[web] [lect]

All scattering techniques have in common that a beam of radiation interacts with the structure of a sample.
The beam can consist of electromagnetic radiation (photons, including x-rays) or quantum-mechanical particle
waves such as neutrons or electrons. The
*incident beam*

is characterised by its
*energy*, $E_0$, in J,

and its
*wave vector*, $\vec{k}_0$, in $\mathrm{m^{-1}}$.

Following the scattering event, the energy of the scattered wave is $E_1$ and its wavevector $\vec{k}_1$. The
*momentum transfer*, $\vec{q}=\vec{k}_1-\vec{k}_0$, in $\mathrm{m^{-1}}$,

is the vector difference between the two wave vectors. The name comes from the fact that the momentum of a particle
wave is proportional to its wave number, $p=\hbar k$. During any scattering event,
*momentum conservation*
$$\hbar\vec{q}=\hbar(\vec{k_1}-\vec{k_0})$$
as well as
*energy conservation*
$$\hbar\omega=E_1-E_0$$
have to be maintained, *i.e.* any energy or momentum gained or lost by the wave must have been exchanged with
the structure of the scattering medium. During
*elastic scattering*, $|\vec{k_1}|=|\vec{k_0}|$,

no energy is transferred, and the wave number of the scattered wave and the incident wave are identical. Therefore,
the momentum transfer reflects only the change in direction of the wave due to the scattering event (scattering angle).
In
*inelastic scattering*

on the other hand, both energy and momentum are transferred, and the length as well as the direction of the scattered
wave vector are different from the incident wave vector.

While both electromagnetic and particle waves can be used for scattering experiments,
there is an important difference in the relationship between wavelength and energy for both types
because particles have a rest mass. For
*x-rays* (photons)

the wavelength is simply the ratio of the speed of light and its frequency, *i.e.*
$$\lambda_{\mathrm{photon}}=\frac{c}{\nu}=\frac{hc}{E}\qquad.$$
For
*neutrons*

on the other hand, the de Broglie wavelength is calculated from the momentum, which
includes the rest mass of the particle:
$$\lambda_{\mathrm{neutron}}=\frac{h}{m_nv}=\frac{h}{\sqrt{2Em_n}}\qquad.$$
As a result, the slope of the line representing the wavelength as a function of energy in a
logarithmic plot is twice as steep for photons as it is for particle waves. The relationship
for electrons is the same as for neutrons, but since their rest mass is only about one two-thousandth
of the neutron rest mass, the same energy produces as wavelength about forty times as large.

*Diffraction*

experiments probe structure of the order of the wavelength of the probe ray, while
*diffuse scattering*

probes structures significantly larger than the wavelength. Therefore, x-rays can be
used for diffraction at atomic lattice planes or to probe nano-structure by diffuse scattering.
Visible light can be diffracted at periodic structures on a sub-micron lengthscale, as is
evident in opalescent materials. Since the wavelength of neutrons can be adjusted via
their velocity, they, like x-rays, can diffract at atomic structures or scatter diffusely
at nano-structures. Electron diffraction is rarely used as a stand-alone technique but
is sometimes used in conjunction with transmission electron microscopy, using the same
instrument but a different optical configuration.

The
*scattered intensity*

recorded by a detector in a scattering experiment is the product of five factors,
$$\Delta I=\frac{\partial^2\sigma}{\partial\Omega\partial E}NI_0\Delta E\Delta\Omega\qquad:$$
The
*energy* of the scattered wave, $E$,

the
*solid angle*, $\Omega$,

captured by the detector, the
*incident intensity*, $I_0$,

coming from the source, the
*number of scattering objects*, $N$,

in the sample, and the
*scattering cross section*, $\sigma$.

The latter describes how likely a scattering process is, given all the ingredients listed above
are present, *i.e.* the efficiency of the scattering process. A particular scattering
experiment will be using monochromatic radiation and an angle-resolving detector; therefore we need to consider how
the cross section changes with these parameters, *i.e.* use the
*differential scattering cross section*, $\frac{\partial^2\sigma}{\partial\Omega\partial E}$.

Image © D Jacobson

There are two contributions to this quantity:
$$\frac{\partial^2\sigma}{\partial\Omega\partial E}=\left(r_0^2|f(q)|^2\right)S(\vec{q},\omega)\qquad,$$
one reflecting the efficiency of the individual scattering elements and the other their
spatial arrangement.
For x-rays, the
*atomic form factor*, $f(q)$,

represents the strength of the interaction between an incident wave and the electron cloud
of an atom in the sample. It is multiplied by the classical electron radius, $r_0$. Because of
the shape of the electron cloud (which is, in turn, determined
by the shape of the occupied electronic states of the atom), the atomic form factor depends on the scattering angle
and on the number of electrons in the atom. In the limit of zero angle (zero momentum transfer), the
atomic form factor becomes simply the atom number in the periodic table,
$$f(0)=Z\qquad.$$
Since neutrons interact with the nuclei rather than with the electron cloud of atoms, the
*scattering length*, $b$,

is used
for neutrons

instead of the atomic form factor term, $r_0|f(q)|$.
There is no relationship between the position of an element in the periodic table and its neutron
scattering length. In fact, protons are the most effective scatterers of all nuclides. Since the
scattering length is a property of the nucleus rather than the atom, it is generally different for
different
*isotopes*

of the same element. It is even possible for scattering lengths to be negative.
Because of these differences, scattering cross sections for x-rays and neutrons can be very different
for different samples, or even for different structures in the same sample.

The second component of the differential scattering cross section is the
*structure factor*, $S(\vec{q},\omega)$.

Determining this quantity is the goal of any scattering experiments, because its
double *Fourier transform* in space and time

is the
*correlation function*, $G(\vec{r},t)$,

giving the spatial distribution of whatever feature it is that causes the wave to scatter -
atoms in the case of diffraction, nano-structures such as grains or layers in the case of
diffuse scattering.

Because the incident beam is
*collimated*

by passing through a system of slits, the rays
contained in it are all essentially parallel. The
incident
wave is therefore a
*plane wave*,
$${\rm e}^{{\rm i}\vec{k}_0\vec{r}}\qquad.$$
On the other hand, the
scattered
wave is a
*spherical wave*,
$$r_0f(q)\frac{{\rm e}^{{\rm i}\vec{k}_1\vec{r}}}{r}\qquad,$$
whose amplitude is given by the atomic form factor term, $r_0f(q)$, or the
scattering length, $b$, for neutrons.
A detector will see the superposition of these two waves

As we have seen, the scattering of electromagnetic radiation by electron clouds and the scattering
of neutrons by atomic nuclei can be treated in exactly the same way to reveal the structure of the
sample (via the structure factor, $S(q,\omega)$) even though the interactions causing the scattering
event are very different. We will next look at
x-ray and neutron sources

before moving on to experimental scattering techniques using either source to reveal different aspects
of structure and morphology.