Non-equilibrium microstructures were obtained in underwater welding due to the fast cooling rates which resulted in the formation of martensite and bainite in the heat affective zone (HAZ) adjacent to the fusion line. The HAZ of under water-welds was not wide as that of similar air welds. The width of coarse grains zone of air welds was much smaller than the width of the corresponding zone of under-water welds. This was because of higher arc and metal temperatures. It was found that the microstructure was dependent upon the waterproof coating used, type of electrode and the number of passes used.

Micro-examination of the welds was conducted in 1971 by Silva which reveal ferrite - pearlite structures in the weld metal and a narrow band of bainite/martensite adjacent to the fusion boundary in the HAZ. With rutile electrode, the martensite band was wider (0.2 - 0.6 mm) than with ironpower type (0.1 mm).

Grubbs and Seth in 1972 reported the presence of a martensitic band adjacent to the fusion boundary with austenitic deposits. According to them alloying elements like chromium and nickel diffused into the base material to give compositions which readily transformed to martensite on cooling.

Masumoto et al. in 1971 reported similar results with 4 mm iron powder electrode at 180 amps. Maximum hardness of 300 Hv (1 kg) in a band less than 1 mm and a partially hardened weld bead and a heat affected-zone of 4 mm, GMA welds at 120 amps and 26 volts showed a peak hardness values of 400 HV (1 kg) and a heat-affected zone width of 6 mm. Hasui, et al. in 1972 reported that for single pass welds the micro hardness approach 400 VHN (200 gm) in a narrow region of 0.5 mm adjacent to opposite side of the plate reduced the original peak hardness to 300 VHN (200 gm). Total heat-affected zone extended for a total of

4.5 mm from the fusion line.

Stalker et al. 1974, indicated that there was a wide range of measured hardness values within one sample and from one weld to another which was partially because of a mixed (hard and soft) microstructure which is typical of mild steel heat affected zone. Despite these differences there was no trend for the heat affected zone at toe of the weld (closer to water) to be harder than the under bead position of the weld. They also reported that there was no apparent relationship between the incidence of cracking and the level of hardness in the heat affected zone.

Brown et al. in 1974 expressed the opinion that the best comparative measure for pre­dicting cooling rate would come from measuring the heat input per unit weld bead size. E 6013 rutile electrodes appeared to result in the lowest heat input while E 7014 rutile iron-powder electrodes were slightly hotter than E-6013, E-7024 rutile iron powder heavy coated and E - 6027 super heavy coated iron powder, both gave much higher heat inputs than E-7014 and were approximately equal to each other in heat input. Localized martensitic transformations appeared in almost all underwater welds immediately adjacent to the fusion line, but extended upto 0.5 mm or less into the heat affected zone. Maximum hardness of 400 HK (100 gm) in SP and 500-600 HK (100 gm) in RP was obtained with E-6013 electrode 4 mm diameter with an energy input of 10-13 kJ/in and 9-11 kJ/in respectively.