ebook img

Molecular zero-range potential method and its application to cyclic structures PDF

0.27 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Molecular zero-range potential method and its application to cyclic structures

Molecular zero-range potential method and its application to cyclic structures 1 1 0 2 Dmitry V. Ponomarev , Sergey B. Leble ∗ n a J Gdansk University of Technology, Theoretical Physics andQuantumInformatics department, 2 ] ul. Narutowicza 11/12, 80-223, Gdansk,Poland l l a h - s Abstract e m The zero-range potentials of the radial Schrodinger equation are investigated from a point of Darboux transformations . t a scheme. The dressing procedure is realized as a sequence of Darboux transformations in a way similar to that used to m obtain the generalized zero-range potentials of Huang-Derevianko by specific choice of a family of parameters. In the - d n present approach we stay within the framework of conventional zero-range potential method whilst the potential parameter o (scatteringlength)ismodifiedtakenintoaccountspectralmolecularproperties. Thisallowstointroducemolecularzero-range c [ potential oncethecorresponding discrete spectrum is known. The results are illustrated on exampleof flat cyclic molecular 1 structures,withparticularfocusonabenzenemolecule,whichboundedstatesenergiesarefirstfoundusingatomiczero-range v 9 potentials, compared with the Huckel method, and then used to introduce single zero-range potential describing the entire 3 4 molecule. Reasonablescatteringbehaviorfornewlyintroducedpotentialgivesapossibilitytotacklemany-moleculeproblems 0 . representing molecules as appropriate single zero-range potentials. 1 0 1 1 : v 1 Introduction i X r Zero-rangepotential(ZRP)methodisprovedtobeanefficienttoolindescribingquantumstructureswitharbitrary a geometry at sufficiently low energies such that detailed structure and interactions of atoms (replaced with point- centers) between themselves can be neglected [13, 5, 2, 9, 12]. Simplicity of the approach makes it very attractive and over the years it has been further developed into generalized zero-range potential (gZRP) method that takes into account effects of higher order partial waves in describing scattering properties [14, 15, 10, 4, 18, 19, 16]. However,the scope of applicationofthe regularZRP method canbe extended to describe not only small,typically atomic, structures but larger molecular systems. ∗Currently atMcMasterUniversity,MathematicsandStatisticsdepartment. 1 2 Method of zero-range potentials 2 In the present work, after giving brief but self-consistent review of Darboux transformation (DT) and gZRP ideas, we enhance standard ZRP approach by means of application of a dressing procedure [11, 18, 19]. We notice and make use the fact that a certain class of DTs of spherically symmetric ZRPs yields also a potential of zero rangebut,ingeneral,withdifferenteffectivecharacteristic(s-wavescatteringlengthoftheoriginalZRPisaltered). Transformation parameters of a single atomic potential can be chosen based on solution of conventional bounded state problem solved for a system of ZRPs (molecule) such that resulting solution mimics scattering behavior for the molecule as a whole at low energies. Described ZRP treatment is demonstrated on example of a generic cyclic structure with focus on a particular case of benzene molecule. 2 Method of zero-range potentials The idea of the ZRP method is to replace real potential with a point potential placed in the origin that mimics physical scattering behavior in the far zone. Therefore the equation in question is the radial Schrodinger equation for the free space (r >0) 2 l(l+1) ψ′′+ ψ′ + ψ =k2ψ, (1) − r r2 (cid:18) (cid:19) where k2 = 2µE is wave number, µ, E are mass and energy of a particle (electron), respectively, ~ is the Planck’s ~2 constant, ψ =ψ(r) is the radial part of the wave function. Itiswell-knownthatthegeneralsolutioncanbeformedasalinearcombinationofsphericalBesselandNeumann functions, j (kr) and y (kr), respectively, l l ψ(r)=C (j (kr) tanη y (kr)), (2) l l l l − · or as a combination of spherical Hankel functions ψ(r)=C˜ s h(1)(kr) h(2)(kr) , (3) l l l − l (cid:16) (cid:17) with s =exp(2iη) being a scattering matrix. l l Taking into account the following asymptotes at kr 0 [1] → (kr)l j (kr) , (4) l ≈ (2l+1)!! (2l 1)!! y (kr) − , (5) l ≈− (kr)l+1 2 Method of zero-range potentials 3 written with notion of the odd factorial (2l+1)!!= (2l+1) (2l 1) ... 3 1, ( 1)!! = 1, we obtain asymptotic · − · · · − behavior of finite-energy solution at the origin (kr)l (2l 1)!! (2l 1)!! ψ(r) C +tanη − C tanη − . (6) ≈ l (2l+1)!! l (kr)l+1 !≈ l l (kr)l+1 From here, the constant C can be expressed as l (2l+1)!! d2l+1 C = rl+1ψ . (7) l kl(2l+1)! dr2l+1 (cid:12)r=0 (cid:0) (cid:1)(cid:12) (cid:12) (cid:12) Elimination of C yields l 1 d2l+1 (2l+1)! 1 k2l+1 rl+1ψ = , (8) rl+1ψdr2l+1 (2l+1)!! · (2l 1)!! · tanη (cid:12)r=0 − l (cid:0) (cid:1)(cid:12) (cid:12) (cid:12) This is to say that imposing the boundary condition 1 d2l+1 2ll! rl+1ψ = 1/a2l+1, (9) rl+1ψdr2l+1 −(2l 1)!! l (cid:12)r=0 − (cid:0) (cid:1)(cid:12) (cid:12) with (cid:12) tanη a2l+1 = l. (10) l −k2l+1 results in the desired physical behavior of solution at far distance from the point-center. Introduced quantity a is termed as the partial wave scattering length and is independent of particular energy l value at low energies. To justify this fact, it is enough to consider zero-energy solution to the equation (1) 1 ψ(r)=A rl+A . 1 2rl+1 By matching this with (2), we conclude 1 tanη A /A k2l+1, l 2 1 ≈− (2l+1)!!(2l 1)!! − const that demonstrates energy-independence of scat|tering lengt{hzparameters}a . l In most of the cases, at low energies, it is enough to consider ZRP as a spherical point, thus neglecting higher partial waves. Therefore, solution to the (1) should be a subject to the simple boundary condition dlog(rψ) = β, (11) dr − (cid:12)r=0 (cid:12) (cid:12) (cid:12) 3 Darboux transformation and its application 4 where β =1/a is the s-wave inverse scattering length. 0 3 Darboux transformation and its application Consider one-dimensional Sturm-Liouville equation Lψ=λψ, (12) where d2 L= +u(x). (13) −dx2 Now we apply such algebraic-differentialtransformation(referred as Darboux transformation [8]) d D = σ(x) (14) dx − that the transformed equation preserve the same form L[1]ψ[1] =λψ[1], (15) d2 where ψ[1] =Dψ, L[1] = +u[1](x) and the transformed potential is −dx2 u[1](x)=u(x) 2σ′(x). (16) − It is straightforwardto check that ′ Φ σ(x)= 1 (17) Φ 1 with Φ being a particular solution to the original equation, i.e. it satisfies 1 ′′ Φ +u(x)Φ =λ Φ − 1 1 1 1 for some λ . 1 It can be shown that in case of multiple application of the Darboux transformation, the expression (17) is generalized into d σ(x)= [logW(Φ , ..., Φ )] 1 N dx resulting in the following transformationof the potential 3 Darboux transformation and its application 5 d2 u[N](x)=u(x) 2 [logW(Φ , ..., Φ )]. (18) − dx2 1 N Corresponding transformationof the solution is given by the Crum’s formula [7, 20] W(Φ , ..., Φ , ψ) ψ[N] =D[N]ψ = 1 N , (19) W(Φ , ..., Φ ) 1 N where W denote determinants of the Wronskian matrices. Having considered one-dimensional Cartesian case, we move towards the radial Schrodinger equation (1) and notice that it can be brought to the form (12) eligible for direct application of obtained transformation formulas. Namely, performing substitution ψ =χ/r, one readily obtains l(l+1) χ′′+ χ=k2χ. (20) − r2 That is to say, that we can apply Darboux transformation to the equation (1) meaning that all original wave functions ψ should be multiplied by r whereas the potential term l(l+1) u(r)= (21) r2 remains unchanged. Thus, the formulas (18), (23) in spherical coordinates should be replaced with d2 u[N](r)=u(r) 2 [logW(rΦ , ..., rΦ )] (22) − dr2 1 N and W(rΦ , ..., rΦ , rψ) ψ[N] =D[N]ψ = 1 N . (23) rW(rΦ , ..., rΦ ) 1 N Following an idea of the previous works [11], we first show how gZRP can be induced by application of DT. We start by choosing a spherical Bessel function as the seed solution ψ(r)=j (kr) (24) l l and apply N-th order Darboux transformation by taking spherical Hankel functions with specific parameters κ m as prop functions Φ (r)=h(1)( iκ r), m=1, ..., N. (25) m l − m 3 Darboux transformation and its application 6 We can employ Crum’s formula (19) and consider asymptotic behavior of spherical functions at r →∞ sin(kr lπ/2) j(kr) − , (26) l ≈ kr cos(kr lπ/2) y (kr) − , (27) l ≈− kr exp(ikr) h(1)(kr)=j(kr)+iy (kr) ( i)l+1 , (28) l l l ≈ − kr exp( ikr) h(2)(kr)=j (kr) iy (kr) il+1 − . (29) l l − l ≈ kr Then the Wronskians turn into Vandermond determinants, hence, exp(ikr)∆(κ , ..., κ , ik) exp( ikr)∆(κ , ..., κ , ik) ψ[N](r)=C ( i)l 1 N il − 1 N − , (30) l − kr ∆(κ , ..., κ ) − kr ∆(κ , ..., κ ) (cid:20) 1 N 1 N (cid:21) where C represents here and later on a generic constant without prescribing it any specific value such that it can absorb constant multipliers where their meaning is not important without changing notation. We note that for l =0 this expression is not asymptotic, but exact. A Vandermond determinant in (30) can be computed by noticing that k = iκ (for m = 1, ..., N) are the m − rootsofpolynomialwithrespecttok equationthatis obviousduetothe factthatreplacementik κ yieldszero m → determinant due to linear dependencies of the rows, thereby allowing the following factorization 1 κ κ2 ... κN 1 1 1 (cid:12) (cid:12) (cid:12)(cid:12) 1 κ2 κ2 ... κN2 (cid:12)(cid:12) N (cid:12) (cid:12) ∆(κ1, ..., κN, ik)=(cid:12)(cid:12) ... ... ... ... ... (cid:12)(cid:12)=C (κm−ik). (cid:12) (cid:12) m=1 (cid:12) 1 κ κ2 ... κN (cid:12) Y (cid:12) N N N (cid:12) (cid:12) (cid:12) (cid:12)(cid:12) 1 ik (ik)2 ... (ik)N (cid:12)(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) Denoting N N (κ ik) (k+iκ ) m m s = − = , (31) l (κ +ik) − (k iκ ) m m m=1 m=1 − Y Y we recognize in the expression (30) the asymptotes of spherical Hankel functions, hence ψ[N](r)=C s h(1)(kr) h(2)(kr) . (32) l l l − l h i 3 Darboux transformation and its application 7 As one can notice, the expression (32) coincides with (3) if N (κ ik) m exp(2iη)= − , (33) l (κ +ik) m m=1 Y or, taking into account (10), N (κ ik) N (κ +ik) tanη = a2l+1k2l+1 = i m=1 m− − m=1 m . (34) l − l − N (κ ik)+ N (κ +ik) Qm=1 m− Qm=1 m Q Q Thus we conclude that for the directcorrespondenceto the gZRPbehavior,the number of consequenttransfor- mations N =2l+1 should be taken. Given some quantity a= a eiφa, the parameters κ1, ..., κ2l+1 needs to be chosen in a way that | | 2l+1 (κ +ik)=(ik)2l+1 a. (35) m − m=1 Y This is equivalent to the ik= κ (m=1, ..., 2l+1) being the roots of the equation m − (ik)2l+1 =a, that is to say κ = (2l+1√)a, or m − (φ +2πm) κ = a1/(2l+1)exp i a , m=1, ..., 2l+1. (36) m − | | 2l+1 (cid:18) (cid:19) In the similar manner we obtain 2l+1 (κ ik)=( ik)2l+1 a. (37) m − − − m=1 Y Substitution of (35), (37) into (34) results in a=( 1)l+1/a2l+1. − l Therefore, providing a is a real number, (36) yields l l+2m+1 κ = 1/a exp iπ , m=1, ..., 2l+1. (38) m l − · 2l+1 (cid:18) (cid:19) Now alternatively to gZRP instead of taking into account impact of higher harmonics, we consider the trans- formation of spherically symmetric solution (i.e. l=0) with the parameters κ to be chosen according to spectral m data of the entire molecule obtained, for example, from the discrete spectrum of molecule modeled by regularZRP method. 3 Darboux transformation and its application 8 We get back to (34) and since N N N N N N N N (κ +ik)= κ +ik κ +(ik)2 κ +...+(ik)N−1 κ +(ik)N, m m m m n m=1 m=1 n=1m=1 j=1n<j m=1 n=1 Y Y XmY6=n XXm6=Yn,m6=j X we continue the last equality as ik N N κ +...+(ik)N n=1 m=1 m a2l+1k2l+1 = i m6=n . (39) l − NPκ +Q...+(ik)N−1 N κ m=1 m n=1 n Q P Limiting ourselves with l=0 case, we consider low-energylimit neglecting terms of higher order than one with respect to k to obtain N N κ n=1 m=1 m a = m6=n . (40) 0 P NQ κ m=1 m Alternatively, recalling notation β =1/a , the last eQxpression can be rewritten as 0 −1 N β = κ−1 (41) m ! m=1 X On the other hand, we can readily observe that Darboux transformation of the seed solution (24) with the prop functions (25) results in the solution corresponding to the ZRP for arbitrary choice of N and parameters κ , m m=1, ..., N. Indeed, for l =0 N logW(rΦ , ..., rΦ )= κ r+C, 1 N m m=1 X thus, according to (22), u[N](r) 0, r >0. ≡ Freedom of choice of transformation parameters gives a possibility to induce desired poles of scattering matrix (31)and, thereby,performtransitionfromatomic to molecularZRP by choosingκ , m=1, ..., M suchthat new m potential allows M bounded states obtained by solving conventional formulation of the discrete spectrum problem for the set of M standard ZRPs making up the molecule. Molecular ZRP can be qualitatively characterized by effective scattering length computed according to (40) and thus used in such simplified scattering calculations in complicated problems involving, for example, a chain of molecules. Below we demonstrate the presented approach by applying it to a general cyclic molecule with eventual focus on a benzene molecule. 4 Case study: a cyclic molecule 9 4 Case study: a cyclic molecule Inordertodescribeacyclicstructure,weconsiderasystem(molecule)ofM point-centers,typicallyatoms,modeled by ZRPs which positions are kept fixed in the plane in space in a cyclic order. Spatial coordinates are chosensuch that all atoms lie in the xy-plane with the origincoinciding with the center of the molecule. Atoms are numerated counterclockwise and the x-axis is chosen along the direction to the first atom. Described geometry is illustrated on the Fig. 1. Fig. 1: Illustration of the geometry by an example of 3 atoms. 4.1 Bounded state problem According to the idea of ZRP method, we write solution to the bounded state problem as M ψ(~r)= C g ~r R~ , κ , (42) j j · − Xj=1 (cid:16)(cid:12) (cid:12) (cid:17) (cid:12) (cid:12) (cid:12) (cid:12) subject to the conditions (11) ∂log ~r R~ ψ(~r) i − · = β, i=1, ..., M, (43) (cid:16)∂(cid:12)(cid:12) ~r R~(cid:12)(cid:12)i (cid:17)(cid:12)(cid:12) − (cid:12)(cid:12) − (cid:12)(cid:12) (cid:12)(cid:12)|~r−R~i|=0 (cid:12) (cid:12) (cid:12) where (cid:12) (cid:12) (cid:12) e−κr g(r, κ)= . (44) r Let us introduce the following notation. We will denote a distance between nearest neighboring atoms as ∆R , 1 4 Case study: a cyclic molecule 10 between every second atom from each given one as ∆R , every third as ∆R , and so on (see the Fig. 2). 2 3 Fig. 2: Illustration of the distance notation. From simple geometrical considerations it follows that 2π j ∆R =2R sin =2R sin(φ /2), (45) j 0 0 j+1 M · 2 (cid:18) (cid:19) 2π where φ = (j 1) is azimuthal angular coordinate of a j-th atom, R is radius of the molecule. We note that j 0 M − the expression is formally valid for j = 1, ..., M, however we note that there is only [M/2] different values since ∆R[M/2]−j = ∆R[M/2]+j for j = 1, ..., [M/2] 1 where [], here and later on, marks integer part of an argument − · (i.e. the floor function). We develop conditions (43) to show existence of the derivatives ∂log ~r R~ ψ(~r) i 1 − · (cid:16)∂(cid:12)(cid:12) ~r R~(cid:12)(cid:12)i (cid:17)(cid:12)(cid:12) = Ci [(...+Ci−1·g(∆R1, κ)−Ci·κ+Ci+1·g(∆R1, κ)+...)+ (cid:12)(cid:12) − (cid:12)(cid:12) (cid:12)(cid:12)|~r−R~i|=0 (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) ∂ ψ(~r) g ~r R~ , κ i + ~r R~ − − . (cid:12)(cid:12)(cid:12) − i(cid:12)(cid:12)(cid:12)· (cid:16) ∂(cid:12)(cid:12)~r(cid:16)−(cid:12)(cid:12)(cid:12) R~i(cid:12)(cid:12) (cid:12)(cid:12)(cid:12) (cid:17)(cid:17)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)|~r−R~i|=0 (cid:12) (cid:12) (cid:12) To show that the last term in the square bracketed expression is well-defined and equal to zero due to the

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.