2.

11 CRACKING PHENOMENA IN WELDED JOINTS

2.11.1 Cold cracking

2.11.1.1 Introduction to cold (hydrogen-induced) cracking:

Hydrogen-induced cracking, also known as cold cracking, delayed cracking or underbead cracking, is the
most common problem encountered when welding ferritic steel structures. It is not confined to welding,
but can occur in steels during manufacture, during fabrication and in service. The major variables
influencing the incidence of cracking in carbon steels have been defined for many years and the design of
welding procedures in these materials is dominated by the need to incorporate appropriate safeguards
against hydrogen-induced cracking.
A significant number of costly failures of steel structures in the past originated at small pre-existing cracks
in the weld metal or the HAZ of the base metal adjacent to the weld. The most common HAZ failures in
steels are those resulting from the presence of hydrogen in the weld. Hydrogen-induced cracking often
occurs some time after welding and, although extensive, may be difficult to detect. For this reason a heavy
responsibility is placed on the fabricator to match the welding procedure with the material for each
application to prevent cracking.
When hydrogen cracking occurs as a result of welding, the cracks may be situated either in the HAZ of the
base material, or in the weld metal itself.

2.11.1.2 Cracking in the HAZ:

Cracks in the HAZ are usually situated either at the weld toe, the weld root or in an underbead position.
These positions are shown schematically in Figure 2.11.1 for fillet welds and butt welds. In fillet welds, HAZ
cracks are usually oriented along the weld length, but in butt welds subsurface cracks can be transverse to
the weld. Hydrogen cracks examined under a microscope may be transgranular, intergranular or a mixture
with respect to the transformed microstructure. Intergranular cracking is more common in harder, higher
carbon and more highly alloyed steels. Cracks may vary in length from a few microns to several millimetres.

Figure 2.11.1. Hydrogen-induced cracks in the HAZ of (a) fillet, and (b) butt welds.

Hydrogen-induced cracking occurs when the following four requirements are satisfied simultaneously:

• Hydrogen is present to a sufficient degree:

Hydrogen is inevitably present during welding. It is absorbed by the weld pool from the arc atmosphere
during welding. During cooling, much of this hydrogen escapes from the solidified weld by diffusion,
but a certain amount also diffuses into the HAZ and the surrounding base metal. The amount of
hydrogen that diffuses into the HAZ depends on several factors, including the original amount
absorbed, the size of the weld, the decreasing solubility of hydrogen during cooling, and the cooling
rate. In general, the more hydrogen present in the metal, the greater the risk of cracking. Control over
the absorbed hydrogen level may be achieved either by minimising the amount initially absorbed, or by
ensuring that a sufficient amount of hydrogen is allowed to escape by diffusion before the weld cools.
Frequently a combination of both measures provides the best practical solution.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |73
 The principal sources of hydrogen in welding consumables are:
− Moisture in the coating of SMAW electrodes, in the flux used in SAW, or in flux-cored wires.
− Any other hydrogenous compounds in the flux or coating. The moisture in fluxes may be present as absorbed
water, loosely combined water of crystallisation, or more firmly bound molecules or trapped hydroxyl ions in
the silicate structure.
− Oil, dirt and grease, either on the surface or trapped in the surface layers of welding wires.
− Hydrated oxides, such as rust, on the surface of welding wires.

The principle sources of hydrogen from the material to be welded are:

− Oil, grease, dirt, paint, rust, etc. on the surface and adjacent to the weld preparation. These hydrocarbon
compounds can decompose to produce hydrogen in the arc atmosphere.
− Degreasing fluids used to clean surfaces before welding may likewise decompose to produce hydrogen.
− Hydrogen from the base metal, either remaining from the original casting process (particularly in the interior
of heavy sections), following service at high temperature and high hydrogen partial pressures, or as a result
of corrosion processes, particular in sour (i.e. H2S) service.

In addition, a small contribution to the total hydrogen pick-up may arise from moisture in the ambient
atmosphere.
Direct laboratory measurement of the moisture or hydrogen level of any consumable produces a result
which is termed the potential hydrogen level of the process. This term has been selected because not
all the measured hydrogen in the consumable is absorbed by the weld pool. It is potentially available
and, in general, the higher the potential hydrogen level, the higher the actual weld hydrogen content.
Other factors may, however, affect the extent to which potential hydrogen appears as weld hydrogen.
These include resistance heating of the wire in continuously fed wire processes, current type and
polarity for SAW, and the effect of CO2 generation from carbonates in fluxes which reduce the partial
pressure of hydrogen in the arc atmosphere.
Control over the hydrogen potential of welding consumables once they are received from the
manufacturer depends on the conditions under which they are stored and used. Storerooms should be
dry and warm to minimise moisture pick-up by electrodes and fluxes, and care should be taken that the
welding operation does not deposit oil, grease or moisture on the wire, or put moisture into electrode
coatings and fluxes.
Hydrogen potential figures should be used with caution when welding is carried out under hot, humid
conditions. Such situations can lead to weld hydrogen levels somewhat higher than the standardised
conditions under which most consumable manufacturers assess their products.
The hydrogen potential of electrode coverings and fluxes can be lowered by drying or baking, but the
manufacturers' recommendations should always be observed. In most cases, exceeding the
manufacturers' recommended drying temperatures will increase the fragility of the cover. In the case of
basic coverings, over-heating can also lead to the loss of shielding compounds and alloying elements
from the flux coating, resulting in porous weld metal with incorrect composition and properties.

• Tensile stresses act on the weld:

Tensile stresses inevitably arise from thermal contraction during cooling and may be supplemented by
other stresses developed as a result of rigidity in the parts being joined. Stresses developed by thermal
contraction of the cooling weld must be accompanied by strain in the weld metal. In rigid structures the
natural contraction stresses are intensified because of the restraint imposed on the weld by the
different parts of the joint. These stresses are concentrated at the toe and root of the weld and also at
notches constituted by inclusions and other defects. The presence of hydrogen appears to lower the
stress level at which cracking occurs. Stresses applied externally to a weld soon after completion will
often supplement these residual stresses and may temporarily increase the risk of cracking.
The stress acting upon a weld is a function of the weld size, joint geometry, fit-up, external restraint
and the yield strengths of the base metal and weld metal. Root gaps of 0.4 mm and greater in fillet
welds markedly increase the risk of cracking. Stresses are therefore reduced by guaranteeing good fit-

Welding Metallurgy NSW700

Study Theme 2.11 - Page |74
 up and by selecting the lowest strength weld metal allowable by the design. Hydrogen embrittlement is
strain-rate dependent and the risk of cracking is the greatest at slow strain rates. As the strain rate is
low during the final stages of cooling in the weld, the susceptibility to crack formation is high at this
time.

• A susceptible HAZ microstructure is present:

The part of the HAZ which experiences high enough peak temperatures during welding for the base
metal to transform rapidly to austenite, may harden during subsequent cooling as a result of rapid
cooling (quenching) by the surrounding base metal. Hydrogen cracks, when present, are invariably
found in these transformed regions. Close to the fusion boundary, the HAZ is raised to a sufficiently
high temperature to produce a coarse grain size. This high temperature region, because of its coarse
grain size, is not only more hardenable, but also less ductile than regions further away from the fusion
boundary. This is the region where the greatest risk of cracking exists. As a general rule, for both
carbon-manganese and low-alloy steels, the harder the microstructure, the greater the risk of cracking.
Softer microstructures can tolerate more hydrogen before cracking occurs.
The microstructure produced in any steel is essentially dependent upon:
− the cooling rate through the transformation temperature range of the steel in question,
− the composition and hardenability of the steel, and
− the prior austenite grain size before transformation.
The cooling rate is governed by the heat supplied during welding, the initial temperature of the parts to
be joined, their thickness and their geometry. In arc welding, the heat supplied during welding is
characterised by the heat input. Control over the cooling rate in a particular fabrication can therefore
be achieved by varying heat input and preheat temperature.
The hardenability of a steel is governed by its composition, and a useful way of describing hardenability
is by assessing the total contribution of all the elements present in the material. This is done using
empirical carbon equivalent (CE) equations. The widely used CE formula given in equation (2.11.1) is
generally known as the IIW formula.
Mn Cr + Mo + V Ni + Cu
CE = C + + + …(2.11.1)
6 5 15
Another commonly used carbon equivalent equation is the Pcm formula, equation (2.11.2), developed in
Japan for steels of low carbon content, whose behaviour with regard to hydrogen cracking is not well
described by the IIW formula.
Si Mn Cu Ni Cr Mo V
Pcm = C + + + + + + + + 5B …(2.11.2)
30 20 20 60 20 15 10

• Temperature:
The greatest risk of cracking occurs when temperatures near ambient is reached (as shown in Figure
2.11.2) and cracking may therefore take place several hours after welding has been completed.
Cracking is unlikely to occur in structural steels above about 150°C, and in any steel above about 250°C.
It is therefore possible to avoid cracking in a susceptible microstructure by maintaining the component
at a sufficiently high temperature until enough hydrogen has diffused from the weld. The
microstructure can also be softened by tempering to render it less susceptible, and this principle is
employed in multiple-pass welding and when using PWHT (postweld heat treatment).
Provided the HAZ has completed its transformation from austenite, an increase in temperature
increases the rate of diffusion of hydrogen sufficiently to accelerate its removal from the weld. This
effect is particularly marked in the range of 20°C to 150°C, as shown in Figure 2.11.3. Any measure
which slows the weld cooling rate is helpful in reducing the hydrogen level. Preheat, by slowing the
cooling rate, not only softens the microstructure but also helps hydrogen to escape.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |75
 Figure 2.11.2. The notch tensile strength of steel containing hydrogen compared to that of hydrogen-free steel.

Figure 2.11.3. Diffusion rate of hydrogen through ferritic steel as a function of temperature.

For welds in those steels with hardenability so high that soft microstructures cannot be produced at all
and where preheat cannot remove sufficient hydrogen, a weld interpass temperature or a postheat
temperature, high enough to prevent cracking, must be maintained for a sufficiently long time after
welding to allow hydrogen to diffuse away before the weld cools.
If postheating cannot be employed, it is possible, in principle, to soften the uppermost base metal HAZ
with a temper bead. This method should not be used unless careful control can be exercised. The
temper bead is an extra weld run deposited so that its toe is a small fixed distance (typically 3 mm)
from the fusion boundary. The temper bead tempers the uppermost base metal HAZ without creating a
new hardened region.

2.11.1.4 Cracking in the weld metal:

Hydrogen cracking can occur in the weld metal as well as in the HAZ. Weld metal cracks can be orientated
longitudinally or transverse to the weld length, while in the transverse orientation they can be either
perpendicular or angled, typically at approximately 45° to the weld surface (often referred to a chevron
cracks). The cracks may be buried or may break the weld surface. Under a microscope they are
predominantly transgranular. The same factors which influence the risk of cracking in the HAZ also apply to
weld metal, namely stress, hydrogen, susceptible microstructure and temperature. However, weld metal
hydrogen cracking can occur at much lower weld metal hardness levels than is generally the case for HAZ
cracking.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |76
2.11.1.5 Preventing hydrogen cracking during welding:
Graphical methods (welding diagrams or nomograms) can be used to derive welding procedures for
avoiding HAZ cracking in a wide range of steels. Annex C of BS EN 1011-2:2001 “Welding -
Recommendations for welding of metallic materials -Part 2: Arc welding of ferritic steels” proposes three
methods for preventing hydrogen cracking in ferritic steel welds:
• Method A is applicable to plain-carbon, carbon-manganese, low-alloy and fine-grained steels with CE values in the
range of 0.3 to 0.7. It is based on experience and data obtained mainly, but not exclusively, for carbon-manganese
type steels. Method A is described in more detail in §2.11.1.6.
• Method B is applicable to plain-carbon, fine-grained and low-alloy steels, and covers the arc welding of steels in
groups 1 to 4, as specified by ISO 15608 (see Table 2.11.4 for a description of the grouping system according to
ISO 15608). Method B is based on experience and data obtained mainly, but not exclusively, for low-alloy, high-
strength steels. Method B is considered in §2.11.1.7.
• Clause C.4 describes the prevention of hydrogen cracking in creep-resisting and low-temperature steels. The
methods contained in clause C.4 are described in Study Themes 2.13 and 2.14.
Steels with very high hardenability may need to be treated using isothermal transformation methods, or
welded with austenitic or Ni-base consumables. These methods are not covered in detail in BS EN 1011-
2:2001. A brief description of alternative methods for prevention of hydrogen-induced cracking in these
steels is given in §2.11.1.8.

2.11.1.6 Method A proposed by BS EN 1011-2:2001 for preventing hydrogen cracking:

Method A is recommended for welding steels with carbon equivalents in the range of 0.3 to 0.7. The
applicable range of chemical compositions is given in Table 2.11.1.
Table 2.11.1. Applicable composition range for Method A of BS EN 1011-2:2001. (Percentage by mass, balance Fe).
C Si Mn Cr Cu Ni Mo V
0.05 - 0.25 0.8 max 1.7 max 0.9 max 1.0 max 2.5 max 0.75 max 0.2 max

Method A aims to prevent the formation of hard HAZ microstructures. The occurrence of hydrogen cracking
depends on a number of factors, including the composition of the steel (in particular the hardenability), the
welding procedure, the welding consumables and the stress involved. If ∆t8/5 (the cooling time between
800°C and 500°C after welding) is too short, excessive hardening due to martensite formation can occur in
the HAZ. If the hydrogen content of the weld is above a critical level, this hardened zone can crack
spontaneously under the influence of contraction stress after the weld has cooled to near ambient
temperature. The hydrogen content of the weld can be controlled by using hydrogen-controlled welding
processes and consumables. Welding conditions may be selected to avoid cracking by ensuring that the
HAZ cools sufficiently slowly, by controlling the weld dimensions in relation to metal thickness, and if
necessary, by applying preheat and by controlling interpass temperature. Procedures for avoiding hydrogen
cracking, as well as for selecting cooling times through the transformation temperature range to avoid
hardened or susceptible microstructures, may involve controlled cooling in the lower part of the thermal
cycle, typically from 300°C to 100°C, thereby promoting the evolution of hydrogen from the welded joint.
This is normally achieved by postheating on completion of welding.
Similar considerations apply to hydrogen cracking in the weld metal where, although hardening will be on a
reduced scale, actual hydrogen and stress levels are likely to be higher than in the HAZ. In general, welding
procedures selected to avoid HAZ hydrogen cracking will also prevent cracking in the weld metal. However,
under some conditions such as high restraint, low CE steels, thick sections or alloyed weld metal, weld
metal hydrogen cracking can become the dominant mechanism.
• Selecting appropriate values in applying Method A:
In order to select suitable welding procedures for different steel types using graphs and nomograms, it is
necessary to know how to select appropriate values for the carbon equivalent, the combined thickness of
the joint, the weld metal hydrogen level, and the heat input. These values, which vary with each welding

Welding Metallurgy NSW700

Study Theme 2.11 - Page |77
problem, provide points of entry to the nomograms. Guidelines to aid in the selection of the required
values are given below.
− Steel composition: The determination of preheat levels for the prevention of hydrogen cracking is
critically dependent on an accurate knowledge of the parent metal composition and the carbon
equivalent, CE. Where material test certificates are available for the material to be welded, these
should be consulted to determine the CE appropriate to the steel. If only carbon and manganese levels
are stated on the mill sheet, as is common in the case of carbon and carbon-manganese steels, then
0.03 should be added to the calculated CE value to allow for residual elements. Where test certificates
are not available, the specification for the steel will provide data on the maximum CE for a particular
steel grade. In the absence of detail information on the chemical composition of the steel, it may be
necessary to use the specification maxima for the particular steel type. When only a typical
composition is available, the CE value must be calculated with due allowance for likely variations above
the typical composition. Carbon equivalent levels are usually quoted to two decimal places, but the
significance of the second decimal depends on the precision of the original chemical analysis.
− Welding dissimilar steels: When joining dissimilar steels, welding procedures should be devised to
avoid cracking in both steels. This usually involves selecting a procedure to suit the steel with the higher
CE. However, both steels should be examined if one is of higher carbon but lower alloy content than
the other. This is particularly important for alloy steels, or for thick plate, because fully hardened heat-
affected zones are likely to be formed in both. The steel with the higher carbon content will probably
produce the harder HAZ and thus be subject to the greater risk of cracking. In the selection of
consumables for a dissimilar steel joint, it is only necessary to match the strength of the weaker steel.
In fact, it may be advantageous to use the lowest strength weld metal possible to minimise the stresses
in both the HAZ and weld metal.
− Hydrogen potential of the electrode: Measurement of potential hydrogen levels is intended to reveal
the amount of hydrogen that is potentially available for absorption by the weld pool during welding. It
therefore provides a means of characterising the quality of a consumable with respect to hydrogen. The
hydrogen scales specified in Method A are mainly determined by the weld diffusible hydrogen content,
as given in Table 2.11.2. This value should be stated by the consumable manufacturer in accordance
with a relevant standard (or independently determined) in conjunction with a specified condition of
supply or treatment. The consumables should be used in the manner recommended by the supplier,
and should be stored, dried or baked to the appropriate temperature levels and times.
Table 2.11.2. Hydrogen scales according to Method A in BS EN 1011-2:2001.
Diffusible hydrogen content,
Hydrogen scale
ml/100 g of deposited weld metal
> 15 A
10 ≤ 15 B
5 ≤ 10 C
3≤5 D
≤3 E

− Selection of carbon equivalent axis: The risk of cracking increases with increasing hardness of the HAZ
for a particular level of hydrogen and joint restraint. For a given restraint, lowering the weld metal
hydrogen level allows harder heat-affected zoned to be tolerated without cracking, and this forms the
basis of the different hydrogen scales specified in Method A. The following guidelines can be used in
selecting an appropriate hydrogen scale:
→ SMAW: Basic electrodes can be used with scales B to D, depending on the electrode manufacturer’s
classification of the consumable. Rutile or cellulose electrodes should be used with scale A.
→ FCAW or MCAW: Flux cored or metal cored consumables can be used with scale B to scale D, depending on
the manufacturer’s classification of the consumable.
→ SAW: Combinations of SAW wire and flux can have hydrogen levels corresponding to scales B to D, although
most typically these will be scale C. Assessment is needed in the case of each product combination and

Welding Metallurgy NSW700

Study Theme 2.11 - Page |78
 condition. SAW fluxes can be classified by the manufacturer, but this does not necessarily mean that a
practical flux/wire combination will also meet the same classification.
→ GMAW and GTAW: Solid wires for GMAW and GTAW conform to scale D unless specifically assessed and
shown to meet scale E.
→ Scale E may also be appropriate for certain cored wires and some SMAW basic electrodes, but only after
specific assessment. On achieving these low levels of hydrogen, consideration should be given to the
contribution of hydrogen from the shielding gas and from atmospheric humidity.
− Combined thickness of the joint: The combined thickness of the joint is the total thickness (mm) of the
plates meeting at the joint line. It is used to assess the heat sink of the joint for the purpose of
determining the cooling rate, and should be determined as the sum of the base metal thicknesses
averaged over a distance of 75 mm from the weld centreline. If the thickness increases greatly just
beyond 75 mm from the weld line, it may be necessary to use a higher combined thickness value. On
the other hand, if the part thins or terminates just beyond 75 mm, there is a possibility that less
stringent welding conditions can be employed. In both instances a joint simulation test should be
performed to establish safe conditions. For the same metal thickness, the preheating temperature is
higher in a fillet weld than in a butt weld because the combined thickness, and therefore the heat sink,
is greater. Examples illustrating the determination of combined thickness for various joint types are
shown in Figure 2.11.4.
d1 d2 d1
d3 = 0 d3 = 0

75 mm
Combined thickness d2
= d1 + d2 + d3
d3 d3

d1 d2 d1 d2

Combined Special case: For simultaneously

thickness = deposited directly opposed twin
½ (D1 + D2) fillet welds. Combined thickness
= ½(d1 + d2 + d3)
D1 D2 Maximum
diameter
40 mm

Figure 2.11.4. Examples illustrating the calculation of combined thickness.

− Heat input: This term is used to define the heat input to the plate during welding. Heat input, together
with combined thickness and preheat temperature, controls the cooling rate of the weld, the
microstructure, and hence the susceptibility of the HAZ to hydrogen cracking. The heat input is
calculated according to equation (2.11.3):
ηVI
HI = in kJ/mm …(2.11.3)
v
The following arc efficiency factors are specified in clause 19 of BS EN 1011-1:1998:
→ SAW: η = 1.0
→ SMAW: η = 0.8
Welding Metallurgy NSW700
Study Theme 2.11 - Page |79
 → GMAW: η = 0.8
→ FCAW or MCAW: η = 0.8
→ GTAW or PAW: η = 0.6

Electrode run out lengths can be used to estimate heat input during SMAW. These run out lengths are
available in the form of tables, or can be calculated from equation (2.11.4):
d2LF
Run out length (mm) = …(2.11.4)
HI
where: d is the electrode diameter (mm),
L is the consumed length of the electrode (mm) (normally the original length less 40 mm for the stub
end), and
F is a factor in kJ/mm3 with a value dependent on the electrode efficiency:
F = 0.0368 efficiency approximately 95%
F = 0.0408 95% < efficiency ≤ 110%
F = 0.0472 110% < efficiency ≤ 130%
F = 0.0608 efficiency > 130%.

• Applying Method A to prevent hydrogen cracking:

The most effective way of preventing hydrogen cracking is to reduce the hydrogen content of the weld
metal through the use of low hydrogen processes or welding consumables with lower potential hydrogen
levels. Table 2.11.3 shows how filler materials with lower hydrogen levels can extend the combined
thickness range over which steels can be welded with no preheat > 20°C.
Table 2.11.3. Examples of maximum combined thickness weldable without preheat.
Maximum combined thickness
Diffusible hydrogen
Maximum CE of 0.49 Maximum CE of 0.43
content, ml/100 g of
deposited metal Heat input Heat input Heat input Heat input
1.0 kJ/mm 2.0 kJ/mm 1.0 kJ/mm 2.0 kJ/mm
> 15 25 mm 50 mm 40 mm 80 mm
10 ≤ 15 30 mm 55 mm 50 mm 90 mm
5 ≤ 10 35 mm 65 mm 60 mm 100 mm
3≤5 50 mm 100 mm 100 mm 100 mm
≤3 60 mm 100 mm 100 mm 100 mm

Welding conditions for preventing hydrogen cracking in a range of steels are shown in Figures 2.11.5(a) to
2.11.5(m). These conditions are appropriate for avoiding cracking in the weld metal and HAZ in the majority
of welds. The following should be considered:
− Preheating: The recommended preheat temperature can be obtained from Figures 2.11.5(a) to 2.11.5(m) by
reading the preheat temperature line immediately above or to the left of the point corresponding to the
appropriate heat input and combined thickness. If local preheat is used, it must be applied along the weld line, so
that the temperature on both sides of the joint, measured at least 75 mm from the weld centreline on the
opposite side of the plate to that being heated, is equal to the recommended preheat temperature. If the
temperature can be measured only on the heated side, the heat source should be removed and sufficient time (1
minute per 25 mm thickness) allowed for temperature equalisation before measurement. If general preheat is
used, a lower preheat temperature may be adequate, but this should be confirmed by joint simulation tests.
− Interpass temperature: The preheat temperature is often specified as the minimum recommended interpass
temperature for multiple-pass welds. A lower interpass temperature may be permitted where subsequent runs
are made with higher heat inputs than the root run. In these cases, the interpass temperature should be
determined from Figures 2.11.5(a) to (m) for the larger runs.
− Hydrogen reduction by postweld heat treatment: When a high risk of hydrogen cracking exists, hydrogen release
can be accelerated by maintaining the minimum interpass temperature, or by raising the temperature to 200°C to
300°C immediately after welding, before the weld region cools to below the minimum interpass temperature. The
duration of this postweld heat treatment should be at least 2 hours, and is a function of the thickness. Large
thicknesses require temperatures at the upper end of the stated range, as well as prolonged postheating times.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |80
 Postheating is also appropriate where a partially filled weld cross-section is to be cooled. The width of the band
heated to the postweld temperature during localised heat treatment is not as important as ensuring that the weld
and HAZ do not fall below the specified temperature.
− Misalignment: Misalignment between two butt welded plates gives rise to a higher stress concentration. The root
edges or root faces of butt joints should not be out of alignment by more than 25% of the thickness of the thinner
material (for material up to 12 mm in thickness), or by more than 3 mm for material thicker than 12 mm.
− Fit-up: Edges and surfaces to be joined by a fillet weld should be in close contact. The gap should generally not
exceed 3 mm.

Figure 2.11.5(a). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.30 0.34 0.38 0.44 0.46

Figure 2.11.5(b). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.34 0.39 0.41 0.46 0.48

Figure 2.11.5(c). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.38 0.41 0.43 0.48 0.50

Welding Metallurgy NSW700

Study Theme 2.11 - Page |81
Figure 2.11.5(d). Conditions for welding steels
with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.41 0.43 0.45 0.50 0.52

Figure 2.11.5(e). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.43 0.45 0.47 0.53 0.55

Figure 2.11.5(f). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.45 0.47 0.49 0.55 0.57

Welding Metallurgy NSW700

Study Theme 2.11 - Page |82
Figure 2.11.5(g). Conditions for welding steels
with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.47 0.49 0.51 0.58 0.60

Figure 2.11.5(h). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.49 0.51 0.53 0.60 0.62

Figure 2.11.5(i). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.51 0.53 0.55 0.62 0.64

Welding Metallurgy NSW700

Study Theme 2.11 - Page |83
Figure 2.11.5(j). Conditions for welding steels
with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.53 0.55 0.57 0.64 0.66

Figure 2.11.5(k). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE 0.55 0.57 0.59 0.66 0.68

Figure 2.11.5(l). Conditions for welding steels

with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE - - 0.60 0.68 0.70

Welding Metallurgy NSW700

Study Theme 2.11 - Page |84
 Figure 2.11.5(m). Conditions for welding steels
with defined carbon equivalents.

Hydrogen scale to be used for a CE not

exceeding:

Scale A B C D E
CE - - 0.62 0.70 -

• Conditions requiring more stringent procedures:

The preheating conditions presented in Figures 2.11.5(a) to (m) have been found to provide a satisfactory
basis for deriving safe welding procedures for many welded fabrications. However, the risk of hydrogen
cracking is influenced by several parameters and these can sometimes exert an adverse influence greater
than that accounted for in Figures 2.11.5(a) to (m). Although modifications to the procedures to deal with
such welds can, in principle, be obtained through a change in heat input or preheat, the most effective
modification is to lower the weld hydrogen level. This can be done either directly, by lowering the hydrogen
input to the weld (through the use of lower hydrogen welding processes or consumables), or by
accelerating hydrogen losses from the weld through the use of higher postheat for a period of time after
welding. The required postheat time will depend on many factors, but a period of 2 to 3 hours has been
found to be beneficial in many instances. It is recommended that the required modifications to the
procedures be derived by the use of adequate joint simulation testing. A number of conditions which may
require more stringent procedures are listed below:
− Joint restraint is a complex function of section thickness, weld preparation, joint geometry and the stiffness of the
fabrication. High restraint conditions, such as welds made in section thicknesses above approximately 50 mm and
root runs in double bevel joints, require additional precautions against hydrogen cracking. These can take the
form of increased preheat levels (25 to 75°C higher than preheat temperatures determined from Figures 2.11.5(a)
to (m)), or the use of a hydrogen scale corresponding to a higher weld hydrogen level.
− The use of higher strength alloyed weld metal or carbon-manganese weld metal with a manganese content above
approximately 1.5% can lead to higher operative stresses. Whether or not this causes an increased risk of HAZ
cracking, the weld deposit itself would generally be harder and more susceptible to cracking.
− Lowering the inclusion content of the steel, principally by lowering the sulphur content (but also the oxygen
content) can increase the hardenability of the steel. This can result in an increase in the hardness of the HAZ and
possibly a small increase in the risk of HAZ hydrogen cracking. Very high sulphur levels are also detrimental. Free-
cutting or free-machining grades of steel normally contain up to 0.2% of sulphur, although up to 0.5% is possible.
Other elements such as lead or selenium may also be present. In welding these steels special precautions are
necessary. Reducing the sulphur level in the steel generally increases both the HAZ hardenability and the risk of
cracking. This phenomenon appears to be linked to the number of inclusions, rather than the actual sulphur
content. The change in procedure required from a low sulphur steel to one of higher sulphur content of the same
CE, may be as significant as an increase in CE of 0.03, or an increase in preheat temperature of 50°C to 100°C,
depending on the steel and the welding circumstances.

• Conditions which may allow relaxation of welding procedures:

− General preheating: If the whole component is preheated, it is generally possible to reduce the preheat
temperature by a limited amount.
− Limited heat sink: If the heat sink is limited in one or more directions (e.g. when the shortest heat path is less than
10 times the fillet leg length) especially in the thicker plate (e.g. in the case of a lap joint where the outstand is
only marginally greater than the fillet weld leg length), it is possible to reduce preheat levels.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |85
− Austenitic consumables: If austenitic consumables are used, preheat is not always necessary, especially if the
condition of the consumable is such that it deposits weld metal containing very low levels of hydrogen.
− Joint fit-up: Close fit fillet welds (where the gap is 0.5 mm or less) may justify relaxation of the welding procedure.

2.11.1.7 Method B proposed by BS EN 1011-2:2001 for preventing hydrogen cracking:

Method B covers the arc welding of steels in groups 1 to 4, as specified by ISO 15608. A summary of the
steels covered in groups 1 to 4 in ISO 15608 is given in Table 2.11.4.
Table 2.11.4. Steels covered by groups 1 to 4 in ISO 15608.

Group Sub-group Type of steel

1 Steels with a specified minimum yield strength ≤ 460 MPa and with the following analysis:
C ≤ 0.25%
Si ≤ 0.60%
Mn ≤ 1.70%
Mo ≤ 0.70% (A higher value may be accepted provided that Cr+Mo+Ni+Cu+V ≤ 0.75%)
S ≤ 0.045%
P ≤ 0.045%
Cu ≤ 0.40% (A higher value may be accepted provided that Cr+Mo+Ni+Cu+V ≤ 0.75%)
Ni ≤ 0.5% (A higher value may be accepted provided that Cr+Mo+Ni+Cu+V ≤ 0.75%)
Cr ≤ 0.3% (A higher value may be accepted provided that Cr+Mo+Ni+Cu+V ≤ 0.75%)
Nb ≤ 0.05%
V ≤ 0.12% (A higher value may be accepted provided that Cr+Mo+Ni+Cu+V ≤ 0.75%)
Ti ≤ 0.05%
1.1 Steels with a specified minimum yield strength ≤ 275 MPa
1.2 Steels with a specified minimum yield strength > 275 MPa and ≤ 360 MPa
1.3 Normalised fine-grained steels with a specified minimum yield strength > 360 MPa
Steels with improved atmospheric corrosion resistance whose analysis may exceed the
1.4
requirements for the single elements as indicated above under 1.
2 Thermomechanically treated fine-grained steels and cast steels with a specified minimum
yield strength > 360 MPa
Thermomechanically treated fine-grained steels and cast steels with a specified minimum
2.1
yield strength > 360 MPa and ≤ 460 MPa
Thermomechanically treated fine-grained steels and cast steels with a specified minimum
2.2
yield strength > 460 MPa
3 Quenched and tempered steels and precipitation-hardened steels (except stainless steels)
with a specified minimum yield strength > 360 MPa
Quenched and tempered steels with a specified minimum yield strength > 360 MPa and ≤
3.1
690 MPa
3.2 Quenched and tempered steels with a specified minimum yield strength > 690 MPa
3.3 Precipitation-hardened steels except stainless steels
4 Low-vanadium alloyed Cr-Mo-(Ni) steels with Mo ≤ 0.7% and V ≤ 0.1%
4.1 Steels with Cr ≤ 0.3% and Ni ≤ 0.7%
4.2 Steels with Cr ≤ 0.7% and Ni ≤ 1.5%

Method B is based on the observation that cold cracking can often be avoided by preheating the weld to
higher temperatures to delay cooling and thereby promote hydrogen diffusion from the weld in a shorter
time and to a greater extent than welding without preheat. Preheating furthermore may reduce the state
of internal residual stress. For multiple-pass welds it is possible to start without preheating if a sufficiently
high interpass temperature is maintained by a suitable weld sequence. The lowest temperature before
welding the first run and below which the weld should not fall during welding, is designated the preheat

Welding Metallurgy NSW700

Study Theme 2.11 - Page |86
temperature, Tp. In the case of multiple-pass welding, the term used for this temperature in reference to
the second and all ensuing runs, is the minimum interpass temperature, Ti. Both temperatures are
generally the same.
The susceptibility of welded joints to hydrogen cracking is influenced by the chemical composition of the
parent plate and the weld metal, the plate thickness, the hydrogen content of the weld metal, the heat
input during welding and the stress level. An increase in alloy content, plate thickness and hydrogen
content increases the risk of cracking, whereas an increase in heat input renders the weld more resistant to
hydrogen cracking. Each of these factors is considered in more detail below.

• Base material:
The influence of chemical composition on the hydrogen cracking behaviour of steels is characterised by
means of a carbon equivalent, CET, calculated according to equation (2.11.5). Equation (2.11.5) applies to
the range of chemical compositions shown in Table 2.11.5. Note that the contribution of the alloying
elements relative to that of carbon is different from the IIW CE equation. It is therefore not advisable to
covert CET values into CE values, or vice versa.
Mn + Mo Cr + Cu Ni
CET = C + + + …(2.11.5)
10 20 40

Table 2.11.5. Applicable composition range for Method B of BS EN 1011-2:2001. (Percentage by mass, balance Fe).
C Si Mn Cr Cu Ni Mo V Nb Ti B
0.05 – 0.32 0.8 max 0.5 – 1.9 1.5 max 0.7 max 2.5 max 0.75 max 0.18 max 0.06 max 0.12 max 0.005 max

A linear relationship exists between the carbon equivalent, CET, and the preheat temperature, Tp (or the
interpass temperature, Ti), as shown in Figure 2.11.6. An increase of approximately 0.01% in the carbon
equivalent, CET, leads to an increase of about 7.5°C in the preheat temperature. This can also be presented
algebraically in the form of equation (2.11.6).
TpCET = 750 x CET –150 (°C) …(2.11.6)

Figure 2.11.6. Preheat temperature as a function of carbon equivalent, CET.

• Plate thickness:
The relationship between plate thickness, d, and preheat temperature, Tp, is shown in Figure 2.11.7. For
thinner material, a change in plate thickness results in a greater change in preheat temperature. With
increasing plate thickness, however, this effect is reduced, and above 60 mm the change in preheat

Welding Metallurgy NSW700

Study Theme 2.11 - Page |87
temperature with thickness is minor. The influence of plate thickness on preheat temperature can also be
presented algebraically in the form of equation (2.11.7).
Tpd = 160 tanh (d/35) – 110 (°C) …(2.11.7)

Figure 2.11.7. Preheat temperature as a function of plate thickness, d.

• Hydrogen content:
The effect of the weld metal hydrogen content, HD, on the recommended preheat temperature is shown in
Figure 2.11.8. An increase in hydrogen content necessitates the use of a higher preheat temperature. A
change in hydrogen content has a greater effect on preheat at lower hydrogen concentrations than at
higher concentrations. The influence of weld hydrogen content on the required preheat temperature can
also be presented algebraically in the form of equation (2.11.8).
TpHD = 62 X HD 0.35 –100 (°C) …(2.11.8)

Figure 2.11.8. Preheat temperature as a function of weld hydrogen content, HD.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |88
• Heat input:
The influence of heat input, HI, on the preheat temperature is shown in Figure 2.11.9. An increase in heat
input during welding permits a reduction in preheat temperature. Furthermore, the influence of heat input
on preheat is dependent on alloy content, and is more pronounced for low carbon equivalent welds than
for higher carbon equivalents. The influence of weld heat input on the required preheat temperature can
also be presented algebraically in the form of equation (2.11.9).
TpHI = (53 x CET – 32) x HI – 53 x CET + 32 (°C) …(2.11.9)

Figure 2.11.9. Preheat temperature as a function of heat input, HI.

• Calculation of the preheat temperature:

The effects of chemical composition, as characterised by CET, the plate thickness, d, the hydrogen content
of the weld metal, HD, and the heat input, HI, can be combined in equation (2.11.10) to calculate the
preheat temperature, Tp:
Tp = TpCET + Tpd + TpHD + TpHI (°C) …(2.11.10)

The preheat temperature can also be calculated according to equation (2.11.11):

Tp = 697 x CET + 160 x tanh (d/35) + 62 x HD 0.35 + (53 x CET –32) x HI – 32 (°C) …(2.11.11)

This relationship is valid for structural steels with yield strengths up to 1000 MPa and:
CET = 0.2% to 0.5%
d = 10 mm to 90 mm
HD = 1 ml/100g to 20 ml/100g
HI = 0.5 kJ/mm to 4.0 kJ/mm

The preheat temperatures calculated with the aid of equations (2.11.10) or (2.11.11) apply, provided the
following conditions are fulfilled:
• The carbon equivalent, CET, of the base metal exceeds that of the weld metal by at least 0.03%. Otherwise,
calculation of the preheat temperature should be based on the CET of the weld metal increased by 0.03%.
• Single-pass fillet, tack and root welds have a minimum length of 50 mm. If the plate thickness exceeds 25 mm,
tack and root passes are deposited in two layers using a mild ductile weld metal.
• In the case of filler pass welding, which also includes multiple-pass fillet welds, no interpass cooling takes place as
long as the weld thickness has not obtained one third of the plate thickness. Otherwise, the hydrogen content of
the weld needs to be reduced by means of a postweld heat treatment.
• The welding sequence shall be selected in such a way that the excessive plastic deformation inherent to partially
filled fillet welds is avoided.
Welding Metallurgy NSW700
Study Theme 2.11 - Page |89
Where there is an increased risk of hydrogen cracking, e.g. when steels with yield strengths of more than
460 MPa and in thicknesses greater than 30 mm are submerged arc welded, it is advisable to reduce the
hydrogen content of the weld by means of soaking, e.g. 2 hours at 250°C, immediately after welding.
In deriving equation (2.11.11) it was assumed that the residual stresses in the weld are equal to the yield
strength of the base metal and the weld metal respectively.

2.11.1.8 High hardenability steels not sufficiently covered by BS EN 1011-2:2001:

High hardenability steels form hard HAZ microstructures even at the slowest cooling rates achieved in
welding. In some instances the methods described in BS EN 1011-2:2001 are unsuitable (for example when
adequate preheat is impractical), or do not provide sufficient resistance to hydrogen cracking. An
isothermal heat treatment after welding, or the use of austenitic welding consumables, may then be
advisable to prevent hydrogen cracking.

• Isothermal transformation method:

The isothermal transformation method is particularly suitable for steels of higher carbon content showing
high hardenability, where it is necessary to produce a softer HAZ microstructure directly after welding and
without recourse to postweld tempering.
The isothermal transformation method relies on knowledge of the isothermal transformation
characteristics of the steel, as represented in diagrams of the type shown schematically in Figure 2.11.10 (IT
or TTT diagrams). The object is to control the cooling of the HAZ so that it transforms in an approximately
isothermal manner and produces a softer, less crack-sensitive microstructure than martensite. A
temperature is selected which promotes transformation, usually to lower bainite, in a reasonable time and
over a temperature range that can be controlled in practice. Unless the transformation diagram has been
assembled from data involving austenitising temperatures higher than 1250°C (which produce coarse grain
sizes), the minimum holding time after welding should be at least double that indicated on the diagram for
100% transformation. In Figure 2.11.10 the arrow superimposed on the transformation diagram indicates
the particular isothermal treatment to be followed to produce a softer HAZ, and shows that a minimum
preheat, interpass and postheat temperature of 360°C±20°C should be specified. Isothermal
transformation characteristics may be obtained from the steel maker's data sheets or from one of the
collections of such data (Woolman, J. and Mottram, R.A. “The mechanical and physical properties of the
British Standard EN steels”, vol. 1, 2 and 3. BISRA, London; “Atlas of isothermal transformation diagrams",
US Steel Corporation, Pittsburgh, USA; Vander Voort, G.F. (ed.) “Atlas of time-temperature diagrams for
irons and steels”. ASM, 1991.)

Figure 2.11.10. Schematic IT-diagram of a steel. The diagram can be used to select heating times and temperatures
after welding so that a relatively soft microstructure is obtained.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |90
It must be noted that, although the microstructures produced in this way will be softer and tougher than
martensite, they will almost certainly be harder and less tough that the tempered martensitic structures
produced by the temperature control methods (described in §2.11.1.6 to §2.11.1.7) in hardenable steels.
For this reason, this method is usually adopted when tempering cannot be performed. The principal
advantage of the isothermal transformation method over the temperature control method lies in the
greater ease in deciding how to control the transformation compared to the difficulty of deciding how
much hydrogen needs to be removed.

• The use of austenitic and nickel alloy weld metal:

Where, for various reasons, it is not possible to use preheat temperatures greater than 150°C, the
temperature control methods are severely restricted and the isothermal transformation method cannot be
used. The only alternative is to use a combination of welding process and consumable that prevents the
introduction of hydrogen into the HAZ and produces weld metal that is insensitive to hydrogen. This is
achieved by the use of austenitic stainless steel or nickel-based electrodes.
At ambient temperature, both austenitic stainless steel and nickel alloy weld metals have much higher
solubilities for hydrogen, much slower hydrogen diffusion rates and very low sensitivities to hydrogen
embrittlement and cracking in comparison to ferritic weld metals. The higher solubility means that once
hydrogen has diffused from the HAZ to the fusion boundary, it can easily enter weld metal that has a high
solubility for hydrogen. As a result of the slow diffusion rate in the weld metal, any hydrogen that reaches
the weld metal close to the fusion boundary will stay in its vicinity. It is, therefore, advantageous to use
consumables that result in low weld metal hydrogen contents to reduce the risk of local saturation of the
weld metal near the fusion boundary.
In assessing the hydrogen levels of such consumables it is the total (not the diffusible) weld metal hydrogen
content that is important and should be measured. This is because hydrogen analysis of austenitic and
nickel weld metals at ambient temperature measures very small amounts of diffusible hydrogen, as
diffusion rates in these materials are so slow that virtually all the hydrogen present would be measured as
residual, even though it still active and diffusing.
When selecting austenitic stainless steel or nickel alloy fillers, it is necessary to ensure that dilution from
the base metal can be satisfactorily accommodated. The consumables normally selected for SMAW are
types 316L (20Cr-9Ni-3Mo), 309L (23Cr-12Ni) or 312 (29Cr-9Ni). Type 309L is most commonly used and
produces deposits containing sufficient ferrite to suppress solidification (hot) cracking, with little or no
martensite in the bulk deposit. Type 312 is preferred for high dilution runs to avoid fully austenitic deposits.
However, in low dilution situations, the weld metal will contain a high ferrite level (up to 35%). Although
this is beneficial in tolerating the pick-up of sulphur from the parent metal, and also in giving a high yield
strength, this type of high ferrite weld deposit should not be subjected to postweld heat treatment, since it
will show marked embrittlement as a result of the formation of sigma phase during heat treatment.
Nickel alloy fillers have the advantage of lower coefficients of thermal expansion than stainless steels. This
can reduce shrinkage strains and reduce the risk of cracking in highly restrained joints. They are, however,
more sensitive to solidification cracking than stainless steels. If postweld heat treatment is required, a
nickel alloy filler may well become the preferred choice to avoid intermetallic formation and embrittlement
during heat treatment.
When austenitic electrodes are used, preheat is not normally required for steels containing up to 0.2%
carbon, although they are liable to form hard regions of crack-sensitive alloyed martensite at the fusion
boundary. Here incomplete mixing often leads to small regions alloyed sufficiently to transform to hard
martensite on cooling, but not sufficiently alloyed to remain austenitic. Such regions are prone to hydrogen
cracking and very difficult to detect. At 0.4% carbon and above, a minimum preheat temperature of 150°C
is required to prevent embrittlement and normal HAZ cracking. Buttering the surfaces to be welded may
reduce the level of preheat required.
Normal HAZ cracking is also possible, particularly where the use of large electrodes (more common with
the stainless steel types) results in wide heat-affected zones. These give long diffusion distances for any
hydrogen that has diffused into the HAZ when it was austenitic at high temperatures (or may even be

Welding Metallurgy NSW700

Study Theme 2.11 - Page |91
present in the parent steel), to diffuse back into the weld metal. Whether in the fusion boundary or in the
HAZ, cracking can usually be prevented by applying some preheat (normally about 150°C) to increase the
HAZ cooling time and to allow longer time for hydrogen to diffuse out of the HAZ while it is still too warm to
be embrittled by hydrogen. In all cases, freedom from cracking will be more easily achieved if processes
and consumables giving low total weld metal hydrogen levels are used. High heat input is often helpful, and
techniques that give low dilution should be employed to minimise the formation of martensite in the weld
metal. Hard heat-affected zones will normally be produced and it is usually advantageous to temper after
welding. Although postweld heat treatment tempers the HAZ, it is usually ineffective in giving a high degree
of stress relief, because of the difference in thermal expansion coefficients between the austenitic weld
metal and the ferritic parent plate. In this respect, nickel alloys are likely to be advantageous.

2.11.2 Hot cracking:

2.11.2.1 Solidification cracking:

Solidification cracking occurs in weld deposits during cooling of the weld metal. Examples of solidification
cracks are shown in Figures 2.11.11 and 2.11.12. These cracks occur preferentially at the weld centreline or
between columnar grains along grain boundaries. Solidification cracking is often associated with crater or
end cracking, as illustrated in Figure 2.11.13. Cracking is typically observed to occur at temperatures of
about 200°C to 300°C below the melting temperature.

Figure 2.11.11. Examples of solidification cracks at columnar grain boundaries.

(a) (b)
Figure 2.11.12. Centreline cracking in welds: (a) Macrograph of a steel weld exhibiting a centerline crack; and (b)
schematic illustration of the mechanism of centerline cracking in steel welds.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |92
 Figure 2.11.13. Examples of end cracking in butt welds associated with high restraint joints.

The susceptibility of weld metal to solidification cracking appears to depend on three factors:
− the coarseness of the solidification microstructure,
− the amount and species of segregation, and
− the geometry of the joint.

• Solidification structure:
As a result of epitaxial solidification, the coarseness of the weld microstructure is inherited from the grain
growth zone of the HAZ. High energy welds, such as submerged arc welds, therefore generally demonstrate
the greatest amount of grain growth and hence the coarsest weld metal microstructures.
Solidification patterns are also affected by the welding speed. Low speeds tend to allow the columnar
grains to follow the arc, curving in behind the moving heat source. This has a grain refining effect, because
new grains have to nucleate in order to maintain growth along the preferential crystallographic growth
directions. High speed welds, on the other hand, tend to produce solidification patterns in which the
columnar crystals grow in parallel, straight rows to the weld centerline (see Study Theme 2.7). This may
result in sudden and abrupt changes in the growth direction at the centreline as the final part of the weld
that solidifies attempts to keep up with the moving arc. In general, a long, straight-sided columnar grain
structure tends to be weaker under stress than the more equiaxed, finer grain structure of low travel speed
welds.
The onset of cellular solidification due to constitutional supercooling has only a minor effect on cracking
susceptibility in that it tends to reduce segregation at the cell boundaries. The coarser the cell structure,
the higher the segregation tendency. High speed welding tends to produce finer cell spacings than slower
welding speeds.
The cellular to dendritic transition, associated with the final stages of solidification at the centreline of high
energy welds, is likely to affect cracking susceptibility. This is because dendritic growth in weld metals
occurs due to high levels of constitutional supercooling and relatively slow cooling rates, leading to high
degrees of segregation. The problem is greatest at end-craters where constitutional supercooling is at its
highest.

• Segregation:
Segregation cannot be prevented during the solidification of alloyed weld metal. The amount of
segregation, however, can be controlled to some extent. Segregation occurs due to partitioning of
elements during the initial stages of solidification. The extent of segregation at this stage depends on the
partitioning coefficient, k, defined by equation (2.11.12):
Xs
k= …(2.11.12)
X
where: Xs is the mole fraction of solute in the solid, and
X is the mole fraction of solute in the liquid at a given temperature as defined by the phase diagram.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |93
Different alloying combinations therefore have different k values, the greatest partitioning occurring for the
smallest values of k. In the case of steels, for example, any alloying constituent in the weld deposit that
exhibits a wide solidification range with iron is likely to have a low k value. Some approximate values of k
for iron alloys, as determined from their binary equilibrium solubilities, are given in Table 2.11.9. According
to this table, the elements most likely to segregate in steel are S, O, B, P, C, Ti, N and H, in that order. Of
these elements, sulphur is often considered to be the most dangerous because it readily combines with Fe
and Mn to form (MnFe)S, a compound with a low melting point which easily spreads along grains
boundaries.
Table 2.11.9. Estimations of partitioning coefficients of elements in δ-iron.
Element Al B C Cr Co Cu H Mn Mo
k 0.92 0.05 0.13 0.95 0.90 0.56 0.32 0.84 0.80
Element Ni N O P Si S Ti W V
k 0.80 0.28 0.02 0.13 0.66 0.02 0.14 0.95 0.90

In carbon steels, the amount of segregation is also a function of the carbon content of the alloy. Figure
2.11.14 shows the upper left-hand corner of the Fe-Fe3C phase diagram. When the carbon content is below
0.1%, the metal solidifies as primary δ-ferrite. At higher carbon contents, the primary crystals are δ-ferrite,
but just below 1500°C, a peritectic reaction occurs and the remainder of the weld solidifies as austenite. As
shown in Figure 2.11.15, the maximum solubility of sulphur in δ-ferrite is relatively high (0.18%), but in
austenite it is significantly lower (0.05%). Consequently, there is a possibility at carbon contents higher than
0.1% that sulphur will be rejected to the grain boundaries of primary austenite grains, promoting
intergranular weakness and solidification cracking. Figure 2.11.16 shows the effect of sulphur and carbon
on the brittle temperature range of carbon steels. This diagram illustrates the cooperative action of
increased sulphur and carbon contents in promoting solidification cracking. As shown in Figure 2.11.16,
phosphorus has much the same effect on the brittle temperature range, and hence on solidification
cracking. The formation of a phosphide eutectic is unlikely, but phosphorus segregates to the grain
boundaries and could act either by lowering the melting point of the interdendritic regions or by reducing
grain boundary cohesion.

Figure 2.11.14. A section of the Fe-Fe3C equilibrium phase diagram showing the peritectic reaction.

The beneficial effect of manganese is shown in Figure 2.11.17, and the influence of other elements in Figure
2.11.18. Boron increases susceptibility in the same way as phosphorus; nickel, however, acts in the same
way as carbon in promoting the formation of austenite as a primary structure. Because of a high risk of
solidification cracking, SMAW of carbon steels containing more than 4% Ni is not normally possible, but
GMAW with 9% Ni wire is practicable.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |94
 Figure 2.11.15. Part of the Fe-S binary equilibrium diagram.

P(%) S(%)
 0.010 0.004
 0.020 0.004
 0.010 0.020

Figure 2.11.16. The brittle temperature range of S and P containing steels as a function of carbon content.

Figure 2.11.17. The effect of Mn:S ratio and C content on the susceptibility of carbon steel weld metal to hot cracking.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |95
 Figure 2.11.18. The relationship between weld metal alloying element content and hot cracking susceptibility,
measured as a percentage of the weld run cracked, for a single V-groove with a 60° joint angle.

• Residual stresses and joint geometry:

The thermal cycle of the welding process always results in residual stresses in the weld after cooling. The
severity of these stresses depends on the degree of restraint offered by the welded joint. As a general rule,
the thicker or stronger the plates being welded, the higher the level of restraint and the greater the
residual stress. At its maximum level, the tensile stresses at the weld may reach the yield point of the
metal. These stresses are usually greatest at the end sections of welded plates.

• Mechanism of solidification cracking:

The causes of solidification cracking are well understood. The partitioning and rejection of alloying
elements at columnar and dendritic grain boundaries and ahead of the advancing solid-liquid interface
cause considerable segregation, as illustrated in Figure 2.11.19. The segregated elements form low melting
phases or eutectics with the metal to produce highly wetting films at grain boundaries. These films weaken
the structure to the extent that cracks form at the grain boundaries under the influence of residual tensile
stresses that develop during cooling.

Shrinkage strain

Direction of
solidification

Partially liquid region

Figure 2.11.19. Schematic illustration of the mechanism of solidification cracking.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |96
Any alloying process that reduces the effect of segregation is advantageous. For example, adding Mn where
possible to form MnS instead of FeS is usually beneficial (as shown in Figure 2.11.17). This is because FeS
has a much lower wetting angle than MnS, so that inclusions, rather than grain boundary films, form. FeS
and its eutectics also solidify at lower temperatures than their manganese equivalents. FeS solidifies at
1190°C, FeS-FeO eutectics solidify at 940°C, MnS solidifies at 1600°C and MnS-MnO eutectics solidify at
approximately 1300°C. In some cases, higher oxygen contents in weld deposits may also be advantageous
in that oxysulphide inclusions are formed, instead of low melting grain boundary films.
The mechanism by which solidification cracking occurs is one in which cracks nucleate at the sulphide-
matrix interfaces and spread along the grain boundaries under the influence of the residual tensile stresses.
The criterion as to whether cracks nucleate within the compound or at the compound-matrix interface is
determined by the relative fracture strengths of the compound and the matrix (unless the phase is still
liquid).
The risk of solidification cracking can be minimised by:
− Maintaining a low carbon content in the weld deposit.
− Keeping sulphur and phosphorus contents as low as possible.
− Ensuring that the manganese content is high enough to allow for possible dilution (and ingress of sulphur) from
the plate material.

2.11.2.2 Liquation cracking:

An example of liquation cracking is shown in Figure 2.11.20. The causes of this type of cracking are fairly
well understood and are associated with grain boundary segregation aggravated by localized melting of
grain boundaries near the fusion line. High residual stresses that develop as the weld cools then tend to
rupture these impurity-weakened boundaries. The influence of the solidification sequence on liquation
cracking is illustrated in Figure 2.11.21 and the fracture mechanism is shown schematically in Figure
2.11.22. While the weld deposit is still liquid, compressive stresses tend to close liquation cracks, but as the
weld metal solidifies, tensile stresses develop in the HAZ that tend to open up the cracks. Note that the
origin of liquation cracking is quite different from that of solidification cracking. Localised melting of grain
boundaries occurs at the fusion line at temperatures between the solidus and liquidus. Since melting
nucleates preferentially at high energy crystal defects, such as surfaces and grain boundaries, there is a
gradual increase in melted boundary width up to the melt zone. Impurities with low solubility in the matrix
or melted back inclusions near the fusion line tend to introduce impurity elements that diffuse to the
melted boundaries. The impurities most likely to segregate under these conditions are S, C, B, P, N, Sb and
Sn.

Weld

Figure 2.11.20. Example of liquation cracking or hot tearing in the heat-affected zone of a shielded metal arc weld.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |97
 Figure 2.11.21. Schematic illustration of the formation of hot cracks and liquation cracks in weld metal due to the
formation of low melting phases during solidification; ε refers to deformation by shrinkage and d refers to the crystal
growth direction.

Figure 2.11.22. Schematic illustration of the mechanism of liquation cracking.

On being cooled, the segregated elements tend to form films of low melting point grain boundary
compounds or even low melting point eutectics. The criteria that govern liquation cracking are therefore
very similar to those discussed in the previous section concerning solidification cracking.
Liquation cracking is dependent on the amount and type of impurities in the base metal, the volume
fraction and density of inclusions, and the degree of restraint. The latter effect is in turn dependent on the
strength and thickness of the plates being welded. Another factor of importance concerns the dwell time at
high temperatures; high energy welding processes thus increase susceptibility to this problem.

2.11.3 Lamellar tearing:

Lamellar tearing is a form of cracking that occurs in the base metal of a weld due to the combination of high
localised stress and low ductility of the plate in the through-thickness direction. An example of lamellar
tearing is shown in Figure 2.11.23, in which it is evident that the crack appears to be closely associated with
the edge of the HAZ. The horizontal and vertical step-like cracking of the base metal is a very typical feature
of lamellar tearing. The problem occurs particularly when making T- and corner joints in thick plates, where
the fusion boundary of the weld is more or less parallel to the plate surface (as shown in Figure 2.11.24).
The cracks appear close to or a few millimetres away from the fusion boundary, and usually consist of
planar areas parallel to the surface joined by shear failures at right angles to the surface.
The susceptibility to lamellar tearing depends on the type of joint and the inherent restraint, on the base
metal sulphur and oxygen contents, on the type and morphology of inclusions (which affect the through-
thickness ductility) and on the hydrogen content of the weld. Lamellar tears initiate by separation or void
formation at the interface between inclusions and the metal matrix, or by shattering of the inclusion itself.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |98
The voids so formed link together in a planar manner by necking, microvoid coalescence, or cleavage.
Subsequently, these planar discontinuities are joined by vertical shear walls.

(a) (b)
Figure 2.11.23. (a) An example of lamellar tearing adjacent to a T-butt weld in a structural steel. (b) The stepped
nature of lamellar tearing in relation to the banded microstructure of the plate.

Figure 2.11.24. Typical assemblies in which lamellar tearing occurs.

Silicate and sulphide inclusions both play a part in initiating lamellar tearing. At its simplest, lamellar tearing
can occur as a result of a very low through-thickness or short transverse ductility owing to the presence of
elongated (MnFe)S inclusions (shown in Figure 2.11.25). These arise in the steel processing stage, beginning
with segregation bands during solidification of the ingots, followed by rolling and spreading of non-metallic
inclusions. The latter can be in the form of large (several hundred microns) (FeMn)S particles or long
stringers of oxide or silicates. Under the action of weld residual stresses, particularly in high restraint
geometries, the inclusion-matrix interface ruptures in a number of places, with the last stage of fracture
causing vertical tearing between planes. This gives the crack its characteristic step-like appearance.
Another feature of lamellar tearing is that, in some cases, the phenomenon may occur after a certain time
delay, taking weeks or even years of service before the cracks appear or failure occurs. It is this aspect that
has provoked some researchers to speculate that hydrogen embrittlement may also be involved in this
problem. In high-strength steels that form martensite in the HAZ, hydrogen-induced cold cracks will
generally form preferentially, but in plain-carbon steels of low hardenability, hydrogen markedly increases
the susceptibility to lamellar tearing.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |99
 Figure 2.11.25. Elongated (MnFe)S stringers in hot rolled steel plate.

On the basis of these factors, lamellar tearing can be avoided using steels with good transverse properties,
such as Ce-treated grades, which affect the sulphide-iron properties such that the inclusions remain
spheroidal even after hot rolling. Such Ce-treated steels are, however, comparatively expensive and other
means of solving the problem are often preferred. Lamellar tearing may, in principle, be avoided by
ensuring that the design does not impose through-thickness contraction strains on steel with poor through-
thickness ductility. Some possible design modifications are illustrated in Figure 2.11.26. Such changes will
usually entail an increase in cost and therefore need to be justified by experience. It is possible to grind or
machine away the volume of material where tearing is anticipated, and replace the cut-away portion with
weld metal, a process known as buttering. In severe cases the assembly is then stress relieved before
welding on the attachment. The risk of tearing may be further reduced by specifying a material of high
through-thickness ductility, which is usually achieved by limiting the sulphur content to a low value
(typically less than 0.007%). Preheating may also reduce the risk of lamellar tearing in some cases.

(a)

(b)

(c)

(d)

Figure 2.11.26. Redesign to avoid lamellar tearing: (a) and (b) replace fillets with solid weld metal or forged sections;
(c) buttering; and (d) modify the preparation of corner joints.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |100
2.11.4 Reheat cracking:
Reheat or stress relaxation cracking may occur in the HAZ of welds in low-alloy steel during postweld heat
treatment or during service at elevated temperature. An example of a reheat crack is shown in Figure
2.11.27. There may be several reasons for reheating the weld metal, and in multipass welding, multiple
reheating cycles may be involved. However, reheat cracking generally occurs when the weld is reheated to
temperatures in the range of approximately 400°C to 600°C to relieve residual tensile stresses. The
phenomenon appears to be associated with the grain growth zone of the HAZ, rather than the weld
deposit.

Figure 2.11.27. An example of a reheat crack in the HAZ of a Cr-Mo-V butt weld.

Most alloy steels suffer from some degree of embrittlement in the grain growth zone of the HAZ when
heated to temperatures in the range of approximately 400 to 600°C. Elements that promote embrittlement
are Cr, Cu, Mo, B, V, Nb and Ti, while sulphur, and possibly phosphorus and tin, influence the brittle
intergranular mode of reheat cracking. Among the low-alloy steels the most susceptible alloys are the creep
resistant grades of the type Cr-Mo-V, 2½Cr-1Mo and 1Cr-Mo, especially if the vanadium content is higher
than 0.1%. Mo-B steels are also particularly susceptible. The relative effect of the various elements has
been expressed quantitatively in formulae such as equations (2.11.13) and (2.11.14).
P = Cr + 3.3Mo + 8.1V - 2 …(2.11.13)
P = Cr + Cu + 2Mo + 10V + 7Nb + 5Ti - 2 …(2.11.14)

When the value of the parameter P is equal to or greater than zero, the steel may be susceptible to reheat
cracking. The cracks are usually intergranular relative to the prior austenite grains (as shown in Figure
2.11.28) and occur preferentially in the grain growth zone of the HAZ, usually in the parent metal, but also
sometimes in the weld metal. There are two distinct fracture morphologies: low-ductility intergranular
fracture and intergranular microvoid coalescence. The former is characterised by relatively smooth
intergranular facets with some associated particles, and occurs during heating to temperatures between
400°C and 600°C, whereas the latter shows heavily cavitated surfaces and occurs at temperatures above
600°C. The brittle intergranular mode is initiated by stress concentrators such as pre-existing cracks or
unfavourable surface geometry. In the absence of stress intensifiers, the intergranular microvoid
coalescence type of fracture is dominant. In the latter case, particles within cavities are either nonmetallic
inclusions containing sulphur or Fe-rich M3C-type carbides. Microcracks that form during postweld heat
treatment are likely to extend during service at elevated temperature.
Reheat cracking is thought to be closely related to the phenomenon of creep rupture. The microstructure
of the grain growth zone is likely to be relatively hard, particularly in alloy steels and high CE steels.
Furthermore, during reheating, reprecipitation of carbides is likely to occur, especially inside the grains on
dislocations, which further increases the hardness. Because of this, creep deformation tends to be confined
to grain boundaries. In general the observed cracks are intergranular along the prior austenite grain
boundaries in transformable steels, and along any of the grain boundaries in austenitic steels. In coarse

Welding Metallurgy NSW700

Study Theme 2.11 - Page |101
grained material stress concentrations at the grain boundaries can cause them to rupture and it is this
phenomenon which appears to occur in the grain growth zone during stress relieving.

Figure 2.11.28. A typical reheat crack in the grain growth zone of the HAZ of a Cr-Mo-V steel weld.

The most researched creep resistant alloy is the Cr-Mo-V type steel, with a typical composition of ½Cr-
½Mo-¼V. These steels are extremely prone to reheat cracking. The creep strength and stability of these
alloys is due to a refined microstructure with a fine dispersion of very stable carbides, including NbC, VC,
Cr23C6 and Mo2C. During the weld thermal cycle, these precipitates tend to dissolve in the austenite at
temperatures above approximately 1200°C, and grain growth can occur virtually unrestricted. Cooling
typically produces a mixed bainite-ferrite microstructure and some fine continuous precipitation of VC. The
weld cycle is thus expected to produce a hard, coarse grained HAZ structure. Reheating the alloy to
approximately 600°C brings about further precipitation of VC and some additional precipitates of NbC,
which tend to form on dislocations inside the grains. The combined effect of cooling and reheating is thus
to harden the grains even more. Reheating also causes impurity segregation to grain boundaries, with
segregation being more concentrated for large grain sizes. Elements with the highest segregation factors in
ferrite are (in order) sulphur, carbon and boron. The relative effect of these impurities on reheat cracking is
shown in Figure 2.11.29 in Cr-Mo-V steels as a function of grain size. The grain size is therefore important,
both as a sink for impurities and also because of its effect on the mechanism of residual strain relief in
welds.

Figure 2.11.28. The effect of impurities and grain size on reheat cracking in various ½Cr-Mo-V steels.

In metals that fail in a brittle, intergranular manner, fracture appears to originate within the grain
boundaries. Cracks can either nucleate at inclusions and propagate as a result of reduced grain boundary
cohesion because of impurity segregation, or the cracks can nucleate at discontinuities in the grain
boundary such as grain corners (triple junctions). This type of barrier can cause large stress concentrations.
Because of the hard matrix, plastic accommodation of these stress concentrations is unable to occur, and
microcracks tend to form in order to relieve the high stress levels.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |102
Experimental studies of the complete process of intergranular creep fracture have revealed three
sequential, but overlapping stages:
− Nucleation of cavities on grain facets and triple junctions.
− Growth of these individual cavities to form cracks of one grain facet in length.
− Interlinkage of these single facet cracks to form cracks several grain diameters long which then rapidly lead to
failure.
A major fraction of time can be spent linking up single facet cracks to obtain the critical crack size needed
for fracture. On this basis, it has been found experimentally that coarse grained material fractures much
faster than fine grained material. This emphasises the detrimental influence of grain growth in the HAZ.
It is common when using materials susceptible to reheat cracking to prescribe a preheat treatment in order
to impose a slower cooling rate on the microstructure and thereby avoid an overhardened microstructure.
It should be pointed out that a potential negative effect of preheating is that, owing to the longer time
spent at high temperature, grain growth and carbide solution can increase substantially. In view of the
importance of grain size to creep rupture failure, this could have serious consequences. On this basis, it is
better to reduce the weld heat input, if possible, when employing preheat.
Therefore, the important factors that govern the cause and mechanism of reheat cracking are:
− The development of large grain sizes in the HAZ, which during welding tend to increase segregation, to increase
the likely number of ledges per boundary and to reduce time to fracture.
− The reheating temperature, which tends to encourage grain boundary segregation and fine reprecipitation within
grains.
− The joint geometry and weld heat input, which determine the amount of relaxation strain during reheating.
− The presence of impurities, which can reduce grain boundary cohesive strengths.
− The presence of grain boundary particles, which may be detrimental or beneficial depending upon such factors as
their size, interfacial energy, etc.
− The use of preheat, which can substantially increase grain size.

Any improvement in reheat cracking resistance has to be made by balancing these various factors against
possible losses in creep strength of the alloys concerned. Reheat cracking can be avoided and/or detected
by the following means:
− Material selection. For heavy sections, limit the alloy content as indicated by equations (2.11.13) and (2.11.14)
and limit the vanadium content to a maximum of 0.10%.
− Designing to minimise restraint. Where restraint is unavoidable, consider performing a stress relief treatment
when the component is partially welded.
− Using a higher preheat temperature (combined with a lower heat input); dressing the toes of fillet and nozzle
attachment welds; and using a lower strength weld metal.
− Carrying out ultrasonic and magnetic particle testing after PWHT.

Welding Metallurgy NSW700

Study Theme 2.11 - Page |103