Difracción de rayos X

Relationship between diffracted peak positions and crystal structure:

When light waves of sufficiently small wavelength are incident upon a crystal lattice, they diffract from the lattice points. At certain angles of incidence, the diffracted parallel waves constructively interfere and create detectable peaks in intensity. W.H. Bragg identified the relationship illustrated in Figure 1 and derived a corresponding equation:

nλ = 2d_hkl sin θ [1]

Here λ is the wavelength of the X-rays used, d_hkl is the spacing between a particular set of planes with (hkl) Miller indices*, and θ is the angle of incidence at which a diffraction peak is measured. Finally, n is an integer that represents the 'harmonic order' of the diffraction. At n=1, for example, we have the first harmonic, meaning that the path of X-rays diffracted through the crystal (equivalent to 2d_hkl sin ) is exactly 1λ, while at n=2, the diffracted path is 2λ. We can typically assume n=1, and, in general, n=1 for θ < sin^-1(2λ/d_h'k'l'), where h'k'l' are the Miller indices of the planes that show the first peak (at the lowest 2θ value) in a diffraction experiment. Miller indices are a set of three integers that constitute a notation system for identifying directions and planes within crystals. For directions, the [h k l] Miller indices represent the normalized difference in the respective x, y and z coordinates (in a Cartesian coordinate system) of two points along the direction. For planes, the Miller indices (h k l) of a plane are simply the h k l values of the direction perpendicular to the plane.

In a typical XRD experiment in reflection mode, the X-ray source is fixed in position and the sample is rotated with respect to the X-ray beam over θ. A detector picks up the diffracted beam and has to keep up with the sample rotation by rotating at twice the rate (i.e. for a given sample angle of θ, the detector angle is 2θ). The geometry of the experiment is schematically shown in Figure 1.

Figure 1: Illustration of Bragg's Law.

When a peak in intensity is observed, equation 1 is necessarily satisfied. Consequently, we can calculate d-spacings based on the angles at which these peaks are observed. By calculating the d-spacings of multiple peaks, the crystal class and the crystal structure parameters material sample can be identified using a database such as the Hanawalt Search Manual or database libraries available with the XRD software being used.

We will be assuming that the sample being investigated is not a single crystal. If the sample were a single crystal with a particular (h*k*l*) plane parallel to the sample surface, it would need to be rotated until the Bragg condition for the (h*k*l*) is satisfied in order to see a peak in diffracted intensity (for n=1) with potentially higher harmonic (h*k*l*) peaks (e.g. for n=2) also detectable at higher angles. At all other angles there would be no peaks in a single crystal sample. Instead, let's assume that the sample is either polycrystalline or that it is a powder, with a statistically significant number of crystalline grains or powder particles illuminated by the incident X-ray beam. Under this assumption, the sample consists of randomly oriented grains, with a similar statistical probability for all possible lattice planes to diffract.

The relationships between the d_hkl and the unit cell parameters are shown below in Equations 2-7 for the 7 crystal classes, cubic, tetragonal, hexagonal, rhombohedral, orthorhombic, monoclinic and triclinic. The unit cell parameters consist of lengths of (a,b,c) and the angles between (α, β, γ) the edges of the unit cells for the 7 crystal classes (Figure 1x shows the example of one of the crystal classes: the tetragonal structure where a=b≠c, and α=β=γ=90⁰). Using multiple diffracted peak positions (i.e. several distinct d_hkl values), the values of the unit cell parameters can be solved uniquely.

Figure 2: The tetragonal structure as one of the seven crystal classes.

Cubic (a = b = c; α = β = γ = 90⁰):

  [2]

Tetragonal (a = b ≠ c; α = β = γ = 90⁰):

  [3]

Hexagonal (a = b ≠ c; α = β = 90⁰; γ = 120⁰):

  [4]

Orthorhombic (a ≠ b ≠ c; α = β = γ = 90⁰):

  [5]

Rhombohedral (a = b ≠ c; α = β = γ = 90⁰):

  [6]

Monoclinic (a ≠ b ≠ c; α = γ = 90⁰ ≠ β):

  [7]

Triclinic (a ≠ b ≠ c; α ≠ β ≠ γ ≠ 90⁰): 

        [8] 

Relationship between diffracted peak intensities and crystal structure:

Next we examine the factors that contribute to the intensity in an XRD pattern. The factors can be broken down as 1) the contribution to scattering that results directly from the unique structural aspects of the material (the specific types and locations of scattering atoms in the structure) and 2) those that are not specific to the material. In the former, there are two factors: the 'absorption factor' and the 'structure factor'. The absorption factor primarily depends on the ability of the material to absorb X-rays on their way in and out. This factor does not have a θ dependence as long as the samples are not thin (the sample should be > 3 times thicker than the attenuation length of the X-rays). In other words, the contribution by the absorption factor to the intensity of different peaks is constant. The 'structure factor' directly affects the intensity of specific peaks as a direct result of the structure. The remaining factors, the 'multiplicity', which accounts for all the planes that belong to the same family because they are symmetrically related, and the 'Lorentz-Polarization' factor, which comes from the geometry of the XRD experiment, also affect the relative intensity of the peaks but they are not specific to a material and can easily be accounted for with analytical expressions (i.e. XRD analysis software can remove them with analytical functions).

Figure 3: Three diffraction ray paths, of which rays 11' and 22' satisfy the Bragg condition, while ray 33' results from scattering by an atom (red circle) at an arbitrary position.

As the only factor that carries the unique structural contribution of a material to the relative intensities of XRD peaks, the structure factor is very important and requires a closer look. In Figure 2, let us assume that the 1^st order Bragg diffraction condition (remember, that this corresponds to n=1) is satisfied between ray_11' and ray_22' which are scattered on two atomic planes in the h00 direction (using the Miller indices notation described earlier) separated by a distance d. Under this condition, the difference in path length between ray_11' and ray_22' is δ_(22'-11') = SA + AR = λ. The phase shift between the diffracted rays 1 and 2 is, therefore, Φ_22'-11' = (δ_(22'-11')/λ) 2π = 2π (assuming a cubic symmetry and, therefore, d = a/h in the h00 direction].

With a few steps in analytical geometry, it can be shown that the phase shift, Φ_(33'-11'), with ray 3 diffracted by an arbitrary plane of atoms that are spaced an arbitrary distance x, is given by: Φ_(33'-11') = 2πhu, where u=x/a (a is the unit cell parameter in the (h00) direction.) Taking the two other orthogonal directions, (0k0) and (00l), and v=y/a and w=z/a as fractional coordinates in the y- and z-directions, the expression for the phase shift extends to Φ = 2π(hu+kv+lw). Now, the X-ray wave scattered by the j-th atom in a unit cell will have a scattering amplitude of f_j and a phase of Φ_j, such that the function describing it is . The structure factor we seek, therefore, is the sum of all the scattering functions due to all the unique atoms in a unit cell. This structure factor, F, is given as:

  [9]

and the intensity factor contributed by the structure factor is I = F².

Based on the positions (u,v,w) of atoms on particular planes (h,k,l), there is the possibility of interference between scattered waves that is constructive, destructive, or in-between, and this interference directly affects the amplitude of the XRD peaks representing the (hkl) planes.

Now, a plot of intensity, I, versus 2θ is what is measured in an XRD experiment. The determination of the of crystal type and the associated unit cell parameters (a, b, c, α, β, and γ) can be arrived at analytically by observing systematic presence/absence of peaks, using the equations 2-9, comparing values against databases, using deduction and a process of elimination. Nowadays, this is process is fairly automated by a variety of software linked to crystal structure databases.