Introduction

Cladding the surface of a steel sub-strate with a layer of nickel (Ni) alloyembedded with hard tungsten carbide(WC) particles can significantly im-prove the wear and corrosion resist-ance of the steel. Various processeshave been used for hardfacing with Ni-WC, for instance, plasma transferredarc welding (PTAW), submerged arcwelding (SAW), laser beam welding

(LBW) and cladding, and gas metal arcwelding (GMAW) (Ref. 1). The PTAWprocess is the most common and effi-cient method for shop production, butit is impractical for field welding appli-cations (Ref. 2). Gas metal arc weldingis a low-cost alternative to PTAW andLBW. It is suitable for cladding oversmall areas that need protection byhardfacing, cladding small internal di-ameters, or repairing small damagedareas in existing cladding. Choi et al. (Ref. 2) made single-bead

Ni-WC cladding using a Lincoln Elec-tric Power Wave 455M/STT powersource for GMAW. The filler metal was1.6- mm-diameter Arctec TungcoreFCS cored wire. The heat input variedwith the filler metal transfer mode:lowest with the short-circuiting mode,higher with the globular mode, andhighest with the spray mode. It wasshown that increasing the heat inputincreased both dilution and carbidedissolution. Vespa et al. (Ref. 3) madesingle-bead Ni-WC cladding using aJetline Engineering process controllercalled CSC-MIG along with a MillerElectric XMT 350 CC/CV powersource. The filler metal was a 1.6-mmPolyTung NiBWC cored electrode wiremanufactured by Polymet Corp. It wasalso reported that increasing the heatinput increased WC dissolution. In general, the microstructure ofNi-WC cladding made by variousprocesses consists of “chunks of un-melted tungsten carbide in a nickel-base alloy matrix,” according toISO/TR Technical Report 13393 (Ref.4). The microstructure of Ni-WCcladding made by GMAW has been dis-cussed by Choi et al. (Ref. 2) and Vespaet al. (Ref. 3), which will be mentionedsubsequently in the present study. The cladding of Ni-WC has excep-tional abrasion resistance and moder-ate corrosion resistance (Ref. 4). How-ever, increasing the heat input can in-crease the dissolution of tungsten car-bide particles and reduce the abrasionresistance. It also increases the dilu

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 451-s

SUPPLEMENT TO THE WELDING JOURNAL, DECEMBER 2016Sponsored by the American Welding Society and the Welding Research Council

Ni­WC Hardfacing by Gas Metal Arc Welding

GMAW­CSC was used to widen the operation window, 3D printing approach to select weaving pattern, advanced EPMA to analyze microstructure,

and thermodynamics to show WC loss

BY P. YU, X. CHAI, D. LANDWEHR, AND S. KOU

ABSTRACT Nickel cladding reinforced with hard tungsten carbide particles was deposited onsteel by both conventional gas metal arc welding (GMAW) and GMAW controlled shortcircuiting (GMAW­CSC). Single­bead hardfacing was deposited at various heat inputs. Thecurrent/voltage waveforms were recorded during welding. It was found that the windowof welding parameters for making a smooth cladding without much spatter was signifi­cantly wider with GMAW­CSC than conventional GMAW. While the cladding dilution bysteel increased with increasing heat input as expected, the dilution increase was lesswith GMAW­CSC than conventional GMAW. A low­cost 3D printer­based substrate manip­ulation table was built to move the steel plate under a stationary welding gun for deposi­tion on a designated area on steel. Thus, even when a robot is unavailable to weave thegun over the steel, the effect of the welding gun’s weaving pattern on the resultantcladding can still be examined, e.g., the uniformity of the cladding thickness and the par­ticle distribution in the cladding. A state­of­the­art FESEM+EPMA with a very small beamdiameter of 80 to 100 nm was used to identify various phases inside both the tungstencarbide particles and the matrix. An eight­layer square cladding was deposited and itsoverall composition measured. Thermodynamic calculations based on this compositionshowed that substituting Ni with Cr can degrade the cladding by causing brittle Cr­con­taining carbide to form at the expense of WC.

KEYWORDS • Hardfacing • Cladding • Controlled Short Circuiting • Tungsten Carbide • Nickel

P. YU and S. KOU are with the Department of Materials Science and Engineering, University of Wisconsin, Madison, Wis. X. CHAI is withNovelis Global Research & Technology Center, Kennesaw, Ga. D. LANDWEHR is with Fisher Barton Technology Center, Watertown, Wis.

Yu 12-16_Layout 1 11/9/16 8:25 AM Page 451

tion of the cladding by the substrate.Dilution can have a significant effecton the performance of the cladding. For corrosion resistance applica-tions, the corrosion performance ofsome cladding can be rather sensitiveto the change in dilution. For wear re-sistance applications, dilution can affect the severity of cracking encoun-tered. For erosion plus corrosion appli-cations, cracking of the hardfacing al-loy can provide a corrosion path to thesubstrate. Hence, a two-layer claddingmay help, with the first layer being thecorrosion resistant alloy and the sec-ond layer being the hardfacing alloy.A crack-free hardfacing alloy may alsohelp in some cases. The present study used GMAW-CSC and 1.6-mm PolyTung NiBWCcored welding wire to make thecladding. However, it differed fromthe study of Vespa et al. (Ref. 3) inseveral ways. First, conventionalGMAW was also used and the resultswere compared with those of GMAW-CSC. Second, instead of just single-bead cladding, a 3D printer-basedsubstrate manipulation table wasbuilt to make the cladding over asquare area, both single layer andmultiple layers. Third, the overallcomposition of the cladding wasmeasured. Fourth, the effect of Cr onWC dissolution was shown by ther-modynamic analysis.

Experimental Procedure

Materials

The workpiece was 1018 steel, 150mm long, 50 mm wide, and 6.4 mmthick. Single-bead welding was con-ducted in the length direction alongthe centerline of the workpiece, atthe travel speed of 15 mm/s and vari-ous heat inputs. The filler metal fordepositing the cladding was a Poly-Tung NiBWC flux cored wire of 1.6-

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95452-s

Fig. 1 — 3D printer (lower left corner) supporting and generating 3D motionof the steel workpiece under the welding gun to allow printing filler metalon steel.

Table 1 — Nominal Chemical Composition (wt­%) of PolyTung NiBWC (Ref. 5)

W Si B Ni

Wt­% 38–45 2.2 1.0 Balance

Table 2 — Welding Parameters for Conventional GMAW

Samples Welding Feeding Traveling Speed Electrode Extension Heat Input Gas Flow Rate Shielding Gas Voltage (V) Speed mm/s mm/s (mm) (J/mm) (m3/h)

52 18 29.6 15 22.5 99.5 1.1 75% Ar 25% CO2

53 19 29.6 15 22.5 107 1.1 75% Ar 25% CO2

54 20 29.6 15 22.5 111 1.1 75% Ar 25% CO2

55 21 29.6 15 22.5 120 1.1 75% Ar 25% CO2

Yu 12-16_Layout 1 11/9/16 8:25 AM Page 452

mm outer diameter. The nominalchemical composition (wt-%) provid-ed by the manufacturer is shown inTable 1 (Ref. 5).

Welding Processes

Two different versions of gas metalarc welding (GMAW) were used toclad. The first version was convention-al GMAW, consisting of an Invision456P power source, an XR-M wirefeeder, and a conventional weldinggun. The welding conditions areshown in Table 2. The second versionwas a controlled-short-circuiting (CSC)type GMAW process called GMAW-CSC, consisting of an Invision 456Ppower source, a CSC process con-troller, and a special welding gunhooked up to the controller. Theprocess controller coordinates thefeeding and speed of the wire elec-trode with the level of welding currentdelivered by the power source (Ref. 6).

Briefly, the welding process has twoprimary phases: the arc phase duringwhich heat is generated to melt thebase metal, and the short-circuit phaseduring which the filler metal droplet isdeposited when the welding wiremakes contact with the weld pool (Ref.6). The controller monitors the voltagebetween the electrode and the work-piece to determine which phase theprocess is in at any given time. Thecontroller clears the short by retract-ing the wire to the preset arc lengthlevel. Once the arc is established again,the controller begins feeding the wiretoward the weld pool, and the cycle re-peats. GMAW-CSC was originally de-veloped and called “CSC-MIG” byMiller Electric Manufacturing Co. andsubsequently manufactured by JetlineEngineering, Irvine, Calif. The waveforms of the welding cur-rent in GMAW-CSC can be tailored ingreat detail in order to optimize thewelding process and reduce spatter.Examples of the operating parameters

that can be specified include: 1) cur-rent levels and durations (for the startand mid periods) of the arc phase andthose of the short-circuit phase, 2) thewire down speed, the delay before wiredown, the wire up speed and the delaybefore wire up, 3) the arc length, and4) the penetration delay. The weldingconditions are shown in Table 3. Thedesign and selection of the welding pa-rameters were based on the experienceof the authors in using the GMAW-CSC process in the past few years. Ascompared to conventional GMAW,GMAW-CSC has been reported to re-duce heat input and spatter (Refs.7–9). Reducing heat input is desirablebecause it can help reduce dilution ofthe cladding by the base metal. Reduc-ing spatter is also desirable to maxi-mize deposit efficiency. The waveforms of the welding cur-rent and arc voltage were recorded usinga computer data acquisition system to-gether with LabView software. Thedata-sampling rate for each signal was15,000 Hz. The heat input per unitlength of the weld Q was calculated us-ing the following equation (Ref. 8):

(1)

where I is the current, E is the voltage,t is the welding time, and U is the travel speed.

3D Printer

A cladding may need to be de-posited on a designated area on acomponent that requires highwearresistance or where the damagedcladding requires repairing. In such asituation, the pattern of motion ofthe GMAW welding gun relative tothe workpiece may affect the qualityof the resultant cladding. To studythe effect of the motion pattern, a3D printer can be useful, especiallywhen a much more expensive robotor 3-axis CNC machine is not avail-able. So, a low-cost open-source 3Dprinter was built as shown in Fig. 1,similar to that developed by Anza-lone et al. (Ref. 10). Essentially, theplatform that supports the workpiececan move the workpiece horizontallyaccording to the specific motion pat-tern programmed by the computer,and it can also move vertically to

Q = I E( )dt0

t/ (t U )

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 453-s

Fig. 2 — PolyTung NiBWC tubular wire. A — Transverse cross section; B — particles re­moved from inside wire; C — identification of particles by EDS analysis.

ABC

Yu 12-16_Layout 1 11/9/16 8:26 AM Page 453

keep the arc length constant as the cladding getsthicker and thicker during deposition. The weldinggun stays stationary while the workpiece movesaround. So, in addition to single-bead cladding, single-layersquare cladding was made over a 27 ¥ 27 mm area onthe steel. Cladding was also made with four layers andeight layers. The eight-layer cladding was used to de-termine its overall chemical composition.

Overall Cladding Composition

In order to provide enough material for the manysamples needed for chemical composition analysis, aneight-layer cladding was made over a square area of31.5 ¥ 31.5 mm on the 1018 steel plate. The claddingwas about 6.7 mm thick. After welding, the claddingwas cut off by electrical discharge machining (EDM)for the purpose of chemical analysis, from 1 mm abovethe steel plate. In other words, the amount of samplefor the chemical analysis was 31.5 ¥ 31.5 ¥ 5.7 mm involume. The cut-off sample was then broken up into

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95454-s

Fig. 3 — Waveforms of current and voltage. A — Sample 53 (conven­tional GMAW); B — Sample 55 (conventional GMAW); C — Sample 50(GMAW­CSC).

Fig. 4 — Top views of welds made by conventional GMAW atvarious voltages and heat inputs. A — Weld No. 52; B —Weld No. 53; C —Weld No. 54; D — Weld No. 55. Travelspeed: 15 mm/s.

A

A

B

B

C

C

D

Yu 12-16_Layout 1 11/9/16 8:27 AM Page 454

granules and dissolved in variouschemicals for chemical analysis. Thechemical analysis was conducted atthe National Analysis Center of Ironand Steel (NACIS) in Beijing, China,according to the National Standards

of China (GB/T)and NACIS stan-dards (Refs.11–16).

MicrostructuralAnalysis of theCladding

SXFiveFE, astate-of-the-artfield emissionelectron probe mi-cro analyzer(EPMA) for quan-titative analysisand x-ray map-ping at high-spa-tial resolution,was used (Ref.17). Up to fivewavelength dis-persive spectrom-eters (WDS) couldbe fit into the mi-croprobe for high-precision quanti-tative analysis.The beam diame-ter used was80–100 nm(called “0 m”),the voltage andcurrent being 8kV and 20 nA, re-spectively. Thevolume below thesample surface af-fected by thebeam was about250 nm in diame-ter (Ref. 18). Thesamples were pol-ished but notetched, in ordernot to affect thecompositionmeasurements byEPMA. A plasmacleaner (IBSSGV10X) was usedto clean the sam-ple surface. Thestandards for cali-bration for EPMA

included pure WC, W, C, Ni, Fe, Si,and B. EPMA was done using Ka x-raylines of B, O, C, and Si, whereas theW Ma line and the Ni Ll lines wereused for those elements. Crystalsused were: LPC2 for C and B; LPET

for W and Si; and LLIF for Fe, TAP forNi, and PCO for 0. Counting timeswere 10 s on peak and 5 s each ontwo backgrounds. During the initialstage of the present study, SEM andenergy dispersive spectrometry(EDS) were also used to identify theparticles removed from inside the tu-bular filler metal. A Bruker D8 Dis-cover diffractometer along with a mi-crofocus x-ray source and a Vantecarea detector was also used to identi-fy phases in the cladding.

Results and Discussion

Figure 2 shows a transverse crosssection of the PolyTung NiBWC flux-cored wire and the particles removedfrom inside. EDS analysis indicates thetube sheath as Ni and the main parti-cles as Ni, WC, and W.

Single­Bead Cladding

Figure 3 shows example waveformsof current and voltage recorded duringwelding. Consider conventionalGMAW first. Welds No. 53 was madeat 19 V. Its waveforms in Fig. 3A showthe voltage approaches zero periodi-cally. This suggests the short-circuit-ing mode of metal transfer. Weld No.55 was also made by conventionalGMAW but at a slightly higher voltageof 21 V. As shown by its waveforms inFig. 3B, the voltage approaches zeroonly occasionally and not quite asclose to zero. This suggests the modeof metal transfer is more like globularthan short circuiting. As for theGMAW-CSC, the waveforms of WeldNo. 50 in Fig. 3C are typical of con-trolled short circuiting. When shortcircuiting occurs, the maximum cur-rent is always kept low (< 150 A) in-stead of being allowed to have a sud-den surge to cause spatter (e.g., about400 A in Fig. 3B). It was found that the window ofwelding parameters was significantlywider with GMAW-CSC than conven-tional GMAW. Consider conventionalGMAW first. As shown in Fig. 4A,Weld No. 53 made at the arc voltage19 V (107 J/mm) is smooth with onlyslight spatter. However, Weld No. 52made at the arc voltage 18 V (99.5J/mm) is highly irregular and discon-tinuous in shape. For Weld No. 55

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL455-s

Fig. 5 — Top views of welds made by GMAW­CSC at various volt­ages and heat inputs. A — Weld No. 46; B — Weld No. 47; C —Weld No. 48; D — Weld No. 49; E — Weld No. 50; F — Weld No. 51.Travel speed: 15 mm/s. Window for welding is significantly wider inGMAW­CSC than in conventional GMAW (Fig. 4).

A

B

C

D

E

F

Yu 12-16_Layout 1 11/9/16 8:28 AM Page 455

made at 21 V (120 J/mm), the beadwas continuous but with large spatter.With GMAW-CSC, on the other hand,a smooth single bead with slight spat-ter could be made over the much widerrange of 56.3 to 120 J/mm as shownin Fig. 5. Examples of the transverse cross-sections of single-bead cladding madeby conventional GMAW are shown inFig. 6. Similar results are shown in Fig.7 for cladding made by GMAW-CSC. Ineither case, the cladding is not verydensely packed with large tungstencarbide particles. To increase the frac-tion of large tungsten carbide particlesin the cladding, a tubular welding wirewith a thin sheath and small crimpoverlap can be used as pointed out byMendez et al. (Ref. 1). The PolyTungNiBWC flux-cored wire of 1.6-mm out-er diameter, which has a relativelythick sheath and large crimp overlapas can be seen in Fig. 2A, was selectedfor the present study mainly becauseof the intended comparison with thestudy of Vespa et al. (Ref. 3), in whichthe same wire was used. From the transverse cross section ofthe cladding, the extent it is diluted bythe melted base metal can be deter-mined. Table 4 summarizes the dilu-tions in the single-bead cladding. Asshown in Fig. 6, with conventionalGMAW, the dilution varies from 2.6% at107 J/mm to 12% at 120 J/mm. WithGMAW-CSC, as shown in Fig. 7, the di-lution varies from 0.6% at 75 J/mm to5.4% at 107 J/mm and 5.7% at 120

J/mm. Thus, the dilu-tion varies less inGMAW-CSC than con-ventional GMAW. Thedissolution of WC par-ticles appears to in-crease with increasingheat input as can be seen by comparing,for instance, Fig. 6A with 6B.

Square Cladding Made with 3D Printer

Since the dilution increases lesswith increasing heat input in GMAW-CSC than conventional GMAW, theformer was used to deposit claddingover a square area (about 27 ¥ 27mm). Three single-layer squarecladdings, made at the travel speed of15 mm/s, are discussed below. The pattern of the welding gun/steel relative motion was found to af-fect the thickness uniformity andoverall dilution of the resultantcladding. Figure 8 shows a single-lay-er square cladding (No. 13) made witha square spiral pattern. The claddingbecomes thinner and thinner as itscenter is approached. Based on theenlarged transverse cross section ofthe cladding (not shown), the overalldilution is 2.28%. The micrographstaken at three different locations inthe transverse cross section show WCparticles are fewer and smaller nearthe center of the cladding. Figure 9shows a single-layer square cladding

(No. 14) made with a short serpen-tine motion pattern that covers theleft one-third of the square cladding,reverses its direction to cover themiddle one-third, and then reversesits direction again to cover the rightone-third. As compared to claddingNo. 13 in Fig. 8, its thickness is moreuniform but the dilution was higherat 6.01%. Figure 10 shows a single-layersquare cladding (No. 11) made with alonger serpentine motion patternthat covers the entire square claddingwithout reversing its direction like incladding No. 14 — Fig. 9. The overalldilution is at 4.6%. As compared tocladding No. 13, No. 11 appears to bemore uniform in thickness and have amore uniform distribution of tung-sten carbide particles. It is likely that,with a closed-loop control strategy,the pool size can be better controlledto reduce variations in the uniformityand thickness of the cladding and itsdilution by the substrate. However,the same pool size on the surface maynot necessarily mean the same pooldepth, which affects the level of dilu-tion. By the way, some soot is visiblein the middle of the top surface ofcladding No. 11 — Fig. 10B. The soot

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95456-s

Fig. 7 — Transverse cross sections of single­bead cladding de­posited by GMAW­CSC. A — Sample No. 47; B — sample No.50; C — sample No. 51. The average of three dilutions atthree different locations along each sample is shown.

Fig. 6 — Transverse cross sections of single­bead cladding de­posited by conventional GMAW. A — Sample No. 53; B —sample No. 55. The average of three dilutions at three differ­ent locations along each sample is shown.

A

B

A

B

C

Yu 12-16_Layout 1 11/9/16 8:28 AM Page 456

was likely caused by surface oxidationafter welding as the current and volt-age waveforms indicated the weldingprocess was stable. The results of the single-layersquare cladding indicate that, out ofthe three motion patterns investigat-ed, the long serpentine pattern shownin Fig. 10A seems to yield the best re-sults, that is, uniform cladding thick-ness and moderate dilution of 4.6%.Thus, a four-layer cladding was madewith the serpentine pattern, that is,cladding No. 9. As shown in Fig. 11,the cladding is essentially uniform inthickness and in the distribution ofWC particles. To help reduce overheat-ing, the travel speed was gradually in-creased from 16.7 mm/s at the begin-ning of the first layer to 26.7 mm/s atthe end of the first layer. It was fur-ther increased gradually to 33.3 mm/sbefore the end of the second layer. Tohelp distribute heat more uniformly,the pattern was rotated 90 deg aftereach layer. It was noticed that the tip of the Nisheath was deformed into a bell shape,as shown in Fig. 12. The foam-like ma-terial visible outside the bell is likelyto be a flux or binder added during thefabrication of the welding wire. Ascompared to GMAW-CSC with a solidwire (Refs. 8, 9), the level of spattershown in Figs. 5 and 12 is much high-er even though GMAW-CSC is knownto be very effective in eliminatingspatter. When the tip of a solid wireforms a liquid bridge with the weldpool, the controller can clearly sense avoltage of nearly zero — Fig. 3C. How-ever, with the tip of a tubular wire likethat shown in Fig. 12, the liquid bridgemay not be as easy to form and con-trol. Consequently, more spatter canbe expected with a tubular wire. Some

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 457-s

Fig. 8 — Single­layer square cladding No.13 made by GMAW­CSC. A — Pattern ofmotion of workpiece relative to weldinggun; B — top view; C — transverse cross­section; D, E, F — optical micrographs. Di­lution = 2.28%.

A

D E F

C

B

Fig. 9 — Single­layer square claddingNo. 14 made by GMAW­CSC. A — Pat­tern of motion of workpiece relative towelding gun; B — top view; C— trans­verse cross section; D, E, F — optical mi­crographs. Dilution = 6.01%.

A

D E F

C

B

Yu 12-16_Layout 1 11/9/16 8:29 AM Page 457

distortion of the workpiece is visible,which probably can be minimized byusing a stronger frame and heaviermotors for the 3D printer system toallow the use of a metal fixture forclamping down the workpiece tightlybefore welding.

Overall Composition ofCladding To prepare a sample for determin-ing the overall chemical compositionof the cladding, an eight-layer squarecladding was made with the long ser-pentine motion pattern — Fig. 9A. Asmentioned previously, the upper halfof the cladding was cut off by electri-cal discharge machining (EDM), bro-ken up into particles, and dissolved invarious chemicals for chemical analy-

sis. As shown in Table 5, the overallcomposition of the cladding is 53.36± 2.80 wt-% Ni, 42.09 ± 0.93 wt-% W,1.15 ± 0.003 wt-% C, 0.64 ± 0.004 wt-% B, 0.34 ± 0.01 wt-% Fe, and 0.33 ±0.004 wt-% Si. Due to accumulationof errors, the summation of theseconcentrations is 97.91 wt-% insteadof 100 wt-%, the difference being

2.09 wt-%. For simplicity, the compo-sition will be taken as Ni-42.09W-1.15C-0.64B-0.34Fe-0.33Si by wt-%,which implies the balance is 55.45 wt-% Ni (= 53.36 + 2.09). This composi-tion is more reliable than that provid-ed by the supplier of the weldingwire. As shown in Table 1, the carboncontent was not even measured.

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95458-s

Table 3 — Welding Parameters for GMAW­CSC

Sample 46 47 48 49 50 51 Arc Time Initial arc current (A) 40 50 60 65 70 77 Time (ms) 4 4 4 4 4 4 Mid arc current (A) 80 100 120 130 140 154 Time (ms) 4 4 4 4 4 4 End arc current (A) 40 50 60 65 70 77 Short Circuit Start short current (A) 40 40 40 40 40 40 Time Time (ms) 7 7 7 7 7 7 Mid short current (A) 40 40 40 40 40 40 Time (ms) 7 7 7 7 7 7 End arc current (A) 40 40 40 40 40 40 Ramp Rates Rise (A/ms) 250 250 250 250 250 250 Fall (A/ms) 250 250 250 250 250 250 Wire feeding Down wire feed 8 8 8 8 8 8 Rate speed (mpm) Delay before wire down (ms) 0 0 0 0 0 0 UP1 wire feed speed (mpm) 8 8 8 8 8 8 Delay before wire up (ms) 0 0 0 0 0 0 Up2 wire feed speed (mpm) 8 8 8 8 8 8 Arc length (mm) 0.7 0.7 0.7 0.7 0.7 0.7 Penetration delay (ms) 0 0 0 0 0 0 Start Sequence Preflow time (s) 0.5 0.5 0.5 0.5 0.5 0.5 Data Run­in wire feed speed (mpm) 1.0 1.0 1.0 1.0 1.0 1.0 Process starting current (A) 50 50 50 50 50 50 Initial arc length (mm) 2.0 2.0 2.0 2.0 2.0 2.0 Preheat current (A) 60 60 60 60 60 60 Preheat time­start delay (ms) 50 50 50 50 50 50 Stop Sequence Stop arc length (mm) 2.0 2.0 2.0 2.0 2.0 2.0 Data Stop time (ms) 50 50 50 50 50 50 Arc stop current (A) 50 50 50 50 50 50 Postflow time (s) 0.5 0.5 0.5 0.5 0.5 0.5Electrode Extension (mm) 16 16 16 16 16 16 Shielding Gas 75% Ar 75% Ar 75% Ar 75% Ar 75% Ar 75% Ar 25% CO2 25% CO2 25% CO2 25% CO2 25% CO2 25% CO2

Gas Flowing Rate (m3/h) 1.1 1.1 1.1 1.1 1.1 1.1 Travel Speed (mm/s) 15 15 15 15 15 15 Heat Input (J/mm) 56.3 74.8 90.6 99.5 107.1 119.5

Table 4 — Dilutions in Single­Bead Cladding Made by GMAW­CSC and ConventionalGMAW

GMAW­CSC Conventional GMAW

Sample 46 47 48 49 50 51 53 54 55 Dilution 1 0.124 1.25 2.95 1.38 5.76 1.24 1.05 9.81 14.21 (%) 2 0.098 0.49 2.35 1.81 4.50 5.37 4.53 2.35 3 0.247 0.072 1.0 3.68 5.98 10.37 2.23 0.82 9.21 Average 0.16 0.6 2.1 2.29 5.41 5.66 2.60 4.33 11.71

Yu 12-16_Layout 1 11/9/16 8:30 AM Page 458

Optical Micrographs Figure 13 shows the microstruc-ture of single-bead cladding No. 47(74.8 J/mm) and No. 50 107 J/mm).As shown by the transverse opticalmicrographs in Fig. 13A and B, thelarge tungsten carbide particles arethose from the filler metal that werenot melted completely during weld-ing. Dendrites of Ni-rich primary so-lidification phase -Ni are visible inthe matrix between large tungstencarbide particles. The secondary den-drite arm spacing is larger withcladding No. 50 than No. 47. It is wellknown that the higher the heat inputis, the slower the cooling rate duringwelding (Ref. 19). It is also wellknown that the slower the coolingrate is, the more time is available fordendrite arms to coarsen and increasethe secondary dendrite arm spacing

(Refs. 19, 20). So, the higher the heatinput is, the larger the secondarydendrite arm spacing can be expect-ed. Thus, the higher heat input usedfor depositing cladding No. 50 can beexpected to result in larger dendritearm spacing.

Local CompositionMeasurements by High­Resolution EPMA

Figure 14 shows a big tungsten car-bide particle in cladding No. 9 (Fig.11), and the various phases in it iden-tified by EPMA. As mentioned previ-ously, with the high-resolution EPMAused in the present study, the diame-ter of the electron beam was only80–100 nm in diameter and the vol-ume below the sample surface affectedby the beam was 250 nm in diameter.

The original electron image takenduring EPMA has been replaced by aSEM image taken after EPMA. TheSEM image, which is shown in Fig. 14,has a higher contrast, which is neededto show the different phases in theparticles more clearly. WC exists along the edge of the bigparticle as shown by the compositionsat Points 1, 2, and 3. The presence ofWC along the edge is consistent withthe observation of a WC “shell”around big tungsten carbide particlesby Vespa et al. (Ref. 3), who used thesame welding wire used in the presentstudy, that is, PolyTung NiBWC. It isalso consistent with the observationof a “halo” around big tungsten car-bide particles by Choi et al. (Ref. 2),although they used a different weld-ing wire, that is, an Arctec TungcoreFCS cored wire. Figure 14 shows that inside the bigtungsten carbide particle WC existsat Points 4 and 5, and W2C exists atPoints 10, 11, and 12. The presenceof both WC and W2C inside tungstencarbide particles was also reported byVespa et al. (Ref. 3) and Choi et al.(Ref. 2). According to Fig. 14, W3C2also exists inside the big tungstencarbide particle, which was not re-ported by Vespa et al. (Ref. 3) or Choiet al. (Ref. 2). Further identification,e.g., by x-ray diffraction (XRD), is

needed to confirm the presence ofW3C2. It is interesting to note thatHuang et al. (Ref. 21) showed XRDpeaks corresponding to -Ni, WC,W2C, and WC1-x. If x = 1/3, WC1-x be-comes WC2/3 or W3C2. Figure 15 shows the compositionmeasurements in an interdendritic areain cladding No. 9. At Points 1, 2, and 3,the -Ni dendrites are Ni-rich as expect-ed. However, they also contain signifi-cant amounts of W and C. The inter-dendritic eutectic is composite-like,consisting of a small lighter contrast-phase like -Ni and a small darker-con-trast phase. The darker-contrast phaseat Points 4 and 5 appears to be Ni3B,but it also contains a significantamount of C. The lighter-contrastphase at Points 6 and 7 are similar to -Ni in composition. Figure 16 shows the x-ray diffrac-tion pattern obtained by directing thex-ray over an area of 0.1 mm diameteron the transverse cross section ofcladding No. 9 (Fig. 11). The diffrac-

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 459-s

A

C

D E F

B

Fig. 10 — Single­layer square cladding No. 11 made by GMAW­CSC. A — Pattern of motionof workpiece relative to welding gun; B — top view; C— transverse cross section; D, E, F —optical micrographs. Dilution = 4.6%.

Yu 12-16_Layout 1 11/9/16 8:30 AM Page 459

tion angles of various phases areshown at the bottom of the diffrac-tion pattern, including Ni, Ni-0.09W,WC, W2C, WC1-x, and Ni3B. As shown,the presence of Ni-0.09W, WC, W2C,and Ni3B is confirmed. It is interest-ing that Ni-0.09W matches the dif-fraction pattern better than Ni. Asshown previously in Fig. 15, the -Nidendrites contain 7.86 to 9.31 at-%W. The small peak at 37 deg matchesWC1-x, but it also matches Ni3B. Like-wise, the small peak at 61.8 deg

matches WC1-x butit also matchesW2C. So, the pres-ence of WC1-x stillneeds further confirmation.

According to thebinary W-C phasediagram in Fig. 17(Ref. 22), however,-WC1-x does existfrom about 38 to50 at-% C or arange of the C/Wratio of about 0.6to 1.0. This rangedoes cover WC1-xwith x = 1/3, whichis equivalent toWC2/3 or W3C2 .

Figure 18 showsthe electron im-ages of claddingNo. 47 taken dur-ing compositionmeasurements byEPMA. This smallbeam size helpeddetermine thecompositions ofsmall interden-dritic features.These features areaway from the bigtungsten carbideparticles from the

filler metal, suggesting their forma-tion from the liquid pool during so-lidification. As shown by Fig. 18Band the compositions at Points 1 and2 in the table, the angular bright-contrast particle is WC. Other bright-contrast particles are also W-rich asshown by the compositions at Points3 through 5. The small Ni contents ofthese particles might be real or at-tributed to the neighboring Ni-richmatrix. Similar results are shown in Fig. 19

for cladding No. 51. As shown, the an-gular bright-contrast particle is againWC as indicated by the compositionsat Points 8 and 9. However, unlike theW-rich interdendritic particles in Fig.18 for cladding No. 47, the bright-contrast interdendritic phases incladding No. 51 contain about 18 to30 at-% Ni and some Fe and B in addi-tion to more than 40 at-% W and 20at-% C. As shown previously (Fig. 7),cladding No. 51 was made with a heatinput (120 J/mm) higher than thatfor weld No. 47 (70 J/mm). Thus, it islikely that under a higher heat inputbig tungsten carbide particles dis-solved to a greater extent duringwelding and formed these interden-dritic particles during solidification.

Thermodynamic Analysis

The thermal spray and hardfacingindustries have included Cr in manycommercially available Ni-based alloysas it aids in hardenability and increas-es corrosion resistance. However, ithas been noted (Ref. 1) that brittle Cr-containing phases form in claddingcoatings when WC/W2C particles par-tially dissolve into the matrix duringPTA, laser, or GMAW processing.These brittle phases will be detrimen-tal to performance when impactevents occur during service. The ef-fects of Cr additions to the Ni alloyneed to be understood and consideredfor cladding coatings. To help understand the effect ofthe Cr addition on Ni-WC cladding,thermodynamic software and data-base were used to calculate the solidi-fication path of the cladding based onits composition. Obviously, the nomi-nal composition of the filler metalprovided in Table 1 by the suppliercannot be used because it does noteven show the C content. This is whyan eight-layer square cladding was

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95460-s

Fig. 11 — Four­layer square cladding No. 13 made by GMAW­CSC. A — Pattern of motion of workpiece relative to weldinggun; B — top view; C — transverse cross section; D, E, F — opti­cal micrographs. Dilution = 3.61%.

Table 5 — Chemical Composition (wt­%) of Eight­Layer Square Cladding

W Si B Ni C Fe

Chemical Composition, 41.82 0.337 0.640 54.01 1.149 0.349 wt­% 41.83 0.337 0.645 54.93 1.145 0.334 42.75 0.332 51.14 0.334 41.65 42.40 Average 42.09 ± 0.93 0.33 ± 0.004 0.64 ± 0.004 53.36 ± 2.80 1.15 ± 0.003 0.34 ± 0.01

AB

C

D E F

Yu 12-16_Layout 1 11/9/16 8:31 AM Page 460

prepared for detailed chemical analy-sis. As mentioned previously, theoverall composition of the claddingwas Ni-42.09W-1.15C-0.64B-0.34Fe-0.33Si by wt-%. Since the weld pool and hence theresultant cladding contained the un-

melted portion of tungsten carbideparticles as well as liquid metal, thecomposition of the liquid metal need-ed to be found in order to calculate itssolidification path. To do this, two dif-ferent methods were tried in the pres-ent study. The first method was to de-

termine the local average compositionof the Ni-rich matrix away from un-melted tungsten carbide particles, thatis, to determine the local average com-position in an area significantly largerthan dendrites and away from bigtungsten carbide particles. The newSXFiveFE EPMA machine allowed a50-m-diameter beam size to be usedto measure the average composition ata location. However, it was found thatthe measured composition of the Ni-rich matrix varied significantly fromlocation to location in the claddingand an overall composition of the ma-trix as a whole could not be deter-mined accurately. The second method, on the otherhand, was based on the overall com-position of the cladding measured bychemical analysis. As mentioned pre-viously, an eight-layer square claddingwas cut off by EDM from 1 mm abovethe steel substrate and broken up anddissolved in chemical solutions forwet chemical analyses. Thus, the dis-solved cladding included not only theNi-rich matrix, but also the unmeltedportions of the big tungsten carbideparticles, which need to be excludedin order to have an accurate composi-tion of the liquid metal in the weldpool to calculate its solidificationpath. According to image analysis byImage J (Ref. 23), tungsten carbideparticles larger than 15 m in equiva-lent diameter occupied 13.53% of thecross-sectional area of the cladding.Examination of representative exam-ples of such tungsten carbide parti-cles revealed that the “halo” aroundthe particle occupied about 25.7% ofthe area of the particle. Thus, the un-melted portions of such tungsten car-bide particles represented about 10%of the cladding, that is, 13.53% ¥(100% – 25.7%). As an approximation, assume thetungsten carbide particles are WC.Since the measured W content of thecladding is 42.09 wt-% (Table 5), theW content in the liquid portion of theweld pool should be 42.09 wt-% ¥(100% – 10%) = 37.88 wt-%. Similarly,since the measured C content of thecladding is 1.15 wt-% (Table 5), the Ccontent in the liquid portion of theweld pool should be 1.15 wt-% ¥(100% – 10%) = 1.035 wt-%. Thus,based on 37.88 wt-% W and 1.035 wt-% C and the measured contents of all

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 461-s

Fig. 13 — Microstructure of cladding. A, B — Optical micrographs; C, D — SEM imagesshowing W­rich interdendritic features (light­contrast). Secondary dendrite arm spacing ishigher with higher heat input.

Fig. 12 — Expansion at the lower end of the tubular welding wire after GMAW­CSC ofcladding No. 9 (Fig. 11). Such expansion is likely to contribute to spatter, which usually iseffectively eliminated in CSC­GMAW with a solid welding wire.

A B

B

C D

Yu 12-16_Layout 1 11/9/16 8:32 AM Page 461

other elements in Table 5, the totalcontents of all the elements in theweld-pool liquid is now 95.675 wt-%(= 37.88 wt-% W + 0.33 wt-% Si + 0.64wt-% B + 55.45 wt-% Ni + 1.035 wt-%C + 0.34 wt-% Fe). Since the totalshould be 100%, each alloying contentshould be multiplied by a factor of1.045 (= 100 / 95.675). Thus, the com-position of the liquid in the weld poolis as follows: 39.58 wt-% W, 0.34 wt-%Si, 0.67 wt-% B, 57.95 wt-% Ni, 1.08wt-% C, and 0.36 wt-% Fe or just Ni-39.58W-1.08C-0.67B-0.36Fe-0.34Si by wt-%. Figure 20A shows the solidificationpath of the liquid metal in the weldpool based on the composition of Ni-39.58W-1.08C-0.67B-0.36Fe-0.34Si.The solidification path is the curve oftemperature T vs. fraction of total sol-id fS during solidification. The curvewas calculated using thermodynamicsoftware Pandat (Ref. 24) and Ni-alloy

database PanNickel(Ref. 25) of Com-puTherm, LLC.The arrowheadsindicate the pointsat which solidphases start toform from the liq-uid during cooling.As shown, before-Ni starts to format about 1380°C,WC forms from the liquid at about1650°C. This WC formation probablyoccurs on the unmelted portions ofthe tungsten carbide particles, that is,by epitaxial growth as “halos” aroundthe unmelted portions of the particles.Epitaxial growth is much easier sinceWC does not have to nucleate from themelt. Figure 20A shows -Ni starts toform just below 1400°C. So, in Fig.20B the big hump in the WC curve just

below 1400°C indicates WC formingwhile -Ni dendrites grow, and thisWC is likely to exist in the interden-dritic areas. At about 1300°C M2Bboride begins to form, where M con-tains Ni, Fe and W. At about 1060°C,Ni3B boride begins to form in the formof eutectic Ni-Ni3B. Figure 20B showsthe fractions of solid phases during so-lidification. The Ni-rich phase -Ni,which has a much higher fraction thanall other solid phases, is not includedin order to show other solid phases

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95462-s

Fig. 14 — Composition measurements of a big tungsten carbideparticle in cladding No. 9 (Fig. 11) showing W3C2 as well as thepreviously reported WC and W2C inside partially melted tungstencarbide particles. Points of measurements are superimposed onan SEM image for better contrast between phases.

Fig. 15 — Composition measurements in interdendritic area incladding No. 9 (Fig. 11).

A

B

Yu 12-16_Layout 1 11/9/16 8:32 AM Page 462

more clearly. As shown, the fraction ofWC is about 0.125. The effect of substituting Ni withCr on the fraction of WC is shown asfollows. Figure 20C and D shows thecalculated results for 39.58 wt-% W,0.34 wt Si, 0.67 wt-% B, 52.95 wt-%Ni, 1.08 wt-% C, 0.36 wt-% Fe, and 5wt-% Cr or just Ni-39.58W-1.08C-0.67B-0.36Fe-0.34Si-5Cr. In otherwords, 5 wt-% Ni is substituted with 5wt-% Cr and the contents of other al-loying elements such as C, B, Fe, and Siremain unchanged. As shown, WC stillstarts to form from the weld-pool liq-uid at about 1650°C. However, the for-mation of M6C carbide soon follows atabout 1500°C. The amount of WC af-ter solidification is only about 7.5 wt-%, instead of 12.5 wt-% as in the casewithout Cr. On the other hand, theamount of M6C is as high as 15 wt-%. With 10 wt-% of Ni substituted by10 wt-% Cr, the composition of theweld-pool liquid becomes 39.58 wt-%W, 0.34 wt Si, 0.67 wt% B, 47.95 wt-%Ni, 1.08 wt-% C, 0.36 wt-% Fe, and 10wt-% Cr or just Ni-39.58W-1.08C-0.67B-0.36Fe-0.34Si-10Cr. As shownin Fig. 20E and F, WC no longer formsduring solidification. Instead, M2C car-bide forms at about 1630°C and M6Ccarbide about 1570°C. The resultantcontent of M6C carbide is as high as 30wt%. These results suggest that Crtends to form undesirable Cr-contain-ing carbide M6C at the expense of WC. Liyanage et al. (Ref. 26) used threeNi-rich matrix alloys containing 0, 8.4,and 13.8 wt-% Cr in PTAW of Ni-WCcladding. It was shown that the WCvolume fraction in the cladding waslower with than without Cr. The solidi-fication paths of the three matrix al-loys were calculated based on theScheil solidification model. No frac-tions of solid phases were shown. Itwas pointed out that the solidificationpaths showed that the stability of WCis reduced when Cr is added, which re-duces WC and promotes Cr-containingcarbides. Landwehr (Ref. 27) also showed Craddition increased WC dissolution.Ni-WC claddings were made by usinga defocused laser-beam and addingNi-WC powder to the weld pool. Thecladding made with a powder contain-ing 14 wt-% Cr showed significantlymore WC dissolution than one madewith a Cr-free powder. Although the

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 463-s

Fig. 16 — X­ray diffraction pattern of cladding No. 9 (Fig. 11) identifying the presence ofphases, including Ni, Ni­0.09W, WC, W2C, WC1­x and Ni3B.

Fig. 17 — Tungsten­carbon phase diagram (Ref. 22).

Yu 12-16_Layout 1 11/9/16 8:33 AM Page 463

Ni-WC differed from the filler metalused in the present study, this effect of Cr on WC dissolution is consistentwith that shown in Fig. 20 based onthermodynamic analysis.

Conclusions

1) The operation window forpreparing a smooth cladding of Ni-WCwithout severe spatter is significantlywider with controlled-short-circuitingGMAW, i.e., GMAW-CSC, than withconventional GMAW. 2) The undesirable melting of thesteel substrate, that is, the dilution ofthe hard Ni-WC cladding by the melt-ed steel, can be better controlled withGMAW-CSC than with conventionalGMAW. Increasing the heat inputtends to increase dilution much moresignificantly with conventionalGMAW than with GMAW-CSC. 3) The weaving pattern of the weld-ing gun can significantly affect theuniformity of the thickness of thecladding and its dilution by the basemetal. The optimum pattern can befound by using a low-cost 3D printerto move the workpiece under a sta-tionary gun, which is especially con-venient when a much more expensivewelding robot or 3-axis CNC millingmachine is unavailable. 4) High-resolution electron probemicroanalysis (EPMA), with a verysmall electron beam of 80–100 nm indiameter, can allow the compositionsof various phases in partially meltedtungsten carbide particles and smallW-rich interdendritic features to bedetermined. 5) As the heat input increases,which increases the partial melting oftungsten carbide particles duringwelding, the W-rich interdendritic fea-tures can change from essentially Ni-free to containing 18–30 at-% Ni. 6) Thermodynamics analysisdemonstrates that increasing the Crcontent can cause Cr-containing car-bide to form at the expense of the WCneeded to provide the wear resistance.

The authors would like to thankDr. John H. Fournelle of the Depart-ment of Geoscience, UW-Madison for

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95464-s

Fig. 18 — Compositions of interdendritic particles in single­bead cladding No. 47 (Fig.7A) made with a lower heat input. A — Overview; B — enlarged.

Fig. 19 — Compositions of interdendritic particles in single­bead cladding No. 51 (Fig.7C) made with a higher heat input. A — Overview; B, C, D — enlarged.

A

A

C D

B

B

Acknowledgments

Yu 12-16_Layout 1 11/9/16 8:34 AM Page 464

composition measurements by high-resolution EPMA and CompuTherm,LLC in Madison, Wis,. for kindly provid-ing Pandat 2013 and PanNickel 2013 forcalculating the solidification paths.They would also like to thank Polymet,Cincinnati, Ohio, for donating the Poly-Tung NiBWC wire. The authors wouldlike to thank Bruce Albrecht, ToddHolverson, Rick Hutchison, and JoeFink of Miller Electric ManufacturingCo. and ITW Global Welding Technolo-gy Center, both located in Appleton,Wis., for donating the CSC process con-troller and drive assembly, Invision 456power source, XR-M wire feeder, andwelding gun used in the present study.This work was supported initially by theIndustry/University Collaborative Re-search Center (I/UCRC) for Integrated

Materials Joining Science for EnergyApplications, and subsequently by theNational Science Foundation underGrant No. DMR 1500367.

1. Mendez, P. F., Barnes, N., Bell, K., Bor-le, S. D., Gajapathi, S. S., Guest, S. D., Izadi,H., Gol, A. K., and Wood, G. 2014. Weldingprocesses for wear resistant overlays. Journalof Manufacturing Processes 16(1): 4–25.

2. Choi, L., Wolfe, T., Yarmuch, M., andGerlich, A. September 2011. Effect of weld-ing parameters on tungsten carbide-metalmatrix composites produced by GMAW.Proc. CWA Conf., Banff, AB, Canada.

3. Vespa, P., Pinard, P. T., Gauvin, R., and

Brochu, M. 2012. Analysis ofWC/Ni-based coatings depositedby controlled short-circuit MIGwelding. Journal of Materials En-gineering and Performance 21(6):865–876.

4. ISO/TR 13393, 2009, Weld-ing consumables — hardfacingclassification — microstructures,Switzerland.

5. Data Sheet, PolyTung NiB-WC, WC-W-007. April 2010.Cincinnati, Ohio: Polymet Corp.

6. CSC-MIG Weld Process Con-troller, Jetline Engineering,Irvine, Calif., pdf.directindustry.com/pdf/jet-line-engineering/csc-mig-weld-process-controller/27577-54392.html.

7. Gilles, K. 2006. An anti-spat-ter matter. Welding Design and Fab-rication,weldingdesign.com/process-es/anti-spatter-matter.

8. Wagner, D. C., Yang, Y. K.,and Kou, S. 2013. Spatter andporosity in gas-metal arc weldingof magnesium alloys: mecha-nisms and elimination. WeldingJournal 92: 347-s.

9. Chai, X., Yang, Y. K., and Kou,S. 2015. Oxide films, high crownsand fingers in gas-metal arc weldsof Mg alloys: Mechanisms of for-mation and mitigation. WeldingJournal 94: 16-s to 33-s.

10. Anzalone, G. C., Zhang, C.,Wijnen, B., Sanders, P. G., andPearce, J. M. 2013. A low-costopen-source metal 3-D printer.IEEE Access 1: 803–810.

11. GB/T 223.60. 1997. Meth-ods for chemical analysis of iron,steel and alloy — The perchloricacid dehydration gravimetricmethod for determination of sili-con content, Beijing, China:Standardization Administration

of China.12. GB/T 7731.1. 1987. Methods for

chemical analysis of ferrotungsten — Thecinchonine gravimetric method for the de-termination of tungsten content, Beijing,China: Standardization Administration ofChina.

13. GB/T 24583.4. 2009. Vanadium-Ni-trogen alloy — Determination carbon ofcontent — The infrared absorptionmethod, Beijing, China: StandardizationAdministration of China. 14. NACIS/C 048. 2013. Iron, steel andalloy — Determination of nickel content-Dimethylglyoxime precipitation-EDTAtitration method, Beijing, China: NationalAnalysis Center for Iron and Steel. 15. NACIS/C H 070. 2013. Determina-tion of iron, cobalt, nickel and copper con-tent-ICP-AES, Beijing, China: National

WELDING RESEARCH

DECEMBER 2016 / WELDING JOURNAL 465-s

Fig. 20 — Solidification paths and resultant fractions of phases. A, B — Ni­39.58W­1.08C­0.67B­0.36Fe­0.34Si; C, D — Ni­39.58W­1.08C­0.67B­0.36Fe­0.34Si­5Cr, showing loss of WC caused by substituting 5wt­% Ni with 5 wt­% Cr; E, F—Ni­39.58W­1.08C­0.67B­0.36Fe­0.34Si­10Cr, showing complete loss ofWC caused by substituting 10 wt­% Ni with 10 wt­% Cr. Arrowheads indicate the points at which newphases start to form from the liquid during cooling. Fraction of -Ni is not included in order to showother solid phases more clearly.

References

A

B

C

F

E

D

Yu 12-16_Layout 1 11/9/16 8:34 AM Page 465

WELDING RESEARCH

WELDING JOURNAL / DECEMBER 2016, VOL. 95466-s

Analysis Center for Iron and Steel. 16. NACIS/C H 121. 2013. Iron, steeland alloy — Determination of Boron con-tent- ICP-AES, Beijing, China: NationalAnalysis Center for Iron and Steel. 17. CAMECA SXFiveFE: Field EmissionElectron Probe Microanalyser.cameca.com/instruments-for-research/sx-fivefe.aspx. 18. Fournelle, J. H. May 2016. Privatecommunications, Department of Geo-science, University of Wisconsin, Madison,Wis. 19. Kou, S. 2003. Welding Metallurgy, 2ndedition, John Wiley and Sons, Hoboken, N.J.

20. Flemings, M. C. 1974. SolidificationProcessing, McGraw-Hill, New York, N.Y. 21. Huang, S. W., Samandi, M., andBrandt, M. 2004. Abrasive wear perform-ance and microstructure of laser cladWC/Ni layers. Wear 256: 1095–1105. 22. Kurlov, A. S., and Gusev, A. L. 2013.Tungsten Carbides. Springer, Switzerland. 23. J Image, Image processing andAnalysis in Java, National Institute ofHealth. imagej.nih.gov/ij/. 24. Pandat. Phase Diagram Calculationsoftware package for Multicomponent Sys-tems. 2014. CompuTherm, LLC, Madison,Wis.

25. PanNickel – Thermodynamic data-base for Commercial Aluminum Alloys.2014. CompuTherm, LLC, Madison, Wis. 26. Liyanage, T., Fisher, G., and Gerlich,A. P. 2012. Microstructure and abrasivewear performance of PTAW deposited Ni-WC overlays using different Ni-alloychemistries. Wear 274–275: 345-354. 27. Landwehr, D. 2015. Investigation ofcharacterization techniques for phasesformed during laser cladding of WC/Ni-al-loy wear resistant coating, senior project,University of Wisconsin, Madison, Wis.

www.aws.org

American Welding Society®

BRING BRAND AWARENESSTO YOUR COMPANYBy placing your product video on the AWS website.

Contact AWS for more information at 800-443-9353Sandra Jorgensen at Ext. 254, email: sjorgensen@aws.org / Annette Delagrange at Ext. 332 email: adelagrange@aws.org

Yu 12-16_Layout 1 11/9/16 8:35 AM Page 466