xiao-Guang Wu •Lin Zhu,3 • Wen-Qing Yang,3 • Gui-An Yang • Bei-Ju Guan •Hong-Yu Zhang • Yu-Bin Zhao • Zhi-Jia Sun,3 • Yuan-Bo Chen,3

ABSTRACT The small-angle neutron scattering (SANS)instrument, one of the first three instruments of the China Spallation Neutron Source (CSNS), is designed to probe the microscopic and mesoscopic structures of materials in the scale range 1-100 nm. A large-area 3He tube array detector has been constructed and operates at the CSNS SANS instrument since August 2018. It consists of 120 linear position-sensitive detector tubes, each 1 m in length and 8 mm in diameter,and filled with 3He gas at 20 bar to obtain a high detection efficiency. The 3He tubes were divided into ten modules, providing an overall area of 1000 mm × 1020 mm with a high count rate capability.Because each tube is installed independently, the detector can be quickly repaired in situ by replacing damaged tubes.To reduce air scattering,the SANS detector must operate in a vacuum environment (0.1 mbar). An all-metal sealing technique was adopted to avoid high-voltage breakdown by ensuring a high-voltage connection and an electronic system working in an atmospheric environment. A position resolution of 7.8 ± 0.1 mm (full width at maximum) is measured along the length of the tubes, with a high detection efficiency of 81 ± 2%at 2 A˚.Operating over the past four years, the detector appears to perform well and with a high stability, which supports the SANS instrument to finish approximately 200 user scientific programs.

Keywords Neutron detectors · 3He tubes · Gaseous detectors · Small-angle neutron scattering

1 Introduction

Neutron scattering has become a powerful technique used in materials research and industry thanks to its high penetration characteristics, sensitivity to light elements,discrimination ability of isotopes, and detectivity of magnetic fields in matter. Neutron scattering can be used to study the interplay between spin fluctuations and superconductivity in high-temperature superconductors. The wavelength of low-energy neutrons is close to the atomic spacing, which allows the probing of the static microstructure of matter and the study of its dynamic mechanisms.Neutron scattering also allows nondestructive testing to reveal the inner structure and deformation of engineering devices. The high penetration of neutrons facilitates in situ measurements in a high-temperature,high-pressure, or strong magnetic field environment. The expansion of neutron-scattering applications has led to the rapid development of neutron scattering sources and instruments. Spallation neutron sources have the potential to provide a much higher neutron flux than reactor neutron applications. In addition, they provide neutron flight-time information and a low-background environment. These benefits have motivated the development of advanced neutron scattering sources worldwide. The China Spallation Neutron Source (CSNS) [1] is the fourth built pulsed spallation neutron source in the world, after the ISIS Neutron and Muon Source in the UK, the Spallation Neutron Source (SNS) in the US, and the Japan Proton Accelerator Research Complex (J-PARC). The CSNS consists of an 80 MeV negative hydrogen ion linear accelerator, a 1.6 GeV rapid cycling proton synchrotron,two beam transport lines,and a target station.With a pulse frequency of 25 Hz and power of 100 kW, the CSNS has been in public operation since August 2018 [2, 3]. In future,the power of the CSNS will be upgraded to 500 kW and a second target station will be built. During the first phase of the project, three instruments were built: a smallangle neutron scattering (SANS) instrument, a generalpurpose powder diffractometer (GPPD), and a multipurpose reflectometer (MR).

The SANS instrument at the CSNS is the first SANS instrument based on a pulsed source in China. It was designed to probe the internal structure of bulk, powder,and liquid materials on length scales ranging from 1 to 100 nm[4,5].The SANS instrument utilizes the first beam port of the CSNS target station, which is adjacent to a coupled liquid hydrogen moderator.Neutrons are produced when beams of high-energy protons (accelerated to 1.6 GeV)are fired at a tungsten target,and then moderated to lower energies by liquid hydrogen.A T0 chopper placed downstream of the shutter eliminates fast neutrons by blocking. A double-disk bandwidth chopper installed behind the T0 chopper selects neutron wavelengths between 0.5 and 12 A˚. In addition, a commercial frameoverlapping mirror removes background long-wavelength neutrons.The SANS instrument beamline is designed to be short and straight. A set of collimation segments with different apertures were assembled to shape and focus the incident neutron beam. With its classic point-focusing beamline geometry, SANS provides good neutron transmission at short wavelengths and a wide range of scattering vector measurements.

The scattering vector (q) can be acquired from the neutron scattering angle(θ)and wavelength (λ),which are obtained from the detection position and time of flight(TOF) of the scattering neutrons [6]. Achieving the range and accuracy of q measurements requires a detector with an effective area of more than 1 m × 1 m, a detector efficiency of over 50%at 2 A˚,and a position resolution along the tube of more than 10 mm. The detector must be operated in a vacuum to reduce air scattering.The distance between the detector and the sample is adjustable from 2 to 4 m to measure a wide range of scattering angles. The detector must also be operatable at both high and relatively weak count rates, depending on the sample scattering experiment. For high count rates, the detector must have a small dead time and a high data-transmission speed. Low count rates crucially require a low background, in particular, low gamma-ray sensitivity. A larger-area3He tube array detector was constructed using SANS instruments,-consisting of 1203He tubes with an effective detection area of 1000 mm × 1020 mm. Because each tube is installed independently, the detector can be quickly repaired in situ by replacing damaged tubes. An all-metal sealing technique with a good sealing capability and stable reliability was applied to the vacuum operation capacity.The design,assembly and performance of the detector are described below.

2 Detector design

Owing to their superior neutron detection efficiency and neutron gamma discrimination ability,3He gas neutron detectors have been used in SANS instruments worldwide[7]. A3He gas detector is sensitive to neutrons via the nuclear reaction

Multiple wire proportional chambers (MWPC) and linear position-sensitive3He tubes are commonly used3He gas detectors.An MWPC filled with3He gas was once used as the main detector for the SANS instruments of other neutron sources [8].The MWPC consists of an anode wire plane and two orthogonally oriented cathode wire planes.The anode wire plane with a positive bias voltage is located at the center of the detection volume. The cathode wire planes are arranged on both sides of the anode wire plane and connected to the preamplifiers for signal readout. The neutron position is determined by XY coincidence, and the spatial resolution depends on the cathode wire pitch. The sensitive area of the MWPC can be tailored for various applications, and the spatial resolution can reach the millimeter level. However, MWPCs suffer from a low count rate capability and are easily damaged by excessive beam exposure. Subsequent repair is expensive and time-consuming.A thin entrance window in the MWPC is necessary to reduce neutron scattering.However,when the MWPC is operated in a vacuum,such a window becomes unilaterally deflected by the atmospheric pressure. This causes a deviation of the neutron trajectory in the MWPC, which is undesirable for scattering-angle measurements. To overcome the above shortcomings,3He tube detector arrays have been increasingly used in SANS instruments because of their high count rate, high detector efficiency, and easy maintenance. For this study, we adopted an array design.

2.1 3He tube array

The detector and electronics groups of the CSNS conducted numerous studies on neutron detectors [9-14].Based on international expert advice, a detector scheme with a3He tube array was chosen.The3He tube is a single-wire proportional tube with a length of 1 m, a diameter of 8 mm,and filled with3He gas at 20 bar.Thinwalled stainless steel is used as the detector cathode. A resistive wire (7.8 kΩ) with a diameter of 13 μm, running along the tube length, serves as the detector anode. As shown in Eq. (1), the3He tube captures neutrons through the nuclear reaction n(3He, P)T with a total kinetic energy of 764 keV. Primary electrons are produced in3He gas by the ionization of protons and tritons.To increase the signalto-noise ratio (SNR), a gas multiplication effect is produced by a strong electric field near the anode wire. These primary electrons are multiplied by a Townsend avalanche.The charge cloud generates a pulse current that is transmitted along the resistance wire to both ends of the tube.The hit position of neutrons can be acquired using the charge center of the gravity method[15].The high3He gas pressure (20 bar) achieves a theoretical average detection efficiency of more than 80% for 2 A˚ neutrons, which is adequate for the detector requirements of SANS. A stopping gas of argon was also added to minimize the wall effect and improve the spatial resolution by reducing the range of protons and tritons.The small diameter of the3He tube provides a good resolution perpendicular to the tube length and can also improve the tube count rate by reducing the drift time of positive ions.

To achieve the detection area, 1203He tubes were used to construct a two-dimensional position-sensitive detector.The small3He tube diameter and the high-speed readout system enable the detector to reach a much higher total count rate capability than an MWPC. Because each3He tube detects neutrons independently, the detector can still operate when some3He tubes are damaged. The detector can be repaired easily by replacing a defective tube. The reliability and easy maintenance of the detector significantly improve experimental efficiency.

2.2 All-metal sealing technique

The detector is housed in a vacuum tank at a pressure of 0.1 mbar.Since the vacuum level is sufficient to satisfy the neutron scattering background requirements, pushing to achieve an even higher vacuum level is unnecessarily timeconsuming. However, high-voltage breakdown is inevitable in this rough vacuum.A high-voltage breakdown experiment was performed to confirm the sparking region of the high-voltage connection under different vacuum conditions.Figure 1a shows the experimental setup,which included a vacuum chamber, a vacuum pump, and a highvoltage power supply. One end of the high-voltage cable was exposed to the vacuum chamber,and the other end was connected to a power supply, which also monitored the current in real time. Breakdown is deemed to occur when the current exceeds 1 μA. The breakdown voltage was measured at different vacuum levels, and the results are shown in Fig. 1b.The breakdown voltage was significantly higher at both very low and high gas pressures. At 0.1 mbar, high-voltage breakdown occurs when the high voltage exceeds 800 V. The working voltage of the3He tube was set in the range of 1450-1600 V, far above the safe high voltage.

Isolation from the vacuum environment of the highvoltage connection is necessary to ensure the safe operation of the detector. Plastic pipes were used in the SANS2d instrument of ISIS to encase high-voltage connections to provide an atmospheric environment [16]. Plastic pipes simplify the installation process; however, they are vulnerable to wear and tear.A fully engineered approach with the all-metal sealing technique was employed at the CSNS to ensure that the high-voltage connections and electronic system could operate under atmospheric pressure. The allmetal sealing system consists of stainless-steel right-angled connectors, tailored copper connecting tubes, copper tube fittings, an aluminum-sealed cavity, and vacuum hoses.The copper connecting tubes enclose the high-voltage cables of the3He tube, and a sealed aluminum cavity was used to install the electronics. The right-angled connectors and tube fittings ensured that the3He tube installation tightly sealed and leak-free. As the high-voltage cables pass through these metal parts and connect to the preamplifier in the sealed cavity, this technique simultaneously provides good electromagnetic shielding. The atmospheric environment is achieved using vacuum hoses, which connect the sealed cavity to the wall of the vacuum tank.Another problem associated with vacuum is that the heat generated by electronics is difficult to dissipate.To sustain good heat-dissipation performance for the electronics, a ventilating system was connected to the sealed cavity through vacuum hoses.

2.3 Detector structure

As shown in Fig. 2a, 120 tubes are mounted onto four support panels, which are in turn mounted onto an assembly jig.The four support panels provide four support points for each tube, which are aligned within an error of 0.3 mm to guarantee that the wire is located at the center of the tube. Near the two inner support points, two thin stainless-steel strips are attached to the bronze springs that pull the tube down tightly onto the support. The other two support points are at both ends of the tube, where pairs ofright-angled connectors are welded. The right-angled connectors enclose the high-voltage cable and provide an installation interface for the3He tube. The right-angled connectors are mounted into the mounting hole of the external support panel using bulkhead unions (Fig. 2b),which can bear the force of the following installation to protect the tube. Two different lengths of right-angled connectors are installed individually to achieve the staggering assembly (Fig. 2c). This assembly method provides ample space for installation and repair. The copper connecting tubes connect the right-angled connectors to the sealed cavity (Fig. 2d) to ensure a good sealing performance. Tube fittings are used to clamp tailored copper tubes.

Twelve KF vacuum hoses are connected to the sealed cavity.Ten of these hoses are used to feed the high-voltage cables, T0 signal, electronic low-voltage cables, and detector output signals. The remaining two hoses supply the cooling gas.A refrigerator machine,combined with one water-gas heat exchanger, further lowers the temperature of the cooling gas.The cooling gas effectively removes the heat generated by the electronic system. A good working state of the electronics can thus be maintained in a lowtemperature environment. To reduce the back-scattered neutrons, the backs of the tubes are shielded by a layer of boron aluminum alloy. A 240 mm × 240 mm hole at the center of the sealed cavity serves to avoid any high-energy neutron background along the path of the neutron beam,to achieve a high central resolution.A regulating mechanism,consisting of vertical and horizontal adjusting nuts, facilitates the installation of the detector.The adjusting nuts can change the detector position in small steps to ensure mounting accuracy. The total weight of the detector is ～800 kg, with a width, height, and depth of 1300 mm × 1420 mm × 1000 mm.

3 Readout electronics and data acquisition system

The readout electronics system was customized for the SANS detector by the detector and electronics group of the CSNS, to achieve a high response speed, a high charge resolution, and a low noise level. This system consists of ten modules designed for easy installation and maintenance(Fig. 3). Each electronic module contains two front-end electronics (FEE) boards and one digital readout electronics(DRE)board to record the neutron events in the twelve3He tubes [17]. An FEE board contains a high-voltage distribution circuit, preamplifiers, blocking capacitors, and a shaping circuit. Each FEE board has only one highvoltage input port but supplies the same high voltage to the twelve3He tubes. The pulse signal passes through the blocking capacitors and is acquired by the preamplifier when a neutron is incident on a3He tube.Charge-sensitive amplifiers are used to enhance the small pulse currents. A shaping circuit is used to broaden the signal to improve the charge accuracy. The charge signal accumulated in the amplifiers is then converted to a voltage signal.This analog signal is converted to a digital signal by the ADCs on the DRE board, and is then transmitted to the FPGA for data processing. The signal pulse amplitude is acquired using the peak method. Neutron signals are screened out in the FPGA by determining whether the signal amplitude exceeds the previously set threshold. The neutron event is time-stamped according to the T0 signal,which is an event start time synchronously generated in the T0 board with the bombarding of protons onto the tungsten target. A preliminary calculation was performed to reduce the amount of transmitted data.The neutron event information was simplified and only includes the T0 number, TOF,baselines(VB1and VB2),and pulse amplitudes(V1and V2)from both ends of the3He tube.The neutron event location can be calculated from the electronic system output as

Fig. 3 Diagram of the readout electronics

Raw data from 240 channels are acquired by the data acquisition (DAQ) system through twelve pairs of optical fibers. The DAQ system is based on the Scientific Linux CERN 6.8 operating system,and the hardware architecture performs front-end data readout, operation control, and back-end data processing. The application software was built using the Qt framework to enable the experimental configuration, control, monitoring, alarm, display, and other functions to be performed using a graphical user interface rather than a command terminal. After pre-processing and local storage, the data are uploaded to the analysis system of the SANS instrument. Optical fibers allow a high transmission speed. Because one pair of optical fibers can transmit the data output from twelve3He tubes, the number of cables needed to enter the detector cavity is significantly reduced. With a count rate of 200 kHz per module, the detector provides a total counting-rate capability of 2 MHz.

4 Lab assembly

The assembly was conducted in the lab,in a clean room customized specifically for the detector. The main process includes right-angled connector welding,3He tube mounting, copper tube connection, and electronic installation. A special welding platform was set up to ensure welding accuracy. As shown in Fig. 4a a right-angled connector was fixed to the end of a3He tube using a customized fixture. A thermally insulating ceramic was placed inside the right-angled connector at the welding position to protect the high-voltage cable. After passing the high-voltage cable through the ceramic and the connector, welding was performed using a Swagelok welding system (M200),which executes the pre-set welding program automatically.The welding current was optimized to ensure the quality of the welding point(Fig. 4b).Cooling nitrogen gas was used to ensure that the temperature was below 50 °C near the welding point. When the welding was completed, the leak rate was tested for each3He tube to verify the sealing properties of the welded joints. A helium mass spectrometer leak detector was used to perform the high-precision leak measurements. A special adapter port connects the right-angled connector to the leak detector. The right-angled connector was cut off and rewelded if the leak rate was above 10-8Pa·m3/s.The leak rates of all the3He tubes were measured to be approximately 10-9Pa·m3/s, well below the threshold. To check whether the high-voltage cable was scalded, the high-voltage leakage current was tested. Only one3He tube had a large leakage current caused by scalding. We therefore conclude the welding process is sufficiently safe for the metal extension connection of3He tubes.

Fig. 4 (Color online) Welding of tube and right-angled connector:a welding process,b welding joint.Detector assembly:c copper tubes used to connect the right-angled connectors to the sealed cavity, d assembly of readout electronics. Detector installed at the SANS instrument: e front view, f rear view

3He tubes with right-angled connectors were mounted in a staggered manner on the support panels, as shown in Fig. 4c. Right-angled connectors were inserted into the bulkhead unions.For each3He tube,two thin stainless steel strips were wound around the tube and attached to a bronze spring.The other end of the bronze spring was then fixed to the screws installed on the assembly jig. The bronze springs tightly pulled the tube down to the mounting grooves of the support panels. The bulkhead unions were then tightened to ensure that the tube was fixed rigidly and did nor bend.Easy to be processed,each copper connecting tube was customized for the corresponding3He tube to reduce stress during installation. Then, the tailored copper tubes connected the right-angled connectors to the sealed cavity through the tube fittings. The tube fittings were tightened by the torque spanner with the optimized torque to prevent the tube from bearing excessive force.As shown in Fig. 4d,the FEE boards and DRE boards of the modules were mounted in the sealed cavity. Each FEE board was installed in a closed aluminum box to ensure good electromagnetic shielding. The corresponding DRE boards is installed between the two FEE boards in the same module.When the assembly was completed, two protection covers were installed on the two outer support panels to protect the connectors.

The leak rate of the entire detector was tested using a helium mass spectrometer leak detector after assembly.Each welding joint and tube fitting interface was sprayed with helium using an air gun to measure the local leak rate.The maximum testing result was approximately 10-10-Pa·m3/s, which confirms the good sealing property of the tube connection. The detector was encased in a large helium-filled plastic bag to test its sealing performance. A total leak rate of approximately 10-7Pa·m3/s showed adequate sealing performance under vacuum conditions.The detector was then moved to a SANS instrument and installed in a vacuum tank.A customized mounting system with one mounting trolley and two locating guide rails was used to ensure smooth installation. The detector was first mounted onto the trolley in the laboratory, and the trolley was then loaded onto the guide rails using hoisting equipment. The trolley moved slowly along the guide rails to deliver the detector to the mounting platform in the vacuum tank. The regulation mechanism was used with a high-precision position-test system to adjust the detector position to achieve a high installation accuracy. Figure 4e,f shows the front and rear views, respectively, of the detector in the vacuum tank.Twelve corrugated tubes were used to connect the sealed cavity to the vacuum tank to pass through the cables and cool the air. The corrugated tubes were threaded through a drag chain to enable the movement of the detector. With a vacuum motor, the detector can be moved in a vacuum along the beam path over a distance of 14 to 16 m from the moderator. A refrigerating machine with a heat exchanger was installed outside the scattering room to improve the heat-dissipation efficiency of the detector. With a cooling air flow of 300 SLPM (standard liters per minute) and a temperature of 1 °C, the temperature in the detector-sealed cavity can be kept below 30 °C even under a high heating load due to the electronics.

5 Commissioning and performance

Preliminary tests were conducted at BL20 of the CSNS to evaluate the position resolution and detection efficiency of the3He tubes. To ensure a high position accuracy for sample measurement, the position calibration of the entire detector system was performed at the SANS instrument.

5.1 Position resolution

Position resolution is an important property for a detector that is not only related to the parameters of the3He tube, but also affected by the electronic system. To facilitate the measurements, a testing prototype, including the same type of3He tube,was constructed(Fig. 5a).A boron/aluminum (B/Al) alloy plate with nine slits (2 mm wide)was attached to the front of the tubes to collimate the neutron beam. The distance between adjacent slits was 100 mm, such that the slits were in the coordinate range 100-900 mm. The pulse-height distribution of the neutron signals is shown in Fig. 5b with a high voltage of 1540 V.Three peaks are clearly visible in this distribution, corresponding to the triton, proton, and full energy peaks. The triton- and proton-escaping events account for a relatively small percentage of neutrons count; therefore, the triton and proton peak amplitudes should be much lower than the full energy peak. However, the space-charge effect becomes more apparent when the high voltage and the associated gain of the3He tube are sufficiently high.Owing to the space-charge effect,the charge generated by the total deposition energy of the secondary particles decreases,and the peak-to-valley ratio of the full energy peak decreases.The amplitudes of the triton and proton peaks in Fig. 5b are similar to that of the full energy peak,which indicates that the space-charge effect is clearly manifested at the supplied high voltage.

Owing to testing time limitations, only five slits (at positions 100, 300, 500, 700, and 900 mm) of the B/Al alloy plate were measured.The full width at half maximum(FWHM) of the measured position distribution can be calculated using a Gaussian fit (FWHM = 2.35 × sigma,where sigma is the standard deviation). Because of the impedance of the FEE board and the uniformity of the tube anode wire,the measured position TestX deviates from the actual position RealX of the slit. To remove this discrepancy, a calibration was performed by linear fitting. RealX is plotted against TestX in Fig. 5c,with the fit results given in the legend.The formula RealX = 1.294 × TestX-133.8,was used in the subsequent position reconstruction for this3He tube. The recalculated position distributions of all the five slits are shown in Fig. 5d. All the FWHM position resolutions of the slits after calibration are shown. The measured position resolution is approximately 8 mm, in good agreement with the typical position resolution,which is within 1% of the tube length [18]. The best resolution was 7.8 ± 0.1 mm at the center of the tube, and the error was the statistical error. Owing to the influence of uncorrelated noise, the resolution worsens as the measured position approaches the tube end [19]. The position resolutions along the length are also comparable to the tube diameter, which determines the position resolution across the3He tube array detector.Consideration of the tested3Hetubes suggests that the detector provides a relatively uniform position resolution in both the X- and Y-directions.The3He tube provides a better position resolution under a higher voltage, albeit at the expense of the count rate.

5.2 Detection efficiency

A high neutron detection efficiency is required to improve the SNR and reduce the experimental time, particularly for large neutron scattering vector measurements.To accurately measure the detection efficiency, a low-flux monochromatic neutron beam is required.Figure 6 shows a schematic of the detection-efficiency test. A mica monochromator was used to provide neutrons with a wavelength of 2 A˚and to reduce the neutron flux.To avoid count-rate saturation, an additional slit (20 mm × 2 mm)was placed in front of the3He tube to limit the input area of monochromatic neutrons. A standard3He tube with a diameter of 25 mm and a3He gas pressure of 20 bar was used to measure the input neutron flux at the same testing position. A beam detector (ORDELA 4562 N) was placed at the neutron beam port to normalize neutron counts.When the standard3He tube was being measured, the count-rate saturation test was performed by partially shielding the slit. The neutron counts, denoted countA and countB, were measured as the slit was semi-shielded on each side. After normalization, countA and countB were summed and compared to the count acquired under the fully open slit. The same count result for 2 A˚ neutrons indicated that the count rate of the3He tube was not saturated. The detection efficiency of the testing tube was measured to be 81 ± 2% at 2 A˚, compared to that of the standard tube.The measured efficiency was consistent with the theoretical calculation and simulation results at the center of the3He tube. The uncertainty of the measured value included the influence of neutron wavelength broadening (± 0.5%), the slit width (± 0.5%), and the wall-thickness absorption of the3He tube (± 1%). A high detection efficiency is sufficient to meet the design requirements, and the sample experiment time can be significantly reduced.

5.3 Calibration of the detector system

Neutron scattering experiments require a high precision in the scattering vector (q) measurement. The detector should therefore provide not only a high spatial resolution but also sufficient position accuracy. The position calibration of the entire detector system was performed using the SANS instrument. A piece of 10-mm-thick polyethylene was used to diffuse the neutrons to cover the entire detector area.Six 7-mm-thick B/Al alloy plates,each with a 2 mm slit in the middle, were mounted at the front of the detector at 4 mm intervals (Fig. 7a). This interval was set to ensure sufficient neutron detection statistics if the calibration experiment time was limited. 11 slits were used to calibrate the measured positions. The charge balance effect shifted the measured position distribution toward the middle of the3He tube,resulting in an effective length of less than 1 m. The charge balance effect was weak in the middle but worsened near the tube end. The calibration function was optimized to correct for the charge balance effect.Compared with the above linear calibration,the improved calibration function added an exponential component for positions near the tube end. Figure 7b shows the position distributions of all 11 slits after the calibration. Thanks to this calibration improvement, even the position distributions of the edge slits show good linearity, and the position distributions of the 1203He tubes are mutually consistent. The gap in the middle of the central slit was due to neutron absorption by the neutron beam stop.

With a good position resolution, detection efficiency,and global uniformity, the detector meets the design requirements of the SANS instrument. The detector has been operating steadily for four years since 2018 and has successfully completed six cycles of operation open to the public. Many scientific user programs (in fields including nanocomposites, nanoparticle clusters, ionic solid solutions, metallic glasses, and cancer immunotherapy) have been completed using the SANS instrument [20-27].

Fig. 6 (Color online)a Schematic and b experimental setup of the detection efficiency test

Fig. 7 (Color online)a Calibration plates fabricated by B/Al alloys with 2 mm wide slits and b position distributions after the improved calibration

6 Summary and future work

To conclude,a3He tube array detector with an effective detection area of 1000 mm × 1020 mm was installed and stably operated on the SANS instrument. All the 1203He tubes were assembled independently,enabling rapid in situ maintenance. Using the all-metal sealing technique, the detector operates under the vacuum condition of the SANS and also meets higher vacuum requirements. With its minimum position resolution of 7.8 ± 0.1 mm FWHM and its efficiency of 81 ± 2% at 2 A˚, the detector has been producing quality data for many users. In future, a new electronic system based on ASIC will be developed to achieve a higher count rate, much lower power consumption, and much better resolution.

Author contributionsAll authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Xing-Fen Jiang, Jian-Rong Zhou, Hong Luo,Liang Xiao, Xiao-Juan Zhou, Hong Xu, Yuan-Guang Xia, Xiao-Guang Wu, Lin Zhu, Wen-Qing Yang, Gui-An Yang and Bei-Ju Guan.The first draft of the manuscript was written by Xing-Fen Jiang and all authors commented on previous versions of the manuscript.All authors read and approved the final manuscript.