betaFeedback

WorksPhysics and Astronomy

Fast electrons and high energy photons produced in strong-field ionization of atoms as diagnostics of multi-petawatt laser beams

Abstract

We calculate numerically energy spectra and angular distributions of photoelectrons and X-ray photons emitted in the process of tunnel photoionization of argon by a focused laser beam of multi-petawatt power. By scanning the beam power and varying the degree of focusing and the field polarization, we show these distributions to be sensitive both to the peak laser intensity and to the field configuration in the focal spot. Investigated parameters of the laser beam correspond to those of several recently commissioned lasers of multi-petawatt power. These findings, in line with the earlier reported proposals, provide a basis for the development of a reliable ionization diagnostics of extremely intense laser fields.

Introduction

The development of super-intense laser sources of electromagnetic radiation has demonstrated considerable progress over the past two decades. Several multi-petawatt laser facilities have been put into operation [1, 2, 3, 4, 5]. Even more ambitious projects of sub-exawatt sources are currently under development and construction [6, 7]. Among the multiple potential applications of petawatt (PW), multi-PW and exawatt sources of ultrashort laser pulses, studies of fundamental effects of the interaction of electromagnetic field with charged particles in nonlinear and non-perturbative regimes are of primary significance for fundamental physics. These include investigations of the radiation reaction force in classical and quantum regimes, relativistic ionization, effects of nonlinear quantum electrodynamics, various schemes of laser acceleration of electrons and ions, and more. The current status of this thriving research area is outlined in reviews [8, 9, 10, 11] and the literature quoted there. The precise knowledge of the peak intensity, which can be achieved in the focus of a super-intense laser beam, and of the intensity distribution and field configuration in the focal area is of key significance for the reliable analysis of experimental data as well as for the determination of principal requirements for the search of certain effects with the use of a given laser source. This raises a high demand for methods of direct diagnostics of extreme electromagnetic fields in the foci of multi-PW laser beams. To clarify terminology, note that we call diagnostics direct if it is based on measurements of effects, which take place in the laser focus in the full power mode of the beam. Several nonlinear effects, which can serve as sensitive tools for field diagnostics, have been examined in theory and, partially, experimentally. An incomplete list includes tunnel ionization of multielectron atoms [12, 13, 14, 15, 16, 17], acceleration of electrons and ions created in the laser focus through ionization [18, 19, 20, 21], nonlinear Thomson and Compton scattering of photons of an intense laser pulse on a beam of fast electrons [22, 23, 24]. For further references, we direct the reader to review [11].

Among these effects, tunnel ionization of heavy atoms and ions possesses two significant advantages: it is highly nonlinear and local. Each of these features potentially allows us to probe the laser intensity and measure its peak value with good accuracy [15, 16]. However, the method has also certain shortcomings: (a) the tunnel ionization diagnostics appears most sensitive to the value of the laser intensity only when the latter is close to the ionization threshold of s-electrons [17] and (b) in the non-relativistic domain, which for tunnel ionization spreads up to ≃1024−1025 W/cm2\simeq10^{24}-10^{25}\,\mathrm{W}/\mathrm{cm}^{2} the ionization probability is solely determined by the electric field amplitude so that the method probes intensity and not the electric and magnetic fields vectors. At the same time, in an experiment targeted at measurements of production of highly charged ions [12, 13, 14], one could register also fast electrons produced through nonlinear ionization [20, 21] and high-energy photons emitted by such electrons while they leave the focus. Such three-fold diagnostics realized in the frame of one experiment could essentially improve the accuracy of measurements, both for the intensity peak amplitude and also for the field distribution in the focal area.

In this paper, we present results of numerical calculations of electrons and photons energy-momentum distributions for laser parameters corresponding to the current status of the multi-PW Apollon laser facility [5]. We change the hitherto unknown parameters of the beam including its degree of focusing and the peak value of the field amplitude in the focus center to trace the evolution of electron and photon distributions triggered by this change. To this end, we developed a Monte Carlo (MC) code, which probabilistically simulates tunnel ionization and allows calculating electron trajectories from the instant of birth near the parent ion to the escape from the laser focus. These trajectories are then used to find the spectra of emitted photons according to the method applied in our earlier work [25]. The MC code developed in this work allows greatly accelerating the calculations, compared to a particle-in-cell (PIC) code, because the statement of the problem excludes the necessity of calculation of the back reaction of charges on the electromagnetic field configuration.

The paper is organized as follows. In the next Section, we describe the statement of the problem and introduce basic equations used for the calculation of electron and x-ray spectra. Section III describes the numerical set-up including the Monte Carlo code. Electron and photon spectra and their evolution with the beam peak power and the focal field configuration are discussed in Section IV. The last Section contains conclusions and outlook.

Statement of the problem

We consider the dynamics of photoelectrons created through tunnel ionization of argon atoms in the focus of an intense laser pulse and calculate momentum distributions of such electrons after escape from the laser field and angular-resolved spectra of X-ray photons emitted by these electrons during their motion. Parametrically, we focus on the characteristics of the Apollon multi-PW laser in its present status [5]: the FWHM pulse duration is τ=22fs\tau= 22\mathrm{fs}; the pulse energy is W=70JW = 70\mathrm{J}; the peak power Pmax⁡=3.5PWP_{\max} = 3.5\mathrm{PW} at the wavelength λ=0.8μm\lambda= 0.8\mathrm{\mu m}; the expected focal waist is w0=2.1μmw_0 = 2.1\mathrm{\mu m} (considering Gaussian beam; FWHM waist 2.5 mum); the ellipticity is ρ≈0.15\rho\approx0.15. At these parameters, the peak intensity in the central part of the focus is Imax⁡≈3.5⋅1022W/cm2I_{\max} \approx3.5 \cdot10^{22}\mathrm{W}/\mathrm{cm}^{2}. However, neither the peak power at the target, nor the waist w0w_0 are accurately known, and the uncertainties have to be minimized using diagnostics, whose possible realization we discuss here. Thus our aim is to study the sensitivity of the photoelectron and photon distributions to the peak intensity, which we will vary by changing either the peak power Pmax⁡P_{\max} or the spot size w0w_0.

For the gas target, we choose argon as an element widely used in strong-field ionization experiments. Considering the interval of intensities Imax⁡=1022−1023W/cm2I_{\max} = 10^{22} - 10^{23}\mathrm{W}/\mathrm{cm}^{2}, which is likely to be explored at Apollon for the current level of the peak power, xenon and krypton could look more suitable targets [15], because there are levels there to be efficiently ionized in this intensity interval. However, for the first proof-of-principle experiments, we suggest argon, which is cheaper, and its ionization potentials allow probing the domain Imax⁡=1021−5⋅1022W/cm2I_{\max} = 10^{21} - 5 \cdot10^{22}\mathrm{W}/\mathrm{cm}^{2}, which can be reliably achieved with a few-PW level source. The concentration of atoms is assumed to be low, nat≃1011cm−3n_{\mathrm{at}} \simeq10^{11}\mathrm{cm}^{-3}, which allows one to neglect space charge effects and interaction between photoelectrons and other ions.

We consider three configurations of the field. The first one is a linearly polarized electromagnetic field of a Gaussian laser pulse propagating in the direction of the xx axis with zero longitudinal fields.

E(r,t)=E01+x2/xR2exp⁡(−r2w02[1+x2/xR2])×cos⁡(kx−ωLt−arctan⁡(xxR)+kr22x[1+r2/xR2])×f(ωLt−kx),r=y2+z2,H=n×E,n=kk.(1)\begin{aligned} \mathbf{E}(\mathbf{r},t) &= \frac{\mathbf{E}_0}{\sqrt{1+x^2/x_R^2}} \exp\left(-\frac{r^2}{w_0^2[1+x^2/x_R^2]}\right) \\ &\quad\times\cos\left(kx-\omega_L t-\arctan\left(\frac{x}{x_R}\right)+\frac{kr^2}{2x[1+r^2/x_R^2]}\right) \\ &\quad\times f(\omega_L t-kx), \qquad r=\sqrt{y^2+z^2}, \\ \mathbf{H} &= \mathbf{n}\times\mathbf{E}, \qquad\mathbf{n}=\frac{\mathbf{k}}{k}. \tag*{(1)} \end{aligned}

where E0E_0 is the electric field peak amplitude, ff is the time envelope of total duration τ=2Mπ/ωL≡MTL\tau= 2M\pi/\omega_L \equiv MT_L, xR=πw02/λx_R = \pi w_0^2/\lambda is the focal Rayleigh length, ωL\omega_L and TLT_L are the frequency and the laser period, correspondingly. The vector E0\mathbf{E}_0 is directed along the yy-axis. The time envelope has the form:

f(ωLt−kx)={cos⁡2(ωLt−kx2M),ωLt−kx∈[−πM,πM],0,ωLt−kx∉[−πM,πM].(2)f(\omega_L t-kx)= \begin{cases} \cos^2\left(\dfrac{\omega_L t-kx}{2M}\right), & \omega_L t-kx\in[-\pi M,\pi M],\\ 0, & \omega_L t-kx\notin[-\pi M,\pi M]. \end{cases} \tag*{(2)}

Note that Eq.(1) is valid in the paraxial approximation when the pulse duration is sufficiently long, ωLτ≫kw0\omega_L\tau\gg kw_0 [26]. For the parameters introduced above, one obtains ωLτ≈40\omega_L\tau\approx40, and kw0≈20kw_0 \approx20 so that Eq.(1) is on the edge of its applicability.

The second field configuration is also a linearly polarized Gaussian pulse but with but with the magnetic field defined as

H=−c⋅rot⁡∫tE(t′) dt′(3)\mathbf{H}=-c\cdot\operatorname{rot}\int^t \mathbf{E}(t')\,\mathrm{d}t' \tag*{(3)}

and has a non-vanishing longitudinal component. For the considered parameters, ∣Hx∣≈0.05E0|H_x| \approx0.05E_0.

Finally, the third field configuration is a Gaussian pulse with elliptical polarization at small ellipticity ρ=0.15\rho= 0.15, which we assume relevant for realistic focusing optical elements. In this case, the field component EyE_y is given by Eq. (1) divided by 1+ρ2\sqrt{1+\rho^2}, and Ez(r,t)=ρEy(r,t−π/2ω)E_z(r,t)=\rho E_y(r,t-\pi/2\omega). The magnetic field is elliptically polarized accordingly, however, we omit the longitudinal component. This way we study the impact of the longitudinal field and ellipticity independently.

The real structure of the electromagnetic field in the focal area is by far more complicated than that of the Gaussian beam. A realistic field distribution, which can be obtained by solving numerically the Maxwell equations, e.g. using a PIC code in the particle-free mode [27], demonstrates a diffraction structure in the form of Airy rings. However, as we are solely interested here in ionization of inner atomic levels with binding energies of hundreds and thousands eV (the ionization potential of the hydrogen-like Ar17+^{17+} is Ip=4426I_p = 4426 eV) and in the dynamics and radiation of photoelectrons accelerated up to relativistic gamma-factors γ≃102−103\gamma\simeq10^2-10^3, the central part of the focal area, where the field is relativistically strong, is only of interest for our calculations. Significance of relativistic effects is essentially determined by the value of the dimensionless invariant field amplitude

a0=eE0mcωL,(4)a_0 = \frac{eE_0}{mc\omega_L}, \tag*{(4)}

which is a0=102a_0 = 10^2 for a linearly polarized field of wavelength λ=1μm\lambda= 1\mu\mathrm{m} and intensity I=2⋅1022 W/cm2\mathcal{I}=2\cdot10^{22}\,\mathrm{W/cm^2}. For the description of the central part of the focal area, the approximation of Gaussian beams works reasonably well for w0/λ≈2−3w_0/\lambda\approx2-3 and higher. For tighter focusing, different exact solutions for stationary beams [26, 28] can be used instead of the Gaussian family.

Tunnel ionization

Here we introduce the analytic formalism for description of the ionization step. The model is based on the theory of non-relativistic ionization in static and low-varying electromagnetic fields proposed by Smirnov and Chibisov (SC) [29] and Perelomov, Popov and Terentiev (PPT) [30, 31, 32, 33]. In this subsection, we use atomic units e=m=ℏ=1e=m=\hbar=1.

Ionization proceeds in the tunnel regime specified by smallness of the Keldysh parameter [34, 32].

γK=2IpωLE0,(5)\gamma_K = \frac{\sqrt{2I_p\omega_L}}{E_0}, \tag*{(5)}

For laser fields with wavelengths λ≃1 μm\lambda\simeq1\ \mu\mathrm{m} and intensities I≳1018 W/cm2I \gtrsim10^{18}\ \mathrm{W/cm^2} γK≃10−2−10−3\gamma_K \simeq10^{-2}-10^{-3} (see [15] for estimates), so that the tunnel regime is well secured. Relativistic effects in the rate of ionization play no significant role up to intensities ≃1024 W/cm2\simeq10^{24}\ \mathrm{W/cm^2} [35, 36].

Under these conditions, the ionization probability is determined by the instant electric field amplitude at the jj-th atom position E(Rj,t)\mathbf{E}(\mathbf{R}_j,t). Atoms are considered at rest during the femtosecond pulse duration. We assume that the electrons are removed sequentially with increasing ionization potentials. Deviations from this simplest sequence of ionization were identified in a recent paper [17]. However, they do not considerably affect the number of photoelectrons released at a given intensity. According to the SC and PPT theories [29, 37, 32], the probability for the outer electron with orbital and magnetic numbers $l and mm to be removed from its atom (ion) with residual charge ZcZ_c (in these notations, Zc=1Z_c=1 for a neutral atom and Zc=NZ_c=N for A(N−1)+A^{(N-1)+} ion) during the time interval dtdt is given by the PPT static rate [29, 38, 31, 32]

dW(t)=w(Ip,ν,l,m;E(t))dt,dW(t)=w(I_p,\nu,l,m;E(t))dt,
w=22ν+1Cνl2BlmIpF1+∣m∣−2ν(t)exp⁡(−23F(t)).(6)w=2^{2\nu+1}C_{\nu l}^{2}B_{lm}I_pF^{1+|m|-2\nu}(t)\exp\left(-\frac{2}{3F(t)}\right). \tag*{(6)}

Here

F(t)=E2(t)(2Ip)3/2(7)F(t)=\frac{\sqrt{E^2(t)}}{(2I_p)^{3/2}} \tag*{(7)}

is the reduced electric field scaled by the characteristic field Ech=(2Ip)3/2E_{\mathrm{ch}}=(2I_p)^{3/2} of the atomic level,

ν=Zc2Ip(8)\nu=\frac{Z_c}{\sqrt{2I_p}} \tag*{(8)}

is the effective principal quantum number,

Cνl2=22ν−2νΓ(ν+l+1)Γ(ν−l)(9)C_{\nu l}^{2}=\frac{2^{2\nu-2}}{\nu\Gamma(\nu+l+1)\Gamma(\nu-l)} \tag*{(9)}

is the squared asymptotic coefficient of the bound state single-electron wave function calculated in Hartree approximation [32] and the factor

Blm=(2l+1)(l+∣m∣)!22∣m∣∣m∣!(l−∣m∣)!(10)B_{lm}=\frac{(2l+1)(l+|m|)!}{2^{2|m|}|m|!(l-|m|)!} \tag*{(10)}

reflects the angular structure of the bound state.

For the simulations of this work, we secure several important limitations of the SC and PPT rates. Firstly, we focus on dynamics and radiation of high-energy photoelectrons produced in the central part of the focus, where the maximal local laser intensity is close to its peak value. This allows us to consider only ionization of the inner shells of argon, starting from the 2p2p levels. By this, we exclude from consideration the low-energy electrons, which predominantly appear due to ionization of the outer shell in a large low-intensity part of the focal volume. Secondly, the 22fs pulse duration is still sufficient to make sequential tunneling the main ionization mechanism at high intensities [39]. Finally, by modeling the ionization step (see Section III below for details) we check that the given level is being depleted before the reduced field (7) reaches 0.1. For higher values of FF the rate (6) highly overestimates the probability of ionization, and empirical extensions of (6) have to be used to secure quantitative accuracy [32, 40, 41].

Post-ionization photoelectron dynamics and radiation

The post-ionization motion of photoelectrons is described by solving numerically the classical equation of motion

dpdt=eE+ecv×H(11)\frac{d\mathbf{p}}{dt}=e\mathbf{E}+\frac{e}{c}\mathbf{v}\times\mathbf{H} \tag*{(11)}

with the electric and magnetic fields given by Eqs. (1)–(3). Initial conditions are determined by the tunnel ionization mechanism and take into account that the tunnel barrier width is small compared to the laser wavelength:

r(tjn)=Rj,v(tjn)=0.(12)\mathbf{r}(t_{jn})=\mathbf{R}_j,\quad\mathbf{v}(t_{jn})=0. \tag*{(12)}

Here, jj denotes the atom number and nn that of the electron in this atom.

Photoelectron trajectories corresponding to the initial conditions (12) in extremely strong fields are of specific spatial shape [42, 25]. Due to a strong acceleration along the propagation direction of the laser pulse, the electrons make a few or even no single full oscillation before leaving the laser focus. For lower ionization potentials, the electrons predominantly leave the focus through its lateral surface, while the electrons removed from the inner shells at time instants when the laser field was close to its maximum, experience no oscillations and leave the focus through its back surface.

In a plane wave of finite duration, the initial condition (12) leads to a simple relation between the asymptotic value of the longitudinal pxp_x and lateral p⊥p_\perp photoelectron momenta [42, 25]. In terms of the polar emission angle θ=tan⁡−1(p⊥/px)\theta= \tan^{-1}(p_\perp/p_x) and the relativistic gamma-factor γ\gamma this relation reads

γ≈2tan⁡2θ.(13)\gamma\approx\frac{2}{\tan^{2}\theta}. \tag*{(13)}

Note that in a linearly polarized field of type (1) the photoelectron motion is confined in the (x,y)(x,y) plane, despite the focusing. In fields with a longitudinal component (3) or with nonzero ellipticity, the electron trajectories become three-dimensional, which may reduce the accuracy of estimate (13), see Section IV for results and discussion.

The spectral-angular distribution of radiation emitted by an electron moving along the trajectory rjn(t)\mathbf{r}_{jn}(t) is calculated using the pivotal expression of classical electrodynamics [43, 44]

dEjndωdΩn=e2ω24π2c3∣∫tjn∞dt [n×[n×vjn(t)]]eiω[t−n⋅rjn(t)/c]∣2.(14)\frac{d\mathcal{E}_{jn}}{d\omega d\Omega_n} = \frac{e^2\omega^2}{4\pi^2 c^3}\left|\int_{t_{jn}}^{\infty}dt\,[\mathbf{n}\times[\mathbf{n}\times\mathbf{v}_{jn}(t)]]e^{i\omega[t-\mathbf{n}\cdot\mathbf{r}_{jn}(t)/c]}\right|^2. \tag*{(14)}

Here dΩnd\Omega_n is the solid angle in the direction defined by the unit vector n\mathbf{n}.

Numerical set-up and method

The numerical method consists of two sequential stages: (a) Monte-Carlo modeling of ionization and (b) numerical integration of electron trajectories and calculation of the radiation integrals along these trajectories. Atoms are randomly distributed in a rectangular box centered at the focal spot. The box size is 110w0110w_0 along xx axis and 16w016w_0 in the lateral direction, with w0=2.65λw_0 = 2.65\lambda. Such an extended computational volume allows one to ensure that all parts of the space where the 2p2p and lower shells can be ionized at the peak intensity ≃1023\simeq10^{23} W/cm2^2 in the center are included. Assuming the concentration nat≃1011n_{at} \simeq10^{11} cm−3^{-3}, we set the total number of atoms in the box equal to Na=5⋅104N_a = 5\cdot10^4. For the initial configuration, we chose the ground 2p62s21s22p^6 2s^2 1s^2 state of Ar8+^{8+}. This helps to exclude ionization of the 3p63s23p^6 3s^2 shells which proceeds at low intensities beyond the tunneling regime and gives rise to low energy photoelectrons, which are out of interest for this study.

We consider multiple ionization as a sequence of independent steps, which allows us to use parallel methods and speeds up the numerical calculation. The electric field at the position of the atom is evaluated at discrete time steps ωLΔt=0.001\omega_L\Delta t = 0.001, and the ionization probability of the considered level is calculated using the rate (6) multiplied by this time interval securing the condition that this probability always appears small compared to unity. Then a random number uniformly distributed in the interval [0,1)[0,1) is generated, and an ionization event is registered when this number is smaller than the calculated probability. After a successful ionization event, we update the ionic state, and the atom position Rj\mathbf{R}_j and the instant of time tjnt_{jn} become the initial conditions for the post-ionization electron motion.

To find the electron trajectory rjn(t;Rj,tjn)\mathbf{r}_{jn}(t;\mathbf{R}_j,t_{jn}) we solve the equation of motion (11) rewritten in the form of two first-order equations by Runge-Kutta (RK) method of order 4(5). It is realized by Python solve_ivp method, where Runge-Kutta is one of available options. For parallel calculations, the pool.imap is utilized.

First, we calculate the final momenta of the electrons by integrating their equations of motion from the instant of ionization to ωLtout=104\omega_L t_{\mathrm{out}} = 10^4, without storing information about the trajectories. The laser pulse leaves the target at ωLt=103\omega_L t = 10^3, so the integration continues long after that. The trajectories are recorded at uniformly spaced time points with a step ωLδt=0.1\omega_L \delta t = 0.1. Under RK4(5) method the error is controlled assuming accuracy of the fourth-order method, but steps are taken using the fifth-order accurate formula, so the internal integration step size is independent of the time interval δt\delta t used to record the trajectory data. We use the final momenta arrays for plotting angle-resolved photoelectron spectra and the trajectories for the calculation of the emission spectra.

Results and discussion

Photoelectron spectra

We start from a discussion of the photoelectron distributions in the (γ,θ)(\gamma,\theta) plane with θ\theta being the angle between the propagation direction of the beam and the photoelectron final momentum. In a linearly polarized plane wave pulse of finite duration and for a0≫1a_0 \gg1, the γ\gamma factor of an escaping electron is connected to the angle according to (13). This function itself does not depend on a0a_0 and therefore bears no information on the peak laser intensity. Instead, the minimal angle θmin⁡\theta_{\min} clearly seen for the distributions shown in the upper and lower rows of Figure 1, depends on the value of a0a_0 Within this plane wave model, for an electron born at the peak value of the field,

Photoelectron $(\gamma,\theta)$ distributions for different peak intensities and field configurations

Figure 1. Photoelectron (γ,θ)(\gamma,\theta) distributions for different peak intensities and field configurations. The peak intensity is I=5×1021 W/cm2\mathcal{I} = 5 \times10^{21}\,\mathrm{W/cm^2} (left column), I=2.15×1022 W/cm2\mathcal{I} = 2.15 \times10^{22}\,\mathrm{W/cm^2} (middle column) and I=1×1023 W/cm2\mathcal{I} = 1 \times10^{23}\,\mathrm{W/cm^2} (right column). The field configuration is: linear polarization, no longitudinal field (1) (upper row), linear polarization with a longitudinal field (3) (middle row), elliptical polarization with ρ=0.15\rho= 0.15 and no longitudinal field (lower row). The plane wave dependence (13) is shown by a solid blue line. The number of the argon shell is color-coded starting from 2p62p^6 (light-yellow) to 3s13s^1 (black). Note that a considerable contribution of 3s3s electrons is only present at peak intensities I>2⋅1022 W/cm2\mathcal{I} > 2 \cdot10^{22}\,\mathrm{W/cm^2} (middle and right columns). In each panel, the values of γmax⁡\gamma_{\max} and θmin⁡\theta_{\min} are shown as extracted from the simulation results and (on parentheses) calculated along (15).

θmin⁡≈2a0,γmax⁡≈a022(15)\theta_{\min} \approx\frac{2}{a_0}, \qquad\gamma_{\max} \approx\frac{a_0^2}{2} \tag*{(15)}

Blue solid lines in Figure 1 show the analytic scaling (13), which, for the fields without a longitudinal component, reproduce the high energy parts of the distributions remarkably well.

In all cases, there are a low-energy (γ<2\gamma< 2) and a mid-energy (γ<102\gamma< 10^2) haloes populated by the electrons escaping through the lateral surface of the focus [25]. Note that the focal parameter w0=2.7λ≈2.2 μmw_0 = 2.7\lambda\approx2.2\,\mu\mathrm{m} is quite small, so that the approximation of such a focus by a plane wave is essentially rough. However, results in Figure 1 (a-c) and (g-i) show that for entirely transverse field configurations this approximation describes the electron dynamics quite accurately, for the high-energy part of the spectrum. Thus, observation of high-energy electrons near the bounds (15) could provide a better diagnostics of the field near the focal center.

In contrast, the field distribution with a longitudinal magnetic component (3) accelerates the electrons in a different way, so that the angular spread of high-energy electrons appears much broader, and the scaling (13) no longer gives a qualitatively correct description of this part of the spectrum. An experimental observation of such angular-resolved spectra in their high-energy γ>102\gamma> 10^{2} part should in particular give a clear evidence of the presence of longitudinal fields and their relative magnitude.

The main message of the spectra shown in Figure 1 and Figure 2 is in the sensitivity of γmax⁡\gamma_{\max} and θmin⁡\theta_{\min} to the peak laser intensity and in the dependence of this sensitivity on the field configuration in the focus. As is seen from the numbers shown on the panels, a good agreement between the theory and simulations takes place at the lowest intensity. This can be explained by the fact that, in contrast to (13), which does not contain a0a_0, the limiting values (15) do, so that the theoretical estimates are valid under assumption that the most energetic electrons were created at the peak intensity. This is the case for I<5⋅1021 W/cm2\mathcal{I} < 5 \cdot10^{21}\,\mathrm{W}/\mathrm{cm}^{2}, the value determined by the highest ionization potential of argon [15]. A position of the spectral cut-off γmax⁡\gamma_{\max} appears the most reliable measure of the peak intensity even in the presence of the longitudinal field (see (a), (d) and (g) in Figure 1), but under a condition that this intensity is close to the ionization threshold for highly charged ions of argon.

Plots of angular distributions for different beam waists

Figure 2. Same as in Figure 1(h) for different values of the beam waist, w0=2.7λw_{0}=2.7\lambda (a), 2w02w_{0} (b), 3w03w_{0} (c) and 4w04w_{0} (d) at fixed pulse energy W=70W=70 J and duration τ=22\tau=22 fs. The smallest waist (a) corresponds to the peak intensity I=2.15⋅1022\mathcal{I}=2.15\cdot10^{22} W/cm2^{2}, the largest one (d) – to I=2.28⋅1021\mathcal{I}=2.28\cdot10^{21} W/cm2^{2}.

A complementary diagnostics with photoelectrons can be developed by measuring their angular distributions in the plane perpendicular to the beam propagation direction. Such distributions integrated over the energy interval γ=102..103\gamma= 10^{2}..10^{3} are shown in Figure 3 with angle θ\theta as a radial coordinate and the azimuthal angle φ\varphi – as an azimuthal coordinate. Intensities and field configurations correspond to those of the second and the third rows of Figure 1. In the case of a linearly polarized transversal field, the electron motion is precisely confined in the (Ey,k)(E_y,\mathbf{k}) plane so that the angular distributions appear one-dimensional. The central message of this plot is in a strong influence of the longitudinal field on the distributions. Even being as small as 0.05 of the transversal field amplitude, it leads to a considerable spread of the distributions, particularly, at the highest intensity. The distributions shown in the third row and corresponding to the field parameters expected at the Apollon laser facility [5] demonstrate a clear spread along the angle φ\varphi with the peak intensity increasing.

Angular distributions of photoelectrons

Figure 3. Angular distributions of photoelectrons integrated over energy in the range γ=102…103\gamma=10^{2}\ldots10^{3}. The angle θ\theta between the propagation direction of the beam and the electron momentum is shown as a radial coordinate; the azimuthal angle φ\varphi – as an angular coordinate. The panels correspond to those of the second and the third rows of Figure 1.

Radiation spectra

Finally, using Eq. (14) we calculate angular distributions of the emitted radiation. The distributions are integrated in the frequency interval 102ωL<ω<3⋅102ωL10^{2}\omega_L < \omega< 3 \cdot10^{2}\omega_L, where the maxima of the synchrotron-like spectra of radiation are located [25] for all intensities selected for our calculations. As in the plots discussed above, it is clearly seen that the longitudinal field (second row) gives rise to diffuse distributions with a broader spread over the azimuthal angle φ\varphi and a weaker dependence on the peak laser intensity. Apart from this feature, the distributions of the upper (linear polarization) and the lower (elliptical polarization) rows demonstrate a clear tendency of narrowing with the intensity increasing. Note that the distributions of Figure 4 are normalized, so that a growth in the integral photon signal with the intensity increasing is not reflected.

Angular distributions of the emitted photons

Figure 4. Angular distributions of the emitted photons in the frequency interval 102<ω/ωL<3×10210^{2}<\omega/\omega_{L}<3\times10^{2} for laser intensities I=5×1021 W/cm2\mathcal{I}=5\times10^{21}\ \mathrm{W}/\mathrm{cm}^{2} (left column), I=2×1022 W/cm2\mathcal{I}=2\times10^{22}\ \mathrm{W}/\mathrm{cm}^{2} (middle column), I=1×1023 W/cm2\mathcal{I}=1\times10^{23}\ \mathrm{W}/\mathrm{cm}^{2} (right column) and different field configurations: Gaussian beam (upper row), Gaussian beam with longitudinal field (middle row), elliptically polarized Gaussian beam (lower row). Azimuthal angle φ\varphi is measured in the plane perpendicular to the laser wave vector k\mathbf{k}, relative to the electric field vector E\mathbf{E} directed along the yy-axis in the case of linear polarization; the radial distance represents the angle θ\theta between k\mathbf{k} and the photon propagation direction. All distributions are normalized to unity for better visibility.

The total number of photons emitted by the whole ensemble of atoms in the focal volume in the selected frequency interval can be estimated as [25] Nγ≃102−104N_\gamma\simeq10^{2}-10^{4}. Although the brightness is not high, such photon flux can be experimentally detected and resolved in angles1.

The photon diagnostics provides a probe, which is complementary to that based on the measurement of photoelectrons. In particular, both the photoelectron and the photon angular distributions appear highly sensitive to the presence of the longitudinal field component. For purely transverse fields, the width of the photon angular distribution can be used as measure of the peak intensity in line with the values γmax⁡\gamma_{\max} and θmin⁡\theta_{\min} extracted from angular-resolved spectra of photoelectrons.

Conclusions and outlook

Summarizing, we have presented results of numerical calculations of photoelectron and X-ray photons energy-angular distributions for the case, when the photoelectrons are produced in the central part of a focused high-power laser beam with parameters close to the currently achieved at several multi-PW laser facilities, including Apollon [5], whose beam energy and pulse duration have been chosen as the core parameters of the calculation. The distributions are shown to be essentially sensitive both to the peak laser intensity and to the field configuration in the focus.

Our results show that the presence of longitudinal field components should be considered as the main obstacle for the reliable intensity diagnostics. At the same time, both the photon and the electron angular distributions are shown to be efficient tools for the estimation of the longitudinal field relative amplitude.

The values of γmax⁡\gamma_{\max} and θmin⁡\theta_{\min} appear good measures of the peak intensity, when there are levels with ionization potentials, which correspond efficient ionization at such intensities. This statement holds, although with lower quantitative accuracy, even in the presence of longitudinal fields This returns us to the problem of the intensity-dependent composition of the gas target for ionization-based diagnostics of extreme laser fields [15, 16].

A correlative analysis of momentum and angular distributions of photons and photoelectrons by means of neural networks may be helpful to make considerable further advances of this problem.

Acknowledgment

AAM and SVP acknowledge fruitful discussions with R. Taïeb and S. Turban. NVM and SVP acknowledge support of the Russian Science Foundation through Grant No.25-22-00308.

References

  1. [1]Yoon J W, Kim Y G, Choi I W, Sung J H, Lee H W, Lee S K and Nam C H 2021 Optica 8 630–635 URL https://opg.optica.org/optica/abstract.cfm?URI=optica-8-5-630
  2. [2]ELI Beamlines https://www.eli-beams.eu/
  3. [3]Radier C, Chalus O, Charbonneau M, Thambirajah S, Deschamps G, David S, Barbe J, Etter E, Matras G, Ricaud S and et al 2022 High Power Laser Science and Engineering 10 e21
  4. [4]Yoon J W, Kim Y G, Choi I W, Sung J H, Lee H W, Lee S K and Nam C H 2021 Optica 8 630–635 URL https://opg.optica.org/optica/abstract.cfm?URI=optica-8-5-630
  5. [5]Yao W, Lelièvre R, Cohen I, Waltenspiel T, Allaoua A, Antici P, Ayoul Y, Beck A, Beluze A, Blancard C et al. 2025 Physics of Plasmas 32 043106
  6. [6]Kostyukov I Y, Khazanov E, Shaikin A, Litvak A and Sergeev A 2023 Bulletin of the Lebedev Physics Institute 50 S635–S640
  7. [7]Peng Y, Xu Y, Yu L, Wang X, Li Y, Lu X, Wang C, Liu J, Zhao C, Liu Y et al. 2021 The Review of Laser Engineering 49 93
  8. [8]Blackburn T 2020 Reviews of Modern Plasma Physics 4 5
  9. [9]Gonoskov A, Blackburn T, Marklund M and Bulanov S 2022 Reviews of Modern Physics 94 045001
  10. [10]Fedotov A, Ilderton A, Karbstein F, King B, Seipt D, Taya H and Torgrimsson G 2023 Physics Reports 1010 1–138
  11. [11]Popruzhenko S V and Fedotov A M 2023 Phys. Usp. 66 460–493 URL https://ufn.ru/en/articles/2023/5/b/
  12. [12]Chowdhury E A, Barty C A and Walker B C 2001 Physical Review A 63 042712
  13. [13]Yamakawa K, Akahane Y, Fukuda Y, Aoyama M, Inoue N and Ueda H 2003 Physical Review A 68 065403
  14. [14]Link A, Chowdhury E A, Morrison J T, Ovchinnikov V M, Offermann D, Van Woerkom L, Freeman R R, Pasley J, Shipton E, Beg F et al. 2006 Review of Scientific Instruments 77
  15. [15]Ciappina M, Popruzhenko S, Bulanov S, Ditmire T, Korn G and Weber S 2019 Physical Review A 99 043405
  16. [16]Ciappina M, Peganov E and Popruzhenko S 2020 Matter and Radiation at Extremes 5
  17. [17]Mironov A, Gelfer E, Tupitsyn I, Beck A, Jirka M, Klimo O, Meuren S, Oberreit G, Smorodnikova T, Taïeb R et al. 2025 Physical Review E 112 055202
  18. [18]Vais O and Bychenkov V Y 2018 Applied Physics B 124 211
  19. [19]DiChiara A, Ghebreghziahber I, Sauer R, Waesche J, Palaniyappan S, Wen B and Walker B 2008 Physical review letters 101 173002
  20. [20]Yandow A, Toncian T and Ditmire T 2019 Physical Review A 100 053406
  21. [21]Yandow A, Ha T, Aniculaesei C, Smith H, Richmond C, Spinks M, Quevedo H, Bruce S, Darilek M, Chang C et al. 2024 Physical Review A 109 023119
  22. [22]Har-Shemesh O and Di Piazza A 2012 Optics Letters 37 1352–1354
  23. [23]He C, Longman A, Pérez-Hernández J, De Marco M, Salgado C, Zeraouli G, Gatti G, Roso L, Fedosejevs R and Hill III W 2019 Optics Express 27 30020–30030
  24. [24]Mackenroth F, Holkundkar A R and Schlenvoigt H P 2019 New Journal of Physics 21 123028
  25. [25]Makarenko N and Popruzhenko S 2026 Physical Review A 113 043120
  26. [26]Narozhny N and Fofanov M 2000 Journal of Experimental and Theoretical Physics 90 753–768
  27. [27]Derouillat J, Beck A, Pérez F, Vinci T, Chiaramello M, Grassi A, Flé M, Bouchard G, Plotnikov I, Aunai N, Dargent J, Riconda C and Grech M 2018 Computer Physics Communications 222 351–373 ISSN 0010-4655 URL https://www.sciencedirect.com/science/article/pii/S0010465517303314
  28. [28]Thiele I, Skupin S and Nuter R 2016 Journal of Computational Physics 321 1110–1119
  29. [29]Smirnov B and Chibisov M 1966 Sov. Phys. JETP 22 23
  30. [30]Perelomov A, Popov V and Terent’ev M 1966 Zh. Eksp. Teor. Fiz 50 1393–1410
  31. [31]Perelomov A, Popov V and Terent’ev M 1967 Zh. Eksp. Teor. Fiz 52 514
  32. [32]Popov V S 2004 Physics-Uspekhi 47 855
  33. [33]Popruzhenko S 2014 Journal of Physics B: Atomic, Molecular and Optical Physics 47 204001
  34. [34]Keldysh L 1965 Sov. Phys. JETP 20
  35. [35]Popruzhenko S 2023 Bulletin of the Lebedev Physics Institute 50 S928–S932
  36. [36]Baranov D, Khudov A and Popruzhenko S 2026 Physical Review A 113 033108
  37. [37]Perelomov A, Popov V and Terent’ev M 1967 Sov. Phys. JETP 24 207–217
  38. [38]Perelomov A, Popov V and Terent’ev M 1966 Sov. Phys. JETP 23 924–934
  39. [39]Tyurin D, Strelkov V and Popruzhenko S 2025 arXiv preprint arXiv:2504.20583arxiv.org/abs/2504.20583
  40. [40]Tong X and Lin C 2005 Journal of Physics B: Atomic, Molecular and Optical Physics 38 2593
  41. [41]Kostyukov I Y and Golovanov A 2018 Physical Review A 98 043407
  42. [42]Popruzhenko S V and Kalymbetov E 2021 Quantum Electronics 51 801
  43. [43]Jackson J D 2021 Classical electrodynamics (John Wiley & Sons)
  44. [44]Landau L and Lifshitz E 1967 Theoretical physics. Field theory (Nauka)

Paper details

vlnc:W_6jjwyyes
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents