Fast high power laser power measurement
Laser optical power measurement is performed by using a sensor to transform the optical power into a measureable current or voltage. The physical principles of the sensor will determine the functionality of the whole instrument. Two kinds of standard sensors are available to measure the power of laser radiation (see table).
A thermopile is a sensor comprising an arrangement of many thermocouples. The individual thermocouples are connected thermally in parallel and electrically in series. For practical sensors, this arrangement is necessary as the thermal sensitivity (V/°C) of a single thermocouple is very low.
The detector surface is coated with a dull, deep black absorbent material. The purpose of the coating is to absorb as much incident laser beam power as possible, independent of the wavelength.
Considering these facts about the construction of a thermopile, certain characteristics are evident:
1. Thermopiles have a fairly low sensitivity to light.
3. The absorbent material is crucial for the measurement. However, this coating fades over time, which leads to loss of calibration.
4. Thermopiles react very slowly as the measurement is based on heat flow. Typical response times vary from one to several seconds.
A photodiode is a semiconductor device designed in such a way that an electric potential gradient exists between its two electrodes (anode and cathode). These two electrodes are electrically contacted via thins filaments which are led out of the device by two pins. Since this structure is mechanically sensitive, the device is enclosed in a metal housing comprising a protective window through which light can enter.
The functional properties of the photodiode are apparent from its construction:
1. The photodiode is very sensitive to light, as a direct quantum transfer of photons to current takes place. Typically the quantum efficiency can be close to 100%. This allows power measurements down to the femtowatt range. On the other hand, the maximum power is limited to a few milliwatts above which the photodiode goes into saturation: the current generated is no longer proportional to the irradiated power.
2. Silicon – the material of choice for visible range measurements – is abundant and cheap. However, germanium and InGaAs which are required for NIR photodiodes are expensive. The available sensors are thus greatly limited in their size.
3. Semiconductors have a high index of refraction which leads to a relatively high reflection of the incoming beam. Since the surface is very flat, the device acts as a mirror to some extent which may pose a safety hazard or be troublesome in a measurement setup.
4. The protective window acts as a weak etalon. This means that, depending on the angle and location of incidence of the light, the photodiode may show a different total sensitivity .
It would seem therefore, that a photodiode is not suitable for high power laser measurements, since the detector area is small (many high power lasers are at NIR wavelengths) and the device can only measure up to mW of power. Thus many practitioners just accept the compromise and use thermopile sensors.
But what if you want a higher resolution of measurement, wide dynamic range and high speed?
The integrating sphere offers a remedy for the above-mentioned deficiencies of the naked photodiodes and the thermopiles. An integrating sphere is a passive component comprising a hollow sphere with openings ("ports") that allow the laser radiation to enter and escape.
The inner surface of the hollow sphere is made of a material with a high degree of reflection for the wavelength range to be measured. The surface is manufactured in such a way that the incident radiation is highly scattered. Consequently, such a structure allows an incident laser beam to be evenly distributed over the entire sphere surface via the multiple, strongly diffuse reflections.
Hollow spheres made from a special polymer are suitable for the wavelength range between 250 nm to 2.5 µm. Barium Sulphate (BaSO4) coated aluminium spheres are somewhat less expensive but they tend to tint yellow over time and are thus unsuitable for precise laser power measurement. For longer radiation in the wavelength range of 700 nm – 20 µm, a gold coating is used on a rough, metal surface. Many high power lasers fall into this spectral range and so solid copper or aluminium is suitable as a good heat-conducting substrate material.
A photodiode built into the wall of an integrating sphere sees only a part of the incident laser power entering into the sphere with the following important changes:
1. The power density is fully homogeneous.
2. The radiation is unpolarized, even if the input radiation was polarized.
3. The power incident upon the sensor is greatly weakened.
We see that the combination of an integrating sphere and a photodiode allows the design of a laser power sensor that responds as fast as a photodiode but can measure considerably more power. By selecting the size of the integrating sphere, the overall sensitivity of the system can be adjusted. In addition, the detector is now independent of the inhomogeneities of the power density and polarization. The detector is also independent of the location and angle of incidence of the laser radiation.
The integrating sphere can be used for relatively large beam diameters as the size of the photodiode itself is not a limiting factor. The power density is also significantly less on the inner wall of the sphere than on an absorbent thermopile because the inner surface of the sphere is at least 20 times larger than the input aperture. The wall material can thus tolerate a higher power density and it does not change significantly over time.
Additional measurement ports can be located in the sphere wall offering further benefits. For example, a fibre-optic port may be used for simultaneous measurement of the spectrum of the laser.
As an example of application, a 100mm inner diameter, gold coated copper integrating sphere with water cooling was used to measure the real time power fluctuations of a 5kW disc laser used for materials processing. Since the integrating sphere becomes warm at these power levels, the photodiode was not installed in the integrating sphere itself. Temperature changes of the photodiode can lead to inaccuracy of the power measurement. Instead, the sphere was equipped with an SMA fibre port which leads to an optical power meter equipped with an SMA-fibre receptacle. The complete system (sphere-fibre-photodiode) was calibrated as a single unit to ensure accurate power measurement. The power meter is USB powered and controlled which limits the cabling required for the measurement (one USB cable and two water lines).
With this setup it was found that the laser power was very stable up to 2500W. However, when the power was increased to the full rating of the laser (5000W), a long term fluctuation of about 1.5% was seen.
In addition, a faster fluctuation of about 0.7% was seen in the output power. Note that this faster fluctuation is on a time scale which would not be measureable with a thermopile detector.
The integrating sphere combined with a photodiode represents a virtually perfect sensor for measuring laser power. For applications with high power lasers, this combination allows the operator to see fluctuations which are too fast for a thermopile detector to measure. This includes fluctuations during CW operation, transients and overshoot on starting the laser and short term power drop-outs during operation.
In addition, since the measurement is virtually independent of beam divergence, integrating spheres can be used for laser-based measurements such as transmission on refracting and scattering objects. For example, the integrating sphere can be used to measure the transmission of laser-weldable plastic materials to determine the optimum operating parameters of the welding laser.