Laser parameters: their measurement and control
Continuously operating or CW lasers are invariably rated by their power output, measured in watts or kilowatts. This rating refers to the power generated at the output mirror or window of the laser and is usually measured by an internal power meter and displayed on a monitor. For some cases, some fraction of the beam is split from the beam delivered to the workpiece and monitored on a power meter, so there is always a display present. In other cases, a power reading can be taken only when the beam is not delivered to the workpiece; that is, the beam either goes to the power meter or to the workpiece, but not to both simultaneously. Note that high power laser power meters are usually thermal; the beam causes something to heat and the rate of heating of the object is related to the energy-incident beam. Because thermal conduction is a relatively slow process, such power meters have a slow response time, usually seconds or longer. These meters would not detect modulation on the beam or start-up transients, which may have a deleterious effect in welding.
It is not the power at the laser, but the power at the position of the workpiece that is of interest in the welding process. Although gold-coated mirrors may have a theoretical reflectivity of 99%, it is usually assumed that there is a 4% loss on each mirror surface. In a multi-axis motion system, there may be a number of mirrors and lenses in the beam path between the laser and the workpiece. Consequently, the power at the position of the workpiece can be considerably lower than the laser output power and should be verified and used in any data records of the process. Similarly, with lasers in which the beam is delivered by fiber optics from the laser to the workpiece, there may be losses in introducing the beam into the fiber and extracting it for focusing on the workpiece.
For welding with pulsed lasers, the relevant parameter is not the power but the energy per pulse, which is similarly diminished between the laser output and the workpiece. The weld is performed by individual pulses. The average power with pulsed lasers is proportional to the pulse repetition rate, for a given laser pulse energy and determines the rate at which spot welds are performed on the workpiece. The dimensions of the weld are determined largely by the energy of individual pulses, however.
Whether using CW or pulsed lasers, however, the power and pulse energy are not as important as the power density, measured in units such as watts per square centimeter, or the energy density, measured in joules per square centimeter. To determine these quantities one needs a method of measuring the size of the beam, as discussed in a later section.
The term mode describes the distribution of laser intensity within the beam. For industrial lasers, the term is short for ‘transverse modes’ since the other type of modes, longitudinal modes, are relevant only to lasers used for precision sensing. The transverse mode, or distribution of intensity in the plane perpendicular to the optic axis, is determined by the nature of the mirrors used in the laser construction. There are four types of laser modes: stable, unstable, waveguide and hybrid stable-unstable.
Any light beam, by its very nature, tends to spread out or ‘diffract’ as it passes through space. A stable mode is formed when the light radiation bouncing back and forth between two mirrors of the laser is refocused when one or both of the mirrors has a curved surface. The refocusing counteracts the tendency of diffraction to spread the beam out and confines the beam near the axis of the two mirrors. One of the mirrors must be partially transmitting, to allow some fraction of the beam to emerge from the laser to perform useful work. The remainder of the beam is reflected back into the laser, is amplified by the medium of the laser to compensate for the amount lost by transmission through the mirror, and is retroreflected by the second mirror. The two retroreflecting mirrors are said to form a ‘resonator’, because the amplified light resonates between them. The stable resonator is one in which the curvature of the mirrors is such that the light is confined to near the axis defined by the two retroreflecting mirrors.
There are a number of transverse modes, or distributions of laser radiation, that can be formed by a stable resonator. The modes are solutions of the mathematical equations which describe the propagation of light, with the boundary conditions established by the two resonator mirrors. The preferred mode is one that is strongest along the axis, with the intensity decaying in a Gaussian fashion with distance away from this axis. This is called the TEM00 mode. The other modes, or solutions of the mathematical equations, can also be realized in practice. High power lasers often operate in a multimode fashion, with a variety of the modes operating simultaneously.
The intensity distribution in the TEM00 beam is circularly symmetric and is given by
A hybrid stable-unstable laser is one which has a stable resonator configuration in one direction (e. g. the x-direction) and an unstable resonator configuration in the y-direction, where the axis of the laser beam is along the z-direction. One method of scaling of gas laser excitation to higher power is to extend the electrodes in the direction of the transverse gas flow; this leads to an excitation region that is rectangular shaped. A hybrid stable-unstable resonator has been investigated as one way of extracting energy from the rectangular excited region. More recently, the diffusion cooled laser has been extended to a ‘slab’ geometry, in which two extended electrodes are placed on either side of the rectangular shaped excited region. In this case, the laser resonator results in waveguide laser modes in the direction between the electrodes and a stable mode in the direction perpendicular to this. Special optical arrangements are used with these lasers to produce a beam of high quality and minimum divergence. They are used for welding applications involving sheet metal and polymers.
The beam width is defined as the diameter of a circle that includes 1 - 1/e2 ~ 85% of the total power of the beam. For the TEM00 beam described above, this definition of beam diameter corresponds to twice the beam radius w.
The quality of a beam is a measure of its ability to be focused to a small spot size, raising the intensity, or power per unit area (or energy per unit area, for pulsed beams) to a high value to do useful work. The quality of the beam is determined by the resonator design and the choice of the retro - reflecting mirrors. The measurement of beam quality is called the M2 factor. Note that this is a different M from the one mentioned above for the magnification of an unstable resonator. European laser practitioners use a K factor that is related to the M2 factor by K = 1/M2. For the lowest order Gaussian mode, M2 = 1. Most laser suppliers list the M2 value for their laser. The M2 can be determined by determining the beam waist wB of the laser, via multiple measurements of the beam radius, and the divergence of the laser. The quality factor of the beam is then given by the ratio of the beam divergence to the value that the beam divergence would have if it was the TEM00 mode; that is,
M2 = eM/(1/nmB) [4.4]
where is the measured value of the half-angle beam divergence.
Why is the beam quality important? A low quality beam diverges more rapidly than a high quality beam, and is focused to a minimum beam radius a factor of M2 larger than a similar low order mode can be focused. This means the intensity at the focus is lower by a factor of M4 than that of a similar TEM00 mode and the ability of the laser to do work is diminished.
The depth of focus, or the range of distance over which the beam maintains a minimum value, is also smaller by a factor of M2. Consequently, the criticality of maintaining focus in a welding operation is more severe with a beam with a higher order mode.
Note that the beam quality is determined by the characteristics of the laser. As the beam propagates through space, if it is enlarged or focused by perfect lenses, its quality remains the same. If the beam passes through imperfect focusing systems, its quality can be worsened. In some cases, beam quality has been improved by focusing the beam through a pinhole and recollimating the transmitted beam. The pinhole absorbs the fringes of the beam that represent power in higher order modes.
The beam diameter at the focus, also called the spot size, determines not only the fineness of the features that can be cut or welded, but also determines the intensity, or power per unit area, at the focus. Laser material interactions are determined by the intensity, hence the focused beam size is a very important parameter.
To the zeroth order approximation, the beam radius is controlled by the quality of the beam, as reflected in the divergence в of the beam prior to the beam striking the lens:
Wf ~ eF [4.5]
Here F is the focal length of the lens or mirror used to concentrate the beam on the workpiece. However, an actual measurement of the power distribution in the region of the focus provides a more direct and reliable source of information. This actual measurement takes into account any aberrations that may be produced by the focusing lens. Since most metalworking lasers have the power to vaporize any known substance, measuring the focused spot size of the laser is a challenging operation. It has been achieved using a variety of commercially available equipment based on light scattering or light collection by a rapidly scanning wire or hollow needle through the focal volume. The spinning needle survives since it does not spend enough time in the focal region to accumulate sufficient heat.
From the signal received, the beam radius can be calculated. If the detection system is moved along the direction of propagation through the focus, the beam size can be found as a function of position. Then the minimum beam waist can be found and the beam divergence at positions away from the beam waist. From this data, the beam quality can be evaluated using Equation [4.2], where z0 = pw02 /1M2 is the parameter that describes the expansion of the beam away from its minimum value.
Using this method, it was found that the output of a Mitsubishi 1.6kW carbon dioxide laser, focused by a 6.3 cm focal length lens, had a minimum radius of 0.33, 0.35 and 0.62mm if the laser was operated at 200, 800 or 1600 W, respectively. The measured beam quality factor was M2 = 2.6, 2.4 and 4.7 at the three power levels, indicating that the beam quality at the workpiece degraded as the power level was increased. The intensity, or power per unit area, was actually higher if the laser was operated at 800 W than if the laser was operated at 1600 W. It is likely that at least some of the beam degradation was due to distortion of the focusing lens due to heating as the zinc selenide lens material absorbs a small amount of energy from the beam. This behavior of the beam would have been very difficult to ascertain without actual measurements of the beam spot size.
The average power output of pulsed lasers represents the average power delivered by the laser; for example, if the laser delivered pulses with 7 J of energy at a repetition rate of 10 Hz, the average power is 70 W. If the pulses last a millisecond, often the peak power is calculated by dividing the pulse energy by the duration of the pulses. For example, if the 7 J pulses last a millisecond, the peak power is said to be 7 kW. But it is more correct to consider this the average power during the pulse; the pulse itself may have fluctuations that can only be observed with a fast detector and oscilloscope. Consequently, the peak power may be significantly higher than the average power during the pulse.
Most pulsed lasers used for metalworking applications at the time of writing are Nd:YAG lasers. The neodymium atoms in the yttrium aluminum garnet rod absorb energy from the flashlamps; some of this energy is extracted in the laser pulse, but the remainder is conducted through the YAG rod to the water-cooled walls. Consequently, there is a temperature distribution across the rod, which then behaves as a lens. The amount of lensing in the rod is believed to be determined by the average power input to the rod. Since the rod becomes a lens, it affects the propagation of light between the resonator mirrors with the result that the mode of the output depends on the average power to the rod. The focused spot size can be measured using the spinning wand or hollow needle technique discussed above, but the firing of the pulsed laser must be synchronized with the revolution of the rod (Graham and Weckman, 1995). The peak intensity at the workpiece surface does increase with average power output, but the increase is less than proportional to the increase in power output.
Equally important as the change in minimum beam waist of the focused beam is the shift in the position of best focus as power is increased. Clearly, the position of best focus should be found for every average power level for which the laser is operated, otherwise welding may be attempted with the beam considerably out of focus.
A later generation of Nd:YAG lasers used a rectangular solid piece of crystal, with the beam redirected through the excited medium in such a way that any thermal distortion of the beam would be cancelled out. This style of laser was more expensive, however, and found limited market acceptance. The output of many Nd:YAG lasers is delivered by a fiber optic to the focusing head, which concentrates the beam on the workpiece. It is widely reported that the fiber ‘homogenizes’ the beam, and that the spot size on the workpiece is the image of the end of the fiber. The implication of these reports is that any energy-input dependence of the mode produced by the Nd:YAG laser is not important when the beam is fiber delivered. This is not necessarily true. Boechat et al. (1993) have reported that the length of fiber required for the output beam to be independent of the launch conditions is far longer than the normal fiber lengths used in welding lasers. Moreover, recent measurements showed a dependence of the measured spot sizes of fiber-delivered lasers on the laser operating conditions (MacCallum et al., 2000).
As mentioned above, there was a shift in the position of the focus as a function of laser power when using an Nd:YAG laser. Similar effects have been noticed with carbon dioxide lasers. For example, with a 20 kW CO2 laser focused by a 70 cm focal length mirror, a shift in position in excess of a centimeter was observed between spot size measurements undertaken at 2 kW and at 20 kW. The shift, which was accompanied by a change in focal spot size, is attributed to beam-induced thermal distortion of the window of the laser chamber.
Thermal distortion of the focusing lenses or output windows does not occur instantly. It was ascertained that the focus position of the output of a
1.6 kW carbon dioxide laser shifted in a time period of the order of 60 seconds after turn-on. Similar shifts were observed with an Nd:YAG laser, with 120 s required for stabilization. This shift was attributed not to warm up of the rod, which ought to take place on a time scale of a few seconds, but to temperature changes in the entire water cooling circuit.
It is assumed that the laser reaches its programmed power instantaneously after the command is given. This was investigated on the 1.6 kW laser described above. Measurement of the turn-on transients using a non-thermal detector with a fast response time showed that at times the laser overshot its programmed value by 25%, settling down to the steady state value in approximately a 10 s time period. Repeating the measurements several days later, however, the power was observed to increase gradually to the programmed value over a 10 s time period. The reason for the different response in the two cases was not ascertained. This laser was sealed, but the gas was replenished on a weekly basis. Possibly the different response was related to the age of the laser gas. Regardless of the reason for the time variation, successful welding with this laser was achieved only when the laser action was activated with the beam directed off the workpiece, into a beam dump, and welding commenced after a sufficient time period for the power to stabilize. This time dependence was not displayed on the power meter built into the laser and the laser had been successfully used for cutting operations for many years. Since significant energy input for the cutting operation comes from the heat of reaction of steel with the oxygen assist gas, the cutting operation is possibly less dependent on the laser power level than is the welding operation.
In controlling or specifying a laser-welding process, a number of other parameters affect the process. Welding procedure specifications generally class these as ‘essential variables’ and ‘non-essential variables’. An essential variable is one that can have a major impact on the weld quality; if a substantial change is made in an essential variable, the welding process must be requalified, which may be an expensive and time-consuming procedure. Essential parameters include, for example, the laser power (or pulse energy, pulse duration and pulse repetition rate for pulsed lasers), beam mode profile, lens focal length, focal point position, raw beam size, motion speed, number of passes, angle of incidence, welding position, nozzle gas type or composition, auxiliary gas type or change in composition, backing gas type or change in composition, plume reducing gas jets including orientation, flow rate or pressure of various gases, change in material or in filler metal type or size, joint design and joint gap.
Gas shielding is usually used for laser welding. With Nd:YAG welding, argon is the preferred shield gas, as it is heavier than air and falls onto the workpiece. In welding with the carbon dioxide laser, however, helium is normally used as it has a higher ionization potential than argon. Plasmas that absorb and scatter the laser radiation are more easily created at the carbon dioxide laser wavelength than at the Nd:YAG laser wavelength. Several different types of gas shields may be used. In a weld that completely penetrates the workpiece, a backing or underbead shield may be necessary. For very high speed welding, an auxiliary or trailing gas shield may be necessary. For spot welding, a simple gas flow through the nozzle may be sufficient; this serves not only to protect the weld from contamination, but also to protect the lens from fumes. For high power carbon dioxide laser welding, a plasma suppression jet may be required to blow the plasma out of the beam path. Gas flow must be carefully controlled. For cost purposes it is desirable to keep gas flows low, and indeed too high a gas flow may aspirate air into the gas stream resulting in weld contamination. Gas flows should be high enough, however, to provide adequate shielding in spite of random drafts.
Note that some steel types can be welded in air, without an inert gas shield. Other materials, such as titanium, which are often arc welded inside sealed and purged chambers can be successfully laser welded in air with only a nozzle gas shield. This is because the spot welds produced by a repetitively pulsed Nd:YAG laser are often fairly small and cool rapidly so there is no hot weld pool extending beyond the region covered by the gas flow.
Laser welding is most often carried out without filler metal; this process is called autogenous welding. There are two reasons why one might use a filler metal. As described above, the small size of the focused laser beam means that very good edge preparation must be used. However, the requirements for edge preparation can be relaxed if a filler metal is used. The second reason to use a filler metal is to control the metallurgy of the weld metal.
Filler metal can be applied in the form of wire, powder, or preplaced inserts such as rings or discs. The feeder for filler metal should be integrated with the laser control circuits, but note that many wire feeders are not sophisticated enough to produce reliable welds. In particular, at the end of weld, the wire feeder should be turned off a short interval before the laser beam is. Otherwise, the wire will freeze into the weld pool requiring a manual operation to free it. Moreover, at the start of the weld the position of the wire must be carefully set and the advance of the wire integrated with the laser turn-on time. Co-ordination between the feeder and the laser control circuits is less critical when using a powder feeder rather than wire feeder, as discussed in a later section on laser cladding and weld repair.
Laser welding is usually carried out with the laser beam directed at the seam between the two parts to be welded. There are two reasons for welding off the seam. The first is to control the metallurgy of the weld metal when, for example, welding a low carbon steel to a high carbon steel. This would happen when welding a formed component to a machine component, for example. To prevent cracking, it is beneficial to attempt to lower the carbon content of the weld metal. The position of the bead is judiciously located so the major part of the weld metal originates with the low carbon steel, in such a way that the total weld penetration is unaffected.
A second reason for laser welding with the beam positioned away from the seam between the two materials being welded is to enhance the absorption of the beam. For example, 12mm thick copper which is normally highly reflective to the output of a carbon dioxide laser, has been successfully welded to 12 mm thick nickel with a 9 kW laser beam by locating the beam approximately 0.25 mm onto the nickel side of the seam.