Laser beam welding
V. MERCHANT, Consultant, Canada
4.1 Introduction: process principles
The first operating laser, built by Thomas Maiman in 1960, was a pulsed ruby laser producing millisecond long pulses with a low repetition rate, in the far red region of the spectrum. This laser used the intense light from flashlamps to excite the chromium atoms doped into a crystalline aluminum oxide rod; it is these chromium atoms that give ruby and synthetic ruby their distinctive colour. Typical rods may be 6 to 10 mm in diameter and 20 mm long. The excited chromium atoms radiate their excess energy as the red light, which is repeatedly reflected by carefully aligned mirrors at each end of the ruby rod, is passed multiple times through the rod and is amplified by the process of stimulated emission. Within a few years it was found that greater powers could be achieved by using, instead of ruby, a rod consisting of neodymium ions doped in otherwise very pure glass. Although the Nd:glass laser continues to be used where very high pulsed energies are required, it has been superseded in many applications by lasers built with rods consisting of neodymium atoms doped into a crystal of yttrium aluminium garnet, or Nd:YAG (Koechner, 1976).
The carbon dioxide laser, a gas laser with the potential to be scaled to high average powers, was invented by C. K.N. Patel in 1966. The simplest carbon dioxide lasers consist of a tube from which the air has been evacuated and replaced with a low pressure mixture of carbon dioxide, helium and nitrogen gases. Electrical current from a high voltage power supply or a radio frequency (RF) generator passes through the gas, exciting the carbon dioxide molecules. The mechanisms by which energy is transferred to the carbon dioxide molecule and optical output produced are discussed by Patel (1969) and by DeMaria (1976). High power carbon dioxide lasers require complex gas flow systems to circulate the gas excited by the electrical discharge through heat exchangers that extract the waste heat.
Early in the history of lasers, it was discovered that the laser beam output could heat, melt and vaporize metals. If the laser output was carefully controlled, the melting and subsequent solidification would result in welds between adjacent pieces of metal. Thus laser beam welding was born and announced almost simultaneously by three different suppliers of laser equipment who were seeking to expand their markets (Banas, 1972; Locke et al., 1972; Bolin, 1976).
As of the early years of the twenty-first century, most laser beam welding is conducted by the output of either the carbon dioxide laser or the Nd:YAG laser. Both of these lasers, depending on the electrical excitation circuitry, can emit their output either continuously, as a single pulse, or as a repetitive series of pulses. Laser beam welding has been conducted with both continuous and pulsed lasers.
A very large number of materials have been found to give laser output. The output of the light source known as a laser is very special and differs in many important aspects from the output of any non-laser source. The most important aspect of a laser source is its coherence; coherence implies a definite relationship between the output observed in different places and different times (on a microscopic scale) that results from the process of stimulated emission. In stimulated emission, one molecule with excess energy is stimulated to give up this energy when it is impinged by light of a particular wavelength. The light the excited molecule emits is in the same direction, the same polarization, the same phase and the same wavelength as the light that stimulated the emission. Since this light is reflected by the laser mirrors, all succeeding light that is emitted is in the same direction, the same polarization, the same phase and the same wavelength.
Of all the properties mentioned above, for laser welding the only relevant property is that the light emitted by the stimulated emission, or laser action, is in the same direction. By contrast, the light emitted by a fluorescent or incandescent bulb spreads out all over a room and is useful for a different purpose, that is, illuminating the room. The laser light is unidirectional and can be steered by a series of mirrors to a workpiece located a considerable distance from the source. And because the light is unidirectional, most of the output from the source can be collected by a focusing lens, focused to a very small spot, resulting in localized heating of a selected target material.
As described above, the process of stimulated emission leads to a light that is monochromatic, or consists of a single wavelength. By contrast, the light from a light bulb or from the sun consists of many wavelengths, including ultraviolet wavelengths that do not travel a great distance through air, all the visible wavelengths from violet to red, and some infrared wavelengths. The output from the common lasers used in industrial applications, the carbon dioxide and the Nd:YAG lasers, is monochromatic. That is, the energy output of the laser consists of light with a very narrow band of wavelengths. In both cases the central wavelength is in the infrared beyond the range to which the human eye responds. The wavelength of the Nd:YAG laser is at 1.06 mm and the wavelength of the carbon dioxide laser is further into the infrared, at 10.6 mm. This difference in wavelength has some important consequences, as will be discussed later.
A heated surface can lose heat that is carried away by three different means; conduction, convection and radiation. The element of an electric stove glows red hot, but cools very quickly when a pot of water is placed upon it, illustrating the loss of heat from the heating element by conduction. Since heated air rises, a hot water radiator heats the room because of the flow of air over its surface; this is an example of the convection of heat. The electrically heated filament of a lamp radiates visible light energy, which can illuminate a whole room.
When a focused laser beam is incident on a metal surface, a number of factors come into play. Obviously the incident laser energy will heat the surface on which it is absorbed. If one envisions that the tightly focused laser is a point source of heat on the surface, the temperature at that point is a balance between the rate at which heat is input at the surface, given by the power of the laser source and the fraction of the energy that is absorbed, and the rate at which heat is lost from the surface. At temperatures characteristic of the welding process, it is usually assumed that heat loss by radiation is negligible and that heat loss by convection through removal by the surrounding gases is a secondary effect. The primary means of heat loss is by conduction away into the metallic material being welded. Thus the temperature reached is a balance between the laser power input and the rate of heat conduction.
The temperature at the surface can reach the point at which the metal liquefies. The liquid pool resolidifies when the source of heat, the laser beam, is removed and heat is distributed by conduction through the solid material surrounding the liquid pool. If the laser was incident near a joint between two different pieces of material, both of which melt due to incident energy, a join or weld is established between the two pieces of material when the molten material solidifies.
Note that the molten pool is not stagnant, but is stirred rapidly. The primary force that causes the motion of the liquid pool is known as the marangoni force and is related to the surface tension. The fluid flow is controlled by the spatial variation of surface tension that exists on the weld pool surface. The surface tension gradient arises from the spatial variation in surface temperature and the temperature dependence of surface tension. The spatial variation of surface tension causes the molten metal to be drawn along the surface from the region of lower surface tension to that of higher surface tension and this may result in very large surface flows (Zacharia et al., 1990). For pure metals and alloys the temperature coefficient of surface tension dg/dT is negative. Thus the surface tension is highest near the solid-liquid interface, where the temperature is lowest. The flow of the liquid pool is outward and away from the center of the pool. For metals with impurities, flow in the opposite direction may occur. The role of impurity elements and the spatial distribution of the laser energy in influencing the flow of liquid in the molten pool has been extensively investigated.
As the heat input is increased, the temperature increases until the vaporization temperature of the metal is reached. The laser beam drills a hole through the liquid pool, a hole which is filled with metal vapor. The laser passes through the metal vapor, contacts the liquid at the bottom of the hole and continues the drilling process. If the beam is moving with respect to the metal surface, the drilling process is not destructive, but forms a weld. As the beam moves, it continually melts more material at the front of the hole in the material. The molten material moves around the side of the beam and resolidifies at the rear of the hole. The flow of material is primarily in the liquid phase rather than in the vapor phase (Dowden et al., 1983).
There are two different modes of welding. The first is designated the conduction mode, in which the size of the weld pool is limited by the conduction of the heat away from the point that the beam impinges on the workpiece surface. This mode of welding can be produced by either pulsed or continuous beams. If a pulsed beam is used, the molten pool and hence the weld nugget produced on a flat surface or a butt joint is approximately hemispherical. A repetitively pulsed beam can be used with a moving part to produce a series of overlapping weld nuggets that form a hermetic seal.
The second mode of welding is called deep penetration or keyhole welding. It occurs when the beam is intense enough to cause a hole filled with metal vapor to occur in the workpiece surface. It is generally considered that a laser power of one MW/cm2 is required for keyhole welding in steel workpieces. A somewhat higher power is required for aluminium workpieces. Depending on the welding conditions, the hole may extend either part way or entirely through the workpiece. This mode of welding is most often performed with a continuously operating laser, although there has been some work done with repetitively pulsed or modulated beams. The welds produced by the deep penetration mode of welding have a high aspect ratio; that is, they are relatively deep and very narrow. Aspect ratios as high as 10 to 1 are not uncommon.