Compact Thomson Scattering Source Based on a Mixed Injection Assisted Laser Wakefield Accelerator

In order to establish a compact all-optical Thomson scattering source, experimental studies were conducted on the 45TW Ti: sapphire laser facility. By including a steel wafer, mixed gas, and plasma mirror into a double-exit jet, several mechanisms, such as shock-assisted ionization injection, ionization injection, and driving laser reﬂection, were integrated into one source. So, the source of complexity was remarkably reduced. Electron bunches with central energy ﬂuctuating from 90 to 160MeV can be produced. Plasma mirrors were used to reﬂect the driving laser. The scattering of the reﬂected laser on the electron bunches led to the generation of X-ray photons. Through comparing the X-ray spots under diﬀerent experimental conditions, it is conﬁrmed that the X-ray photons are generated by Thomson scattering. For further application, the energy spectra and source size of the Thomson scattering source were measured. The unfolded spectrum contains a large amount of low-energy photons besides a peak near 67keV. Through importing the electron energy spectrum into the Monte Carlo simulation code, the diﬀerent contributions of the photons with small and large emitting angles can be used to explain the origin of the unfolded spectrum. The maximum photon energy extended to about 500keV. The total photon production was 10 7 /pulse. The FWHM source size was about 12 μ m.


Introduction
When photons are scattered by relativistic electrons, the socalled omson scattering process can increase the photon energy by approximately 4c 2 times [1,2], where c is the normalized electron energy. Based on this process, the omson scattering source possesses a set of valuable features, including compactness, tunable photon energy, potential for the duration of femtosecond time scale, and high brightness [3,4]. In many application areas, such as phasecontrast imaging of organisms [5], shadow imaging of high-Z materials [6], stimulating nuclear resonance fluorescence (NRF) of specific isotopes [7][8][9], and ultra-fast pump-probe detection of transient physical and chemical processes [10], the omson scattering source may play a unique role.
For established, conventional omson scattering sources [11][12][13][14][15], where the electrons are produced by a classical electron accelerator, the huge size and ultra-expensive cost required by the large-scale classical electron accelerators restrict further applications. As an improvement, an all-optical omson scattering X-ray source has been proposed by Catravas et al. in 2001 [16], where the electrons are produced by the laser-plasma wakefield accelerator.
is accelerator can produce quasimonoenergetic electron beams with low divergence, small energy spread, and ultra-short pulse duration [17][18][19]. Due to the ultra-high accelerating gradient, electrons can be accelerated toward tens of MeV to several GeV within a distance of mm scale [20][21][22][23]. So, the all-optical omson scattering source is a very promising way to establish a compact, ultra-fast, quasimonoenergetic X-ray source. e first successful experimental results of all-optical omson scattering sources were obtained in 2006 [24], where photons with energies ranging from 400 eV to 2 keV were produced by two counter-propagating laser pulses. Based on these two laser collision models, several studies have been conducted, including the improvement of the source stability [25,26], extending the electron maximum energy to 18 MeV through nonlinear omson scattering [27], and a way to obtain sources with tunable energy [28]. High-order multiphoton (n, 500) omson scattering signals were also obtained in 2017 [29]. In 2012, a compact alloptical omson scattering source was reported [30], where a plasma mirror (PM) was used to directly re ect the driving laser. In this method, the temporal and spatial synchronization between laser and electrons can be realized more easily. Several compact all-optical omson scattering sources have been established based on this method [31][32][33][34].
In order to improve the stability of the generation of wake eld electrons and omson scattering sources, the shock assistant ionization injection technique has proved its validity [35,36]. e studies on the cascaded acceleration technique [37][38][39] also show the advantages of this technique for producing high-quality electrons.
e above studies show that the integration of multi-injection technologies within one accelerator has the potential to generate omson scattering sources with better stability and higher production. But at present, the studies on the all-optical omson scattering source containing multi injection and acceleration techniques are insu cient. e integration of multitechnologies within one accelerator requires proper designing on the compactness of the accelerator. In our previous studies [40], we have established a compact laser wake eld accelerator where shock-assisted ionization injection and ionization injection were integrated into a double-exit gas jet. Based on this accelerator and a plasma mirror attached to the gas jet, a compact all-optical omson scattering source can be established. In the following paragraphs, the experimental layout and the obtained electron energy spectra are shown. en, after the driving laser was re ected by the plasma mirror, the scattering of the re ected laser on the electron bunch led to the generation of X-ray photons. rough comparing the X-ray spots under di erent experimental conditions, it is con rmed that the X-ray photons were produced by omson scattering. e location of the plasma mirror was also optimized. At last, the energy spectra and source size of the omson scattering source were measured for further application. Research Center of CAEP in Mianyang. e laser is linearly polarized in the horizontal direction. e laser focal spot is 11.9 × 7.8 μm full width at half maximum (FWHM) with 60% energy in the spot. In the experiments, the initial input power was set at 28 TW. e total energy coupling e ciency to the target is 70%. So, the peak laser intensity at the focus is I 1.81 × 10 19 W/cm 2 , corresponding to a normalized intensity of a 0 2.9.

Experimental Parameters
e experimental layout is shown in Figure 1(a). A gas jet containing two nozzles is used as the target.
e rst nozzle is a cylindrical hole with an exit diameter of 1 mm and higher gas pressure. e second nozzle is a conical hole with an exit diameter of 2 mm and lower gas pressure. e shock front is produced by the impact between the gas jets from the two nozzles. e addition of a wafer can reduce the exit size of nozzle 1, leading to higher jet density and a shock front with higher density and a steeper falling edge. Mixed gas containing 97.5% He and 2.5% N 2 is used. In the rst nozzle, the electron is injected from shock-assisted ionization injection. en, the accelerated electrons are injected into the second nozzle. A 75 m thick plasma mirror (PM) made of polyethylene terephthalate (PET) is attached to the exit edge of the second nozzle. After being re ected by the plasma mirror, the driving laser interacts with the wake eld electrons for the second time, generating X-ray photons.
A spectrometer consisting of a permanent magnet (0.8 T with a length of 230 mm) and a phosphor screen imaged using a 16-bit CCD camera is used to de ect the electrons and measure the electron energy spectra.
e produced X-ray photons emit out of the vacuum chamber through a 300 μm beryllium window. An X-ray camera consisting of a CsI: Tl scintillator, a ber taper, and a 16-bit CCD sensor records the spatial distribution of the X-ray photons. e X-ray spectra are measured by using a lter stack spectrometer [41]. e X-ray source size is measured using the knife-edge imaging method [42].

Experimental Results
In our previous works, we established a laser wake eld accelerator with a mixed injection mechanism [40]. Based on this accelerator, stable monoenergetic wake eld electron beams with a central energy of 60 MeV were experimentally obtained. To increase the electron energy, the gas pressure of the jet was increased from 650 kPa to 1100 kPa. Under the higher gas pressure, electrons can be steadily produced, but the electron central energy fluctuates from 90 MeV to 160 MeV, as shown in Figure 2. All of the electron energy spectra in this paper are on the same color bar as in Figure 2.
e instability of electron spectra in Figure 2 comes from the instability of the gas pressure and laser energy. e nonlinear evolution of the laser in the relatively high gas pressure (1100 kPa) also leads to the fluctuation of electron energy. Actually, when the gas pressure was adjusted to about 600 kPa as in our previous experiments [40], the electron spectra with better stability and lower production were obtained. In our further studies, the optimization of the gas jet may lead to a better control of the injection process and smaller energy spread. e improvement in the stability and output power of the laser facility may lead to more stable electron production and higher electron energy.
Based on this accelerator, several kinds of X-ray sources can be produced. A omson scattering source can be produced through the scattering of the reflected driving laser on the electrons. Betatron radiation with photon energy ranging from hundreds of electronvolts to several tens of kiloelectron volts can be produced through the oscillation of the electrons in the wakefield. Bremsstrahlung can be produced when the electrons penetrate the plasma mirror. e direct interaction between the driving laser and the plasma mirror may also produce bremsstrahlung. To distinguish the signals of omson scattering from the other sources, the X-ray spots under different conditions are recorded as shown in Figure 3. In Figure 3(a), a pristine plasma mirror is used, and wakefield electrons are efficiently produced. All of the possible signals, such as signals from the omson scattering, Betatron radiation, and bremsstrahlung, could be produced and recorded by the X-ray camera. Figure 3(a) shows that strong and clear X-ray signals were recorded. en, when there is no plasma mirror, the driving laser cannot be reflected. Bremsstrahlung and omson scattering radiation cannot be generated. Betatron radiation can be generated, but only photons with energy higher than 2 keV can penetrate the beryllium window and get recorded by an X-ray camera. e obtained result is shown in Figure 3(b), where no obvious signal is recorded. When the plasma mirror was placed farther from the gas jet, defocusing of the laser reduced the corresponding intensity. So, the laser reflection on the plasma mirror and the radiation intensity from the omson scattering were drastically reduced. Only Betatron radiation and bremsstrahlung can be produced as a result of electron-plasma mirror interaction. But similar to Figure 3(b), the recorded result in Figure 3(c) shows no obvious result. So, it is clear that the contributions of betatron radiation and bremsstrahlung can be ignored.
Finally, the variability of the electron injection at higher pressures causes instability of the wakefield acceleration. Some shots under the same conditions as in Figure 3(a) resulted in poor electron beam generation. Figure 3(d) gives an example of such a result with no obvious X-ray images from omson scattering. All of the results in Figure 3 show that only with pristine, properly placed plasma mirror, and efficient laser wakefield acceleration, distinct X-ray signals can be recorded. Only effective omson scattering can result in the X-ray spot as shown in Figure 3(a). In Figure 3(a), the FWHM emitting angle of the X-ray source is 30 mrad.
To find the optimized position of the plasma mirrors, the average intensity of the X-ray spots recorded by CCD at different plasma mirror positions is shown in Figure 4. In this figure, position 0 corresponds to the exit edge of the gas jet. When the plasma mirror moves away from the gas jet, it moves from the edge part of the gas jet to a vacuum. e optimized plasma mirror position of 0.1 mm shown in Figure 4 reflects the position in the edge part of the gas jet with the highest laser reflectivity. en, in vacuum, when the plasma mirror moves away from the gas jet, the laser beam size increases at the colliding point, leading to a decrease in laser intensity, reflectivity, and X-ray intensity. e incidence angle of laser photons in vacuum may be larger than the incidence angle of the guided laser in the gas jet. So qualitatively, the energy spread of the X-ray spectra in vacuum will be larger than that in the gas jet. But when the plasma mirror moves in vacuum, the interaction angles between electrons and laser photons stay unchanged, leading to a nearly unchanged spectra profile.
Next, we used the previously established filter stack spectrometer [41] to measure the energy spectra of the obtained omson scattering source. For the electron bunch with the energy spectrum as in Figure 5(a), the different X-ray profiles behind a set of stacked filters are shown in Figure 5(b). Based on the results in Figure 5(b), the input X-ray spectrum within the emitting angle of 20 mrad can be unfolded. e unfolded spectrum in Figure 5(c) shows that the spectrum contains a large number of low-energy photons besides a peak near 67 keV. e maximum photon energy extends to about 500 keV. In the electron energy spectrum in Figure 5 500 keV can be attributed to the omson scattering of laser photons on these high-energy electrons. e contribution of the nonlinear omson scattering e ect should also be considered. In our experiments, the peak laser intensity at the focus is I 1.81 × 10 19 W/cm, corresponding to a normalized intensity of a 0 2.9. e laser spot at 0.1 mm away from the jet exit is nearly unchanged. If the re ectivity on the plasma mirror is assumed to be 50%, the intensity of the re ected can be estimated as 0.9 × 10 19 W/cm, corresponding to normalized intensity a 0 2.05. So, under this intensity, the nonlinear Compton scattering e ect may lead to an inapparent contribution. Based on the obtained spectrum in Figure 5(c), the total photon production can be estimated as 10 7 /pulse. In order to nd the origin of these low-energy photons, the electron energy spectrum in Figure 5(a) can be imported into the Monte Carlo simulation code CAIN [43] to obtain the theoretical radiation spectrum. e calculation parameters are the same as the experimental parameters. e relation between photon energy and emitting angle is shown in Figure 6(a), which is a traditional angular-energy distribution of the omson scattering source. e red and yellow dots in Figure 6 with the highest weight, re ecting the electrons in the peak of the electron energy spectrum in Figure 5(a). Due to momentum and energy conservation, the energy of the photons scattered toward a larger angle is always smaller. On the contrary, photons with higher energy mainly emit out within an angle smaller than 5 mrad. e integrated photon energy spectra for integration angles of 5, 10, and 20 mrad are shown in Figure 6(b). e red solid line in Figure 6(b) shows that the photons emitting out at an angle of 5 mrad mainly contribute to the higher energy part of the spectrum. When photons within a larger emitting angle are included in the integration, the black and blue dashed lines in Figure 6(b) show that the low-energy part of the spectra is drastically enhanced. e di erent contributions of the photons with small and large emitting angles can be used to explain the similar spectral shapes in Figures 5(c) and 6(b). But restricted by the limited lter numbers in the lter stack spectrometer, the photon energy of the unfold spectrum in Figure 5(c) is a bit lower than the calculated results in Figure 6(b). In our further studies, the energy spectra may be more precisely measured with the improvement in the lter stack spectrometer. e obtained source can be used for radiography, where the resolution is one of the most important parameters. For a knifeedge made of 100 μm lead foil, its radiograph image is shown in Figure 7(a). If the source pro le is set as a Gaussian shape, the unfolded source size (FWHM) is 12 μm, as shown in Figure 7(b). It also shows that a gold grid used in the electron microscope with a grid size of 2000 mesh × 12.5 μm pitch per inch and a diameter of 3.05 mm was also used. e thickness of

Summary and Discussion
is work reports our studies on the compact all-optical omson scattering source. By including a steel wafer, mixed gas, and plasma mirror into a double-exit jet, several mechanisms, such as shock-assisted ionization injection, ionization injection, and driving laser re ection, are integrated. In this compact source, only one gas supply system is required. e source complexity is remarkably reduced. e laser wake eld accelerator with a mixed injection mechanism can steadily produce electron bunches with a central energy uctuating from 90 MeV to 160 MeV. en, the driving laser is re ected by the plasma mirror and scatters on the electron bunch, leading to the generation of X-ray photons. By comparing the X-ray spots under di erent experimental conditions, it is con rmed that the recorded signals by the X-ray camera are produced by omson scattering. For further application, we used a lter stack spectrometer to measure the X-ray energy spectra. e unfolded spectrum has a similar spectral shape to the calculated results, which can be explained by di erent contributions of photons at small and large emitting angles. e maximum photon energy extends to about 500 keV. e total photon production is 10 7 /shot. By using the knife-edge method, the FWHM source size can be measured as about 12 μm.
In our further studies, the optimization of the gas jet may lead to better control of the injection process and smaller energy spread. e improvement in the stability and output power of the laser facility may lead to a more stable electron production and higher electron energy. Moreover, the energy spectra and source size can be more precisely measured with the modi cation of the corresponding diagnostic techniques.
Data Availability e data, models, and code generated or used in this study are included within the article.

Conflicts of Interest
e authors declare that they have no con icts of interest.