Alumunium and Alumunium alloys

Literature
Alumunium-Zentrale (Hrg) : Alumumium-Taschenbuch, dzt. 14
Edition 1983, updating 1988, Alumunium-Verlag, Düsseldorf
Hufnagel, W : Alumunium-Schlüssel, Normenbezeichnungen,
Zumanmensetzungen, Markenwörter von Alumuniumwerkstoffen, 4.
Edition 1991, Alumunium-Verlag, Düsseldorf
Altenpohl, D. Alumunium von innen, Eine Einführung in die Metallkunde der
Alumuniumverarbeitung, 5. Edition 1994, Alumunium-Verlag Düsseldorf

1. Physical Properties

atomic number 13
specific gravity 2,7 g/cm3
linear expansion coefficient 24*10-4 1/K
modulus of elasticity 70000 N/mm2
modulus of elasticity in shear 26000 N/mm2
Poisson’s ratio 1/3
caloric conductibility 200 W/m*K
Electric conductibility 30 m/Ω*mm2
Paramagnetic half-life value 10 sek
2. The metal alumunium

Reseaches enabled Oersted (1825) in Denmark and W öhler (1827) in Germany

enabled Sainte-Claire Deville for the first time to produce small quantities of
alumunium in 1850 in France. With the development of the dynamo by Siemens
(1866) in Germany a sufficient source for the production of electric current in a
great scale existed and the use of the electrolytic method to produce primary
alumunium reducing it from alumunium oxide became practical. Heroult in France
and Hall in the US applied at the same application date independently for patents for
the production of alumunium using the electrolytic method in 1886.
The process to purity alumunium oxide from the ore bauxite was found by Bayer in
Austria (1887). The first industrial production began in Neuhausen in Switzerland
and in Pittsburgh in the US in 1888.

Alumunium, is after silicon the metal which occurs most often in the earths crust.
Ores, which can be economically processed, are available in quantity and the known
sources today can be estimated up to 40.109 tons. Taht enough to ensure with an
average growth of production of 2,5 % for at least 150 years. Due to the stronger
drive to recycling this period will be increased. A moreimportant problem in certain
regoains is the provision of sufficient and cheap current for electrolysis. This
significally effects locations for the production of primary alumunium.
The raw material for primary alumunium production is the “alumunium ore”
bauxite, whose alumunium content is between 20 and 30%. In Europe bauxite is
mined in Mediterranean areas of France, former Yugoslavia and Greece; the great
concentrations are in the tropical ares of West Africa, Brazil, in theCarribean and in
West Australia. Here lies almost 90% of the total. Additionally these ores are more
easily processed than those in Europe and thus more economically usable, despite
the higher transport costs. The trend goes to transfer the production of primary
alumunium to place, where cheap energy is available by the use of water power.

2.1. An inside view

When the molten mass of the metal and its alloying constituents is cooling, nuclea of
the crystals are formed . Around these nuclea alumunium atoms begin to place
therselves and form crystals. Inside the growing crystals gratings develop and at the
end the crystals fit completely. At the boundaries of crystals the gratings are
disturbed.
For the atoms of the alloy components two possibilities are open : they can remain
dissolved in the crystal structural arrangement, the crystals remains homogeneous
and the crystal a isomorphous structure,
or
they form their own crystals and the crystal structural arrangement is homogeneous,
these crystals are placed at the boundaries and also within the alumunium crystals.
2.1. An inside view

A simulation of the events in a simplified presentation can help understand the process :
Every atom has many elastic legs, equipped with suckers to connect them to the neighbouring
atoms.

Forces acting from outside this structure stretch the lags and after loadallevation the lags go
back to the original length. The deformationwithin the material is elastic.
But if the forces are so strong, that the sucker cannot transmit the load and pops off from the
connected atom to find a new hold at the next atom, the move is irreversible.
The deformation has a plastic nature. Slip planes formed and they are oriented according the
direction of acting forces. But they cannot have the same orientation in all crystals and slip
impediments at the grain boundaries develop. These impediments produce better mechanical
properties and smaller ductile yield. These effect are the work hardening.
Work hardening can be done unalloyed alumunium and on all alumunium alloys. Cold rolled
sheets can be so processes and cold drawn bars and tubes are possible.
This is the only method to raise the strength of alloys, that are not heat treatable.
Heating the metal over 250°C, the hardening begin to soften and at 350°C the effect are
totally reduced. A new hardening is not possible without new cold-work.
The high temperature during welding produce a soft state in the heat affected zone.
A slow warming up from 250°C to 350°C produces the recrystallization, causing in
the texture change crystals to grow, and has detrimental effects, such as producing
brittleness.
If the prescribed thickness is produced through rolling, we name the state end-rolled.
Different hardness is found across the sheet thickness, it is maximal at the surfaces
and the minimum is in the middle of the thickness. But rolling to a high hardness
and thermally treated to the desired hardness bring another distribution, the maximal
hardness is in the middle of the thicknessand the minimal are on the surface. This
state is end-glued and because of this hardness distribution bending or edging is
posssible with smaller bending radia.

2.3. Thermal treatment

The thermal treatment produces high strength in alloys, that are suitable for such
processes. Pure alumunium and alloys with only one alloy component cannot be
trated in such a way (magnesium + manganese behave as one component)
Heating to 460° - 540°C (according to the alloy involved), the atoms of the alloy
component dissolve and originate a homogeneous crystal structural arrangement.
The process is the solution annealing. A folowing fast cooling – quenching-prevents
the change to a heteregenous state.
The heteregenous state with the dissolved alloy atoms has the smallest slip
resistance, the strength is at an minimum and total softening exists. This state gives
the best conditions for bending or edging and totally contradicts, if steel is so
processed.
In room temperature the strength rises within the first 3 days rapidly and then
always slower to a maximum after one month.
This process is the cold ageing.
The temperature during welding are so high, that solution annealing takes palces,
but due to the high thermal conductivity, a soft quenching effect is produced. The
following cold ageing raises the strength from the soft state, but the improvement is
not great and differs between the thermal treateable alloys.
Ageing for a period of about 9 to 12 hours at temperatures between 140 ° and 180°C
the improvement in strength is much higher. This process is the warm ageing
(Fig. 2.23-1)
Temperature rise above 200°C brings a softening and at 350°C total softening is
reached. In the range of 250° to 350°C. When the temperature are rised slowly,
recrystallization occurs. The thermal treatment can be repeated but is in practice not
employed. Welded structures with their residual stresses, freed by such a treatment,
are distorted, bolted connections have no thermal treatment during assembly and no
softened zones.
In both ageing processes very complicated intercrystaline processes act. Dispersion processes
in the crystals, slip impediments take part i the hardening. The ageing processes affect the
corrosion properties, in some case strongly and unfavourably.
 Fig. 2.23-1 Time/Temperature-Relationship in Thermal Treatment

T [°C]
Lösungsg ] Ühtemperatur
500

400
Warmauslagerung
300

200 Auslagerungstemperatur
Kaltauslagerung
100
Raumtemperatur
20
t
5h gh

β2 [N/mm2]
300 warmausgelagert
kaltausgelagert
200
weich

100

t
1 Monet
2.4. Alumunium alloys
Alloys be judge from two points of view :
a) The cast alloys must have a suitable castability, the wrought alloys a good warm

and cold ductility.

b) Metallurgically alloys are divided into systems with one alloy component, that can
be only precipitation hardened, and with more alloy component, thatallow also
thermal treatment. In the first group are the alloysof the AA 5-Series with
manganese, magnesium and magnesium + manganese components, the second
group includes alloys of AA 6 – and AA 7- series with silicon + magnesium and
zinc + magnesium components.

Allloys for bearing structure never use copper components, because copper reduces corrosion
resistance heavily and the weldability is rather bad. Sheet used in aircraft are protected on the
surface by plating with pure alumuium, so the copper component in the alloy is isolated.

The designation of alumunium products include the alloy and condition as delivered. The
designations are determined in national standards. Internationally accepted is the International
Register of Alloys maintened by the Alumunium Association in the US.
According to the International Register of Alloys the most important alloy component is
defined by the first number of the total designation comprising 4 digits :

1.... pure alumunium

2.... copper
3.... manganese
4.... silicon
5.... magnesium, magnesium+manganese
6.... silicon+magnesium
7.... zinc, zinc+magnesium
8.... other

The second number is kept for possible change of an already registered alloy and the two last
numbers are the registration numbers.
The conditions are defined as follows :
H stands for work hardening and the following 1 means end-rolled, 2 end-glued. The grade of
hardening is given by the last number and is :
0 soft
1 ⅛ hardened
2 ¼ hardened
4 ½ hardened
6 ¾ hardened
8 1/1 hardened

Example : 5083 H24 alumunium-magnesium-manganese alloy

with 4,5% magnesium
0,5% manganese
end-glued, half hardened
:
T means thermally treated and the following conditions are :

1 soft
4 cold aged
5 no solution annealing, quenched by blast or sprayed with water
at the exit of the extruder (ca. 400°C)
6 warm aged
7 warm aged and stabilized
(against selective corrosion)
8 warm aged and work hardened before ageing
9 warm aged and work hardened after ageing

Example : 6082 T6 alumunium-silicon-magnesium alloy

with 1,0% silicon
0,6% magnesium
solution annealed, quenched and
Beside these most important designation other possible conbinations exist, but are not
included in materials for civil engineering.

In JIS, the Japaness industry standard, wrought alloys are listed according to the International
Register, in addition 5 alloys more, that seem to be used often in Japan, have Japanese
designations. The first digit is the same as used in the international Register, followed by N
for Nippon and a registration number. The alloys and their chemical composition are :
2NO1 1,52,5% Cu 0,5-1,3% Si 1,2-1,8% Mg
5NO1 0,2-0,6% Mg
6NO1 0,8% Si 0,5% Mg
7NO1 4,0-5,0% Zn 1,2-2,0% Mg 0,2-0,7% Mn
7N11 1,0-3,0% Zn 3,0-4,6% Mg 0,2-0,7% Mn

The Austrian standard M 3430, alumunium wrought alloys uses the same designation as the
German standard DIN 1725/1. It is very clear and gives the chemical composition with the
most important alloy components. The thermal treated alloys have F and the work hardened F
for end-rolled and G for the end-glued state.
The strength in kN/cm2 completes the designation.
Examples :
AlMg 3 G 24 work hardened alloy with 3% magnesium end-glued,
strength 240 N/mm2
AlMgSi 0,5 F 22 thermal treated alloy with 0,5% silicon, warm aged,
strength 220 N/mm2

2.5. The stress-strain curve and strength properties

The stress-strain curve for alumunium alloys is defined by Ramberg and Osgood
Є = σ / E + 0,002. (σ / Rpo.2)n (Fig. 2.23-2)

and with Є in %

Є = 100.σ / E + 0,2. (σ / Rpo.2)n

Є ............ elongation
Rpo.2 ....... 0,2%-proof stress = stress with a plastic strain of 0,2 %
n ............ exponent, depending on the alloy and condition
Rm.......... ultimate stress
σp ........ elastic limit
Є ............ Elongation at rupture
The exponent n is following Ramberg-Osgood formula to be derived by calculating the plastic
strain at the 0,1%-proof stress :

0,001 = 0,002. (Rpo.1 / Rpo.2 )n

n = log 2 / log (Rpo.2 / Rpo.1 )

As statistically verified values for Rpo.1 do not exist , the proposal of Prof. Steinhard used in
the German standard for alumunium structures, is a very good and practical value :
n= Rpo.2 / 10 mit Rpo.2 in N/mm2
 elastischer plastischer Fig. 2.23-2 Stress/Strain-Curve of Alumunium-Alloys up to Rpo.2
O[N/mm2] Anteil

Rp 0,2

Rp arctan Et
0,1

arctan E

0,1 0,2 ε (%)

Differentiating the formula for the stress-strain curve the tangent modulus of elasticity is
given by :
ET = dσ/dЄ
And with n = Rpo.2 / 10
σ = σ / Rpo.2
Is ET = E / (1+14. σn-1)
[more information in the Appendix]

The elastic limit according to DIN 4113 Teil 1 :

σp = Rpo.2 – 2.V (10. Rpo.2) and in N/mm2

Rpo.2 σp ET bei σp
280 174 69997
250 150 69995
200 111 69987
160 80 69970
140 65 69953
80 23 69819
The stress-strain curve is well defined by this formula up to the 0,2 %-proof stress. Above this
value up to the elongation corresponding to the ultimate stress in the tensile test a modified
formula can be adopted.
 nach Nach modifiziertem
Ramberg-Osgood Ramberg-Osgood

EI

EI.2 Teil 2 Teil 2

Teil 1

Teil 1

E1 0,2 0,75 * A1 ε [%]

At the .................... Of range stress is Rpo.2 and the tangent modulus ....................
elasticity must be the same, as in the first range ................................................ of
validity of the Ramberg-Osgood formula.
At an ........................................................, at the beginning of necking, the stress is
the ............................................... These condition can be satisfied,
when ................................................ Outside its validity begins at an elongation
Є1, ................................................. Є = 0 for Ramberg-Osgood formula. The
elongation ..................................... of the range is Є2 and, corresponding to the
alloy ........................................, tensile tests show it located between 0,6 x ε
and .............................................statistically good avarage is Є2 = 0,75 x ε

The ......................................... formula is

... = Є1 + 100.σ/E + (Є2 - Є1) . (σ/Rm)m
... = Rpo.2 / Rm Fig. 2.23-3

... = Є1 + 100.σ/E + (Є2 - Є1) . pm . (σ/ Rpo.2)m

... = (Є2 . pm – 0,2) / ( pm – 1)
the tangent modulus of elasticity is
ET = E / (1 + 70 . (Є2 - Є1) . (m / n) . pm . (σm-1 / Rpo.2 )m-1

And the conditions are fulfilled for ET with

0,2 = (Є2 - Є1) . (m / n) pm

The constant values for the curve in this range are obtained by iteration with
0,2 = (Є2 - Є1) . (m / n) pm / (1 - pm )
[more information in the Appendix]
Values for the stress-strain curve in the second range and tangent moduli of elasticity at the
end ofAlloy
the range follow
β :
0.2 β z A 5 M Є 1 E * t

N/mm 2
% % N/mm2
AlZn4, 5Mg1 275 350 10 11,6 -0,274 386

AlMgSi1 255 315 10 14,6 -0,149 281

AlMgSi0,5 160 215 12 11,9 -0,069 198
AlMg4, 5Mn 205 310 10 5,9 -0,494 650
Almg3 160 240 10 7,0 -0,259 441
At the elongation Є2 the tangent modulus should be 0, but as seen
from Fig. 2.23-4 the true curve evaluated from tensile tests
considering the diminution of the crossection due to necking and the
concentration of the yielding in the zone of necking, has a tangent
modulus > 0.

The European Recommendation for Alumunium Alloy Structure

recommend alloys with mechanical properties of the parent material :

Alloy Rm Rp0.2 plate, sheet profil

7020T6, T5 340 280 x x
6082 T6 310 260 x x
6061 T6 260 240 x x
6060 T5 200 150 - x
50830 270 120 x -
5454 H24 270 200 x -
 Fig. 2.23-4 Real and Deformed Stress/Strain-Curve through TensilenTest
0 [N/mm ] 2

Verfestigungbereich
Einschnürungsbereich Bruchbereich
300

R wahrer Verlauf

200
Rp 0,2

Verzerrter Verlauf
bezogen auf
A und 1 = 5d
100

Rp

Elastischer Bereich

0,2 0,75*A5 10 A5 20 ε [%]

In the German DIN 4113/1, Alumunium structures under predominantly static loads, and the
Austrian recommendations the alloys and their mechanical properties are :

parent material HAZ

plates and sheets profiles all

Legierung Rm Rp0,2 Rm Rp0,2 Rm Rp0,2

AlZnMg1 F35 350 275 350 280 250 175
AlMgSi1 F31 315 255 310 250 180 125
AlMgSi0,5 F22 - - 215 160 110 80
AlMg4,5Mn F31 310 205 - - 250 130
AlMg4,5Mn F27 - - 270 140 250 130
AlMg3 F23 230 135 230 135 180 80
AlMg3 F18 - - 180 80 180 80

note : HAZ is the heat affected zone resulting through welding, the strengths in the
heat affected zone are higher for thermal treated alloys than in the soft condition due to
cold ageing. AlMg3 and AlMgMn4,5 are not used for extrusions.
The effects of cold hardening or thermal treatment are seen in stress-strain curves in Fig. 2.23-5
0 [N/mm2] Bz, kaltverfestigt, hart

Bz, kaltverfestigt, ½ hart

Bz, weich

0,2 A5 A5 A5 ε [%]
2.6. Ductility

A ductile behaviour allows plastic deformations in case of overloading and secure

the structure against collapse.
Design and engineering, load conditions and insufficient toughness of the material
can cause brittle fracture. Biaxial stressesobstruct yielding and triaxial equal stresses
cause a brittle fracture evenin avery tough material. Notches have always more axial
stress conditions at the bottom. In bolted connections the bore holes weaken locally,
the heat affected zoneof a transversal weld is also small in comparison to a long
bearing part. Plastification is restricted, if overloading occurs, to this zone and the
main length is not affected. So the failure seems to be brittle. Impacts
impedeyielding, as the time to build up sufficient plastic deformations is not given.
With rising speed of loading the 0,2%-proof stress is lifted, the difference to the
ultimate stress becomes smaller and the elongation at rupture also. For these reasons
impact loads cause brittle behaviour. The notched bar test (Charpy test) gives for
steel an information to the ductile behavior. To judge the numerical values, the long
experience available for in steel structures are pretty new in thier application, such
criteria of judgment are not existing. As the impact is physical work, a specific work
calculated from the stress-strain curve can be helpful.
2.7. Long-term behaviour
Under longtime constant loads a small creeping is found in alumunium and alloy
structural members already at room temperatures, and in the German standard DIN
4113/1 this effect is taken into account by lower allowable stresses for such
loadings. Other standards neglect it.

If fluctuating loads produce over the whole design life load cycle in an extent of
more than ~ 104, depending on the stress amplitude, fatique design must be adopted.
The loading assumed for ultimate strength design is no longer applicable (Fig. 2.23-
10).
Further information can be found in 3.11 and the Appendix.

2.8. Bahaviour at low and elevated Temperatures

In case of low temperatures down to -200°C ultimate stress, 0,2%-proof stress and
rupture elongation improve, contrary to steel, which becomes sensitive to brittleness
under ~ -50°C. Temperature above 70°C and still more above 100°C lessen the
strength values, over ~ 200°C alumunium should not be used in load-bearing
structure.
The effects of higher temperatures are different because :

for temporary heating

strength reduction at that temperature
for heating over a longer time
strength reduction at that temperature and creeping

Temporary heating influence the strength at room temperature, if softening takes

place.
Under the elevated temperature conditions reductions in the allowable strength with
the temperature in °C and the upper limit at 250°C are
σ = σRT * [1-0,004. (T-100) ] for work hardened material

σ = σRT * [1-0,005. (T-85) ] for thermal treated material

σRT is the allowable stress at room temperature

the modulus of elasticity is
ET = E. [1-0,044. ( (T-20) / 100)1.75]
These formulas were deduced from values in the literature by the author and are not
covered by any standard.

High temperature over a longer period cause creeping in addition to reduced

strength. The deformation is time dependent and criteria are the same as used for
steel. The values for alumunium and alumunium alloys are different and the
reductions from allowable stresses at room temperature are severe.

2.9. Thermal expansion

The linear expansion coefficient αT is in the temperature range
< 100°C ~ 2,4 x 10-5 1/°C and that is twice the value for steel. For middle European
climates as an avarage the elongations over the whole year (summer/winter) are +
0,5 to 1,0 mm/m. If the elongation are totally blocked, thermal stresses, calculated
with numerical values of αT and E for alumunium, are

σ = 1,68. T N/mm2
and in steel
σ = 2,47. T N/mm2
2.10. Corrosion behaviour
The “love” of alumunium for oxygen is so “intanse” that it is really a totally unstable
material. Because of this affinity, the high expenditure of energy in theelectrolytic process is
required. Air oxidation on the surface begins within seconds and gives a hard and dense oxide
film. This prevents the progress of the corrosion; the film is renewed when scratched and
secures in this way corrosion resistance. The passivating oxide film is a corrosion product.
The rust on the steel surface is porous, not adherent and opens during the progress new layers
to corrosion, but to alumunium the oxide film adheres intensively. It is necessary to
differentiate between corrosion and corrosion damage.

Corrosion damage depends on the condition given by the application. For visually important
items even color changes of the thickening oxide film is damage. For bearing strucyures a
loss of the effective area diminishes carrying capacity and if loads have an oscillating nature,
notch effects deteriorate the life time. Leakage is a serious damage in vessels, tanks and
packaging material.

Corrosion damage can be divided into 3 groups :

1) Selective corrosion
This corrosion damage acts locally in the texture, and occurs in alloys with two alloy
components.
 Intercrystalline corrosion proceeds along the grain boundaries, produces severe
notches and diminishes the effective area.Alloys with higher component shares and
cold aged alloys are more susceptible. Warm aged alloys, used in civil engineering,
are not sensitive, ifthermal treatment is done at correct temperature.

Stress corrosion has the same effect, but a constant stress must be acting to initiate
this kind of corrosion. Mostly the stress results from a non adequate construction e.g.
by contraints. For alloys and conditions the same is valid as case of intercrystalline
corrosion.

Filament corrosion proceeds in layers parallel to the surface and under unfavourable
conditions rapidly leads to destruction. Alloys with zinc components are more
sensitive than alloys with silicon components, and a correct thermal, as already
described, can avoid such corrosion.

2) Corrosion by aggressive media

This corrosion damage is caused by chemical attack. The damages depends on aggreesivity,
different for the distinct media and their concentration.
Surface corrosion is not dangerous and has only optical effects, can be only rarely
encountered, mostly under reactions with alkaline solutions. It increases the thickness
of the oxide film, and in this way improve the protection.

Pitting corrosion is the most encountered corrosion, araising under reactions of

aggresive load from the air or water. It acts locally on the surface, and comes in most
cases to a stop. If the loss of material is schematically seen in the form of a cone and
assumed, that the volume lost per time unit isconstant, the deepness of the progress
into the material from the surface is growing with a rate given by the cubic root of
the time. The growth progress is smaller as time goes on, at the end it becomes
negligible (cubic root rule).
The ventilation in the bottom in the bottom of the cone can renew the protective film
and helps stop the corrosion.

Crevice corrosion is caused by enrichment of the concentration of the corrosive

mediumeven if this is at the beginning lightly aggresive. The chemical reaction
consumes the oxygen in the gap and the ventilation in the gap is insufficient. Both
events produce conditions favouring the corrosion the corrosion. The most dangerous
gap width is 0,02 to 0,5 mm. With an adequate construction this sort of corrosion can
be prevented.
Green and not set concrete has a high pH-value and is aggressive.
The value diminishes, when concrete binds and if dry, alumunium will no longer be affected.
As moisture is always possible, the interface should be protected with an bitumen coat on the
alumunium surface. Quick setting cement and antifreeze comounds in the concrete are in
most cases very aggresive, also after setting and should not be used in case of alumunium.

3) Biametallic corrosion
Different metals have differences in their potentials, which cause decomposition of the metal
with the lower potential, when an electrolyte exists and both metals have electrical contact.
The potential difference is decisive.
Potential differences of most important metals (passivation by oxide films taken into account)
metal Potential pH=6 in mV

silver + 194
copper +140
nickel +118
alumunium -169
tin -175
steel -350
cadmium -574
zinc -794

A galvanic element with metals of different potentials develops

The intensity of the process depends also on the electrolyte and the relationship of the areas
that the different metals in the conection have. Experience proves, that combination of
alumunium with steel, zinc coated steel, zinc and cadmium have no galvanic effects. Stainless
steel has a very high potential, but if its surface area is small in relation to the area of
alumunium no damage occurs. In the opposite relationship the alumunium part is endagered.
Combination with copper, brass and bronze are dangerous and should be avoid.
Corrosion resistance is slighlty different for alumunium for alumunium alloys, but in practice
no significant different in behaviour is found, alloys with copper contents excepted.
Work hardening has no influence on the corrosion resistance. Cold ageing brings risks of
selective corrosion and according to the standard only warm aged material should be used.
To prevent corrosion, the design and construction should take care that no constraints
producing constrant stresses exist, drainage and ventilation is given and gaps, if inavoidable,
are sealed at the interfaces with paint or foil.

3. Alumunium and alumunium semiproducts

3.1. Primary alumunium
The production of primary alumunium consists of two processes, independent of each other.
In the first the ore bauxite isrefined with the Bayer process to the pure alumunium oxide
Al2O3, alumina. This is used in the second step to gain pure alumunium in the electrolyte
process, reducing the oxygen. The problem of different melting points of Al 2O3 at 2050°C and
alumunium at 660°C is overcome with the use of the catalyst cryolite, which allows a process
temperature of 950°C. The catalyst contains flurine, which causes pollution, but as it is very
expensive, it is recovered to an extent of over 95% and recycled.
The production of 1 ton primary alumunium requires :
4 tons bauxite giving about 2 ton alumina
20 to 50 GJ thermal energy in the Bayer process
14 MWh electric energy in the electrolys
500 kg carbon anodes
95 kg cryolite

3.2. Rolled products

The alumunium billet is preheated to 400 °C and rolled at that temperature in the hot rolling
duo mill. Emulsion is poured over the rollers to prevent plating alumunium on them. The
billet is up to 500 mm thick and at the end the thickness is reduced to 6 to 12 mm.

In a cold rolling quarto mill the required thickness is obtained and work hardening done.
Thermal treatment is adopted, to get the prescribed hardening. Roller levelling in
straightening rollers, edge planing and cutting to length, punching of round plates and
stretching, which relieve the residual stresses, finishes the product.
3.3. Extruded products
The alumunium slug, preheated to 400°C is transformed according to the die of the extrusion
press to bars, tubes and profiles. 3 extrusion processes are known, the most important and
most applied is the direct, in which the slug in the recipient is moved by the plunger toward
the die. The indirect process avoids the friction at the walls of the recipient, as the die is
moved into the recipient towards the slug, which stands still. It need a lesser pressure force
and allows a higher extrucing rate. The hydrostatic procedure has a liquid under high pressure
in the recipient, preheating is not required but obtained by the high extrusion rates are
possible, as the friction on the walls of the recipient is avoided. It is possible to rotate the slug
and to obtain in this way profiles with helical webs. This process is only adopted, because it
is expansive, for small profiles with minimum tolerances or for special extrusion shapes.

The basic principle is common to all 3 extrusion processes.The material is pressed with high
pressure from the recipient into the die and emerges through the opening of the die, which has
the form cooresponding allows, due to the dies, a great freedom in the choice of profile
crossections. But this freedom must allow a good extrudability, care should be taken to select
the alloy and the degree of severity.
The opening in the die corresponds to the desired shape, when open profiles are extruded. For
hollow profiles a mandrel, mounted on a frame named spyder gives the contour of the inner
shape in addition to the oppening for the outer shape. Such dies are due to the mandrel and
the spyder more expensive.
(Fig. 2.23-6). The circumcircle of the profile is an important factor for the costs of the die and
the pressure force required. The circumcircle must be kept for at least 60 mm smaller as the
diameter of the recipient. To obtain a flow trough the opening of the die a pressure in average
of 400 N/mm2 is needed and certain alloys or high degree of severity the pressure must be
higher. The capacity of the extrusion press is reckoned from the diametre of the recipient and
calculated with the required pressure.
 Fig. 2.23-6 Shaping of Extruded Profiles by the die

Schnitt a-a
a
Profilöffnung

PreBrichtung

a
Following information are only a general orientation and must be taken only as a reference.
For an open profile with a low degree of severity the die costs are ~ 1000 $ for the
circumcircle of 100 mm and ~ 2000 $ for 200 mm, for a circumcircle of 350 mm the die cost
is ~ 5000 $. A hollow shape in a circumcircle of 200 mm will cost ~ 3600 $. Profile that have
a circumcircle > 350 mm needs dies, that are rapidly more and more expensive and for a Ø >
450 mm 20 000 $ must be calculated. Such big profiles are economical only, when great
quantities are required. In mostcases smaller sections connected to a bigger shape by
longitudinal welds are more cost efficient.

4. Welding of alumunium
4.1. Welding processes
The alumunium weld must be protected by a shield of inert gas to prevent oxydation. Argon is
used, rarely also helium. The welding arc cracks the oxide film and floats it off. Two method
are applied :
The weld arc is acting between weld and a tungsten electrode and the filler metal is given by a
filler rod (TIG). The welding bead is smooth and glossy. The process is suited for manual
welding and is adopted for short welds or welds around conners. One pass is sufficiently
penetrating thickness of 5 mm or less and over 5 mm two or more passes are required.
The weld arc is acting between weld and the filler wire serving as the electrode (MIG). The
process is suitable for mechanized welding and very economical. The high penetration allows
welding in one pass thickness of 8 to 10 mm. The welding bead is mat, gray and scaly. A
better quality of welding is given with MIG pulsed-arc method. This method is used also for
stud welding ofbolts with diametre > 8 mm.
Electron-beam welding needs no filler metal, is done in vacuum and is used for welding of
plates with great thickness. Even if the thickness is very big, only one or two passes are
required.
Other fusion welding methods, like gas welding with fluxing agents are not permissible for
alumunium structures.
Pressure welding such as spot or roll seam welding is seldomadopted and only for thickness <
1,5 mm. Flash-butt welding is adopted for joints in the corners of window frames.
Friction welding is a very economical solution, welding of parts of different metals is possible
and post-clenching avoide the softening effects by work hardening. It can be employed, when
one or both parts is round (round rods or tubes).
High-frequency induction welding is used for longitudinal welds in the production of thinn
walled tubes.

4.2. Filler metal

Generally the filler metal has higher alloy components than the parent metal to avoid cracks
at elevated temperatures. Welding of different alloys of parent metals is possible.
The metal of the filler rod/wire must be compatible to the parent metal according to the
International Register :

parent 5454 5083 6060 6082 7020

metal 6061
7020 5356 5356 5356 5356 5356
5183 5183 5183 5183
5754
6082 5356 5356 5356 5356
6061 5183 5183 5183
6060 5356 5356 5356
5754 5754
5083 5356 5356
5183 5183
5754
5454 5356
5183
5754
and according to the German DIN standard :
parent metal AlMg3 AlMg4,5Mn AlMgSi0,5 AlMgSi1 AlZn4,5Mg

S- S- S- S- S-
AlMg5 AlMg5 AlMg5 AlMg5 AlMg5
S- S- S- S- S-
AlZn4,5Mg AlMg4,5Mn AlMg4,5Mn AlMg4,5Mn AlMg4,5Mn AlMg4,5Mn
S- S- S-
AlMg3 AlSi5 AlSi5
AlMg3 S- AlSi5 AlSi5
S- AlMg5 S- S-
AlMg5 S- AlMg5 AlMg5
S- AlMg4, 5Mn S- S-
AlMgSi1 AlMg4, 5Mn AlMg4, 5Mn AlMg4, 5Mn
S- S- S-
AlMg3 AlSi5 AlSi5
S- S- S-
AlMg5 AlMg5 AlMg5
S- S- S-
AlMgSio,5 AlMg4, 5Mn AlMg4, 5Mn AlMg4, 5Mn
S-
AlSi5
S- S-
AlMg5 AlMg5
S- S-
AlMg4, 5Mn AlMg4, 5Mn AlMg4, 5Mn
S-
AlMg3
S-
AlMg5
S-
AlMg3 AlMg4, 5Mn
S-
AlMg3
The filler rods/wires must be stored under dry conditions in the original package, which is
opened immediately before welding. Damp rods/ wires, or rod/ wires that are not grease free,
cause weld defects.
4.3. The heat affected zone
The significant effect of welding, deteriorating the mechanical properties in the weld and the
adjancent weld junction, are caused by the high temperature. During welding they are higher
than the melting point in the weld, but in the weld junctions, due to the high thermal
conductility, 30 mm from the toe of the temperature is already down to below 200°C. This
region is called the heat affected zone (HAZ) or strength reduce zone (SRZ) and she is 1 inch
wide right and left of the weld toe.
Work hardened alloys are in this zone totally softened, in case of thermal treating, this
softening exists also immediately after welding. The rapid cooling produces a soft quenching
and there fore the strength is due to cold ageing raised from the soft state to better values, but
not so good as with a correct there mal treatment. This rise takes places within one month
after welding and the alloys of the 7-series with the zinc component have the highest cold
ageing effect. The metallurgical efficiency factor k is the relationship of the lower 0,2 % -
proof stress in the HAZ to the unaffected 0,2 % - proof stress in the parent metal and is 0,5 in
case of most alloys.
The temperature distribution is simplified in the definition of the HAZ, with the
temperatures in out side region having no effect, and constant high temperatures giving
constant strength reduction inside the zone (Fig. 2.23.7). The half width is in the US and GB
Canada fixed with 1: = 25,4 mm. And in Germany and most European state with 30 mm.
Studies in Great Britain have shown, that the width can ber ever larger if a sufficient heat
elimination is not possible e.g. the weld lies near to the edge, or if welds with many passes in
case of thickness > 15 mm were necessary.
 4.4. Preparations for welding
 No edge preparation is needed for thickness < ~ 5 mm, when welding in one pass, or < ~ 12 mm in case of
two passes. To avoid faults in the bottom of the toe, edges should have a bevel cut. A gap < 2 mm is
permissible. When thickness are bigger, edges with and aperture of 70° in case of MIG and 90° for TIG
welding must be cut. K-weld have a aperture of 60° at one side.
 Preheating normally is not necessary. Only if the thickness are very different and the temperature, required
for welding, tends to be insufficient, the thicker edge should be heated to maximally 250° C by a burner
with a slightly reduces flame.
 Before welding care should be taken that the edges are dry and grease free, and immediately before starting
to weld they must be brushed with a stainless steel
 brush to remove a part of the oxide film. The shield of inert gas may noy be
distributed by wind.
 4.5. Weld arrangements
 Backing help the welder, with predominantly loads acting, backing strips or weld
backs integrated in the extraction are used. In case of oscillating loads such
precautions are not preferred due to their notch effects. Then backing strips of
stainless steel are used, which can be removed after welding.
 As totally fault-free welds cannot be oroduced, allowable faults must be determined
and classificated. For static loads such rules exist, for fatigue loads not yet.
 The worst fault is weld cracking. If occurs, if an improper combination of pattern
and filler alloy is chosen. But it most cases at the beginning and the end of the
weld. Transverse welds with start – and end – plates avoids these cracks,as the
plates with cracks are cut off (Fig. 2.23.8). Welds can be started from a place with
low strain and end in similar place (Fig. 2.23.9). Critical faults can also occur at the
toe of the weld, and rounding of the edges, as mentioned before, is a great help, as
it allows the escape of oxides.
Aluminum has no change of colors at high temperatures and the welder cannot
optically control, if the danger “to weld holes” or to “fall through” exists, and less
experienced welder will try to avoid such events by insufficient penetration.
Porosity is more often encountered in aluminium welds than is steel welds, because
the weld temperature is much lower. The inert gases of the shield, air of hydrogen are
included I the molten mass of the weld, and due to the high welding velocity cannot
escape. Dense porosity is always caused by too high welding velocity in case of TIG
welding and too strong current when welding with MIG. Poor cleaning at the edges,
condensing water film on the surface and disturbance of the protective gas shield by
wind can also be responsible for poor welding quality.
Defect in welds, causes and avoidance :
Porosity :
Weld deposit too cool or too hot

Humidity from the air, condensed water, on the oxide film

Room temperature lower than on the work pieces, preheating to ~ 40°C

Leakage of coolant in the burner

Turbulences in the molten mass

Too much or too little cover gas

Draughts

Wrong burner handling

Hot welding cracks :

Wrong combination of parent and filler metal too high temperature in the molten mass

Correct combination of parent and filler metal

Higher welding velocity

Start – and end plates, program of filling the craters at the beginning and the end.
Lack of fusion :

Oxide film only partially removed, impurities

Better brushing, better cleaning, degreasing

Thickness of the connecting parts ate too different

Higher weld energy, preheating the thicker part

Defects in the toes of the weld :

Insufficien bevel cuts

Correct bevel cuts, backing strips

4.6. Residual stresses and distortions

The high shrinkage occurring at the transition from the molten to the solid state in the weld
is hindered by the unaffected adjancent zones of parent metal, and residual stresses and
distortions occur. The residual stresses of ~ 120 N/mm2, found in ususal bearing members,
are lower, than such stresses in steel structures due to the lower modulus of elasticity.
The distortion in aluminium structures are greater than in the case of steel, although the
difference in shrinkage is not very relevant. Straigtening is far more difficult as the resilience
is great, again due to the low modulus of elasticity. Local buckling often prevents it, and
aligment at high temperatures is not preferable. To prevent such obstacles, a good design can
be of great importance.
4.7. Effect of welding on the load-bearing capacity of the structure
The softening in the HAZ to an extent of 50% is fully effective in transverse welds, which
comprise the whole cross-section area, and the construction only can be help to reduce this
unwanted effect (see Fig. 3.11-2) . Longitudinal welds are allowed to plastify, and are the-
refore have not so much influence (see fig. 3.11-2,7,8,9,11).
Oscillating loads, which produce a high number of load cycles in the whole design life must
be treated under fatigue conditions. The important decrease of fatigue behaviour in welds can
be judged by decrease of allowed stress amplitude based on safe-life limit state 2 x 106 cycles
from 130 – 70 N/mm2 for parent material for the different alloys to 64 – 17 N/mm2 for all
alloys but depending on the notch effect. Scarsely this can be diminished by constructional
precautions. This is the reason why welding is avoid in structure. Which are highly loaded,
and with a high number of load cycles and in long design life, e.g. the aircraft construction.
The S/N-curve (Wohler-line) has like the S/N-curve for steel 2 slopes up to and from the
stress amplitude at 5 x 106 cycles when presented in the log-log scale (Fig. 2.23-10) and the
effects over the whole range can be calculated (see Appendix)