8.
04 Quantum Physics
Lecture XVII
For the single -potential we have exactly one bound state (symmetric state), for the double -potential we always have one symmetric bound state, and we may have (depending on the potential strength) also an antisymmetric bound state. For the nite-size potential well we may have several (but always a nite number) of bound states.
Bound states in potential well
Figure I: Solutions in dierent regions for bound states in a potential well. iqx + Ce iqx , we have Here, instead of writing the solutions as exponentials, Be already written them in a form that reects the symmetry of the potential. We 1 du match u at x = a: dx For even solutions: C = 0 q sin(qa) ea = cos(qa) ea = q tan(qa) (17-1) (17-2)
For odd solutions: D = 0 q cos(qa) = sin(qa) = q cot(qa) Massachusetts Institute of Technology (17-3) (17-4) XVII-1
8.04 Quantum Physics
Lecture XVII
Even solutions
Let us introduce y = qa, =
2m V a2 h2 0
a =
2ma2 2 |E | h 2ma2 2ma2 = V (V0 |E |) 0 2 h 2 h = q 2 a2 = y2
(17-5) (17-6) (17-7) (17-8)
Figure II: Graphic solution of the eigenvalue equation (17-2) for symmetric bound states.
m There is always at least one solution, more if = 2 V a2 is larger (potential deeper h2 0 and/or wider). For 1, the lowest energy solutions are approximately located at h2 q n 2 h2 2 1 2 y = qa = n + 1 , or V | E | = = n + , similar to innite well. 0 n 2 2 2m 2ma 2 The existence of at least one bound state is typical of 1D problems, but not of 3D problems that behave more like odd solutions.
Odd solutions
y = cot(y ) = tan +y y 2 (17-9)
The looks similar to the previous plot, but with shifted RHS. For large , the solutions 2 2 2 0a are qn a = n . For small , a solution exists only if 0 or 2mV 4 . 2 2 h Massachusetts Institute of Technology XVII-2
8.04 Quantum Physics
Lecture XVII
Figure III: Graphic solution of the eigenvalue equation (17-4) for antisymmetric bound states.
Figure IV: Graphic construction of an odd-state solution, or of a solution in 3D, where the wavefunction must vanish at the origin. Condition for the existence of odd solutions. In 3D, we will require that a (modied) wavefunction vanishes at the origin, therefore the solutions will look like odd-parity solutions. (It is as if the wavefunction were continued at r.) Odd solutions do not always exist because the wavefunction needs to bend around suciently to match a decaying exponential, this requires high KE.
Figure V: If the well is not deep enough, the odd solution cannot bend down suciently to match (with continuous slope) a decaying exponential at the edge of the well.
Massachusetts Institute of Technology
XVII-3