Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

External hardness limits for welds in cathodically protected C-Mn steel pipelines carrying sour products (September 2000)

R J Pargeter and B J Ginn

Metallurgy, Corrosion, Arcs and Surfacing Technology Group, TWI Ltd

Presented at 'Past Successes - Future challenges', EUROCORR 2000, 10-14 September 2000, Queen Mary and Westfield College, London, UK; published by IOM Communications Ltd


Hardness limits are commonly used to control the risk of cracking in ferritic steel pipelines carrying sour products. The necessary hardness limits are dependent on the hydrogen concentration in the steel, which, with single sided exposure, is much higher adjacent to the sour environment than near the free surface. Thus it has been demonstrated in laboratory work that considerably higher hardnesses are tolerable on the outsides of pipelines carrying sour products, and this effect is recognised, in terms of relaxed limits for capping passes, in BS4515 and EFC 16. The hydrogen concentration near the outer surface of such a pipe will, however, be affected by cathodic polarisation, which would be expected to restrict hydrogen egress. A programme of work has therefore been carried out to determine safe hardness levels on the outsides of pipelines carrying sour products, in the presence of externally applied cathodic protection. Despite severe test conditions, high hardness levels were shown to be resistant to cracking, and the increased confidence given by the new data has allowed further relaxations to the current hardness levels permitted for cap passes in standards to be recommended.


For a pipeline carrying a sour product, the hydrogen concentration in the steel due to corrosion will be greatest at the inner surface. It would be expected that diffusion through the wall will take place with evolution of hydrogen from the external surface, so that the peak hydrogen level in this region will be appreciably lower. Hence, the risk of cracking at the external surface may also be reduced relative to the inner wall, and a higher hardness limit may be acceptable.

Studies [2-4] have confirmed these effects and have formed the basis of the hardness limits propounded in BS4515 [1] , which permits the external hardness of linepipe welds carrying sour products to be up to 275-300HV (depending on wall thickness). Similarly, 275HV is recommended as a safe external hardness in EFC publication No 16 [5] . The earlier work employed full scale ring samples, stressed by internal jacking and containing a severe environment, and in fact, external hardness levels well above 300HV were found to be tolerable without causing cracking, so that this limit has been considered conservative. However, it must be recognised that in service, the external surface of linepipe would normally be subject to cathodic polarisation (CP) to provide protection from the seawater or other external environment. This may restrict hydrogen egress from the outer surface, and thus, lead to a higher hydrogen content.

In the absence of published data on the significance of this effect, the present work was carried out to determine whether current limits are still conservative.

Experimental procedure


The experimental programme comprised two phases. The first phase consisted of the production, exposure to sour NACE solution and examination for cracking of five ring test specimens. These specimens had a number of longitudinal, single pass bead on plate (BOP) welds on the external surfaces. In the second phase a validation test was performed under identical test conditions to those of phase 1, but using a specimen incorporating a centrally located full penetration girth weld. Additional hard zones were created on the second phase ring test piece by the deposition of further BOP welds and from the attachment of studs by underwater friction welding.


Five different pipe samples were selected for inclusion in the phase 1 exposure trials, following single bead weld deposit trials, designed to ensure a hardenability response covering the range 300-400HV5. The pipe samples selected had wall thicknesses from 17 to 25mm, diameters from 356 to 914mm, and had been manufactured to either API 5L X52 or X60 specifications. Details of the pipe dimensions, strength and chemical compositions are given in Table 1.

The validation test was undertaken using a sample from one of the five pipes (IC4040) included in the phase 1 trials. The wall thickness of this pipe, as supplied, was 18mm. This was reduced to 9.5mm by machining the outer wall surface after a light skimming cut had been made to centralise the bore.


Table 1 Summary of Pipe Materials

Ident. No.API 5L GradeChemical Composition Wt %
1B766 X60 0.13 0.002 0.009 0.41 1.51 0.02 0.01 <0.005 <0.002 <0.005 0.029 <0.002 0.048
1B823 X60 0.11 <0.005 0.012 0.32 1.33 0.02 0.04 0.01 0.05 0.03 0.031 <0.002 0.031
API 5L X60 0.26
1C4039 X52 0.15 <0.002 0.014 0.28 1.19 0.05 0.09 0.01 0.03 0.09 0.032 <0.002 0.034
1C4040 X52 0.16 0.004 0.014 0.37 1.33 0.04 0.03 0.01 0.04 0.08 0.034 0.002 0.044
EX-621982 X52 0.15 0.004 0.012 0.21 1.25 0.13 0.14 0.03 0.05 0.23 0.021 0.005 0.021
API 5L X52 0.29

Table 1 Summary of Pipe Materials - Continued

Ident. No.Yield Strength, MPaSize, mm
1B766 423 700 24
1B823 413 914 22
API 5L 413    
1C4039 378* 356 25
1C4040 404 457 18
EX-621982 399* 457 17
API 5L 358    


Environmental testing

Construction of test cells

Test cells were produced from pipe sections assembled to produce a leak-proof container with a rigidly fixed lid. Short lengths of pipe 350-600mm long, were sawn, then lathe turned, to produce square end faces. One end of each ring was further machined to reduce the wall thickness to 3mm for a distance of approximately 100mm. These reduced wall thickness ends, which would ultimately form the base of the test cell, were then capped with a 3mm thick steel disc, a leak-proof joint being obtained by circumferential welding. It was established, using strain gauges, that the stress distribution within the ring test piece was not influenced by this sealing technique.

The lids of the test cells were constructed from either 10mm thick steel or 20mm thick perspex. The steel lids were centrally fitted with a 200mm diameter, 10mm thick, perspex window. This window allowed inspection of the interior of the test cell during the exposure period, as well as providing access for filling with solution and for gas inlet and outlet pipes. A gas-tight seal was obtained on the test cells by pulling the lids down against a flexible sealing ring after stressing, using bolts tapped into the wall of the test pipe, or by applying heavy weights.

Production of hard zones

BOP deposits
In phase 1, each ring test piece was produced with four different hard zones on the outer pipe surface, [ Fig.1]. Each hard zone consisted of three separate deposits approximately 100mm long. The locations of these deposits were such that different, but known, stress levels could be applied to each weld in any one quadrant. The hard zones had aim HAZ hardness levels of 300, 350, 375 and 400 HV5. The welding process was mechanised MAG using 1.0mm diameter Bostrand LW1 filler and CO2 shielding. The welding conditions were selected from initial trials, to give appropriate hardness levels.

Fig.1. Test piece design for Phase 1
Fig.1. Test piece design for Phase 1

Bead on plate deposits were also made on the validation test piece of phase 2. Two beads approximately 150mm long were deposited at three locations on the outer pipe surface. The bead locations corresponded with the 90% yield position when the ring was stressed. Each pair of weld beads was displaced from the girth weld by at least 25mm to avoid reheating. The aim hardness levels were 350, 375 and 450HV5. The welding process was the same as for the phase 1 deposits.

Girth weld
The full penetration girth weld used to manufacture the validation ring test piece, Fig.2, was produced using a narrow groove weld preparation following established practice. This preparation had a root face of 1.5mm, a root radius of 2.5mm and an included angle of 6° with no root gap. A copper backing bar with a shallow groove was fitted. The welding process was mechanised MAG and welding position was pipe horizontal (1G) i.e. rotated. Bostrand LW1 filler was used with CO2 shielding.

Fig.2. Validation test piece after exposure
Fig.2. Validation test piece after exposure

Three passes were required to fill the joint. On completion of the capping pass, and after the weld had cooled, a further two circumferential beads were deposited, one at each weld toe on the outer pipe surface. The arc energies of these two additional beads were adjusted such that the associated HAZ regions had hardness levels, estimated from welding trials, in one case of ~350HV, in the other 375-400HV. Similarly, arc energy and preheat conditions for the root pass were adjusted to give an estimated HAZ hardness level of <250HV.

Underwater friction welds
For the validation test only, additional hard zones were produced by attaching studs using the underwater friction welding process. Six, threaded (½" UNC), solid carbon steel studs were attached to the outer surface of the assembled ring piece. These were located in two groups of three each side of the girth weld at the 90% yield position and in a line parallel to the pipe length, [ Fig.2]. One group of studs was welded following an established procedure for this type of welding in which a polystyrene shield produces an insulating gas shield when heated. The second set were welded with the shield missing and had as a result a faster cooling rate and significantly higher HAZ hardness levels. The studs were placed approximately 75mm apart and such that none were within 25mm of the toes of the circumferential weld. All the underwater friction welding was performed by a specialist contractor, Circle Technical Services Limited.

Method of stressing ring pieces
Stressing of the ring test pieces was by means of an internally fitted screw jack mechanism, which contacted the pipe wall through two diametrically opposed anvils. For the phase 1 test pieces the anvils were 150mm long and had a square section with 50mm sides. The anvils used for the validation ring test were, of necessity, more substantial with cross section dimensions of 75mm and length 450mm. The screw jacks and anvils were made of carbon steel. One face of each anvil was 'V' shaped (included angle of 120° and a radius of 2mm) to give a line contact with the pipe wall. The larger anvils had a shallow groove ground across the 'V' shape at the mid-length to prevent localised contact with the girth weld root pass penetration. The steel screw jacks remained in place throughout the exposure period and were not protected against corrosive attack.

Individual ring pieces were stressed by manual operation of the screw jacks before the test cells were sealed. Loading continued until it was determined, from strain gauges bonded to the outer pipe wall in the immediate vicinity of the anvils, that yield stress had been reached at that location. The micro strain requirement at the pipe surface was calculated for each pipe material from yield stress values [ Table 1] reported by the manufacturer, or determined, as part of this project using small tensile test pieces.

Exposure testing
The assembled test cells, comprising circumferentially stressed ring pieces and gas tight lids, were loaded into tanks for exposure to artificial seawater solution. The volume of seawater was at least equal to the displacement volume of the test cell. The tanks were filled with artificial seawater to a level approximately 50mm below the top of the test cell and covering the weld deposits. Cathodic protection (see below) was initiated and the potential allowed to stabilise for a period of 48 hours. The test cells were filled to within 10mm of the lids with sour test solution to NACE TM0177 Method A, using 5% sodium chloride with 0.5% acetic acid and sealed. De-oxygenation of the NACE solution was undertaken in-situ in the phase 1 tests because of the relatively large volumes used. This was achieved by streaming nitrogen gas through the test cells for approximately 72hrs. For the phase 2 validation test, de-oxygenation was carried out in a separate vessel and the solution transferred to the test cell under a covering of nitrogen. The H 2S gas flow through the test cells then commenced. The flow rate was initially approximately 5 litres/min for the first 24hrs. This was subsequently reduced to approximately 0.5 litres/min for the remainder of the 720hr exposure period. Exposure testing was conducted at room temperature. On completion of exposure testing the cells were purged with nitrogen gas for 24-48hrs, to flush out H 2S from the NACE solution and gas lines and enable safe handling, then drained and dismantled.

Cathodic protection
All the test cells in both phases were individually subjected to potentiostatically controlled cathodic protection at an impressed potential of -1100mV, relative to a Ag/AgCl reference electrode. A 2-3 metre length of 0.5mm diameter platinum wire was used as an anode. This wire was suspended in the seawater, using plastic hangers, around the outer circumference of the test cell and at a similar depth to the weld deposits. The distance of the anode wire from the ring surface varied from 50 to 100mm. The impressed potential was allowed to stabilise over a 48hr period before H 2S gas was introduced into the NACE solution. Cathodic protection was maintained throughout the 720hr exposure period.

Monitoring during the exposure period
All the test cells were checked daily, excluding weekends, to ensure that the original set conditions were maintained. These checks included ensuring the continuation of H 2S gas flow, of cathodic protection at the required potential and the monitoring and recording of temperature variations in the test containment laboratory. For the phase 1 ring test series the pH of the NACE solution at the start of testing, prior to H 2S streaming, was 2.5, that of the seawater in the outer cells 8.2. On completion of the 30 day exposure period the pH of the NACE solution in two of the test cells (4 & 5) were measured, and values of 3.4 and 3.6 respectively were recorded; the corresponding pH of the seawater was 6.4. Hydrogen sulphide concentrations measured in the NACE solution of the same two test cells after 10 days exposure were 3614ppm and 4100ppm respectively. For the phase 2 validation test the pH of the NACE and seawater solutions at test initiation was 2.8 and 8.3 respectively; on completion of the exposure period these were 3.5 and 7.3 respectively. Monitoring of the time for hydrogen evolution in individual ring pieces was not undertaken. However it is known from previous work undertaken at TWI [4] and elsewhere [6] that, for the type of test conditions employed in this project, hydrogen evolution occurs some 12-48 hours after the initiation of H 2S flow.

Examination of hard zones after exposure

Phase 1 - Ring test pieces
Segments of the ring test pieces from phase 1 containing hard zones were extracted by flame cutting from the test cells and subjected to magnetic particle inspection (MPI) for evidence of cracking. Care was taken during the flame cutting to prevent any excessive heating of the weld zone regions, which might have influenced their physical properties, in particular, the hardness levels. Following MPI inspection and visual examination of the inner and outer pipe surfaces in the weld regions, individual weld beads and associated HAZ, were extracted from the pipe segments by cold sawing. Initially, a single transverse section from the mid-length position of each weld deposit was prepared to a 6µm diamond finish. One further transverse section was taken from the hardest HAZ region of each pipe. Examination of all sections for cracking and blistering was conducted firstly with the prepared surface in the 'as polished' condition, then secondly, after light etching in 2% nital. On completion of the metallographic examination, hardness surveys, (Vickers HV5), were made of the hard zones.

Phase 2 - Validation ring test piece
The girth weld region of the validation test piece was examined ultrasonically after exposure testing, and before sectioning. The ring piece was then sectioned and subjected to MPI inspection of both inner and outer weld surfaces. Transverse weld sections were extracted from the girth weld, BOP deposits and friction welds. The girth weld sections were taken from regions subjected to 100% and 90% yield stress at the weld cap surface. The BOP sections were extracted at approximately mid-length position. Transverse section through the friction welds were taken centrally. Sections were prepared as previously for metallographic examination and hardness surveys were conducted in the hardened regions.


Phase 1 - ring test pieces

No crack like type defects were identified by either visual or MPI examination of the outer pipe surface in the weld regions, although indications believed to be due to weld cold laps were reported in the toe regions at the weld start positions in some deposits. None of the weld zones in this phase of the project were subjected to ultrasonic examination. The results of metallographic sectioning confirmed the presence of weld toe intrusions in several of the sections examined. It was evident from the nucleation of ferrite on the surfaces of these features that they arose from the welding operation, and are not corrosion related.

No evidence of any hydrogen assisted stress corrosion cracking was revealed by any of the various examinations carried out. The results also indicate that, for the test conditions employed in this phase of the project, the five pipe materials assessed have substantial resistance to hydrogen pressure induced cracking (HPIC). Indeed, the worst example of this type of cracking was an isolated 340µm crack associated with W17 in pipe No.Ex-621982. This particular feature was located beneath the root of the weld deposit at approximately 10mm from the outer pipe surface and 7mm from the weld fusion boundary.

The results of hardness surveys are presented in Table 2. Hardness levels of over 390 HV were achieved and stressed to 90% yield in all pipes.


Table II Summary of results of Phase 1 tests

Material Type API 5L -Ident. No.Arc Energy kJ/mmWeld No.Stress Level % YSHardness HV5MPI InspectionMetallographic Examination
X52 1C4039 0.4 W8 90 399 349 - 399 NC NC TD
0.4 W7 50 - - NC TD NE
0.4 W9 25 - - NC NC NE
0.8 W5 90 369 303 - 369 NC NC NE
0.8 W4 50 - - NC NC NE
0.8 W6 25 - - NC NC NE
1.0 W2 90 367 329 - 367 NC NC NE
1.0 W1 50 - - NC NC NE
1.0 W3 25 - - NC NC NE
1.8 W10 90 329 298 - 329 NC TD NE
1.8 W12 50 - - NC NC NE
1.8 W11 25 - - NC NC NE
X52 EX - 621982 0.7 W17 90 401 347 - 401 NC NC (HPIC?) TD
0.7 W16 50 - - NC NC NE
0.7 W18 25 - - NC NC NE
1.0 W23 90 374 345 - 374 NC NC NE
1.0 W22 50 - - NC NC NE
1.0 W24 25 - - NC NC NE
1.3 W21 90 379 325 - 379 NC NC NE
1.3 W20 50 - - NC NC NE
1.3 W19 25 - - NC NC NE
1.9 W14 90 315 268 - 315 NC NC NE
1.9 W15 50 - - NC TD NE
1.9 W13 25 - - NC NC NE
X52 1C4040 0.5 W31 90 394 339 - 394 NC NC NC
0.5 W30 50 - - NC NC NE
0.5 W32 25 - - NC NC NE
0.8 W28 90 367 339 - 367 NC NC NE
0.8 W27 50 - - NC NC NE
0.8 W29 25 - - NC NC NE
1.2 W26 90 358 299 - 358 NC NC NE
1.2 W25 50 - - NC TD NE
1.2 W26A 25 - - NC NC NE
0.6 W34 90 396 339 - 396 NC NC NE
0.6 W33 50 - - NC NC NE
0.6 W35 25 - - NC NC NE
X60 1B823 0.2 W46 90 396 343 - 396 NC NC NC
0.2 W47 50 - - NC TD NE
0.2 W48 25 - - NC NC NE
0.6 W37 90 364 321 - 364 NC NC NE
0.6 W38 50 - - NC NC NE
0.6 W39 25 - - NC TD NE
1.1 W40 90 329 280 - 329 NC TD NE
1.1 W41 50 - - NC TD NE
1.1 W42 25 - - NC TD NE
1.5 W43 90 296 252-296 NC NC NE
1.5 W45 50 - - NC NC NE
1.5 W44 25 - - NC NC NE
X60 1B766 0.8 W49 90 391 367 - 391 NC NC TD
0.8 W51 50 - - NC TD NE
0.8 W50 25 - - NC TD NE
1.1 W59 90 391 329 - 391 NC NC NE
1.1 W60 50 - - NC NC NE
1.1 W58 25 - - NC TD NE
1.6 W56 90 347 288 - 347 NC NC NE
1.6 W57 50 - - NC NC NE
1.6 W55 25 - - NC TD NE
1.8 W52 90 351 288 - 351 NC NC NE
1.8 W54 50 - - NC NC NE
1.8 W53 25 - - NC NC NE
HPIC = hydrogen induced crack   TD = toe defect (intrusion)   NE = not examined   C = cracking   NC = no cracking


Phase 2 - validation test piece

An initial visual examination of the weld zones on both external and internal pipe surface of the test piece did not reveal the presence of any surface breaking cracks although hydrogen blistering was noted on the inner pipe surface, [ Fig.3a]. Further examination by MPI indicated the presence of numerous transverse cracks in the root weld metal, [ Fig.3b]. In places these cracks extend into the weld HAZ. These transverse cracks were present throughout the complete circumferential length of the root pass although the frequency increased in those regions placed in tension during exposure testing. In addition cracking associated with hydrogen blistering was noted, [ Fig.3a]. An ultrasonic examination of the girth weld zone conducted on the outer pipe surface did not indicate any crack like defects in the weld metal or HAZ regions. The results of MPI and ultrasonic examinations made of the validation test piece are summarised in Table 3.


Fig.3. MPI indications showing internal blistering and cracking of girth weld

Table III Summary of results of Phase 2 validation test

Weld DetailsArc Energy
Stress Level
HAZ Hardness, HV5InspectionMetallographic Examination
MPIUTSection No.
Girth Cap 1 0.35 100 447 441-447 NC NC NC NC NC
Cap 2 0.65 100 349 296-349 NC NC NC NC NC
Root 0.48 100 253 241-253 C NC NC NC NC
Stud Shroud ND 90 353 178-353 ND ND NC ND ND
No Shroud ND 90 429 182-429 ND ND NC ND ND
BOP Cap 0.2 90 432 355-432 NC ND NC ND ND
Cap 0.9 90 271 225-271 NC ND NC ND ND
Cap 1.2 90 265 232-265 NC ND NC ND ND
NC = no cracking   C = cracking   ND = not determined   *90% yield stress

No evidence of hydrogen assisted cracking was observed in eternal regions in metallographic sections through the girth, bead on plate, or friction stud welds. Weld toe intrusions were observed in the root pass and capping passes in some girth weld sections, but in all cases they were associated with the welding process.

For the girth weld capping pass, hardness levels up to 447HV5 were measured whilst 429HV5 was measured under one of the friction stud welds. A peak HAZ hardness of 432HV was recorded in the BOP welds.


At the beginning of this programme, a body of data existed which had been generated in earlier research projects. On the basis of these data, relaxations of up to 50HV are currently allowed for external regions of sour service pipelines. The basis of these relaxations is that with single-sided exposure, steel near the surface remote from the environment, from which hydrogen is escaping, will have a lower hydrogen content than that adjacent to the environment, and thus should be able to tolerate higher hardness. The previous experimental data [2-4] had confirmed this theory, and indeed the apparently safe hardness levels were well in excess of the currently permitted 300HV.

In the more recent work [3,4] , hard zones of up to 396HV were created on the outer wall surfaces of samples of API 5L X52 line-pipe containing sour solution. Test pieces were stressed to 100% yield strength. With the exception of one test piece in which sour solution leakage had contaminated the outer pipe surface, no cracking was recorded. Prior to this, Robinson [2] performed a series of similar tests but without applied load. Cracking was generated in some welds but only in those having hardness levels of between 411 and 427HV. Clearly hardness levels of 400HV are unlikely to be acceptable for pipeline fabrication, and some margin between experimental data and permitted hardness would be advisable. Nevertheless, 100HV is a very large margin, and hardness of up to 350HV are permitted for cap HAZ regions in non-sour service by BS4515.

One consideration in relaxing the 100HV margin between experimentally demonstrated safe hardnesses and permitted hardnesses would be a concern over the possible effect of cathodic polarisation on hydrogen escape from the external surface. In the previous experimental work, the external surfaces had been dry and open to atmosphere, whereas operating pipelines are generally coated and/or wet, with cathodic polarisation for corrosion protection.

The present work has demonstrated that even with -1100mV cathodic polarisation, HAZ hardness levels of above 400HV near the outer surface are resistant to cracking in the presence of NACE TM0177 method A solution inside a pipe. In assessing these results, the pipe wall thickness needs to be taken into account, as the steepness of the hydrogen concentration gradient will be affected by this. Hydrogen concentration at one bead depth below the outer surface will be higher in a thinner pipe, due to the steeper gradient, and thus safe hardness levels would be expected to be lower. For this reason, the relaxation in BS4515 is limited to pipes of wall thickness greater than or equal to the minimum used to generate the experimental data [3,4] , namely 9.5mm. In the first phase of the present work the thinnest pipe wall tested was 17mm, but applicability of the results down to the current limit of 9.5mm has been confirmed by the Phase 2 trial, in which HAZs of over 400HV were shown to be resistant to cracking.

One aspect which has been tested to a limited degree by the Phase 2 trial is weld metal hardness. The two capping runs, added to generate the desired HAZ hardnesses, had weld metal hardnesses of up to 269HV5. These runs will have had high longitudinal residual stresses, and were additionally stressed in this direction by the ovalisation, and did not crack. (All previous trials have employed longitudinal oriented welds with overfill intact, so that the ovalisation will not have stressed the weld metal regions as severely as possible). Although it is not necessarily possible to apply HAZ hardness limits to weld metals, the margin between experimental HAZ limits and the highest weld metal hardness which has resisted cracking is such that one can have some confidence that this weld metal value is some way below the threshold for cracking.

The other hard zone which has been investigated in this work is that induced by underwater friction welded studs. Again, HAZ regions of over 400HV have been found to be resistant to cracking.

One unexpected result of the Phase 2 test was the presence of transverse cracking in the weld root. This was mostly in weld metal, but some extended into HAZ material, generally associated with blisters. The root weld metal hardness was below 250HV (243HV5 maximum). This cracking did not affect the testing of the external region, which was the focus of the project, and has not been investigated in any detail. It may be remarked, however, that there have previously been indications that hardness limits for C-Mn weld metals may be lower than for HAZs in sour service [7] and therefore lower limits may need to be specified for weld metal directly exposed to sour environments. The cracking in the present case does provide some confirmation that the test environment was aggressive and hydrogen charging.

Practical implications

The results of this project have been encouraging in that they do not indicate that the present practice of allowing higher hardnesses on the outsides of pipelines is in any way unsafe in the presence of cathodic polarisation. Indeed the additional data suggests that some relaxation in this hardness limit may be appropriate. In considering the acceptable hardness limits, there are three values to be taken into account, namely the current limit for sour service, the current limit for sweet service, and the experimental threshold for cracking. The current situation with regard to BS4515 for the cap region may be summarised as follows:

The simplest option in all cases would be to allow the same cap hardness for sweet and sour service. For HAZs in pipe of 9.5mm wall thickness and above, the quantity of data, and the size of the difference between the current sweet service limits and experimental thresholds are such that this approach can be recommended.

LocationSour limit (HV10)Sweet limit (HV10)Experimental threshold (HV5)
HAZ, ≥9.5mm 300 325 (cellulosic MMA)
350 (other processes)
HAZ, <9.5mm 275 325 (cellulosic MMA)
350 (other processes)
Weld metal 275 275 ≥ 269 (9.5mm wall)


The simplest option in all cases would be to allow the same cap hardness for sweet and sour service.  For HAZs in pipe of 9.5mm wall thickness and above, the quantity of data, and the size of the difference between the current sweet service limits and experimental thresholds are such that this approach can be recommended.

Below 9.5mm wall thickness there are no experimental data, and greater caution would be advisable. Nevertheless, the fact that cathodic polarisation has little effect on hardness thresholds at higher wall thickness is encouraging, and at least does not imply that permitted hardness levels should be reduced for cathodically protected pipe.

For weld metals, experimental thresholds are not well defined, and there is a concern that weld metals may be more susceptible to cracking than HAZs. Nevertheless, the data generated in this work do not cause any concern over the current 275HV10 limit.

The results of this project also gave confidence in the use of underwater friction welding as a means of attaching studs. There are currently no specific hardness limits for this process in either sweet or sour service, and only limited service experience, however, and it is therefore premature to set a hardness limit.

Summary and conclusions

A programme of work has been carried out to investigate tolerable hardness levels on the outsides of X52 and X60 pipelines of 9.5-25mm wall thickness carrying sour media, which are also subjected to external cathodic polarisation. The internal test medium was NACE TM0177 method A solution, and cathodic polarisation was applied at -1100mV, Ag/AgCl. Test welds consisted of bead on plate runs, friction welded studs, and a girth weld. Samples were stressed, by ovalising, to 100% actual parent material yield stress in hardened areas. No cracking was induced in hardened external regions in any of the test specimens, and as a result the following conclusions can be drawn.

Peak HAZ hardness levels of over 400HV are resistant to hydrogen cracking under the combined influence of NACE TM0177 method A solution on the opposite steel surface, and cathodic polarisation of -1100 mV, Ag/AgCl on the hardened surface, at applied stress levels of up to 100% actual parent material yield stress. This conclusion is applicable to steels of 9.5mm wall thickness and above.

Peak weld metal hardnesses of up to at least 269HV are resistant to cracking under the same conditions as for conclusion 1.

Threshold weld metal hardnesses for root regions in contact with NACE TM0177 method A solution are <243HV5.


The following maximum hardness levels are proposed for welded regions on the outsides of C-Mn steel pipelines of up to X60 strength operating in sour service:


Work is required to explore safe weld metal hardness limits for operation in sour service.

Material/wall thicknessPeak hardness, HV
HAZ, ≥9.5mm 325 (cellulosic welds)
350 (other arc welded processes)
HAZ, <9.5mm 275
Weld metal 275


The authors acknowledge the assistance of Mike Bennett and Ian Wallis with the sour gas exposure testing.


1 BS4515:1996. Specifications for welding of steel pipelines on land and offshore.
2 Robinson J L: 'A preliminary assessment of the risk of hydrogen cracking in service of external welds on vessels with sour contents'. Report LD22532, May 1981. The Welding Institute.
3 Walker R A: 'The significance of local hard zones on the outside of pipeline girth welds'. Report No.34025/10/89, July 1989, for Welding Supervisory Committee, Pipeline Research Committee. PRC Contract No.PR-164-84.
4 Ginn B J: 'The Significance of local hard zones on the outside of pipeline girth welds - further studies'. Report No.34074.18/91, October 1991 for Welding Supervisory Committee, PRC Contract No.PR-164-008.
5   'Guidelines on material requirements for carbon and low alloy steels for H 2S-containing environments in oil and gas production'. European Federation of Corrosion Publication No.16. Published by The Institute of Materials 1995. ISSN1354-5116.
6 Kushida et al: 'Development of line pipe for sour service and full ring test evaluation' The Sumitomo Search No.49 April 1992.
7 Pargeter R J and Gooch T G: 'Welding C-Mn steels for sour service', Corrosion '95.