Measuring beam power from multi-kilowatt lasers

by Keng H. Leong

Laser processing is widely used in diverse manufacturing processes including aerospace and automotive components, consumer and healthcare products, and even frames for windows. Different laser-specific processes are utilized: cladding, cutting, drilling, marking, and welding. For each process, beam power and intensity are critical parameters to ensure the quality of the process; consequently, laser beam power is the primary parameter that is monitored or controlled. Other beam parameters such as beam mode or quality may also be monitored as they affect the spot size at the point of application. The beam intensity is the ratio of the power to the area of the beam spot. Industrial laser systems provide a continuous readout of the beam power for continuous wave (cw) outputs. The laser system usually samples a small fraction of the total beam power for monitoring whereas many users prefer to sample the beam near the point of use.

Thermopile-based measurement

The first commercial power and energy meter for laser beams was marketed by Scientech (Boulder, CO) in the late 1960s. Scientech was founded by Dr. Robert Zimmerer who was associated with the National Bureau of Standards (now known as National Institute of Standards and Technology, or NIST). The first models arose from the US Army's needs to determine the power of their lasers and were limited to a few watts. These meters were based on prototypes developed at the National Bureau of Standards using the thermopile principle -- the temperature difference across the thermopile junctions, between the hot side exposed to the beam and the cooled side, results in the generation of a voltage signal that is proportional to the absorbed power. A water-cooled version for 5 kW beams was offered in the 1980s. FIGURE 1 shows the system that uses a water-cooled tube to absorb the beam power. The inlet water and the outlet water pass through a thermopile detector and the generated voltage was interpreted as the power absorbed.

FIGURE 1: 5 kW power probe system from Scientech.

Calorimetric method

Another pioneer in laser power measurements is John Macken of Macken Instruments who provided simple and inexpensive probes starting in the mid 1970s. These probes are metal pucks with an absorbing surface and an analog gage on one end (FIGURE 2). The principle of measurement is calorimetry similar to the Scientech method. However, no water cooling is used and the temperature increase of the puck after a timed exposure to the beam was interpreted by the gage in terms of beam power. The coating absorptivity varies with the beam wavelength and slightly different exposure times are used for CO₂ and Nd:YAG beams. Laser powers up to 10 kW can be easily measured. One limitation is that beam intensities >1 kW/cm2 would burn off the absorptive coating. The 10 kW conical probe (3 in diameter) shown in FIGURE 2 provides a larger exposed surface to the beam than a flat probe, increasing this intensity limit. The early power probes were all based on calorimetric methods.

Radiometric principle

In the 1970s, Walter Spawr of Spawr Industries used a radiometric principle to build water-cooled 100 kW probes and beam dumps for the Wright Patterson Air Force Base's 100 kW CO₂ laser. The probe was based on an integrating sphere with a sensor sampling a small portion of the total input. The probe shown in FIGURE 3 is approximately 17 in. high and much bulkier than the simple puck but still portable. Inlet and outlet connections are for cooling water and allow continuous operation as opposed to the limited time sample for the puck type probe.

In the late 1980s, several other companies started offering water-cooled puck-like probes based on the thermopile principle. An example is shown in FIGURE 4.

FIGURE 4: 10 kW water-cooled thermopile probe. (Courtesy of Ophir)

In the 1970s and '80s, accurate calibration of such high-power probes was not generally available. However, based on calorimetry (the temperature rise of the puck or cooling water) and the absorption coefficient of the surface coating, reasonable values of ±5% can be expected. Low powerthermopile probes (initially developed at NIST) had NIST-traceable calibrations with better accuracy of ±3%. For high power, the NIST calibrations at low power are extrapolated to higher powers based on electrical and thermal principles. The resulting accuracy is reduced to ±5%.

Evaluating today's high-power probes

In the early 1990s, Argonne National Laboratory carried out an evaluation of different types of high-power probes available using a 6 kW CO₂ laser.¹ At power levels below 2 kW, all the different types of probe discussed above were found to exhibit excellent linearity in response. At higher powers, the Scientech and Spawr probes maintained their linearity, but the puck and thermopile probes tend to deviate from linearity. Although linearity was excellent, the slope of the response was significantly different such that disagreements at lower power levels exceeded the variability in the stated accuracy of ±5%. In addition, the water-cooled surface absorber thermopile probes were slightly sensitive to beam mode and position of the beam on the absorber. Water-cooled probes were sensitive to coolant flow rates with higher flow rates required for optimal response at higher power levels. Optimal response times varied from 5-20s.

Today, several companies offer high-power probes. The table below contains a selection of manufacturers listed alphabetically with the standard power handling range (higher power levels are also available on a custom order basis).

It is obvious that current probes available more than adequately cover the industrial range of interest. Software and electronic integration are much improved from the early days and digital readouts are available for the puck type probes. The use of ceramic components in the coatings has extended the intensity limit to 10 kW/cm². However, accuracies are still ±5% and all calibrations, though NIST-traceable, are still based on thermal/electrical principles and/or carried out at laser powers below 1 kW. Note that at 10 kW, there can be a discrepancy of up to 1 kW within the accuracy stated. Variation in the coating absorptivity with temperature and nonlinearity of the thermopile response at higher powers are the main contributors to the uncertainty. A NIST-traceable calibration based on the actual measurement of beam power for the full range of the probe (which may be cost prohibitive) is needed to improve accuracies and consistency across different types of probe.

Selecting a multi-kilowatt laser power probe

For a researcher involved in defining a laser process on a workpiece, high accuracy of power probes will result in precise parameter definition of the process. In practice, is improved accuracy really critical? In many instances, the beam power for a process is set by dialing in the desired output power from the laser system. On most multi-kilowatt lasers, the output beam power is monitored using the leakage from a mirror on a low-power water-cooled thermopile probe. For example, a 100 W probe may be used in a 10 kW laser system. For these lowpower thermopile probes, there is higher accuracy and repeatability among different manufacturers. A recent test (by the author) of thermopile probes from Ophir and Gentec showed agreement to within 2%. Consequently, accounting for any losses in the beam delivery system used, one can expect precise delivery of desired beam power across laser systems. Nevertheless, there is still a need for accurate, repeatable high-power probes to confirm power on the workpiece, and for diagnostics as optics used in the laser system may degrade over time affecting the power output monitored.

Reference

[1] K.H. Leong, D.J. Holdridge, and K.R. Sabo, "Characteristics of power meters for high power CO₂ lasers," J. Laser Applications (1994) v6 n4, 231-236.

Dr. Keng H. Leong ([email protected]) is a laser materials processing consultant and an Editorial Advisor to Industrial Laser Solutions.