The Metallurgy of Welding
The Metallurgy of Welding
The Metallurgy of Welding
1 INTRODUCTION
Steels form the largest group of commercially important alloys for several reasons:
The great abundance of iron in the earths crust
The relative ease of extraction and low cost
The wide range of properties that can be achieved as a result of solid state transformation
such as alloying and heat treatment
A sound knowledge of what is meant by the word weld is essential to an understanding of both
welding and weldability. A weld can be defined as a union between pieces of metal at faces
rendered plastic or liquid by heat, or pressure, or both, with or without the use of filler
metal. Welds in which melting occurs are the most common. The great majority of steels welded
today consist of low to medium carbon steel (less than 0.4%C).Practical experience over many
years has proved that not all steels are welded with ease. For example, low carbon steels of less
than 0.15%C can be easily welded by nearly all welding processes with generally high quality
results. The welding of higher carbon steels or relatively thick sections may or may not require
extra precaution. The degree of precaution necessary to obtain good quality welds in carbon and
alloy steels varies considerably. The welding procedure has to take into consideration various
factors so that the welding operation has minimal affect on the mechanical properties and
microstructure of the base metal.
The application of heat, generally considered essential in a welding operation, produces a variety
of structural, thermal and mechanical effects on the base metal being welded and on the filler
metal being added in making the weld. Effects include:
Expansion and contraction (thermal stresses etc.)
Metallurgical changes (grain growth etc.)
Compositional changes (diffusion effects etc.)
In the completed weld these effects may change the intended base metal characteristics such as
strength, ductility, notch toughness and corrosion resistance. Additionally, the completed weld
may include defects such as cracks, porosity, and inclusions in the base metal, heat affected
2
zone (HAZ) and weld metal itself. These effects of welding on any given steel are minimized or
eliminated through changes in the detailed welding techniques involved in producing the weld.
It is important to realize that the suitability of a repair weld on a component or structure for a
specific service condition depends upon several factors:
Original design of the structure, including welded joints
The properties and characteristics of the base metal near to and away from the intended welds
The properties and characteristics of the weld material
Post Weld Heat Treatment (PWHT) may not be possible
As discussed, a steels weldability will be dependent upon many factors but the amount of carbon
will be a principal factor. A steels weldability can be categorized by its carbon content as shown
in Table 1.
In order to understand the physical and chemical changes that occurs in steels when they are
welded, a basic understanding of the metallurgy of steels is necessary.
for several reasons such as size, in situ, economics etc. In these cases, the weld metal and repair
procedure have to be carefully considered to ensure, as near as possible, matching mechanical
properties are obtained (or superior) with HAZ hardness taken into account.
The type of microstructure formed in the coarse-grained region of a steel depends upon:
The carbon content
The alloy content
The time at elevated temperature
The cooling rate
For any given steel, the greater the weld heat input the longer the time spent above the grain
coarsening temperature of the steel, and the coarser the grain size. Steels containing grain
refining additions such as titanium, vanadium, niobium, and aluminium are exceptions in that a
fine HAZ grain size may be achieved right up to the fusion boundary. Titanium nitride is very
stable and may not completely dissolve in the HAZ even at the temperatures immediately
adjacent to the fusion boundary. This can be advantageous with high heat input welds such as
submerged arc welding.
Figure 44 illustrates four welds in a carbon steel that have been welded with different heat inputs.
Alongside each weld the HAZ transforms to a microstructure dependent on the cooling rate of
that weld. For higher heat input welds, the cooling rate will be slower. In Figure !!, for the small
rapidly cooled welds, martensite is formed. For the large, slowly cooled welds the HAZ structure
is pearlite. The hardness of the HAZ is much higher in those welds in which martensite is present
as illustrated.
Adjacent to the weld the base metal undergoes various changes according to peak temperature
and cooling rate experienced at various locations away from the weld joint. Close to the fusion
zone the peak temperature will be high enough to cause complete transformation to austenite and
some grain growth.
At some distance away from the fusion zone the temperature is not sufficient to
cause any microstructural changes although other effects such as strain aging
(plastic deformation) may occur. In between a range of mixed structures may be
observed as illustrated in Fig.38.
SEFERIAN GRAPH
The Seferian graph shown in Fig. 43 takes into account CE, and restraint in calculating preheat.
WHAT NEXT?
BASE METAL CONSIDERATIONS
Recognize potential problems such as:
HAZ embrittlement
Loss of strength hardness in Q&T steels
Reduction in corrosion resistance
Porosity generation from base metal chemistry
Contraction cracking
Consequence of fracture
Locate specification
Spark analysis
Component function
Weldability Factors
Other factors for consideration:
Cost
Weldability of base metal
Preheat
Postheat
Base metal properties
10
Correlation of CCT and TTT Diagrams With Jominy Hardenability Test Data
for an 8630 Type Steel
11
Figure 33 illustrates the TTT diagram for a common chrome-molybdenum steel (4137) with a
Jominy end quench test superimposed. Thus the microstructure and hardness can be correlated
on the one diagram.
The cooling rate curves represent the same cooling rate conditions located along the Jominy endquench test bar. At the top of Figure 33, the measured hardness curve has been superimposed
over a schematic of the end-quenched bar. Four representative locations (A, B, C, D) along the
bar have been related to the representative cooling curves(CCT) and isothermal transformation
(TTT) curves. Thus location A on the bar experienced a fast cooling rate resulting in austenite
transforming to martensite producing the high hardness indicated. Similar cooling rate effects
need to be considered from a weldability viewpoint.
The addition of alloying elements (for example Mo, Cr, Mn) to steel increases the hardenability
by slowing down the rate of austenite transformation. The data is plotted as shown in Figure 34
for a 0.45%C steel with different alloying additions.
Figure 33 Typical End-Quench Curves for Several 0.45%C Low Alloy Steels
Several formulae have been developed which assign a contributing factor to each element
addition and its effect on hardenability and conversely weldability. The maximum hardness
attainable (and therefore its weldability characteristics) in carbon and low-alloy steels, however,
is still almost exclusively dependent upon the carbon content.
12
Another linear model has been used to predict the hardness of the HAZ for different cooling rates
in low alloy steels and is illustrated in Figure 36.
13
CEs are used widely in industry as measures of weldability. Several different formulae have
been developed and some are even incorporated into national codes and specifications. In general
terms, other factors being equal, as the carbon content increases, so does the difficulty in
weldability. In practice, this means generally using higher preheats until cracking and restraint
problems are overcome.
Using an engineering/analytical approach becomes very useful when confronted with unknown
material compositions, and weld repairs can become challenging where reverse engineering must
be utilized to develop a repair procedure. The engineering approach may involve evaluating
composition, hardenability, service conditions, size, restraint conditions, and PWHT feasibility.
One of the popular methods for determining weldability is to review the hardenability of the base
material. As discussed earlier the CE formula(s) have been developed as a convenient method of
normalizing the chemical composition of a material into a single number to indicate its
hardenability. Review of the literature indicates no less than a dozen different formulas have
been developed. One of the most commonly used formulas for calculating the CE is the IIW
formula:
It must be stated that low carbon steel and carbon manganese steels generally behave in a
predictable manner and are successfully welded with preheat and PWHT criteria outlined in
codes such as
AWS D1.1, Structural Welding Code Steel. The CE is not usually evaluated on these
materials. Medium carbon, HSLA, and Q&T Steels, however, present different challenges where
consideration of CE, restraint, hydrogen control, PWHT not practicable, weld filler chemistry
14
mismatch, weld heat input etc. can be critical to successful repair welding. These factors can be
summed up as a materials weldability, and it is these factors that will be considered in Section 8.
15
16
procedure adopted as well as the type of electrode flux used can affect the hydrogen content in a
weld or HAZ.
With hot cracking, embrittlement occurs in carbon and low alloy steels by a chemical reaction
occurring between hydrogen and carbides which causes irreversible damage either
decarburization or cracking or both. Of much greater importance in welding is hydrogen
entrapped in the weld or HAZ causing embrittlement. Hydrogen cracking can subsequently occur
at some later time (sometimes days) once a weld repair is complete, generally at service
temperatures between 100C and 200C. This embrittlement is due to physical interactions
between hydrogen and the crystal lattice structure of the steel and is reversible by removal of
hydrogen by stress relieving allowing the ductility of the steel to revert back to normal.
Hydrogen cracking can occur in either the weld metal, HAZ, or base metal and be either
transverse or longitudinal to the weld axis. The level of preheat or other precautions necessary to
avoid cracking will depend on which region is the more sensitive. In carbon - manganese
medium strength steels the HAZ is usually the more critical region and weld metal rarely causes
a problem.
Cracking due to dissolved hydrogen is now thought to occur by decohesion. Where there is a
defect, discontinuity or pre - existing crack and a tensile stress applied, hydrogen is considered to
diffuse preferentially to the region of greatest strain i.e. near to the stress concentration such as
near a crack tip. The presence of a relatively large concentration of hydrogen reduces the
cohesive energy of the crystal lattice structure to the extent that fracture occurs at or near the
stress concentrator. This view is consistent with observations that cracking can occur slowly (the
crack velocity being dependent on the diffusion rate of hydrogen) and is quite often
discontinuous.
In welding, the region most susceptible to hydrogen cracking is that which is hardened to the
highest degree (areas where the welding residual stresses is greatest) although regions of coarse
grain growth can be a contributing factor. The most crack-sensitive microstructure is high carbon
martensite.
Hot or cold cracking in the weld metal or HAZ depends on the same fundamental factors as in
the base metal, i.e. hydrogen content, microstructure and residual stress. In practice the
controlling variables are usually strength, hydrogen content, restraint, stress concentrations, and
heat input.
In single pass welds and root runs of multiple pass welds the root pass may provide a stress
concentration which can lead to longitudinal cracks in the weld metal. High dilution of the root
run (high heat input) can often result in a harder weld bead more likely to crack (this is
commonly seen in such applications as pipeline welding). Figure 40 illustrates the physical
appearance of hydrogen cracking in welds.
18
In Figure 41 the crack has initiated at the root of the weld where a lack of fusion can be seen.
The crack has then traveled through the HAZ mainly in the coarse grained region. In heavy
multiple pass welds cracking will generally be transverse to the weld direction, sometimes
running through the weld itself since the maximum cooling rate is along the weld axis. Many
HSLA steels in critical repair situations where PWHT is impracticable are welded using a filler
metal of good toughness and ductility and in such cases the HAZ may be more crack sensitive.
The risk of hydrogen-induced cold cracking in the weld can be minimized by:
Reducing hydrogen pick-up (low hydrogen flux chemistries)
Maintaining a low carbon content
Avoiding excessive restraint
Control of welding procedures (preheat; heat input; PWHT etc.)
Developing a non-sensitive weld microstructure
In carbon or carbon-manganese steels (i.e., those with steep hardening curves as shown in Figure
35!!) welding conditions can be selected to avoid the cooling rates at which martensite is
produced. This could include preheat; high heat input welding; slow cooling etc.
In low alloy steels or those where a hard HAZ cannot be avoided, other steps must be taken to
prevent cracks. These often involve applying preheat and interpass temperatures to allow the
diffusion of hydrogen out of the weld metal. Figure 43 shows that quite moderate temperatures
are highly effective in removing hydrogen.
19
The freedom of selecting a suitable welding solution is sometimes limited. The solution must be
practicable and economic. Further constraints may be applied by the job such as base metal
condition, size, location, PWHT not practicable, equipment availability etc. In such cases, the
welding engineer may need to consider the steels CE and M s temperature by referring to its TTT
and CCT curves in providing a weld procedure.
The cracks appear close to or a few millimetres away from the HAZ at right angles to the weld
interface as shown in Figure 44. In HSLA steels that form martensite in the HAZ, hydrogen induced cold cracking will generally form preferentially, but in plain carbon steels of low
hardenability, hydrogen increases the susceptibility to lamellar tearing quite markedly due to
20
HAZ stresses. There is no correlation between heat input and the incidence of lamellar tearing,
but in the presence of hydrogen a low heat input might tip the balance towards hydrogen
cracking because of a lack of time for hydrogen to dissipate away from the weld area.
Lamellar tearing may, in principle, be avoided by:
Design modification
Buttering weld runs and temper bead welding
Control of welding procedures (preheat; heat input; PWHT etc.)
6. CONCLUSION
Modern structural steels with their demands for strength, toughness, and good welding behavior
have evolved to depend less on carbon content as a strengthening agent and more on fine grain
size and precipitation hardening. This has meant that welding (specifically weld repair)
procedures may now have to utilize consumables which meet stringent property requirements as
well as avoiding cracking and other defects during welding.
21