Stainless steel MMA welding - fume safety under the spotlight

TWI Bulletin, January/February 1992

Graham Carter joined TWI in 1965 as an analytical chemist in the chemical laboratory. Following a change in emphasis from chemical to instrumental methods of analysis he was responsible for installation and calibration of TWI's first optical emission and X-ray equipment which formed the basis of its current analytical service for members.

In 1983 Graham additionally became involved in welding fume research with a particular interest in reducing fume emission rate and toxicity through changing consumable formulation. Knowledge of fume measurement techniques and rules and regulations required to conduct the work has led to his presence on International and European Standards Committees concerned with fume testing and to the provision of 'on-site' surveys for member companies.

Fumes from MMA welding of stainless steel contain hexavalent chromium and must therefore be controlled to low levels. Control would be made easier if smaller quantities of less toxic fume were emitted. Graham Carter explains.

When welding stainless steel using manual metal arc (MMA) consumables, exposure standards for chromium compounds may be exceeded even though total fume has been controlled to 5 mg/m ³. An OES of 0.5 mg/m ³ has been assigned to trivalent chromium compounds and a lower value of 0.05 mg/m ³ to hexavalent compounds because of their suspected carcinogenic properties.

Calculation ^[3] shows that when fume concentrations of Cr ³ exceed 10% or Cr ⁶ concentrations exceed 1%, control of total fume concentrations below 5 mg/m ³ is required.

Typically, commercial stainless steel consumables produce up to 10% chromium in the fume with 60 to 90% in the hexavalent form so that control of total fume to around 1 mg/m ³ is necessary. Additionally, MMA consumables for welding other alloy types, e.g. hardfacing and nickel alloys may also produce significant quantities of hexavalent chromium in the fume, and fume control to similar low levels may be required. However, sustained control of fume to low levels is difficult and ventilation requirements could be reduced if less fume were produced at source or a change in fume composition could be effected, reducing the concentration of toxic elements.

Work was carried out using stainless steel MMA consumables to investigate the possibility of changing electrode formulation to reduce fume emission rate and the concentration of hexavalent chromium compounds. The influence of general electrode type was examined, as were the effects of a strong deoxidant addition (aluminium) because chromium metal present in the electrode must undergo oxidation before occurring as hexavalent chromium compounds in the fume. Investigators in Japan ^[4] had shown that reduced quantities of sodium and potassium in the coating resulted in lower concentrations of hexavalent chromium in the fume, but commercial consumables had not followed. Evaluation of effects of sodium and potassium was carried out therefore with the area between 0 and 0.6 weight % alkali metal in the coating receiving attention.

Experimental

Comparison of consumables was made on the basis that they deposited weld metal of similar composition. All consumables deposited weld metal having an approximate composition of 18%Cr, 13%Ni, 2.5%Mo, 1%Mn, 0.8%Si, 0.03%C.

Fume emission rates were measured using the Swedish Fume Box technique, ^[5] the effects of formulation being isolated by conducting all tests with parameters held as nearly constant as possible. Voltages used were 'normal' for each consumable tested, as determined by the welder and a test current of approximately 150A AC was used throughout, this being a typical value for the 4mm electrodes used.

Electrodes

Experimental electrodes corresponded to the four commonly available commercial types, i.e. basic, basic rutile, acid rutile and synthetic. A synthetic formulation is taken to be one where alloying elements are present only in the coating. A base formulation, of which all other consumables were variants, was manufactured for each consumable type ( Table 1).

Table 1 Base formulations of different electrode types

All core wires were 4mm diameter. For the basic, basic rutile and acid rutile types, the core wire was BS 970: Part 1 1983 304 S11 stainless steel, but for the synthetic type a low carbon iron was used. The synthetic consumable was extruded through an 8.7mm die, whilst a 6.7mm die was used for the other types. A 1:1 mixture of sodium and potassium silicates was used as a binder. The electrodes were baked at 420°C and used as received.

Variables

The variables studied were:

• Effect of electrode type;
• Effect of deoxidation with aluminium;
• Effect of alkali metal additions.

Electrode type

Basic, basic rutile, acid rutile and synthetic formulations, as shown in Table 1, were studied. The synthetic formulation most closely resembled a rutile type.

Deoxidation

Aluminium metal power was added to basic, basic rutile and acid rutile formulations in quantities ranging from 0.5 to 2% ( Table 2). Iron and ferro-silicon additions were varied to permit incorporation of the aluminium and to allow for expected improvements in weld metal recovery.

Table 2 Formulations of electrodes used to study effects of deoxidation

All other compounds were as in the base formulations.

Alkali metal additions

The influence of sodium and potassium on fume emission rate and composition was studied using basic rutile formulations. The electrodes were made using a silica binder system and additions of sodium and potassium were made as the titanate compound which replaced part of the rutile present in the base formulation ( Table 3). Additionally, it was necessary to replace feldspar in the base formulation with alumina and silica (in the same ratio as in feldspar) because feldspar contains significant quantities of sodium and potassium. In consumables, BR5, 6, 7, 8, sodium titanate was absent, but potassium titanate was varied from 0-3%, whilst in another consumable, BR10, potassium titanate was omitted, but 2.8% sodium titanate was added. A further consumable, BR9, contained neither sodium nor potassium compounds, but used a lithium silicate binder system.

Table 3 Formulations used to study effects of alkali metals

Results

Electrode type

In considering effects of electrode type, it was found that fume emission rate increased in the order synthetic, basic rutile, acid rutile, basic ( Table 4). This was also the order of increased arc voltage, although the relationship was not linear ( Fig.1). It is well known that higher arc voltages produce increases in fume emission rate so that the observed differences in emission rate could be attributed to changes in arc voltage, stemming from differences in formulation, but the results may equally well be associated with coating volatility. Detailed information on vapour pressures of coating compounds at arc temperatures is not available, but the rutile type electrodes, which contained the largest quantities of refractory materials, provided lower fume emission rates than the basic electrodes.

Table 4 Fume emission rate results for electrode type tests

Table 5 shows that the total chromium concentrations varied between 4.3 and 6.5%, the lowest value being obtained using the basic consumable. Concentrations of hexavalent chromium were very similar for all the consumables tested. Thus, there was no significant hygienic advantage in one formulation method over another, and control of total fume to approximately 1 mg/m ³ would be required for all electrodes.

Table 5 Fume compositions for electrode type tests

In principle, the net emission rate of hexavalent chromium may be obtained by combining emission rate data with fume composition. For the present data, the emission rate of hexavalent chromium varied from 0.1 mg/sec using the synthetic or rutile consumables to 0.28 mg/sec ( Table 4) when the basic consumable was used. Thus, the synthetic consumable, where the toxic elements had been incorporated in the flux, was hygienically equivalent to, or better than, the other consumables.

Deoxidant additions

Emission rate was reduced marginally from 8 to 6 mg/sec ( Table 6) when aluminium levels in the coatings of basic electrodes were increased from 0-2%, but similar aluminium additions made to basic rutile and acid rutile types had little effect.

Table 6 Fume emission rate results for deoxidant tests

Small, progressive, but hygienically insignificant reductions in fume concentrations of total chromium, which could be correlated with aluminium coating additions, were observed for all electrode types ( Table 7). However, concentrations of hexavalent chromium remained virtually unchanged.

Table 7 Fume compositions for deoxidant tests

Thus, it must be concluded that significant changes in fume composition cannot be effected by increasing aluminium additions in stainless steel consumables. The reason for this is not fully understood, but may be connected with the fact that the deoxidant addition was effectively made directly into the high temperature arc region, and other work ^[6] indicates that hexavalent chromium is probably formed away from the immediate arc area.

Alkali metal additions

Significant changes in fume emission rate did not result from the small changes in alkali metal content used ( Table 8). However, their effect on fume composition was marked ( Table 9). As the amount of potassium in the consumables increased, the concentration of potassium and fluorine in the fume rose, and was accompanied by a decrease in the concentrations of other fume elements ( Fig.2).

Table 8 Fume emission rate results for alkali metal tests

Table 9 Fume compositions for alkali metal tests

Conversely, as the amount of potassium in the coatings decreased, the concentration of total chromium in the fume rose to a level where, when potassium was omitted from the coating, the concentration of trivalent chromium in the fume approached a level where control of fume to 5 mg/m ³ would barely maintain trivalent chromium below its own OES.

However, as the concentration of trivalent chromium rose the concentration of hexavalent chromium fell to a level where control of total fume to recommended limits would maintain hexavalent chromium below its OES ( Fig.3). When 0.2% potassium was present in the coating, control of total fume to 5 mg/m ³ would mean that hexavalent chromium was at its OES. In tests using consumables containing 0.4% sodium, BR10, or using a lithium silicate binder, BR9, similar levels of hexavalent chromium were obtained to those from consumables using 0.2% potassium in the coating. However, all the consumables produced had friable coatings and as such, were not suitable for commercial use.

General implications

In this work stainless steel MMA consumables depositing similar composition 19.12.3L weld metals gave rise to fume with similar levels of chromium. This occurred when acid rutile, basic rutile, basic and synthetic consumables were tested. Regulations relating to fume control dictate that fume from all the consumables would require control to approximately 1 mg/m ³, this low level being dictated solely by the concentrations of hexavalent chromium present.

It seems probable that other consumables depositing 19.12.3L weld metal will produce fume with similar chromium levels to those obtained here and that consumables depositing weld metals of different stainless steel alloy types will produce fume with compositions specific to the grade of weld metal concerned. ^[7] In this case, it would be possible to ascertain the level of fume control required simply from a knowledge of the weld metal composition deposited, although such a system would tend to ignore future benefits gained from reduced fume toxicity as a result of improved electrode formulation.

At the same time, the consumables tested emitted fume at different rates, the basic consumable providing emission rates which were three times those observed for the basic rutile and synthetic types. Thus, if fume is a major consideration, rutile types and possibly the synthetic method of formulation would seem preferable, although it should be remembered that the all positional properties of synthetic consumables are likely to be inferior to those of the 'normal' rutile types and that the weld bead shape may be rather different.

Production of hexavalent chromium in the fume was greatly reduced by minimising the quantities of sodium and potassium in the coating. When sodium and potassium were excluded from the coating, the concentration of hexavalent chromium was reduced to a level where control of total fume to its OES would maintain hexavalent chromium below its own OES. However, in order to manufacture consumables without sodium and potassium, it was necessary to use either a silica or a lithium silicate binder system and soft, friable electrodes resulted where the coating did not adhere adequately to the core wire. The principles involved in reducing hexavalent chromium formation are not, therefore, commercially viable until a superior binder system is found. Literature is now available ^[8] claiming an improved lithium silicate binder which makes possible manufacture of robust electrodes for commercial use but the knowledge required for successful use of the binder has not been revealed.

Summary and conclusions

Tests were carried out using stainless steel MMA consumables to determine effects of electrode formulation on fume emission rate and composition. Assessment was made of acid rutile, basic rutile, basic and synthetic formulations using the Swedish Fume Box test and a nominal current of 150A. All test parameters were standardised, as far as possible, thus making changes in fume emission rate and composition attributable to differences in electrode formulation. From the studies carried out, the conclusions were:

• Fume emission rates varied by a factor of three when using stainless steel consumables of different type. Basic consumables produced the most fume but the amount decreased in the order acid rutile, basic rutile and synthetic, the synthetic consumable being a 'rutile type' formulation.
• Changes in emission rate observed for the consumable types tested stemmed from differences in working arc voltage and coating volatility. Coatings operating at lower arc voltage and containing larger proportions of refractory materials such as rutile provided lower fume emission rates.
• Acid rutile, basic rutile, basic and synthetic consumables depositing weld metal of similar composition produced welding fume having approximately the same concentrations of hexavalent and total chromium.
• Additions of aluminium as a deoxidant to the coatings of acid rutile, basic rutile and basic consumables made no significant difference to fume emission rate or composition.
• Elimination of compounds containing sodium and potassium from the coatings of basic rutile consumables reduced the fume concentration of hexavalent chromium to a level where normal fume control would maintain its concentration below the exposure limit. However, this method of formulation caused trivalent chromium levels to rise so that control of total fume to 5 mg/m ³ would barely maintain exposure levels of the trivalent chromium below its OES. The coatings used either a silica or lithium silicate binder system and were friable.
• The use of rutile-type coatings can be expected to reduce the hygienic burden in practice, provided that operating and service requirements are fulfilled.

References