Todd Palmer

Fatigue failures in additively manufactured 316L austenitic stainless steel produced using laser powder bed fusion processes are often attributed to the presence of process-related defects or surface roughness. To better investigate the role of surface roughness on fatigue properties, internal pores were mitigated using hot isostatic pressing, and strain-controlled fatigue testing was performed with the surface in the as-deposited condition. In the absence of internal defects, crack initiation was expected to occur at the surface, which displayed arithmetic mean surface roughness (Sa) values between 10 and 40 µm as the build angle decreased from a vertical (90°) to a 45° orientation. Even though a decrease in average fatigue life with this increase in surface roughness was observed, there was no evidence of fatal cracks initiating from as-deposited surface asperities. Instead, widespread brittle intergranular fracture occurred within fine-grained sub-surface regions in the contour passes along the specimen perimeter that were populated by sub-micrometer-sized Cr₂N particles and nanometer-sized α-tridymite oxides that decorated the austenite grain boundaries. The width of these brittle fracture regions increased by 100 µm as the build angle changed from 90° to 45°. At the same time, the fraction of decorated grain boundaries (30–45%) and precipitate length (up to 10 µm) within these wider contour regions increased, driving the observed decreases in the fatigue life.

Graphical abstract

During high power laser welding of Inconel 740H, horizontal solidification cracks form at locations between approximately 70–80% of the weld depth. With the integration of circular beam oscillation, the laser energy density distribution and underlying thermo-mechanical conditions were altered. By coupling three-dimensional heat transfer, fluid flow, and stress modelling tools, the role that circular beam oscillation plays in the formation of the strain rates and stresses driving this cracking phenomenon was identified. While the addition of a circular oscillation pattern appeared to lower the stress levels along the solidification front, it did not eliminate the appearance of horizontal cracking. Only when the beam amplitudes reached 1.6 mm did solidification cracking disappear, but the weld depths were also significantly reduced.

During deep penetration laser welding of nickel alloys, interactions between composition and processing conditions can lead to the formation of defects. Inconel 740H, for example, has demonstrated a susceptibility to horizontal fusion zone cracking at locations between 70% and 80% of the weld depth during laser welding at powers above 5 kW. Coupling three-dimensional heat transfer, fluid flow, and stress modeling tools allowed that both the strain rate and stress normal to the solidification direction to be calculated. In the Inconel 740H welds, cracks formed at locations where the strain rate and stress simultaneously reached critical levels. No cracking was observed in the Inconel 690 welds, since the strain rate and stress did not simultaneously reach these critical levels.

This article focuses on the design optimization of shape memory alloy compliant mechanisms with functionally graded properties to achieve a user-defined target shape. The functional grading is approximated by allowing the geometry and the modulus of elasticity of each zone to vary. The superelastic phenomenon has been taken into account using a standard nonlinear shape memory alloy material model with linear region of higher modulus of elasticity and a superelastic region with much lower modulus of elasticity. A large deflection beam model is integrated with a multi-objective evolutionary algorithm for constrained optimization of the structure’s mechanical properties and geometry. Examples illustrate the trade-offs between the objectives of minimizing shape error, maximum stress, and volume. It is observed that in the optimized designs, the elastic modulus and the geometry work together in regions where large flexibility is required to achieve the target shape.

Fixturing of components during laser cladding can incur significant conductive thermal losses. However, due to the surface roughness at contact, interfacial conduction is impeded. The effective contact conductivity, known as gap conductance, is much lower than the contacting material conductivities. This work investigates modeling conduction losses to fixturing bodies during laser cladding. Two laser cladding experiments are performed using contrasting fixturing schemes: one cantilevered substrate with a minimal substrate-fixture contact area and one with a substrate bolted to a work bench, with a significant substrate-fixture contact area. Using calibrated gap conductance values, error for the cantilevered fixture model decreases from 20.5% to 6.49% in the contact region, while the bench fixtured model error decreases from a range of 60–102% to 11–45%. The improvement in accuracy shows the necessity of accounting for conduction losses in the thermal modeling of laser cladding, particularly for fixturing setups with large areas of contact.

The accurate modeling of thermal gradients and distortion generated by directed energy deposition additive manufacturing requires a thorough understanding of the underlying physical processes. One area that has the potential to significantly affect the accuracy of thermomechanical simulations is the complex forced convection created by the inert gas jets that are used to deliver metal powder to the melt pool and to shield the laser optics and the molten material. These jets act on part surfaces with higher temperatures than those in similar processes such as welding and consequently have a greater impact on the prevailing heat transfer mechanisms. A methodology is presented here which uses hot-film sensors and constant voltage anemometry to measure the forced convection generated during additive manufacturing processes. This methodology is then demonstrated by characterizing the convection generated by a Precitec^® YC50 deposition head under conditions commonly encountered in additive manufacturing. Surface roughness, nozzle configuration, and surface orientation are shown to have the greatest impact on the convection measurements, while the impact from the flow rate is negligible.

Topology optimization is a well-established engineering practice to optimize the design and layout of parts to create lightweight and low-cost structures, which have historically been difficult, or impossible, to make. Additive Manufacturing (AM) provides the freedom to fabricate the complex and organic shapes that topology optimization often generates. In this paper we use topology optimization to create lightweight designs while conforming to additive manufacturing constraints related to overhanging features and unsupported surfaces when using metallic materials. More specifically, we use design for additive manufacturing (DfAM) rules along with topology optimization to study the tradeoffs between the weight of the part, support requirements, manufacturing costs, and performance. The case study entails redesigning an upright on the SAE Formula student racecar to reduce support structures and manufacturing and material cost when using Direct Metal Laser Sintering (DMLS). Manufacturing the optimized design without applying DfAM rules required support material up to 202.4% of the volume of the model. Using DfAM, the upright is redesigned and manufactured with supports requiring less than 15% of the volume of the model. The results demonstrate the challenges in achieving a balance between weight reduction, manufacturing costs, and factor of safety of the design.

Nickel Titanium (NiTi) shape memory alloys (SMAs) exhibit shape memory and/or superelastic properties, enabling them to demonstrate multifunctionality by engineering microstructural and compositional gradients at selected locations. This paper focuses on the design optimization of NiTi compliant mechanisms resulting in single-piece structures with functionally graded properties, based on user-defined target shape matching approach. The compositionally graded zones within the structures will exhibit an on demand superelastic effect (SE) response, exploiting the tailored mechanical behavior of the structure. The functional grading has been approximated by allowing the geometry and the superelastic properties of each zone to vary. The superelastic phenomenon has been taken into consideration using a standard nonlinear SMA material model, focusing only on 2 regions of interest: the linear region of higher Young’s modulus of elasticity and the superelastic region with significantly lower Young’s modulus of elasticity. Due to an outside load, the graded zones reach the critical stress at different stages based on their composition, position and geometry, allowing the structure morphing. This concept has been used to optimize the structures’ geometry and mechanical properties to match a user-defined target shape structure. A multi-objective evolutionary algorithm (NSGA II - Non-dominated Sorting Genetic Algorithm) for constrained optimization of the structure’s mechanical properties and geometry has been developed and implemented.

Laser‐fired contacts (LFCs) are typically fabricated with nanosecond pulse durations despite the fact that extremely precise and costly control of the process is necessary to prevent significant ablation of the aluminum metallization layer. Microsecond pulse durations offer the advantage of reduced metal expulsion and can be implemented with diffractive optics to process multiple contacts simultaneously and meet production demands. In this work, the influence of changes in laser processing parameters on contact morphology, resistance, and composition when using microsecond pulses has been fully evaluated. Simulated and experimental results indicate that contacts are hemispherical or half‐ellipsoidal in shape. In addition, the resolidified contact region is composed of a two‐phase aluminum–silicon microstructure that grows from the single‐crystal silicon wafer during resolidification. As a result, the total contact resistance is governed by the interfacial contact area for a three‐dimensional contact geometry rather than the planar contact area at the aluminum–silicon interface in the passivation layer opening. The results also suggest that for two LFCs with the same size top surface diameter, the contact produced with a smaller beam size will have a 25–37% lower contact resistance, depending on the LFC diameter, because of a larger contact area at the LFC/wafer interface. Copyright © 2014 John Wiley & Sons, Ltd.

The utility of a new laser interferometric technique, inline coherent imaging, for real time keyhole depth measurement during laser welding is demonstrated on five important engineering alloys. The keyhole depth was measured at 200 kHz with a spatial resolution of 22 μm using a probe beam, which enters the keyhole coaxially with the process beam. Keyhole fluctuations limited average weld depth determination to a resolution on the order of 100 μm. Real time keyhole depth data are compared with the weld depths measured from the corresponding metallographic cross-sections. With the exception of an aluminium alloy, the technique accurately measured the average weld depth with differences of less than 5%. The keyhole depth growth rates at the start of welding are measured and compare well with order of magnitude calculations. The method described here is recommended for the real time measurement and control of keyhole depth in at least five different alloys.

Low resistance laser-fired ohmic contacts (LFCs) can be formed on the backside of Si-based solar cells using microsecond pulses. However, the impact of these longer pulse durations on the dielectric passivation layer is not clear. Retention of the passivation layer during processing is critical to ensure low recombination rates of electron-hole pairs at the rear surface of the device. In this work, advanced characterization tools are used to demonstrate that although the SiO2 passivation layer melts directly below the laser, it is well preserved outside the immediate LFC region over a wide range of processing parameters. As a result, low recombination rates at the passivation layer/wafer interface can be expected despite higher energy densities associated with these pulse durations.

In laser-based direct energy deposition additive manufacturing, process control can be achieved through a closed loop control system in which thermal sensing of the melt pool surface is used to adjust laser processing parameters to maintain a constant surface geometry. Although this process control technique takes advantage of important in-process information, the conclusions drawn about the final solidification structure and mechanical properties of the deposited material are limited. In this study, a validated heat transfer and fluid flow laser welding model are used to examine how changes in processing parameters similar to those used in direct energy deposition processes affect the relationships between top surface and subsurface temperatures and solidification parameters in Ti-6Al-4V. The similarities between the physical processes governing laser welding and laser-based additive manufacturing make the use of a laser welding model appropriate. Numerical simulations show that liquid pools with similar top surface geometries can have substantially different penetration depths and volumes. Furthermore, molten pool surface area is found to be a poor indicator of the cooling rate at different locations in the melt pool and, therefore, cannot be relied upon to achieve targeted microstructures and mechanical properties. It is also demonstrated that as the build temperature increases and the power level is changed to maintain a constant surface geometry, variations in important solidification parameters are observed, which are expected to significantly impact the final microstructure. Based on the results, it is suggested that the conclusions drawn from current experimental thermography control systems can be strengthened by incorporating analysis through mathematical modeling.

Laser doping to form selective emitters offers an attractive method to increase the performance of silicon wafer based photovoltaics. However, the effect of processing conditions, such as laser power and travel speed, on molten zone geometry and the phosphorus dopant profile is not well understood. A mathematical model is developed to quantitatively investigate and understand how processing parameters impact the heat and mass transfer and fluid flow during laser doping using continuous wave lasers. Calculated molten zone dimensions and dopant concentration profiles are in good agreement with independent experimental data reported in the literature. The mechanisms for heat (conduction) and mass (convection) transport are examined, which lays the foundation for quantitatively understanding the effect of processing conditions on molten zone geometry and dopant concentration distribution. The validated model and insight into heat and mass transport mechanisms also provide the bases for developing process maps, which are presented in part II. These maps illustrate the effects of output power and travel speed on molten zone geometry, average dopant concentration, dopant profile shape, and sheet resistance.

Laser doping is an attractive way to manufacture a selective emitter in high efficiency solar cells, but the underlying phenomena, which determine performance, are not well understood. The mathematical model developed in Part I solves the equations of conservation of mass, momentum, and energy and is used here to investigate the effects of processing parameters on molten zone geometry, average phosphorus dopant concentration, dopant profile shape, and sheet resistance. The empirically calculated sheet resistance values are in good agreement with independently measured sheet resistance values reported in the literature. Process maps for output power and travel speed show that molten zone geometry and sheet resistance are more sensitive to output power than travel speed. The highest molten zone depth-to-width aspect ratios and lowest sheet resistances for 532 nm laser beams are obtained at higher laser powers (>13 W) and lower travel speeds (<2 m/s). Once the power level is set, the travel speed can be varied for further optimization of dopant concentration and geometry.

The evolution of temperature and velocity fields during laser processing of solar cells to produce an ohmic contact between an aluminum thin film and a silicon wafer is studied using a transient numerical heat transfer and liquid metal flow model. Since small changes in pulse duration, power, and power density can result in significant damage to the substrate and, in extreme cases, expulsion of droplets from the molten zone, the selection of optimal laser processing parameters is critical. The model considers the unusually large heat of fusion of the Al-Si alloy formed during processing and the large composition-dependent two phase region. The calculated size and shape of the fusion zone were in good agreement with the corresponding experimental data, indicating the validity of the model and providing a basis for using the model to develop a better understanding of the laser-assisted fabrication of contacts for solar cell devices. The transient changes in the composition of the Al-Si molten region are found to have a major impact on the heat transfer during the formation of the contact. Consideration of the time-dependent concentration of Al in the molten region is also essential to achieve good agreement between the experimental and computed molten pool sizes. Process maps showing peak temperatures and the depth and width of the molten pool are presented in order to assist users in the selection of safe process parameters for the rapid fabrication of these silicon-based photovoltaic devices.

The austenitisation ( α→ γ) transformation in a 1005 C–Mn steel is monitored in real time at continuous heating rates between 1 and 10°C s⁻¹ using in situ synchrotron X-ray diffraction and validated using dilatometry. Experimental validation is provided for austenitisation models that predict that the austenitisation transformation proceeds through multiple mechanisms. At temperatures below the A1 transformation temperature, the starting microstructure undergoes recovery and recrystallisation to relieve stress imparted during the initial thermomechanical processing of the steel. The austenitisation transformation follows, beginning at the A1 temperature, with the initial transformation proceeding as the pearlite in the microstructure is dissolved and high carbon concentration austenite is formed. Since the carbon is localised near the original pearlite colonies, there is a pronounced heating rate dependent delay before the remaining low C ferrite grains begin to transform. The transformation reaches completion at temperatures above the A3 temperature, and the last ferrite to be transformed is nearly pure iron.

The austenitisation ( α→ γ) transformation in a 1005 C–Mn steel is monitored in real time at continuous heating rates between 1 and 10°C s⁻¹ using in situ synchrotron X-ray diffraction and validated using dilatometry. Experimental validation is provided for austenitisation models that predict that the austenitisation transformation proceeds through multiple mechanisms. At temperatures below the A1 transformation temperature, the starting microstructure undergoes recovery and recrystallisation to relieve stress imparted during the initial thermomechanical processing of the steel. The austenitisation transformation follows, beginning at the A1 temperature, with the initial transformation proceeding as the pearlite in the microstructure is dissolved and high carbon concentration austenite is formed. Since the carbon is localised near the original pearlite colonies, there is a pronounced heating rate dependent delay before the remaining low C ferrite grains begin to transform. The transformation reaches completion at temperatures above the A3 temperature, and the last ferrite to be transformed is nearly pure iron.

Process control in electron beam welding is typically based on control of machine settings, such as accelerating voltage, beam current, focus coil current, and vacuum level. These settings, though important, provide little insight into the characteristics of the beam used to make the weld. With the enhanced modified Faraday cup (EMFC) diagnostic tool, these beam characteristics, including the peak power density, full width at half maximum, and full width at 1∕e2 values, can be quantified. The use of this diagnostic tool in an extended production run at Lawrence Livermore National Laboratory (LLNL) is described. Results show that machine performance, in terms of these measured beam characteristics, varies over time when the EMFC is not used to adjust the machine settings. Testing has shown that the variability of the beam characteristics can be measurably decreased with the use of the EMFC diagnostic tool. With the implementation of this diagnostic tool in the process control procedures, every electron beam weld, which encompassed approximately 90 welds over an 18month time frame, met all of the requirements defined in the weld process specification and passed all of the postweld quality control checks. The results also show that variations in each of the measured beam parameters can be controlled at levels below ±2.2%, which is smaller than the 5% tolerance band suggested by ASME for other welding parameters. Such an enhanced level of control allows product throughput to be increased by decreasing the number of rejected parts through the elimination of unexpected variations in beam characteristics. The benefits of integrating this diagnostic tool into future process control regimes are also discussed.

An experimental test melt of a boron alloyed 9Cr-3W-3Co-V,Nb steel for high temperature applications in the thermal power generation industry was produced by vacuum induction melting. This grade of steel typically displays a homogeneous tempered martensitic microstructure in the as-received, i. e. normalised and tempered, condition. However, after welding, this microstructure is significantly altered, resulting in a loss of its desired properties. The phase transformations during simulated thermal cycles typical of those experienced in the weld heat-affected zone were directly observed by in-situ X-ray diffraction experiments using synchrotron radiation. Heating rates of 10 K s⁻¹ and 100 K s⁻¹ up to a peak temperature of 1300°C are investigated here. The final microstructures observed after both simulated weld thermal cycles are primarily composed of martensite with approximately 4% retained delta ferrite and 4% retained austenite, by volume. With the temporal resolution of the in-situ X-ray diffraction technique, phase transformations from tempered martensite to austenite to delta ferrite during heating and to martensite during cooling were monitored. With this technique, the evolution of the final microstructure through both heating and cooling is monitored, providing additional context to the microstructural observations.

In situ X-ray diffraction methods have been developed at Lawrence Livermore National Laboratory for direct observation of microstructural evolution under quasi-steady state and transient welding conditions. Using intense highly collimated synchrotron radiation, the crystal structures in the weld heat affected and fusion zones are probed in real time to monitor solidification and solid state phase transformations during welding. Here the authors review recent work on the development and use of the time resolved X-ray diffraction (TRXRD) technique during transient welding and illustrate its unique capabilities to: directly observe the solidification mode; discover, in real time, the definitive sequence of phase transformations that lead to the final microstructure; and provide quantitative kinetic data of phase transformations through synthesis of TRXRD data with the temperature history obtained through heat transfer modelling. The TRXRD technique has been used to investigate welding induced phase transformations in titanium alloys, low alloy steels, and stainless steel alloys. The results show some of the first real time observations of the weld solidification mode and the evolution of equilibrium and non-equilibrium phases during rapid heating and cooling. When combined with numerical modelling, quantitative phase transformation kinetic data are obtained, allowing for its use in a wide variety of isothermal and non-isothermal processing. The potential for future applications of this and similar techniques is also addressed.

Using the enhanced modified Faraday cup (EMFC), the differences in the beams produced by two electron beam welders are characterised at different focus settings and work distances. For example, EMFC measurements show that sharply focused beams display different shapes and peak power densities which vary by nearly 20% for the same welding parameters on these two welders. Increases in work distance on each machine were shown to result in decreases in both the peak power density and the resulting weld size and shape. Because of the differences in machine performance, additional differences also arise when comparing the welds produced by each machine. These different weld dimensions are attributed to differences in the beam shape and a 70 mm difference in the theoretical beam crossover location in the upper column of the two welders. The crossover location, which can not be physically measured, is determined using the EMFC by analysing the beam distribution parameters of sharply focused beams over a range of work distances. By combining these results with simplified optics calculations, the magnification of the beam optics can be determined and the machine performance of each welder characterised. The work distance on each machine at which beams with similar peak power density values will be produced can then be determined. With this knowledge, changes in either the beam focus or work distance can be made to attain similar beams from different welders, thus providing a baseline for developing modern weld transfer procedures.

In situ x-ray diffraction was performed while annealing thin film Au∕Cu binary diffusion couples to directly observe diffusion at elevated temperatures. The temperature dependence of the interdiffusion coefficient was determined from isothermal measurements at 700, 800, and 900°C, where Cu and Au form a disordered continuous face centered cubic solid solution. Large differences in the lattice parameters of Au and Cu allowed the initial diffraction peaks to be easily identified, and later tracked as they merged into one diffraction peak with increased diffusion time. Initial diffusion kinetics were studied by measuring the time required for the Cu to diffuse through the Au thin film of known thickness. The activation energy for interdiffusion was measured to be 65.4kJ∕mole during this initial stage, which is approximately 0.4× that for bulk diffusion and 0.8× that for grain boundary diffusion. The low activation energy is attributed to the high density of columnar grain boundaries combined with other defects in the sputter deposited thin film coatings. As interdiffusion continues, the two layers homogenize with an activation energy of 111kJ∕mole during the latter stages of diffusion. This higher activation energy falls between the reported values for grain boundary and bulk diffusion, and may be related to grain growth occurring at these temperatures which accounts for the decreasing importance of grain boundaries on diffusion.

In situ time-resolved x-ray diffraction (TRXRD) experiments were used to track the evolution of the α→β→L→β→α/α′ phase transformation sequence during gas tungsten arc welding of Ti–6Al–4V. Synchrotron radiation was employed for the in situ measurements in both the fusion zone (FZ) and the heat-affected zone (HAZ) of the weld, providing information about transformation rates under rapid heating and cooling conditions. The TRXRD data were coupled with the results of computational thermodynamic predictions of phase equilibria, and numerical modeling of the weld temperatures. The results show that significant superheat is required above the β transus temperature to complete the α→β transformation during weld heating, and that the amount of superheat decreases with distance from the center of the weld where the heating rates are lower. A Johnson–Mehl–Avrami phase transformation model yielded a set of kinetic parameters for the prediction of the α→β phase transformation during weld heating. Corresponding TRXRD measurements were made during weld cooling. In the HAZ, the β→α transformation during weld cooling was shown to initiate at the β transus temperature and terminate below the Ms temperature, resulting in a microstructure containing a substantial fraction of α′ martensite. In the FZ, the β→α transformation during weld cooling was shown to initiate below the Ms temperature, and to completely transform the microstructure to α′ martensite.

In situ time-resolved x-ray diffraction (TRXRD) experiments were used to track the evolution of the α→β→L→β→α/α′ phase transformation sequence during gas tungsten arc welding of Ti–6Al–4V. Synchrotron radiation was employed for the in situ measurements in both the fusion zone (FZ) and the heat-affected zone (HAZ) of the weld, providing information about transformation rates under rapid heating and cooling conditions. The TRXRD data were coupled with the results of computational thermodynamic predictions of phase equilibria, and numerical modeling of the weld temperatures. The results show that significant superheat is required above the β transus temperature to complete the α→β transformation during weld heating, and that the amount of superheat decreases with distance from the center of the weld where the heating rates are lower. A Johnson–Mehl–Avrami phase transformation model yielded a set of kinetic parameters for the prediction of the α→β phase transformation during weld heating. Corresponding TRXRD measurements were made during weld cooling. In the HAZ, the β→α transformation during weld cooling was shown to initiate at the β transus temperature and terminate below the Ms temperature, resulting in a microstructure containing a substantial fraction of α′ martensite. In the FZ, the β→α transformation during weld cooling was shown to initiate below the Ms temperature, and to completely transform the microstructure to α′ martensite.

Understanding the evolution of microstructure in welds is an important goal of welding research because of the strong correlation between weld microstructure and weld properties. To achieve this goal it is important to develop a quantitative measure of phase transformations encountered during welding in order to ultimately develop methods for predicting weld microstructures from the characteristics of the welding process. To aid in this effort, synchrotron radiation methods have been developed at Lawrence Livermore National Laboratory for direct observations of microstructure evolution during welding. Using intense, highly collimated synchrotron radiation, the atomic structure of the weld heat affected and fusion zones can be probed in real time. Two synchrotron-based techniques have been developed for these investigations, known as spatially resolved (SRXRD) and time resolved (TRXRD) x-ray diffraction, and these techniques have been used to investigate welding induced phase transformations in titanium alloys, low alloy steels, and stainless steel alloys. This paper will provide a brief overview of the application of these methods to understand microstructural evolution during the welding of low carbon (AISI 1005) and medium carbon (AISI 1045) steels, where the different levels of carbon influence the starting microstructures and the evolution of microstructures during welding.

In situ spatially resolved x-ray diffraction (SRXRD) experiments were used to directly observe the heat-affected zone phases present during gas tungsten arc welding of a Ti–6Al–4V alloy. The experiments were performed at the Stanford Synchrotron Radiation Laboratory using a 250 μm diam x-ray beam to gather real-time experimental information about the α−Ti→β−Ti phase transition during weld heating. Six different welding conditions were investigated using SRXRD to experimentally determine the extent of the single phase β-Ti region surrounding the liquid weld pool. These data were compared to predicted locations of the β-Ti phase boundary determined by calculated weld thermal profiles and equilibrium thermodynamic relationships. The comparison shows differences between the experimentally measured and the calculated locations of the β-Ti boundary. The differences are attributed to kinetics of the α−Ti→β−Ti phase transition, which requires superheating above the β-Ti transus temperature to take place during nonisothermal weld heating. Analysis of the results reveal that the transition to β-Ti requires an additional 3.96 s (±0.29 s) of time and 169 °C (±25.7 °C) of superheat above the β-Ti transus temperature to go to completion under an average weld heating rate of 42.7 °C/s.

In situ spatially resolved x-ray diffraction (SRXRD) experiments were used to directly observe the heat-affected zone phases present during gas tungsten arc welding of a Ti–6Al–4V alloy. The experiments were performed at the Stanford Synchrotron Radiation Laboratory using a 250 μm diam x-ray beam to gather real-time experimental information about the α−Ti→β−Ti phase transition during weld heating. Six different welding conditions were investigated using SRXRD to experimentally determine the extent of the single phase β-Ti region surrounding the liquid weld pool. These data were compared to predicted locations of the β-Ti phase boundary determined by calculated weld thermal profiles and equilibrium thermodynamic relationships. The comparison shows differences between the experimentally measured and the calculated locations of the β-Ti boundary. The differences are attributed to kinetics of the α−Ti→β−Ti phase transition, which requires superheating above the β-Ti transus temperature to take place during nonisothermal weld heating. Analysis of the results reveal that the transition to β-Ti requires an additional 3.96 s (±0.29 s) of time and 169 °C (±25.7 °C) of superheat above the β-Ti transus temperature to go to completion under an average weld heating rate of 42.7 °C/s.

Ferrite (δ)-austenite (γ) transformations in the heat affected zone (HAZ) of a gas tungsten arc weld in 2205 duplex stainless steel are observed in real time using spatially resolved X-ray diffraction (SRXRD) with high intensity synchrotron radiation. A map showing the locations of the δ and γ phases with respect to the calculated weld pool dimensions has been constructed from a series of SRXRD scans. Regions of liquid, completely transformed γ, a combination of partially transformed γ with untransformed δ, and untransformed γ + δ are identified. Analysis of each SRXRD pattern provides a semiquantitative definition of both the δ/γ phase balance and the extent of annealing, which are mapped for the first time with respect to the calculated weld pool size and shape. A combination of these analyses provides a unique real time description of the progression of phase transformations in the HAZ. Results show that during heating, δ and γ both show signs of annealing as temperatures approach 550 ° C. The δ phase then starts to transform to γ as temperatures approach 700 ° C. Although supported by thermodynamic calculations, this δ → γ transformation during heating has not been directly observed until now. Following this reaction, the HAZ microstructure evolves in three different ways. For peak temperatures less than approximately 1100 ° C, δ retransforms, reverting to its original base metal fraction on cooling. When peak temperatures exceed approximately 1375 ° C, the microstructure completely transforms to δ before retransforming to a mixture of δ and γ during weld cooling. For peak temperatures between 1100 and 1375 ° C, γ is only partially transformed during both heating and cooling. Using these real time observations, important kinetic information about the transformations occurring in duplex stainless steels during both non-isothermal heating and cooling of welding can be determined.

Ferrite (δ)-austenite (γ) transformations in the heat affected zone (HAZ) of a gas tungsten arc weld in 2205 duplex stainless steel are observed in real time using spatially resolved X-ray diffraction (SRXRD) with high intensity synchrotron radiation. A map showing the locations of the δ and γ phases with respect to the calculated weld pool dimensions has been constructed from a series of SRXRD scans. Regions of liquid, completely transformed γ, a combination of partially transformed γ with untransformed δ, and untransformed γ + δ are identified. Analysis of each SRXRD pattern provides a semiquantitative definition of both the δ/γ phase balance and the extent of annealing, which are mapped for the first time with respect to the calculated weld pool size and shape. A combination of these analyses provides a unique real time description of the progression of phase transformations in the HAZ. Results show that during heating, δ and γ both show signs of annealing as temperatures approach 550 ° C. The δ phase then starts to transform to γ as temperatures approach 700 ° C. Although supported by thermodynamic calculations, this δ → γ transformation during heating has not been directly observed until now. Following this reaction, the HAZ microstructure evolves in three different ways. For peak temperatures less than approximately 1100 ° C, δ retransforms, reverting to its original base metal fraction on cooling. When peak temperatures exceed approximately 1375 ° C, the microstructure completely transforms to δ before retransforming to a mixture of δ and γ during weld cooling. For peak temperatures between 1100 and 1375 ° C, γ is only partially transformed during both heating and cooling. Using these real time observations, important kinetic information about the transformations occurring in duplex stainless steels during both non-isothermal heating and cooling of welding can be determined.

Although nitrogen concentrations at levels much higher than Sieverts’ Law predictions during the arc welding of iron and steel are well established, there is currently no commonly accepted methodology to determine this concentration quantitatively. The nature and concentrations of various species in the plasma phase above the weld pool surface are therefore investigated in the present work using both theoretical and experimental techniques. A comprehensive thermodynamic analysis of the nitrogen containing plasma phase of a gas tungsten welding arc shows that ionised species dominate close to the electrode, whereas neutral monatomic and diatomic nitrogen are the primary species near the metal surface at plasma temperatures as low as 5000 K. When oxygen is added to a nitrogen containing plasma, the resulting nitrogen concentration in the weld metal is further enhanced. Definitive proof is provided for a mechanism in which nitrogen and oxygen species interact in the plasma phase at temperatures below 6000 K, resulting in a significant increase in the concentration of monatomic nitrogen. Furthermore, at plasma temperatures as low as 5000 K, the equilibrium monatomic nitrogen partial pressure is sufficiently high to cause nitrogen saturation in the weld metal. Emission spectroscopy of glow discharge plasmas validates both the species density calculations and the presence of NO in the nitrogen and oxygen containing plasmas.