Abstract
A sample-motion-synchronized neutron stroboscope is developed using the RIKEN accelerator-based compact neutron source (RANS). When a sample reaches a specified rotation phase, a phase timing signal is generated, triggering a proton pulse at the proton accelerator. A neutron pulse exposes the imaging detector through the sample. By controlling the delay time between the phase signal and the proton pulse timing, the neutron exposure timing is always synchronized with the specific sample rotation phase. A sample rotating at a speed of 1800 RPM is prepared, and clear still images of neutrons are taken.
Introduction
Automobile manufacturers expect to take still images of engines in operation to observe the behavior of engine oil and fuel injection [11] using a stroboscopic method to improve engine efficiency. A typical stroboscope [6] is illuminated with intermittent light to capture intermittent images of objects in motion with an open camera shutter. A high-brightness light source is necessary to clearly capture high-speed phenomena. For neutron imaging, a high-intensity neutron source and a high-speed neutron imaging detector are required to take still images of an object in motion [3,7,8]. A neutron beam was irradiated onto the sample at a timing independent of the sample’s motion or continuously, and a high-speed camera cut the image at various times and captured still images. In the medical use of x-rays, x-rays are irradiated synchronously with the patient’s motion to capture still images [5].
In this study, we have developed a new sample-motion-synchronized neutron stroboscope for taking still images of a sample with repetitive motion using a low-intensity neutron source [9]. The timing of a neutron pulse is forced to synchronize with the sample rotation phase, and then all images generated by each neutron pulse are projected at same degree of the sample rotation. All images generated by each neutron pulse are integrated to reduce the statistical fluctuation of neutron beam intensity, thus clear and sharp images can be obtained. There is no limitation on the integration time duration for obtaining clear images with a smaller statistical fluctuation of neutron number. The low-intensity compact neutron source allows the collection of good results. However, this method does not capture phenomena with one or fewer repetitions.
Figure 1 shows the timing chart concept of the sample synchronized stroboscope. The top curve indicates the sample rotation phase, and the square wave pulses are neutron pulses. The rotation cycle is 20 ms. The neutron pulses of “A” are synchronized with the sample rotation phase of 0π (0 ms). The pulses of “B”, “C”, and “D” are synchronized with the timing of the sample rotation phase by delaying by π/4 (5 ms) for each step. Because the neutron pulse timing is synchronized with the sample rotation phase, the images obtained for each timing can be integrated.
Timing chart of the sample synchronized stroboscope.
Block diagram of the timing system.
A block diagram of the timing system of the sample synchronized stroboscope with a proton-accelerator-based neutron source is shown in Fig. 2 The sample rotates and a phase timing signal is generated when the sample reaches a specified rotation angle. Then, the phase timing signal triggers proton timing and detector timing signals. The trigger circuit controls delays and widths of both signals independently.
The RIKEN accelerator-based compact neutron source (RANS) accelerates protons to 7 MeV using Accsys PL-7 Linac [2] and injects them into a beryllium target [12] to produce fast neutrons. The timing of proton acceleration can be arbitrarily determined externally at PL-7. A polyethylene moderator converts neutrons into thermal neutrons [10]. The neutron intensity is approximately 5 × 104 neutron/cm2/s at 5 m from the moderator with a proton current of 100 μA. Therefore, the final thermal neutron pulses are synchronized with the proton pulse timing with some delay. In this study, RANS generates a 60-μs-wide proton pulse. The moderation time for conversion in thermal neutrons is typically 30 μs. The maximum repetition rate is limited to 100 Hz owing to the capacity of RF amplifiers and the cooling of accelerating cavities. The total neutron pulse duration at the moderator surface is about 90 μs. The detector is placed 3 m from the moderator. The average flight time of a thermal neutron is approximately 1.36 ms.
The electric motor turns a disk-shaped paper sample of 2 mm thickness and 150 mm diameter at 150 up to 1800 RPM. A 5-mm-diameter hole is drilled on the sample with an infrared LED on one side of the hole and an infrared sensor on the other. The LED is kept on all the time. The sensor turns on only when the sample rotates at the angle of the LED-sensor assembly and generates a phase timing signal. With delays, the phase timing signal fires the proton timing signal for the start of a proton emission and the detector timing signal for the image capture duration timing.
Radiography setup.
The timing of delay and the width at the “trigger circuit” (BRoadD logic synthesizer module from Bee Beans Technologies Co. Ltd.) [1], which is composed of an FPGA, are controlled remotely by a Python API. The circuit blocks proton emission for 10 ms from the previous proton emission to maintain a repetition frequency of less than 100 Hz. The timing control step takes 25 ns, which is sufficiently accurate for this application.
Radiography experiments were performed. Figure 3 shows a photograph of the setup consisting of an imaging detector, a rotating sample and an infrared LED.
A neutron imaging detector manufactured by Toshiba [4] is used for stroboscopic radiography. It consists of a Gd converter, an image intensifier (II), and a CCD sensor. The II amplifies an image signal from the Gd converter and gates it in time. The amplified and gated image is captured by the CCD sensor. The delay and width for the timing need to be set. Rise time and fall time for the II gate are 154 and 90 μs, respectively. Therefore, the width of the II gate is set to 350 μs to maintain a 100 μs signal gating time for a fully efficient time. The rising and falling edges are indicated by the dashed lines. Fully efficient time is indicated by the solid line. The delay time from the proton timing to the detector timing can be varied to choose the time when the neutron reaches the sample. The example of detector timing with the delay of 1 ms is shown in Fig. 4. The temporal resolution of the stroboscope is determined by the detector timing width of 350 μs instead of the thermal-neutron generation duration of 90 μs.
Detector timing example.
Dependence of CCD count on detector timing delay.
The dependence of the CCD counts on detector timing delay from the proton timing is measured as shown in Fig. 5. It has a peak at 1 ms. The delay of the time is set to 1 ms. It takes 204 μs to reach the center of fully efficient time from the detector timing. The difference between the expected flight time of 1.36 ms and the measured peak time of 1 ms + 204 μs = 1.2 ms is 0.16 ms. We consider that the reason for this difference may be due to delays in the II’s circuitry, the causes of which we do not know yet.
Radiography images.
The L/D of the neutron beam is 20, and the distance between the sample and the detector is 20 mm. The expected spatial resolution is 1 mm.
Pieces of 10% boron-containing rubber of 5 mm thickness are taped to the 2 mm-thick paper disk. The sample rotates at 1800 RPM, which means 33 ms per revolution.
Figure 6 shows the images obtained by stroboscopic radiography with an exposure time of 3 minutes at an average proton current of 10 μA. The top is an image obtained without synchronizing with the sample rotation. At the center is the motor, and to its left is an inductor for electrical noise reduction. The inductor is not shown in the other images because it moved to a different location. Only light strips are visible in the image. The middle and bottom are the images obtained by synchronizing with the sample rotation. The bottom is an image obtained with a delay of 3 ms compared with the middle image, corresponding to a rotation angle of 33°. These images clearly show four pieces of boron-containing rubber. The rotation angle measured from the middle and bottom images in Fig. 6 is 35°, which is in good agreement with that calculated from the rotation speed of 33°.
The sample synchronized neutron stroboscope was used to obtain images of rotating samples using a low-intensity compact neutron source. The stroboscope took clear images of the sample rotating at 1800 RPM. In addition to the expected use of taking still images of an engine in operation for observing the behavior of engine oil, we believe that the system could be used in combination with a diffractometer to measure the strain on the cutting edge of a machine tool in operation.
Footnotes
Acknowledgement
Machine Tool Engineering Foundation, Japan, supported this study from 2021 to 2022. The RIKEN incentive research project funded this study from 2019 to 2020. The authors thank Mr. David Williams for helpful discussions on the RANS timing system.