particle Free waves So, we have confirmed the idea that matter behaves like waves, but what is the nature of these waves, and how do we describe them. Let's begin with a free particle, that is, a particle moving in one dimension with no force acting on it, therefore with no potential changes. We start by writing a function that will describe the dynamics of the wave, what we will henceforth refer to as the wavefunction:
( x , t ) = sin (kx - t )
This function describes an oscillating wave with wavelength = 2/k and frequency = /2. Let us try to extract the dynamics of the particle from this function. There are zeros of this function whenever (kx-t)=n or whenever xn=n/k+t/k. The velocity of these zeros is w = d/dt(xn) = /k = 2/(2/) = as we expect for the relationship between the velocity, wavelength and frequency. But, let's now go back to the de Broglie relationship between the wavelength and momentum, and the Einstien relationship between the energy and the frequency: p = h/ and E = h. Then = (E/h)(h/p) = E/p = ()mv2/mv = v. This gives us the puzzling result that the group velocity (the velocity of the group of waves) is of the particle velocity!? So, how can we reasonable describe the particle with such a wavefunction? Well, there is another problem with the wavefunction that we initially chose, that is that there is no defined spatial extent. There are an infinite number of zeros, or, if we choose, an infinite number of maxima. So it would be a stretch to call this a particle wavefunction something that intuitively we would expect to describe a localized particle moving with velocity v. We make one step towards creating a wave packet. Let's look at the sum of two wavefunctions, one with a slightly different wavelength and frequency: ( x, t ) = 1 ( x, t ) + 2 ( x, t ) 1 ( x, t ) = sin (kx - t ) 1 ( x, t ) = sin ((k + dk )x - ( + d )t ) using sinA +sinB = 2cos[(A-B)/2]sin[(A+B)/2], then dk d ( x, t ) = 2 cos x- 2 2 dk d = 2 cos x- 2 2 2k + dk 2 + d t sin x - 2 2 t sin (kx - t ) t
where the last step is a simplification from the fact that d<<2 and dk<<2k. Looking at this wavefunction, we can see that it is the same sin function, moderated by a changing
amplitude (cos function). From the figure below we see now larger "packets of waves" with wavelengths 1/dk moving with velocity g = d/dk.
So, now we have two distinct velocities for the wavefunction: 1) the wave velocity w = /k, and 2) the group velocity g = d/dk. If we now return to the de Broglie relationships: = 2E/h and k = 2/ = 2p/h, then d = 2dE/h and dk = 2dp/h. This gives us for the group velocity, g = d/dk = dE/dp. Using E = mv2/2 and p = mv, dE = mvdv and dp =mdv or, dE/dp = mvdv/mdv = v. OK, so now we have that the group velocity for the wavefunction is the same as the particle velocity, but what about localization? We still have an infinite of number these "groups" moving at v. The next step we make is to sum more waves together. The figure below shows seven different waves with different k values (k/2) = 9, 10 ... 15. The amplitude and phase of each wave has been chosen such that all the waves are in phase at x = 0 and that the sum of all the waves = k (shown at the bottom) gives a packet
k
of particular width and shape. Notice that there is still a repeated pattern at the right side of the plot of the sum.
If we were to extend this process using the exact same k-range (i.e. k = 9-15) but with a greater number of k values (say, k = 9, 9.5, 10, 10.5, ... 14.5, 15) one would find that the width of the packet centered around x = 0 would remain the same, but the distance between the repeated patterns would increase, and the amplitude of the oscillations between the repeated patterns would diminish. We could continue this process of subdividing the k-space and adding waves and we will find that as the number of waves in this interval goes to infinity, the number of packets that are left goes to one. Instead of a sum of waves, we could now write this as an integral:
( x, t ) =
-
(k ) sin(kx - t )dk
where (k) now represents the relative amplitude of the waves in the integral. OK, so now we have a wavefunction that is localized (is non-zero over a limited x-range), travels with a group velocity equal to that of a moving particle, v. But we have arrived at this wave description at a price. In order to make the packet spatial dimension increasingly small, we must increase the range of the k-values that we include. Hence, if we want a localized wave packet around some very small x-value a (in the figure below), then we must in...