Electron beam charge diagnostics for laser plasma accelerators

Electron Beam Charge Diagnostics for Laser Plasma Accelerators K. Nakamura, 1 A. J. Gonsalves, 1 C. Lin, 1, 2 A. Smith, 1 D. Rodgers, 1 R. Donahue, 1 W. Byrne, 1 and W. P. Leemans 1, 3 Lawrence Berkeley National Laboratory, University of California, Berkeley, CA 94720, USA Peking University, Beijing, 100871, P. R. China Department of Physics, University of California, Berkeley, CA 94720, USA (Dated: July 19, 2011) Abstract A comprehensive study of charge diagnostics is conducted to verify their validity for measur- ing electron beams produced by laser plasma accelerators (LPAs). First, a scintillating screen (Lanex) was extensively studied using sub-nanosecond electron beams from the Advanced Light Source booster synchrotron, at the Lawrence Berkeley National Laboratory. The Lanex was cross- calibrated with an integrating current transformer (ICT) for up to the electron energy of 1.5 GeV, and the linear response of the screen was conﬁrmed for charge density and intensity up to 160 pC/mm 2 and 0.4 pC/(ps mm 2 ), respectively. After the radio-frequency accelerator based cross- calibration, a series of measurements was conducted using electron beams from an LPA. Cross- calibrations were carried out using an activation based measurement that is immune to electro- magnetic pulse noise, ICT and Lanex. The diagnostics agreed within ±8 %, showing that they all can provide accurate charge measurements for LPAs. PACS numbers: 06.20.fb, 07.77.Ka, 29.20.-c, 29.40.Mc, 42.79.Pw, 52.38.Kd Keywords: laser plasma accelerator, Lanex, ICT, activation, charge calibration


I. INTRODUCTION
Laser plasma accelerators (LPAs) [1] have shown remarkable progress over the past decade, driven in part by advances in laser technology. The production of quasi-monoenergetic electron beams (e-beams) with energies of the order of 100 MeV in a few millimeters was demonstrated in 2004 [2][3][4]. In 2006, the production of GeV electron beams was demonstrated in just a few centimeters [5,6], using a discharge capillary based guiding structure [7].
Detection by surface barrier detectors and scintillators with photomultipliers were popular methods in early LPA experiments due to their high sensitivity. Scintillating fibers added the capability of imaging, but their cost became significant for large scale systems that cover a broad energy range. Scintillating screens and IPs are now widely used as they allow large areas to be imaged with reasonable sensitivity and cost. Combined with a dipole magnet, they can provide e-beam energy spectrum information over a broad spectral range [26][27][28].
As an IP is capable of accumulative measurements, it can have an advantage in sensitivity. By using mono-energetic e-beams from RFAs, the signal of the IP against relativistic electrons up to 1 GeV of electron energy was experimentally calibrated against the e-beam charge measured by a Rogowsky coil or current transformer [29,30]. The range of applicable charge density was extensively studied with 40 MeV e-beams from an RFA [31] using an ICT and Faraday cup measurements for reference.
As LPAs can have significant shot-to-shot fluctuation in the electron beam properties depending on the input parameters, high repetition-rate single-shot measurements of e-beams are important. Scintillating screens are ideal for typical 10 Hz repetition rate operation since the fluorescence decay time is less than 10 ms. Among many kinds of scintillating screens from various manufactures, the ones with Terbium doped Gadox (Gd 2 O 2 S : Tb) as an active layer have been commonly used in the LPA community. The light yield from the screens has been experimentally calibrated against ICTs by using e-beams from RFA with 3 -8 MeV electron energy [26] and 40 MeV electron energy [32]. By using broadband electron beams from an LPA, sensitivity for 1 to 80 MeV electrons was experimentally calibrated against IPs [28]. Although simulations suggested that the scintillating screens were energy insensitive above a few MeV, a detailed experimental study with electrons above 80 MeV has not been reported until now. Since recent progress in LPA research has pushed attainable energy to 1 GeV and beyond [5,6,33,34], it is important to experimentally explore the applicable energy range of the screens at those energies.
Faraday cups and ICTs have been used as reliable charge diagnostics in the RFA community [19,31]. Since Faraday cups have to physically capture electrons, their size can be fairly large to stop GeV electrons. In contrast, ICTs are non-destructive, energy independent and compact. Despite all of the favorable features of the ICT for LPA, its use for LPA produced e-beams has been questioned in recent studies. It was reported in Ref. 26 that the ICT overestimated the e-beam charge by more than a order of magnitude compared to the measurement based on the RFA-calibrated scintillating screen, and the source of discrepancy was attributed to the electromagnetic pulse (EMP) from the laser-plasma interaction.
Another cross-calibration using LPA produced e-beams was done in Ref. 35, where it was reported that an ICT overestimated the charge by a factor of about 3 -4 compared to the IP based charge measurements. Both studies indicated that further cross-calibration measurements and detailed investigations were necessary regarding the use of the ICT in a harsh laser-plasma environment.
In this paper, we have experimentally studied the sensitivity of Lanex Fast scintillating screen (Kodak, Rochester, NY, United States) using e-beams provided by a booster synchrotron at the Advanced Light Source (ALS), Lawrence Berkeley National Laboratory (LBNL). The energy of the electron beam was varied from 106 MeV to 1522 MeV covering the as yet unexplored energy range. In addition, the linearity of the response against differ-ent charge density and charge intensity has been extensively studied. This study provides essential information for the Lanex to be a charge diagnostic for GeV e-beams.
A comprehensive study of charge diagnostics for LPAs was performed using an ICT, a Lanex Fast screen (Lanex), and an activation based measurement [23]. The activation based measurement is intrinsically RF noise tolerant and independent of the e-beam intensity. Therefore, it can provide an accurate reference for LPA produced e-beams. This RFA based calibration of Lanex was benchmarked against the activation based measurement using LPA produced e-beams. Also the ICT was cross-calibrated against Lanex using LPA produced e-beams. The results show that the Lanex and ICT can be accurate diagnostics for an LPA.
The calibration of the Lanex with RFA produced e-beams is shown in Sec. II, the crosscalibration of the ICT, Lanex, and activation based measurement with LPA produced ebeams is described in Sec. III, and conclusions are provided in Sec. IV.

A. Experimental Setup
The light yield of the Lanex for relativistic mono-energetic electron beams was studied at the booster-to-storage ring (BTS) beamline of the ALS, LBNL. The ALS linear accelerator (Linac) provided 50 MeV e-beams with a micro bunch duration of ≃30 ps in full-width halfmaximum (FWHM) [36], and one or two micro bunches with 8 ns separation determined by the 125 MHz electron gun cathode pulser. The total charge was controlled by changing the bias voltage of the thermionic electron gun (gun bias voltage). Each micro bunch contained up to ≃450 pC giving the maximum macro bunch charge around 900 pC. The bunches were produced at 0.5 Hz of repetition rate so that thermal loading of the screen was not significant.
The e-beam from the Linac was injected into the booster and further accelerated up to 1522 MeV. The e-beam was then extracted from the booster by the extraction magnet system, a part of which is shown in Fig. 1. In normal operation, the e-beam is extracted at the top of the booster magnets' current ramp, which gave 1522 MeV, the highest ebeam energy. By changing the e-beam extraction trigger timing and field strength of the extraction magnets, an e-beam with lower energy was extracted from the booster and sent to BTS beamline.
Inside of the BTS beamline, there was a magnet (FE-B1, see Fig. 1) to send e-beams to the Lanex calibration setup that is shown in Fig. 2, instead of to the storage ring. The FE-B1 magnet was also used to determine the energy of the e-beam using the following procedure. First, with the FE-B1 magnet off, the e-beam was aligned to the center of the beamline by using 4 steering magnets and 2 monitors that were installed before and after the FE-B1 magnet (see Fig. 1 for locations) to establish the injection angle and position offset consistently for e-beams with different energies. The excitation current of the FE-B1 magnet was then scanned to find the e-beam at the Lanex screen installed in the end of the beamline. Using the excitation current for 1522 MeV e-beam as the reference, the energies of e-beams were deduced.
The longitudinal bunch duration, momentum spread, and transverse emittance of the e-beam evolved during the acceleration in the booster. Although they were not measured during the experiments described in the paper, they were modeled theoretically [37]. The longitudinal bunch duration was estimated to be ≃200 ps FWHM for 1522 MeV e-beams.
Note that the bunch duration of the injected e-beam was ≃30 ps. In order to give a lowerlimit estimate to the charge density and charge intensity at the Lanex screen, a bunch duration of 200 ps FWHM was assumed for all the energies explored in this paper. For the charge density estimate, e-beams with 2 micro bunches was considered to be one bunch, because the micro bunch separation time of 8 ns was much shorter than the fluorescence decay time. For the charge intensity, only e-beams with 1 micro bunch were considered.
The root-mean-square (rms) momentum spread of the e-beam was estimated to be ≃ 0.4 % to ≃ 0.063 %, so hence e-beams were considered to be mono-energetic.
The incident angle of the electrons on the Lanex screen can in principle affect the light yield as well, because it defines the interaction length between the electron and the active layer of the Lanex. The last focusing quadrupole magnet was located before the second pair of the steering magnets, and this was ∼5 m away from the Lanex -ICT setup. The magnet was tuned to provide a focus at the Lanex. The largest possible convergence angle of the e-beam was estimated to be 3.2 mrad based on the clear aperture of the quadrupole magnet, which is 32.5 mm in diameter. For the normalized transverse emittance, the evolution during the acceleration is estimated to be in the range between ≃ 3 mm mrad and ≃ 3 × 10 −2 mm mrad [37]. Based on the low transverse emittance and the small convergence angle, all electrons in a beam were considered to be parallel to the beamline throughout the paper.
Two types of Lanex Fast were studied: Lanex Fast Front (thinner) and Lanex Fast Back The upstream surface of the Lanex was covered by a ≃ 40 µm thick aluminum foil to minimize any contributions from scattered laser light which copropagates with the electron beam when exiting from LPA [27]. To mimic the LPA setup, all RFA based calibration experiments were also carried out with the aluminum foil in front of the LANEX.
For charge measurement using the ICT, the temporal waveform of the ICT signal was recorded for each shot, a typical example of which is shown in Fig. 3. For each shot, the background level was evaluated by averaging the signal between 0 ns (when the external trigger was applied for the scope acquisition) and the first vertical dotted line in Fig. 3. The evaluated background level is indicated by the dashed line. To obtain charge, the signal between the second and third vertical dotted line was integrated. The signal to noise ratio was defined by SN ict = V peak /V σBG , where V peak is the peak signal voltage and V σBG is the standard deviation of the background signal.
In order to estimate the number of photons emitted by the Lanex, the following optical properties of the system were measured. The transmission of the borosillicate window was measured to be (0.95 ± 0.02), and of the objectives was (0.88 ± 0.02) for the wavelength of 543 nm. Note that the emission peak of the Lanex is at 546 nm. The quantum efficiency of the camera at 550 nm was 0.46, and the analog to digital unit (ADU) for 2x2 binning mode was 9.3 (9.3 electrons for one digital count) [38]. Based on the distance between the Lanex and the camera, the f-number, and the focal length, the solid angle of the system was (3.6 ± 0.1) ×10 −4 sr. Assuming the Lambertian distribution for the directionality of the emitted light [26,32], the light collection efficiency of the system was (1.
The total efficiency of the system, which was defined by the ratio of the total counts on the CCD camera to the number of photon emitted from the Lanex to a hemisphere, was (4.8 ± For the image processing, several effects were taken into account: 1) darkening at the edges of acquired images due to the finite collection solid angle, 2) darkening/brightening due to the 45 degree orientation of the Lanex. The imaging system used the objectives instead of single lenses to realize a compact observation system [27], which made theoretical modeling of the finite solid angle effect difficult. Therefore, that effect was experimentally measured.
In order to evaluate finite solid angle effects, the transverse and longitudinal light location dependences were measured off-line by using a green light emission diode (LED). A piece of Lanex was placed in front of the LED to imitate scattering from the Lanex, and position of the LED was scanned transversely to measure the darkening effect. Shown in Fig. 4 (a) is the applied compensating factor for the image processing as a function of the number of pixels from the center pixel. The measurements were done with two different distances between the camera and the Lanex, confirming that the normalized edge darkening did not depend on the distance. One can see that about 10 % darkening was observed at the edge.
To compensate distance dependence due to the 45 degree orientation of the Lanex, the distance dependence was measured by longitudinally scanning the LED location. The compensating factor as a function of the Lanex -camera distance is shown in Fig. 4 (b), where the factor was normalized to the distance from the Lanex at the center of the view to the camera (213 mm). The measured distance dependence was also used for the relative calibration between this setup and others with different distances [27].
Typical processed images for 1289 MeV and 106 MeV electron energy are shown in Fig. 4 (c) and (d). As can be seen, the size of the electron beam on the Lanex depended on the energy of the electron. The compensation of the finite solid angle effect was essential for accurate measurements. The signal to noise ratio for each image was defined by SN img = C peak /C σBG , where C peak is the peak counts of the image and C σBG is the standard deviation of the background image.

B. Results
Measurements of e-beam charge and light yield from the Lanex Fast Back were conducted for 15 different electron energies. An appropriate ND filter, which was calibrated with a spectrophotometer, was chosen for each measurement, so that the highest charge density could be explored. For each measurement, more than 100 shots were acquired scanning the gun bias voltage to explore a wide range of the total charge. Shown in Fig. 5 (a) is a case for 889 MeV electron energy with 1 micro bunch. The horizontal axis is the charge measured by the ICT, and the vertical axis is the total counts on the CCD camera. Note that the large error seen for low charge (< 100 pC) points is not an intrinsic limitation of the Lanex measurement. It was due to the ND filter used and the dynamic range of the camera. The Lanex measurement can be set for much lower charge as shown in the Sec. III.
From all the scans, offsets were observed as indicated in Fig. 5 (a). Since these offsets were found to be systematic, and always negative for CCD camera counts, it was attributed to a threshold bias of the CCD camera electronics. This threshold would give a negative offset proportional to the beam size. The observed offsets as a function of the beam size are shown in Fig. 5 (b). The offsets and beam size were well-correlated, which was consistent with a threshold-induced offset. In the analysis, the bias based on the modeled offset was applied for the total counts on the CCD camera. Note that this problem can be solved by employing a camera with higher sensitivity.  on the CCD camera were biased based on the modeled offset shown in Fig. 5 (b), and only points with the signal to noise ratios for ICT SN ict and CCD camera image SN img above 10 were taken into account. The linear fitting was done for the combined dataset, and the slope is shown in the inset. To evaluate the quality of the fit, the fit errorǫ is shown in the lower right box and is defined as the mean of the normalized standard deviation, where g is the number of samples, CCD i and ICT i are the total counts on the CCD camera The electron energy dependence of the Lanex light yield was not observed in previous works [26,28,32]. Since the previously explored energy range was small enough for the difference to be within the measurement error, the observed energy dependence is not con-tradictory to previous works. Although the material used was different, it is noteworthy that the IP showed similar energy dependence [29,30], where the sensitivity of IP decreased ∼ 4% per 100 MeV increase of the energy.
There are a few possible scenarios that may explain the observed energy dependence.
Relativistic electrons ionize material, exciting a certain level followed by the radiative relaxation, and also produce Bremsstrahlung γ-rays, that can create knock-on electrons which also ionize the material. Since higher energy photons get absorbed less in the material [39], the contribution from Bremsstrahlung γ-rays may become less for higher energy electrons.
Another possible scenario is that the electron scattering angle is less for higher energy electrons, resulting in less interaction with the material. Since the observed energy dependence was fairly small, the effect from the different beam size on the screen cannot be ruled out for the cause of this. Although all the geometrical effects were taken into account such as the edge darkening and the distance dependence, there still could be small effects that were not modeled in the analysis. In this section, a study of three charge diagnostics using LPA produced e-beams is presented. The activation based measurement [23] is intrinsically immune to EMP noise, and electron charge intensity independent. Therefore, it can provide reliable reference for LPA produced e-beam charge measurements. The Lanex has been calibrated using RFAs [26,32], but has not been benchmarked with a reliable diagnostic for LPA produced e-beams. The use of an ICT has been questioned for LPA produced e-beams charge measurement, while it has been used as a reliable reference for RFAs. The experimental setup is shown in Sec. III A, the Lanex -activation cross-calibration in Sec. III B, and the Lanex -ICT cross-calibration in Sec. III C. It is shown that all the diagnostics can provide accurate charge measurements for LPA produced e-beam.

A. Experimental Setup
Cross-calibrations between the Lanex and activation based method, and between the Lanex and ICT were conducted by using LPA produced e-beams at the LOASIS facility, LBNL [40]. The laser that was utilized was a short pulse, high peak power and high repetition The schematic drawing of the setup is shown in Fig. 7 with a scale to indicate the distance from the interaction point. The laser pulse came from the left, and was focused onto a supersonic gas jet. A magnet for electron spectrometer was located about 0.7 m downstream of the interaction point. For the cross-calibrations, the magnet was turned off to send e-beams to charge diagnostics located at further downstream. The laser pulse was reflected by the aluminum coated mylar foil toward the laser beam dump, and only e-beams went through following vacuum tubes.
An aperture with a 36 mm inner diameter was installed about 3.6 m away from the interaction point. An aluminum coated 5 µm thick pellicle foil was located about 3.9 m for an optical transition radiation experiments (OTR foil). The identical ICT unit described in The light from Lanex was observed by an identical CCD camera described in the Sec. II A (not shown in Fig. 7) through the reflection of an aluminum coated 5 µm thick pellicle foil installed in the outside of the vacuum tube. The reflectivity of the foil was measured to be (0.97 ± 0.02). For the activation measurement, the target was placed behind the pellicle foil. The target consisted of 6 mm thick lead as a γ-ray generator followed by the 25 mm thick copper as an activation material, and is illustrated in Fig. 7 inset.

B. Lanex -Activation Cross-Calibration
The activation based charge diagnostic was used as follows. The target material was irradiated by e-beams for ≃ 1 hour. After the irradiation, the target was transferred to an ultra-low background counting facility at LBNL, where γ-ray spectroscopy was conducted.
Based on the γ-ray spectroscopy results, the activity in terms of the average production rate R exp (atoms/minute) for a specific isotope was calculated. In order to estimate the charge, a Monte-Carlo simulation was carried out to calculate the yield of an isotope for a unit charge with a certain e-beam energy spectrum Y sim (atoms/electron). The e-beam energy spectrum was separately measured during the experiments. The number of electron per minute irradiated to the targetN e was obtained fromN eAct = R exp /Y sim . As the result, The laser pulse was focused onto the hydrogen jet to produce high charge relativistic e-beams. The plasma density was measured by transverse interferometry. The peak plasma density was measured to be ∼ 1.9 × 10 19 /cm 3 , and plasma longitudinal length was ∼ 0.9 mm. The e-beam energy spectra were measured by a single shot magnetic electron spectrometer [27] before sending the e-beam to charge diagnostics, and reproducible broadband e-beams up to 250 MeV were observed. The reference e-beam spectrum, which is the average of 50 shots, is shown in Fig 8 (a). The solid line shows the average spectrum, and the grey area shows the ± standard deviation. The normalized standard deviation of the total charge was found to be 32 %.
The e-beams were sent to the activation target by turning off the magnet of the electron spectrometer. A total of 2700 shots was incident onto the target in 60 minutes. The aluminum coated 5 µm thick pellicle foil was placed between the Lanex and the activation target to separate the light from the Lanex and e-beams with minimal disturbance.
During the activation, the e-beam charge was simultaneously measured by the Lanex.
The average e-beam profile is shown in Fig. 8 (b). The e-beam charge was measured to be (36.6 ± 1.3) pC/shot, or the number of electrons in a minute measured by the Lanex N eLnx = (10.3 ± 0.4) × 10 9 electrons/minute. The error of the Lanex charge measurement was 3.4 %, obtained from the convolution of the average fit error described in Sec. II B and the error in the reflectivity of the foil. The horizontal beam size in FWHM was measured to be 11.5 mm, and the vertical size was 13.7 mm. The sharp edge on the top right was due to the retaining ring of the OTR foil (see Fig. 7 for location).
After the irradiation, the target was transferred to the counting facility, and γ-ray spectroscopy was conducted by using a p-type HPGe detector. The 1345 keV photons from 64 Cu decay (half life time 12.7 hour) was used to determine the yield of the isotope. Assuming the constant activity during the irradiation, the average production rate R exp was measured to be (9.79 ± 0.59) × 10 6 atoms/minute, where the error was given by the ±1σ (standard deviation).
A Monte-Carlo simulation was carried out to estimate isotope production based on the e-beam reference spectrum using MCNPX code [41] and NNDC database [42]. An axisymmetric, three-dimensional distribution of the isotope production was calculated. The incident angle of electrons were assumed to be 0 degree, and the transverse profile of the e-beam was assumed to be a Gaussian shape with 13 mm FWHM based on the Lanex measurements. In the simulation, energy loss of electrons due to the interactions with the foil, Lanex, vacuum window, and air were included. Based on the 3-D activation distribution, self attenuation of emitted γ-ray was taken into account. The simulated yield of the isotope Y sim was 1.02×10 −3 atoms/electron, givingN eAct = R exp /Y sim = (9.6 ± 0.6) × 10 9 electrons/minute. The charge measured by the activation based measurementN eAct = (9.6 ± 0.6) × 10 9 and the Lanex measurementN eLnx = (10.3±0.4)×10 9 agreed with each other within the error of each measurement. This was the first time that the charge diagnostic based on a phosphor screen was benchmarked against a reliable method for LPA produced e-beam, and indicated that the Lanex can be an accurate charge diagnostic for LPA produced e-beams.

C. Lanex -ICT Cross-Calibration
In order to produce relativistic e-beams, the laser pulses were focused onto the downstream edge of a gas jet comprised of 99% helium and 1% nitrogen. The gas jet backing pressure was varied to change charge yield from the LPA. The peak plasma density was measured by transverse interferometry, and was found to be from 4 to 10 ×10 18 /cm 3 , and the longitudinal plasma length was about 0.5 mm. Reproducible e-beams up to 60 MeV were observed, and the e-beam spectra as a function of the peak plasma density are shown in Fig. 9. Higher plasma density resulted in producing e-beams with the higher charge and the broader spectrum.
The Lanex -ICT cross-calibration was carried out in the same manner as the calibration with RFA-generated e-beams. Electron beams were sent to the various charge diagnostics by turning off the magnet for the spectrometer. A total of 320 shots was recorded while varying the plasma density, and e-beams with the charge up to 16 pC were observed. Shown in Fig. 10   As shown above, the ICT can measure the LPA produced e-beam charge accurately, while previous works showed that the ICT overestimated the charge up to two orders of magnitude. The excellent agreement we observed can be explained by the special attentions paid to the following three possible noise sources of the ICT measurement: (1) EMP from the laser plasma interaction, (2) direct particle/radiation hit on the ICT, and (3) low energy electrons.
There were two kinds of EMP noises observed, one was directly on the scope, and the other was on the cable and/or the ICT. The noise on the scope was separated from the signal in the time domain by extending the cable length. To minimize the noise on the cable, well-shielded cables (Heliax FSJ1-50A, CommScope, Hickory, NC, United States) were employed, and the route of the cables was carefully arranged to reduce the noise. Since the higher frequency part of the EMP noise was visible from the waveform, it was used as an indicator while optimizing the route. As can be seen from Fig. 10 (b), the obtained ICT signals did not contain high frequency spikes.
The direct hit of the laser pulse or e-beam onto the ICT can potentially create secondary electrons and/or ionize the material, possibly contributing to the noise. The laser pulse was separated from e-beams to prevent it from hitting the ICT. An aperture was utilized for e-beam transverse size to be smaller than the acceptances of the ICT and Lanex, assuring that the e-beams did not hit the ICT or vacuum tube. The ICT was installed outside of the vacuum tube over a ceramic gap so that e-beams propagate in vacuum with minimum disturbance.
The low energy electrons could cause a large discrepancy between the ICT and Lanex measurements because of the following reasons. (1) Non-linear beam size evolution due to the space charge effect can lead to the acceptance mis-match between the two diagnostics.
(2) The response of the Lanex against low energy electron (< 1 MeV) is not clear. In this experiment, the ICT was installed 4.2 m away from the interaction point to assure that the low energy electrons diverge enough to minimize their contribution. Furthermore, the small residual magnetic field (< 0.4 mT) of the magnetic spectrometer, and absorptions/scatterings at the foils used for the laser beam separation may have contributed to eliminate low energy electrons. The distance between ICT and Lanex was kept at a minimum to avoid acceptance mismatch.
Although no quantitative evaluation was performed for each noise source, it was considered to be critical to provide a low-noise environment for the charge measurement. Note that the accuracy of the measurement can be improved by a more sensitive camera for Lanex measurements, and by more sensitive electronics for ICT measurements.

IV. SUMMARY AND CONCLUSIONS
The electron energy dependence of the Lanex Fast light yield was studied from 106 MeV to 1522 MeV e-beams. The Lanex was observed to be 1% less sensitive for every 100 MeV increase in the energy. The charge density and the charge intensity on the Lanex was explored up to 160 pC/mm 2 in ∼ 8 ns, and 0.4 pC/ps/mm 2 , respectively. The linear response of the Lanex Fast Back was verified up to those parameters.
A comprehensive study of the charge diagnostics for LPA produced e-beams was conducted. The cross-calibration between the Lanex and the activation-based charge diagnostic showed good agreement within the error of each diagnostic. This was the first benchmark of the Lanex against a reliable method for LPA produced e-beam charge measurement. The cross-calibration between the Lanex and the ICT showed good agreement as well. The result of the cross-calibrations can be summarized as follows, Q Act = 93 98 87 , Q Lnx = 100 103 97 , Q ICT = 108 116 100 , where they are normalized to the Lanex result, and super and sub scripts show result error of the Lanex -ICT cross-calibration using LPA produced e-beams. This study showed that all diagnostics can provide accurate charge measurement for LPA produced e-beams.
The guideline for accurate measurements with the ICT under the harsh LPA environment was discussed. This study will provide essential information for charge measurements of LPA produced e-beams.