1. Introduction

In the inertial confinement fusion (ICF) experiment, microchannel plate (MCP) gated x-ray framing cameras are currently used to measure the peak x-ray emission and image the shape of the implosion to improve the implosion performance [1–6]. Such cameras have been developed over the last several decades and contain four basic components, a pinhole array, a MCP imager, a MCP gating pulse generator, and a charge-coupled device (CCD). The pinhole array is placed between the object and the camera, producing an array of x-ray images on the photo-cathode (PC). Then, the x-ray images are converted into electron images and amplified at different times to achieve time-dependent information by using a gated MCP. The electrons outputted from the MCP are accelerated and then hit a phosphor screen, the resulting visible light output corresponding to the incident x-ray image is captured by a CCD [7–9]. The temporal resolution of the camera is limited by the transit time spread of electron through the MCP pore, and currently a temporal resolution of 35-100 ps is achieved [10,11]. Typically, the duration of the x-ray emission “hotspots” produced by the imploding ICF cores is about 100 ps [12]. Therefore, the detailed time history of the x-ray emission “hotspots” cannot be captured by the traditional gated x-ray camera. A new camera with better than 20 ps temporal resolution is desired to characterize the implosion performance more accurately [13]. Fortunately, a new generation of x-ray framing cameras named DIlation X-ray Imager (DIXI) with a temporal resolution of 5 ps has been developed by using pulse-dilation technology [14–18]. The pulse-dilation technology was suggested firstly by Prosser to improve the bandwidth of the electronic detectors by imparting a velocity dispersion of the electron beam [19–22], and Hilsabeck et al. introduced it into the framing camera [14]. In the DIXI, the PC is separated with the MCP, and an anode mesh is formed between them. The PC is applied with a pulse voltage plus a negative DC high voltage, while the anode mesh is held at ground potential, which leads to a time-dependent electric field between PC and anode mesh. The electric field strength will be decreased with time during the photoelectron birth, while the incident light is synchronized with the rising edge of the pulse voltage applied on PC. Then, an electron energy dispersion is achieved. The electrons born early in time have larger energy than those coming later, and transit faster in the electron drift space between anode mesh and MCP, resulting in a temporal magnification of the electron pulse. Finally, the dilated electron pulse is sampled by the gated MCP. In other words, a slice of the incoming signal, which is temporally dilated up to 50 times, is captured by the gated MCP [16]. The temporal resolution is greatly improved.

In this work, a dilation x-ray detector is presented, and an experiment is carried out to demonstrate the single area synchronization between PC pulse and MCP gating pulse in the detector without 1:1 image ratio [23]. It has the following different points compared to the DIXI. The first one is the imaging system for the electrons from PC onto MCP. A uniform coaxially guiding magnetic field is achieved by four magnet coils, and an image de-magnification of three times is obtained by the large uniform axial magnetic field electron lens in DIXI [14,16,23]. The photoelectrons with a large energy spread can be imaged onto the MCP in this magnetic field. However, a large current through the solenoid coils should be used to generate the axial magnetic field. The large current may lead to a current thermal effect. Fortunately, the thermal effect could be reduced by using a pulsed current. Engelhorn et al. [13] have successfully used the pulse with a peak current of ~1 kA and a duration of ~1-10 ms to produce a required magnetic field strength [13]. In this detector, an axially symmetric non-uniform magnetic field produced by two large aperture magnetic lenses is used to image electrons from PC onto MCP with the image ratio of 2:1. Such magnetic lens is typically used in the electron microscope and streak camera to improve the focusing ability and imaging quality of the optical electronic [24–26]. The magnetic lens imaging system can obtain high spatial resolution and small current thermal effect. However, it has chromatic aberration. The photoelectron energy spread will be limited by the imaging system. In additional, the spatial resolution is worse with the increasing off-axis distance. Four or more magnetic lenses with larger diameter will be formed in our next generation imaging system to reduce the chromatic aberration and improve the spatial resolution on the PC edge. Moreover, the size of this detector is smaller than that of DIXI. In DIXI, there are four 40 cm diameter coils. The axial length of each coil is 8 cm and the gap between each adjacent two of them is 15 cm [14]. In this detector, two magnetic lenses with an outer diameter of 256 mm, an inner diameter of 160 mm, and an axial length of 100 mm are used. Furthermore, the temporal resolution measurement methods are also different. In DIXI, the temporal resolution is measured using a Mach–Zehnder interferometer which outputs a pair of laser pulses. A vertical arrow and a horizontal arrow are generated by the laser pulses. The arrival time of the horizontal arrow is fixed and centered within the DIXI gating interval. The vertical arrow arrival time is changing by varying its optical path length. The vertical arrow appears and then disappears over a temporal delay adjustment of ~5 ps, which demonstrates that the temporal resolution of the DIXI is 5 ps [14]. In this paper, the temporal resolution is measured by using a fiber bunch, which is made up of thirty fibers with different length [23].

2. Detector description

The dilation x-ray detector, shown in Fig. 1, is consisted of four main components, a pulse-dilation device, a magnetic lens imaging system, a gated MCP imager, and a pulse generator. The pulse-dilation device is made up of three transmission x-ray photo-cathodes, an anode mesh, and a drift space. The x-ray photo-cathodes are coated with 80 nm Au on a 90 mm diameter C₈H₈ film, shown in Fig. 2. Each PC has 12 mm width, and there is a gap of 10 mm between two neighboring photo-cathodes. The excitation pulse transmits from input port connector to a Cu tapered transmission line etched onto the printed circuit board, and then to the PC. The Cu tapered transmission line is used to transform the impedance and is also formed in the output side to guide the excitation pulse from PC to output port connector. A piece of gold sheet is used as an electrical contact between the PC and the Cu tapered transmission line. The accelerating gap between the PC and the anode mesh is 1.5 mm. The PC is applied with an excitation pulse plus a negative DC high voltage bias, while the electroformed nickel mesh with 10 lp/mm is grounded, which leads to an electron energy dispersion. The length of the drift space between anode mesh and MCP is 50 cm. The electrons travel in the drift space with electrons born earlier moving faster than those born later, which produces a temporal magnified of the electron pulse.

 

Fig. 1 Time-resolved x-ray detector. (a) Schematic diagram. (b) Photograph.

Download Full Size | PDF

 

Fig. 2 Photograph of the transmission x-ray PC. The labels A and B represent different positions on the middle PC. Point A is 25 mm from the PC left-most, point B is 40 mm from A.

Download Full Size | PDF

The magnetic lens imaging system consisting of two identical large aperture magnetic lenses is used to image the photoelectrons from PC onto MCP. An image de-magnification of two times is obtained. The distance between the PC and the center of the first magnetic lens is 12.5 cm, and the second lens to the MCP is 9 cm. The DC current of the first lens is 0.198 A, and that is 0.31 A for the second lens. The annular magnetic lens composes of a soft iron frame and a 1320-turn copper coil. A 256 mm outer diameter, 160 mm inner diameter, and a 100 mm axial length are formed. A circular slit with 4 mm width in the inner cylinder is used to leak the magnetic field from soft iron to drift space to form a magnetic lens. An axially symmetric non-uniform magnetic field is produced by this magnetic lens. The on-axis distribution of the magnetic field strength produced by each magnetic lens is an approximate Gaussian distribution.

The gated MCP imager contains a MCP, a phosphor screen, and a CCD. The MCP is an array of parallel electron multiplier channels. A MCP with 56 mm in diameter and 0.5 mm in thickness is used. The pore diameter is 12 μm, and the MCP bias angle is 6◦. The MCP imager is formed with a microstrip transmission line structure to transmit the fast gating pulse [7]. Three microstrip transmission lines are deposited on the MCP input surface with 500 nm Cu overlaid by 100 nm of Au. The whole MCP output surface is also deposited, and used as a ground plane. The width of each microstrip line is 8 mm, and the gap between two neighboring microstrip lines is 3 mm. The gating pulse is transmitted from imager input port to MCP microstrip line by a tapered transmission line. The MCP microstrip line and tapered transmission line are connected by a gold foil. The MCP gate width and gain of the signal are achieved by passing a gating pulse along the microstrip line [8,9]. The phosphor screen coated on a fiber-optic faceplate is placed 0.5 mm from the MCP output surface, and is applied with a 4 kV positive DC voltage. A cooled CCD is in contact with the faceplate output to capture the visible images.

The pulse generator produces the PC excitation pulse and the MCP gating pulse. Firstly, there are several avalanche transistors with a Marx bank configuration to generate high voltage fast step pulses. Then, one of the output pulses is used as the PC excitation pulse, and the other pulse is shaped by using avalanche diodes to obtain the MCP gating pulse [27]. The PC excitation pulse with gradient of approximately 3.1 V/ps is obtained and shown in Fig. 3(a). The MCP gating pulse shown in Fig. 3(b) has width of 225 ps and amplitude of −1.8 kV. The incident light should be synchronized with the PC excitation pulse rising edge to produce pulse dilation. In this paper, the ultraviolet (UV) laser pulse is applied at about 200 ps after the beginning of the PC excitation pulse.

 

Fig. 3 Wave forms of the pulses. (a) PC excitation pulse with gradient of 3.1 V/ps. (b) MCP gating pulse with amplitude of −1.8 kV and width of 225 ps.

Download Full Size | PDF

The time domain reflectometry (TDR) curves of the MCP microstrip line and the PC transmission line are measured using a high speed sampling oscilloscope (LeCroy WE9000). The TDR curve shows that the time of the gating pulse propagating on the microstrip line deposited on the MCP with 56 mm length is about 300 ps. Then, the measured velocity of the gating pulse traveling on the MCP microstrip line is about 1.87 × 10⁸ m/s [28]. It also can be measured that the excitation pulse travels on the PC transmission line with a velocity of about 1.81 × 10⁸ m/s, which is almost the same as the MCP gating pulse velocity. The schematic diagrams of the synchronous gating are shown in Fig. 4. The PC excitation pulse transmits on the PC and the gating pulse on the MCP. The photoelectrons emitted from PC at points A and B are imaged onto MCP at A' and B', respectively. Assuming that the time of flight (TOF) of the electrons from PC to MCP is T. That is to say, the time of electrons transiting from A to A' is T, and that is also T for the electrons from B to B'. The excitation pulse travels from PC left side to the right, and the gating pulse propagates from left to right on the MCP. Assuming the traveling distance from A to B is 2L.

 

Fig. 4 The schematic diagrams of the synchronous gating. The excitation pulse transmits on the PC and the gating pulse on the MCP. (a) The image ratio is 1:1, the transmission directions of the PC excitation pulse and the MCP gating pulse are the same. (b) The image ratio is 2:1, the transmission directions of the two pulses are the same. (c) The image ratio is 2:1, the two pulses travel with opposite directions.

Download Full Size | PDF

While the image ratio is 1:1, the distance between A' and B' is 2L, shown in Fig. 4(a). Assuming that at the moment t₁, PC excitation pulse arrives at point A to produce electron pulse-dilation in this area. The dilated electron signal will arrive at A' on MCP at t₁ + T due to the TOF. The MCP gating pulse is timed relative to the PC excitation pulse to gate the dilated electron signal, so the MCP gating pulse arrives at point A' at t₁ + T. Assuming the time of the PC excitation pulse or the MCP gating pulse traveling distance with L is Δt. Then, at t₁ + 2Δt, the PC excitation pulse arrives at B. The MCP gating pulse will arrive at B' at t₁ + T + 2Δt. The electrons emitted from B at time t₁ + 2Δt are synchronized with the PC excitation pulse. The dilated electron pulse will arrive at B' at t₁ + 2Δt + T, which is also the moment of the MCP gating pulse arriving at B'. Then, the dilated electron pulse is synchronized with the MCP gating pulse, which can produce gating image. In the whole microstrip line, the MCP gating pulse can be timed relative to the PC excitation pulse to gate the dilated electron signal.

While the image ratio is 2:1, the distance between A' and B' is L, shown in Fig. 4(b). Assuming that at the moment t₂, PC excitation pulse arrives at point A. Then, the dilated electron signal will arrive at point A' at t₂ + T. Assuming the MCP gating pulse arrives at A' at t₂ + T, and synchronizes with the dilated electron signal to produce the gating image in area A'. The PC excitation pulse will arrive at point B at t₂ + 2Δt, and the MCP gating pulse will arrive at point B' at t₂ + T + Δt. The dilated electron pulse arrives at B' at t₂ + 2Δt + T, which is not the moment of the MCP gating pulse arriving at B'. The MCP gating pulse arrives at B' Δt earlier than the dilated electron pulse. Then, the dilated electron pulse is unsynchronized with the MCP gating pulse at B'. The electrons in area B' will be absorbed by the MCP, which leads to no signal outputting in this area. Actually, at t₂ + 2Δt + T, the dilated electron pulse arrives at B', whereas the MCP gating pulse arrives at point C'. In short, the MCP gating pulse can be timed relative to the PC excitation pulse to gate dilated electron signal in single area, and they are unsynchronized in the other area. For instance, they are only synchronized in the area A', and they are unsynchronized in the area B'. Therefore, single area with very narrow range can produce gating image, and other large areas are no outputting signal. While the image ratio is other proportions but not 1:1, it can be obtained the same results.

While the image ratio is 2:1 and the transmission directions of the pulses are opposite, the schematic diagram of the synchronous gating is shown in Fig. 4(c). Assuming that at t₃, PC excitation pulse arrives at point A. Then, the dilated electron signal arrives at A' at t₃ + T. Assuming the MCP gating pulse arrives at A' at t₃ + T, and synchronizes with the dilated electron signal. The PC excitation pulse will arrive at B at t₃ + 2Δt, and the dilated electron pulse arrives at B' at t₃ + 2Δt + T. However, the MCP gating pulse arrives at B' at t₃ + T - Δt, which is 3Δt earlier than the dilated electron pulse arriving time. Actually, at t₃ + 2Δt + T, the dilated electron pulse arrives at B', whereas the MCP gating pulse arrive at C'. The point C' is 3L from B'. Therefore, the dilated electron pulse cannot synchronize with the MCP gating pulse at B', it also has single area synchronization effect.

3. Experimental verification

It can be seen from Fig. 4(b) that, in the dilation x-ray detector with image ratio of 2:1, while the MCP gating pulse synchronizes with the dilated electron signal at point A', they will be unsynchronized at other area, such as B'. The dilated electron signal arriving time at point B' is t₂ + 2Δt + T, whereas the MCP gating pulse arrives at B' at t₂ + T + Δt. The arriving time at B' of the MCP gating pulse is Δt earlier than that of the dilated electron signal. In this paper, an experiment is presented to validate the phenomenon in Fig. 4(b). A fiber bunch consisting of thirty fibers with different length is used to obtain the earlier MCP gating pulse arriving time of Δt at B'. The experimental setup is shown in Fig. 5(a). The application of this detector is mainly detecting x-rays. However, we lack x-ray source for the resolution measurement and the synchronous gating demonstration. Therefore, the resolution test and the demonstration are performed using a UV pulse, which has been utilized in the framing or streak camera test [11,14,29]. The incident x-ray or UV pulse is converted into photoelectrons. The electron energy spread produced by UV light is smaller than that by x-ray [29]. Thus, the test using UV light underestimates the transit time dispersion for application in x-ray detection. The ultimate temporal resolution is the transit time dispersion, which is dependent upon the initial electron energy spread and the space charge effect [30]. However, the transit time dispersion is much less than the temporal resolution in this paper. Therefore, the temporal resolution difference due to the UV light or the x-ray test could be negligible. The temporal resolution is mainly determined by four factors, the PC bias voltage, the gradient of the PC excitation pulse, the drift length, and the MCP gate width [14–16]. The work function of bulk Au is 5.1 eV, which is larger than the photon energy of the 266 nm light used in the testing (4.66 eV). However, the contamination of the photocathode surface could reduce the work function to 4.2-4.3 eV [29]. In addition, for the Au thin film below 100 nm thick by the physical vapor deposition, the work function maybe 4.2 eV [31]. A third harmonic of a Ti-sapphire laser system outputting two laser beams with wavelengths of 266 and 800 nm is used. The UV 266 nm laser pulse with width of 130 fs is utilized to excite the photoelectrons from PC, and the 800 nm laser beam illuminates the p-i-n detector to generate an electrical trigger signal. The 266 nm UV laser is firstly reflected by mirror M₂, and then illuminates the fiber bunch with thirty different length fibers. The length of each fiber is measured, which is increased by a step of 2 mm, and the length error is within 0.2 mm. The fiber bunch output thirty laser pulses at different time. The measured delay time of 10 ps between each two adjacent fibers is obtained by using a streak camera [23]. The thirty laser pulses are imaged on the PC by the lenses L₁ and L₂ to excite photoelectrons with different birth time. In the fiber bunch output port, thirty fibers are arrayed as shown in Fig. 5(b). The fiber bunch has three rows, and there are ten fibers in each row. The shortest fiber is labeled number one. While the fiber number adds by one, the delay time will be increased by 10 ps. The spatial distance of each two neighboring fibers is 0.2 mm. The long side of the fiber image is vertical to the traveling direction of the PC excitation pulse. The laser beam with 800 nm wavelength is reflected by mirror M₁ and illuminates the p-i-n detector to generate an electrical signal. The electrical signal is delayed by a delay circuit and used to trigger the pulse generator to output PC excitation and MCP gating pulses. The delay circuit is adjusted precisely to synchronize the UV laser pulses with the rising edge of the PC excitation pulse. Then, the photoelectrons excited from PC will be accelerated by a time varying electric field between the pulsed PC and the grounded mesh, and an electron velocity dispersion will be achieved. The electron signal arriving at MCP will be dilated. Finally, the MCP gating pulse is timed to gate the dilated electron signal. The delay time between the PC excitation pulse and the MCP gating pulse is about 17.3 ns.

 

Fig. 5 (a) Schematic diagram of the experimental setup. (b) The array of the fiber bunch output port. The shortest fiber is labeled number one. The delay time is increased by 10 ps, while the fiber number adds by one.

Download Full Size | PDF

The excitation pulse travels from PC left side to the right, and the gating pulse propagates on MCP from left to right, corresponding to Fig. 4(b). While the UV laser pulse excites the photoelectron from the gold cathode at point A shown in Fig. 2, the measured static image, and gating image with pulse-dilation are shown respectively in Figs. 6(a) and 6(b). The fiber image at the top right has the shortest delay time and there is a spatial distance of 0.25 mm between each two adjacent images. The static image in Fig. 6(a) is measured with a static DC bias of −3 kV applied to the PC, and a static DC bias of −700 V on the MCP. The gating image in Fig. 6(b) is measured with a DC bias of −3 kV overlapped by a 3.1 V/ps gradient excitation pulse on the PC, and a gating pulse plus a −400 V DC bias on the MCP. Figures 6(a) and 6(b) are the original images obtained by the CCD. The signal out of the gating image in Fig. 6(b) is shown in Fig. 6(c). The horizontal coordinate in Fig. 6(c) is the time of the laser pulse arriving at PC, which is related to the arriving time of the dilated electron pulse on the MCP. The time of the fiber image at the top right is set as 0 ps, and it is increased by a step of 10 ps while the number of the fiber image increases one. Each fiber image on the MCP is a dilated electron pulse. Therefore, there are thirty electron pulses with different arriving time on the MCP. While one fiber image on the MCP is synchronized with the gating pulse, this fiber image can be gated and gained by the MCP. The CCD could capture the output signal from this fiber image. Otherwise, if one fiber image on the MCP is unsynchronized with the gating pulse, the fiber image will be absorbed by the MCP, and there is no signal outputting to the CCD. The final results in Fig. 6(c) are calibrated by the static results from Fig. 6(a). The solid points in Fig. 6(c) are the experimental results, and the solid line is the Gaussian fitting curve for the points. The full width at half maximum (FWHM) of the gain versus time curve is regarded as temporal resolution. It can be seen from Fig. 6(c) that the temporal resolution of the detector is about 14 ps by using the pulse-dilation technology. Moreover, in the gain versus time curve, the time of the peak gain could represent the synchronous position of the dilated electron pulse with the MCP gating pulse. In Fig. 6(c), the time of the peak gain is about 183.2 ps. Therefore, while the photoelectrons are emitted from point A, the MCP gating pulse is synchronized with the electron pulse from fibers 19 and 20.

 

Fig. 6 (a) Static image, while −3 kV DC bias without excitation pulse is applied on the PC, and the MCP is applied with −700 V. There is a spatial distance of 0.25 mm between each two adjacent images. (b) Gating image with pulse-dilation in point A, the PC is applied with an excitation pulse plus −3 kV DC bias, and a DC bias of −400 V overlapped by a gating pulse are applied on the MCP. (c) Signal out of the gating image in (b).

Download Full Size | PDF

Then, the thirty laser pulses are imaged onto point B to excite photoelectrons in this PC area. Point B is 40 mm from point A, shown in Fig. 2. The delay circuit is adjusted precisely to ensure that the synchronous position of the UV laser pulses with the PC excitation pulse in point B is the same as that for A. The delay time of 17.3 ns between the PC excitation pulse and the MCP gating pulse is remained unchanged. In brief, all of the experimental conditions in point B are the same as that for A. If the image ratio is 1:1, the experimental results will be the same as that in point A, which is that the MCP gating pulse is synchronized with the dilated electron pulse from fibers 19 and 20. However, the image ratio is 2:1, and it can be seen from Fig. 4(b) that the time of the MCP gating pulse arriving at B' is Δt earlier than that of the dilated electron signal from fibers 19 and 20. From Fig. 4(b), it can be obtained that Δt is the time of the PC excitation pulse traveling length with 20 mm, which is about 110 ps. If the total temporal width of the dilated electron pulses arriving at B' is shorter than 110 ps, the dilated electron pulses could not synchronize with the MCP gating pulse. A fiber bunch made up of thirty different length fibers is used to achieve a long temporal width of the electron pulses. The total temporal width of the fiber bunch is about 290 ps on PC, and the temporal width of the dilated electron pulses on MCP will be larger due to the pulse-dilation. Therefore, the MCP gating pulse could be synchronized with the dilated electron signal, and the number of the synchronized fiber will be less than 19 according to the theoretical results of Fig. 4(b). The measured gating image without an electron pulse-dilation, and gating image with a pulse-dilation technology are shown respectively in Figs. 7(a) and 7(b). The gating image in Fig. 7(a) is measured with a static −3 kV DC bias on the PC, and a gating pulse plus a −400 V DC bias on the MCP. The gating image in Fig. 7(b) is measured with a DC bias of −3 kV overlapped by the excitation pulse on the PC, and the voltages applied to the MCP are the same as that for Fig. 7(a). The signals out of the gating images in Figs. 7(a) and 7(b) are shown in Figs. 7(c) and 7(d), respectively. Figures 7(c) and 7(d) show that the temporal resolution of the detector without pulse-dilation is about 82 ps, and is improved to 22 ps by using pulse-dilation technology. The temporal magnification factor is about 3.7 ×. Furthermore, the time of the peak gain in point B is about 156.2 ps, which is 27 ps earlier than that in point A. The 27 ps is the time difference of the laser pulses arriving at PC, it will be magnified to 100 ps on the MCP due to the 3.7 × temporal magnification factor. Therefore, in point B', the MCP gating pulse is synchronized with the dilated electron pulse 100 ps earlier than that in point A'. The 100 ps experimental result is agreed well with the theoretical result of 110 ps.

 

Fig. 7 (a) Gating image without pulse-dilation in point B. (b) Gating image with pulse-dilation in point B. (c) Signal out of the gating image in (a). (d) Signal out of the gating image in (b).

Download Full Size | PDF

As a contrast, another experiment is provided to acquire the results while the pulses travel with the opposite directions. The excitation pulse travels from PC left side to the right, whereas the gating pulse propagates from MCP right to left, corresponding to Fig. 4(c). While the UV laser pulses synchronize with the PC excitation pulse at point A, the measured gating image with pulse-dilation is shown in Fig. 8(a). Then, the thirty laser pulses are image onto point B, and the gating image in this point is shown in Fig. 8(b). The experiment conditions in point B are the same as that for A. The signals out of the gating images in Figs. 8(a) and 8(b) are shown in Fig. 8(c). It can be seen from Fig. 8(c) that the time of the peak gain is about 184.3 ps in point A, and that is 103 ps in point B. The time in point B is 81.3 ps earlier than that in A on the PC, and it is 300.8 ps on the MCP because of the 3.7 × temporal magnification factor. Therefore, the time of the dilated electron pulse which can be synchronized with the MCP gating pulse in point B' is 300.8 ps earlier than that in point A'. The experimental result of 300.8 ps is agreed well with the 330 ps theoretical result.

 

Fig. 8 The results while the pulses travel with the opposite directions. (a) Gating image with pulse-dilation in point A. (b) Gating image with pulse-dilation in point B. (c) Signals out of the gating images in (a) and (b).

Download Full Size | PDF

Both of the theoretical and experimental results show that the single area synchronization effect exists in the dilation x-ray detector without 1:1 image ratio. Although the experimental verification is performed by an ultra-short pulse with width of 130 fs, the single area synchronization effect will still exist while the temporal width of the optical signal is longer, such as 100 ps. In the dilation x-ray detector, photoelectrons are imaged from PC onto MCP by the magnetic lens imaging system. While the currents of the two lenses are respectively 0.198 A and 0.31 A, the photoelectrons with a certain energy range could be imaged on the MCP. The measured energy range is about 2.37 keV to 2.43 keV with the energy difference of about 60 eV. The photoelectrons with other energy cannot be imaged. While the PC bias voltage is changed, the currents of the two lenses should be changed to insure the photoelectrons imaged. In short, the PC voltage with a certain range corresponds to the fixed lenses currents. The photoelectron energy is varying with time because of the time-dependent electric field between PC and anode mesh. The photoelectrons with energy difference of 60 eV can be imaged, and the PC excitation pulse gradient is 3.1 V/ps. Then, the temporal width of the effective photoelectron is about 19 ps, although the incident light lasts 100 ps or longer. From Fig. 4(b), it is easy to obtain the results that the MCP gating pulse can be timed relative to the PC excitation pulse to gate dilated electron signal in single area, and they are unsynchronized in the other area. Along the PC, the single area with maximum length of 6.9 mm can produce gating image, and other areas are no outputting signal. The 6.9 mm length is corresponding to the pulse traveling on the PC with 38 ps. Therefore, the single area synchronization effect still exists while the optical signal has longer temporal width, and the synchronized area is larger while the optical signal is longer. However, because of the 19 ps effective photoelectron, the maximum length along PC which can produce gating image is about 6.9 mm, and that is 3.45 mm on MCP. The single area synchronization effect will be non-existent in the DIXI, because the four magnet coils used in DIXI allow photoelectrons with large energy spread imaged from PC onto the MCP [14].

The single area synchronization effect will lead to a single area with very narrow range can obtain gating image, and other large areas are no outputting signal. It is a problem for accurate x-ray measurement in ICF experiments. In the ICF experiment, gated x-ray detectors have been developed to measure the time-dependent shape of the implosion. A pinhole array or Kirkpatrick–Baez microscope is placed between the object and the detector, producing an array of x-ray images in the PC [32–34]. The time of the x-ray images are different in the different PC positions to achieve time-dependent information. While the detector has single area synchronization effect, only an x-ray image in one PC position can be captured, and the x-ray images in other PC positions cannot be obtained. A lot of information about x-ray emission will be missed. Therefore, the single area synchronization effect should be overcome to obtain all of the information. While the image ratio is non 1:1, the detector could use a MCP gating pulse with width larger than 500 ps to reduce the single area synchronization effect. Moreover, a 1:1 image ratio can avoid this effect. Furthermore, the electron TOF from PC to MCP is unchanged in the above theoretical analysis and experimental demonstration. The TOF could be decreased to eliminate the single area synchronization effect. The photoelectrons could be synchronized with the different part of the PC excitation pulse rising edge to obtain a higher acceleration voltage, which can lead to a smaller TOF. Then, the earlier MCP gating pulse arriving time of Δt could be compensated, and the dilated electron pulse may be synchronized with the MCP gating pulse. The magnetic lens imaging system allows photoelectrons with the energy from 2.37 keV to 2.43 keV can be imaged onto MCP. Then, the Δt which can be compensated is within about 215 ps [14]. It is worth noting that different parts of the PC excitation pulse rising edge being synchronized may bring measurement error for accurate time-dependent x-ray measurement in ICF experiments.

4. Conclusions

A time-resolved x-ray detector consisting of a pulse-dilation device, a magnetic lens imaging system, a gated MCP imager, and a pulse generator is described. The temporal resolution of the detector is about 14 ps by using the pulse-dilation technology, better than that of 82 ps for the gated MCP. Furthermore, a theoretical analysis of the single area synchronization effect is provided. While the image ratio of the detector is not 1:1 and the TOF of the electrons is identical, the MCP gating pulse can be timed relative to the PC excitation pulse to gate dilated electron signal in single area, and they are unsynchronized in the other area. An experiment is carried out to validate this effect. A fiber bunch consisting of thirty fibers with different length is used to obtain the earlier MCP gating pulse arriving time comparing to that of the electron pulse. The experimental results show that while the photoelectrons are emitted from point A, the MCP gating pulse is synchronized with the electron pulse from fibers 19 and 20. However, in point B, the MCP gating pulse is synchronized with the electron pulse 100 ps earlier than that in point A. The 100 ps experimental result is agreed well with the theoretical result of 110 ps. The single area synchronization effect leads to a single area with very narrow range can obtain gating image, and other large areas are no outputting signal, which may result in an information missing. Therefore, the single area synchronization effect should be overcome to obtain all of the information. The magnetic lens imaging system that only allows photoelectrons with a certain energy range can be imaged onto the MCP, which may lead to the single area synchronization effect. An imaging system enabling photoelectrons with large energy spread imaged onto MCP could be used to reduce this effect. In addition, the single area synchronization effect can also be reduced by using a MCP gating pulse with width larger than 500 ps. Moreover, the detector developed with a 1:1 image ratio can avoid this effect. Furthermore, the decreasing electron TOF can compensate the earlier MCP gating pulse arriving time, and then the single area synchronization effect could be eliminated. The electrons generated from one point on PC have continuously changing TOF. The electrons with smaller TOF can be synchronized with the gating pulse in the area near the MCP microstrip line output end. Therefore, the single area synchronization effect will be avoided in the dilation x-ray detector.