2-4 The Particle in an Infinite “Box” with

Chapter 2 Quantum Mechanics of Some Simple Systems
2-4 The Particle in an Infinite “Box” witha Finite Central Barrier
Another example of barrier penetration in a stationary state of a system is provided by inserting a barrier of finite height and thickness at the midpoint of the infinite square well of Section 2-1 (see Fig. 2-12).
The boundary conditions for this problem are easily obtained by obvious extensions of the considerations already discussed. Rather than solve this case directly, we shall make use of our insights from previous systems to deduce the main characteristics of the solutions. Let us begin by considering the case where the barrier is infinitely high.
Figure 2-12 (a) Solutions for identical infinite square wells. (b) Effect of finite partition on half waves. (c) Symmetric combination of half waves. (d) Antisymmetric combination of half waves.
Section 2-4 The Particle in an Infinite “Box” with a Finite Central Barrier
Then the problem becomes merely that of two isolated infinite square wells, each well having solutions as described in Sections 2-1 and 2-2.
Now, as the height of the barrier is lowered from infinity, what happens? The levels lying deepest in the two sections should be least affected by the change. They must still vanish at the outer walls but now they can penetrate slightly into the finite barrier.
Thus, the lowest state in, say, the left-hand section of the well will begin to look as given in Fig. 2-12b. The solution on the right side will do likewise, of course. As this happens, their energies will decrease slightly since their wavelengths increase.
However, since the two wells are no longer separated by an infinite barrier, they are no longer independent. We can no longer talk about separate solutions for the two halves.
Each solution for the Schr¨odinger equation is now a solution for the whole system from x = −L to +L. Furthermore, symmetry arguments state that, since the hamiltonian for this problem is symmetric for reflection through x = 0, the solutions, if nondegenerate, must be either symmetric or antisymmetric through x = 0.
This requirement must be reconciled with the barrier-penetration behavior indicated by Fig. 2-12b, which is also occurring. One way to accomplish this is by summing the two half waves as shown in Fig. 2-12c, giving a symmetric wavefunction. Alternatively, subtraction gives the antisymmetric form shown in Fig. 2-12d. Both of these solutions will be lower in energy than their infinite-well counterparts, because the wavelengths in Fig. 2-12c and d continue to be longer than in 2-12a. Will their energies be equal to each other? Not quite. By close inspection, we can figure out which solution will have the lower energy.
In Figs. 2-12b to 2-12d, the slopes of the half wave, the symmetric, and the antisym metric combinations at the finite barrier are labeled respectively m, m, and m. What can we say about their relative values? The slope m should be less negative than m because the decaying exponential producing m has an increasing exponential added to it when producing m. Slope m should be more negative than m since the decaying exponential has an increasing exponential subtracted from it in case d, causing it to decay faster. This means that the sine curve on the left-hand side of Fig. 2-12c cannot be identical with that on the left side of Fig. 2-12d since they must arrive at the barrier with different slopes. (The same is true for the right-hand sides, of course.) How can we make the sine wave arrive with a less negative slope m?—by increasing the wavelength slightly so that not quite so much of the sine wave fits into the left well (see Fig. 2-13a). Increasing the wavelength slightly means, by de Broglie’s relation, that the energy of the particle is decreased. Similarly, the sine curve in Fig. 2-12d must be shortened so that it will arrive at the barrier with slope m, which corresponds to an energy increase. Of course, now that the energy has changed outside the barrier, it must change inside the barrier too. This would require going back and modifying the exponentials inside the barrier. But the first step is sufficient to indicate the qualitative results: The symmetric solution has lower energy. In Fig. 2-13a is a detailed sketch of the final solution for the two lowest states.
There is a simpler way to decide that the symmetric solution has lower energy. As the barrier height becomes lower and lower, the two solutions become more and more separated in energy, but they always remain symmetric or antisymmetric with respect to reflection since the hamiltonian always has reflection symmetry. In the limit when the barrier completely disappears we have a simple square well again (but larger), the lowest solution of which is symmetric. (See Fig. 2-13b.) This lowest symmetric solution must

Chapter 2 Quantum Mechanics of Some Simple Systems
Figure 2-13 (a) Detailed sketch of the two lowest solutions for the infinite square-well divided by a finite barrier at the midpoint. The waves are sketched from a common energy value for ease of comparison. Actually, the symmetric wave has a lower energy. (b) A correlation diagram relating energies when the barrier is infinite (left side) with those when the barrier vanishes. Letters A and S refer to antisymmetric and symmetric solutions, respectively.
“come from” the symmetric combination of smaller-well wavefunctions sketched at the left of Fig. 2-13b; similarly, the second lowest, antisymmetric solution of the large well correlates with the antisymmetric small-well combination (also at left in Fig. 2-13b).
A figure of the kind shown in Fig. 2-13b is called a correlation diagram. It shows how the energy eigenvalues change throughout a continuous, symmetry-conserving process.