1992 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012 High-Energy Bremsstrahlung Diagnostics to Characterize Hot-Electron Production in Short-Pulse Laser-Plasma Experiments Anthony L. Meadowcroft and Ray D. Edwards Abstract—Short-pulse (< 10 ps) high-intensity (1019 − 1021 W · cm−2 ) laser-target interactions produce high-density plasmas, in which a hot (nonthermal) collisionless electron population is generated via a number of energy absorption mechanisms. A key requirement in many applications of these interactions is to achieve maximum conversion efficiency of laser energy into hot electrons of a specific energy. Measurement of the hot-electron temperature and the associated hot-electron fraction supports the understanding of energy absorption mechanisms present in the 1019 −1021 W · cm−2 intensity regime. One highly useful signature of hot-electron generation is the bremsstrahlung radiation emission due to the hot electrons interacting electromagnetically with cold target atoms. We present the design, characterization, and modeling of two target diagnostics to measure the high-energy (100 keV–2 MeV) bremsstrahlung emission from hot electrons in laser-plasma experiments; a thermoluminescent dosimeter (TLD) array and a hard X-ray spectrometer (HXRS). These diagnostics will exploit new techniques to determine the hot-electron distribution generated with the short-pulse beamlines of the upcoming high-power Orion laser system at AWE. Past bremsstrahlung dose measurements obtained with a TLD array (of similar design to the Orion one) are used to demonstrate how the bremsstrahlung production efficiency, an indicator of hot-electron generation, can change with laser parameters, target type, and experimental geometry. In addition, we use results from characterization of the HXRS, a diagnostic which collects channel charges in defined bremsstrahlung spectral ranges, to establish a new method enabling the diagnosis of hot-electron temperatures. TABLE I D IAGNOSTIC T ECHNIQUES TO M EASURE THE H OT-E LECTRON E NERGY D ISTRIBUTION Index Terms—Bremsstrahlung dose, high-energy bremsstrahlung emission, laser-plasma experiments, plasma diagnostics, short-pulse laser–target interactions, X-ray detectors. I. I NTRODUCTION S HORT-PULSE (< 10 ps) laser-target interactions at intensities of 1019 −1021 W · cm−2 produce a hot (nonthermal) collisionless electron population via a number of energy absorption mechanisms, primarily by Brunel absorption (or vacuum heating) at low to moderate intensities and by direct ponderomotive acceleration (or j × B heating) at the highest intensities. These mechanisms, like resonance absorption, are collisionless in nature. Unlike resonance absorption, however, Manuscript received July 21, 2011; revised February 27, 2012; accepted May 4, 2012. Date of publication June 21, 2012; date of current version August 7, 2012. The authors are with the Department of Plasma Physics, AWE Aldermaston, RG7 4PR Berkshire, U.K. (e-mail: [email protected]; [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TPS.2012.2201175 they are applicable to very short scale-length plasmas of the kind common in short-pulse laser-matter interactions. Accurate measurement of the hot-electron energy distribution supports the understanding of various energy absorption mechanisms present in the 1019 −1021 W · cm−2 intensity regime. Past measurements of the hot-electron energy distribution have been obtained by a number of methods, including electron spectrometers [1]–[3], Kα X-ray emission [4], [5], bremsstrahlung radiation emission [6]–[9], and fast ion [10] measurements. The advantages and disadvantages of using these diagnostic techniques in the measurement of hot electrons are summarized in Table I [11]–[14]. For short-pulse laser-target interactions at intensities of 1019 −1021 W · cm−2 , electron energies may be on the order of several megaelectronvolts, with associated hot-electron temperatures of > 100 keV. These hot electrons interact with the cold background material of the target to produce energetic bremsstrahlung emission of > 100 keV. Unlike measurements 0093-3813/$31.00 © 2012 IEEE MEADOWCROFT AND EDWARDS: DIAGNOSTICS TO CHARACTERIZE HOT-ELECTRON PRODUCTION 1993 made using electrons, fast ions, and Kα X-rays, the > 100-keV bremsstrahlung emission clearly reveals the presence of two or more hot-electron temperature components [15]. In order to measure the hot-electron energy distribution, information on both the hot-electron temperature and the hot-electron fraction (which may be determined from the hot-electron temperature [16]) is required. The short-pulse beamlines of the upcoming high-power Orion laser system at AWE will last ≈1 ps, and they will focus onto targets at 1019 −1021 W · cm−2 intensities. Two target diagnostics will be used to measure the resulting hot-electron energy distribution by recording bremsstrahlung emission in the range of 100 keV–2 MeV: a thermoluminescent (TL) dosimeter (TLD) array and a hard X-ray spectrometer (HXRS). In order to diagnose the hot-electron production efficiency in a laser–target interaction, the bremsstrahlung dose is modeled and compared to the dose recorded by the TLD array. Hotelectron temperatures will be determined from measurements of the bremsstrahlung emission in defined spectral ranges with the HXRS diagnostic, using the difference in collected channel charges in a technique similar to that utilized by a Ross pair spectrometer. In this paper, we present the design, characterization, and modeling of the TLD array and the HXRS diagnostics to acquire hot-electron distribution data from impending shortpulse laser-plasma experiments on the Orion laser. II. TLD A RRAY D IAGNOSTIC A. TLD Array Design The Orion TLD array measures bremsstrahlung radiation dose levels generated in the electromagnetic pulse environment of a laser-plasma interaction. An overview of the diagnostic is shown in Fig. 1. The Orion TLD array consists of nine TLD collimators. The purpose of the collimator is to restrict the lineof-sight to the target chamber center (TCC). Each collimator is fielded at the same radial distance from the TCC but at different angles with respect to the laser axis to measure the anisotropy of bremsstrahlung radiation emission (see Section II-B). Ten-inch manipulators (TIMs) provide a practical means of deploying sets of collimators over a total angular range of ±60◦ with respect to the laser axis, within the Orion target chamber and at equidistant location from the target. The collimators each encloses a series of pots containing TL detector powder and metal filters or “slugs.” A combination of TL lithium fluoride (LiF) phosphor and various thicknesses of tungsten slugs enables bremsstrahlung radiation dose measurements over an energy range of ∼0.1–20 MeV. B. Modeling Bremsstrahlung Emission From Short-Pulse Laser–Target Interactions In order to be able to predict the dose expected at a given location from a particular target configuration, one must first be able to model the bremsstrahlung emission arising from laser-target interactions. Data for experiments on the PETAL laser at the Atomic Energy and Alternative Energies Commission (CEA), France, have been applied to simulations of Fig. 1. Overview of the Orion TLD array: (a) Breakdown of structure, (b) TIM, (c) deployment of TLD collimators on a TIM, (d) TLD collimator, and (e) TLD collimator components. bremsstrahlung emission on Monte Carlo N-Particle Code [17] in conjunction with two different codes, CALDER at the CEA [18] and PSC at the AWE [19], which are used to predict electron generation for a laser beam incident on a gas jet target (detailed as follows). The codes have been benchmarked against published results, which have been obtained using accelerators with monoenergetic electron beams [20]. The experiments considered use two different target configurations (see Fig. 2): a gas jet with a separate bremsstrahlung 1994 Fig. 2. Solid target and gas jet–solid target configurations for short-pulse laser–target experiments. converter and a solid metal plate. Electrons are accelerated to produce bremsstrahlung radiation by different mechanisms in each target configuration. In the solid target configuration, ponderomotive acceleration occurs, whereas in the gas jet–solid target configuration, self-modulated laser wakefield acceleration occurs. A gas jet-solid target produces more efficient bremsstrahlung emission for two reasons. First, self-modulated laser wakefield acceleration generates a self-collimated and almost monodirectional (< 5◦ off-axis) beam of electrons incident on the target; in ponderomotive acceleration, the electrons are incident on the target over a broad range of angles. Second, a laser electromagnetic wave does not propagate very far into the solid target material since the refractive index of the plasma becomes too high; in laser wakefield acceleration, high-amplitude plasma waves generated by the interaction of laser light with the plasma medium accelerate the electron beam from the gas jet. At energies beyond ∼1 MeV, the bremsstrahlung emission becomes rather anisotropic due to Compton scattering, and the angular distribution of scattered radiation is described by the Klein–Nishina formula [21]. This leads naturally to anisotropy in the recorded dose. As the gas jet-solid targets produce more bremsstrahlung emission, they allow the dose anisotropy to be more accurately measured, and so, simulations of the emitted bremsstrahlung spectrum, including anisotropy contributions from Compton scattering, were obtained for those experiments. The simulation used for the dose model (see Section II-C) included the maximum possible output of the PETAL as input parameters: 3.6 kJ in 0.5 ps, a focal spot diameter of ≈ 20 μm, a laser wavelength of 1.053 μm, and an intensity of ≈ 2 × 1021 W · cm−2 , incident on a helium gas jet–tungsten target of 4-mm thickness. The simulated PETAL bremsstrahlung spectrum is shown in Fig. 3. Note that the maximum bremsstrahlung yield is shown in Fig. 3 (detection along the target normal direction only, which is aligned with the laser axis). Also, energies are shown up to 400 MeV, rather than the 20-MeV upper energy limit considered for the diagnostics in this paper. However, most of the bremsstrahlung flux is emitted around energies of 1–10 MeV (at 10 MeV, the bremsstrahlung yield is already ∼10% of that at 1 MeV). IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012 Fig. 3. Simulated bremsstrahlung spectrum produced by a gas jet-solid target for the PETAL, generated with codes that have been validated against experimental data. The laser irradiance was ≈2 × 1021 W · cm−2 , incident on a gas jet-tungsten plate target of 4-mm thickness. Fig. 4. Schematic of the experimental configuration, with the relevant physical quantities identified. that are important at each location. This information can be used to help derive a simple model for the absorbed dose. The calculations of how the bremsstrahlung dose D scales with laser and target parameters assume that D ∝ Iλ2 [9], [22] (for laser intensity I and laser wavelength λ), D ∝ Z [21], [23] (for target atomic number Z), and D is dependent on the type (gas or solid) of target. Dose has been reported to change with solid target thickness for short-pulse experiments when using target thicknesses in the range of 0.1–2.0 mm [23], but these measurements were not filtered from high-energy electrons emitted from the target. High-energy electrons contribute significantly to the dose measurement. Further experimental data [24], [25], with filtering of high-energy electrons, have been collected for target thicknesses within the range of 0.5–5.0 mm considered in this paper. These results indicate some variation of dose with target thickness but with no clear correlation. As the laser-target interaction yields a spectrum of hot electrons and the exact relationship between dose and target thickness is unknown, the model initially assumes that D is independent of target thickness. Comparison of the model calculation with experimental data may then be used to determine how the target thickness affects the bremsstrahlung production efficiency and, thus, D. Since bremsstrahlung dose is total bremsstrahlung energy absorbed per unit mass of absorbing material, it follows that C. Bremsstrahlung Dose Model Fig. 4 shows the geometry of the general type of experiment being considered, as well as identifying the physical quantities 1 D(φ, θ, E) = m ∞ NX (φ, θ, E) dE 0 (1) MEADOWCROFT AND EDWARDS: DIAGNOSTICS TO CHARACTERIZE HOT-ELECTRON PRODUCTION where D m NX E φ θ measured in grays; mass of LiF phosphor in kilograms; number of bremsstrahlung photons incident on the LiF phosphor; bremsstrahlung photon energy in joules; angle between the laser axis and the target normal; angle between the emitted bremsstrahlung photon and the target normal. In order to estimate NX , we use the simulated gas jet-solid target bremsstrahlung spectrum for the PETAL (see Fig. 3). Additionally, we implement the solid angle of πa2 /b2 for each TLD pot (where a is the radius of the collimator channel and b is distance of the pot from the target) and the filtering or absorption of bremsstrahlung by tungsten slugs and LiF phosphor (E). Including normalization of giving a spectral response SR the bremsstrahlung spectrum, the number of bremsstrahlung photons incident on the LiF phosphor for a solid target using the laser and target scaling described earlier, with a point source, is given by NX (φ, θ, E) = 1.0 × 10−23 Zs(φ, Iλ2 ) × Iλ2 SR (E) a 2 b ψ(φ, θ, E) (2) where s(φ, Iλ2 ) I λ Z ψ(φ, θ, E) dose scaling for a gas jet–solid target to a solid target; laser intensity in watts per square centimeter; laser wavelength in micrometers; target atomic number; bremsstrahlung yield per solid angle in photons per steradian. An important consideration when measuring the dose from a short-pulse laser-target interaction is the laser prepulse. Prepulses may form preplasmas on the surface of a solid target prior to arrival of the main pulse. The preplasma has a significant effect on laser-solid target interaction because the refractive index of the plasma inhibits propagation of the laser electromagnetic wave (see Section II-B). In comparison, a gas jet-solid target interaction results in a larger proportion of the incident laser energy producing a self-collimated monodirectional electron beam incident on the target. Therefore, a larger amount of preplasma may be produced for the solid target configuration in comparison to the gas jet-solid target configuration at increased laser intensity. Increasing the laser intensity, however, may also produce either a larger hotelectron fraction or higher hot-electron temperatures in either target configuration, which results in increased bremsstrahlung emission. The resulting change in dose may be explained by a change in bremsstrahlung production efficiency, and this is implemented in the model using the dose scaling s(φ, Iλ2 ). Therefore, the dose scaling s(φ, Iλ2 ) can be used to infer the hot-electron fraction for a given laser-target interaction. Using (1) and (2), the calculation giving the bremsstrahlung dose 1995 for collimator pot i [see Fig. 1(e)] at 1 m in milliradians is given by Di (φ, θ, E) = 1 × 10−18 Zs(φ, Iλ2 ) m ∞ × Iλ a 2 2 ψ(φ, θ, E)SR (E)dE (3) 0 where m is in milligrams and a is in millimeters. As the angular dependence of the dose scaling depends on the mechanism for bremsstrahlung production in the target, it is assumed that the scaling takes the form s(φ, Iλ2 ) = f (φ)f (Iλ2 ), where f is a function. In order to minimize the preplasma contribution to the bremsstrahlung production efficiency, s(φ, Iλ2 ) was initially evaluated from Vulcan 100-TW laser experimental data [26] (with laser intensities up to ≈ 7 × 1019 W · cm−2 · μm2 ) rather than from PETAL experimental data (with laser intensities up to ≈ 2 × 1021 W · cm−2 · μm2 ). An expression for f (φ) was determined from filtered TLD measurements using the gas jet-solid target and solid target experiments described in [26]. The value of s(φ, Iλ2 ) is then normalized by setting f (Iλ2 ) = 1 for the Vulcan 100-TW experiments. For other short-pulse laser-target experiments, where prepulse characteristics are likely to differ (e.g., longer pulse duration), the dose scaling is adjusted by setting f (Iλ2 ) such that the dose calculation has a best fit to experimental results. It should be noted that, since prepulse effects and bremsstrahlung production efficiency for a solid target are inextricably linked, prepulse contributions are subsumed in the evaluation of f (Iλ2 ). D. Dose Model Validation In order that (3) could be evaluated, a new code, bremsstrahlung radiation dose calculator (BRADDOC), was written in IGOR Pro [27]. BRADDOC consists of an algorithm in which the integral in (3) is solved analytically and other parameters are changed until a solution is found. Besides handling the complexity of the calculation, IGOR has a number of useful features that can be implemented easily, such as a simple user interface for data entry and altering code parameters, real-time monitoring of the calculation, and clear graphical output. BRADDOC results were initially generated using actual Vulcan 100-TW laser experimental parameters given in Table II. A best fit was subsequently obtained by allowing for uncertainty in the focal spot size (laser spot for the solid target, a focused electron beam for the gas jet–solid target), which primarily affects the dose magnitude, as well as the angle of the laser axis with respect to the target normal, which primarily affects the shape of the dose distribution (see Fig. 5). Subsequent comparison of Vulcan PW experimental data with BRADDOC results (see Fig. 6) confirmed that the dose scaling takes the form s(φ, Iλ2 ) = f (φ)f (Iλ2 ). As mentioned in Section II-C, this can be explained by the increase in preplasma formation, the hot-electron fraction, or hot-electron temperature at increasing laser intensity. Also, using a target thickness of less than 1996 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012 TABLE II BRADDOC I NPUT PARAMETERS FOR F IG . 5 Fig. 6. Experimental data in markers and BRADDOC results in solid/dashed lines. TLD pot 2: (Red) Best fit BRADDOC results to Vulcan 100-TW data with initial gas jet-to-solid target scaling s(φ, Iλ2 ), BRADDOC results with no scaling applied to f (Iλ2 ) for (solid green, pink, and gray lines) 1-mm · W targets with target normal along the laser axis and for (solid blue line) a 4-mm · W target with target normal at 15◦ to the laser axis, and (dashed lines) best fit BRADDOC results to Vulcan PW data with scaling of f (Iλ2 ) as given in Table III. TABLE III S UMMARY OF THE S CALINGS A PPLIED TO f (Iλ2 ) TO P RODUCE B EST F IT BRADDOC R ESULTS TO V ULCAN PW DATA Fig. 5. (Markers) Bremsstrahlung dose results for the Vulcan 100-TW laser [5] and (solid line) BRADDOC results for (a) solid tantalum target 1.75 mm thick and (b) helium gas jet-and-solid gold target. Note that actual experimental errors are not shown; they have been assigned to show dose values with 10% uncertainty to check code reproducibility (given uncertainty in prepulse contribution) and experimental alignment to within ≈ ±2.5◦ . ≈1.5 mm leads to an increase in the bremsstrahlung production efficiency (see Table III). This indicates that, at a given laser intensity, there is an optimum target thickness required to max- imize the bremsstrahlung production efficiency. Increasing the target thickness stops more hot electrons, and this yields more bremsstrahlung. However, bremsstrahlung self-absorption also rises with increased target thickness. If the target thickness is reduced too much, not all hot electrons will be stopped, and the bremsstrahlung yield is lower. As a closing note, there will be some ambiguity in using the scaling of f (Iλ2 ) as an indicator of the hot-electron fraction. In best fitting BRADDOC results to TLD experimental data, uncertainties will primarily result from experimental data (∼10%), prepulse contribution (∼20%), and focal spot size (∼10%). If the root-mean-square value of these uncertainties is taken, then there is a typical uncertainty of ∼25% in deducing the hot-electron fraction. Although prepulse cannot be fully mitigated, knowledge of the prepulse conditions could be used to further reduce the uncertainty of experimental data serving as BRADDOC input. MEADOWCROFT AND EDWARDS: DIAGNOSTICS TO CHARACTERIZE HOT-ELECTRON PRODUCTION Fig. 7. 1997 Functional breakdown of the HXRS. III. HXRS D IAGNOSTIC A. HXRS Design The HXRS consists of an array of filtered detectors that provides a time-integrated measurement of the hard X-ray spectrum over ∼100 keV–2 MeV generated by laser-plasma interactions. The HXRS has eight channels, each comprising a filter, scintillator, and a photomultiplier tube (PMT) detector. This configuration has been adopted to meet the required spectral coverage, satisfy the engineering constraints, and maximize the dynamic range of the diagnostic. The choice and thickness of the filter and the scintillator, as well as the intensity of the hard X-ray spectrum, determine the channel spectral response. An outline of the operation of the HXRS is given in Fig. 7. In order to prevent any significant background signals arising from X-ray or particle interactions, each channel has a line-of-sight to the target. These lines-of-sight are collimated and shielded and also avoid the main directions of propagation of electrons and protons emitted by the laser–target interaction. Sufficient temporal resolution provides discrimination against background signals from electrons, protons, and neutrons generated in the target chamber walls of Orion. Since photon radiation generated in the Orion target chamber wall has a cross-chamber transit time of ≈14 ns, it is desirable to have an overall temporal resolution of < 14 ns. Background signals resulting from X-ray scattering interactions are minimized by using a high-Z material (lead) to collimate the HXRS channels. Bremsstrahlung background is minimized by using a set of sweeper magnets at the front of the HXRS, to deflect stray electrons from the target into a plastic dump absorber. A crosssectional view of the HXRS design is shown in Fig. 8. In addition to having collimated lead channels, the HXRS channels use a column of air to separate the filter and scintillator pair and to minimize X-ray scattering and fluorescence contributions detected by the scintillators. The length of the air column (30 cm) is determined from calculations and simulations of the HXRS response. Aluminum disks included in each channel isolate the HXRS from the target chamber vacuum and filter out soft X-ray radiation produced by secondary interactions. In order to minimize the impact of electron scattering on the response signal, these disks are located after the sweeper magnet section. Electromagnetic interference originating from laser–target interactions is a problem, as it can overwhelm the detector signal. To mitigate this effect, the HXRS is designed as a fully metallic enclosure, with collimators serving as waveguides beyond cutoff. Fig. 8. Cross-sectional view of the HXRS design. B. Characterization of the HXRS The HXRS was characterized using the miniature linear accelerator (MINAC), a bespoke radio frequency linear accelerator, at AWE. The MINAC uses microwave energy to create electric fields, which are then used to accelerate electrons into a tungsten target, producing bremsstrahlung X-rays. The MINAC is capable of producing variable energy X-rays, with bremsstrahlung end-point energies in 2-MeV steps from 2 to 10 MeV. Variable dose rates can also be achieved by varying the pulse repetition frequency (PRF) and the pulsewidth, where the PRF can be set at 40–200 pulses per second and the pulsewidth can be set at 100 ns–3 μs. In order for all eight channels to be fully exposed to X-rays, the front of the HXRS was positioned 3 m away from the source. The source emission is not isotropic but in the form of a beam with a maximum of 5◦ divergence. Cables from the rear of the HXRS were routed into a control room containing high-voltage power supply units (for the PMT detectors) and oscilloscopes to record the data. The test setup for the HXRS is shown in Fig. 9. Throughout the tests, the PRF was set to either 40 or 200 pulses per second, with bremsstrahlung end-point energies of 2, 4, or 6 MeV used (so the majority of X-rays were in the HXRS energy range of ∼100 keV–2 MeV), and the dose rate ranged from ∼6 to 250 rad/min. The low-dose rate was necessary to minimize additional background from the MINAC, specifically scattered X-rays and high-energy electrons. Examples of oscilloscope traces recorded during these tests are shown in Fig. 10. These traces demonstrate increase in the signal level at an applied PMT bias of 0.9 kV, as the end-point energy is increased from 2 to 6 MeV for channels 5–8 with LuAP scintillators (see Table IV). This is explained by the increased efficiency of bremsstrahlung emission from the MINAC as the end-point energy rises. Furthermore, the individual channel signals are more clearly distinguished at higher end-point energies. The effect of increasing the MINAC bremsstrahlung endpoint energy was only investigated for the channels containing LuAP scintillators as the channels containing Bicron BC-422 scintillators do not have sufficient stopping power for X-rays at the upper end of the ∼100-keV–2-MeV energy range for the HXRS. Therefore, channels 1–4 containing Bicron BC-422 scintillators were only tested at the 2-MeV end-point energy of 1998 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012 TABLE IV O PTIMIZED C HANNEL D ESIGN AND PMT G AIN S ETTINGS U SED IN HXRS R ESPONSE S IGNAL C ALCULATIONS Fig. 9. Photograph showing the HXRS in position, some 3 m from the MINAC. The shielding and collimation requirements for the HXRS give the HXRS a high mass (∼400 kg). Cranes and a compressor table were required to move the diagnostic into position. signals due to PMT light collection efficiency to be deduced and implemented into an unfold of hot-electron temperatures. C. Unfolding Hot-Electron Temperatures With the HXRS Fig. 10. Oscilloscope traces recorded for the HXRS using the MINAC bremsstrahlung X-ray source. Channels 5–8 (LuAP scintillators, Table IV) were used, with an applied bias of 0.9 kV and bremsstrahlung end-point energies of 2, 4, and 6 MeV. No attenuators were used with the oscilloscope, and the oscilloscope input impedance was set to 1 MΩ. the MINAC. As expected, the output signal level increased as the applied PMT bias (and, thus, gain) was increased from 1.4 to 1.6 kV. A separate calibration of each PMT and scintillator was also conducted using the MINAC X-ray bremsstrahlung source, including the installation cabling that allowed the attenuation due to the cabling to be investigated. For these new tests, the PMT and scintillator were housed in a lead tube configuration similar to the channel geometry used in the HXRS. Testing of the PMT and scintillator combinations was only effective at 2 MeV, due to the absence of the HXRS channel filters. All channel scintillators were coupled to the same PMT detector, and only the scintillators were changed. Data collected from this calibration enabled relative changes in the channel Characterization of the HXRS has shown that, in experiment, the HXRS will require different values of PMT gain to be used rather than the values shown in Table IV. Also, the channel signal level is not a suitable measurement, as the difference in signal level between a pair of the more sensitive channels (containing Bicron BC-422) can be comparable to the noise level. Therefore, we use collected channel charge rather than the output current in order to unfold hot-electron temperatures using the HXRS. A statistical analysis of errors and assumed error values, which may vary significantly between laser shots, has been used before in this approach [6]. Here, we present a new method which entails establishing expressions for the channel charge, comparing this to the integrated area for an oscilloscope pulse and rearranging the result to obtain hotelectron temperature data. First, we consider calculation of the HXRS channel charges due to the bremsstrahlung emission at given hot-electron temperatures. Assuming that a temporal resolution of ≥ 17 ns is attained during the laser shot (corresponding to the maximum scintillator decay time) and the X-ray flux is isotropic over the front face of the HXRS, the charge collected in a channel may be expressed as Qj (E, λv ) = Ij (E, λv )τj = eφSj (E)Aj ηj (E)Lj Eχj s(λv )j εj Gj z (4) where e φS E A s(λv ) ε = −1.6 × 10−19 C; X-ray flux incident on the scintillators; X-ray energy; active area of the PMT; relative sensitivity of the PMT photocathode at the scintillator maximum wavelength of emission λv ; quantum efficiency of the PMT photocathode of ≈0.20; MEADOWCROFT AND EDWARDS: DIAGNOSTICS TO CHARACTERIZE HOT-ELECTRON PRODUCTION 1999 the distribution of bremsstrahlung photons in the energy range of E to E + dE is given by E dE (5) f (E)dE = κ exp − kTH where κ kTH Fig. 11. X-ray spectral bands for HXRS using the difference in collected channel charges over the ∼100-keV–2-MeV energy range. G τ z η, χ, and L calculated gain of the PMT; electron pulsewidth, equal to the sum of the scintillator decay time (∼2–50 ns) τd and the electron transit time spread (≈0.1 ns) in the PMT tube ≈ τd ; signal attenuation due to cabling between the PMT output and the oscilloscope input; scintillator absorption efficiency, absolute light yield, and light collection efficiency, respectively. The variables L, A, and ε have to be taken as being specific to each channel, rather than as constants for the instrument as, for instance, the input face of the thickest LuAP scintillator has a smaller area than the other scintillators. Also, each of these variables is not absolutely known, and therefore, it is assumed that L = 0.5, ε = 0.2, and z = 1 to enable evaluation of (4), while the net relative differences between the channels due to all the variables are implemented by the adjustment of each previously calculated gain G (in Table IV) to an “effective” value GC . The X-ray spectral bands for the HXRS (Fig. 11) are determined by calculating (4) for each channel and taking the difference between the results for two channels. The two channels are selected to optimize the 100 keV–2 MeV spectral coverage, with the peak charges appearing at sufficiently well-separated X-ray energies. Since the HXRS only detects the cumulative X-ray yield and the channels are not energy dispersive, Fig. 11 may be regarded as the typical instrumental energy resolution. On this basis, the minimum resolution would be ∼100 keV within the 100-keV–2-MeV range, with a typical instrumental resolving power of ∼1. It is noted that φS uses a bremsstrahlung spectrum from a specific short-pulse laser-target interaction. For simplicity, it will be assumed that the distribution of hot electrons producing bremsstrahlung X-rays from a short-pulse laser-target interaction is a Maxwellian. Non-Maxwellian functions could also be used in the unfold method, although it may not be possible to extract an analytical expression to evaluate hotelectron temperatures. In the case of a Maxwellian distribution, relative spectral intensity; hot-electron temperature. Total channel charges are then determined as a function of hotelectron temperature by including this distribution function in (4) and integrating the result over all energies. Using the channel differences shown in Fig. 11, combined with the effective gain values, gives positive spectral bands over the entire 0–2-MeV range. Therefore, the assumptions that have got us to this point seem valid, suggesting that we can use the charge differences between channels to obtain hotelectron temperatures. In order to use this approach to obtain kTH , consider two channels j and k. In Section II, it was noted that, at energies > 1 MeV, bremsstrahlung emission becomes anisotropic. Such anisotropy is evident at 2-MeV energies for the TLD array which has ≈ 15◦ intervals in the line-of-sight to the target, but not for the HXRS channels where the lineof-sight ranges from ≈ 1◦ to 4◦ . Therefore, it is reasonable to assume isotropic X-ray flux and the same input X-ray spectrum from the laser-target interaction for two HXRS channels. As a result, the hot-electron source is also assumed to be the same for both channels. From Fig. 11, the integrand range may be defined as 0–2 MeV, and from (5), the difference in charge collected by the channels is given by ΔQT (E, λv ) = Qj,T (E, λv ) − Qk,T (E, λv ) 2 MeV E E exp − kTH = κez 0 × φSj (E)ηj (E)χj s(λv )j Lj GCj Aj εj − φSk (E)ηk (E)χk s(λv )k Lk GCk Ak εk dE. (6) Now, we consider the measured channel charge. Using Ohm’s law for a channel j, the charge collected from an oscilloscope pulse is given by Qj,exp = 1 R Vj (t)dt (7) pulsej where Qj,exp R Vj (t) charge collected from the oscilloscope pulse; oscilloscope input impedance; pulse amplitude (voltage) at time t. In order to account for differences between the PMT gain used in the experiment GE and the effective PMT gain used in the 2000 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012 calculation GC to define the X-ray spectral bands, we introduce a gain compensation factor δ = GC /GE in (7): Qj,exp = δj R Vj (t) dt. (8) pulsej Setting the corresponding difference in measured charge, ΔQT,exp = ΔQT , and assuming that the two channels use the same oscilloscope input impedance yields the following result: ⎛ 2 MeV ⎞ E κez ⎝ Fjk (E) exp − dE ⎠ kTH 0 ⎛ 1 ⎜ = δj Vj (t) dt − ⎝ R pulsej where Tj (E) Φ(E) ⎞ ⎟ δk Vk (t) dt⎠ (9) pulsek filter transmission for channel j; X-ray flux incident on the front end of each HXRS channel Fjk (E) = EΦ(E) Tj (E)ηj (E)χj s(λv )j Lj GCj Aj εj − Tk (E)ηk (E)χk s(λv )k Lk GCk Ak εk . The solution to (9) requires numerical integration, with initial estimates for the unknowns κ and kTH provided as input parameters in a computer program. However, the absolute values of Φ(E), L, z, and ε are not known as they have not been measured, nor is it possible to measure all of them. Relative channel signals due to changes in L and ε may be determined from tests using just the PMTs and scintillators, as described in Section III-B. Consider two different channels for the HXRS, l and m, where the charge difference is taken. An equation similar to (9) may be written. Taking the ratio of (9) with this result then enables evaluation of kTH from 2 MeV 0 2 MeV 0 Fjk (E) exp − kTEH dE Flm (E) exp − kTEH dE = δj Vj (t) dt − pulsej pulsel formation on the bremsstrahlung spectrum from the short-pulse laser-target interaction. δl Vl (t) dt − δk Vk (t) dt pulsek δm Vm (t) dt (10) pulsem in which Fjk (E) = EΦ(E)[Tj (E)ηj (E)Γj −Tk (E)ηk (E)Γk ], where Γ corresponds to the peak signals measured from the PMT and scintillator tests and it is assumed that Γ ∝ χs(λυ )LGC Aε. Since a ratio is given in (10), Γ does not have to be in absolute units. As a result, (10) only has two unknowns to solve for, which are the hot-electron temperature kTH and the incident flux on the HXRS channels Φ(E), providing in- IV. C ONCLUSION The design, characterization, and use of two high-energy bremsstrahlung target diagnostics, a TLD array and the HXRS, to measure hot-electron production from Orion laser-plasma experiments, have been presented. Modeling the TLD array doses as a function of laser, target, and geometric parameters provides information on the bremsstrahlung production efficiency and fraction of hot electrons generated in a laser-target interaction. At Iλ2 exceeding ≈ 8 × 1019 W · cm−2 · μm2 , bremsstrahlung production is optimized for high-Z targets less than ≈1.5 mm thick. Using targets less than ≈1.5 mm thick gives an increase of bremsstrahlung production efficiency with increasing Iλ2 , as reducing the target thickness reduces bremsstrahlung selfabsorption within the target when the hot electrons are stopped. Increasing Iλ2 produces either a larger hot-electron fraction or higher hot-electron temperatures. The HXRS has been characterized with the MINAC high-energy bremsstrahlung X-ray source at AWE, prior to installation on the Orion target chamber. The testing produced results that invalidated previous assumptions as to how the data could be unfolded to determine hot-electron temperatures in laser-plasma experiments, and a new unfold method using the HXRS channel charges was subsequently derived. R EFERENCES [1] T. Feurer, W. Theobald, and R. Sauerbrey, “Onset of diffuse reflectivity and fast electron flux inhibition in 528 nm laser–solid interaction at ultrahigh intensity,” Phys. Rev. E, vol. 56, no. 4, pp. 4608–4614, Oct. 1997. [2] H. Chen, P. K. Patel, D. F. Price, B. K. Young, P. T. Springer, R. Berry, R. Booth, C. Bruns, and D. Nelson, “A compact electron spectrometer for hot electron measurement in pulsed laser–solid interaction,” Rev. Sci. Instrum., vol. 74, no. 3, pp. 1551–1554, Mar. 2003. [3] K. B. Wharton, S. P. Hatchett, S. C. Wilks, M. H. Key, J. D. Moody, V. Yanovsky, A. A. Offenberger, B. A. Hammel, M. D. Perry, and C. Joshi, “Experimental measurements of hot electrons generated by ultraintense (> 1019 W/cm2 ) laser-plasma interactions on solid-density targets,” Phys. Rev. Lett, vol. 81, no. 4, pp. 822–825, Jul. 1998. [4] G. Pretzler, T. Schlegel, E. Fill, and D. Eder, “Hot electron generation in copper and photopumping of cobalt,” Phys. Rev. E, vol. 62, no. 4, pp. 5618–5623, Oct. 2000. [5] A. Rousse, P. Audebert, J. P. Geindre, F. Falliès, and J. C. Gauthier, “Efficient Kα X-ray source from femtosecond laser-produced plasmas,” Phys. Rev. E, vol. 50, no. 3, pp. 2200–2207, Sep. 1994. [6] W. Priedhorsky and D. Lier, “APACHE: A low background hard X-ray spectrometer for 10.6 μm laser fusion experiments,” Rev. Sci. Instrum., vol. 53, no. 8, pp. 1189–1197, Aug. 1982. [7] M. Schnürer, M. P. Kalashnikov, P. V. Nickles, T. Schlegel, W. Sandner, N. Demchenko, R. Nolte, and P. Ambrosi, “Hard X-ray emission from intense short pulse laser plasmas,” Phys. Plasmas, vol. 2, no. 8, pp. 3106– 3110, Aug. 1995. [8] F. Beg, A. R. Bell, A. E. Dangor, C. N. Danson, A. P. Fews, M. E. Glinsky, B. A. Hammel, P. Lee, P. A. Norreys, and M. Tatarakis, “A study of picosecond laser-solid interactions up to 1019 W · cm−2 ,” Phys. Plasmas, vol. 4, no. 2, pp. 447–457, Feb. 1997. [9] J. Yu, Z. Jiang, J. C. Kieffer, and A. Krol, “Hard X-ray emission in high intensity femtosecond laser–target interaction,” Phys. Plasmas, vol. 6, no. 4, pp. 1318–1323, Apr. 1999. [10] B. Luther-Davies, “X-ray bremsstrahlung and fast ion measurements from picosecond laser-produced plasma,” Opt. Commun., vol. 23, no. 1, pp. 98–104, Oct. 1977. [11] M. H. Key, M. D. Cable, T. E. Cowan, K. G. Estabrook, B. A. Hammel, S. P. Hatchett, E. A. Henry, D. E. Hinkel, J. D. Kilkenny, J. A. Koch, W. L. Kruer, A. B. Langdon, B. F. Lasinski, R. W. Lee, B. J. MacGowan, MEADOWCROFT AND EDWARDS: DIAGNOSTICS TO CHARACTERIZE HOT-ELECTRON PRODUCTION [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] A. MacKinnon, J. D. Moody, M. J. Moran, Q. A. Offenberger, D. M. Pennington, M. D. Perry, T. J. Phillips, T. C. Sangster, M. S. Singh, M. A. Stoyer, M. Tabak, G. L. Tietbohl, M. Tsukamoto, K. Wharton, and S. C. Wilks, “Hot electron production and heating by hot electrons in fast ignitor research,” Phys. Plasmas, vol. 5, no. 5, pp. 1966– 1972, May 1998. S. Coda, “Diagnostic techniques for measuring suprathermal electron dynamics in plasmas,” Rev. Sci. Instrum., vol. 79, no. 10, p. 10F501, Oct. 2008. M. Schnürer, R. Nolte, T. Schlegel, M. P. Kalachnikov, P. V. Nickles, P. Ambrosi, and W. Sandner, “On the distribution of hot electrons produced in short-pulse laser-plasma interaction,” J. Phys. B: At. Mol. Opt. Phys., vol. 30, no. 20, pp. 4653–4661, Oct. 1997. A. Rubenchik and S. Withowski, Physics of Laser Plasma, 3rd ed. New York: Elsevier, 1991. H. Schwoerer, P. Gibbon, S. Düsterer, R. Behrens, C. Ziener, C. Reich, and R. Sauerbrey, “MeV X-rays and photoneutrons from femtosecond laser-produced plasmas,” Phys. Rev. Lett., vol. 86, no. 11, pp. 2317–2320, Mar. 2001. P. Gibbon, Short Pulse Laser Interactions With Matter: An Introduction, 1st ed. London, U.K.: Imperial College Press, 2005. T. E. Booth, F. B. Brown, J. S. Bull, R. A. Forster, J. T. Gurley, H. G. Huges, R. Martz, R. E. Prael, A. Sood, J. E. Sweezy, and A. Zukaitis, MCNP—A General Monte Carlo N-Particle Transport Code, 5th ed. Belmont, CA: Wadsworth, Apr. 2003. E. Lefebvre, private communication, 2008. M. Ramsay, private communication, 2008. E. Lefebre, N. Cochet, S. Fritzler, V. Malka, M. M. Al Onard, J. F. Chemin, S. Darbon, L. Disdier, J. Faure, A. Fedotoff, O. Landoas, G. Malka, V. M. Ot, P. Morel, M. Rabec, L. Gloahec, A. Rouyer, C. Rubbelynck, V. Tikhonchuk, R. Wrobel, P. Audebert, and C. Rousseaux, “Electron and photon production from relativistic laserplasma interactions,” Nucl. Fusion, vol. 43, no. 7, pp. 629–633, Jul. 2003. G. F. Knoll, Radiation Detection and Measurement, 3rd ed. New York: Wiley, 2000. 2001 [22] T. E. Cowan, M. Roth, J. Johnson, C. Brown, M. Christl, W. Fountain, S. Hatchett, E. A. Henry, A. W. Hunt, M. H. Key, A. MacKinnon, T. Parnell, D. M. Pennington, M. D. Perry, T. W. Phillips, T. C. Sangster, M. Singh, R. Snavely, M. Stoyer, Y. Takahashi, S. C. Wilks, and K. Yasuike, “Intense electron and proton beams from Petawatt laser–matter interactions,” Nucl. Instr. Methods A, vol. 455, no. 1, pp. 130–139, Nov. 2000. [23] Y. Hayashi, A. Fukumi, K. Matsukado, M. Mori, H. Kotaki, M. Kando, L. M. Chen, I. Daito, S. Kondo, S. Kanazawa, A. Yamazaki, K. Ogura, M. Nishiuchi, M. Kado, A. Sagisaka, S. Nakamura, Z. Li, S. Orimo, T. Homma, and H. Daido, “Estimation of photon dose generated by a short-pulse high-power laser,” Radiat Prot Dosimetry, vol. 121, no. 1, pp. 99–107, Dec. 2006. [24] D. Neely, R. Clarke, P. Brummitt, J. Collier, C. Danson, A. Frackiewicz, J. Govans, S. Hancock, P. Hatton, S. Hawkes, R. Heathcote, C. Hernandez-Gomez, P. Holligan, C. Hooker, M. H. R. Hutchinson, A. Kidd, W. Lester, D. McAllister, J. McLaughlan, D. Neville, P. Norreys, D. Pepler, M. Pitts, C. Reason, I. Ross, R. Wellstood, B. Wyborn, T. Winstone, P. Wright, R. Wyatt, and C. Ziener, “Radiological commissioning of the Vulcan Petawatt facility,” RAL Rep. RAL-IR-2003-018, Central Laser Facility Annu. Rep., 178-180, 2002/2003. [25] R. J. Clarke, private communication, 2009. [26] R. D. Edwards, M. A. Sinclair, T. J. Goldsack, K. Krushelnick, F. N. Beg, E. L. Clark, A. E. Dangor, Z. Najmudin, M. Tatarakis, B. Walton, M. Zepf, K. W. D. Ledingham, I. Spencer, P. A. Norreys, R. J. Clarke, R. Kodama, Y. Toyama, and M. Tampo, “Characterisation of a gamma-ray source based on a laser-plasma accelerator with applications to radiography,” Appl. Phys. Lett., vol. 80, no. 12, pp. 2129–2132, Mar. 2002. [27] WaveMetrics Inc., IGOR Pro Version 5.02, 2004. Authors’ photographs and biographies not available at the time of publication.
© Copyright 2026 Paperzz