Beam profiler

A beam profiler answers the questions a power meter cannot: how wide the beam is, where its centre sits, whether it is round and whether its profile is Gaussian. Those numbers feed almost every other beam calculation on a bench, from the peak irradiance and the coupling efficiency into a fiber to the divergence and the $M^2$ beam quality factor, which is measured by profiling the beam at several positions through a focus.

Types

Camera profilers place a CCD or CMOS sensor directly in the beam and record the two-dimensional intensity map. Silicon sensors cover roughly 350 to 1100 nm, InGaAs cameras the 900 to 1700 nm telecom range, and pyroelectric or microbolometer arrays the mid and far infrared. They show the full shape, including side lobes, hot spots and diffraction rings, and can capture single pulses. Their limits are the pixel size, a few micrometres, which sets the smallest beam they can measure, about ten pixels across (roughly 50 µm for 5.5 µm pixels); the sensor size, which sets the largest; and saturation, which requires calibrated attenuation ahead of the sensor, usually reflective neutral-density filters or wedge pickoffs so that the attenuator does not distort the beam.

Scanning-slit and knife-edge profilers sweep a slit or edge across the beam in front of a single large detector and reconstruct the profile, or its integral, from the transmitted power. Because the detector is a single photodiode or thermal sensor, they work at high powers, at wavelengths without good cameras, and for beams down to a few micrometres. They measure one-dimensional profiles along the scan axes rather than the full two-dimensional map, so they can miss structure between the axes.

Width definitions

The width a profiler reports depends on the definition chosen. The $1/e^2$ width, where the intensity falls to 13.5% of the peak, is the traditional one and matches the $w$ of Gaussian beam formulas. ISO 11146 defines width by the second moment of the intensity distribution, $d = 4\sigma$, which equals the $1/e^2$ diameter for a Gaussian, applies to any shape, and is what $M^2$ is defined with; it is very sensitive to background noise in the wings, so the standard prescribes background subtraction and an integration area about three times the beam width. Knife-edge profilers usually report the 10–90% or 16–84% clip width, which converts to the $1/e^2$ width for a Gaussian by a known factor.

Good practice

Subtract a background frame taken with the beam blocked; set the exposure so the peak sits near but below saturation; check the result against the known total power; and use the second-moment width for anything that will be compared with a standard or used to compute $M^2$. The full procedure for $M^2$ is in the M² measurement article.

References: ISO 11146-1:2021, Lasers and laser-related equipment: Test methods for laser beam widths, divergence angles and beam propagation ratios; ISO 11146-3:2004, Intrinsic and geometrical laser beam classification, propagation and details of test methods; A. E. Siegman, Proc. SPIE 1224, 2 (1990).