Nd:YAG Directed Gas Jet Laser Welding of Titanium Alloys

Laser Processing Research Centre, The University of Manchester, Manchester M60 1QD, United Kingdom

 

Chris Allen and Paul Hilton
TWI Ltd., Granta Park, Abington, Cambridge CB21 6AL, United Kingdom

Lin Li
Laser Processing Research Centre, The University of Manchester, Manchester M60 1QD, United Kingdom

Copyright 2010, Laser Institute of America, Orlando, Florida. Reproduced with permission from Journal of Laser Applications, Volume 22, Number 2, pages 71-78. The Laser Institute of America disclaims any responsibility or liability resulting from the placement and use in the described manner.

Paper published in Journal of Laser Applications, Vol.22. No.2. 2010. pp. 71-78.

The increasing utilization of titanium alloys in the aerospace industry, a direct result of socioeconomic pressures, has created the need for a production process which can produce high quality near-net-shape titanium alloy components. Keyhole laser welding is a joining technology which could be utilized for this requirement. In general, when laser welding titanium alloys, a jet of inert gas, directed at the region of the laser beam/material interaction point is utilized to achieve the weld quality required. A statistical study has been performed in order to determine the optimum position and flow rate of this directed gas jet, with respect to reducing the weld metal porosity and optimizing the weld profile, for autogeneous Nd:YAG laser welding of 3.25 mm thickness Ti-2.5Cu and Ti-6Al-4V. As a result, butt welds have been reproducibly made with a quality that exceeds the most stringent aerospace weld quality criteria. High speed imaging and spectroscopic analysis of the welding process have revealed that, when correctly set-up, the directed inert gas jet disperses the formation of excited metal vapor above the keyhole and also significantly changes the hydrodynamic behavior of the weld pool. © 2010 Laser Institute of America.

Key words: laser, weld, titanium, Nd:YAG, porosity, aerospace

I. Introduction

The increasing utilization of titanium alloys in the aerospace industry, a direct result of socioeconomic pressures, has created the need for a production process which can produce high quality near-net-shape titanium alloy components. Keyhole laser welding is a joining technology which could be utilized, after appropriate classification of the resultant weld qualities, for this requirement. As a result of their high specific strength, corrosion resistance, fatigue resistance, and their ability to operate at elevated temperatures, titanium alloys are already exploited throughout the aerospace industry. The specific alloy employed depends upon the exact service requirements of the component, but α, α / β, and β alloys are all exploited.^[1] Current demand for titanium products in the aerospace industry is primarily being driven by the commercial sector, which is utilizing an increasing amount of titanium alloys to achieve weight savings.^[2] The adoption of large volumes of carbon fiber reinforced polymers (CFRPs) into the fuselages of modern airframes is adding further to the use of titanium alloys because of their galvanic compatibility with graphite.

High quality components can be produced by forging and subsequent machining. However, this process is labor intensive and finished components can have uneconomical buy-to-fly ratios. The production of near-net shape components with a high integrity joining process could significantly reduce material wastage and increase production rates. Keyhole laser welding is a high energy density joining technology that produces deep penetration welds with a relatively low heat input when compared to inert gas arc welding. Furthermore, it can be performed at atmospheric pressure and the fiber optic delivery of near infrared laser beams, provides increased flexibility when compared to CO₂ laser systems and other joining technologies. However, the formation of porosity in the weld metal is of particular concern when laser welding titanium alloys with a 1µm wavelength laser source. For high-performance, fatigue-sensitive components whose weld profiles are dressed, pores can break the surface of the dressed weld and reduce its fatigue resistance.^[3] A direct result of this is the stringent weld quality criteria that are pplied to welded components in the aerospace industry.^[4]

During fusion welding, the formation of porosity in the weld metal can occur as a result of soluble gases dissolved in the weldpool, such as hydrogen, being rejected upon solidification of the molten material. The presence of hydrogen in the joint zone can be significantly reduced by shielding the process with low moisture content shielding gas and a suitable preweld purge time. The hydrogen content in modern titanium alloys is typically <150 ppm and is not of concern. However, since TiO₂ is hygroscopic it will adsorb moisture from the atmosphere and a hydrated layer will form on the surface of the parent material. Removal of this layer prior to welding has been reported to reduce weld metal porosity in laser welded titanium alloys.^[5]

Table I. Chemical compositions of the titanium alloys investigated

Material	Element, wt% (except where stated ppm)
	C	Al	Cu	Fe	H (ppm)	N	O	V	Ti
3.25 mm Ti-2.5Cu	0.006	<0.01	2.34	0.07	17	0.006	0.16	<0.01	Bal
3.25 mm Ti-6Al-4V	0.011	6.33	<0.01	0.20	61	0.007	0.16	3.83	Bal

It has been reported that keyhole instability can lead to metal vapor and/or inert shielding gases being trapped in the weld metal.^[6] This occurs when the forces trying to hold the eyhole open (i.e. vaporization pressure and radiation pressure) are not in equilibrium with those trying to close it (i.e. surface tension, hydrostatic pressure, and hydrodynamic pressure).^[7] The vaporization pressure can be affected by a vapor plume exiting the keyhole that attenuates the incident laser radiation. The specific attenuation mechanism(s) depends upon the wavelength and power density of the laser beam, as well as the chemical properties of the workpiece.^[8] As the dynamic behavior of the vapor plume is not constant the power density delivered to the workpiece can fluctuate.^[9] Consequently, changes in vaporization pressure can lead to an unstable keyhole, periodic collapses of the keyhole, and the generation of porosity in the weld metal.

When welding C-Mn steel with a CO₂ laser it has been demonstrated that the primary mechanisms responsible for beam attenuation are absorption and defocusing of the laser radiation through a partially ionized plasma with a gradient refractive index.^[8] For CO₂ laser welding of titanium alloys, it has been reported that shielding of the weld pool with helium gives a higher weld quality than when using argon shielding gas.^[10] This is a result of the lower ionization potential of argon, which promotes inverse Bremsstrahlung absorption of the incident 10.6 µm laser radiation at lower irradiances than when helium shielding is used. The use of helium shielding gas can therefore reduce the variation in the vaporization pressure and increase keyhole stability/internal weld quality. Conversely, for Nd:YAG laser welding, of C-Mn steels at least, beam attenuation takes place by absorption and scattering of the laser light by a population of Fe nanoparticles (~50 nm in diameter).^[8,11]

Previous studies have shown that an inert gas jet, directed toward the laser beam material interaction point, can reduce the occurrence of weld metal porosity when CO₂ or Nd:YAG laser welding of 304 stainless steel.^[12-14] Furthermore, recent publications have reported that butt welds in titanium alloys with a low level of weld metal porosity can be produced using a near-infrared laser beam.^[4,15] In order to achieve a low level of weld metal porosity, both Hilton et al.^[4] and Mueller et al.^[15] reported the use of an inert assist gas. No information regarding the parameters associated with this assist gas jet were disclosed by the authors. It is not known whether these parameters have been optimized and what the set-up tolerances are if a consistently low level of weld metal porosity is to be achieved.

In this paper, the directed gas jet parameters for Nd:YAG laser welding of 3.25-mm-thick titanium alloys have been optimized, with respect to minimizing weld metal porosity and achieving a favorable weld profile, using a parametric approach. Furthermore, the interaction between the directed gas jet and the welding process has been observed with a high speed camera, and the vapor plume has been characterized.

II. Experimental procedures

A. Laser welding procedures

Experimental trials were performed on 3.25-mm-thick Ti-2.5Cu and Ti-6Al-4V (Table I). The materials had been chemically pickled after rolling, although the time between pickling and welding was not controlled. Immediately prior to performing bead-on-plate (BOP) tests, the surfaces of test pieces were cleaned with an abrasive paper and acetone degreased. An identical procedure was followed when performing butt welding trials, except that the abutting edges of the test pieces were dry machined prior to abrading and degreasing.

A Trumpf HL4006D Nd:YAG laser source was utilized for all welding trials. Relevant information for the laser source and its focused beam is shown in Table II. The process head was attached to a Kawasaki JS30 robot, which allowed accurate control of the focal plane position. The average laser powers at the workpiece were measured using a water cooled Ophir power meter. A measured power of 4.1 kW (at workpiece) was used in all trials. The focusing optic was protected with a high pressure air-knife and a cover slide positioned between the lower surface of the focusing lens and the workpiece. All BOP tests and butt welds were performed in the 1 G position.

Table II. Trumpf HL4006D Nd:YAG laser source information and focused beam properties

Delivery fiber diameter, µm	600
Nominal beam parameter product, mm.mrad	23.2
Collimating lens focal length, mm	200
Focusing beam focal length, mm	200
Nominal beam diameter at waist position, µm	600
Calculated beam focusing angle, °	4.42
Nominal power density (at beam waist for 1 kW), kW/mm2	3.54
Calculated Rayleigh length, mm	±3.88
Calculated depth of focus (5%), mm	±2.47

Since titanium reacts with oxygen and nitrogen at relatively low temperatures (~500°C), a stainless steel trailing shoe was used to prevent oxidation and embrittlement of the weld metal. The trailing shield and a 10 mm² cross-section copper efflux channel were both supplied with a regulated supply of pureshield argon (gas type I1 to BS EN 14175).^[16] A directed jet of pureshield argon was used in the majority of experiments. The positional parameters of the argon jet could be manipulated by the use of a dedicated bracket, and the flow rate of gas was measured using a Platon flow meter. Figure 1 details the experimental setup.

Fig.1a) Experimental setup; and
Fig.1b) schematic of the directed gas jet configuration

B. Parametric study

A three stage parametric study, consisting of preliminary screening trials and two separate design of experiments (DOE), was carried out in order to determine the significant processing variables and their optimum values with respect to minimizing weld metal porosity (the response variable investigated). Initially, the effects of seven processing variables on the weld metal porosity were evaluated. The variables investigated are given in Table III. Evaluation of the weld metal porosity and profile results from these experiments allowed the range of variables examined in the first DOE to be defined. A two-level factorial experiment was created, using DESIGN-EXPERT software, for four process variables (i.e. 2⁴) associated with the directed gas jet. The results from this DOE were subsequently used to define the range of variables investigated in the second DOE - a general factorial experiment used to determine the optimum position of the directed gas jet and its operating windows if a consistently low-level of weld metal porosity is to be achieved.

Table III. Process variables investigated

Variables	Fixed value or range investigated
	Initial trials	First DOE	Second DOE
Welding speed (mm/s)	10 to 100	67	67
Laser focal plane position (mm)	+3 to -3	0	0
Directed gas type	Argon, helium	Argon	Argon
Directed gas jet pipe internal diameter (mm)	1.4, 2.0, 4.0	4.0	4.0
Measured gas flow rate (l/min)	5 to 26	16, 20	16, 18, 19
Directed gas jet angle off plate surface (deg)	30 to 55	35, 55	35, 45, 55
Y-axis impingement point of gas jet axis (mm)	0 to 3.0	1.5, 2.5	1.5, 1.75, 2.0, 2.25
X-axis impingement point of gas jet axis (mm)	0	-0.5, 0.5	0

C. Weld quality assessment

All BOP tests and butt welds were assessed visually for oxidation. Radiographic examination was performed, according to BS EN 1435:1997,^[17] to determine the weld metal porosity content, which allowed indications ≥0.05 mm in diameter to be detected. Selected samples were also sectioned, transverse to the welding direction, at representative positions, to assess their weld profile. Weld profile and weld metal porosity results were compared to typical aeroengine quality criteria, as reported previously by the authors.^[4] Table IV details the weld quality criteria.

Table IV. Typical aeroengine weld quality criteria for 3.25-mm-thick titanium alloys. Derived from Hilton et al. (Ref.^[4]).

Weld profile		Subsurface porosity
Undercut (Ca) (mm)	0.16	Maximum dimension for a single pore, mm	1
Excess weld metal (R) (mm)	0.69	Accumulated length in 76mm of weld, mm	1.7
Excess penetration (r) (mm)	0.69
Incompletely filled groove (Cr) (mm)	0.33
Root concavity (cr) (mm)	0.33
Shrinkage groove (ca) (mm)	0.33
Face weld width (L) (mm)	<4.2
Minimum weld width (I0) (mm)	1.1<I0<2.1
Root weld width (I) (mm)	1.1<I<4.2

D. Process characterization

High speed video observations and optical spectroscopy were performed during selected welding experiments. A Photron Fastcam MC1 high speed video camera was utilized to observe the behavior of the keyhole and the weld pool during selected welding experiments. Temporal and spatial filtering effectively eliminated the inherently bright process emissions and a CAVILUX^® Smart fiber-delivered diode laser (808 nm) provided illumination. An Ocean Optics HR4000 spectrometer was utilized to record the spectrum of the vapor plume. The spectrometer was calibrated such that it recorded in the range 400-1000 nm with a spatial resolution of 0.3 nm and a sampling frequency of ~30 Hz. The spectrometer was controlled using the OOIBASE32 software package and the recorded data analyzed using MATLAB. Furthermore, typical condensed vapor plume particle sizes were determined using field emission gun (FEG) scanning electron microscopy (SEM). The elements contained in the vapor plume were determined using conventional SEM with energy dispersive x-ray (EDX) analysis.

III. Results and discussions

A. Parametric study

1. Initial BOP tests

Autogeneous BOP tests were produced at welding speeds between 10 and 100 mm/s, and at different focal plane positions without using a directed gas jet. The weld face and root of all the BOP tests were bright silver in color, indicating that they had been adequately shielded. A welding speed of 67 mm/s was the fastest welding speed that gave a consistent fully penetrating weld, with a root wide enough for practical application.

Numerous BOP tests were performed at a welding speed of 67 mm/s with different combinations of parameters shown in Table III. Not all permutations of the parameters were performed. With certain parameter combinations, the weld quality criteria stipulated in Table IV could be achieved. In general, lower weld metal porosity results were achieved with the laser beam focal plane positioned on the surface of the workpiece, when using argon rather than helium, and with a 4.0 mm internal diameter pipe. Henceforth, these parameters were fixed to the above values. A large variation in the weld metal porosity, both in terms of the number and size of pores, was observed with relatively small adjustments in the remaining parameters.

2. First design of experiments

In order to determine the significance of single factors as well as two and three level factor interactions, on the weld metal porosity, a two-level factorial experiment for four variables (shown in Table III) was designed using the DESIGN-EXPERT software. The variables examined were the measured gas flow rate (variable A), the jet angle off the plate surface (variable B), the impingement point of the jet axis ahead of laser impingement point (variable C), and the impingement point of the jet axis with respect to either side of the joint line (variable D). The values of the variables examined in this DOE were chosen based on the results of the earlier trials. Experimental noise was reduced by choosing parameter values that were expected to produce fully penetrating BOP tests, performing three replications of each parameter set, and randomization of the run order.

Radiographic evaluation of the weld metal porosity revealed a range of contents depending on the gas jet condition used, with 25 of the 48 BOP tests meeting the internal weld quality criteria stipulated in Table IV. Using the DESIGN-EXPERT software, statistically significant factors were selected, including those supporting hierarchical terms. An analysis of variance (ANOVA) model including these terms was fitted to the natural logarithm of the response variable. A single BOP test did not fit this model and was excluded from further analysis. The t values of the factors, and their interactions, are shown in Table V. The t values suggest that variable C is several times more significant in its effect on internal porosity content than the next most significant factors, B and A. However, it should be noted that these values are only relevant to the high-low range of values used in the study, as chosen from initial screening trials, whereby some parameters may have been closer to their optimum settings than others.

Table V. ANOVA table of weld metal porosity model from the first DOE

Source	Σ squares	Df	Mean square	F value	Prob >F	Significance
Model	27.21	11	2.47	63.43	<0.0001	Significant
A-flow rate	0.79	1	0.79	20.21	<0.0001	Significant
B-angle	1.38	1	1.38	35.49	<0.0001	Significant
C-Y offset	23.04	1	23.04	590.80	<0.0001	Significant
D-X offset	0.094	1	0.094	2.41	0.1299	Not significant
AB	0.031	1	0.031	0.79	0.3793	Not significant
AC	0.29	1	0.29	7.44	0.0099	Significant
BC	0.064	1	0.064	1.64	0.2088	Not significant
BD	0.40	1	0.40	10.15	0.0030	Significant
CD	0.018	1	0.018	0.45	0.5049	Not significant
ABC	0.46	1	0.46	11.91	0.0015	Significant
BCD	0.36	1	0.36	9.15	0.0046	Significant
Lack of fit	0.14	4	0.034	1.14	0.3553	Not significant
Pure error	0.92	31	0.030
Cor total	30.41	46

All but one of the BOP tests made with a flow rate of 16l/min were silver colored in appearance, indicating adequate shielding. Conversely, most of the BOP tests produced with a flow rate of 20l/min were oxidized, ranging from light straw to blue colored. This is a result of turbulence in the shielding shoe and air being drawn in, allowing the weld metal to oxidize. Fine spatter was invariably present on the underside of all the BOP tests but could be much reduced on the top side if certain gas jet conditions were used.

3. Second design of experiments

A general factorial parametric study was designed in order to determine the optimum position of the gas jet and the flow rate through it, with respect to minimizing weld metal porosity content. The values used for these three factors were chosen based on the results of the first DOE and are given in Table III. The directed gas jet was pointed directly at the joint line, to conserve symmetry. All other parameters were fixed at the same values as used in the first DOE. A total of 36 individual BOP tests were made. The order in which these were made was again randomized. The first DOE had indicated that results were sufficiently reproducible to forgo repeats, particularly when a given combination of parameters led to a low internal porosity content weld. All BOP tests produced were fully penetrating and bright silver in color.

Statistical analysis of the porosity formation data was performed using the DESIGN-EXPERT software package. An appropriate power transformation (in this case a natural log transformation) was determined using a Box-Cox plot. Factors to be included in the polynomial model were selected using a backward selection with an Alpha out value of 0.100. Further diagnostic graphical plots (e.g. normal plot of studentized residuals, externally studentized results, and the studentized residuals versus predicted values) were utilized to check for any time-related variables unaccounted for and any data points having an excessive influence and/or leverage on the model. None of the data points were found to be problematic. An ANOVA analysis was performed and a calculated Model F-Value of 16.56 implied that the model was significant. This analysis confirms that the statistically significant terms in this model are; the jet angle off the plate surface (B), the impingement point of the jet axis ahead of the laser impingement point (C), and the measured gas flow rate (A). Additionally the factor BC was found to be significant suggesting that the height above the keyhole at which the gas jet passes is important. The terms AB, AC, and any high order terms were determined to be statistically insignificant. A cube plot of the three factor interactions is shown in Fig.2. The final equation in terms of the natural log of the accumulated length of porosity and for the actual factors is shown below;

Fig.2. Cube plot of the three factor interactions from the second DOE

It should be remembered that this equation is only valid within the range of parameters investigated in the second DOE. Comparison of the relative significances of factors determined from the first and second DOEs revealed some differences. In the first DOE, factor C was the most significant. In the second DOE, factor C was only the second most significant. However, in the second DOE a smaller range for factor C had been chosen deliberately, this being perceived as a parameter requiring tight control for achievement of low internal porosity content. Therefore, using such a small range, the second DOE determined C to be of less absolute significance than it actually was. A similar situation also held true for factor A, in that both DOEs determined this to be a less significant factor. Nevertheless, the range of factor A was tightly controlled deliberately in both DOEs, as the earlier screening trials had indicated this was necessary.

B. Optimized weld quality

Butt welds in 3.25 mm thickness Ti-2.5Cu and Ti-6Al-4V were made with the optimized directed gas jet conditions determined from the second DOE. For comparison, butt welds were also made under the same conditions without using a directed gas jet. All welds were fully penetrating and bright silver in color. Any spatter on the weld face was avoided when using the directed gas jet. The porosity content of the weld metal is shown in Fig.3. The maximum pore diameter of 0.4 mm was reduced to 0.1 mm by using the optimized directed gas jet. The accumulated length of porosity was reproducibly reduced in all cases to within internal aeroengine standards. These results suggest that these gas jet conditions can be used successfully when butt welding both Ti-2.5Cu and Ti-6Al-4V, producing welds that easily meet all the subsurface porosity criteria stipulated in stringent aeroengine standards. Figure 4 details the cross sections of butt welds made with and without a directed gas jet in both Ti-2.5Cu and Ti-6Al-4V. The welds made with a directed gas jet are free of top bead undercut, but both have a small amount of root undercut, to a maximum defect depth of 0.06 mm. Conversely, in the welds made without a directed gas jet, an unacceptable level of both top bead and root undercut is observed. Those BOP tests or butt welds produced with an optimized directed gas jet met all of the weld quality criteria stipulated in Table IV.

Fig.3. Weld metal porosity in Ti-2.5Cu BOP tests and butt welds and Ti-6Al-4V butt welds - for optimized directed gas jet conditions and no directed gas jet

 

Fig.4. Weld profiles of Ti-6Al-4V and Ti-2.5Cu butt welds produced - for optimized directed gas jet conditions and no directed gas jet

 

C. High speed video observation

Selected images from high speed observations of BOP melt runs made in Ti-2.5Cu without and with an optimized directed gas jet are shown in Fig.5. Without the gas jet the keyhole frequently collapses and much spatter is ejected from the front of the keyhole. This is most likely the result of the keyhole momentarily being covered over by molten metal (keyhole collapse), vapor pressure in the keyhole then leading to the ejection of this overlying molten material. This also explains the presence of top surface weld spatter, undercut and underfill observed in welds made without the argon gas jet. Conversely, when welding with a gas jet, the keyhole is markedly more constant in size and quiescent, and the keyhole does not collapse. Consequently, no top bead weld spatter is observed, either in the high-speed video or in the resulting welds. In addition to considerations of spatter and weld profile, a more stable keyhole process will reduce the tendency of the process itself to generate pores, from collapses of the keyhole.

 

Fig.5. High speed images of Nd:YAG laser welding of Ti-2.5Cu without (left) and with (right) an optimized a directed gas jet, 0.5 ms intervals

 

Differences in the length and width of the weldpool when using a directed jet of gas can be seen by comparing the images in Fig.5. When using the gas jet the weld pool is narrower (~0.9 mm) immediately behind where the laser beam impinges on the material surface and is also longer (~10 mm with the gas jet and ~8 mm without). After some lag, related to the low thermal conductivity of titanium, the pool then widens out to a width similar to that of the pool formed without using an Ar gas jet (~1.9 mm with the gas jet and ~2.2 mm without). Weld pool lengthening would prolong solidification time, in turn allowing more time for any gas bubbles trapped in the weldpool to escape prior to freezing of the weldpool.

D. Vapor plume analysis

An example of the emission spectrum taken when Nd:YAG laser welding of Ti-2.5Cu is shown in Fig.6. Only TiI or ArI emission lines were detected in the data. The atomic lines shown in Table VI were used to calculate the TiI electron temperature, T_e , using the equation below and the Boltzmann plot method; ^[18,19]

Fig.6. Emission spectra of the vapor plume when Nd:YAG laser welding Ti-2.5Cu within the range 450-520 nm

where I_mn is the spectral line intensity, λ_mn is the wavelength of the transition, g_m is the statistical weight at level m, A_mn is the transition probability, N is the atomic number density, h is Planck's constant, c is the velocity of light, E_m is the energy at level m, and k is the Boltzmann constant. An average vapor plume temperature of ~3100 K was calculated from the gradient of the Boltzmann plot. According to the Saha equation, and assuming a quasineutral vapor plume at atmospheric pressure, the ionization fraction of a titanium vapor plume at this temperature is <1% and the vapor plume can be categorized as a negligibly ionized vapor. Therefore any attenuation of the laser beam by inverse Bremsstrahlung absorption and/or defocusing by a gradient of electron density can be considered negligible.

Table VI. TiI emission lines used to calculate electron temperature (Ref.^[18])

Transition wavelength λ mn	Transition probability Amn (108 s-1)	Statistical weight gm	Upper energy level Em (eV)
453.324	8.83 x 10-1	11	3.5826549
468.192	2.35 x 10-2	11	2.6953807
498.173	6.60 x 10-1	13	3.336502
517.375	3.80 x 10-2	5	2.395745
521.039	3.57 x 10-2	9	2.4268626

Particle diameters collected from the condensated Ti-2.5Cu vapor plume ranged from 30-60 nm, with the majority being in the 40-50 nm range. The EDX analysis indicates that both alloying elements from the parent material (titanium and copper) were present in the vapor plume. A degree of oxidation of one or both of these elements may have also have taken place. Carbon was also detected, which most likely originated from inadvertent sampling of the adhesive film beneath the particles.

The Mie solution to Maxwell's equations has been applied to determine the extinction coefficient (combination of scattering and absorption) of the laser radiation through the vapor plume. The simplified solutions can be applied in this case as the diameter of the spheres is less than 140 nm.^[20] For calculation purposes it is assumed that the particles are all identical in size (either 40 or 50 nm) and composition and that they are isotropic and optically linear. Additionally, the particles have been assumed to be pure titanium. The complex index of refraction for titanium at 1 µm is 3.8 +4.0i (Ref.^[21]);

 

where Q_ABS and Q_SCA are the absorption and scattering terms, r is the radius of the particles, and m is the complex refractive index. The calculated absorption and scattering efficiencies for vapor plumes containing either 40 or 50 nm particles are shown in Table VII.

Table VII. Calculated Q_ABS and Q_SCA efficiencies

Particle diameter (nm)	QABS	QSCA	QEXT
40	0.0486	5.2 x 10-4	0.049
50	0.061	1.3 x 10-3	0.062

It can be seen that the absorption efficiency is the dominant beam attenuation mechanism at these particle sizes and compositions. The effect of preventing this attenuation by using the directed gas jet has been revealed through high speed video observation. However, it should be noted that these are only the efficiency values and in order to calculate the actual attenuation values the particle density needs to be calculated. These could be empirically determined through observation of the vapor plume and measurement of the vaporization mass loss.

IV. Conclusions

A three stage study was performed in order to minimize weld metal porosity content, using a directed gas jet, in Nd:YAG welding of 3.25-mm-thick Ti-2.5Cu and Ti-6Al-4V titanium alloys. Initial trials determined a suitable welding speed for full penetration and outlined the influence of a number of gas jet parameters on weld porosity content. Subsequently, two DOEs were performed to optimize the directed gas jet parameters with respect to minimizing internal porosity content, as well as weld face and root undercut. The working tolerances of parameters have been determined which allow a weld quality to within internal aerospace standards to be reproducibly achieved.

Additionally, the vapor plume above the keyhole when Nd:YAG laser welding Ti-2.5Cu has been characterized using optical spectroscopy and vapor plume sampling. The plume temperature was calculated to be ~3100 K. The main mechanisms of Nd:YAG laser beam attenuation by the vapor plume have been found to be primarily absorption but also scattering of the incident laser light by a population of 30-60 nm titanium particles. High speed observation of the welding process with and without an optimized directed gas jet has revealed that it disperses the formation of the vapor plume above the keyhole. This prevents fluctuations in the keyhole's vaporization pressure and hence increases keyhole stability thus reducing weld metal porosity.

Acknowledgments

This work was funded by the Industrial Members of TWI, as part of the Core Research Programme. Part of the work was also funded by EPSRC Grant No. C537750. Materials for the work were kindly provided by Aeromet International PLC and The Boeing Company. The high speed camera was loaned from the EPSRC Engineering Instrument Pool.

1. R. R. Boyer, 'An overview on the use of titanium in the aerospace industry,' Mater. Sci. Eng., A 213, 103-114 (1996).
2. www.boeing.com/commercial/aeromagazine/articles/qtr_4_06/AERO_Q406.pdf, date accessed, November 25, 2009.
3. D. V. Lindh and G. M. Peshak, 'The influence of weld defects on performance,' Weld. J. (London) 48, 45s-56s (1969).
4. P. Hilton, J. Blackburn, and P. Chong, 'Welding of Ti-6Al-4V with fiber delivered laser beams,' Proceedings of ICALEO 2007, Orlando, Florida, 29 October-1 November 2007 (Laser Institute of America,Orlando, 2007), pp. 887-895.
5. C. Li, H. Lunji, and G. Shuili, 'A study on the porosity of CO₂ laser welding of titanium alloy,' China Welding 15, 1-5 (2006).
6. A. Matsunawa, M. Mizutani, S. Katayama, and N. Seto, 'Porosity formation and its prevention in laser welding,' Welding International 17, 431-437 (2003).
7. W. W. Duley, 'Heat transfer and modelling in laser welding,' Laser Welding (Wiley & Sons, New York, 1999), p. 83.
8. J. Greses, 'Plasma/Plume effects in CO₂ and Nd:YAG laser welding,' Ph.D. thesis, University of Cambridge, 2003.
9. W. M. Steen, 'Laser welding,' Laser Material Processing, 2nd ed. (Springer-Verlag, London, 1998), p. 155.
10. F. Caiazzo, F. Curcio, G. Daurelio, and F. C. M. Minutolo, 'Ti6Al4V sheets lap and butt joints carried out by CO₂ laser: mechanical and morphological characterization,' J. Mater. Process. Technol.149, 546-552 (2004).
11. D. Lacroix, G. Jeandel, and C. Boudot, 'Solution of the radiative transfer equation in an absorbing and scattering Nd:YAG laser-induced plume,' J. Appl. Phys. 84, 2443-2449 (1998).
12. H. Mauro, I. Miyamoto, and Y. Arata, 'The roles of assist gas in CO₂ laser welding,' Q. J. Jpn. Weld. Soc. 3, 276-283 (1985).
13. D. Douay, F. Daniere, R. Fabbro, and L. Sabatier, 'Plasma blowing in deep penetration CO₂ laser welding,' Proceedings of ICALEO 1996, Detroit, Michigan, 14-17 October 1996 (Laser Institute of Laser, Orlando,1996), pp. 54-63.
14. K. Kamimuki, T. Inoue, K. Yasuda, M. Muro, T. Nakabayashi, and A. Matsunawa, 'Prevention of welding defect by side gas flow and its monitoring method in continuous wave Nd:YAG laser welding,' J. Laser Appl. 14,136-145 (2002).
15. S. Mueller, E. Stiles, and R. Dienemann, 'Study of porosity formation during laser welding of Ti6Al4V,' Proceedings of ICALEO 2008, Temecula, California, 20-23 October 2008 (Laser Institute of America,Orlando, 2008), pp. 133-138.
16. BS EN ISO 14175:2008, 'Welding consumables. Gases and gas mixtures for fusion welding and allied processes' (2008).
17. BS EN 1435:1997, 'Nondestructive examination of welds. Radiographic examination of welded joints' (1997).
18. physics.nist.gov/PhysRefData/ASD/index.html, date accessed March 2009.
19. H. R. Griem, 'Spectroscopic temperature measurements,' Principles of Plasma Spectroscopy (Cambridge University Press, Cambridge, England, 1997), p. 281.
20. H. C. van de Hulst, 'Non-absorbing spheres,' Light Scattering by Small Particles (Dover, Mineola, 1981), p. 144.
21. American Institute of Physics Handbook, edited by B. D. E. Gray (MacGraw-Hill, New York, 1972).