FluxAssisted Gas Tungsten Arc and Laser Welding of ...files.aws.org/wj/supplement/WJ_2014_10_s379.pdf · The use of titanium alloys has been expanding in the aerospace, power-generation,

Introduction The use of titanium alloys has beenexpanding in the aerospace, power-generation, medical, chemical plants,and marine applications due to their

superior mechanical and corrosionproperties. The main welding processfor titanium alloys is gas tungsten arcwelding (GTAW), also known as tung-sten inert gas (TIG) welding. The ma-jor shortcomings of the GTAW

process, even though it produces high-quality welds, are the slow productionrate and shallow weld penetration.Single-pass welds with GTAW can havearound 2–3 mm penetration creating abottleneck in the fabrication process(Refs. 1, 2). In the 1960s, researchersat the E. O. Paton Institute inventedthe A-TIG process for welding titani-um alloys by using oxygen-free activat-ed fluxes that significantly increasedweld penetration (Refs. 3, 4). The ob-served deep penetration was attrib-uted to arc constriction and increasein current density. Fluorides of alkaliand alkali/earth metals were found tobe effective in increasing weld pene-tration in titanium alloys with obser-vation of arc constriction (Refs. 5, 6). Welding arc plasma is complex anddifficult to investigate. One of the bet-ter techniques that can determine theconcentrations of arc constituents andestimate arc temperature is arc emis-sion spectroscopy. The concept of arcspectroscopy is to measure the intensi-ty (emission coefficient) of the light(at specific wavelengths) emitted bythe arc plasma. An atom, an ion, orelectrons at high temperature will be-come excited, shifting from one ener-gy state to another of higher energy,e.g., from L shell to M shell in the caseof the electrons. Following the excita-tion process, the metastable particlesdecay, dropping from the higher ener-gy state to a lower energy state. Underset conditions, the decay of an elec-tron from a shell at a higher energystate (E1) to a lower energy shell (E2),as shown in Fig. 1, will be accompa-

WELDING RESEARCH

FluxAssisted Gas Tungsten Arc and Laser Weldingof Titanium with CryoliteContaining Fluxes: ArcSpectroscopy and Corrosion Resistance Studies

Arc temperature and arc force were found to contribute to deeper weld penetration influxassisted gas tungsten arc welding

BY T. ALSABTI, A. ALSHAWAF, AND S. LIU

ABSTRACT This research compares autogenous welding, welding with a solid weldingwire, and welding with tubular welding wires on CPtitanium using gas tungsten arc (GTA) and laser beam (LB) welding. Commercially pure titanium (CPTi Grade 2) solid wire and two newly designed fluxcored tubular wires containing cryolite (Na3AlF6) and MgF2 were used. The comparison covers theaspects of weld penetration, arc constriction, arc force, arc emission spectrallines, arc temperature, and constituent concentrations in the arc. Laserbeam welding was performed for comparison since it does not involve arcplasma as in the GTAW process. Also, this research investigated two mechanisms for weld penetration, arc constriction, and Marangoni convection,which were suggested in previous literature as the driving mechanisms forthe deeper weld penetration observed with the use of fluxes. Fluxassistedgas tungsten arc welding is also known as activated tungsten inert gas (ATIG) welding. This work concluded that mainly arc temperature and arc forcecontributed to the deeper weld penetration. Surfacetensiondriven flow(Marangoni convection) was found not to contribute to the weld penetrationincrease observed with increasing cryolite content. Immersion corrosion testresults on the ATIG weld specimens in separate 3.5% sodium chloride (NaCl)and 0.1% phosphoric acid (H3PO4) solutions for 800 h, particularly in termsof mass loss, compared favorably with the autogenous welds.

KEYWORDS • Gas Tungsten Arc Welding (GTAW) • ATIG • CryoliteContaining Fluxes • Marangoni Convection • Laser Beam Welding • Arc Spectrometer • Immersion Corrosion Test

T. ALSABTI is with Consulting Services Department, Saudi Aramco, Kingdom of Saudi Arabia. A. ALSHAWAF is with Saudi Basic Industrial Corp. (SABIC), Al Jubail,Kingdom of Saudi Arabia. S. LIU is with Colorado School of Mines, Center for Welding, Joining and Coatings Research, Golden, Colo.

OCTOBER 2014 / WELDING JOURNAL 379-s

AlSabti Supplement layout_Layout 1 9/11/14 5:00 PM Page 379

nied by the emission of a monochro-matic light according to the followingequations (Refs. 7, 8):

E1 – E2 = E = hv (1)

c = v (2)

where E is the change in energy, h isPlanck’s constant, v is the frequency ofthe emitted light, is the wavelength,and c is the speed of light. The emittedlight will have a specific wavelength asdetermined by the energy change as-sociated with the electron during its

decay. Each decayevent in a givenatom has a charac-teristic wavelengthof light associatedwith it. It is theunique emittedspectra by certaindecay events that allow for detectionof the elements within an arc. It has been shown by Bang,Chirieleison, and Liu (Ref. 1) that in-creasing amounts of a simple fluoride,MgF2, both in paste and flux coredwire form, increased weld bead pene-

tration in titanium by nearly 200%over fluxes without the fluoride. Intheir work, MgF2 was used as an addi-tion into CaF2, BaCl2, AlF3, and NaF.The mechanism in this case was attrib-uted to arc constriction. They ex-plained that as the amount of flux va-por entering the weld zone increases,the thermal conductivity of the arc in-creases and the arc constricts. This ex-planation was supported by researchconducted by Zamkov, Prilutsky, andGurevich (Ref. 5). In addition, the elec-trical conductivity of the arc will de-crease with an increase in flux vapordue to the trapping of electrons in theouter peripheral region of the arc. Inthis region, temperatures are generallynot high enough to fully dissociateatoms into ions. The flux vapor existsas molecules that can attract free elec-trons to form negatively charged parti-cles. The degree of arc constriction willthen depend on how effectively the va-porized molecules can trap these elec-trons. Bang, Chirieleison, and Liu (Ref.1) also noticed the presence of titani-um spectra lines when welding withMgF2-containing fluxes. Without theaddition of MgF2, the spectra lines of

WELDING RESEARCH

WELDING JOURNAL / OCTOBER 2014, VOL. 93380-s

Fig. 1 — Illustration of excitation and decay of electron in energystates in electronic structure in atoms or ions (Ref. 7).

Fig. 2 — Aswelded GTA (autogenous and solid wire) and ATIG welds (Wires 15 and 45).

Fig. 3 — Arc constriction observed in the GTA and ATIG welds.

Fig. 4 — Crosssection macrograph of the GTA and ATIGwelding.

Table 1 —GTA Welding Parameters

GTA Welding Parameters

Machine Miller Dynasty 350 (GTA)Current 180 ATravel Speed 3 mm/sShielding Gas Type and Flow Argon (UHP G5.0), 0.275 L/sElectrode Type and Size 2% Thoriated, 3.2 mmElectrode Tip Preparation 45-deg included angleArc Length 5 mm

Table 2 — Laser Welding Parameters

Laser Welding Parameters

Type/Mode Yb-Fiber Laser continuous wave mode/ Conduction modePower/Focus 1100 W/5 mm (under plate surface)Travel Speed 3 mm/s

Shielding Gas Type and Flow Argon (UHP G4.5), 0.157 L/sWire Feed Rate 13 mm/s (Solid CP Ti-G2 wire)

4 mm/s (Wires 15 & 45)


titanium were absent. Hillier, Liu, andRoepke (Ref. 2) investigated the ef-fects of cryolite, which contains moreF- ions in its formula than MgF2.Compared to MgF2 flux pastes, weldsusing cryolite pastes proved to bemore effective, netting up to twotimes deeper penetration and depth-to-width ratio. Thus, cryolite can beconsidered a strong ingredient in anactivating flux. This work attempted to character-ize the effects of cryolite and distin-guish the effects of arc constriction,arc force, and Marangoni convectionon Ti weld penetration. Parallel to thewelding study, a corrosion study wasalso carried out to determine the cor-rosion resistance of these A-TIG weldsby comparing them to the autogenouswelds and base metal.

Experimental Procedure Welding experiments were con-ducted to examine the effects of se-lected fluxes on penetration in com-mercially-pure titanium (CP-Ti).Commercially-pure titanium (Grade2) solid wire and two tubular coredwires containing cryolite in differentproportions [15% (Wire 15) and 45%(Wire 45)] in a base flux that con-tained 15% MgF2 were manufacturedand tested. CaF2, BaCl2, AlF3, andNaF were the remaining ingredientsin the fluxes. Autogenous GTAW wasdone as reference (Table 1). Compari-son between these two sets of weldswould clarify the effects of the acti-vating flux on weld penetration.Aside from A-TIG, laser beam welding(LBW) was also done (Table 2). SinceLBW does not involve any significantarc plasma, it was hoped comparison

of the two process-es would elucidatethe effects of arcplasma on weldpenetration. AnOcean Optics USB4000 spectrometerwas used to recordthe arc emissionwavelengths andrelative intensities.The arcs were alsophotographed forsubsequent analy-sis. The detectorrange was between200 and 1100 nm.The collimatinglens size was 5 mm.Other experimen-tal data are listedin Table 3.

Results and Discussion

Results of GTAW Experiments

The as-welded GTA and A-TIGwelds are shown in Fig. 2. The welds ingeneral showed good bead morpholo-gy. Some evidence of oxidation and ni-

trogen pickup can also be seen indicat-ing that impurities such as oxygen andnitrogen may be present and negative-ly affect the mechanical and corrosionproperties of the A-TIG welds. The im-mersion corrosion test results are dis-cussed at the end of this section. The arc plasmas for the autogenous,Wire 15, and Wire 45 welds were pho-

WELDING RESEARCH


Fig. 6 — Arc force vs. weld penetration.Fig. 5 — Weld penetration comparison for different fluoride fluxsystems.

Table 3 — Experimental Data

Plate Thickness 5.1 mmWire Diameter 1.57 mmWire Feed Rate 13 mm/s (Solid CP Ti-G2

wire)4 mm/s (Wires 15 & 45)

Position of Feed Introduced into the frontof arc

Table 4 — Arc Constriction Data for the Experimental Welds

Weld Case Arc Constriction %

Area Method Line Method

Autogenous 0 0Wire 15 21.0 7.5Wire 45 5.5 0

Fig. 7 — Aswelded condition with the laser welding process.

B

A


tographed for comparison, as shown inFig. 3. Arc constriction was measured bytwo methods. The first was measuringthe area of the cross section of the arc,as marked in Fig. 3. The second methodmeasured the length of the horizontalline, i.e., the arc width, in the middle ofthe arc length. Wire 15 showed greaterarc constriction than Wire 45 in bothmethods, as shown in Table 4. Wire 15exhibited 21% arc constriction as com-pared to 5.5% for Wire 45 in the areameasurement method. Using the linemeasurement method, Wire 15 showed7.5% arc constriction whereas no arcconstriction was observed for Wire 45. The welds were cross sectioned,polished, and etched to measure theweld penetration, as shown in Fig. 4.Kroll’s reagent, containing 6 vol-%HNO3, 2 vol-% HF, and 92 vol-% dis-tilled water, was used for etching. Theweld penetration for each case was

measured andrecorded in Table 5.Wire 45 was foundto have producedthe highest weld penetration of 3.5mm followed by Wire 15 of 2.9 mm.The autogenous weld penetrationmeasured only 2.2 mm. So, as thecontent of the complex fluoride (cry-olite) increased, the weld penetrationalso increased, which confirmed re-search work by Hillier et al. (Ref. 2). Figure 5 shows a summary of weldpenetration of this research work usingA-TIG wires (15% and 45% cryolite +15% MgF2), the A-TIG wires (15, 30, 45,and 60% cryolite) reported by Hillier etal. (Ref. 2), and the A-TIG wires (15, 30,and 60% MgF2) from the research workof Bang, Chirieleison, and Liu (Ref. 1).The weld penetration of Wire 15 (com-bined 15% cryolite and 15% MgF2) washigher than the 15% MgF2 and 15%cryolite, respectively. On the otherhand, the weld penetration of 30% cryo-lite was higher by around 1 mm thanthe 30% MgF2 weld. Also, the weld pen-etration of 45% cryolite is much higherthan Wire 45 (45% cryolite + 15%MgF2) and 60% MgF2, respectively.Thus, increases in Na3AlF6 played adefinite role in penetration. Overall,mixed Na3AlF6 plus MgF2 fluxes in-creased penetration over the MgF2 ad-dition alone, but did not promote deep-er penetration than single cryolite ormagnesium fluoride additions at higherconcentrations. Cryolite promotedgreater penetration than the other fluxingredients when added at 30 and 45%.

Effect of Arc Force

The arc force mechanism was inves-tigated by measuring the R2/R1 ratiosfrom the photographed arcs of the au-togenous, Wire 15, and Wire 45 weldsfrom this research work as well as thearcs from the welds by Hillier et al.The R2/R1 ratios were calculated byusing the following arc force equation(the equation Converti) (Refs. 8–10):

where I is the welding current, 0 isthe permeability of free space, R1 isthe radius of the arc where it contactsthe welding electrode, R2 is the radiusof the arc where it contacts the baseplate. The arc force was plotted vs.the weld penetration of each case inFig. 6. Despite the scatter in the data,a clear relationship can be seen be-tween arc force and weld penetration.Increasing arc force increased weldpenetration. This finding is contraryto the results reported by Burleighand Eagar (Ref. 11) who did not findany correlation between the arc forcesand weld penetration of several fluxsystems (Refs. 12–15).

Results of Laser WeldingProcess Experiments

The welds made by the laser weldingprocess in conduction mode are shownin Fig. 7. The laser welding process waschosen to eliminate the effect of arc

FI

lnR

R81 2 (3)0

2 2

1

= μπ

+⎛⎝⎜

⎞⎠⎟

WELDING RESEARCH


Fig. 8 — Crosssection macrograph of the laser welding process.

Fig. 9 — Variation of the weld depthtoautogenous welddepth ratio as a function of cryolite addition in the flux.Table 5 — Weld Penetration Data for the

Experimental Welds

Weld Case Width (mm) Penetration Depth, mm

Autogenous 5.5 2.2Ti-CP G2 7.0 0.6Wire 15 6.0 2.9Wire 45 7.1 3.5

Table 6 — Weld Penetration for Laser WeldingProcess

Weld Case Width (mm) Penetration Depth (mm)

Autogenous 7.5 3.9Ti-CP G2 5.4 3.0Wire 15 4.8 2.3Wire 45 5.5 2.7


constriction in an attempt to clarifywhether Marangoni convection is pres-ent in these welds. Increasing weld pen-etration in these experiments would in-dicate a contribution of surface-tension-driven flow. The consumable in theform of a wire was laid on top of theplate prior to laser irradiation. The laserbeam was then rastered over the wireand the plate. Total melting of the wireand wetting of the plate was observed.The welds were sectioned, polished, andetched, and the cross sections areshown in Fig. 8. Table 6 lists the weld penetration ofeach case. It was observed that the weldmade using Wire 45 had slightly deeperpenetration (2.7 mm) than the Wire 15weld (2.3 mm). However, weld penetra-tions that resulted from both of the A-TIG wires were lower than those of thesolid Ti-CP G2 wire (3.0 mm) and theautogenous laser weld (3.9 mm). Thereason for having lower weld penetra-tion for all three wires than for the au-togenous laser weld is likely because ofthe placement of the wire on top of thetest plate. By positioning the wire ontop of the test plate, most of the laser

power was consumed in the melting ofthe wire and little was left for the melt-ing of the base plate to form the weldpool, thus the shallower welds. It alsoappeared that the flux cored wires ab-sorbed more energy from the laserbeam than the solid wire, which wouldexplain why the A-TIG welds have lowerweld penetration than the solid Ti-CPG2 wire. Figure 9 shows the calculated welddepths normalized by the autogenousweld depth for both A-TIG and laserwelding. The normalized depths forthe laser welding remained practicallyunchanged despite increasing cryolitecontent. Contrarily, the A-TIG weldsshowed normalized weld depths in-crease with increasing cryolite addi-tion. Thus, it can be concluded thatMarangoni convection did not haveany significant contribution as a driv-ing mechanism in increasing weld pen-etration. This finding is supported bythe findings of Howes and Lucas (Ref.16). Using A-TIG pastes for stainlesssteel by laser and electron beam weld-ing, they concluded that A-TIG fluxesare effective only for welds done by arc

and plasma welding processes.

Temperature and ConcentrationCalculation Results

The spectral lines data collected byarc spectrometer from the autogenousGTA weld, and the A-TIG welds withWires 15 and 45 were processed andmapped in Figs. 10–14. Using a Math-ematica program, the spectral lineemission data collected from each weldcase were processed and used to deter-mine the temperature of the arc andthe concentrations of arc constituentsby the off-axis peak method (Fowler-Milne Method). The asymmetrical be-havior in the spectral peaks in thesefigures is most likely due to the singlescanning of the arc position. The emis-sion data were collected from individ-ual slices of the arc along the line ofsight of the spectrometer. Note thatcontinual scanning of the entire crosssection of the arc would have allowedfor the Abel Inversion calculation andcaptured the entire arc emission sig-nals to result in a more symmetricaldistribution in the spectral peaks.

WELDING RESEARCH


Fig. 11 — Major spectra peaks across autogenous weld.

Fig. 13 — Major spectra peaks across Wire 15 weld.Fig. 12 — Emission spectral line peaks from Wire 15 weld.

Fig. 10 — Emission spectral line peaks from autogenous weld.


However, single scans like the ones de-scribed in this work are still useful togain insight to the arc during titaniumwelding. Neutral argon (Ar I) spectrallines were dominant for the autoge-nous welding arc, as indicated in Fig.8. In addition, very low intensity spec-tral lines of ionized argon (Ar II) andtitanium vapor (Ti I) are visible inFigs. 10 and 11. Despite the molten ti-tanium weld pool, little titanium vaporwas absorbed above the weld. Figure11 plots the emission coefficients forspecific wavelengths for Ar I from oneedge of the arc to the other. In com-parison, Wire 15 has more spectrallines for other constituents such asfirst ionized titanium (Ti II), neutralvanadium (V I), neutral sodium (Na I),and neutral magnesium (Mg I), asshown in Figs. 12 and 13. Spectral lines of other constituentssuch as neutral aluminum (Al I) mightbe hidden behind the lines of the oth-er elements mentioned earlier. This

limitation is partly due to the low res-olution of the specific arc spectrome-ter used. The collimating lens size of5.0 mm was much larger than thecondition of “not more than 0.25mm” that Hiraoka (Ref. 17) had sug-gested in order to limit the error ofline intensities within 3%. The spec-tral lines of fluorides were difficult toobserve because of their higher ener-gy levels and that the transition prob-abilities are lower than those ofmetallic constituents such as magne-sium or sodium. As for Wire 45, moreconstituents are found in the arc aswell as higher intensities as shown inFigs. 14 and 15. First ionized titani-um (Ti II) lines were noticed and theyhad higher intensity than the neutraltitanium (Ti I) lines. Also, neutralsodium (Na I) lines were increasedacross the weld in general. The neu-tral sodium came mainly from thedissociation of cryolite (Refs. 18–20).The asymmetric distribution of the

different species, e.g., Ar I and Na I, isa reflection of the fact that spectraldata collection was from a slice of thearc and not considering the entire arccolumn. (Refs. 21–23). Figure 16 shows the arc tempera-ture of each weld case. At the centerof the weld, the autogenous weldingarc had the lowest temperature ofaround 17,500 K. The Wire 15 weld-ing arc temperature was approximate-ly 2800 K higher at the center of thearc and the weld penetration wasaround 0.6 mm deeper than the auto-genous welds. The Wire 45 weldingarc was approximately 5000 K higherthan the Wire 15 arc and the weldpenetration was around 0.6 mmdeeper. A possible explanation forthis behavior is the greater contribu-tion of arc temperature observed inWire 45. The arc force calculated forthe two welding wires were quite sim-ilar, 0.0042 N for Wire 15 and 0.0045N for Wire 45, implying that the

WELDING RESEARCH


Fig. 15 — Major spectra peaks across Wire 45 weld.

Fig. 17 — Autogenous welding arc constituent concentrations.Fig. 16 — Arc temperature profile.

Fig. 14 — Emission spectral line peaks from Wire 45 weld.


contribution of arc force in welddepth for these two welds would besimilar. As for concentrations of thearc constituents, the autogenous weldcontained approximately 80% neutralargon (Ar I) and 20% first ionized ar-gon (Ar II) in the center of the weld.The second ionized argon (Ar III) andneutral titanium (Ti I) made up about0.2% of the arc in the center of theweld, as shown in Fig. 17. The neutralargon (Ar I) percentage in the arc in-creased with increasing distance fromthe center of the arc — Fig. 17. Thefirst ionized argon (Ar II) decreasedgradually until 3.6 mm away from thecenter of the arc. The concentration of the con-

stituents of the Wire 15 arc (Fig. 18)shows decreasing neutral argon con-centration from 70% at the center ofthe weld to 55% at 3.3 mm away fromthe center, mainly due to the increaseof neutral sodium to around 24%. Atdistances greater than 5.0 mm fromthe center of the arc, only Ar I wasobserved. The concentrations of neu-tral and first ionized titanium, neu-tral vanadium, and neutral magne-sium all increase in the core of thearc. The concentration of the first

ionized titanium also increased to5.5% as compared to 0.6% concentra-tion of neutral titanium. The arc dataon ionized titanium support the high-er calculated arc temperatures. TheWire 45 weld (Fig. 19) shows a largernumber of constituents present andhigher concentration percentages forthe metallic constituents in the arcthan Wire 15 (Table 7). Neutral andfirst ionized barium were also ob-served in the arc due to its high tem-perature (Refs. 24, 25).

WELDING RESEARCH


Fig. 18 — Wire 15 welding arc constituent concentrations. Fig. 19 — Wire 45 welding arc constituent concentrations.

Table 7 — Comparison between Wires 15 and 45

Wire 15 Wire 45

Arc Constriction 21% 5.5%Arc Force 0.0042 N 0.0045 N

Arc Temperature 20177 K 25177 KWeld Penetration 2.9 mm 3.5 mm

(A-TIG)Weld Penetration 2.3 mm 2.7 mm

(Laser)

Table 8 — Corrosion Rate Based on Equation 4in mm/year

0.1% H3PO4 3.5% NaClBase Plate 0.1 0.05

HAZ Weld HAZ

GTAW-AUT 0.1 0.1 0.1GTAW-Ti-CP G2 0.2 0.2 0.1

A-TIG-45 0.2 0.2 0.1A-TIG-15 0.2 0.2 0.1

LAS-Ti-CP G2 0.2 0.2 0.1LAS-AUT 0.1 0.1 0.1LAS-45 0.1 0.2 0.1LAS-15 0.2 0.1 0.1

Fig. 20 — The average mass gain in both solutions compared to the titanium base plate.


Immersion Corrosion Test ofTitanium in NaCl and H3PO4Environment at 25°C

The corrosion resistance of titani-um is based on the presence of a sta-ble, continuous, and adherent oxidelayer. This layer forms spontaneouslyand immediately upon exposure tooxygen. If damaged, it re-forms readilyas long as there is some source of oxy-gen (air or moisture) in the environ-ment (Ref. 26). The immersion corrosion test wasconducted according to NACETM0169/G31-12a. Samples were ex-tracted from both the laser and GTAwelding processes and then preparedby grinding on No. 120 abrasive SiCpaper. Each zone of the weld was stud-ied separately because some welds maybe cathodic or anodic to the base metaland may affect the corrosion rate. Cor-rosion of welded specimens is normal-ly localized and not representative ofthe entire surface. The test durationwas 800 h. Samples were weighed us-ing an electronic analytical balance.The weights of polished and driedsamples before immersion and washedand dried samples after immersionwere measured and the mean valuesdetermined. All samples gained massbecause of the titanium oxide forma-tion. The average mass gain is 2 to 5mg depending on the solution, asshown in Fig. 20. The corrosion ratesof the samples were also calculated us-ing the following equation:

Corrosion rate = (K W)/(A T D) (4)

where K is a constant equal to 8.76 104, T is time of exposure in hours, A isarea in cm2, W is mass loss in grams,and D is density (Ti = 4.54 g/cm3). Thecorrosion rates calculated according tothis equation for titanium in sodiumchloride solution and in phosphoric acidsolution are shown in Table 8. As a re-sult of the sample mass gain, the corro-sion products of the titanium must havedeposited and attached on the Ti sam-ple surface. The above results seem tocorroborate the formation of titaniumoxide phosphate on titanium surface inphosphoric acid. This observation issupported by a number of experimentalresults in literature as follows: by nu-clear reaction analysis, Ferdjani et al.(Ref. 27) found phosphate incorpora-

tion in titanium oxide formed by an-odization in 3.0 mol/L H3PO4 solution;by qualitative EDS analysis, Krasicka-Cydzik (Ref. 28) found titanium phos-phate deposit in the form of an amor-phous thin layer in anodized commer-cially pure titanium in 0.5 mol/L phos-phoric acid; by FT-IR study, Narayananet al. (Ref. 29) also showed the presenceof titanium oxide phosphate in the tita-nium surface anodized in 0.3 mol/Lphosphoric acid. From the experiments, the corro-sion resistance of the titanium baseplate and all laser and GTA welded ti-tanium samples was higher in theNaCl solution than in the H3PO4 solu-tion. The mass gain in the NaCl solu-tion was around 2 mg while it wasaround 5 mg in the H3PO4. It was no-ticed that GTA welded Ti-CP G2 sam-ples had the highest mass gain (~ 5mg) in the H3PO4 solution followedby the fluoride-containing wires andthen the base plate with the autoge-nous weld samples. The mass gain forthe laser weld and heat-affected zonesamples for Ti-CP G2 was about 4 mgwhile it reached 5 mg when GTAW wasused. It was noticed previously in theGTA and laser welding discussion sec-tions that the solid wire (Ti-CP G2) ab-sorbed more energy (compared to theflux cored wires) from the laser beamto result in deeper penetration. Lesspenetration resulted from the GTAWprocess. It seems that the level of en-ergy absorption from both weldingprocesses does not much affect thecorrosion resistance of the solid wire(Ti-CP G2). Rather, it is more relatedto the chemical composition and thepurity of the alloy grade. In general,the higher the purity of CP Ti, thegreater the corrosion resistance. Thesolid wire Ti-CP G2, with very low im-purity levels, has been widely used be-cause it is capable of performing wellin many critical corrosion applicationssuch as marine environments andchemical processing. In seawater, it isfully resistant to corrosion at tempera-tures up to 315ºC (600ºF) (Ref. 30).Hence, A-TIG welds showed a slightlyhigher corrosion rate than the laserwelds. The autogenous welds in bothwelding processes showed the lowestcorrosion rate. It can be concludedthat the corrosion resistance of thewelds produced with Wires 45 and 15are higher than those with Ti-CP G2.

Conclusion

Arc constriction, arc force, and arctemperature are confirmed to be thedriving mechanisms of the A-TIGwelding process. As the cryolite con-tent in the flux increases, the weldpenetration, arc force, and arc temper-ature increase as well. The increase inarc temperature led to increases in arcconstituent concentrations. The arcconstricts with the increase of cryolitecontent in the flux up to a certain per-centage and then decreases with fur-ther cryolite addition. Comparison be-tween the depths of the activated andautogenous welds showed Marangoniconvection does not have a significantcontribution as a driving force in theweld penetration. It was found that arcconstriction and arc temperature arethe two dominant driving mechanismsin weld penetration for fluxes of lowercryolite contents. As for fluxes withhigher cryolite contents, the arc tem-perature is the main mechanisms forweld penetration. Finally, Wires 45and 15 produced welds with much bet-ter corrosion resistance than Ti-CPG2. For all three wires and autogenouswelds, the phosphoric acid solutioncaused greater corrosion than thesodium chloride solution.

1. Bang, K. S., Chirieleison, G., and Liu,S. 2005. Gas tungsten arc welding of titani-um using flux cored wire with magnesiumfluoride. Science and Technology of Weldingand Joining 10(5): 617–623. 2. Hillier, C., Liu, M., Roepke, C., andLiu, S. 2009. A-TIG welding of CP titaniumplates using cryolite-containing flux pastesand flux-cored wires. Proceeding of the AWSWelding Show, Chicago, Ill. 3. Howse, D. H., and Lucas, W. 2000. Aninvestigation into arc constriction by ac-tive fluxes for TIG (A-TIG) welding. Scienceand Technology of Welding and Joining 5(3):189–193. 4. Kumar, V., Lucas, B., Howse, D.,Melton, G., Raghunathan, S., and Vilarin-ho, L. O. 2009. Investigation of the A-TIGmechanism and the productivity benefit inTIG welding. Proceedings of the 15th Inter-national Conference on the Joining of Materi-als and 6th International Conference on Edu-cation in Welding, Helsingr, Denmark. 5. Gurevich, S. 1970. U.S. Patent

WELDING RESEARCH


References


3,551,218. 6. Liu, S., and Perez, F. 2001. Manufac-turing titanium flux-cored arc welding elec-trodes — Feasibility study. U.S. Navy Con-tract N00167- 01-M-0053, Final report. 7. Chandler, J. H. 2004. Spectroscopicmonitoring of hydrogen in welding arcs.Colorado School of Mines, Golden, Colo.,master’s thesis. 8. Converti, J. 1981. MIT, Cambridge,Mass., PhD thesis, mechanical engineer-ing, p. 57. 9. Modenesi, P. J., Apolinario, E. R., andPereira, I. M. 2000. TIG welding with sin-gle-component fluxes. Journal of MaterialsProcessing Technology, pp. 99, 260–265. 10. Lucas, W., and Howse, D. 1996. Ac-tivating flux — Increasing the perform-ance and productivity of the TIG and plas-ma processes. Welding and Metal Fabrica-tion 64(1): 11–17. 11. Burleigh, T. D., and Eagar, T. W.1983. Measurement of the force exerted bya welding arc. Metallurgical Transactions A14A: 1223, 1224. 12. Chirieleison, G. M. 2004. The effectof flux ingredients on flux-cored arc weld-ing on titanium. Colorado School of Mines,Golden, Colo., MS thesis. 13. Lu, S., Fujii, H., Sugiyama, H., Tana-ka, M., and Nogi, K. 2002. Weld penetra-tion and Marangoni convection with oxidefluxes in GTA welding. Materials Transac-tions, Japan, 43(11): 2926–2931. 14. Yang, C., Lin, S., Liu, F., Wu, L., andZhang, Q. 2003. Research on the mecha-

nism of penetration increase by flux in A-TIG welding. Journal of Materials Scienceand Technology 19(1): 225–227. 15. Leconte, S., Paillard, P., Chapelle, P.,Henrion, G., and Saindrenan, J. 2007. Ef-fects of flux containing fluorides on TIGwelding processes. Science and Technology ofWelding and Joining 12(2): 120–126. 16. Howse, D. H., and Lucas, W. 2000. Aninvestigation into arc constriction by activefluxes for TIG (A-TIG) welding. Science andTechnology of Welding and Joining 5(3): 189–193. 17. Hiraoka. K. 1998. Plasma structuresof Ar-H2 mixed gas tungsten arcs deter-mined by spectroscopy measurements.Weld. Int. 12(3): 186–194. 18. Kobayashi, M., and Suga, T. 1979. Amethod for the spectral temperature meas-urement of a welding arc. Arc Physics andWeld Pool Behavior-International ConferenceProceedings, London, UK. Ed. W. Lucas,Abington, UK: The Welding Institute,1980, pp. 25–37. 19. Griem, H. R. 1997. Principles of Plas-ma Spectroscopy. Cambridge, UK: Cam-bridge University Press. 20. Lancaster, J. F. 1986. The Physics ofWelding, 2nd Edition. International Insti-tute of Welding. 21. Deutsch, M., and Beniaminy, I. 1982.Derivative-free inversion of Abel’s integralequation. Appl. Phys. Lett. 41(1): 27, 28. 22. Liu, L. M., Cai, D. H., and Zhang, Z.D. 2007. Gas tungsten arc welding of mag-nesium alloy using activated flux-coated

wire. Scripta Materialia, pp 57, 695–698. 23. Liu, L. M., Cai, D. H., and Zhang, Z.D. 2008. Magnesium alloy weld using man-ganese chloride coated wire. Science andTechnology of Welding and Joining 13(1):44–48. 24. Matsushita, M. 2001. Clarificationof hydrogen reduction mechanism in steelweld metal by means of fluoride additionsin welding flux. Colorado School of Mines,Golden, Colo., PhD thesis. 25. Liu, S., and Perez, F. 2001. Manufac-turing titanium flux-cored arc weldingelectrodes — Feasibility study. U.S. NavyContract N00167- 01-M-0053, final re-port. 26. Handbook of Corrosion Data, 2ndedition. 1995. Materials Park, Ohio: ASMInternational. 27. Ferdjani et al. 1993. Anodic oxida-tion of titanium in phosphoric acid baths:Phosphorus incorporation into the oxide.Journal of Alloys and Compounds, 200:191–194. 28. Krasicka-Cydzik, E. 2004. Gel-likelayer development during formation ofthin anodic films on titanium in phosphor-ic acid solutions. Corrosion Science 46:2487–2502. 29. Narayanan, R., and Sesharderi, S. K.2007. Phosphoric acid anodization of Ti-6Al-4V, Structural and corrosion aspects.Corrosion Science 49: 542–558. 30. http://cartech.ides.com/datasheet.aspx?i=101&E=266. Sept. 25,2013.

WELDING RESEARCH


Call for Papers JOM18

18th International Conference on Joining MaterialsInstitute for the Joining of Materials in association with IIW

Helsingør, Denmark, April 26–29, 2015

Download the brochure detailing topics, expenses, and registration form at:www.aws.org/wj/JOM-18-CallForPapers.pdf

Review the brochure for conference topics. E-mail a title and short abstract of your paper before Nov. 2, 2014.You will receive author guidelines for preparation of the full paper by Nov. 30. The full paper for publication in the Con-ference Proceedings must be received by Jan. 15, 2015. E-mail to [email protected].


Documents

FluxAssisted Gas Tungsten Arc and Laser Welding of ...files.aws.org/wj/supplement/WJ_2014_10_s379.pdf · The use of titanium alloys has been expanding in the aerospace, power-generation,