Is becoming a welding engineer agood career choice today? Weldingis at the heart of many great engi-

neering achievements. It is essentialfor mechanization of agriculture, gen-eration of energy, distribution of cleanwater, and production of medical de-vices. True, in this digital age, self-driving cars, robots for remote sur-gery, and other products of emergingtechnologies seem more exciting. But,today’s dizzying pace of progress inengineering often merges the time-tested mature fields such as weldingwith the new fields like digital dataprocessing into a powerful stream forthe benefit of all people. Welding today is much more ad-vanced than it was just a few decadesago, and I hope this article will helpyou make an informed career choice. From about 3000 welds in a car tonumerous joints in large buildings,bridges, and other important struc-tures (Fig. 1), welds are everywhere.Transportation, manufacturing, elec-tronics, and other industries that sup-port our standard of living depend onit. Welding-related expenditures bythese industries in the United Stateswere about $34 billion in 2000, whichwas about one tenth the cost of all carssold annually. While the joining of metals hasbeen practiced since prehistoric times,modern welding technology began af-ter electricity became available in the19th century. Since then, both its en-gineering practice and the underlyingscientific knowledge base have ma-tured. Today, structurally sound andreliable joints of numerous engineer-ing alloys, including many that werepreviously considered difficult to weld,can be made with confidence.

WELDING JOURNAL / APRIL 201558

Welding in the Digital AgeA professor asks young people to choosea career in welding engineering

BY TARASANKAR DEBROY

Fig. 1 — The stunning Grand Canyon Skywalk was built with round-the-clock welding in10-h shifts. Designed to withstand large earthquakes, this huge structure consists of threesteel plates, 3200 lb each. The structure houses a 3-in.-thick, heat-strengthened glasswalkway (Ref. 1).

DebRoy Feature April 2015_Layout 1 3/12/15 1:51 PM Page 58

APRIL 2015 / WELDING JOURNAL 59

Because of its close connection toconstruction projects, many viewwelding as a primitive and dangerousart. At any sizable construction site,sparks fly, fumes spread their mustyaroma, and large skull and crossboneswarning posters speak volumes aboutwelding’s hazards. However, the realityis very different from the perception.The welding industry has an excellentsafety record, robots now performmany of the repetitious and difficulttasks, and a mature and sophisticatedscientific knowledge base supportswelding practices. Most exciting, sincethe 1970s, the expanding digital dataprocessing capabilities have been com-bined with the well-established tech-nological knowledge base of welding,totally transforming both its practiceand the underlying analytical capabili-ty that supports it. Analytical capability is importantfor problem solving because problemsin welding often affect life and proper-ty. This article focuses on uncovering along-standing mystery in welding thatremained elusive until the tools of thedigital age were used. In a larger con-text, it shows how the renaissancebrought about by the fusion of matureand new technologies has taught engi-neers powerful lessons while providingsignificant benefits to all people.

A Welding Primer As you likely know, the purpose ofwelding is to combine two parts —metallic materials in most cases —into a strong joint. Several commonterms used in describing different re-gions of the weld are shown in Fig. 2(Ref. 2). In fusion welding processes,the joint forms by the melting and so-lidification of the metal parts. The re-gion under the heat source meltsforming a liquid metal puddle calledthe weld pool. A small solid regionnext to the weld pool, where the struc-ture and properties of the workpieceare changed by heat, is called the heat-affected zone (HAZ). The size andshape, or geometry, of the weld pool is

affected by how much heat is absorbedand distributed within it. Tempera-tures vary within the weld pool, andheat flows by conduction from high to low temperatures and by convectionfrom the motion of the hot liquid metal. The molten metal within the weldpool circulates under the action of sev-eral forces. The most important,Marangoni force, is named after Ital-ian scientist Carlos Marangoni, whoshowed that liquids move from re-gions of low to high surface tension.The nature of this force can be easilyunderstood from the motion of pepperin water — Fig. 3. When a cotton swabdipped in household soap is immersedin water, the pepper moves away fromthe swab (Ref. 3) (Fig. 3, right).Adding soap to water reduces its sur-face tension locally. Water flows awayfrom the low surface tension region towhere the surface tension is relativelyhigh (Ref. 3). A similar effect causesweld metal to flow within the weldpool. Surface tension of the weld metaldepends on temperature. So its valuejust under the heat source is differentfrom that in other regions and this dif-ference drives the flow of weld metal. The gravitational force tends tosink the colder, heavier liquid near theedge of the weld pool and raise thehotter, lighter liquid metal in the mid-dle of the weld pool. In addition, dur-ing arc welding, an electromagneticforce is generated from the interactionbetween the current path in the weldpool and the magnetic field it gener-ates. Of these three forces, the gravita-tional force is by far the weakest, andduring arc welding, the electromagnet-ic force is comparable to the Marango-ni force only at fairly high currents. Inmost cases, the Marangoni force pro-vides the main driving force for theflow of weld metal within the weldpool. The rolling streams of weld met-

Fig. 2 — A schematic of the fusion welding process is shown at left and a transverse crosssection, perpendicular to the direction of welding, is shown at right (Ref. 2).

Fig. 3 — The photo on the left shows pepper on the water’s surface and a cotton swab thatwas dipped in soap. The photo at right shows the pepper moves away from the center afterthe swab is dipped into the water (Ref. 3).

Fig. 4 — Progression of computational hardware from mechanical to digital devices.

DebRoy Feature April 2015_Layout 1 3/12/15 1:52 PM Page 59

WELDING JOURNAL / APRIL 201560

al carry heat from underneath the heatsource to all other locations within theweld pool. Its circulation determinesthe melting pattern of the various re-gions of the workpiece, the shape andsize of the weld pool, and the structureand properties of the welded joint. Butthe weld metals are opaque and veryhot. As a result, the actual velocitiesand temperatures within the weld poolhave not been experimentally meas-ured so far. A welcome recourse emerged in the1970s. Advancements in computerhardware and software (Fig. 4) madefluid flow and heat transfer calcula-tions accurate and affordable. Engi-neers now routinely use these calcula-tions in critical designs in aeronauti-cal, aerospace, civil, and other engi-neering disciplines. In welding, thereare many important problems thatcannot be solved without these calcu-lations, at least not easily.

Billions of EquationsSolved Instantly Evolution of computational hard-ware and software from mechanical toanalog to digital calculations has im-proved both the theory and practice ofwelding. The combination of digitalcomputers and robots has improvedjoint quality, enhanced safety, and tak-en the boredom out of repetitiouswelding in automotive and other in-dustries. Clearly, a new manufacturing

paradigm has emerged. What is lessapparent but equally important is theadvancement of analytical ability forproblem solving and design based onfundamental principles. There are compelling reasons fordetailed understanding of heat trans-fer and fluid flow in welding. Both thetemperatures and velocities at all loca-tions in the weld pool affect not justits shape and size, but the mixing ofthe filler metals, cooling rates at dif-ferent locations, vaporization of alloy-ing elements, weld metal composition,and the structure and properties ofthe joint. Local temperatures and ve-locities can be calculated by solvingequations of conservation of mass,momentum, and energy (Ref. 4). Sincethese equations are too complex to besolved analytically, an appropriate nu-merical method is needed. A typicalnumerical solution procedure starts bydividing the workpiece into many

small volumes or cells, typically about250,000 cells. For each cell, an algebra-ic equation relates the local values of avariable with its values at the neigh-boring cells (Ref. 5). Typically, thevariables include three components ofvelocities, enthalpy or temperature,and pressure, which are solved repeat-edly until correct solutions are ob-tained. For these five variables, a totalof 5 × 250,000 or 1.25 million equa-tions have to be solved for each at-tempt at solution, commonly called aniteration. In most cases, several thou-sand iterations are needed before cor-rect solutions for the variables at allcells are obtained. So, several billionequations are solved cumulatively toget temperatures and velocities in theentire workpiece. Today, about a bil-lion such linear algebraic equationscan be solved in about two minutes us-ing inexpensive laptops. Typical computed temperature andvelocity fields during gas tungsten arcwelding are shown in Fig. 5. The figureshows regions of different tempera-tures by specific color bands. Since theheat source is moving, the tempera-ture changes rapidly in the cold work-piece ahead of the moving weld pool.Behind the weld pool where the mate-rial has already been heated, the metalcools slowly in air and the tempera-tures change more gradually. On theweld pool surface, liquid metal movesaway from the low surface tension re-gion under the heat source to other re-gions where the surface tension ishigher. The surface is depressed belowthe arc because it exerts pressure onthe liquid surface and forms a smallhump behind the arc. The velocitiesrange from a few tens of centimetersper second to about a meter per sec-ond, and the liquid metal carries a sig-nificant amount of heat from underthe heat source to all other locationswithin the weld pool.

Fig. 5 — Computed flow of weld metal during arc welding. The colors represent tempera-tures in K and the dotted lines represent the lines of flow of liquid. The two loops shownnear the surface are from the Marangoni flow and the two loops below the surface resultfrom electromagnetic force (Ref. 4).

Fig. 6 — Weld cross sections of 15-mm-thick, high-speed steel plates containing 0.9%C,3.9%Cr, 6.3%W, 4.8%Mo, 1.8%V, 4.6%Co, 0.2%Mn, 0.5%Si by weight containing 20 ppmsulfur (left) and 150 ppm sulfur (right) spot welded at a laser power of 5200 W for 5 s (Ref.11).

DebRoy Feature April 2015_Layout 1 3/12/15 1:52 PM Page 60

APRIL 2015 / WELDING JOURNAL 61

An Enduring Mystery

The Puzzle and Its Importance

Failure to reproduce experiments isunacceptable in science and so, whenthe same grade of steels are weldedunder the exact same welding condi-tions, it would be absurd to expectweld-to-weld variations in geometry.In reality, this totally unexpected be-havior was the norm when the samegrade of steel with minor variations incomposition was welded (Refs. 6–10).Figure 6 shows cross sections of twowelds fabricated using the exact sameprocedure from the same grade of steelthat are strikingly different (Ref. 11).The main difference in the steels wasthe amount of sulfur, which differedby 130 parts per million (ppm) byweight. Finding a solution to this long-standing puzzle (Refs. 6–10) was im-portant because the weld geometry af-fects its performance; however, the so-lution remained elusive for decades(Refs. 8–10).

A Promising Hypothesis

A team of scientists at the RockyFlats plant, a former nuclear weaponsproduction facility near Denver, Colo.,(Refs. 8–10) first presented a promis-ing solution to this mystery in the ear-ly 1980s. They proposed a hypothesisto explain why a small amount of sele-nium (Ref. 6) or sulfur (Ref. 7) in steelsignificantly increases the depth ofpenetration. They considered how sulfur affectsthe surface tension of liquid steel,weld metal spin, convective heat trans-fer, and the resulting weld pool geome-

try. Figure 7 shows the surface tensionof pure iron decreases with tempera-ture. The same trend is observed forsteels (very low sulfur). However,when a small amount of sulfur is pres-ent, the surface tension decreasesoverall and increases with temperatureas shown in the figure. At tempera-tures close to the boiling point, thesurface tension decreases with in-crease in temperature. Sulfur andmany other alloying elements such asoxygen, nitrogen, selenium, and tel-lurium have a tendency to migrate tothe surface of the liquid steel. They allaffect the surface tension in a mannersimilar to sulfur and are called surface-active elements (Ref. 12). Directly under the heat source, theliquid metal has the highest tempera-ture and lowest surface tension whenthe steel contains practically no sulfur.Since liquids flow from low to highsurface tension regions, hot liquidsteel moves sideways from the middleto the edge of the weld pool and meltsmetal there. It then turns downwardas shown in Fig. 8A. As a result, theweld pool becomes wide and shallow. Small additions of sulfur changethe flow pattern completely. Hot liq-uid under the heat source now has ahigher surface tension than that in thecooler regions — Fig. 7. So, on the sur-face of the weld pool, the weld metalrushes to the middle then movesdownward to the bottom of the weldpool. The downward flow of the hotmetal in the middle of the weld poolworks like a thermal drill and a deepweld pool forms, as shown in Fig. 8B.The Rocky flats team also showed thatselenium affected the shape of theweld pool just like sulfur. The hypothesis the Rocky Flatsteam proposed provided a plausible ex-planation. However, in order for theirtheory to gain traction, direct proof of

changes in the flow of liquid weld met-al was required. Since metals areopaque and the liquid weld metal isvery hot, this was not an easy task.They added some tiny alumina parti-cles that floated on the surface of theweld (Ref. 9) and used a high-speedcamera to film their motion duringwelding. The evidence was now athand. Sulfur does change the flow pat-tern of liquid metal (Ref. 9). Insightfuland elegant, their work inspired manyother researchers.

Helpful but Incomplete

The work at Rocky Flats explainedwhy a small amount of sulfur or sele-nium changed the shape of welds forthe conditions of their welding. But af-ter more than a decade, when experi-ments were conducted to cover moreextensive welding conditions, it wasfound that sulfur does not alwayschange the shape and size of the weldpool, although it does so in many cases(Ref. 11). So, the mystery actuallydeepened. The powerful spark of anew idea incubated at the Rocky Flatsplant still required more work for adeeper understanding of when sulfurchanges the shape and when it doesnot, and why. The Rocky Flats team explained therole of selenium and sulfur assumingconvection as the mechanism of heattransfer. This mechanism, valid onlywhen velocities within the weld poolare large, was indeed valid for their ex-periments. However, the assumed con-vective heat transfer mechanism is notalways valid, because the velocities aresmall for certain welding conditions. Arigorous understanding of the role ofsurface-active elements for a specificwelding condition requires mechanis-tic insight of heat transfer achievablethrough a combination of experiments

Fig. 8 — A — Pure iron flows sideways from the middle, making the weld pool wide andshallow; B — when a small amount of sulfur is added, the alloy goes downward in themiddle of the weld pool resulting in a deep weld pool.

Fig. 7 — Variation of surface tension withtemperature (Ref. 12).

DebRoy Feature April 2015_Layout 1 3/12/15 1:52 PM Page 61

WELDING JOURNAL / APRIL 201562

and mathematical modeling. The first computational study (Ref.13) of the effects of surface-active ele-ments in welding was published by ateam at MIT in 1983 that investigatedconvection in arc weld pools. Theyshowed that, in many cases, the sur-face tension driven flow dominatesthe convection in the arc weld pool.But more important, they also provedthat a small amount of sulfur or sele-nium influenced the direction andmagnitude of the liquid metal flow, abehavior the Rocky Flats team ob-served experimentally. These calcula-tions, with assumed weld pool shapeand size, were a giant step forward be-cause they provided a world of insightthat could not be obtained by any oth-er means.

Only the Numbers Reveal theWhole Truth

The influence of selenium or sulfurdepends on the mechanism of heattransfer which, in turn, is determinedby the magnitude of the velocities andthe thermal conductivity of the liquidmetal. If the velocities are small forthe conditions of welding, the direc-tion of liquid metal circulation doesnot affect the shape of the weld pool.As a result, sulfur and other surface-active elements do not always affectthe weld pool shape. Only comprehensive computer sim-ulations can reveal the velocity fieldsand the mechanism of heat transfer.Such calculations show when surface-active elements affect weld pool geom-etry and when they do not. Two weld-ing experiments and their computersimulations are presented here toshow how the same level of sulfur mayor may not affect the weld pool geom-etry depending on the weldingconditions. A side by side comparison of experi-mentally determined and numericallycomputed weld pool cross sections(Ref. 11) of two high-power laser spotwelds of steels is shown in Fig. 9. For asulfur content of 20 ppm, calculationsshow a strong convection currenttransports liquid metal from the mid-dle of the weld pool sideways. The sur-face velocities are fairly large, on theorder of about 20 cm/s, and at this ve-locity, convection is the main mecha-nism of heat transfer. In comparison,heat transfer by conduction is negligi-ble (Ref. 11). The molten metal flows

sideways from the middle, forming ashallow weld pool, as shown in Fig. 9A. When the sulfur content is 150ppm, the circulation pattern is oppo-site to what was observed for the 20ppm sulfur steel weld shown in Fig.9B. The surface velocities are fairlylarge, higher than 20 cm/s. So the heatis carried mostly by convection, andconduction heat transfer is unimpor-tant (Ref. 11). Hot weld metal flowsdownward under the heat source, thebase metal melts near the root, and adeep weld pool forms. The computedweld pool geometry agrees well withthe experimentally determinedgeometries in both cases. But sulfur does not always changethe weld pool geometry (Ref. 11). Fig-ure 10 shows no perceptible differencein the cross sections of low-powerlaser welds in steel containing 20 and150 ppm of sulfur. The numerical sim-ulation of heat transfer and fluid flowreveals why. The computed results show fairlylow peak temperatures and lower ve-locities in the weld pool for these smallwelds. Convection did not carry muchheat since the velocities in both cases

were weak. As a result, conduction wasthe main mechanism of heat transfer.The direction of spin of the weld metalwas opposite in the two cases as ex-pected, but since conduction was themechanism of heat transfer, the op-posing spin did not result in any dif-ference in geometry (Ref. 11). Themechanism of heat transfer was themost important factor, not the con-centration of sulfur or the direction ofweld metal spin. Dramatic effects of sulfur, seleni-um, and other surface-active elementsknown for many decades led many tobelieve these elements always affectedweld pool geometry. In fact, only whenconvection is the dominant mecha-nism of heat transfer can the surface-active elements play an important rolein affecting weld geometry. Solving a compelling problem oflack of reproducibility of the weldgeometry has made the world a saferplace for all people. But does the inno-vation and discovery stop once a long-standing mystery is solved? Not at all,because new welding problems that af-fect life and property arise frequently.The following example shows that the

Fig. 9 — Experimentally determined and theoretically calculated weld pool geometries in a15-mm-thick, high-speed steel plate spot laser welded for 5 s. The welds had 20 and 150ppm sulfur on the left and right sides, respectively (Ref. 11).

Fig. 10 — Comparison of the computed and experimental weld pool geometries at a laserpower of 1900 W for steels containing A — 20 ppm; B — 150 ppm sulfur (Ref. 11).

DebRoy Feature April 2015_Layout 1 3/12/15 1:53 PM Page 62

powerful tools of the digital age canaccelerate the pace of solution andhelp in producing reliable and better-engineered welds.

Sulfur Strikes Back When each half of the joint con-tains steels with significantly differentconcentrations of sulfur, a totally un-expected result is observed. Since thearc is positioned just above the origi-nal joint of the two plates, it is expect-ed that both plates melt equally. In-stead, melting occurs mainly in thelow-sulfur plate (Ref. 14) — Fig.11. The point of maximum pene-tration, B, is laterally shiftedfrom the location of the originaljoint, A. The figure shows theweld bead has clearly shifted to-ward the plate with lower sulfurcontent. The extent of shift dependedon the difference between thesulfur concentrations of the twoplates and the heat input perunit length (Refs. 14, 15). Fur-thermore, the arc was asymmet-ric with a flare toward the lowsulfur side (Ref. 15). To examinethe role of arc flaring and theMarangoni convection, experi-ments were also done with alaser beam to avoid the effect ofarc flaring (Ref. 16). Pronouncedcenterline shift and selectivemelting of the low-sulfur steel werestill observed when a laser beam wasplaced directly above the original jointinterface. Numerical modeling estab-lished that Marangoni convection wasan important factor in transportingmetal from high- to low-sulfur steel,which caused selective melting of thelow-sulfur plate (Ref. 14). A pro-nounced rotational asymmetry (Ref.16) of the weld bead during laser weld-ing of steels with dissimilar sulfur con-centrations was also observed. Amechanistic understanding of the ro-tational asymmetry still remains to bedeveloped (Ref. 17). So, the enduringmystery of the surface-active elementsstill persists in a different form.

Welding Engineering as a Career The welding mystery described hereshows the diversity of scientific sub-fields within welding engineering.

Many of today’s welding engineershave academic degrees in metallurgyand materials, mechanical, electrical,and several other branches of engi-neering. But a degree is just the begin-ning. From heat transfer to robotics,computers, control theory, corrosion,materials performance, and proper-ties, there are many technical areas of opportunity for lifelong on-the-joblearning. Highly sought after in the automo-tive, aerospace, construction, energy,shipbuilding, electronics, and appli-ance industries, welding engineers

perform many important tasks. Weld-ing process selection, quality control,code compliance, and design of hard-ware and software are just a few exam-ples of the crucial tasks for welding en-gineers in diverse activities rangingfrom underwater construction to nu-merous manufacturing processes tobuilding a spaceship. Some welding en-gineers work in research and develop-ment in universities, national labs,and industrial labs to solve problemsand advance the knowledge base thatsupports the practice of welding. After four years of engineering col-lege, you will be among a highly selectgroup of people, much smaller than1% of the population, with technicalskills that are essential in today’sworld. If you select the fascinatingfield of welding engineering as a ca-reer, you will have an awesome oppor-tunity to assimilate new contemporarytechnologies into welding and improveour world in numerous ways. Betterwelding can build more reliable ma-

chinery for better agriculture, hous-ing, energy, clean water, transporta-tion, health care, and practically allequipment that support our standardof living. I hope you will consider theexciting field of welding engineeringas a career.

Epilogue and Lessons Synthesis of the knowledge base ofa mature and important field such aswelding with the emerging awesomedigital data processing capabilities hashelped in the production of better

welds and made the world a bet-ter place. Apart from more rigor-ous analysis and solution ofcomplex problems, the synthesisof welding and computationalcapabilities has also incubated atransformative new technology.Additive manufacturing, whichhas been hailed as the future ofmanufacturing, evolved fromthis merger. It starts with a digi-tal picture of a part in a comput-er and builds it by adding liquidmetal, layer by layer. Machinesthat use an electron beam weld-ing gun and deposit 7 to 20 lb ofmetal per hour to make largeparts are already available (Ref.18). This disruptive additivemanufacturing process is an ex-ample of welding’s evolution inthe digital age and proof that

better welding can build a better worldfor all.

References

1. Welding the world’s highest walkway.2006. Welding Journal 85(10): 40, 41. 2. DebRoy, T., and David, S. A. 1992.Physical processes in fusion welding. Sci-ence 257: 497–502. 3. http://video.mit.edu/watch/the-marangoni-effect-how-to-make-a-soap-pro-pelled-boat-13540/ Snapshots from a videodownloaded on 30 June 2014. 4. DebRoy, T. 1995. Role of interfacialphenomena in numerical analysis of weld-ability. Mathematical Modelling of Weld Phe-nomena II. London, UK: The Institute ofMaterials, pp. 3–21. 5. Kou, S., and Sun, D. K. 1985. Fluidflow and weld penetration in stationary arcwelds. Metallurgical Transactions A – Physi-cal Metallurgy and Materials Science 16(2):203–213. 6. Linnert, G. E. 1967. Weldability ofaustenitic stainless steel as affected byresidual elements. Effects of Residual Ele-ments on Properties of Austenitic Stainless

APRIL 2015 / WELDING JOURNAL 63

Fig. 11 — Weld geometry when welding two stainless steelplates with different sulfur contents. The white vertical linepassing through point “A” indicates the original interface ofthe two plates. The location of maximum weld penetrationis indicated by point C, and AB indicates the shift of themaximum penetration from the original joint of the twoplates (Ref. 14).

WJ

DebRoy Feature April 2015_Layout 1 3/12/15 1:53 PM Page 63

Steels. Am. Soc. for Testing and Materials,Publication No. 418, pp. 105–119. 7. Bennett, W. S., and Mills, C. S. 1974.CTA weldability studies on high man-ganese stainless steel. Welding Journal53(12): 548-s to 553-s. 8. Heiple, C. R., and Roper, J. R. 1981.Effect of selenium on GTAW fusion zonegeometry. Welding Journal 60(8): 143-s to145-s. 9. Heiple, C. R., and Roper, J. R. 1982.Mechanism for minor element effect onGTA fusion zone geometry. Welding Journal61(4): 97-s to 102-s. 10. Heiple, C. R., and Roper, J. R., Stag-ner, R. T., and Aden, R. J. 1983. Surface ac-tive element effects on the shape of GTA,laser, and electron beam welds. WeldingJournal 62(3): 72-s to 77-s. 11. Paischeneder, W., DebRoy, T.,Mundra, K., and Ebner, R. 1996. WeldingJournal 75(3): 71-s to 80-s. 12. Sahoo, P., DebRoy, T., and McNal-lan, M. J. 1988. Metallurgical Transactions B19: 483–491. 13. Oreper, G. M., Eagar, T. W., andSzekely, J. 1983. Convection in arc weldpools. Welding Journal 75(3): 307-s to 312-s. 14. Mishra, S., Lienert, T. J., Johnson,M. Q., and DebRoy, T. 2008. Acta Materi-alia 56: 2133–2146. 15. Rollin, A. F., and Bentley, M. J.1984. Proceedings of the International Con-ference on the Effects of Residual, Trace andMicro-Alloying Elements on Weldability andWeld Properties. Cambridge, UK: TWI, p. 9. 16. Lienert, T. J., Burgardt, P., Harada,K. L., Forsyth, R. T., and DebRoy, T. 2014.Weld bead centerline shift during laserwelding of austenitic stainless steels withdifferent sulfur content. Scripta Materialia71: 37–40. 17. Pal, S., Manvatkar, V., DebRoy, T.,and Lienert, T. J. 2014. Rotational asym-metry in steel welds with dissimilaramounts of sulfur, unpublished documen-tation, Department of Materials Scienceand Engineering, June. 18. Additive Manufacturing, SciakyInc., www.sciaky.com/additive_manufactur-ing.html accessed on 2 May 2014.

Acknowledgments

Many thanks to Ashwin Raghavan,Jared Blecher, Dr. Chobi DebRoy, Dr.Thomas J. Leinert, Dr. John W. Elmer,Dr. David G. C. Robertson, Dr. Stan A.David, Dr. Harry K.D.H. Bhadeshia,Dr. Ronald M. Latanision, Dr. KwadwoOsseo-Asare, and Dr. Thomas W. Ea-gar for their interest.

WELDING JOURNAL / APRIL 201564

TARASANKAR DEBROY (rtd1@psu.edu) isprofessor, Materials Science and Engineering,

Penn State University, University Park, Pa.

For info, go to www.aws.org/ad­index

For info, go to www.aws.org/ad­index

DebRoy Feature April 2015_Layout 1 3/12/15 1:54 PM Page 64