_{1}

^{*}

With a view to getting further insight into the solutions of one-dimensional analogous Schr?dinger equation for a non-hermitian (complex) Hamiltonian system, we investigate the quasi-exact
PT- symmetric solutions for an octic potential and its variant using extended complex phase space approach characterized by
x=x_{1}+ip_{2}, p=p_{1}+ix_{2}, where (
x_{1}, p_{1}) and (
x_{2}, p_{2}) are real and considered as canonical pairs. Besides the complexity of the phase space, complexity of potential parameters is also considered. The analyticity property of the eigenfunction alone is found sufficient to throw light on the nature of eigenvalue and eigenfunction of a system. The imaginary part of energy eigenvalue of a non-hermitian Hamiltonian exist for complex potential parameters and reduces to zero for real parameters. However, in the present work, it is found that imaginary component of the energy eigenvalue vanishes even when potential parameters are complex, provided that
PT-symmetric condition is satisfied. Thus
PT- symmetric version of a non-hermitian Hamiltonian possesses the real eigenvalue.

In the recent years, one-dimensional complex Hamiltonians have generated lot of interest for the understanding of several newly discovered phenomena in various science-streams [1,2], but such studies in mathematical terms have not been reached to the desired extent. Several attempts have been made to obtain the solutions of Schrödinger equation (SE) for different anharmonic potentials in real domain. However, the study of complex octic potential has become of considerable interest due to the peculiar nature of the eigenvalue spectrum. Further, besides some general studies of complex Hamiltonians in nonlinear domain [1,3], efforts have been made to study both classical and quantum aspects [4-7] of a system. At the classical context, becomes the function of two complex variables and the analyticity property of leads to a class of integrable systems in the associated twodimensional real systems and, where, act as new Hamiltonian and is a second integral of motion. The possible connection between and is sought in terms of anti-Bäcklund transformation [

where,. Here, the operators and are real, the commutator is invariant under operators and. It is interesting to note that commutation relation still remain invariant even if and becomes complex, provided that above transformation hold. There are various ways of complexifying a given Hamiltonian [

if, we define , then and can be defined as

The presence of variables in the above transformations may be regarded as some sort of co-ordinate momentum interactions of a dynamical system. Note that, in this complexifying scheme, the degrees of freedom of the underlying system just become double. The -symmetric condition for the above transormation becomes Though complex potentials are in practice for a long time, such as in optical model of nucleus, delocalization transition in condensed matter system—such as vortex flux line dippening in type-II superconductors, yet the quantum mechanics of complex potentials has not been studied to a desired level. It is since last few years that the study of complex potentials has become important enough for better theoretical understanding of the detailed properties of some newly discovered phenomena in physics and chemistry, like the phenomena pertaining to resonance scattering in atomic, molecular, and nuclear physics and to some chemical reactions [14-17]. The complex Hamiltonian is used in several other theoretical context likestudies of complex trajectories with regard to the calculation of semiclassical coherent-state propagator in the path integral method have attracted particular interest in laser physics [12,13]. The -symmetric non-hermitian Hamiltonians have many applications in various fields of physics-like superconductivity, population biology, quantum cosmology, condensed matter physics, quantum field theory etc.

Transformations similar to Equation (2) have also been used in the study of nonlinear evolution equations in context of amplitude-modulated nonlinear Langmuir waves in plasma [

The paper is organized as follows: in Section 2, we are devoted with the mathematical formulation of the ECPSA for computing the ground state and excited state eigenvalue spectra of some one-dimensional complex systems. Under the same mathematical prescription, ground state solutions are presented in Section 3 and excited state solutions are described in Section 4. Finally, concluding remarks are presented in Section 5.

For a complex Hamiltonian system in one dimension, the ASE (for) is given by

where

Here, Equation (3) departs from the conventional and mathematical setting of the standard Schrödinger equation [

Note that, the momentum operator of the conventional quantum mechanics under the transformation (2) reduces to This relation yields,. Also, the complex co-ordinate transformation (2) preserves the fundamental commutation relations, , which can be easily verified with the help of Equations (2) and (5). To express the ASE (3) into a pair of coupled partial differential equation, the complex forms of and E are written as

where, subscripts “r” and “I” denote the real and imaginary parts of the corresponding quantities and other subscripts to these quantities separated by comma will denote the partial derivatives of the quantity concerned. Thus, after inserting Equations (2), (4) and (6a)-(6c) in Equation (3) and separating the real and imaginary parts in the final expression, one gets the following pair of partial differential equations

The analyticity property of the wavefunction, in terms of Cauchy-Riemann conditions, implies

Under the analyticity condition (8), Equations (7a) and (7b), reduces to

The ansatz for the wavefunction is taken as [

where, and are the polynomial functions of the complex variable, which can be expressed as

After utilizing Equations (6b), (9b), (11a) and (11b) in Equation (10), the real and imaginary parts of the wavefunction are expressed as

In view of the analyticity condition (8), and satisfies the relations

Therefore, with the help of Equations (12a) and (12b), the Equations (9a) and (9b) yield

It is to be noted that for given functional forms of and, the rationalization of Equations (14a) and (14b) yield the real and imaginary components of the energy eigenvalue spectrum for the excited state of a system. On the other hand, if is chosen as constant, then Equations (14a) and (14b) reduces to ground state solutions as

With a suitable ansatz for, satisfying the analyticity condition (19), the rationalization of Equations (15a) and (15b), provides ground state solutions of the ASE for a given complex potential.

Here, we look for the ground state solutions of onedimensional complex octic potential and its variant as:

Consider a generalized octic potential of the form

where, the coupling parameters are complex i.e. and, are constants.

Under the -symmetric condition (1), the potential (24) reduces to

By implying the transformation (2) on the potential (17), the real and imaginary parts of the potential turn out to be

The polynomial forms of and, in conformity with (13) are written as

where, and are real. Now, inserting the above forms of and in Equations (15a) and (15b), the rationalization of the resultant expression yields the following set of non-repeating equations

Here, Equations (21c)-(21j) except (21e) give rise to the constraining relations among the potential parameters. However, the Equations (21e), (21h) and (21j) can be immediately solved for four arbitrary constants i.e., , and. Whereas Equations (21d)-(21j) can be solved for some negative values of say in the potential (17). The results obtained are

where, is real positive. The constraining relations obtained from Equations (21c), (21d), (21f) and (21g) are

The presence of these constraining relations, makes the problem quasi solvable. Such relations can be helpful in definition and approximate sub domain in complex parametric space in which a given complex potential provides real spectra. As from Equation (21b), imaginary part of the energy eigenvalue is zero, while the real part of the energy eigenvalue obtained from Equation (21a) turns out to be

The corresponding eigenfunction becomes

Here, we consider one-dimensional octic potential (16) with the inverse harmonic term as

where, the potential parameters and are complex constants. By implying the -symmetric condition (2) on the potential (26), one gets

Implying the transformation (2), the real and imaginary parts of the potential (26) are written as

where, and are same as given by Equations (18) and (19). The functional form of and complying with the analyticity condition (13) are written as

As before, using these forms of and in Equations (15a) and (15b), the rationalization of the resultant expression yield a set of equations in addition to Equations (21f)-(21j) as

The pair of Equations (30e) and (30f) immediately lead us to

The additional constraining relations given by Equations (30c) and (30d) are

Under the similar prescription as adopted in previous case, the imaginary component of the energy eigenvalue vanishes, whereas the real component of the energy is

The ground state eigenfunction for the potential (44) turn out to be

Here, we elaborate viability of the prescription laid down in Sections 2 and 3 to compute eigenvalue and corresponding eigenfunction for the first excited state. The functional form of for the first excited state is taken as

Then, under the transformation (2), the above equation reduces to

where, and are considered as real constants. In order to compute the corresponding energy eigenvalue and eigenfunction for the first excited state of potential (16), we use the same functional forms of and as mentioned in Equations (20a) and (20b). After inserting the Equations (20a), (20b) and (35) in Equations (14a) and (14b), then equating the coefficients of and their various products to zero, one gets the following set of non-repeating equations in addition to Equations (21f)- (21j)

Inserting Equations (21h)-(21j) in Equation (36e), we have

The other potential parameters are same as described in earlier section, whereas the constraining relations obtained from Equations (36c) and (36d) are

Using the various ansatz parameters in Equation (36a), the real component of energy eigenvalue is written as

whereas the eigenfunction is given by

Again to compute energy eigenvalue and corresponding eigenfunction for the first excited state of potential (26), we use the same functional forms of and as mentioned in Equations (29a) and (29b). Then implying Equations (20a), (20b) and (35) in Equations (14a) and (14b), the rationalization of the final expression yields the following set of non-repeating equations in addition to Equations (21f)-(21j)

The Equations (40e) and (40f) lead us to

The other ansatz parameters are same as for ground state solutions. However, the additional constraining relations given by Equations (40c) and (40d) are

Under the similar prescription as in previous case, the energy eigenvalue and eigenfunction are given by

In the present work, we have investigated quasi-exact -symmetric solutions of the ASE for one-dimensional octic potential and its variants using ECPSA. Besides complexity of the phase space produced by transformation (2), complexity of the potential parameters is also taken into account and ground state as well as excited states solutions are worked out. It is also emphasized that solutions of the ASE in the above said cases are obtained only in the presence of certain constraining relations among potential parameters, such constraining relations give rise to bound states of a system. It is found that imaginary part of the energy eigenvalue always vanishes for the solvable case of ASE, as long as all potential parameters are real. However, for - symmetric potentials, energy eigenvalues are found real, even if concerned potentials possess complex parameters. The interesting aspect of this method is an account of complex coupling coefficients of potential in addition to complex phase space. Thus present method suggests another degree of freedom to obtain the real spectra for non-hermitian operator.

The author expresses his gratitude to Prof. S. C. Mishra and Dr. Fakir Chand, Department of Physics, Kurukshetra University, Kurukshetra (India), for their valuable suggestions regarding the manuscript.