How to reduce edge diffraction?

I have a simple setup with a diode laser and an object placed in front of it. I want to observe the edge's shadow clearly.

However, I get this very annoying effect of diffraction and interference around the edges of the object(see pic below)

Is there a way to reduce this effect? Spatial coherence is mandatory for my setup, but spectral/temporal isn't.

The two solutions I've thought of are:
(a) using a temporally incoherent source, like an arc lamp (which unfortunately won't be as spatially coherent as a laser) and
(b) somehow randomly scrambling the phase through the beam's section

I'm not good with optics though so I don't know if these will work or how to implement them.
Different approaches with regards to the light source are welcome.

• It is a pity that the diffraction pattern annoys you but it is due to the fact that light is behaving like a wave and therefore is not in the realm of geometrical optics which in theory should give you a sharp edge to the shadow. If you use a white light source then you may also be frustrated because there will be a visible penumbra unless the source is made very "small" or the incident light is parallel. – Farcher Nov 16 '18 at 10:23

The diffraction can be greatly reduced by any of several methods. You haven't mentioned the objective of your experiment, so some of these might not be useful.

First method: You already suggested the first: reduce the temporal coherence of the source. For this, you can use an LED rather than a laser, but an LED typically has rather poor spatial coherence. Alternatively, you can put a rotating diffuser in the system which will reduce both spatial and temporal coherence. To get high spatial coherence from a very bright spatially incoherent source, focus it onto a pinhole. You're then trading brightness for spatial coherence.

Second method: Place a lens between the object and the screen to image the object on the screen. This will almost entirely eliminate the diffraction "ringing" around the edges of the image. A zoom lens will allow you to adjust the size of the image. I'm guessing this might work for you because you've said you need a spatially coherent source, which suggests you just want to obtain a "shadow" that amounts to an enlarged silhouette of the object.

Third method: Select your object to have optically "soft" edges rather than abrupt opaque-transparent transitions. This amounts to apodization. I'm guessing this won't serve your purpose.

• I disagree with 2nd method, or you have omitted something. I did try using a microscope with laser illumination and the interference noise is all over the image. The image pictured was already taken with a lens between the object and the (ccd) screen – Manu de Hanoi Nov 16 '18 at 19:54
• If so, the lens was not positioned correctly to focus the image. Please see the Thin Lens equation: [hyperphysics.phy-astr.gsu.edu/hbase/geoopt/lenseq.html] Of course, if the lens is dirty there will be a lot of diffraction from the dirt. If the object is a razor blade and you want an image that is a few inches across, you probably need a lens a couple of inches wide. – S. McGrew Nov 16 '18 at 20:01
• It seems like I'll just have to use classical optics (ie, a lens to project the object). However I'm disappointed that we don't have light sources with high spatial coherence and no temporal coherence. I suppose they would be very useful. – Vasilis Papadimitriou Nov 19 '18 at 12:26
• Actually there are such sources. For example, a "white light" laser or a supercontinuum laser which emits a nearly continuous broadband spectrum. However, diffraction fringes form in the shadow even when the light source is temporally incoherent. If you let sunlight project through a tiny hole into a dark room, and put a straight edge in the beam, the shadow will have diffraction fringes. See [en.wikipedia.org/wiki/White_light_interferometry]. – S. McGrew Nov 19 '18 at 12:54
• @S.McGrew You're right, I was confused, forgetting that a changing phase is equivalent to a broad spectrum emission (with Fourier analysis etc). However I still don't get why you'd get fringes if you have an arrangement with ie a point source that emits monochromatic light with a different, random phase per unit of angle (ie phase is 35° at dead center, 44° at 1° right, 87° at 2° right etc). Shouldn't the secondary (diffraction) sources be in sync with the incoming wave, hence cancelling out? Or am I missing something? – Vasilis Papadimitriou Nov 20 '18 at 23:08

You can reduce wavelength (blue laser) so that the diffraction pattern will shrink

Vibration of the object may help, you would need to get a time averaged measurement. Or can you do a mathematical correction based on observing the maxima and using a formula similar to single slit diffraction.