The reprocessing and re-use of slag as flux in submerged-arc welding - 1

TWI Bulletin, February 1988

Graham Carter, LRSC, is a Principal Chemist in the Materials Department.

In principle, cash savings can be made by reclaiming submerged-arc welding slag for further use as flux. It is shown that this may be feasible for fused acid fluxes for non-critical applications, but should not be used when agglomerated basic fluxes are required.

In submerged-arc welding, the weld pool formed is shielded from atmospheric contamination with a layer of granular flux material manufactured predominantly from minerals. During welding, much of the flux is melted by the high temperatures experienced, and then cools to form a solid slag covering the weld. Unmelted flux is generally collected and re-used, but the slag is discarded. Substantial quantities of slag are produced during welding, the composition of which bears a general resemblance to that of the original flux.

The possibility exists, therefore, of reprocessing and re-using the slag as flux providing there is no serious deterioration in its properties. This is not common practice in Europe, at least for welding, although it is understood that such reclaimed slag is used in some surfacing fluxes. However, trade advertisements have indicated that reprocessing is currently used in North America to produce a range of flux types for structural welding. ^[1] In principle, reuse of welding slag can lead to considerable cost savings, possibly of the order of £2-4/kg of flux, but there is virtually no published information on the practicality and reliability of submerged-arc welding with reclaimed slag as the flux.

A brief study was therefore initiated to explore the feasibility of reprocessing and re-using slag produced in submerged-arc welding as flux, with particular reference to welding ferritic steel. In general, fluxes may be divided into six categories, depending upon chemical composition/properties, and mode of manufacture. ^[2] Two fluxes were chosen for examination to encompass this range, these being an acid fused flux and a basic agglomerated type. Acid fused fluxes are for general purpose use, where there is little demand on the weld metal mechanical properties, usability and ease of operation being an important consideration. Basic fluxes, on the other hand, are intended for use in critical applications and require not only acceptable usability but good weld metal mechanical properties, often with particular emphasis on weld metal toughness.

These two flux types were used on the basis that they would differ appreciably in their physical and chemical behaviour. Bearing in mind their method of manufacture, fused fluxes would be expected to give slags that could readily be reprocessed by grinding and sieving to give a product closely similar to the original flux. This would not necessarily be the case with agglomerated fluxes, for which the temperatures experienced in welding greatly exceed those in manufacture, and cause a considerable change in the physical form of the slag relative to the original flux. From the compositional viewpoint, acid fluxes based on stable silicate systems should show less chemical change on welding than other types, although even in this case transfer of elements such as manganese to or from the flux/slag will occur, together with changes in the levels of minor or impurity elements. However, compositional changes with basic fluxes should be more marked, especially in terms of factors such as reactive flux components and deoxidant additions.

The work undertaken on the chosen flux types was in two phases. First, slag produced during routine welding in the Arc Welding Department of The Welding Institute was used for study of reprocessing procedures. Second, bead-on-plate welds were produced using virgin flux and recycled slag: usability and ease of operation were examined with reference to bead profile and ease of slag detachability and the weld beads were subjected to chemical analysis and metallographic examination.

Experimental procedure

Material types

Oerlikon Welding OP121TT was selected as the basic agglomerated flux, used in combination with Murex Bostrand WB34.0mm diameter submerged-arc welding wire. ESAB OK 10.40, a manganese silicate flux of the manganese alloying type, was chosen as the fused acid flux, and was used together with Oerlikon Welding 4.0mm diameter Si wire. The plate material was 25mm thickness BS4360:1986 Grade:50D.

Chemical analysis

For flux and slag, the principal analytical technique used was X-ray fluorescence. In this case, 1g of finely divided sample was fused with a mixture of 2g lanthanum oxide and 9g lithium tetraborate at 1200°C for 12min. The resultant fusion mixture was cast into a bead of suitable dimensions for X-ray analysis. Analysis for fluoride used a pyrohydrolysis technique and sodium was determined using flame emission spectroscopy. Carbon and sulphur determinations were conducted using a Leco CS-125 combustion infrared analyser.

Wire, plate and weld metal analysis was performed predominantly using optical emission spectrographic techniques. Plate analysis was obtained by averaging the results of several sparks placed on the through-thickness plane. Weld metal analysis consisted of averaging the results from sparks placed on prepared surfaces of the weld cap of the two sections removed for analysis. Wire analysis was performed on remelted buttons. Small solid pieces of plate, wire and weld were analysed for oxygen and nitrogen, using a Leco TC136 combustion inert gas analyser.

Slag reprocessing

Reprocessing trials to convert slag into a suitable form for re-use as flux were conducted on slag obtained from routine welding using the OP121TT/WB3 flux/wire combination in The Welding Institute's Arc Welding Department.

Slag reprocessing was performed using a Christy Laboratory Mill. The grinding member is a four armed steel cross which revolves at high speed inside a drum and shatters the sample, which is fed down a shute into a feeder aperture. The sample is collected in a cloth bag which is suspended beneath a size screen. As supplied, the mill rotated at a high speed (8000rpm) and a crushed product of fine particle size resulted. However, by reducing the operating speed using a different pulley system and fitting a 2.5mm screen, it was possible to produce a sample of reprocessed slag which visually resembled the original flux ( Fig. 1). The operation was fast, and it was possible to reprocess a kilogram of flux in a few minutes.

Following these trials, slags from the two fluxes were crushed for welding trials using the Christy mill. The particle size distribution was determined for both the original flux and crushed slag ( Table 1). It was found that sieving crushed slag to either a 60 or 85 mesh sieve and discarding the fines provided a reasonable approximation to the particle size of the original flux and that the recovery rate was about 80%. Each sieved fraction was analysed as described earlier. No attempt was made to remove trapped spatter when reprocessing either flux.

Table 1 Sieve size analysis of flux and crushed slag for OP121TT and OK 10.40.

Welding equipment

The welding equipment was an ESAB LAE 1600 thyristor controlled transformer rectifier, and a beam mounted ESAB A6 single wire submerged-arc head ( Fig.2). Welding was carried out using DC electrode positive polarity (DCEP). A PAMS II (portable arc monitor system) briefcase monitor ^[3] was used to monitor wire feed speed, welding current and voltage, the last being measured between the contact tip and the workpiece. Travel speed was set using a stopwatch and ruler.

Welding trials

OP121TT/WB3

Welding trials were conducted using reprocessed slag obtained from routine welding in the Arc Welding Department. Initially, three bead-on-plate welds were produced using undried reclaimed slag. Slag retained on an 85 mesh sieve was used to make welds with the plate at room temperature and at 130°C. Additionally, a weld was made on heated plate (130°C) using slag which had been retained on a 60 mesh sieve, i.e. more fine particles had been removed. To allow direct comparison of operability and weld metal properties with those normally obtained using the OP121TT/WB3 combination, welds were deposited at room temperature and 130°C using virgin OP121TT flux. The five weld beads thus produced were made on a single 600 x 300 x 25mm panel of plate material. Welding parameters were set whilst using the virgin flux with the plate at room temperature, and no further adjustment was made throughout the tests ( Table 2).

Table 2 Welding parameters for trials using OP121TT/WB3

All welds were produced using DC+ polarity and a travel speed of 450 mm/min

Using the recycled slag, welding behaviour in this first series of tests was impaired in that some arc instability was manifest, while the resultant slag showed a 'dragonsback' ( Fig.3). A second series of trials was performed with reclaimed slag which had been dried at 350°C for 1 hr. On this occasion, welding parameters were adjusted for each test so that current and voltage remained nominally the same. Four welds were deposited; two were made with the plate at room temperature and two with the plate at 130°C. At each plate temperature, bead-on-plate welds were produced using crushed slag which had been retained on either a 60 mesh or 85 mesh sieve. The welds were made on the same panel of parent material.

Slag produced from each weld in both series of trials was collected and analysed. A summary of the tests performed and welding conditions used is provided in Table 2.

OK 10.40/S1

Welding was carried out using OK 10.40 flux to produce adequate quantities of slag for reclamation and the welding trials. Following crushing, slag retained on a 60 mesh sieve and dried at 150°C was used for further bead-on-plate welding. No differences were apparent between welding using the reclaimed material and welding using the original flux. Welding was continued, therefore, until sufficient quantities of slag were produced for a second reclamation. This new slag was processed in a similar manner to that previously obtained, and a further series of welds was produced. The final slag, together with slag from each previous stage, was collected and analysed. Welding conditions are given in Table 3.

Table 3 Welding parameters for trials using OK 10.40/S1

All welds were produced using DC+ polarity and a travel speed of 450 mm/min

Welding behaviour and weld metal assessment

Assessment of welding behaviour and weld metal deposited was made on a comparative basis, results from trials conducted using reclaimed slag being compared with those obtained using the original flux. Welding behaviour is a largely subjective quality and was assessed by observing the stability and ease with which welding proceeded. Following welding, slag detachability was noted and general weld bead appearance recorded.

The welds were sectioned transversely, and two samples were removed for chemical analysis and three for metallurgical examination. Macrographs prepared from the metallurgical sections were used to examine weld bead shape, weld area, dilution, depth-to-width ratio and penetration.

Microstructure was examined following polishing to a 1µm finish and etching in nital. A visual semiquantitative assessment was made of the constituents of the microstructure.

References