Next Article in Journal
Photoionization and Electron-Ion Recombination of n = 1 to Very High n-Values of Hydrogenic Ions
Previous Article in Journal
Magnetic Angle Changer for Studies of Electronically Excited Long-Living Atomic States
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Calculation of the Lowest Resonant States of H and Li by the Complex Absorbing Potential Method

1
Semenov Federal Research Center for Chemical Physics, Russian Academy of Sciences, 119991 Moscow, Russia
2
Department of Biology and Chemistry, Moscow State Pedagogical University, 119435 Moscow, Russia
3
Research Center for Environmental Changes, Academia Sinica, Taipei 115, Taiwan
4
National Research Center, Kurchatov Institute, 123182 Moscow, Russia
*
Author to whom correspondence should be addressed.
Submission received: 1 September 2021 / Revised: 21 September 2021 / Accepted: 27 September 2021 / Published: 30 September 2021

Abstract

:
The analysis of the features of the method of complex absorbing potential (CAP) is carried out for a single-channel problem with an explicit parameterization of the scattering matrix. It is shown that there can be several types of CAP trajectories depending on the choice of the initial conditions. In any case, the estimation of the resonance parameters from the position of the optimal trajectory point can lead to a systematic error or an ambiguous result. In special cases, the search for the optimal point can be replaced by the averaging over a closed section of the trajectory. The CAP trajectories constructed in the H and Li resonance calculations correlate well with the model trajectories, which have a curl around the resonance. The averaging over a closed area of the trajectory leads to better estimates of the energy and width of the resonance in comparison with the technique of searching for the optimal point.

1. Introduction

The possibility of calculating the energy and width of the resonance state by introducing an imaginary potential into the Hamiltonian was first shown more than thirty years ago [1,2,3,4,5]. The procedure proposes constructing the dependence of the complex eigenvalue of the stationary Schrödinger Equation by varying the scaling factor with the subsequent estimation of the resonance position as the geometric center of the closed interval of the obtained dependence (trajectory). This technique was substantiated because using a finite basis set leads to the appearance of a trajectory section closed around the resonance, which should converge to the exact position of the resonance in a complete basis set [5].
Later, an attempt was made to explain the introduction of the absorbing potential into the Hamiltonian by compensating for the incompleteness of the basis in the complex coordinate method (CCM). As in the CCM, in this case, the resonance position was estimated from the “optimal” point of the trajectory, for which the derivative of the energy concerning the scaling factor is equal to zero [6].
The key point in the application of the absorbing potential became the work [7,8], in which a theoretical foundation was given, and the results of calculations of the metastable 2Πg state of the anion of the N2 molecule were presented. Demonstration of the successful application of the method called the “complex absorbing potential method” (CAP) and the relative simplicity of its implementation in comparison with the CCM [9] were the reasons for its application to the calculation of resonances in many electron systems [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47].
The CAP method’s development was aimed at multiconfiguration quantum chemical methods that provide the correct account of electron correlation effects. At present, computational procedures have been tested using the configuration interaction (CI) formalism [7,8,11,15,17,18,21,22], various variants of cluster expansion (EOM-CC, SAC-CI, FSMRCC) [26,27,29,32,33,34,35,36,38,39,40], algebraic diagrammatic approach [24,47] and quasi-degenerate perturbation theory (XMCQDPT2) [41]. The accuracy of the proposed approaches can be compared only for the lowest resonances N 2 (2Πg) and CO (2Π), although simple systems ( H , He , Li, etc.) could be used as test systems, for which the positions of the resonances are known with sufficient accuracy [48,49,50,51,52,53,54].
For the lowest resonance CO (2Π) at an equilibrium internuclear distance, the energy estimates are in the range 1.90–2.08 eV [31,32,33,43,47,55] and exceed the experimental value of 1.50 eV [56]. The resonance widths are estimated in the range 0.38–0.63 eV [31,32,33,43,47,55] at the experimental value of 0.4 eV. The maximum theoretical value of the resonance width (1.3 eV) was obtained in [43]. The sensitivity can explain the error of theoretical results to a quality of the basis of atomic orbitals (AO) and parameters of the absorbing potential [32,33]. For the lowest resonance N 2 (2Πg), the smallest deviations from the reference values of energy (2.316 eV) and width (0.414 eV) [57] are less than 0.1 eV [27,49] and 0.12–0.2 eV [24,26,27,32,43,47,55], respectively. In other calculations, deviations in the energy and width of the resonance have larger values [7,11,18,24,26,30,32,33,41,43,55]. The calculation accuracy depends on the same factors as in CO , i.e., the quality of the AO basis and the parameters of the absorbing potential [18,30,32].
Simultaneously with the search for methods for solving the electronic problem, attempts to improve the accuracy of calculations by changing the methods of building trajectories or estimating the position of the resonance were made. It has been shown that replacing imaginary absorbing potential with a complex one makes it possible to eliminate nonphysical optimal points of the CAP trajectory [28]. Moreover, a new technique for estimating the position of the resonance was proposed, based on the construction of Padé approximants [55,58]. The application of this technique for N 2 (2Πg) and CO (2Π) resonances has improved the accuracy of the calculation [55].
A feature of the implementation of the CAP is the use of a basis of atomic orbitals (AO), square-integrable over the entire three-dimensional space. This method of constructing the AO basis preserves the continuity of the wave function at the boundary of the overlapping region of the absorbing potential but does not allow to separate the errors arising from the electron correlation, as well as the errors of the CAP parameterization. This feature can be removed by stitching the solutions at the boundary of the overlapping region of the absorbing potential using the R-matrix formalism [59,60]. Similar approaches were previously proposed to calculate the parameters of scattering resonances by the complex coordinate method [61,62] and solving problems of reaction scattering [25,63,64,65]. The combination of the R-matrix with the CAP method was also used in the construction of the theory of the CAP [7].
The approach proposed in [7] makes it possible to explicitly parametrize the R-matrix and, therefore, to obtain the CAP-trajectories free from errors of the approximate solution of the Schrödinger Equation in the inner region. By comparing the behavior of the trajectories obtained for an explicitly parameterized R-matrix and real systems, it is possible to formulate the requirements necessary for the successful application of the CAP [46].
This work aims to assess the possibilities of the combined approach [7]. For comparison, we chose the lowest 1S resonance of the H anion and 2P0 resonances of the lithium atom. The theoretical formalism is briefly presented in the second part of the article. The calculation details are given in the third part, followed by a discussion of the main results and a conclusion.

2. Theory

Let there be a radial equation
[ 2 2 m 2 r 2 + V ( r ) E ] ψ ( E , r ) = 0 ,
where is the cyclic Planck constant, m is the particle mass, r 0 is an independent variable, V ( r ) is the potential, E is the eigenvalue, and ψ ( E , r ) is the solution satisfying the initial condition ψ ( E , r = 0 ) = 0 . Here we will assume that for the function ψ ( E , r ) and its derivative, the continuity conditions are satisfied at the point r = R , i.e.,
ψ int ( E , r ) | r = R 0 = ψ e x t ( E , r ) | r = R + 0 ,
ψ int ( E , r ) r | r = R 0 = ψ e x t ( E , r ) r | r = R + 0 .
The int and ext indexes here are introduced for the inner ( 0 r R ) and outer ( r R ) domains.
Let us assume that there are functions { ϕ n ( r ) } in the outer domain and a system { ϕ ˜ n ( r ) } biorthogonally conjugated to it. Then the elements of these systems satisfy the condition
R ϕ ˜ n ( r ) ϕ m ( r ) d r = δ n m ,
where δ n m is the Kronecker symbol and the function ψ e x t ( E , r ) can be represented as series
ψ e x t ( E , r ) = n A n ϕ n = n A ˜ n ϕ ˜ n
with the coefficients A n = R ψ e x t ( E , r ) ϕ ˜ n d r and A ˜ n = R ψ e x t ( E , r ) ϕ n d r .
Multiplying Equation (1) by the functions ϕ ˜ n ( r ) (n = 1, 2, …) on the left and integrating over the range r [ R , ) , we obtain the equation
R ϕ ˜ n ( r ) [ 2 2 m 2 r 2 + V ( r ) E ] ψ e x t ( E , r ) d r = 2 2 m W ( ϕ ˜ n ( r ) , ψ e x t ( E , r ) ) | r = R + R ψ e x t ( E , r ) [ 2 2 m 2 r 2 + V ( r ) E ] ϕ ˜ n ( r ) d r = 0 ,
where W ( f , g ) = f g r g f r is the Wronskian, resulting from integration by parts of the term containing the second derivative. This term can also be represented in the integral form
2 2 m W ( ϕ ˜ n ( r ) , ψ e x t ( E , r ) ) | r = R = R ψ e x t ( E , r ) L ^ b ϕ ˜ n ( r ) d r ,
provided if the Bloch operator [59,60]
L ^ b = 2 2 m δ ( r R ) [ r b ]
is introduced, where δ ( r R ) is the Dirac delta function, and b = ( ln ψ e x t ( E , r ) / r ) | r = R .
Expand the function ψ e x t ( E , r ) by functions ϕ n ( r ) , and substitute the result into Equation (6). Then let us combine the coefficients A n , functions ϕ ˜ n ( r ) , and their derivatives in the obtained equations into the row vectors A = | A 1 A n | , F ˜ = | ϕ ˜ 1 ϕ ˜ n | and F ˜ / = | ϕ ˜ 1 / r ϕ ˜ n / r | whereas the integrals into the matrix H e x t with elements H n m e x t = R ϕ n ( r ) [ 2 2 m 2 r 2 + V ( r ) ] ϕ ˜ m ( r ) d r , which leads to the matrix equation
2 2 m ( ψ e x t ( E , r ) r F ˜ ψ e x t ( E , r ) F ˜ / ) | r = R + A ( H e x t I E ) = 0 .
Multiplying (9) on the right by the column vector G ( E ) F T , where F = | ϕ 1 ϕ n | and G ( E ) = ( H e x t I E ) 1 , we get the equation
2 2 m ( ψ e x t ( E , r ) r F ˜ G ( E ) F T ψ e x t ( E , r ) F ˜ / G ( E ) F T ) | r = R = A F T | r = R = ψ e x t ( E , r ) | r = R ,
which after reducing similar terms transforms to the form
2 2 m F ˜ G ( E ) F T | r = R R ( E ) ( 2 2 m F ˜ / G ( E ) F T 1 ) | r = R = 0 ,
where R ( E ) | r = R = b 1 is the element of R-matrix.
Thus, solving Equation (11) with the given boundary condition b 1 = R ( E )   |   r = R , it is possible to find a value E that satisfies the radial Equation (1). Since the eigenfunction in Equation (1) is assumed to be continuous at the point r = R , the solution of Equation (11) for the given boundary condition can be considered as one of the options for combining the R-matrix with the approximate solution (1) in the basis of functions that are square-integrated over the range r R . The possibility of using the simplified form (11) to combine R-matrix and the complex scaling method was discussed earlier, but the practical side of the issue was reduced to solving the generalized eigenvalue problem [61,62]. Case (11) was also used to solve the model problem by CAP [7].
Let us consider the application of Equation (11) to search the resonant pole by the CAP [7]. We will assume that in the region r R , there is an absorbing potential
V ( r ) = s 2 ( r R ) 2
with a complex scaling factor s ( Im ( s ) < 0 ) [7,25,55]. To expand ψ e x t ( E , r ) , let us choose the Hermite functions
ϕ ˜ n ( x ) = C n H n ( x ) exp ( x 2 / 2 ) ,
where H n ( x ) is the n-th degree Hermite polynomial x = β ( r R ) , β is a parameter and C n = ( β π 2 n n ! ) 1 / 2 is a normalization factor. Functions (13) are eigenfunctions for the harmonic oscillator problem with mass m, potential V ϕ ( r ) = λ 2 ( r R ) 2 , β = m λ ( β 2 = m λ / 2 ), and eigenvalues E n = 2 β m ( n + 1 / 2 ) = λ m ( n + 1 / 2 ) , where n = 0 , 1 , 2 , .
In an interval x ( , + ) , functions ϕ ˜ n ( x ) form a complete and closed orthonormal system for β > 0 [66,67]. Using the orthonormality of the functions ϕ ˜ n ( x ) and representing the integral over the domain x ( , + ) as the sum of the integrals
+ ϕ ˜ n ( x ) ϕ ˜ m ( x ) d x = 0 + ϕ ˜ n ( x ) ϕ ˜ m ( x ) d x + 0 + ϕ ˜ n ( x ) ϕ ˜ m ( x ) d x = δ n m ,
it can be noted that for an even integrand at x [ 0 , + ) the condition
0 + ϕ ˜ n ( x )     ϕ ˜ m ( x )   d x = δ n m / 2
is satisfied, i.e., orthogonality is preserved. For an odd integrand, expressing the Hermite polynomials in terms of the generalized Laguerre polynomials H 2 m ( x ) = ( 1 ) m 2 2 m m ! L m 1 / 2 ( x 2 ) and H 2 m + 1 ( x ) = ( 1 ) m 2 2 m + 1 m ! x L m 1 / 2 ( x 2 ) , where m = 0 , 1 , 2 , [68], we have
0 + ϕ ˜ n ( x ) ϕ ˜ m ( x )   d x 0 .
Thus, it follows from expressions (15) and (16) that for real r and β > 0 Hermite functions of even and odd degrees form two mutually nonorthogonal systems of orthonormal functions in the range x [ 0 , + ) .
When passing to complex values of the parameter β properties (15) and (16) are preserved, which allows defining the biorthogonal conjugate functions by the condition ϕ ˜ n ( x ) = ϕ n ( x ) without restrictions on the values β. It should be noted that the completeness of systems of Hermite functions of even (or odd) degrees on the entire complex plane of x values has not been proven, but it has been shown that these systems are closed but not complete [69,70].
Equation (11) assumes that ϕ n ( R ) 0 and ϕ ˜ n ( R ) 0 (or F | r = R 0 and F ˜ | r = R 0 ). Hence it follows that for the expansion of ψ e x t ( E , r ) that only Hermite functions of even degrees can be used. Then ϕ n ( x ) / r | r = R = 0 , and consequently F / | r = R = F ˜ / | r = R = 0 , which leads to the equation
R ( E ) |   r = R = 2 2 m F G ( E ) F T |   r = R ,
particular cases of which were introduced in [7,60,62].
Building the trajectory E ( s ) using the CAP can be carried out in two ways: fix the value of the parameter λ and solve Equation (17) for all required values of the parameter s, or postulate s = λ at each point of the trajectory. The first method is used in most studies of many-electron systems since it does not require multiple calculations of one- and two-electron integrals; the second one was introduced if considering the model problem in work [7].
Substituting the Hermite functions into the expressions for the elements of the matrix H e x t , we obtain
H n m e x t = E m δ n m + s λ 2 V n m ,
where V n m = K ϕ n ( r ) ( r R ) 2 ϕ m ( r ) d r . Using the recursion relation, H n + 1 ( x ) 2 x H n ( x ) + 2 n H n 1 ( x ) = 0 , it can be shown that H e x t is a tridiagonal matrix with nonzero entries V n n = ( 2 n + 1 / 2 ) / β , V n n + 1 = [ ( n + 1 / 2 ) ( n + 1 ) ] 1 / 2 / β and V n n 1 = [ n ( n 1 / 2 ) ] 1 / 2 / β . Suppose the value of the parameter λ is fixed. In that case, after the substitution of matrix elements (18) and the corresponding values of the Hermite functions into it, the expression (17) takes on its final form.
If s = λ the matrix H e x t becomes diagonal ( H n m e x t = E m δ n m ), which allows replacing the right-hand side of (17) with the sum
R ( E ) |   r = R = 2 2 m n ( ϕ n ( r ) ) 2 E n E |   r = R ,
which includes terms with n = 0 , 2 , 4 , .
Substituting the values of E n and ϕ n ( r ) in (19), taking into account the change in the normalization ( C n C n 2 ), and performing the necessary transformations, we obtain
R ( E ) |   r = R = 1 2 π β l = 0 Γ ( l + 1 / 2 ) l ! ( l + 1 / 4 z ) ,
where Γ ( x ) is a gamma function, z = E 2 m λ = k 2 4 β . k = 2 m E / 2 is a wave number, and l = 0 , 1 , 2 ,   . If the summation in Formula (20) is carried out over an infinite number of terms, then their sum can be represented in a more convenient form
l = 0 Γ ( l + 1 / 2 ) l ! ( l + 1 / 4 z ) = π Γ ( 1 / 4 z ) Γ ( 3 / 4 z ) ,
where z 5 / 4 + n ( n = 0 , 1 , 2 , ). Details of transformation (21) are given in [68,71] (see the Appendix A). The substitution of (21) into (20) followed by multiplication by i k gives the final expression
i k R ( E ) | r = R = i z Γ ( 1 / 4 z ) Γ ( 3 / 4 z ) .

3. Calculation Method

In the numerical solution of Equations (17), (20) and (22), the function ψ int ( E , r ) at r R was taken in the form
ψ int ( E , r ) ~ χ ( ) S ( k ) χ ( + ) ,
where χ ( ± ) = e ± i θ , θ = k r z 0 k ln ( 2 k r ) + arg Γ ( 1 + i z 0 k ) , z 0 is the target charge, k is the wavenumber, and S ( k ) is the scattering matrix. For a system with isolated Breit–Wigner resonances, the scattering matrix can be represented as the sum
S ( k ) = S 0 ( k ) n i T n E E n ,
where S 0 ( k ) = e i 2 δ 0 is the nonresonant component of the scattering matrix, δ0 is the background part of the phase shift, E n = E 0 n i Γ n / 2 is the position of the n-th pole, and T n are some complex numbers [72]. Using the unitarity of the scattering matrix at real values of the energy, (24) can be replaced by the expression
S ( k ) = S 0 ( k ) n E E n * E E n .
Parametrization of the scattering matrix (25) assumed a single resonance with E 0 = 3.42639031 0.00638724 i a.u., which corresponds to the energy of the lowest resonance in the potential [73], and the nonresonant component of the phase shift δ 0 = π / 4 . The particle mass was set equal to m = 1 a.u., R = 20 a.u. Equations (17), (20) and (22) were solved iteratively with an energy convergence threshold of 10−10 a.u. The method [74] was used to calculate the gamma functions of a complex argument. The inversion of the tridiagonal matrix in (17) was carried out by the method proposed in [75].
The CAP method was used to calculate the two lowest 2P0 [1s(2s2p) 3P] and 2P0 [1s(2s2p) 1P] resonances of the lithium atom and 1S resonance of H. For the lowest 2P0 [1s(2s2p) 3P] resonance, there are experimental estimates of energy and width [51,52,53,76], for the next 2P0 [1s(2s2p) 1P] resonance where only the energy is known [77,78]. The experimental results are in satisfactory agreement with the theoretical ones [54,79,80,81,82].
In the calculations of the Li atom, we used a technique previously tested on anions H and He [46,83]. Accordingly, the wave function was represented as a linear combination of three-electron configuration state functions (CSF), including the correlation and asymptotic components. Both components included complete sets of CSFs built from orthonormal AOs of the form
χ n l m ( r ) = N n l ( λ r ) l e λ r / 2 L n 2 l + 2 ( λ r ) Y l m ( r ^ ) ,
where λ is the scaling factor ( λ > 0 ) and N n l is the normalizing factor [83,84,85,86]. The AO basis for the correlation component included the functions χ n l m ( r ) with l 3 , 0 n 4 l , and | m | l , whereas the asymptotic one χ n l m ( r ) with l 1 , 5 l n 26 l , and | m | l (after this referred to as the basis of AO 5s4p3d2f/22s22p). The calculations used the absorbing potential (12) with an imaginary scaling factor ( 8.9 ln ( Im ( s ) ) 8.0 ).
To select the point of superposition of the absorbing potential calibration, calculations were performed with λ = 3.5 , corresponding to the minimum energy of the first excited state of the target (the closed channel with the lowest energy). Analysis of CAP trajectories obtained for 10 R 17 a.u. revealed that for the chosen basis, AO closed sections appear on trajectories at 14 R 16 a.u. Subsequent calculations were carried out for R = 15 a.u. at 3.1 λ 3.7 . The resonance position was calculated using two methods: averaging over a closed section of the trajectory (A) and searching for the optimal point (minimum of the function | s E ( s ) / s | ) (B) [7,8]. The minimum of the function determined the position of the optimal point
ν ( η ) = | η d E ( η ) d η | ,
where η = Im ( s ) . For an independent assessment of energy and width of 2P0 resonances, we used the method of stabilization by a modified Coulomb potential [83,84,85,86,87]. CAP-trajectories for the lowest 1S resonance of H were calculated using an 9s8p7d/26s26p26d AO basis set. The details of calculations can be found elsewhere [46].

4. Results and Discussion

Regarding Hermite functions of even degrees, the expansion of Equation (17) solutions leads to three model problems. The first of them ψ e x t ( E , r ) is a solution to (22) and is expanded in a series in an infinite number of scalable ( λ = s ) functions (I). The second is reduced to solving Equation (20) and uses an expansion of ψ e x t ( E , r ) in a series in a finite number of scalable functions (II). The latter corresponds to the solution (17) with an expansion in a finite number of functions with a constant scaling factor λ ( λ s ) (III).
For positive values of the parameter s, the behavior of the branches of the function E ( s ) depends on the definition of the parameter λ. In cases I and II, the branches E ( s ) converge to the point E ( s 0 ) 0 and finite values at s . In case III at s 0 the branches E ( s ) have nonzero limits and at s converge to the same finite values as in cases I–II. At the parameter value s 0.02 0.025 , the branches E ( s ) , which have been built for the problems I-III, practically coincide (Figure 1).
Since the E ( s ) is a multivalued function at s > 0 , it follows that for the absorbing potential with a complex scaling factor s = ζ i η ( ζ , η > 0 ), with the appropriate values of ζ = Re ( s ) , it is possible to construct several trajectories converging to the same real value E ( s ) at η 0 . It was found that in the case of problem I for an uncharged target ( z 0 = 0 ), in the range of values ζ = Re ( s ) < 1 there are three types of trajectories passing in the vicinity of the resonance (Figure 2). Trajectories of the first type ( ζ < 0.1 ) have a quasilinear section near the resonance pole, whereas the second type ( 0.3 < ζ < 1 ) bypasses the resonance, forming a section that is closed around it. The third type of trajectory ( 0.1 < ζ < 0.3 ) has several closed sections near the pole. Note that the reasons for the appearance of quasilinear trajectory segments in the vicinity of the resonance were discussed earlier, and the corresponding explanations remain valid for the model problem I [55,58].
Trajectories of all types have optimal points (27), which can be considered as estimates of the resonance position. However, in the case with ζ < 0.3 the resonance position cannot be estimated unambiguously since the trajectories have more than one optimal point. At ζ > 0.3 the resonance position can be estimated uniquely, but the optimal point is away from the pole, leading to a significant systematic error (Table 1).
When passing to expansions of finite length (problem II), the shape of the trajectories changes, but the number of optimal points remains the same (Figure 3, Figure 4 and Figure 5). Simultaneously, the accuracy of the estimation of the resonance position is significantly reduced. (Table 2).
For problem III it was found that in the case Re ( s ) = 0 , there are only trajectories of the first type (Figure 3, Figure 4 and Figure 5), while for Re ( s ) > 0 , the type of trajectory depends on the chosen value Re ( s ) like for the problem I. The last conclusion for problem II is also valid for problem III (Table 3). Since the solution of problems I–III for a charged target ( z 0 = 1 ) gives practically the same results as the above case, we give only a representative part of the results (Table 1, Table 2 and Table 3). It should be added that the systematic shift of the resonance energy by ~10−5 a.u. for a charged target (Table 1) is the result of the finite-difference calculation of the minimum of function (27). Thus, two methods of calculating the resonance parameters can be used for the considered model problems. First, one can use the absorbing potential with an imaginary scaling factor and choose one of the optimal points, which will require justifying the possibility of introducing a selection criterion. Second, the form of trajectories can be simplified by using a complex scaling factor for absorbing potential, like in the technique [28]. In this case, the resonance position can be estimated by averaging over a fragment of the trajectory [5]. However, both results will contain a systematic error, the value of which depends on the length of the solution expansion.
In the calculations of H and Li resonances, it was found that CAP trajectories have a curl that is closed around the reference resonance position (Figure 6 and Figure 7). For the H anion, the energy and width of the lowest 1 S 0 resonance, calculated by averaging over the closed section of the trajectory (A), are in better agreement with the reference values than the energy and width of the resonance estimated from the position of the optimal point (27). In addition, there is a dependence of the energy and resonance width on the choice of the value of the AO scaling factor (26) (Table 4), which is consistent with the results of calculations by the CAP and stabilization methods [46,84,86].
For 2P0 resonances of Li, methods A and B provide energy estimates that are close in accuracy. Both methods predict the value of the width of the lowest resonance, which is in poor agreement with the experimental value, but estimates of the width by method (A) are consistent with the independent theoretical results (Table 5). Estimates of the second resonance width by method A agree better with theoretical values than estimates by method B. Calculations of 2P0 resonances by the CAP and stabilization methods give close values of energies and widths for both resonances (Table 6). Comparison of the results of H and Li calculations by the CAP and stabilization methods allows us to conclude that the accuracy of the ab initio calculation is primarily determined by choice of the AO basis and the method for the approximate solution of the stationary equation.

5. Conclusions

In this paper, we analyzed the features of the CAP method for a single-channel problem with an explicit parameterization of the scattering matrix. It was found that there are three types of trajectories, the of which is determined by the choice of a branch of the stabilization graph. The first type of trajectory is characterized by a quasilinear segment close to the resonance pole, and a closed segment around the resonance characterizes the second type of trajectory. There are also trajectories of an intermediate type, with several closed sections near the resonant pole. When passing to a model with a finite number of terms in the expansion of the solution of the stationary equation, the number of optimal points is preserved for the trajectories, but it becomes difficult to classify the type of trajectory.
On the contrary, CAP trajectories for H and Li correlate well with model trajectories of the second type. Accordingly, the estimates of the resonance parameters by averaging over the closed section of the trajectory are in better agreement with the results of independent calculations.

Author Contributions

Conceptualization, S.O.A., D.D.K. and G.V.G.; methodology, S.O.A., D.D.K. and Y.A.D.; software, S.O.A. and D.D.K.; validation, S.O.A., Y.A.D. and G.V.G.; formal analysis, S.O.A. and D.D.K.; investigation, S.O.A., D.D.K. and G.V.G.; resources, S.O.A. and M.G.G.; data curation, S.O.A. and D.D.K.; writing—original draft preparation, S.O.A. and D.D.K.; writing—review and editing, S.O.A., D.D.K., Y.A.D. and M.G.G.; visualization, S.O.A. and D.D.K.; supervision, G.V.G.; project administration, M.G.G.; funding acquisition, M.G.G. All authors have read and agreed to the published version of the manuscript.

Funding

This work was carried out in the framework of State Assignment of the Ministry of Science and Higher Education of the Russian Federation (project No. AAAA–A19–119010990034–5).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Based on the results from the book [68], it is possible to propose several methods for transforming the series
f ( a , b ) = n = 0 Γ ( n + a ) n ! ( n + b ) ,
where a and b are complex numbers. First, we can use the table integrals (7.511) and (7.512.10) and replace the series (A1) with the Gaussian hypergeometric function, and then replace the latter with the ratio of the gamma functions [71]. Secondly, the same result can be obtained using formulas for the sums of the series obtained using the Mellin transformation (§1.4 and §2.4 [68]). However, implicit restrictions on the values of parameters a and b are assumed in both cases, which requires additional study. Alternatively, it is convenient to use the method from §2.1.3 of the book [68]. Replacing the expression b + n = Γ ( b + n + 1 ) / Γ ( b + n ) and entering the Pochhammer symbols ( a ) n = Γ ( a + n ) / Γ ( a ) , series (A1) can be represented as
f ( a , b ) = Γ ( a ) Γ ( b ) Γ ( b + 1 ) n = 0 ( a ) n ( b ) n n ! ( b + 1 ) n = Γ ( a ) Γ ( b ) Γ ( b + 1 ) F ( a , b , b + 1 ; 1 ) ,
where F ( α , β , γ ; x ) is the Gaussian hypergeometric function. If the conditions Re ( γ ) > Re ( β ) > 0 and Re ( 1 α ) > 0 are met, it becomes possible to use the Gauss Formula and replace the function F ( α , β , γ ; 1 ) with the expression
F ( α , β , γ ; 1 ) = Γ ( γ ) Γ ( β ) Γ ( γ β ) 0 1 t β 1 ( 1 t ) γ β α 1 d t = = Γ ( γ ) Γ ( β ) Γ ( γ β ) B ( β , γ β α ) = Γ ( γ ) Γ ( γ β α ) Γ ( γ β ) Γ ( γ α ) ,
where B ( x , y ) = Γ ( x ) Γ ( y ) / Γ ( x + y ) is a beta function. As shown above, the conditions Re ( γ ) > Re ( β ) > 0 and Re ( 1 α ) > 0 can be replaced with weaker ones Re ( γ α β ) > 0 and γ 0 , 1 , 2 , ([68], chapter 2.8).
Finally, substituting (A3) into (A2), we obtain the ratio
f ( a , b ) = Γ ( a ) Γ ( b ) Γ ( 1 a ) Γ ( 1 + b a ) ,
where the parameters a and b must satisfy the conditions Re ( 1 a ) > 0 and b + 1 0 , 1 , 2 , … Using (A4), it is easy to show that the finite sum of the first terms of the series (A1) can be replaced by the difference
n = 0 N Γ ( n + a ) n ! ( n + b ) = f ( a , b ) f ( a , b + N + 1 ) = Γ ( a ) Γ ( 1 a ) ( Γ ( b ) Γ ( 1 + b a ) Γ ( b + N + 1 ) Γ ( 2 + N + b a ) ) .
In a particular case a = 1 / 2 , b = z 1 / 4 the expression (A4) simplifies to the relation
f ( 1 2 , 1 4 z ) = π Γ ( 1 / 4 z ) Γ ( 3 / 4 z ) .
In this case, the condition Re ( 1 a ) = Re ( 1 / 2 ) > 0 is satisfied, but for (A6) remains the restriction z 5 / 4 + n ( n = 0 , 1 , 2 , ).

References

  1. Jolicard, G.; Austin, E.J. Optical potential stabilization method for predicting resonance levels. Chem. Phys. Lett. 1985, 121, 106–110. [Google Scholar] [CrossRef]
  2. Jolicard, G.; Austin, E.J. Optical potential method of calculating resonance energies and widths. Chem. Phys. 1986, 103, 295–302. [Google Scholar] [CrossRef]
  3. Jolicard, G.; Perrin, M.Y. A perturbational treatment of the resonance eigenvalue problem using the optical potential model. Chem. Phys. 1987, 116, 1–10. [Google Scholar] [CrossRef]
  4. Jolicard, G.; Humbert, J. Study of the one-channel resonance states. Method without a stabilization procedure in the framework of the optical potential model. Chem. Phys. 1987, 118, 397–405. [Google Scholar] [CrossRef]
  5. Jolicard, G.; Leforestier, C.; Austin, E.J. Resonance states using the optical potential model. Study of Feshbach resonances and broad shape resonances. J. Chem. Phys. 1988, 88, 1026–1031. [Google Scholar] [CrossRef]
  6. Rom, N.; Lipkin, N.; Moiseyev, N. Optical potentials by the complex coordinate method. Chem. Phys. 1991, 151, 199–204. [Google Scholar] [CrossRef]
  7. Riss, U.V.; Meyer, H.D. Calculation of resonance energies and widths using the complex absorbing potential method. J. Phys. B At. Mol. Opt. Phys. 1993, 26, 4503–4536. [Google Scholar] [CrossRef]
  8. Riss, U.V.; Meyer, H.D. Reflection-free complex absorbing potentials. J. Phys. B At. Mol. Opt. Phys. 1995, 28, 1475–1493. [Google Scholar] [CrossRef]
  9. Reinhardt, W.P. Complex coordinates in the theory of atomic and molecular structure and dynamics. Ann. Rev. Phys. Chem. 1982, 33, 223–255. [Google Scholar] [CrossRef]
  10. Sommerfeld, T.; Riss, U.V.; Meyer, H.D.; Cederbaum, L.S. Metastable C22− dianion. Phys. Rev. Lett. 1997, 79, 1237–1240. [Google Scholar] [CrossRef]
  11. Sommerfeld, T.; Riss, U.V.; Meyer, H.D.; Cederbaum, L.S.; Engels, B.; Suter, H.U. Temporary anions—Calculation of energy and lifetime by absorbing potentials: The N 2   2 Π g resonance. J. Phys. B At. Mol. Opt. Phys. 1998, 31, 4107–4122. [Google Scholar] [CrossRef]
  12. Sommerfeld, T.; Cederbaum, L.S. Long-lived states of N 2 . Phys. Rev. Lett. 1998, 80, 3723–3726. [Google Scholar] [CrossRef]
  13. Dreuw, A.; Sommerfeld, T.; Cederbaum, L.S. Possible long-lived quartet resonance states of CO. Theor. Chem. Acc. 1998, 100, 60–64. [Google Scholar] [CrossRef]
  14. Ingr, M.; Meyer, H.D.; Cederbaum, L.S. Potential energy curve of the X 2 Σ u + resonance state of F 2 computed by CAP/CI. J. Phys. B At. Mol. Opt. Phys. 1999, 32, L547–L556. [Google Scholar] [CrossRef]
  15. Santra, R.; Cederbaum, L.S.; Meyer, H.D. Electronic decay of molecular clusters: Non-stationary states computed by standard quantum chemistry methods. Chem. Phys. Lett. 1999, 303, 413–419. [Google Scholar] [CrossRef]
  16. Sahoo, S.; Ho, Y.K. The complex absorbing potential method (CAP) to study the Stark effect in hydrogen and lithium. J. Phys. B At. Mol. Opt. Phys. 2000, 33, 2195–2206. [Google Scholar] [CrossRef]
  17. Sommerfeld, T. FCN and FNC radical anions. Phys. Chem. Chem. Phys. 2001, 3, 2394–2399. [Google Scholar] [CrossRef]
  18. Sommerfeld, T.; Santra, R. Efficient method to perform CAP/CI calculations for temporary anions. Int. J. Quant. Chem. 2001, 82, 218–226. [Google Scholar] [CrossRef]
  19. Santra, R.; Cederbaum, L.S. An efficient combination of computational techniques for investigating electronic resonance states in molecules. J. Chem. Phys. 2001, 115, 6853–6861. [Google Scholar] [CrossRef]
  20. Santra, R.; Cederbaum, L.S. Non-Hermitian electronic theory and applications to clusters. Phys. Rep. 2002, 368, 1–117. [Google Scholar] [CrossRef]
  21. Sommerfeld, T. Coupling between dipole-bound and valence states: The nitromethane anion. Phys. Chem. Chem. Phys. 2002, 4, 2511–2516. [Google Scholar] [CrossRef]
  22. Sommerfeld, T.; Meyer, H.D. Computing the energy-dependent width of temporary anions L2 from ab initio methods. J. Phys. B At. Mol. Opt. Phys. 2002, 35, 1841–1863. [Google Scholar] [CrossRef]
  23. Dreuw, A.; Sommerfeld, T.; Cederbaum, L.S. Short- and long-lived electronic states of BF. J. Chem. Phys. 2002, 116, 6039–6044. [Google Scholar] [CrossRef]
  24. Feuerbacher, S.; Sommerfeld, T.; Santra, R.; Cederbaum, L.S. Complex absorbing potentials in the framework of electron propagator theory. II. Application to temporary anions. J. Chem. Phys. 2003, 118, 6188–6199. [Google Scholar] [CrossRef]
  25. Muga, J.G.; Palao, J.P.; Navarro, B.; Equsquiza, I.L. Complex absorbing potentials. Phys. Rep. 2004, 395, 357–426. [Google Scholar] [CrossRef]
  26. Sajeev, Y.; Santra, R.; Pal, S. Analytically continued Fock space multireference coupled-cluster theory: Application to the 2Πg shape resonance in e-N2 scattering. J. Chem. Phys. 2005, 122, 234320. [Google Scholar] [CrossRef]
  27. Sajeev, Y.; Santra, R.; Pal, S. Correlated complex independent particle potential for calculating electronic resonances. J. Chem. Phys. 2005, 123, 204110:1–204110:10. [Google Scholar] [CrossRef]
  28. Sajeev, Y.; Vysotskiy, V.; Cederbaum, L.S.; Moiseyev, N. Continuum remover-complex absorbing potential: Efficient removal of the nonphysical stabilization points. J. Chem. Phys. 2009, 131, 211102:1–211102:4. [Google Scholar] [CrossRef] [Green Version]
  29. Ehara, M.; Sommerfeld, T. CAP/SAC-CI method for calculating resonance states of metastable anions. Chem. Phys. Lett. 2012, 537, 107–112. [Google Scholar] [CrossRef]
  30. Ghosh, A.; Vaval, N.; Pal, S.; Bartlett, R.J. Complex absorbing potential based equation-of-motion coupled cluster method for the potential energy curve of CO2. J. Chem. Phys. 2014, 141, 164113:1–164113:6. [Google Scholar] [CrossRef]
  31. Ghosh, A.; Vaval, N.; Pal, S. Equation-of-motion coupled-cluster method for the study of shape resonance. J. Chem. Phys. 2012, 136, 234110:1–234110:6. [Google Scholar] [CrossRef]
  32. Jagau, T.C.; Zuev, D.; Bravaya, K.B.; Epifanovsky, E.; Krylov, A.I. A fresh look at resonances and complex absorbing potentials: Density matrix-based approach. J. Phys. Chem. Lett. 2014, 5, 310–315, J. Phys. Chem. Lett.2015, 6, 3866. [Google Scholar] [CrossRef]
  33. Zuev, D.; Jagau, T.C.; Bravaya, K.B.; Epifanovsky, E.; Shao, Y.; Sundstrom, E.; Head-Gordon, M.; Krylov, A.I. Complex absorbing potentials within EOM-CC family of methods: Theory, implementation, and benchmarks. J. Chem. Phys. 2014, 141, 024102:1–024102:19, Erratum: J Chem. Phys. 2015, 143, 149901:1. [Google Scholar] [CrossRef] [Green Version]
  34. Jagau, T.C.; Krylov, A.I. Complex absorbing potential equation-of-motion coupled-cluster method yields smooth and internally consistent potential energy surfaces and lifetimes for molecular resonances. J. Phys. Chem. Lett. 2014, 5, 3078–3085. [Google Scholar] [CrossRef]
  35. Jagau, T.C.; Dao, D.B.; Holtgrewe, N.S.; Krylov, A.I.; Mabbs, R. Same but different: Dipole-stabilized shape resonances in CuF and AgF. J. Phys. Chem. Lett. 2015, 6, 2786–2793. [Google Scholar] [CrossRef] [Green Version]
  36. Ehara, M.; Fukuda, R.; Sommerfeld, T. Projected CAP/SAC-CI method with smooth Voronoi potential for calculating resonance states. J. Comp. Chem. 2016, 37, 242–249. [Google Scholar] [CrossRef]
  37. Kunitsa, A.A.; Bravaya, K.B. First-principles calculations of the energy and width of the 2Au shape resonance in p-benzoquinone: A gateway state for electron transfer. J. Phys. Chem. Lett. 2015, 6, 1053–1058. [Google Scholar] [CrossRef]
  38. Kanazawa, Y.; Ehara, M.; Sommerfeld, T. Low-lying π* resonances of standard and rare DNA and RNA bases studied by the projected CAP/SAC–CI method. J. Phys. Chem. A 2016, 120, 1545–1553. [Google Scholar] [CrossRef] [Green Version]
  39. Jagau, T.C.; Krylov, A.I. Characterizing metastable states beyond energies and lifetimes: Dyson orbitals and transition dipole moments. J. Chem. Phys. 2016, 144, 054113:1–054113:10. [Google Scholar] [CrossRef]
  40. Jagau, T.C.; Bravaya, K.B.; Krylov, A.I. Extending quantum chemistry of bound states to electronic resonances. Ann. Rev. Phys. Chem. 2017, 68, 525–553. [Google Scholar] [CrossRef] [Green Version]
  41. Kunitsa, A.A.; Granovsky, A.A.; Bravaya, K.B. CAP-XMCQDPT2 method for molecular electronic resonances. J. Chem. Phys. 2017, 146, 184107:1–184107:8. [Google Scholar] [CrossRef]
  42. Ehara, M.; Kanazawa, Y.; Sommerfeld, T. Low-lying π* resonances associated with cyano groups: A CAP/SAC-CI study. Chem. Phys. 2017, 482, 169–177. [Google Scholar] [CrossRef] [Green Version]
  43. White, A.F.; Epifanovsky, E.; McCurdy, C.W.; Head-Gordon, M. Second order Moller-Plesset and coupled cluster singles and doubles methods with complex basis functions for resonances in electron-molecule scattering. J. Chem. Phys. 2017, 146, 234107:1–234107:10. [Google Scholar] [CrossRef] [PubMed]
  44. Thodika, M.; Fennimore, M.; Karsili, T.N.V.; Matsika, S. Comparative study of methodologies for calculating metastable states of small to medium-sized molecules. J. Chem. Phys. 2019, 151, 244104:1–244104:16. [Google Scholar] [CrossRef] [PubMed]
  45. Phung, Q.M.; Komori, Y.; Yanai, T.; Sommerfeld, T.; Ehara, M. Combination of a Voronoi-type complex absorbing potential with the XMS-CASPT2 method and pilot applications. J. Chem. Theor. Comp. 2020, 16, 2606–2616. [Google Scholar] [CrossRef] [PubMed]
  46. Adamson, S.O.; Kharlampidi, D.D.; Golubkov, G.V.; Dyakov, Y.A.; Morozov, I.I.; Shestakov, D.V.; Golubkov, M.G. Calculation of the lowest resonance 1S state of H ion by complex absorbing potential method. Russ. J. Phys. Chem. B 2020, 14, 742–751. [Google Scholar] [CrossRef]
  47. Belogolova, A.V.; Dempwolff, A.I.; Dreuw, A.; Trofimov, A.V. A complex absorbing potential electron propagator approach to resonance states of metastable anions. J. Phys. Conf. Ser. 2021, 1847, 012050:1–012050:10. [Google Scholar] [CrossRef]
  48. Bylicki, M.; Nicolaides, C.A. Theoretical resolution of the H resonance spectrum up to the n = 4 threshold. II. States of 1S and 1D symmetries. Phys. Rev. A 2000, 61, 052509:1–052509:10. [Google Scholar] [CrossRef]
  49. Bylicki, M. The Hermitian representation of the complex-rotation method and its application to the 1s2s2 2S resonance of He. J. Phys. B At. Mol. Opt. Phys. 1991, 24, 413–422. [Google Scholar] [CrossRef]
  50. Gopalan, A.; Bömmels, J.; Götte, S.; Landwehr, A.; Franz, K.; Ruf, M.-W.; Hotop, H.; Bartschat, K. A novel electron scattering apparatus combining a laser photoelectron source and a triply differentially pumped supersonic beam target: Characterization and results for the resonance. Eur. Phys. J. D 2003, 22, 17–29. [Google Scholar] [CrossRef]
  51. Cederquist, H.; Mannervik, S. On the core-excited doublets of Li I: Accurate determination of autoionisation rates. J. Phys. B At. Mol. Phys. 1982, 15, L807–L812. [Google Scholar] [CrossRef]
  52. Mannervik, S.; Cederquist, H. Absolute term energies of core-excited doublets of Li I determined experimentally. Phys. Scr. 1985, 31, 79–82. [Google Scholar] [CrossRef]
  53. Cederquist, H.; Mannervik, S. Autoionizing states in Li I observed in optical-emission spectra. Phys. Rev. A 1985, 31, 171–176. [Google Scholar] [CrossRef]
  54. Cardona, J.C.; Sanz-Vicario, J.L.; Martín, F. Complete Feshbach-type calculations of energy positions and widths of autoionizing states Li-like atoms. Phys. Rev. A 2010, 82, 022501:1–022501:13. [Google Scholar] [CrossRef]
  55. Landau, A.; Moiseyev, N. Molecular resonances by removing complex absorbing potentials via Padé; Application to CO and N2. J. Chem. Phys. 2016, 145, 164111:1–164111:8. [Google Scholar] [CrossRef]
  56. Ehrhardt, H.; Langhans, L.; Linder, F.; Taylor, H.S. Resonance scattering of slow electrons from H2 and CO angular distributions. Phys. Rev. 1968, 173, 222–230. [Google Scholar] [CrossRef]
  57. Berman, M.; Estrada, H.; Cederbaum, L.S.; Domcke, W. Nuclear dynamics in resonant electron-molecule scattering beyond the local approximation: The 2.3-eV shape resonance in N2. Phys. Rev. A 1983, 28, 1363–1381. [Google Scholar] [CrossRef]
  58. Lefebvre, R.; Sindelka, M.; Moiseyev, N. Resonance positions and lifetimes for flexible complex absorbing potentials. Phys. Rev. A 2005, 72, 052704:1–052704:6. [Google Scholar] [CrossRef]
  59. Bloch, C. Une formulation unifiée la théorie des réactions nucléares. Nucl. Phys. 1957, 4, 503–528. [Google Scholar] [CrossRef]
  60. Zvijac, D.J.; Heller, E.J.; Light, J.C. Variational correction to Wigner R-matrix theory of scattering. J. Phys. B At. Mol. Opt. Phys. 1975, 8, 1016–1033. [Google Scholar] [CrossRef]
  61. Halley, M.H.; Delande, D.; Taylor, K.T. The combination of R-matrix and complex coordinate methods: Application to resonances in the diamagnetic Rydberg spectra of Li. J. Phys. B At. Mol. Opt. Phys. 1992, 25, L525–L532. [Google Scholar] [CrossRef]
  62. Halley, M.H.; Delande, D.; Taylor, K.T. The combination of R-matrix and complex coordinate methods: Application to the diamagnetic Rydberg spectra of Ba and Sr. J. Phys. B At. Mol. Opt. Phys. 1993, 26, 1775–1790. [Google Scholar] [CrossRef]
  63. Huarte-Larranaga, F.; Gimenez, H.; Aguilar, A.; Baer, M. On the accuracy of reactive scattering calculations with absorbing potentials: A new implementation based on a generalized R-matrix propagation. Chem. Phys. Lett. 1998, 291, 346–350. [Google Scholar] [CrossRef]
  64. Huarte-Larranaga, F.; Gimenez, H.; Aguilar, A. The application of complex absorbing potentials to an invariant embedding scattering method: I. Theory and computational details. J. Chem. Phys. 1998, 109, 5761–5769. [Google Scholar] [CrossRef]
  65. Huarte-Larranaga, F.; Gimenez, H.; Lucas, J.M.; Aguilar, A. The application of complex absorbing potentials to an invariant embedding scattering method. II. Applications. J. Chem. Phys. 1999, 111, 1979–1987. [Google Scholar] [CrossRef]
  66. Szegö, G. Orthogonal Polynomials; American Mathematical Society: Providence, RI, USA, 1939; p. 432. ISBN 978-147-043-171-6. [Google Scholar]
  67. Akhiezer, N.I.; Glazman, A.I. Theory of Linear Operators in Hilbert Space; Pittman Publishing Inc.: Marshfield, MA, USA, 1981; p. 347. ISBN 027-308-495-X. [Google Scholar]
  68. Bateman, H.; Erdelyi, A. Higher Transcendental Functions; McGraw-Hill Book Company, Inc.: New York, NY, USA, 1953; p. 324. ISBN 089-874-206-4. [Google Scholar]
  69. Kazmin, Y.A. On subsequences of Hermite and Laguerre polynomials. Vestn. Mosk. Univ. Ser. Mat. Mekhaniki 1960, 2, 6–9. [Google Scholar]
  70. Ryndina, V.V. Approximation by Laguerre polynomials in the complex domain. Izv. Akad. Nauk. USSR Ser. Mat. 1962, 26, 143–160. [Google Scholar]
  71. Gradshteyn, I.S.; Ryzhik, I.M. Table of Integrals, Series and Products; Academic Press, Inc.: New York, NY, USA, 1980; p. 1191. ISBN 012-294-760-6. [Google Scholar]
  72. Davies, K.T.R.; Baranger, M. Symmetry properties of the S-matrix with application to resonance reactions. Ann. Phys. 1962, 19, 383–427. [Google Scholar] [CrossRef]
  73. Bain, R.A.; Bardsley, J.N.; Junker, B.R.; Sukumar, C.V. Complex coordinate studies of resonant electron-atom scattering. J. Phys. B At. Mol. Opt. Phys. 1974, 7, 2189–2202. [Google Scholar] [CrossRef]
  74. Zhang, S.; Jin, J. Computation of Special Functions; John Wiley & Sons: New York, NY, USA, 1996; p. 717. ISBN 047-111-963-6. [Google Scholar]
  75. Ting, T.C.T. A method of solving a system of linear equations whose coefficients form a tridiagonal matrix. Quart. Appl. Math. 1964, 22, 105–116. [Google Scholar] [CrossRef] [Green Version]
  76. Pedrotti, K.D. Extinction spectroscopy: A novel laser spectroscopic technique. Opt. Commun. 1987, 62, 250–255. [Google Scholar] [CrossRef]
  77. Cantù, A.M.; Parkinson, W.H.; Tondello, G.; Tozzi, G.P. Observations of Li i and Li ii absorption spectra in the grazing incidence region. J. Opt. Soc. Am. 1977, 67, 1030–1033. [Google Scholar] [CrossRef]
  78. Ederer, D.L.; Lucatorto, T.; Madden, R.P. Autoionization Spectra of Lithium. Phys. Rev. Lett. 1970, 25, 1537–1539. [Google Scholar] [CrossRef]
  79. Davis, B.; Chung, K.T. Widths of 2S, 2Po, and 2D resonances in Li I, Be II, and B III. Phys. Rev. A 1985, 31, 3017–3026. [Google Scholar] [CrossRef]
  80. Chung, K.T. Photoionization of lithium below the Li+ 1s2s3 S threshold. Phys. Rev. A 1997, 56, R3330–R3333. [Google Scholar] [CrossRef]
  81. Zatsarinny, O.; Fischer, C.F. The use of basis splines and nonorthogonal orbitals in R-matrix calculations: Application to Li photoionization. J. Phys. B At. Mol. Opt. Phys. 2000, 33, 313–342. [Google Scholar] [CrossRef]
  82. Verbockhaven, G.; Hansen, J.E. Energies and autoionization widths in the lithium isoelectronic sequence. J. Phys. B At. Mol. Opt. Phys. 2001, 34, 2337–2361. [Google Scholar] [CrossRef]
  83. Adamson, S.; Kharlampidi, D.; Preobrazhenskaya, A.A.; Dementiev, A. Calculation of the lowest 2S resonance state of He by a stabilization method. Russ. J. Phys. Chem. B 2017, 11, 894–902. [Google Scholar] [CrossRef]
  84. Adamson, S.; Kharlampidi, D.; Dementiev, A. Application of the uniformly charged sphere stabilization for calculating the lowest 1S resonances of H. In Advances in Quantum Methods and Applications in Chemistry, Physics, and Biology. Progress in Theoretical Chemistry and Physics; Hotokka, M., Brandas, E., Maruani, J., Delgado-Barrio, G., Eds.; Springer: Cham, Switzerland, 2012; Volume 27, pp. 101–118. ISBN 978-331-901-528-6. [Google Scholar]
  85. Preobrazhenskaya, A.A.; Adamson, S.O.; Kharlampidi, D.D.; Dement’ev, A.I. Calculation of the lowest 1S resonance state of the H anion by the stabilization method. Russ. J. Phys. Chem. B 2014, 8, 22–29. [Google Scholar] [CrossRef]
  86. Preobrazhenskaya, A.A.; Adamson, S.O.; Kharlampidi, D.D.; Dementiev, A.I. Ab initio calculations of lower resonant states of two-electron systems. Russ. J. Phys. Chem. B 2016, 10, 133–142. [Google Scholar] [CrossRef]
  87. Adamson, S.; Kharlampidi, D.; Dementiev, A. Stabilization of resonance states by an asymptotic Coulomb potential. J. Chem. Phys. 2008, 128, 024101:1–024101:7. [Google Scholar] [CrossRef] [PubMed]
Figure 1. Function E ( s ) at s > 0 when the charge of the target z 0 = 0 . The branches of the function E ( s ) correspond to problem I (solid lines), problem II for N = 100 (dashed lines), and problem III for N = 100 , λ = 1.6 (dash-dotted lines). Data are presented in a.u.
Figure 1. Function E ( s ) at s > 0 when the charge of the target z 0 = 0 . The branches of the function E ( s ) correspond to problem I (solid lines), problem II for N = 100 (dashed lines), and problem III for N = 100 , λ = 1.6 (dash-dotted lines). Data are presented in a.u.
Atoms 09 00072 g001
Figure 2. Trajectories (problem I, z 0 = 0 ) for different values of ζ = Re ( s ) , i.e., ○— ζ = 0.01 , □— ζ = 0.12 , ◊— ζ = 0.36 . The dashed line corresponds to ζ = 0 . Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Figure 2. Trajectories (problem I, z 0 = 0 ) for different values of ζ = Re ( s ) , i.e., ○— ζ = 0.01 , □— ζ = 0.12 , ◊— ζ = 0.36 . The dashed line corresponds to ζ = 0 . Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Atoms 09 00072 g002
Figure 3. Trajectories (with z 0 = 0 ) in the case ζ = λ = 0.01 for the problem I (solid line), problem II (dashed lines), and problem III (dash-dotted lines). Calculations were carried out for various values of N, i.e., ○—N = 100, □—N = 600. Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Figure 3. Trajectories (with z 0 = 0 ) in the case ζ = λ = 0.01 for the problem I (solid line), problem II (dashed lines), and problem III (dash-dotted lines). Calculations were carried out for various values of N, i.e., ○—N = 100, □—N = 600. Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Atoms 09 00072 g003
Figure 4. Trajectories (with z 0 = 0 ) in the case ζ = λ = 0.12 for the problem I (solid line), problem II (dashed lines), and problem III (dash-dotted lines). Calculations were carried out for various values of N, i.e., ○—N = 100, □—N = 600. Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Figure 4. Trajectories (with z 0 = 0 ) in the case ζ = λ = 0.12 for the problem I (solid line), problem II (dashed lines), and problem III (dash-dotted lines). Calculations were carried out for various values of N, i.e., ○—N = 100, □—N = 600. Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Atoms 09 00072 g004
Figure 5. Trajectories (with z 0 = 0 ) in the case ζ = λ = 0.36 for the problem I (solid line), problem II (dashed lines), and problem III (dash-dotted lines). Calculations were carried out for various values of N, i.e., ○—N = 100, □—N = 600. Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Figure 5. Trajectories (with z 0 = 0 ) in the case ζ = λ = 0.36 for the problem I (solid line), problem II (dashed lines), and problem III (dash-dotted lines). Calculations were carried out for various values of N, i.e., ○—N = 100, □—N = 600. Signs + and x indicate the position of the resonance and optimal point, respectively. Data are presented in a.u.
Atoms 09 00072 g005
Figure 6. CAP-trajectories in the vicinity of the lowest 1 S 0 resonance of H for the different values of the scaling factor λ (26), i.e., ○—λ = 1.3, □—λ = 1.5, ◊—λ = 1.7. Sign + indicates the position of the resonance [48]. Data are presented in a.u.
Figure 6. CAP-trajectories in the vicinity of the lowest 1 S 0 resonance of H for the different values of the scaling factor λ (26), i.e., ○—λ = 1.3, □—λ = 1.5, ◊—λ = 1.7. Sign + indicates the position of the resonance [48]. Data are presented in a.u.
Atoms 09 00072 g006
Figure 7. CAP-trajectories in the vicinity of the first (a) and the second (b) 2P0 resonance of Li atom for the different values of the scaling factor λ (26), i.e., ○—λ = 3.1, □—λ = 3.3, ◊—λ = 3.5, Δ—λ = 3.7. The horizontal line in panel (a) corresponds to the experimental value of the resonance width [51,53]. Sign + in panel (b) indicates the position of the calculated resonance [54]. Data are presented in a.u.
Figure 7. CAP-trajectories in the vicinity of the first (a) and the second (b) 2P0 resonance of Li atom for the different values of the scaling factor λ (26), i.e., ○—λ = 3.1, □—λ = 3.3, ◊—λ = 3.5, Δ—λ = 3.7. The horizontal line in panel (a) corresponds to the experimental value of the resonance width [51,53]. Sign + in panel (b) indicates the position of the calculated resonance [54]. Data are presented in a.u.
Atoms 09 00072 g007
Table 1. Optimal trajectory points (problem I). All the data are presented in a.u.
Table 1. Optimal trajectory points (problem I). All the data are presented in a.u.
R e ( s ) z 0 = 0 z 0 = 1
I m ( s o p t ) R e ( E o p t ) I m ( E o p t ) I m ( s o p t ) R e ( E o p t ) I m ( E o p t )
0.010.00313.4263900.0063880.00313.4264410.006375
0.011.08033.4263630.0063641.08033.4264430.006411
0.120.17603.4263820.0063880.16083.4264460.006380
0.121.05603.4263610.0063621.05603.4264430.006414
0.361.03243.4263500.0063451.01423.4264410.006435
Table 2. Optimal trajectory points (problem II). All the data are presented in a.u.
Table 2. Optimal trajectory points (problem II). All the data are presented in a.u.
R e ( s ) N z 0 = 0 z 0 = 1
I m ( s o p t ) R e ( E o p t ) I m ( E o p t ) I m ( s o p t ) R e ( E o p t ) I m ( E o p t )
0.011000.00163.4251150.0051850.00173.4268670.008752
1000.73473.4257470.0062160.74383.4268450.006961
6000.00183.4257850.0059740.00183.4266410.007170
6000.65593.4261080.0063240.66223.4266080.006624
0.121000.05793.4255670.0058750.05763.4267960.007439
1000.68633.4257440.0061940.68833.4268350.007100
6000.12273.4260420.0062480.11573.4266050.006735
6000.64943.4261050.0063180.64943.4266040.006629
0.361000.66223.4257540.0061720.66213.4268120.006999
6000.67513.4260960.0063210.67513.4266140.006633
Table 3. Optimal trajectory points (problem III). All the data are presented in a.u.
Table 3. Optimal trajectory points (problem III). All the data are presented in a.u.
R e ( s ) N z 0 = 0 z 0 = 1
I m ( s o p t ) R e ( E o p t ) I m ( E o p t ) I m ( s o p t ) R e ( E o p t ) I m ( E o p t )
0.011000.000853.4251750.0051440.000853.4267230.008766
1001.295223.4251840.0050681.295173.4266330.008886
6000.000063.4258040.0059470.000063.4266000.007182
6001.105473.4257840.0059171.091223.4265970.007227
0.121000.004463.4256450.0057870.004373.4266410.007474
1001.130393.4256280.0057541.130393.4266330.007523
6000.000353.4260590.0061630.000343.4265310.006789
6001.105473.4260350.0061381.091223.4265320.006828
0.361000.008553.4258040.0059470.008623.4265990.007181
6001.105473.4257850.0059191.105053.4265970.007225
Table 4. Energy (a.u.) and width (×10−5 a.u.) of the lowest 1S resonant state of H. Case A: averaging over a closed trajectory section. Case B: search for an optimal point.
Table 4. Energy (a.u.) and width (×10−5 a.u.) of the lowest 1S resonant state of H. Case A: averaging over a closed trajectory section. Case B: search for an optimal point.
λAB
−EresГ−EresГ
1.000.3512175172.730.3512240132.32
1.200.3512303176.650.3511645144.03
1.400.3512224176.730.3510555164.30
1.600.3512307179.650.3513016199.75
1.800.3512312178.950.3511226193.48
Averaged value a0.3512250175.830.3512137164.51
Reference value b0.35122025/173.870
a Averaged values for the interval 0.9 λ 1.8 with Δ λ = 0.1 [46]. b theory, complex eigenvalue Schrödinger Equation method [48].
Table 5. Energy (a.u.) and width (×10−5 a.u.) of the first and the second 2P0 resonant states of Li atom.
Table 5. Energy (a.u.) and width (×10−5 a.u.) of the first and the second 2P0 resonant states of Li atom.
λ2P0 [1s(2s2p) 3P]2P0 [1s(2s2p) 1P]
ABAB
−EresГ−EresГ−EresГ−EresГ
3.15.31225211.555.31218513.465.25696535.745.25680623.22
3.35.31221212.955.3121994.625.25660846.32--
3.55.31207011.705.31200114.085.25606334.585.25590022.91
3.75.31183713.115.3118274.505.25529844.335.25533214.07
Averaged value a5.31210713.335.3120837.675.25624739.975.25617920.02
Experimental and theoretical values
5.31307(3)/9.6(4) b
5.31183/13.5 c5.25624/34.5 c
5.256864/36.79 d
a Averaged values for the interval 3.1 λ 3.7 with Δ λ = 0.1 . b experiment [51,52,53]. c theory, Feshbach projection formalism [54]. d theory, saddle-point technique [79].
Table 6. Energy (a.u.) and width (×10−5 a.u.) of the first and the second 2P0 resonant states of Li atom calculated by stabilization method (1) and CAP (2). For CAP, the results were obtained by method A.
Table 6. Energy (a.u.) and width (×10−5 a.u.) of the first and the second 2P0 resonant states of Li atom calculated by stabilization method (1) and CAP (2). For CAP, the results were obtained by method A.
λ2P0 [1s(2s2p) 3P]2P0 [1s(2s2p) 1P]
1212
−EresГ−EresГ−EresГ−EresГ
3.15.31225314.025.31225211.555.25692040.715.25696535.74
3.35.31220510.085.31221212.955.25659833.975.25660846.32
3.55.31208214.335.31207011.705.25603943.465.25606334.58
3.75.31182810.055.31183713.115.25529232.615.25529844.33
Averaged value a5.31210611.985.31210713.335.25624338.375.25624739.97
a Averaged values for the interval 3.1 λ 3.7 with Δ λ = 0.1 .
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Adamson, S.O.; Kharlampidi, D.D.; Dyakov, Y.A.; Golubkov, G.V.; Golubkov, M.G. Calculation of the Lowest Resonant States of H and Li by the Complex Absorbing Potential Method. Atoms 2021, 9, 72. https://0-doi-org.brum.beds.ac.uk/10.3390/atoms9040072

AMA Style

Adamson SO, Kharlampidi DD, Dyakov YA, Golubkov GV, Golubkov MG. Calculation of the Lowest Resonant States of H and Li by the Complex Absorbing Potential Method. Atoms. 2021; 9(4):72. https://0-doi-org.brum.beds.ac.uk/10.3390/atoms9040072

Chicago/Turabian Style

Adamson, Sergey O., Daria D. Kharlampidi, Yurii A. Dyakov, Gennady V. Golubkov, and Maxim G. Golubkov. 2021. "Calculation of the Lowest Resonant States of H and Li by the Complex Absorbing Potential Method" Atoms 9, no. 4: 72. https://0-doi-org.brum.beds.ac.uk/10.3390/atoms9040072

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop