S Fisher, BNFL, Dr C A Olivier, TWI and S T Riches, TWI
Presented at 7th NOLAMP Conference (Nordic Conference in Laser Processing of Materials),
Lappeenranta, Finland, August 23-25, 1999
With the advent of high power Nd:YAG lasers (>1.5kW), little attention has been paid to optimisation of gas delivery, as most systems rely on gas shielding arrangements based on low power Nd:YAG lasers. The gas serves two main functions: the prevention of oxidation and the control of the plasma formed during laser welding. In a number of applications, one coaxial inert gas flow (normally helium or argon) around the weld suffices for both oxidation prevention and plasma control. However, in some cases, the plasma control obtained using a coaxial gas flow is not satisfactory as the weld penetration and quality may be adversely affected, particularly for high aspect ratio, partial penetration welds.
This paper describes a programme of work undertaken to compare the features of partial penetration Nd:YAG laser welds in stainless steel using coaxial gas flow and a directed flow of gas introduced by a side jet nozzle. The work has shown that the application of helium gas through the side jet nozzle had a significant advantage in the penetration depths and porosity levels that could be achieved in comparison with coaxial helium shielding. The tolerances to position and flow rates using the side jet nozzle for the Nd:YAG laser welding process on stainless steel tubes will also be described.
The development of high average power Nd:YAG lasers (>1.5kW) has opened up a range of welding applications where penetration depths of greater than 2mm are required. For some critical applications, the welds must meet the weld specification requirements, for example, of ASME, in order for the process to be approved. Historically, with Nd:YAG lasers operating in a pulsed mode with an average power of about 500W, little attention had been paid to the gas shielding beyond that of having a shroud of inert gas around the weld zone to prevent oxidation. When this type of system is applied to welding above 2mm penetration, particularly for partial penetration welds, problems have been encountered in meeting the porosity level specifications, where only a limited number of pores of <0.3mm are allowed.
For Nd:YAG laser welding, the processing capabilities are governed by a number of factors, all of which are interrelated and can affect the quality and properties of the welded joints. A summary of their effects on welding performance is presented in Table 1  .
Table 1. Summary of effects of process variables for laser welding
||>10 7W/cm 2- weld spatter
10 6-10 7W/cm 2- keyhole welding
10 5-10 6W/cm 2 - conduction limited welding
<10 5W/cm 2 - no coupling
||Applied Power/welding speed
||Deeper penetration as applied energy increased
Focal length of focusing system
Diameter of fibre optic cable
||Focused spot area = D x Fl/Rl, where
D - diameter of fibre optic cable
Fl = focal length of focusing lens
Rl - focal length of recollimating lens
||Type of gas
||Control of plasma and oxidation
The power density at the workpiece generally controls the formation of a weld. If the power density is too high, the weld shows an amount of spatter and a degree of undercut; whereas if the power density is too low, the penetration and coupling to the workpiece may be lost. In addition, there are two distinct regions of welding which also depend on the applied power density:
- Keyhole welding: where the power density is sufficient to vaporise material and produce a narrow, deep high aspect ratio weld
- Conduction limited welding: where the power density is sufficient to melt the material but not vaporise it, forming a shallow, wider weld.
In Nd:YAG laser welding of sections of >2mm using 2kW of laser power, most welds have some characteristics of both types of weld profile, in that there is a narrow root to the laser weld and a wider top, often referred to as a 'wine glass' shape. The wider top of the weld is normally related to the formation of plasma (of ionised metal vapour) above the weld, which hinders the transmission of the laser energy to the workpiece and creates a more diffuse heat source. For thin section welding (<2mm thick), this plasma effect is hardly noticed as the welding speeds are sufficient (>1m/min) to leave the plasma behind the keyhole. However, as the welding speeds decrease and thickness of metal to be welded increases, the effect of the plasma formation becomes more influential on the weld shape and quality. This can lead to instances where the welding speeds are so slow that the weld pool becomes unstable and leads to the formation of porosity. This is particularly pertinent in partial penetration welding where the pores once produced must escape from the top surface of the weld pool.
Fig.1 Transverse section of Nd:YAG laser weld in stainless steel enclosure without plasma control
In this application, a 2kW Nd:YAG laser was specified to produce hermetic welds in a type 316L stainless steel enclosure, where a weld penetration of >3mm was required and porosity levels were required to be in accordance with the ASME specification. Using a helium coaxial gas shield, it was possible to produce welds of the required penetration depth at a speed of 0.3m/min, but the levels of porosity were not within the ASME specification for the component. The porosity levels were generally around 100 pores of 0.3-0.75mm diameter in a 300mm weld circumference. A typical cross section is shown in Fig 1.
This programme of work was initiated to examine the factors affecting the weld penetration depth and porosity levels in welding of stainless steel enclosures with the objective of meeting the ASME specification.
2. Experimental Details
AISI 316L stainless steel enclosures of wall thickness 10mm and internal diameter 105mm were used for the trials. For the purpose of this exercise, all trials were carried out on melt runs with a view to achieving a penetration depth of >3mm.
Laser and beam delivery
A 2kW continuous average power Nd:YAG laser (Lumonics MW 2000) was used in the experiments. The beam was delivered to the workpiece via a fibre optic cable of 1mm in diameter and length 30m. The focusing optic comprised a 200mm focal length recollimating lens and a 160mm focal length focusing lens, which produced a nominal focal spot diameter of 0.8mm. The focusing lens was protected from weld spatter by a circular cover slide attached to the optics holder. The intensity distribution in a transverse plane of the beam emitted from the fibre was approximately uniform. The power measured at the workpiece was 1.5-1.65kW.
Experiments were carried out using helium and nitrogen process gases separately. The process gas refers to the shielding of the weld bead from oxidation and for controlling the build up of plasma above the weld.
A shielding shoe was provided to protect the solidifying weld bead from oxidation as illustrated in Fig.2.
Helium was used at a flow rate of 80l/min and nitrogen was used at a flow rate of 40l/min.
Fig.2 Schematic of gas shielding and plasma control arrangement for Nd:YAG laser welding of stainless steel enclosures
This system involved the use of a gas jet, which was delivered through a nozzle directed towards the beam workpiece interaction zone, as illustrated in Fig.2
. The gas type, flow rate, nozzle diameter, angle of incidence and impingement position were varied. For the impingement position, a positive displacement indicated that the gas jet was aimed at the workpiece behind the weld, thus interacting with the plasma above the weld.
The parameters investigated are presented in the results section.
Assessment of welds
The melt runs were radiographed by a single wall-single image technique. Prior to radiography, a section was cut for metallographic inspection. The remainder of the weld was divided into four sections, each of length 70mm. The number of pores in each section with a diameter of >0.3mm was recorded.
The radiographic results were interpreted in terms of the accept/reject criterion described in ASME VIII Div 1 UW 51 1995. This code permits a maximum of 12 pores with a diameter between 0.3mm and 0.75mm in a 150mm length of weld. The weld is rejected if any pore has a diameter of >0.75mm. In relation to the pipe sections, a limit of 25 pores total around the weld circumference was used as the basis for whether the weld passed the porosity criteria.
The weld profile was examined for each melt run and the weld penetration depth was recorded. A minimum limit of 3mm was set for weld penetration depth.
3. Results and discussion
A typical transverse section of a weld made with plasma control is shown in Fig.3.
Fig.3 Transverse section of Nd:YAG laser weld in stainless steel enclosure with plasma control
The main variables investigated in the trials are summarised in Table 2.
Table 2. Summary of process variables examined in Nd:YAG laser welding of stainless steel enclosures
A summary of the findings for each of the above parameters is presented below.
|Laser power, kW
||1.5kW and 1.65kW at the workpiece
|Welding speed, m/min
||0.3, 0.4, 0.6, 0.72
|Gas jet to can angle, °
|Nozzle diameter, mm
||1.2, 2, 3
|Plasma jet-can impingement point, mm
||+2.5, +1.5, +0.5, -0.5, -1.5
|Gas flow rate, l/min
||20, 28, 36 (helium only)
Effect of laser power
The trials indicated that the effect of laser power at the workpiece was significant in terms of the penetration depth and porosity levels obtained. For a workpiece power of 1.5kW, it was more difficult to achieve the required penetration depth with low porosity welds, whilst it appeared to be more tolerant at the higher power. This can be related to the greater power density achievable with the higher power at the workpiece, thus leading to a more stable keyhole at the penetration levels required. A higher power laser (>2kW) would lead to more tolerant conditions for this thickness, as the issue of plasma control would become less critical.
Effect of welding speed
In the initial trials with the coaxial gas arrangement, a welding speed of 0.3m/min was required in order to attain the weld penetration depth. With the plasma control system, it was possible to increase the speed up to 0.7m/min with a satisfactory penetration depth. The weld profile changed from a wine glass shape to a more straight walled profile with the adoption of the plasma control system.
Effect of gas type
Helium was used as the primary plasma control gas as it possesses a high ionisation potential, thus reducing the formation of plasma above the weld. However, helium is a very light gas which means that high flow rates are normally required to achieve an adequate control of the plasma. In these trials, a flow rate of 28l/min was found to be adequate with a nozzle diameter of 1.2mm. Nitrogen gas was also considered as an alternative, and it was shown that acceptable welds in terms of penetration depths and porosity levels could be obtained with a flow rate of 14l/min and a nozzle diameter of 1.2mm.
For higher flow rates than those stated, it was observed that the weld top bead became depressed as the gas tended to produce a gouging action.
Effect of gas jet angle
The gas jet angle with respect to the enclosure was set at 40° during the trials. Due to the set up of the focusing head and the gas shielding arrangement, it was not possible to alter this angle much in the tests carried out.
Effect of nozzle diameter
The nozzle diameter for the initial trials was fixed at 1.2mm to produce a high gas flow velocity for the plasma control effect. In the trials with helium gas, it was found that 28l/min appeared to be the optimum flow rate, with higher flows producing gouging of the weld and low flows not being effective in control of the plasma. However, it was also noted that the tolerance to position of this nozzle diameter was critical, with only an impingement point of +1.5mm producing welds of acceptable levels of porosity and penetration.
The nozzle diameter was enlarged in subsequent trials to 2mm and 3mm, where a tolerance to gas jet to beam position was increased to ±1mm from the position of +1.5mm. This was achieved without altering the gas flow significantly.
Effect of plasma jet impingement position
As mentioned above, when the 1.2mm diameter nozzle was used, there was little tolerance to the impingement position of the plasma gas jet. This tolerance was extended using 2mm and 3mm diameter nozzles. It was also noted that any lateral displacement of the nozzle away from the weld line would result in an increase in porosity and the accuracy of positioning must be maintained within 0.5mm of the joint line.
Effect of gas flow
The gas flow using helium was based around 28l/min. Higher flow rates tended to produce some gouging of the weld top bead, whilst the lower flow rates did not effectively control the plasma.
4. Application of results
From the above, the following parameters were selected to produce welds that met both the criteria of weld penetration and porosity:
- Laser power at workpiece: 1.65kW
- Welding speed: 0.6m/min
- Total gas jet to can angle: 40°
- Helium plasma control gas jet flow rate: 28l/min
- Plasma jet - can impingement point: +1.5mm behind beam-can interaction point
- Nozzle diameter: 2mm
These conditions have been successfully applied in production of the stainless steel enclosures. It was also observed that the effective control of the plasma significantly reduced the level of 'sooting' caused by the welding operation. This 'sooting' effect was the formation of oxides from the stainless steel which can condense on the cover slides, thus reducing the power at the workpiece.
The amount of porosity in partial penetration Nd:YAG laser welds in AISI 316L stainless steel could be reduced through the use of a plasma control jet in preference to a coaxial gas shield, in order to meet the requirements of ASME VIII specification, whilst achieving a penetration depth of over 3mm.
The conditions developed have been applied successfully in production.
The authors would like to thank BNFL Engineering Ltd. for permission to publish this paper. The authors would also like to acknowledge the work of F Nolan, TWI for performing the experimental trials and Dr J Ion for the experimental planning.
||Riches S T and Ion J C
||'Guide to high average power Nd:YAG laser processing with fibre-optic beam delivery for metals' TWI Members report 545/1996, April 1996