Skip to content

Finite Potential Barrier

A finite potential barrier is the standard two-interface scattering problem. It extends the potential step by adding a second boundary, so a wave can reflect many times inside the finite region before emerging to the right or returning to the left.

Use the rectangular barrier

V(x)={0,x<0,V0,0<x<a,0,x>a,V0>0.V(x)= \begin{cases} 0, & x\lt 0,\\ V_0, & 0\lt x\lt a,\\ 0, & x\gt a, \end{cases} \qquad V_0\gt 0.

The same model explains two distinct effects: tunneling when 0<E<V00\lt E\lt V_0, and above-barrier reflection when E>V0E\gt V_0. Both are consequences of wave matching at finite jumps.

For a wave incident from the left, the asymptotic regions have the same potential. Define

k=2mEℏ.k=\frac{\sqrt{2mE}}{\hbar}.

The left and right wavefunctions are

ψI(x)=eikx+re−ikx,x<0,\psi_I(x)=e^{ikx}+r e^{-ikx}, \qquad x\lt 0,

and

ψIII(x)=teikx,x>a.\psi_{III}(x)=t e^{ikx}, \qquad x\gt a.

The incident amplitude has been set to one. The reflection and transmission probabilities are

R=∣r∣2,T=∣t∣2,R=\lvert r\rvert^2, \qquad T=\lvert t\rvert^2,

because the left and right wave numbers are equal. If the asymptotic potentials were different, the velocity factor in Reflection and Transmission Coefficients would be required.

For a finite jump in the potential, both ψ\psi and dψ/dxd\psi/dx are continuous. At the two interfaces,

ψI(0)=ψII(0),ψI′(0)=ψII′(0),\psi_I(0)=\psi_{II}(0), \qquad \psi_I'(0)=\psi_{II}'(0),

and

ψII(a)=ψIII(a),ψII′(a)=ψIII′(a).\psi_{II}(a)=\psi_{III}(a), \qquad \psi_{II}'(a)=\psi_{III}'(a).

These four linear equations determine rr, tt, and the two interior coefficients. The calculation is algebraically longer than the single-step problem, but it uses the same boundary rules.

For E>V0E\gt V_0, define the wave number inside the barrier by

q=2m(E−V0)ℏ.q=\frac{\sqrt{2m(E-V_0)}}{\hbar}.

The interior solution is oscillatory:

ψII(x)=Aeiqx+Be−iqx,0<x<a.\psi_{II}(x)=A e^{iqx}+B e^{-iqx}, \qquad 0\lt x\lt a.

Solving the four matching equations gives the transmission amplitude, with the phase convention used above,

t(E)=e−ikacos⁡(qa)−ik2+q22kqsin⁡(qa).t(E) = \frac{e^{-ika}} {\cos(qa) -i\frac{k^2+q^2}{2kq}\sin(qa)}.

Thus

T(E)=[1+V02sin⁡2(qa)4E(E−V0)]−1.T(E) = \left[ 1+ \frac{V_0^2\sin^2(qa)} {4E(E-V_0)} \right]^{-1}.

The barrier can reflect even when the particle energy is greater than the barrier height. Classically, a particle with E>V0E\gt V_0 crosses the region with reduced speed and then continues. Quantum mechanically, the wave number changes at both interfaces, and the reflected amplitudes can interfere.

When

qa=nπ,n=1,2,3,…,qa=n\pi, \qquad n=1,2,3,\ldots,

the sine term vanishes and T=1T=1. The barrier is then transparent because the interior phase makes the multiple reflected amplitudes cancel on the incident side. This is a single-barrier phase-matching effect; sharper quasi-bound resonance physics is developed in Resonant Transmission.

For

0<E<V0,0\lt E\lt V_0,

define

κ=2m(V0−E)ℏ.\kappa=\frac{\sqrt{2m(V_0-E)}}{\hbar}.

The interior solution is evanescent:

ψII(x)=Aeκx+Be−κx.\psi_{II}(x)=A e^{\kappa x}+B e^{-\kappa x}.

The growing term is not a physical divergence because the barrier has finite width. Both exponentials are needed to satisfy matching at x=0x=0 and x=ax=a.

The transmission amplitude can be written as

t(E)=e−ikacosh⁡(κa)+iκ2−k22kκsinh⁡(κa).t(E) = \frac{e^{-ika}} {\cosh(\kappa a) +i\frac{\kappa^2-k^2}{2k\kappa}\sinh(\kappa a)}.

Therefore

T(E)=[1+V02sinh⁡2(κa)4E(V0−E)]−1.T(E) = \left[ 1+ \frac{V_0^2\sinh^2(\kappa a)} {4E(V_0-E)} \right]^{-1}.

This is the exact rectangular-barrier tunneling formula. Its physical interpretation and opaque-barrier approximation are treated in Rectangular Barrier Tunneling.

If

κa≫1,\kappa a\gg 1,

then

sinh⁡2(κa)≈14e2κa,\sinh^2(\kappa a)\approx \frac14 e^{2\kappa a},

and

T≈16E(V0−E)V02e−2κa.T \approx \frac{16E(V_0-E)}{V_0^2} e^{-2\kappa a}.

The exponential factor is the main qualitative lesson:

T∝e−2κa.T\propto e^{-2\kappa a}.

Increasing the barrier width, the particle mass, or the energy deficit V0−EV_0-E suppresses transmission rapidly.

The formulas above appear singular at E=V0E=V_0, because qq or κ\kappa goes to zero. The limit is finite. Using sin⁡(qa)∼qa\sin(qa)\sim qa or sinh⁡(κa)∼κa\sinh(\kappa a)\sim \kappa a gives

T(E=V0)=[1+mV0a22ℏ2]−1.T(E=V_0) = \left[ 1+\frac{mV_0a^2}{2\hbar^2} \right]^{-1}.

The barrier top is therefore not automatically transparent. The finite width still matters.

For the real time-independent barrier, one-channel current conservation gives

R+T=1.R+T=1.

The transmitted amplitude also contains a phase. The probability TT tells how much flux emerges; the phase of t(E)t(E) controls interference when barriers are combined, and it matters for wave packets and resonant structures.

The Transfer Matrix Method packages this same calculation into interface and propagation matrices. It becomes especially useful for multiple barriers, wells, and layered potentials.

The finite barrier should be read as a wave-interference problem, not as a particle deciding at one point whether to cross. At each interface, the wave-number mismatch generates reflected and transmitted components. Inside the barrier, those components propagate or decay, then meet the second interface. The observed rr and tt are the coherent result of all matching conditions at once.

For E<V0E\lt V_0, the interior is classically forbidden but finite. The evanescent solution can connect the left allowed region to the right allowed region. For E>V0E\gt V_0, the interior is classically allowed, but the changes in wavelength still produce reflection unless the phases cancel it.

  • Treating finite-barrier tunneling as the same problem as a semi-infinite step.
  • Computing TT from an interior coefficient instead of the outgoing current.
  • Forgetting above-barrier reflection.
  • Assuming T=1T=1 whenever E>V0E\gt V_0.
  • Discarding the growing exponential inside a finite forbidden region before applying both boundary conditions.
  • Missing the finite limit at E=V0E=V_0.
  • Confusing single-barrier phase transparency with long-lived double-barrier resonances.
  • D. J. Griffiths and D. F. Schroeter, Introduction to Quantum Mechanics, 3rd ed., Cambridge University Press, 2018.
  • R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994.
  • C. Cohen-Tannoudji, B. Diu, and F. Laloë, Quantum Mechanics, Wiley, 1977.
  • J. R. Taylor, Scattering Theory: The Quantum Theory of Nonrelativistic Collisions, Dover, 2006.
  • E. Merzbacher, Quantum Mechanics, 3rd ed., Wiley, 1998.
  1. For E>V0E\gt V_0, show that the barrier is perfectly transmitting when qa=nπqa=n\pi.
Solution

The above-barrier transmission probability is

T(E)=[1+V02sin⁡2(qa)4E(E−V0)]−1.T(E) = \left[ 1+ \frac{V_0^2\sin^2(qa)} {4E(E-V_0)} \right]^{-1}.

If qa=nπqa=n\pi, then sin⁡(qa)=0\sin(qa)=0. The denominator is 11, so

T=1.T=1.

The result comes from destructive interference among reflected amplitudes, not from the absence of interfaces.

  1. Derive the barrier-top limit of TT from the below-barrier formula.
Solution

For E<V0E\lt V_0,

T(E)=[1+V02sinh⁡2(κa)4E(V0−E)]−1,κ2=2m(V0−E)ℏ2.T(E) = \left[ 1+ \frac{V_0^2\sinh^2(\kappa a)} {4E(V_0-E)} \right]^{-1}, \qquad \kappa^2=\frac{2m(V_0-E)}{\hbar^2}.

As E→V0−E\to V_0^-,

sinh⁡(κa)∼κa.\sinh(\kappa a)\sim \kappa a.

Therefore

V02sinh⁡2(κa)4E(V0−E)∼V02κ2a24E(V0−E)=mV02a22Eℏ2→mV0a22ℏ2.\frac{V_0^2\sinh^2(\kappa a)} {4E(V_0-E)} \sim \frac{V_0^2\kappa^2a^2}{4E(V_0-E)} = \frac{mV_0^2a^2}{2E\hbar^2} \to \frac{mV_0a^2}{2\hbar^2}.

Thus

T(E=V0)=[1+mV0a22ℏ2]−1.T(E=V_0) = \left[ 1+\frac{mV_0a^2}{2\hbar^2} \right]^{-1}.
  1. Use the exact tunneling formula to recover the leading opaque-barrier dependence.
Solution

When κa≫1\kappa a\gg 1,

sinh⁡2(κa)≈14e2κa.\sinh^2(\kappa a)\approx \frac14 e^{2\kappa a}.

The denominator of TT is then dominated by

V024E(V0−E)14e2κa=V0216E(V0−E)e2κa.\frac{V_0^2}{4E(V_0-E)} \frac14 e^{2\kappa a} = \frac{V_0^2}{16E(V_0-E)}e^{2\kappa a}.

Taking the reciprocal gives

T≈16E(V0−E)V02e−2κa.T \approx \frac{16E(V_0-E)}{V_0^2}e^{-2\kappa a}.