Infinite quantum well: a coherent state approach

Infinite quantum well: a coherent state approach

P. L. García de León, J. P. Gazeau and J. Queva Laboratoire Astroparticules et Cosmologie, Bâtiment Condorcet, 10, rue Alice Domon et Léonie Duquet, 75205 Paris Cedex 13 Université Paris Est - Institut Gaspard Monge (IGM-LabInfo), 5 Bd. Descartes, Champs-sur-Marne, 77454 Marne-la-Vallée Cedex 2 Université Paris Diderot-Paris7, Bâtiment des Grands Moulins, 75205 Paris Cedex 13,,

A new family of 2-component vector-valued coherent states for the quantum particle motion in an infinite square well potential is presented. They allow a consistent quantization of the classical phase space and observables for a particle in this potential. We then study the resulting position and (well-defined) momentum operators. We also consider their mean values in coherent states and their quantum dispersions.

, Vector coherent states, quantization, infinite square well
03.65.-w, 03.65.Ca

1 Introduction

Even though the quantum dynamics in an infinite square well potential represents a rather unphysical limit situation, it is a familiar textbook problem and a simple tractable model for the confinement of a quantum particle. On the other hand this model has a serious drawback when it is analyzed in more detail. Namely, when one proceeds to a canonical standard quantization, the definition of a momentum operator with the usual form has a doubtful meaning. This subject has been discussed in many places (see [1] for instance), and the attempts of circumventing this anomaly range from self-adjoint extensions [1] to symmetry approaches [2].

First of all, the canonical quantization assumes the existence of a momentum operator (essentially) self-adjoint in that respects some boundary conditions on the boundaries of the well. As has been shown, these conditions cannot be fullfilled by the usual derivative form of the momentum without the consequence of losing self-adjointness. Moreover there exists an uncountable set of self-adjoint extensions of such a derivative operator which makes truly delicate the question of a precise choice based on physical requirements [1, 3].

When the classical particle is trapped in an infinite well of real interval , the Hilbert space of quantum states is and the quantization problem becomes similar, to a certain extent, to the quantization of the motion on the circle . Notwithstanding the fact that boundary conditions are not periodic but impose instead that the wave functions in position representation vanish at the boundary, the momentum operator for the motion in the infinite well should be the counterpart of the angular momentum operator for the motion on the circle. Since the energy spectrum for the infinite square well is , we should expect that the spectrum of should be , like the one for without the null eigenvalue. This similarity between the two problems will be exploited in the present paper. We will adapt the coherent states (CS’s) on the circle [4, 5, 6] to the present situation by constructing two-component vector CS’s, in the spirit of [8], as infinite superpositions of spinors eigenvectors of .

In the present note, we first describe the CS quantization procedure. We recall the construction of the CS’s for the motion on the circle and the resulting quantization. We then revisit the infinite square well problem and propose a family of vector CS’s suitable for the quantization of the related classical phase space. Note that various constructions of CS’s for the infinite square well have been carried out, like the one in [9] or yet the one resting upon the dynamical symmetry [10]. Finally, we present the consequences of our choice after examining basic quantum observables, derived from this quantization scheme, like position, energy, and a quantum version of the problematic momentum. In particular we focus on their mean values in CS’s (“lower symbols”) and quantum dispersions. As will be shown, the classical limit is recovered after choosing appropriate limit values of some parameters present in the expression of our CS’s.

2 The approach via coherent state quantization

Coherent state quantization [11, 12, 13, 14, 15, 16, 17] is an alternative way of representing classical observables into a quantum system. The states used in it include Glauber and Perelomov CS’s but lie in a wider definition that admits a large range of state families resolving the identity. Identity resolution is here the crucial condition.

In fact, these states form a frame of reference well suited to represent classical quantities and, in that sense, work as a natural quantization procedure which is in one-to-one correspondence with the choice of the frame. The validity of a precise frame choice is asserted by comparing spectral characteristics of quantum observables with data from the observational space. Unlike canonical quantization where the whole model rests upon a pair of conjugated variables within the Hamilton formalism [18], here we need the following elements.

First of all let be a set equipped with a measure , and let be the Hilbert space of square integrable functions on :


The set can be taken as the phase space of a particular problem as will be the case in this paper. Next we need a finite or infinite orthonormal set , selected among the elements of . This set spans, by definition, the separable Hilbert subspace and must obey the following condition:


Now let us define the family of coherent states in through the following linear superposition:


where the states are in one to one correspondence with the functions in the set . This is an injective map (which should be continuous with respect to some minimal topology affected to for which the latter is locally compact): These coherent states have two main features: they are normalized, and crucially, they resolve the identity in


The CS quantization of a classical observable on , consists then in associating to the operator


This “diagonal” decomposition (in a topological weak sense) may reveal to be valid for a wide class of operators. The function is called upper (or contravariant) symbol of the operator and is non-unique in general. On the other hand, the mean value is called lower (or covariant) symbol of .

3 Quantization of the particle motion on the circle

The motion in the infinite square well potential can be seen as a particular case of the motion on the circle , once we have identified the boundaries of the well with each other and imposed Dirichlet conditions on them. Functions on this domain will behave as pinched waves on a circle so it is useful to expose first the more general case.

Applying our scheme of quantization we can define the CS’s on the circle. The measure space is the cylinder , i.e. the phase space of a particle moving on the circle, where and are canonically conjugate variables. We consistently choose the measure on as the usual one, invariant (up to a factor) with respect to canonical transformations: . The functions forming the orthonormal system needed to construct CS’s are suitably weighted Fourier exponentials:


where can be arbitrarily small. This parameter includes the Planck constant together with the physical quantities characterizing the classical motion (frequency, mass, etc.). Actually, it represents a regularization. Notice that the continuous distribution is the normal law centered at (for the angular momentum variable ). We establish a one-to-one correspondence between the functions and the states which form an orthonormal basis of some generic separable Hilbert space that can be viewed or not as a subspace of . CS’s, as vectors in , read then as


where the normalization factor


is a periodic train of normalized Gaussian functions and is proportional to an elliptic Theta function. Applying the Poisson summation yields the alternative form:


From this formula it is easy to prove that .

The CS’s (7) have been previously proposed, however through quite different approaches, by De Bièvre-González (1992-93) [4], Kowalski-Rembieliński-Papaloucas (1996) [5], and González-Del Olmo (1998) [6].

3.1 Quantization of classical observables

The quantum operator acting on , associated to the classical observable , is obtained as in (5). For the most basic one, i.e. the classical observable itself, the procedure yields


and this is nothing but the angular momentum operator, which reads in angular position representation (Fourier series): .

For an arbitrary function , we have


where is the -th Fourier coefficient of . In particular, we have for the angular position operator :


The shift operator is the quantized counterpart of the “Fourier fundamental harmonic”:


The commutation rule between (10) and (13) gives


and is canonical in the sense that it is in exact correspondence with the classical Poisson bracket


Some interesting aspects of other such correspondences are found in [19]. For arbitrary functions of the commutator


can arise interpretational difficulties. In particular, when , i.e. for the angle operator


the comparison with the classical bracket is not direct. Actually, these difficulties are only apparent if we consider instead the -periodic extension to of . The position observable , originally defined in the interval , acquires then a sawtooth shape and its periodic discontinuities are accountable for the discrepancy. In fact the obstacle is circumvented if we examine, for instance, the behaviour of the corresponding lower symbols at the limit . For the angle operator we have


where we recognize at the limit the Fourier series of . For the commutator, we recover the canonical commutation rule modulo Dirac singularities on the lattice .


4 Quantization of the motion in an infinite well potential

4.1 The standard quantum context

Any quantum system trapped inside the infinite square well must have its wave function equal to zero beyond the boundaries. It is thus natural to impose on the wave functions the conditions


Since the motion takes place only inside the interval , we may as well ignore the rest of the line and replace the constraints (20) by the following ones:


Moreover, one may consider the periodized well and instead impose the cyclic boundary conditions .

In either case, stationary states of the trapped particle of mass are easily found from the eigenvalue problem for the Schrödinger operator with Hamiltonian:


This Hamiltonian is self-adjoint [7] on an appropriate dense domain in (21). Then


where obeys the eigenvalue equation


together with the boundary conditions (21). Normalized eigenstates and corresponding eigenvalues are then given by




where is the “revival” time to be compared with the purely classical round trip time.

4.2 The quantum phase space context

The classical phase space of the motion of the particle is the infinite strip equipped with the measure: . A phase trajectory for a given non-zero classical energy is represented in the figure 1.

Typically, we have two phases in the periodic particle motion with a given energy: one corresponds to positive values of the momentum, while the other one is for negative values, . This observation naturally leads us to introduce the Hilbert space of two-component complex-valued functions (or spinors) square-integrable with respect to :


We now choose our orthonormal system as formed of the following vector-valued functions , ,




and the half-width is a parameter which has the dimension of a momentum, say with a dimensionless parameter. This parameter can be arbitrarily small (like for the classical limit) and, of course, arbitrarily large (for a very narrow well, for instance).

The functions are continuous, vanish at the boundaries and of the phase space, and obey the essential finiteness condition (2):


The expression of simplifies to :


It then becomes apparent that and can be expressed in terms of elliptic theta functions. Function has no physical dimension whereas has the same dimension as , that is the inverse of an action.

We are now in measure of defining our vector CS’s [8]. We set up a one-to-one correspondence between the functions ’s and two-component states


forming an orthonormal basis of some separable Hilbert space of the form . The latter can be viewed also as the subspace of equal to the closure of the linear span of the set of ’s. We choose the following set of diagonal real matrices for our construction of vectorial CS’s:


Note that . Vector CS’s, , are now defined for each and by the relation


In particular, we single out the two orthogonal CS’s


By construction, these states also satisfy the infinite square well boundary conditions, namely . Furthermore they fulfill the normalizations


and the resolution of the identity in :


where denotes the identity matrix consistently with the Pauli matrix notation to be used in the following.

4.3 Quantization of classical observables

The quantization of a generic function on the phase space is given by the expression (5), that is for our particular CS choice:




For the particular case in which is function of only, , the operator is given by




Note that this operator is diagonal on the basis.

4.3.1 Momentum and Energy

In particular, using , one gets the operator


where is a Pauli matrix.

For , which is proportional to the Hamiltonian, the quantum counterpart reads as


Note that this implies that the operator for the square of momentum does not coincide with the square of the momentum operator. Actually they coincide up to O.

4.3.2 Position

For a general function of the position our quantization procedure yields the following operator:




In particular, for we get the “position” operator


with . Note the appearance of the classical mean value for the position on the diagonal.

4.3.3 Commutation rules

Now, in order to see to what extent these momentum and position operators differ from their classical (canonical) counterparts, let us consider their commutator:


This is an infinite antisymmetric real matrix. The respective spectra of finite matrix approximations of this operator and of position and momentum operators are compared in figures 2 and 3 for various values of the regulator in units , . When takes large values, one can see that the eigenvalues of accumulate around , i.e. they become almost canonical. Conversely, when all eigenvalues become null, which corresponds to the classical limit.

4.3.4 Evolution operator

The Hamiltonian of a spinless particle trapped inside the well is simply . Its quantum counterpart therefore is . The unitary evolution operator, as usual, is given by


Note the appearance of the global time-dependent phase factor with frequency which can be compared with the revival frequency


5 Quantum behaviour through lower symbols

Lower symbols are computed with normalized CS’s. The latter are denoted as follows


Hence, the lower symbol of a quantum observable should be computed as

This gives the following results for the observables previously considered :

5.0.1 Position

In the same way, the mean value of the position operator in a vector CS is given by:


where we can distinguish the classical mean value for the position corrected by the function


This function depends on the parameter as we show in figure 4 with a numerical approximation using finite matrices. As for , we calculate the dispersion defined as


Its behaviour for different values of is shown in figure 6.

5.0.2 Time evolution of position

The change through time of the position operator is given by the transformation , and differs from by the insertion of an oscillating term in the series. Its lower symbol is given by


where this time the series have the form


Note that the time dependence manifests itself in the form of a Fourier series of with frequencies . This corresponds to the circulation of the wave packet inside the well.

5.0.3 Momentum

The mean value of the momentum operator in a vector CS is given by the affine combination:


This function reproduce the profile of the function , as can be seen in the figure 5. We calculate then the dispersion , defined as


using the mean values in a CS . Its behaviour as a function of is shown in figure 7.

5.0.4 Position-momentum commutator

The mean value of the commutator in a normalized state is the pure imaginary expression:


Given the symmetry and the real-valuedness of states (36), the mean value of the commutator when is one of our CS’s vanish, even if the operator does not. This result is due to the symmetric spectrum of the commutator around . As is shown in Part c) of figures 2, the eigenvalues of the commutator tend to as , i.e. , increases. Still, there are some points with modulus less than . This leads to dispersions in CS’s that are no longer bounded from below by . Actually, the lower bound of this product, for a region in the phase space as large as we wish, decreases as diminish. A numerical approximation is shown in figure 8.

6 Discussion

From the mean values of the operators obtained here, we verify that our CS quantization gives well-behaved momentum and position operators. The classical limit is reached once the appropriate limit for the parameter is found. If we consider the behaviour of the observables as a function of the dimensionless quantity , at the limit and when the Gaussian functions for the momentum become very narrow, the lower symbol of the position operator is . This corresponds to the classical average value position in the well. On the other hand, at the limit , for which the involved Gaussians spread to constant functions, the mean value converges numerically to the function . In other words, our position operator yields a fair quantitative description for the quantum localization within the well. The lower symbol behaves as a stair-step function for close to and progressively fits the function when increases. These behaviours are well illustrated in the figures 4 and 5. The effect of the parameter is also noticeable in the dispersions of and . Here, the variations of the full width at half maximum of the Gaussian function reveal different dispersions for the operators. Clearly, if a classical behaviour is sought, the values of have to be chosen near . This gives localized values for the observables. The numerical explorations shown in figures 6 and 7 give a good account of this modulation. Consistently with the previous results, the behaviour of the product at low values of shows uncorrelated observables at any point in the phase space, whereas at large values of this parameter the product is constant and almost equal to the canonical quantum lowest limit . This is shown in figure 8.

It is interesting to note that if we replace the Gaussian distribution, used here for the variable in the construction of the CS’s, by any positive even probability distribution such that the results are not so different! The momentum spectrum is still and the energy spectrum has the form . In this regard, an interesting approach combining mathematical statistics concepts and group theoretical constructions of CS’s has been recently developed by Heller and Wang [20, 21].

The work presented here has possible applications to those particular physical problems where the square well is used as a model for impenetrable barriers [22], in the spirit of what has been done in [23].

The generalization to higher-dimensional infinite potential wells is more or less tractable, depending on the geometry of the barriers. This includes quantum dots and other quantum traps. Nevertheless, we believe that the simplicity and the universality of the method proposed in the present work should reveal itself useful for this purpose.

Author García de León wishes to acknowledge the Consejo Nacional de Ciencia y Tecnología (CONACyT) for its support.


  • [1] Bonneau G, Faraut J and Valent G 2001 Am. J. Phys. 69 322 (Preprint quant-ph/0103153)
  • [2] Znojil M 2001 Phys. Lett. A 285 7
  • [3] Voronov B L, Gitman D M, and Tyutin I V 2006 Self-adjoint differential operators associated with self-adjoint differential expressions Preprint quant-ph/0603187
  • [4] De Bièvre S, and González J A 1993 Semiclassical behaviour of coherent states on the circle Quantization and Coherent States Methods in Physics (Singapore: World Scientific)
  • [5] Kowalski K, Rembieliński J, and Papaloucas L C 1996 J. Phys. A: Math. Gen. 29 4149–4167
  • [6] González J A and del Olmo M A 1998 J. Phys. A: Math. Gen. 31 8841–8857
  • [7] Reed M and Simon B 1972 Methods of Modern Mathematical Physics, Vol. II: Fourier Analysis, Self-Adjointness (Academic Press)
  • [8] Ali S T , Englis̆ M, and Gazeau J P 2004 J. Phys. A: Math. Gen. 23 6067
  • [9] Antoine J P, Gazeau J P, Monceau P, Klauder J R, and Penson K 2001 J. Math. Phys. 42 2349-2387
  • [10] Frank A and Lemus R 2003 J. Phys. A: Math. Gen. 36 4901
  • [11] Garidi T, Gazeau J P, Huguet E, Lachièze Rey M and Renaud J 2004 Symmetry in Physics. In memory of Robert T. Sharp 2002 ed P Winternitz et al. (Montréal: CRM Proc. and Lecture Notes)
  • [12] Berezin F A 1975 Comm. Math. Phys. 40 153
  • [13] Klauder J R 1995 Ann. Phys. NY 237 147
  • [14] Gazeau J P and Piechocki W 2004 J. Phys. A : Math. Gen. 37 6977
  • [15] Gazeau J P, Josse-Michaux F-X, and Monceau P 2006 Int. Jour. Phys. B 20 1778
  • [16] Gazeau J P and García de Léon P L 2007 Phys. Lett. A 361 301
  • [17] Gazeau J P, Huguet E, Lachièze Rey M and Renaud J 2007 J. Phys. A: Math. Theor. 40, 10225 (Preprint quant-ph/0610080)
  • [18] Dirac P A M 1964 Lectures in Quantum Mechanics (New York: Yeshiva University)
  • [19] Rabeie A, Huguet E, and Renaud J 2007 Phys. Lett. A in press (Preprint math-ph/07070431)
  • [20] Heller B and Wang M 2004 Posterior distributions on certain parameter spaces obtained by using group theoretic methods adopted from quantum physics, University of Chicago, Department of Statistics, Technical Report Series 546 26
  • [21] Heller B and Wang M 2006 Group invariant inferred distributions via noncommutative probability Recent Developments in Nonparametric Inference and Probability (IMS Lecture Notes-Monograph Series) 50 pp 1–19
  • [22] Bryant G W 1987 Phys. Rev. Lett. 59 1140
  • [23] Thilagam A and Lohe M A 2005 Physica E 25 625
Figure 1: Phase trajectory of the particle in the infinite square-well.
Figure 2: Eigenvalues of , and for increasing values of the characteristic momentum of the system, and computed for approximation matrices. Units have been chosen such that , so that and . Note that for with small, the eigenvalues adjust to the classical mean value . The spectrum of is independent of as is shown in (43). For the commutator, the values are purely imaginary.
Figure 3: Continued from figure 2: approximation matrices eigenvalues of , and for increasingly larger values of in units , . The spectrum of is independent of as is shown in (43). For the commutator, the eigenvalues are purely imaginary and tend to accumulate around and as increases.
Figure 4: The lower symbol depicted for various values of in units , . Note the way the mean value fits the function when is large, and approaches the classical average in the well for low values of the parameter.
Figure 5: The lower symbol depicted for various values of in units , . The function becomes smoother when is large.
Figure 6: Variance of depicted for various values of in units , . Note how different dispersions are revealed just by changing the width of the Gaussian function of the variable. Low dispersion, close to classical, is found for near and the quantum behaviour is recovered at large values of the parameter.
Figure 7: Variance of depicted for various values of in units , . Consistently with , a well localized momentum is found for low values of the parameter. This is actually expected since the Gaussian becomes very narrow.
Figure 8: Product for various values of in units , . Note the modification of the vertical scale from one picture to another. Again, the pair position-momentum tends to decorrelate at low values of the parameter, like they should do in the classical limit. On the other hand it approaches the usual quantum-conjugate pair at high values of .
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description