Diffraction of Light

When light passes through an opening it is observed to spread out.  This is known as diffraction and becomes more pronounced with narrower openings.

Diffraction of light through wide and narrow openings

A similar effect also occurs when light hits an opaque edge, causing the light to bend around the edge.

Diffraction of light around an edge

The fact that this happens is to be expected.  Light is known to be a wave and waves, such as water waves and sound waves, are also known to expand through openings.

Water and sound waves are disturbances in flexible mediums.  Such waves spread out because they contain areas of pressure within the medium, and regions of high pressure (wave-peaks) always expand into regions of lesser pressure.  But as far as we know, light doesn’t travel though a medium so how can it spread out?

Huygens-Fresnel principle

The explanation of how light can spread is given by the Huygens-Fresnel principle.  It states that every point on a wavefront can be considered a source for another wavefront which will then become a source for further wavefronts as shown:

Here we see a spherical wavefront generating a series of wavelets.  Each wavelet sits like a soap bubble atop a primary soap bubble.  The side parts of the wavelets are said to combine and cancel each other, leaving behind only the front part of the wave.

Problems

This principle is used to predict what kinds of patterns will occur when light passes through different diffraction gratings such as single and double slits.  And while nobody denies that it does a good job of predicting, there are many problems in believing that this represents a true model of what actually happens with light.

First, why would the wavelets be hemispheres rather than full spheres?  Shouldn’t nature produce full spheres rather than half-spheres with edges?

Second, what happens to the original wavefront?  Does it stop its forward motion, then generate the wavelets, and then cease to exist?  And if it doesn’t stop its motion, wouldn’t its velocity (of c) be added to the wavelets making them move at 2c?  According to special relativity theory, light moves only at c, not more or less.

Third, why would the wavelets cancel each other on the side?  There is no particular reason why they should as they are not ‘out of phase’ with each other.  This cancellation is just an assumption required by the theory.

Fourth, if hemispherical wavelets were being produced, this would allow the light to eventually turn back on itself as shown:

Here we see a primary wavefront, which generates a secondary wavelet, which generates a third/fourth/fifth, etc.  Until finally we have a wave that flows backward.  All light passing through empty space should diffract and spread in all directions.  Instead it diffracts only when interacting with an opaque material.

The Edges

So how can diffraction work without a flexible medium, and why doesn’t light diffract in empty space, as the Huygens-Fresnel principle would suggest?

There is something being overlooked here.  And that is the edges of the slit.

As light hits the slit, most of it either gets absorbed into the walls or passes through the opening.  But a small amount also hits the edges.  What will happen to it?

If the material is reflective, light will bounce off in some direction.  Mostly though, the slit will be made of a dark material.  In that case the light might be reradiated from that point via a different process, as described later.  This diagram shows the idea:

Here we see light passing directly through the opening and a small amount reradiated from the edges.  The radiation from the edges will come out spherically and combine with the light passing straight through.  As a result it will produce interference patterns on a screen.

Single slit patterns

As evidence for this idea, consider this typical interference pattern produced by light passing through a single slit:

This pattern also fits with a prediction from the Huygens-Fresnel principle.

But there’s a major problem: a single slit shouldn’t produce an interference pattern.

When a wave passes through a single slit, all that should happen is that the sides of the flat wavefront that has been cut at either end will expand in quarter circles as shown:

Since no waves crossover and interfere with each other, there’s no reason for any interference pattern to emerge.

This photograph shows what happens when water waves pass through a single slit:

This was taken in a ripple tank.  Plane waves come in from the left, pass through an opening, and then emerge in a shape similar to the previous diagram.

Notice there’s no interference pattern!

That fact that light produces an interference pattern indicates we are dealing with multiple ‘sources’ and the additional light sources are likely to be the slit’s edges.

Double slit patterns

What about double slits?  Here’s a typical interference pattern produced by light passing through two slits:

We have what looks like a large-scale pattern overlayed with a small-scale pattern.  This again fits with a prediction from the Huygens-Fresnel principle, but again doesn’t match a ‘normal’ wave.

A wave passing through two slits should produce two widened semicircles like this:

When those semicircles overlap they will produce interference but the gap between interference peaks won’t be much different to the wavelength of the original wave.

A double slit water wave pattern looks like this:

The ‘screen’ shows only a large-scale interference pattern with no secondary small-scale patterns on top.  Given the shape of the water emerging through the slits, there’s no conceivable way that a secondary pattern could appear.

So how can light produce a pattern within a pattern?  The reason is likely that we are dealing with four edges that generate four radiation points as shown:

The pair of points belonging to each slit are close together and these are what would be responsible for the small-scale pattern within the large-scale pattern.

The details

Let’s look now into the details of how the light-bending might occur.  Earlier it was suggested that light was being absorbed and reradiated from the edges.  But this seems unlikely because the amount of radiation would depend on how dark the material was, with fully dark materials (zero albedo) producing almost no diffraction – and this is not what we observe.
 
Understanding how light interacts with edges requires understanding how light can pass though solid materials like glass.  We don’t exactly understand this process, other than to say that light is somehow being absorbed by atoms on one side and reemitted on the other.  This absorption/reemission process is going to require interaction with the electrons surrounding the atoms.

When the atoms are within a solid material, the electric fields from the electrons are going to be heavily overlapping each other on all sides.  If the material is transparent, this allows light to move directly though it in straight lines.

When the atoms are on the surface of a solid material, the fields from the electrons are going to be heavily overlapping each other on all sides but one.  If the material is transparent and the light was coming from within at a low angle, this would allow the light to be internally reflected, i.e. total internal reflection.

We don’t understand how light can be reflected in that manner but presumably it has something to do with how the fields overlap differently at the surface.

Now what would happen if a group of atoms on the surface of a material suddenly came to an end?

The above diagram shows light skimming across the surface of a material.  The light is moving within the material, but though the ‘cloud’ of electrons belonging to the atoms in the top-most layer of the material.  We will assume light is able do this even if the material is opaque as it is not actually passing though the material, i.e. it is not passing between the nuclei in the material.

Since the electron fields overlap each other, the light jumps from one atom to the next.  For the atom on the far right however, which corresponds to the edge of the material, its field overlaps only on its left side.  How would light behave when it emerges on the other side of that atom?  We don’t know, but it’s possible that when light encounters this situation it radiates spherically as shown:

Such spherical radiation would then produce diffraction patterns if combined with other waves.

Conclusions

The way light diffracts is inconsistent with how other waveforms such as water waves diffract.  The reason why light produces interference patterns is not due to the Huygens-Fresnel principle but more likely has to do with how light interacts with the electric fields of atoms on the edges of diffraction gratings.