huygens-principlelaseropticsvisible-lightwaves
According to Huygens wave theory, every point on a wavefront acts as a secondary source of waves.
Using this principle we can never have pretty narrow parallel beams of light right? Like lasers? There will always be "leakage"(wrong terminology?) of waves from the edge of the beam right?
No, because all the off-rectilinear parts are destructively interfered away.
Huygen's principle is often stated differently than how Huygens stated it. He said that if you draw a sphere at every point of the wavefront, the future wavefront is the envelope of the spheres so drawn, and only one of the envelopes (the one that keeps going forward). So from your sphere, you are supposed to imagine drawing a bunch of other spheres, and then looking at the outermost limit of where they reach. This is a larger sphere, of radius ct, and this is the new wavefront.
The modern version of Huygen's principle (I learned from this site) is the fact that the Green's function of the wave equation in 3+1 dimension is a delta-function on the light cone. This is a way of saying that the full propagation of the wave is by superposing the circles from propagating each point on the previous front. The superposition though is cancelling away from the region where the nearby circles have a mutual tangent, and this is the envelope criterion.
Basically Huygens Principle is just a way that was envisioned in the 17th century to describe how wave behaves and is understood as in the following picture. Each point on the wavefront radiates spherical waves which interfere to preserve it during propagation.
This picture is inspired from Rayleigh Scattering where the emitters are actually all the electrons of the atoms in a material which react to the electric field.
Though Huygens principle is not completely correct, in particular one needs to add an obliquity factor to prevent backward propagation. Moreover for large wavelength on small aperture the Huygens principle is definitely no longer valid.
To sum up I would say that Huygens principle is a convenient way to describe wave propagation and particularly diffraction which happens to be adequate in many cases. However this principle is not fundamentally right.
Best Answer
This is wrong. A wide region (width $\gg\lambda$) of many, evenly distributed in-phase spherical wave emitters, like the Huygens equivalent of a laser beam's cross section, behaves like a phased antenna array, with the result that destructive interference between the emitters cancels radiation deviating significantly from the beam and reinforces radiation along the beam. So you do indeed get arbitrarily narrow beams from Huygens's theory.
What you might be thinking is that this theory does not cope well with a unidirectionally propagating beam. If we replace the laser beam's cross section in-phase with spherical emitters, a narrow beam is radiated both forwards and backwards (a phased antenna array does this, for example). This is where one must call on an obliquity factor to restore unidirectionality. Fresnel-Huygens theory multiplies the spherical radiation pattern by $\frac{1}{2}(1+\cos\theta)$, where $\theta$ is the angle between the direction of propagation and the line joining the centre of the Huygens wavefront and the point in question.
If you look at the Huygens-Fresnel diffraction integral in the Wikipedia link above, then consider the beam cross-section $B$ in the $x-y$ plane and then also a point $P$ with co-ordinates $(x, y, z)$ in the farfield (i.e. so that $R = \sqrt{z^2+y^2+z^2} \gg w$, where $w$ is the maximum width of the beam cross section, then a Huygens superposition of spherical emitters spread over $B$ will beget a disturbance at $P$ of:
$$\int_B \frac{\exp\left(i\,k\,\sqrt{\left(x-x^\prime\right)^2+\left(y-y^\prime\right)^2+z^2}\right)}{\sqrt{\left(x-x^\prime\right)^2+\left(y-y^\prime\right)^2+z^2}}\,h\left(x^\prime,y^\prime\right)\,\mathrm{d}x^\prime\mathrm{d}y^\prime\approx $$ $$\frac{e^{i\,k\,R}}{R} \int_B \exp\left(-i \frac{k}{R}\left(x\,x^\prime+y\,y^\prime\right)\right)\,h\left(x^\prime,y^\prime\right)\,\mathrm{d}x^\prime\mathrm{d}y^\prime$$
where $h\left(x^\prime,y^\prime\right)$ represents any beam apodisation and aberration (intensity variation and phase, respectively) and I have simply pulled the denominator out of the integral: this is justified because, as a proportion of itself, $\sqrt{\left(x-x^\prime\right)^2+\left(y-y^\prime\right)^2+z^2}$ does not vary much for $\left(x^\prime,y^\prime\right)\in B$ as $R\rightarrow\infty$, but, as a number of wavelengths it varies a great deal, so we must keep the numerator. Thus we see two results:
This diffraction scheme is called Fraunhoffer diffraction. So the diffracted beam can be made tightly focussed in the farfield if the cross section at $z=0$ is wide. By dint of the Fourier transform, there is an inverse proportionality between the width of the beam at $z=0$ and that in the farfield. This is exactly the behavior of a high quality laser beam - you will always get a dilation proportional to $R$ - the only way to avoid this is to have a plane wave of infinite extent.
To get some intuition for the result, witness that the above theory gives, for the diffraction of the Gaussian beam defined by:
$$h\left(x^\prime,y^\prime\right) = \exp\left(-\frac{{x^\prime}^2+{y^\prime}^2}{2\sigma^2}\right)$$
the diffracted result at the point $P=(x,y,z)$:
$$\frac{e^{i\,k\,R}}{R} \int_B \exp\left(-i \frac{k}{R}\left(x\,x^\prime+y\,y^\prime\right)\right)\,h\left(x^\prime,y^\prime\right)\,\mathrm{d}x^\prime\mathrm{d}y^\prime = \frac{\sigma^2}{R} \exp\left(-\frac{k^2 \sigma ^2 \left(x^2+y^2\right)}{2 R^2}\right)$$