Joining aluminium

Using the opportunities provided by the extrusion process for creative designs gives strong, stable, rapid and effective joints. Whether it is for joining one extrusion to another or for joining an extrusion to another material.

There are many advantages to be obtained by joining several smaller extrusions to a larger unit. Handling is easier. Pressing, surface treatment and a large amount of the machining can be done on a more rational basis. Smaller extrusions can be produced with less material thickness, better accuracy and in many cases lower die costs.

Jump to:

  1. Screw ports
  2. Tracks for nuts or bolt heads
  3. Snap-Fit Joints
  4. Joining Profile to Profile
  5. Telescoping
  6. Latitudinal joining
  7. Hinges
  8. T-joints
  9. Corner joints
  10. Joining with other materials
  11. Riveting
  12. End caps
  13. Adhesive bonding
  14. Fusion Welding
  15. Friction Stir Welding (FSW)

Screw ports

The screw port can be threaded in the normal way for machine screws.

Most commonly, screw ports are used directly for self-tapping screws. In these cases, the screw ports will have projections to centre the screws.

Port diameters for self-tapping screws
Screw no. Port diam. D Wall thickness t,min. Screw head clearance
ST 3.5 (B6) 3.1 ± 0.15 1.5 4.2
st 4.2 (B8) 3.8 ± 0.15 1.5 5.0
ST 4.8 (B10) 4.2 ± 0.2 1.5 5.8
ST 5.5 (B12) 4.9 ± 0.2 2.0 6.6
ST 6.3 (B14) 5.6 ± 0.2 2.0 7.4

Here, a component is being fitted by screwing through a port at right angles to the profile. In such cases, the port should have a shoulder.

Closed screw ports : Where the design requires a more robust screw (e.g. M8), the screw port can be closed. The port is to be dimensioned for thread cutting or for self-tapping metric screws.


Placing screw ports at corners saves material. To ensure that screw head does not protrude beyond the contours of the profile at outer corners, pay special attention to screw head diameter.

A screw port along the length of as profile facilitates "stepless fastening" , i.e. screw joints can be made at any point along the profile. Suitable dimensions are given iun the table below.


Screw port dimensions - screws at 90° to the profile ST 3.5 (B6) 2.6 st 4.2 (B8) 3.1 ST 4.8 (B10) 3.6 ST 5.5 (B12) 4.2 ST 6.3 (B14) 4.7

Upper joint: A hollow profile joined to another profile via a screw port. To avoid unwanted flexing in the joint, the screw is driven directly through the bottom of the hollow profile. A single screw is sufficient - the hollow profile's flanges stabilise the design. After step drilling, the hole through which the screw is introduced can be hidden using a plastic plug. Lower joint : The same solution, but without a hollow profile. The U-profile has tracks for the insertion of, for example, a metal or foil laminate strip.


Solutions with special screws that fill the screw head clearance hole are common in, for example, the furniture industry.

One way of avoiding step drilling and visible holes is to replace the hollow profile with two snap-fit profiles. This solution is often used in handrails.


This placement of the screw ports increases bending strength.

back to top

Tracks for nuts or bolt heads

Continuous tracks enable stepless fastening with no need to machine the profile. Dimensions for various nuts and bolt heads are given below.

If a standard bolt is too long, it is not always necessary to find a shorter bolt. The track for the nut can easily be designed/extruded as shown above.


Screw port dimensions – screws at 90° to the profile
ST 3.5 (B6) 2.6
st 4.2 (B8) 3.1
ST 4.8 (B10) 3.6
ST 5.5 (B12) 4.2
ST 6.3 (B14) 4.7

Using special nuts/bolts, fastning can take place without having to slide the nut/bolt in from the end of the track. There are no accepted standards, but various solutions are available from screw and fastener manufacturers.

If a set c/c distance between the bolt holes is required, a flat bar with precut threads can be put in the track.


The profile can be stamped to fix fasteners longitudinally in position.

Snap-Fit Joints

Aluminium’s elasticity is highly suited to snap-fit joints. These give far quicker assembly than, for example, screw or welded joints.

Snap-fit joints are widely used in a range of industries.

In openable snap-fit joints, the hook angle is α = 45° In permanent snap-fit joints, the hook angle is α = 0° (or negative). The length of the snap-fit joint has an effect on design.

A permanent snap-fit joint. Dimensions and tolerances must be decided on a case-by-case basis. The length of the hooking arm should not be under 15 mm. In some cases, long hooking arms may have to be extruded pre-stressed. This can eliminate the need for special tolerances.


If a design cannot accomodate hooking arms of sufficient length, the sprung part of the profile should be replaced by plastic clips or similar. The same applies if the joint is to be repeatedly opened. Aluminium's fatigue properties do not permit frequent changes in loading.

If a snap-fitting is difficult to assemble/disassemble, punching a section out of the hooking arm may be the solution.


Amongst other factors, the design of the joint is determined by whether or not it is to be openable. This joint can be opened using, for example, a screwdriver in the outer track.

Examples of snap-fit joints.


Plate A has a punched, rectangular hole. Mounting profile B is pushed into the hole until a snap-fit joint is formed. Lamella profile C is then pushed into profile B to form another snap-fit joint. Exploiting the spaceunder the plate makes it possible to have sufficently long hooking arms.

The hinge profile A (cut from a longer profile_ forms a snap-fit joint with main profile B. Punched hole C also provides longitudinal locking. Sufficient spring is generated in the hooking arm by springing the main profile at d.

Joining Profile to Profile

Longitudinal joining

Joining with a standard

Joining with a fluted, sprung profile in purpose-designed channels.

A sprung inner section that compresses to allow assembly. For easy entry, the inner profile (A) is believed and cut parallel to the main profiles. Tolerances are not critical in this solution. The result is a play-free joint.

Anchoring joined profiles by welding - the illustration shows solutions with a solid profile and a hollow profile respectively.


Longitudinal joining via asymmetrically located screw ports and a pre-drilled spacer. The profiles are turned so that the screws do not foul each other.

Longitudinal joining via longitudinal screw joints. A gap slightly longer than the length of the screw is milled in the screw port.


Longitudinal jointing using the spring and friction in a snap-fit design.


To ensure smooth abd silent operation, platic components are often used in telescoping designs. This design features stepless height adjustment using a nut (a threaded flat bar could also be used) that runs freely in its track. Tightening the fasteners locks the height and removes any play in the joint.

Height adjustment where the inner profile has a fixed thread (blind rivet nut) and the outer profile has a punched or extruded channel.


Height adjustment where the outer profile has a fixed thread (blind rivet nut) and the bolt clamps the inner profile in position.

Telescope solution with stepless clamping


Telescope solution with sping locking.

Where a play-free joint is essential (e.g. a single leg stand), plastic gauge blocks are used.


Plastic is often an excellent solution where components have to be able to slide. A plastic profile can be a part of a telescoping assembly.

Plastic wheels used part of the fastening in the outer profile serve as spacers and give smooth, play-free telescoping.

Latitudinal joining

Larger cross-sectional areas can be economically created by joining a number of profiles together. This solution is often chosen because it is easier to machine smaller profiles individually rather than a single construction as a whole.

Mechanical joints, adhesive bonding, fusion welding and, as illustrated above, Friction Stir Welding. can all be used for latitudinal joining.

Using a flat bar, bracket or similar to join profiles together gives good flatness.

Latitudinal joining using screw ports.

Locking using a splined dowel pin.


Locking using a tubular spring pin

Latitudinal joining with a clamp.


Latitudinal joining with a snap-fit.

Latitudinal joining with a snap-fit.


Joining using an end plate that holds the sections together.

Joining by stamping (creates visible deformations).

Latitudinal joining using dovetail tracks. Note the shape – to achieve acceptable precision, sharp-tipped corners must be avoided.

back to top


A simple hinge - the ball's diameter should never be less than 5 mm.

If the hinge has a screw port, it can be easily locked longitudinally using plastic inserts and self-tapping screws.


A hinge with approximately 110 degrees opening.

Two profiles with 180 degrees opening.


Self-locking with approx. 180 degrees opening.

Chamfering the ball enables hinge disassembly as shown above.


Complex hinging for securing lorry tarpaulins. The hinge is made from three profiles joined together.


Both parts of this hinge are made from a single profile.

Three-part hinge made from a single profile.


Two-part hinge made from a single profile and with identical machining.

A pin in each end gives wide opening and a cost-effecient solution.


A longitudinally adjustable hinge.

Hinges can be made from other materials than aluminium. The illustrations shows a solution where a plastic or rubber profile can be used.


A simple T-joint using screw ports.

A strong jpont with flanges to take up torsional stress.


Screw ports used to join tubular and rectangular profiles.

To avoid flexing in the joint, the screws are driven directly through the inner wall. The outer clearance holes are plugged with standard plastic caps.


Fitting to a wall or another profile: The end fastener is cut from a longer profile and secured with screws.

Joining of a round tube and a transverse profile : The transverse tube comprises two profiles held together by snap-fit joint. This fastening avoids troublesome mating of the contours.


A simple and stable soultion for T and corner jointing of square tubes.

In the furniture and interior decoration industry, special fasteners are used where joints must be easy to take apart. The fasteners often run in a nut track and there is this thus a stepless fit with the mounting profile.

Examples of other special fasteners.

A simple T-joint using nut tracks, right-angled brackets and bolts.

Expansion locking using a wedge shape.


Expansion locking using splined pins.

Corner joints

There are various types of brackets that are exremely suitable for corner joints where the strength and rigidity requirements are high. The brackets are usually cut from long aluminium profiles. Brackets are usually designed to allow several fitting methods. The corner bracket above has both screw ports (for side screws) and channels for stamping. Fitting method can then be chosen to suit equipment, series size, etc.

A special machine or an excentric press is used in the stamping method of connecting profiles. The method is particularly common in long production runs.


In picture frames and other light constructions, the corner joint comprises two flat right-angled brackets, one of them with threaded holes.

This corner joint for square tubes uses self-tapping screws in the transverse screw ports.


Cast metal and plastic ties are a solution that is especially common in long runs and where jointing has to be provided in more than two directions. Various ties are available in standard formats.

Tie using sprung steel clips.


Ties are often rectangular. The main profile's contours are, of course, immaterial as long as there is an inner, rectangular chamber.

A torsionally rigid joint using a single screw. As shown, one of the profiles has flanges. This type of corner joint is used in, amongst other things, TV stands.


This stable corner joint, which has precise angles and good design, involves relativelyeasy machining only.

The flanges of the corner profile are bolted to the insides of the frame profiles. The frame profiles need only be cut at 90 degreees to ensure a snug fit. Where corners are visible, a large radius (as shown by the broken line) gives an attractively rounded design.


These frame profiles have screw ports and, to give a snug fit, need only be cut at 90 degrees when used with the corner profile shown in the illustration. The flanges of the corner profile create channels for the fitting of an outer profile (free choice of radius). Plastic caps are used to cover the ends.

A U-section with a punched or sawn cut. The saw cut should go down into the base of the profile. This can then be folded to give a fame with slightly rounded corners. The frame is locked using a joint on either a long or a short side.

A corner joint that can be used in, for example, a table. The plate and joint combination represents a very stable solution.


Corner joint using pre-mounted bolts in two of the profiles. The bolts are tightened from above using a special tool.

Joining with other materials

Printed circuit boards, metal sheets and other plates can be fitted in channels in the profile. A small deformation (catch) in the plate or the channel ensures good locking.

Protrusions punched into the profile/plate ensure radial locking.


Rattle-free locking through having the profile's arms actively grip the plate/sheet.

Glass and metal plates, etc. can be locked in place using a sprung, special plastic profile (the yellow profile)


A snap-fit joint can be used with formed plates.

A standard method of glazing windows and doors. Rubber profiles, which form snap-fits with the aluminium profiles, act as spacers for the glass. This method can also be used for other plates.


The "Christmas tree model" is a simple solution when jointing with wood.

A snap-fit using track in the wooden board.


Special screws with "snap-fit heads" can be used when jointing with woods or metal plates.

Short snap-fit brackets can be screwed/nailed into wood strips.


A snap-fit joint bewtween aluminium and plastic profiles.

To deal with high local surface loads and reduce wear (e.d. from a rolling steel wheel), a steel strip can be inserted in profiles.

back to top


Examples of blind rivet nuts and press nuts.

In a long profile, it is often uneconomic to build extra thickness simply to provide longer threads. Using blind rivet nuts or press nuts, all that is required is a hole.


The bling rivet nut is fitted from the outside using a special rivet gun.

Press nuts: These are fitted from the back using, for example, an excentric press.


Sliding pop riveting in a longitudinal profile channel.

Pop riveting at the end of a screw port.


Self-punching rivets countersink and join in a single operation.

Riveting without rivets : This method, which is highly suitable for long runs, can join different materials of different wall thicknesses. A crimping press is used.

End caps

Screws and screw ports are the most common method of securing metal or plastic end caps to box profiles.

If the end cap and the profile have the same nominal outer dimensions, any departures from tolerance specifications are clearly visible. The places where metal has been cut become particularly prominent if the profile is surface treated. One solution is to make the end cap slightly bigger than the main profile.


If the main profile is long, it is more cost-efficient not to have screw ports in this but in a purpose-designed end cap. Slight displacement of the holes in the main profile (relative to the screw ports) ensures that a force is set up pulling the end cap into the main profile.

This end cap wedges into the main profile. There is a strong press-fit between the end cap's arms and the channels in the main profile.


Channels in the main profile for the fitting of an end cap with a sprung arm.

An end cap with sprung arms - the cap is removable .


This plastic end cap is held in place by stamped catches in the profile.

Cast metal or plastic end caps are suitable for long runs where the shape of the main profile is complex or where a highly rounded end cap is required.


Two end caps can be held together using long screws or draw bars. Screw ports with adequate clearance are a suitable way of giding the screws. The result is one end cap with no visible screws. This is a good solution in, for example, fascias.

Adhesive bonding

After steel, aluminium is the metal that is most frequently bonded. Though, for example, far more cars are produced than aeroplanes, the adhesive bonding of aluminium in the aero-industry has attracted the most detailed research.

Aeroplanes have used bonded joints since the mid 40’s. Nowadays, the bonding of aluminium is even used for load-bearing components in aircraft. Of course, there are many more down-to-earth examples of the use of bonded aluminium joints. Volvo’s roof rack rail is just one of these.

Many different adhesives, pretreatments and bonding methods have been developed. Selecting the right one is not always easy. Nor is it risk-free to simply start bonding without adequate information.

Essential knowledge

The intermolecular forces that determine whether bonding is possible exert their pull over a maximum range of 0.5 nm (one half of a millionth of a millimetre). If the surface is contaminated or is made up of low strength oxides exceeding this critical “thickness”, there will be no attraction between the adhesive and the aluminium profile.
For good and consistent bonds, the joint surface must be known, reproducible and clean.

The adhesive must wet the entire surface that is to be bonded. To do this, it has to have a lower surface tension than the material being bonded. Otherwise, the adhesive will form droplets rather than spread evenly over the surface.

All adhesives wet aluminium. To bond aluminium profiles to another material, the adhesive must be able to wet this material too. If the other material is a plastic, it can sometimes be diffi cult to fi nd an adhesive with a lower surface tension.

[bild 58.1] Traditional tongue and groove.
[bild 58.1] Tongue and groove with a channel into which the “locking hook” can be hammered or rolled.
[bild 58.1] A variant of the “adhesive trap” and “locking hook” method

Joint design

Adhesive bonding involves the formation of a plastic or rubber load-carrying element. The material in the cured adhesive bond is not as strong as the aluminium.
This can be compensated for by designing profile solutions that provide large contact surfaces.

Aluminium profiles can be easily worked into a wide range of shapes. Where tongue and groove type bonded joints are a possibility, they may be the best solution. The illustrations above give some ideas and guidance on joint design.

Adhesives cope best with shearing forces. Joints subjected to tensional forces are often unsuitable for high loads. Peeling and cleaving forces concentrate stress on a small part of the joint and should be avoided whenever possible

Choice of adhesive

Bonded joints distribute stress relatively well. However, very rarely is stress evenly distributed across the entire surface area of a bonded joint. As a rule, stress is greatest at the edges of the joint.

The stiffer the chosen adhesive, the greater the concentration of any subsequent stress. This leads to (sometimes unnecessarily) high stress on the adhesive and the surface that has been bonded to.

Thus, never choose an adhesive that is stiffer than necessary. Thicker bonded joints also reduce the concentration of stress at the edges of the joint.
The choice of adhesive is determined by the way in which the adhesive works and what is required of the bonded joint (filling/sealing, heat resistance, toughness, etc.).

To be able to mould itself to the surface structure of the profi le, the adhesive must have good liquid properties. It must also harden into a material that can transfer stress in the environment where it is used. Furthermore, it is important that the adhesive has time to mould itself to the surface’s micro-profi le. Fast setting, high-viscosity adhesives rarely permit this. In such cases, it may be advisable to first apply a low-viscosity primer.

The change from liquid to solid is effected in three different ways.

Drying Cooling Curing by
The adhesive is liquid when it is hot


– Heating

– Exposure to moisture

– Illumination (UV or blue light)

– In the abseence of oxygen

– Contact between adhesive and hardener (without preliminary mixing).

Solvents and water vaporise. Thus, adhesives containing solvents or water are unsuitable where:
– gap filling is required
– both the materials are unable to let the solvent escape.

Double-sided PSA tape should be regarded as a drying adhesive that never dries.

The material forming the joint is the same as that in the roll. However, if the stress is low, double-sided structural PSA tape may prove suitable for joining aluminium profiles together.

Double-sided PSA structural tapes formed entirely of the adhesive substance itself are available in thicknesses from 0.1 to 6 mm.
There are also double-sided PSA tapes that can be heat cured. The tape holds the components even during curing – other forms of clamping are unnecessary.
Testing of a simple overlap joint has shown a strength after curing of around 10 N/mm2.


Some thermoplastic adhesives have good plasticity when hot. Hot-melt adhesives are the most widely used. However, the thermoplastic hot-melt adhesives usually set too quickly on aluminium. This results in poor contact with the aluminium surface. Hot-melt adhesives also have very low creep and heat strengths. Many thermoplastic hot-melt adhesives become brittle in cold environments.

Moisture-curing hot-melts are applied at lower temperatures and, compared to thermoplastic hot-melts, have excellent properties after curing. They are used for, amongst other things, applying foil coatings to aluminium profiles.

Heat-reactivated adhesive is also used when coating aluminium profiles with foil. An adhesive solution or a water-based adhesive is applied to the material and left to dry completely. In the bonding process, so that it wets the opposite surface, the adhesive is heated.

Moisture-curing hot melts and heat-reactivated adhesives can both give strong, durable bonds.


Curing adhesives make up the large group of structural adhesives. They cure (often with negligible contraction) in one of the following ways:

Curing by mixing of the components

Typical of this group are the epoxy and polyurethane adhesives. They have very good gap fi lling properties. In principle, they can be cast. Modifi ed acrylic adhesives are now also becoming more common.

There are both stiff and elastic, 2-component, epoxy and polyurethane based adhesives. Epoxy adhesives with an elongation at fracture of up to 120% are now available. Elastic epoxy adhesives normally give a bond that is relatively heat-sensitive.

Using epoxy adhesives, higher strength bonds and improved durability are achieved by curing at elevated temperatures. The curing times are also considerably reduced – the curing time halves for each 10°C rise in temperature.

Two-component polyurethane elastomers give “rubber-like” joints that remain elastic even at low minus temperatures (°C).

There are also 2-component silicon adhesives that cure relatively quickly at room temperature.

Curing by contact between hardener and adhesive
(adhesive on one surface – hardener on the other)
These types of adhesives are usually referred to as SGA adhesives. They have excellent peel and impact strengths, but are not particularly suitable where a gap fi lling adhesive is required. These adhesives have been largely replaced by modifi ed acrylic adhesives, which are mixed direct from their packaging and can be used to form thick joints.

Acrylic adhesives of this type that adhere to untreated polyolefi nes (e.g. PE and PP) are now also available.

Curing by heating

Here, the most common adhesives are the 1-component epoxies. These require heat curing at a minimum of 100°C. With induction heating of aluminium profiles, curing times of approx. 60 seconds are possible.

The aero-industry makes extensive use of heat-hardening adhesive films. These require at least 30 minutes to harden at a minimum of 125°C.

One-component polyurethane elastomers can be heat cured at 70°C – 90°C (in 10 – 30 minutes).

Curing by contact with moisture

Cyanoacrylate adhesives harden very quickly in contact with moisture. A bond between two aluminium surfaces takes longer to harden than a bond between aluminium and plastic or rubber materials.

Cyanoacrylate adhesives are best suited for small joint surfaces and thin bonds. Normally, they have low peel and impact strengths. However, there are “rubber-filled” (black) cyanoacrylate adhesives with good peel and impact properties. Colourless, elastic cyanoacrylates are also available, but these are not particularly suitable as structural adhesives for metal.

Cyanoacrylate adhesives may be suitable where, for example, a plastic is to be bonded to an aluminium profile.

One-component polyurethane elastomers can also be cured by the humidity of the air. This type of adhesive is used in, for example, the bonding of car windows and, on a large scale, for aluminium profi les in container and vehicle body manufacturing. Curing is comparatively slow (hours) and dependant on relative air humidity and joint geometry.

Heat-curing polyurethane elastomers have been mentioned above. There are also polyurethane elastomers that harden both with moisture and heat. Two-component type polyurethane elastomer adhesives are also available.

As an alternative to polyurethane elastomers, there are the so-called MS polymers. These also harden with moisture. Two-component MS polymers are primarily chosen for work environment considerations.

Curing in UV light

There have long been 1-component acrylate adhesives that cure in tenths of a second when exposed to UV light (wavelength approx. 350 nm) or blue light (wavelength > 400 nm). Acrylate adhesives are often limpid and very suitable for bonds between aluminium profi les and glass (most of them perform less well with transparent plastics).

Epoxy adhesives that harden in UV light have also been developed. There are many types of these – limpid, fi lled, low-viscosity, hard, elastic, etc. Some of these adhesives can be irradiated prior to bonding and will then cure relatively quickly.

Curing in the absence of oxygen

Such adhesives cure on contact with active metal ions. They are normally referred to as anaerobic adhesives (or “locking fl uids”). They are not particularly suitable for aluminium. Aluminium surfaces should be regarded as passive. An activator has to be used in such cases. This gives a lower strength bond.

Variants of these adhesives that do harden without an activator on aluminium surfaces are now available.

Temperarure limits

With many adhesives, the practical maximum temperature at which stressed bonded joints can be used is between 60 and 80°C. The highest heat-resistance (approx. 150 – 250°C) is achieved with heat-curing adhesives and heat-curing adhesive films. However, silicon adhesives can give heat-resistance of around 250°C without heat curing.

Long-term strength

Aluminium surface at
x 25,000 magnification
(the red bar is 1 µm).

Bonds to aluminium are as strong and durable as the aluminium oxides with which the bond is formed. Aluminium that has had no surface treatment has a large percentage of magnesium in its surface. Aluminium surfaces should normally always be treated in some way.

Used in a dry environment, an untreated aluminium profi le can give an excellent bond. The same bond outdoors in a coastal climate may have a far shorter life. Bond lifetime depends on the synergistic effects of stress, temperature and environment.

Normally, the problem is not the degradation of the adhesive or the failure of adhesion, but the effects of changes in the underlying aluminium. Any good microscope will show that there are no completely fl at or even surfaces. Highly viscous (slow fl owing) and fast setting adhesives will, therefore, most probably only come into limited contact with the surface. This results in a bond with in-built weak points (air pockets) where the adhesive’s properties are not being exploited. In humid environments, this air will eventually be replaced by water. Where the water is salty, the need for surface treatment is even greater.
Aluminium’s durability can be improved by, for example, anodising.

Basic principles for long-lasting bonds

The basic principles for long-lasting bonds are well fi lled joints and resistant oxides. A large number of pretreatment processes have been developed for aluminium.
Some of the most common (and some of the more unusual) are presented here. Choice is determined by the environment where the aluminium is to be used, likely stresses and costs.

Full details of the processes and any risks to the work environment should, of course, be obtained before starting any form of treatment.
The main purpose of priming prior to the bonding of aluminium is to fill (seal) the surface when high-viscosity and/or fast setting adhesives are to be used.
Priming becomes more important where the aluminium is to be used in a corrosive environment and no surface treatment that improves corrosion resistance (e.g. anodising) is contemplated. Primer also “impregnates” and strengthens porous oxides, e.g. after chromating.

Requirement specification

It is advisable to draw up a requirement specification for the properties of the final bond and the use-related aspects of the adhesive. This helps crystallise the demands really being placed on the adhesive. It also makes it easier to specify exactly what is required to the adhesive manufacturer.

Pretreatment operations in bonding

Process Result Use (max.)
Cleaning/ degreasing Minimum requiement for ensuring a clean and defi ned bonding surface. For moderately stressd joints in dry surroundings.
Fine grinding/blast cleaning Removes weak surface layers e/g/ oxides. Safer than degreasing. Highly stressed joints in dry environments. Unstressed joints in fresh water.
Boiling water for 5 – 10 min. after pickling Gives resistant, but moderately

strong oxides.

Lightly stressed joints using

flexible adhesives in humid,

corrosive environments.



Corrosion resistant, but weak,

porous oxides.

Lightly stressed joints using elastic

or very low-viscosity adhesives in

corrosive environments.

Hydrochloric acid

at 20°C for

30 seconds

Quick, can impart a dark-colouring

to the aluminium surface.

Moderately stressed joints, even

in corrosive surroundings. Relatively

uncommon process.

Etching in


sulphuric acid

Thin, strong oxides. Long used

in the American aero-industry.

Highly stressed joints outdoors.

However, cannot withstand strongly

corrosive environments.

Anodising in

sulphuric acid

Thick very resistant oxide. Lightly stressed joints in corrosive

environments. Best with elastic


Anodising in

chromic acid

chromic acid Medium-thick, strong oxide.Used in the European aero-industry

since the 40’s.

Highly stressed joints, even in

corrosive environments

Anodising in

phosphoric acid

Porous, very resistant oxide. Is used

together with low-viscosity primer.

Optimum pretreatment for highly

stressed joints in corrosive


back to top

Fusion Welding

Aluminium is eminently suitable for welding. Although many welding methods are possible with aluminium, only a few are used in practice. Refinements in welding machines, equipment and materials have resulted in welding acquiring increasing importance as a jointing method.

Oxide formation

When welding aluminium, the metal’s reaction with oxygen, and the oxide rapidly generated therein, have to be taken into account. The oxide is strong, has a high melting point (approx. 2,050°C) and can easily cause welding defects. The oxide is heavier than the weld pool and may form inclusions. Thus, before all welding of aluminium, it is important to remove oxides from the joint surfaces. This may suitably be done using a stainless steel wire brush. Thoroughly cleaned, oxide-free joint surfaces are a basic requirement for faultless welded joints

Weld porosity formation

The risk of void formation must also be taken into account. The hydrogen contained in moisture and contaminants on or in the welding materials, work piece or air is highly soluble in molten aluminium. It loses this solubility almost completely when the metal solidifi es. As the weld pool sets, the hydrogen forms bubbles that may become trapped and form voids.

Most aluminium alloys can be welded

Drying Cooling Curing by
The adhesive is liquid when it is hot


– Heating

– Exposure to moisture

– Illumination (UV or blue light)

– In the abseence of oxygen

– Contact between adhesive and hardener (without preliminary mixing).


Nowadays, gas arc welding methods, MIG and TIG in particular, dominate. Argon (Ar) and helium (He) are used as the shielding gases in the MIG and TIG welding of aluminium. Argon and helium are inert gases and do not, therefore, form chemical compounds with other substances. Where there is a high penetration requirement, e.g. in a fi llet weld or when welding very thick work pieces, an argon-helium mixture can be used in MIG welding. The economic threshold for using mixed gases is a material thickness of 10 – 12 mm.

As welds in aluminium are prone to the formation of oxide inclusions and voids, the shielding gas must also meet certain purity requirements. The minimum requirement is 99.5% argon or helium. Besides playing a part in the electrical processes in the arc, the gas also has the jobs of protecting the electrode and the weld pool from oxidation and of cooling the electrode.

MIG welding

As a rule, MIG welding is used for material thicknesses from 1 mm upwards. In special cases, thicknesses under 1 mm can be welded using a pulsed MIG arc. Filler metal is added in the form of a wire fed through the welding torch. MIG welding can be performed in any position and for all joint types. A higher current density than in TIG welding gives higher welding speeds. The high welding speed has a positive effect on distortion and shrinkage (narrower heat-affected zone).

TIG welding

TIG welding is suitable for material thicknesses down to under 1 mm. In practice, there is an upper limit of around 10 mm, but edge preparation is then necessary. Filler metal is normally used and is introduced from the side. TIG welding can be performed in any position and, when performed correctly, gives the most fault-free welds. The welding speed is relatively high, and even higher in mechanical TIG welding.
TIG welding can be recommended where the gap width varies.

Robot welding

Robotised MIG welding can be used with advantage in long production runs. This method noticeably increases productivity and is also advantageous from a work environment point of view. The position of the work piece is easy to control. This facilitates welding from the optimum position and gives good results. Certain problems may arise with very thin materials and uneven gap widths.

Welding economy

Measured on cost per length, MIG welding is normally cheaper than TIG welding. Equipment costs are identical.

Filler metals

The table below gives recommendations for appropriate fi ller metals. AIMg5 generally gives the greatest strength. AISi5 is more stable as regards cracking and easier to use when welding hardenable alloys.

If the welded assembly is to be anodised, Si alloyed fi ller metals cannot be used. When anodising, the silicon is precipitated and imparts a dark grey, almost black, colour.

In order not to compromise weld quality, filler metals should be stored so that the risk of oxidation and the formation of other coatings is avoided.

Parent metal A
Sapa Swedish standard SS-EN-AW Chemical designation EN-AW







105A 1050A








3103 AIMn1 AI995Ti














AIMg52) AIMg52) AIMg52) AIMg3


5083 AIIMg4.5Mn0.7 AIIMg52) AIIMg52) AIIMg52) AIIMg5























AISi5 AISi5 AISi5 AlMg3







7021 7021 AlZn5.5Mg1.5 AlSi5 AlSi5 AlSi5 AlMg4.5Mn









Parent metal B Chemical








AlMn1 AlMg1(B)




AlMg4.5Mn0.7 AlMgSi









Swedish standard







3103 5005




5083 6060






Sapa 1050A 6060






1) Unsuitable where there is to be subsequent anodising.
2) Less suitable material combinations. However TIG welding with stated filler metal is possible.


In welding, the heat treatment to which the material is subjected affects the structure locally around the weld. The illustration is a schematic representation of how strength and hardness vary with distance from a weld in a hardenable alloy. With aluminium profiles, it is easy to compensate for decreased joint strength by increasing the wall thickness locally. Furthermore, edge preparation can be directly incorporated into the profile’s design

Profile design with regard to fusion welding

Appropriately designed profiles can greatly simplify welding. Edge preparation, material compensation, in-built fastening, integral root backing and the minimisation of the number of welds required are all examples of proactive aluminium profile design.

In many cases, aluminium profi les can be designed in a way that reduces the required number of welds. Sometimes, welds can also be located in a low stress section of the cross-sectional area. This will mean fewer welds and improved strength.

Edge preparation integrated into the profile design – the illustration also features material compensation for strength reduction in the weld zone.

Permanent root backing.
In-built fastener – used in dry environments.

Left: Placing welds in lower stress sections of the cross sectional area. This results in fewer welds, and butt rather than fillet welds.

Right: Number of welds reduced from 12 to 4 – butt welds rather than the weaker fillet welds (which are also harder to x-ray). Fewer components, reduced welding (consequently fewer heat-affected zones) and straightening minimsed.

back to top

Friction Stir Welding (FSW)

Friction Stir Welding (FSW) exploits aluminium’s ability to withstand extreme plastic deformation at temperatures that are high, but not above the melting point. In FSW, the clean metal surfaces of the profi les that are to be joined are heated by friction generated by a rotating tool and pressed together at very high pressures. This forms a new, homogeneous structure.

Compared with fusion welding, FSW gives:

# Increased strength.

# Increased leakproofness – entirely void-free, impermeable joints of a higher strength than fusion welded joints.

# Joints that are, in principle, flush with the surface.

# Reduced thermal deformation – only low thermal stress in the material, hence the flat surfaces.

# Increased repeatability – production has few variables and these are easily controlled; the result is tight tolerances.

An established technology

Left: A cross-section of a joint – x 13 magnification.
Right: The homogeneous crystal structure in the centre section of an FSW joint – x 220 magnification.

A rotating tool is pressed into the metal and moved along the line of the joint. No filler metals or shielding gases are used. FSW takes place at a temperature below the metal's melting point. The results include very little thermal deformation, hence the flat surface.

Using FSW rather than traditional fusion welding to join panels together gives, amongst much else, increased flatness and straightness. Strength is also increased (see the Royal Institue of Technology's tests, pages 72-73). The Sapa panel above is 3 x 14.3 metres.


The joint is in principle, flush with the surface and the FSW weld is, to all intents and purposes, completely void-free. The strength properties are also very good.

The FSW weld – homogenous and void-free with no oxide inclusions

To paint a clearer picture of FSW, we have chosen to compare it with the most commonly used method of welding – fusion welding. At the same time, we must stress that, in our production of added-value aluminium profi les, we often use fusion welding (MIG). The old does have its place alongside the new.

Fusion welding, MIG for example, uses fi ller metals and shielding gases.

The filler metal and the parent metal are melted and produce a weld bead that has a solidification structure different from that of the rest of the metal.

In MIG and TIG welding, attention has to be paid to the metal’s reaction with oxygen. The oxide rapidly formed in this reaction can cause weld failure.

The oxide is heavier than the weld pool and may form inclusions. There is also a risk of void formation.

FSW uses no fi ller metals or shielding gases. The joint is formed under the influences of friction generated heat and extreme plastic deformation.

The material being joined never reaches its melting point, but the profiles weld together in a way entirely analogous to the extrusion of hollow profiles.

The result is a homogenous and void-free weld with no inclusions.

FSW stands out in having only a few variables. These can be easily controlled to ensure the same results from one weld to the next.
Fusion welding is a more complicated process. Consequently, results often vary.

To give a fair comparison, the adjacent pictures are of very high quality fusion welds.

Left: Precipitation in a MIG-weld. Right: Precipitation in an FSW weld.

Left: The MIG weld rises above the surface. Furthermore, its chemical composition differs from that of the welded material. Right: The FSW weld is, in principle, flush with the welded material. No filler metals are used.

Left: A MIG weld viewed from above. Right: An FSW weld viewed from above.


Experience and extensive testing have shown that an FSW weld is usually stronger than a fusion weld. The table below shows the standardised values for arc welded butt joints as per SS-EN 288-4 (see also the tests carried out by the Royal Institute of Technology, pages 72 – 73).

The values given for FSW joints are based on a large number of measurements and should be regarded as guideline values.

Since there are, as yet, no standards for FSW joints, the values for fusion welded joints are used in calculating the strength of standardised designs.

Condition of parent

metal before welding

Ageing after


T = Rm (W)
Rm (pm)
T4 Natural ageing Arc

welding 1)

FSW 2)
0.9 0.9
T4 Artificial ageing 0.7 0.9
T5-T6 Natural ageing 0.6 0.7
T5-T6 Artificial ageing 0.7 0.8

1. For example MIG or TIG. 2. Guideline values only.

Ultimate tensile strength, R (w), of the welded test rod normally has to satisfy the following:

Rm (W) = Rm (pm) x Twhere Rm (pm) is the prescribedminimum ultimate tensilestrength of the parent metal and T is the joint’s weld factor.


The pictures on the right are of heat sink units based on solid profiles that are then CNC machined by Sapa. The machined
interior is closed with a cover, welded in place by FSW.
Helium leak testing was used to assess leakproofness.
The result was no loss of impermeability owing to weld failures. FSW joints have also been tested using the water pressure test.
The results are unambiguous – FSW gives a joint that can be used in components with the severest demands for leakproofness.


The experience Sapa has gained in series production since 1996 shows:
– Very small variations from joint to joint throughout a production cycle.
– Very small variations from joint to joint in repeat customer orders.
This is true of all variables – the joint’s structure, its strength, leakproofness and flatness.

Corrosion resistance

The chemical composition of the material in the joint is identical to that of the original material. Thus, in principle, corrosion resistance is unaltered.


FSW requires the work piece to be held securely in place. This means, amongst other things, that repair welding of finished constructions is rarely possible with FSW. Repairs can, of course, be carried out using traditional methods.

All 25,000 units passed helium testing for leakproofness.

Strength of FSW joints
Comparison with MIG and TIG – Reference:
The Royal Institute of Technology, Sweden

FSW welds have higher fatigue strength than MIG and TIG welds. This is the finding documented by Mats Ericsson, graduate engineer, and Rolf Sandström, professor, (both of the Institution for Materials Science at Sweden’s Royal Institute of Technology) in the December 2001 research report, Influence of Welding Speed on the Fatigue of Friction Stir Welds and Comparison with MIG and TIG. Test material and test methods

Test mateial and test methods

This extract from the report gives values for extruded profiles in alloy SS-EN AW 6082 (AlSi1MgMn) Ð temper T6, material thickness 4 mm. The dimensions of the test pieces were as per SS-EN 284-4. FSW was carried out by Sapa in a plant used for series production. Test materials welded at two different speeds were included in testing. To the same high quality standards as those applying in the aero-industry, fusion welding was carried out by CSM Material Technology. TIG and pulse MIG welding were used. Vickers hardness was measured with a load of 10 kg. Fatigue testing was carried out with a stress ratio ( min/ max) of 0.5, the main direction of stress being across the weld.

Profile design with regard to fusion welding

Appropriately designed profiles can greatly simplify welding.

Edge preparation, material compensation, in-built fastening, integral root backing and the minimisation of the number of welds required are all examples of proactive aluminium profile design.

In many cases, aluminium profi les can be designed in a way that reduces the required number of welds. Sometimes, welds can also be located in a low stress section of the cross-sectional area. This will mean fewer welds and improved strength.

The graph shows the variations in Vickers hardness across a cross section of an FSW joint (green) welded at a speed of 1, 400 mm/ min. and across a MIG weld (grey).

Comments: In both welds, hardness in the heat-affected zone decreases. This is clearly more marked in the MIG weld. Hardness is lowest (just under

60 HV) around the centre of the MIG weld. This is because fusion welding involves higher working temperatures, “foreign” filler metals and a less favourable structure in the weld. More heat is supplied in TIG welding than MIG welding. Consequently, the HAZ is a little wider. No significant difference was observed between the HAZs of the two FSW welds carried out at different speeds.

Rp0,2 (MPa)
Rm (MPa
A50 mm (%)
Min. values
for profiles
t < 5 mm
250 295 6 SS-EN 755-2
Pulsed MIG 147 221 5.2 ME,RS1)
TIG 145 219 5.4 ME,RS1)
speed A 2)
150 245 5.7 ME,RS 1)
speed B 2)
150 245 5.1 ME,RS 1)

1) Mats Ericsson and Rolf Sandström, averages of the results in the report in question.

2) Speed A, 700 mm/min. Speed B, 1,400 mm/min.

Fatigue strength

MIG-weld: This SEM micrograph (x 25 magnification) shows the fracture surface. Fatigue fracture developed at several points in the root (to the right).

MIG-weld: as above (2.500 magnification) Fatigue striation in the area close to the root edge.


The graph above shows the results of fatigue tests on MIG welds (grey), TIG welds (blue) and FSW welds (green). Comments: The FSW weld shows the best values throughout. In the study, TIG welds gave considerably better results than MIG welds. For failure at 500,000 cycles, the stress ranges were: MIG approx. 60 MPa, TIG approx. 70 MPa, FSW approx. 90 MPa at 700 and 1, 400 mm/ min (a shade higher at 1, 400 mm/ min).

SW: Fracture surface through the fine-grained section of an FSW weld (root to the right). Fracture probably developed close the root.

back to top