José Greses started work at the Laser and Sheet Processes Group of TWI (UK) in May 2003, after completion of his PhD at the University of Cambridge (UK). He was previously involved in the PTP scheme (EPSRC-DTI) since October 1999, working on vapour-laser interactions when high-power laser welding.
Paul Hilton joined TWI in 1990 as Head of Department of the Laser Centre, where he was responsible for the day to day operation of the department and for the alliance between the laser centres of AEA Culham Laboratory and TWI. In 1994 he became 'Technology Manager - Lasers', with a strategic responsibility for use of laser technology throughout TWI.
When welding mild steel with high-power continuous wave lasers, vapour is ejected from the keyhole. Spectroscopic studies of the vapour emission have demonstrated that the vapour can be considered as a partially ionised plasma or as a thermally excited gas, depending on the laser beam wavelength and the materials being welded. José Greses and Paul Hilton explain the causes and effects of these differences in the vapour, and make some general recommendations for improving the laser welding process efficiency by use of appropriate control gas.
Since the mid 80s it was believed that Nd:YAG laser light (1.06µm wavelength) could offer major advantages over CO2 laser light (10.6µm wavelength) for welding applications, such as enhanced coupling to reflective metals, use of an optic fibre for beam delivery (offering great flexibility in the welding process) and apparent increased process efficiency at the same power. However, only in recent years have Nd:YAG lasers with more than 3kW of power been commercially available. The use of such high-power Nd:YAG laser light for welding has presented new issues and problems when compared to high-power CO2 laser welding. Work at TWI in the mid-70s, with CO2 lasers, led to the use of side jets of high ionisation potential gas ( ie helium) to control the formation of plasma caused by the interaction between the laser beam and the vapour emerging from the keyhole. Early work with high-power continuous wave (cw) Nd:YAG lasers showed that a large and difficult to control vapour plume was formed ( Fig.1), energetic in the orange-yellow part of the spectrum, which was very different to the intensely bright bluish colour of the plasma formed when cw CO2 laser welding ( Fig.2). The mechanism responsible for this difference is the interaction between the laser beam (of different wavelengths) and the vapour ejected from the keyhole. With the detailed description of these physical processes being unclear, a collaborative project between TWI and the University of Cambridge, part of the PTP scheme, was established to develop a better understanding of the processes involved. The major results from this work and some practical recommendations arising from the project are outlined in this article.
Fig.1. Plume formation when Nd:YAG laser welding
Fig.2. Plasma formation when CO2 laser welding with helium side jet
Laser Keyhole Welding
Laser welding works in a fundamentally different way to most arc fusion processes. The laser beam is focused to produce an energy density such that the metal not only melts but vaporises. The outward pressure of the expanding vapour is balanced by inward acting forces, such as surface tension and gravity forces, to form a through thickness column known as a 'keyhole'. The formation of this keyhole, which promotes the overall coupling between the laser and the material via a mechanism known as Fresnel absorption, leads to an associated increase in weld penetration depth. This deep penetration and narrow fusion zone is characteristic of high-power laser welding. As the material is traversed by the laser beam, molten metal flows around the keyhole and resolidifies on the downstream side.
Plasma/plume formation
When CO2 laser welding, the metallic vapour that results can ionise to form a plasma via a process known as inverse Bremsstrahlung (IB) absorption. The degree of ionisation in the vapour depends upon the ionisation potential of the metallic species being vaporised, and the laser beam parameters.
Typically, a partially ionised plasma vapour with an electron temperature between 6500K and 12000K can be expected when welding with high-power cw CO2 lasers. The wide range of temperatures reported, is mainly due to the different spectroscopic methods which have been used to measure the plasma temperature.
Since the vapour starts to disperse as it leaves the keyhole, the temperature across the ejected vapour is not uniform, with the hottest region being just inside the keyhole close to its top. The temperature (and density) gradient creates a variation in refractive index across the vapour plasma, affecting the laser beam propagation by absorbing and defocusing it.
If no control gas is used during laser welding, the plasma formation would effectively absorb most of the laser energy, preventing the laser beam reaching the workpiece. When helium gas is delivered to the welding area, direct beam absorption by IB is usually less than 5% of the laser energy. However, even such small absorption is enough to create a gradient in refractive index, changing the energy density distribution and the focal position of the laser beam, and thus effectively reducing the penetration. Furthermore, defocusing of the beam, causing enlargement of the beam focal spot, is probably the main mechanism responsible for the wine-glass weld shape characteristic of CO2 laser welding cross-sections ( Fig.3).
Fig.3. Wine-glass weld shape characteristic of CO2 laser welding. Cross-sections at 0.75 m/min and 3.5kW
When cw Nd:YAG laser welding, the metallic vapour created consists only of a thermally excited gas (of temperature between 2000 and 3000K). Since inverse Bremsstrahlung absorption, the main mechanism responsible for vapour ionisation, is approximately inversely proportional to the square of the beam wavelength, the shorter the wavelength the less the IB absorption. Spectroscopic results have shown that no ionisation of the metallic vapour was detected when cw Nd:YAG laser welding. Nevertheless, cross-sections of the Nd:YAG laser welds also show the characteristic wine-glass shape of beam defocusing ( Fig.4). What, then, is the mechanism responsible for this shape?
Fig.4. Wine-glass weld shape characteristic of Nd:YAG laser welding. Cross-sections at 0.75m/min and 3.5kW
Particles in the vapour
An investigation of the average size of the particles that constitute the vapour ejected from the keyhole was carried out for both cw CO2 and Nd:YAG laser welding of mild steel. After analysing and measuring the size of the vapour particles, it was evident that different particle sizes could be found when CO2 and Nd:YAG laser welding.
The measured average particle diameter when cw high-power CO2 laser welding was found to be around 4nm ( Fig.5). Although an exact mechanism is unclear, it is believed that the effect of the plasma over the Knudsen layer, which governs the vaporisation process in the keyhole, reduces and homogenises the size of the CO2-induced particles in the welding vapour.
Fig.5. CO2-induced particles
Fig.6. Cluster of Nd:YAG-induced particles
In contrast, the measured average particle diameter when cw high-power Nd:YAG laser welding was found to be around 40nm ( Fig.6). The particle size was distributed over a wide range of diameters up to 160nm, although 98% of all Nd:YAG-induced particles analysed had a diameter smaller than 100nm.
Plume scattering
The interaction between particles in a vapour and an intersecting laser beam is governed by general Mie scattering theory. A full theoretical description is out of the scope of this article, nevertheless, the attenuation caused by such interactions is partially due to beam absorption (notice that this is a different mechanism to IB absorption) and partially due to beam scattering. The total attenuation is also dependent on the diameter of the particles in the vapour and the number of particles per unit volume. For particle diameters smaller than 100nm, Mie scattering can be simplified to the well-known equations of Rayleigh scattering (the mechanism responsible for the blue-like colour of the sky).
The attenuation, that is absorption and scattering, of a laser beam in a vapour plume containing a high number of particles per unit volume, was investigated both theoretically and experimentally. Based on both sets of results, it was found that the number of particles per unit volume, ie the particle density in the plume, had the strongest influence in attenuating the beam. The size of the particles (for particles up to 100nm in diameter) had a minor influence on the attenuation of the laser beam in the plume.
For a Nd:YAG incident laser wavelength, and using a Nd:YAG probe laser, transverse to the beam, the attenuation of the incident laser beam power was experimentally found to be as high as 40%. The part of the plume close to the keyhole opening had the highest concentration of particles per unit volume. In that zone the strongest attenuation of the incident laser beam was measured. The attenuation decreased along the plume, away from the top of the keyhole, as the volume ratio decreased. From the probe laser results and Mie scattering theory, the mass of particles in the Nd:YAG-induced plume was estimated to be between 0.05 and 0.1g. Such a particle mass was in line with a measured vaporisation rate of ~0.025g/s when cw Nd:YAG laser welding with an incident laser energy density of ~1.24MW/cm 2.
The absorption part of the attenuation reduces the energy density of the beam at the zones where a higher particle concentration exits, effectively reducing the total weld penetration. The scattering part of the attenuation defocuses the laser beam over a larger area, further reducing the energy density in the laser beam. It is believed that this mechanism is largely responsible for the wine-glass shape of the weld cross-sections and their reduced penetration, when Nd:YAG laser welding.
Vapour control: increased process efficiency
In order to control the plasma formation when CO2 laser welding at low speeds, a high ionisation potential gas, delivered coaxially, or at an angle from a side jet, has proved to be the most effective way for reducing the plasma size and stabilising plasma behaviour. Helium, with its high thermal conductivity and high ionisation potential, is the only gas that fulfils all the necessary requirements, so that plasma loss, from direct plasma absorption and defocusing of the laser beam, is minimised. Argon can also be used as a control gas for CO2 laser welding, but it is only recommended for medium to low powers (less than 2kW) and/or high speeds (> 1m/min).
The use of a gas side jet to divert the plume from the interaction path with the laser beam, has proved an effective method of reducing beam attenuation when Nd:YAG laser welding at low speeds. Argon, with its heavier molecular mass, is capable of achieving this. It has also proved to be more efficient than helium at diverting the plume and optimising penetration, when delivered at a high flow rate through a small diameter, precisely positioned nozzle. However, very high flow rates could disrupt the keyhole and weld pool. In addition, the inert characteristics of argon are effective at protecting the weld pool from interacting with the atmosphere while hot.
Nitrogen, and probably CO2, could also be used as efficient plume control gases if the weld metal composition would allow this.
The tolerance in the gas side jet position when Nd:YAG laser welding is very small (less than 1mm with respect to the impingement position of the laser beam with the workpiece). This is due to the delicate balance between the two of the roles that the control gas plays. First, the gas side jet diverts the plume ejected from the keyhole from the interaction path with the laser beam. Second, the gas side jet further expands the keyhole opening stabilising the keyhole front wall and therefore reducing porosity.
Conclusions
The differences between the observed temperature and composition in the vapour outside the keyhole for CO2 and Nd:YAG laser welding of mild steel have been explained. These differences in temperature and composition are responsible for the effects that the plasma or plume produces in the weld, being mainly weld shape and weld penetration. When CO2 laser welding, plasma is created by laser beam absorption and defocusing of the laser beam, due to a gradient of electron temperature, density and refractive index in the plasma which changes the beam energy distribution and focal position. This causes intermittent penetration and porosity. Changes in energy distribution are also responsible for the different cross-section weld shapes. For Nd:YAG laser welding, defocusing due to scattering is largely responsible for the weld shape, although the absorption part of the attenuation is responsible for changes in penetration. Table 1 presents a summary of plasma/plume effects when high-power cw laser welding.
Table 1 Summary of plasma/plume effects in CO2 and Nd:YAG high power laser welding
| CO2
| Comments | Nd:YAG | Comments |
| 10.6µm | | 1.06µm | |
| 6500K to 12000K | Partially ionised plasma formed due to inverse Bremsstrahlung absorption in the laser beam | 2000K to 3000K | Thermally excited gas due to the low radiative transfer from attenuation |
| ~4nm | The plasma affects the Knudsen layer (formed between the liquid and vapour phase of the keyhole), minimising and homogenising the particle size. | ~40nm | The particle size is distributed over a large range due to lower temperatures in the vapour, possibly caused by localised high temperature areas in the keyhole or vapour. |
| Not measured in this work | The effect of the plasma on the Knudsen layer also reduces the vaporisation rate. | ~0.024g/s (for ~1.25 M/cm 2) | Depends on the energy density of the laser beam. |
| Plasma absorption causing beam defocusing | Although the plasma absorption is small when using an optimised control gas, the absorbed energy is enough to create a gradient of electron temperate and density, resulting in a gradient of refractive index in the plasma. | Attenuation (absorption and scattering) | Mie particle scattering theory is applicable to calculate the energy loss in the plume. The major part of the energy is absorbed, with a small part being scattered. Attenuation mainly depends on the vaporised mass in the plume and the refractive index of the particles. |
| Helium | A high ionisation potential is important. | Argon | The molecular weight of the gas is important. CO2 and N2 can also be used depending on the material characteristics. |
| 100 % with a helium control gas | Plasma formation may enhance the energy transfer between the beam and the workpiece. The incident energy density is modified due to beam defocusing. | 85-90% with an argon control gas | Plume formation attenuates the energy transfer between the beam and the workpiece. The incident energy density is modified due to beam defocusing caused by Mie scattering. |
| 100 % or close | The laser-plasma interaction does not affect to a great extent the transfer efficiency. | 100% or close | The laser-plume interaction does not effect to a great extent the transfer efficiency. |
| By doing a comparison between CO2 and Nd:YAG laser welding at 3.5kW using optimised control gas conditions, the penetration was higher for the CO2 laser light than for the Nd:YAG laser light. The higher penetration could be due to the higher beam quality of the CO2 laser. |
Practical recommendations
Based on the results and conclusions presented in the above sections, the typical optimum parameters to control the plasma formation when high-power welding with a cw CO2 laser are specified in Table 2. The typical optimum parameters to control the plume formation when high-power welding with a cw Nd:YAG laser are specified in Table 3.
Table 2 Typical control gas set-up parameters when high-power cw CO2 laser welding using mirror focusing systems.
| Side Gas Jet |
A (deg) 45 | a (mm) 5 | b (mm) +1 | c (mm) 50-200 |
Auxiliary feed Co-axial | Gas type Helium Argon (in lens systems) | Flow (l/min) 20-30 30 |
Laser leading/trailing/angle Air knife | Trailing Yes (in mirror systems) | |
Table 3 Typical control gas set-up parameters when high-power cw Nd:YAG laser welding.
| Side Gas Jet |
A (deg) 45 | a (mm) 5 | b (mm) +1 | c (mm) N/A |
Auxiliary feed Co-axial | Gas type Argon | Flow (l/min) 20-30 |
Laser leading/trailing/angle Air knife | Trailing Yes | |
A schematic overview of a possible control gas arrangement is shown in Fig.7, with A being the angle of the side gas jet and a, b and c the distances from the co-axial shielding nozzle, the impingement position and the side gas jet, respectively.
Fig.7. Gas delivery set-up and nomenclature
Use of these recommendations will assist in the repeatable production of high quality keyhole welds, when using high-power CO2 and Nd:YAG lasers.