In the evaluation of longitudinal seam discontinuities in aluminium replacement profiles, the primary metallurgical boundary is the charge weld generated in porthole die hollow extrusions, where divided metal streams are re-joined under compressive hydrostatic stress at temperatures above the recrystallisation threshold. On a 20 MN direct extrusion press processing EN AW-6063 billets of 203 mm diameter and 720 mm length, a four-port porthole die with a weld chamber depth of 25 mm and a bearing depth of 2.5 mm produces an extrusion speed of 3–5 m/min at billet temperatures of 460–480°C and container temperatures of 430–450°C. Failure records from building envelope replacement programmes show that seam openings concentrate at the internal web junctions of curtain wall mullions where the billet-to-billet transverse weld is allowed to advance beyond 15% of the billet length, thereby permitting oxide skins and dispersed intermetallics to enter the deformation zone. The longitudinal seam is assessed against minimum tensile values in EN 755-2:2016 and the tensile test method in ISO 6892-1:2019, but the standard does not provide a dedicated seam efficiency factor for porthole die longitudinal welds. Published data for this specific configuration is limited; therefore, destructive qualification tests are performed on each new die set and on each replacement profile lot before installation. The failure analyst must distinguish between the extrusion charge weld, the transverse billet-billet weld, and field-fabricated butt or fillet welds, because all three discontinuities are colloquially labelled as seams by contractors but exhibit distinct microstructural features, fracture surface morphologies, and process origins.

What process variables govern longitudinal seam integrity in welded aluminium replacement profiles?

For 6000-series alloys the critical process window lies between the solvus temperature of Mg₂Si and the solidus of the alloy. In EN AW-6063, solution treatment of Mg₂Si requires metal temperatures above approximately 520°C, while incipient melting of the Al-Mg₂Si eutectic can begin near 585°C when the silicon excess is poorly controlled. At the die exit, temperatures between 540°C and 570°C are common, and the weld chamber pressure must exceed the flow stress of the alloy by at least one order of magnitude. Flow stress for EN AW-6063 at 550°C is in the range of 20–35 MPa, whereas chamber pressure must develop roughly 300–700 MPa to rupture the oxide film and create metallic contact across the stream interface. If the press stroke is prolonged beyond 30 s, or if billet skin enters the weld chamber through insufficient butt discard, oxide film thickness increases from the natural passive range of 10–20 nm to several hundred nanometres. A speed-cracking defect appears at ram speeds above 10 mm/s when local surface tearing occurs due to excessive strain rate at the bearing exit. Entrapment of billet skin and oxides in the longitudinal seam yields planar arrays of magnesium oxide, aluminium oxide, and intermittent gas pockets. The defect is detected by macro-etch in 10% sodium hydroxide at 60–70°C for 10 min, showing a continuous white line extending along the profile length. Process variables are therefore not additive; they interact through the ratio between available welding pressure and alloy flow stress at the weld point, the welder chamber residence time, and the oxide surface area introduced by billet skin. Welder chamber residence time is controlled by extrusion ratio, ram speed, and chamber design. A reduction in welding pressure below the quality threshold cannot be compensated by raising billet temperature because higher billet temperature lowers flow stress but also promotes grain boundary liquation and magnesium silicate formation. The die configuration must maintain a minimum weld chamber pressure through sufficient chamber volume and port geometry; pressure transducers mounted in the die face provide continuous values for production records, and excursions below 250 MPa are rejected in seam-critical replacement profiles.

Across a series of architectural replacement programmes, batch-to-batch variation in billet feeder stock has been shown to alter seam consolidation even when extrusion parameters remain nominally unchanged. A batch of EN AW-6060 with magnesium 0.45 wt%, silicon 0.40 wt%, and iron 0.18 wt% produced seam ductility in the T5 state below 8%, while a second batch with magnesium 0.38 wt%, silicon 0.42 wt%, and iron 0.15 wt% exceeded 12% in full-section tensile elongation. The difference was traced not to bulk composition alone but to a billet taper heating gradient of 15°C measured by embedded thermocouples across the billet length and to a deviation in homogenisation cooling rate from 400°C/min to 180°C/min. Rapid cooling after homogenisation suppresses coarse β-AlFeSi formation, but if cooling is too slow, β-AlFeSi particles accumulate at the weld interface and reduce the available metal-metal contact area. In replacement profiles with wall thickness below 2.5 mm, the iron content should be kept below 0.25 wt% unless a post-extrusion thermomechanical treatment is used. Manganese additions in the range 0.15–0.30 wt% are beneficial for grain structure control but increase deformation resistance and can reduce welding chamber pressure if the die design lacks sufficient port area. For EN AW-6082 replacement profiles, the higher manganese and silicon contents require billet temperatures closer to the solidus, and the processing window narrows to ±5°C at the die exit. Seam quality in 6082 is sensitive to the free silicon content; silicon above 1.0 wt% creates eutectic liquid films at the welding point if the local temperature exceeds 575°C. The operational boundary is stated as follows: extrusion billet homogenisation must produce a fully spheroidised β-AlFeSi particle population, billet taper must not exceed 10°C end-to-end, and butt discard must remain above 10% of billet length for seam-critical orders.

Seam consolidation is not governed by extrusion ratio alone

Although extrusion ratio influences the degree of plastic work and interfacial expansion, it cannot compensate for insufficient weld chamber length or die temperatures below the critical value. For an extrusion ratio of 35:1 and a weld chamber depth of 20 mm, a porthole die may yield a macroscopic bond line that appears sound but contains fragmented oxide films less than 2 µm below the surface. The true metric is the ratio between available welding pressure and alloy flow stress at the weld point. Profiles produced with a weld chamber pressure of 180 MPa at a stock temperature of 560°C have failed under torsion in fewer than 500 cycles at a surface shear stress amplitude of ±30 MPa, while profiles produced at 450 MPa endured beyond 2×10⁵ cycles. Microhardness traverses across the seam using ISO 6507-1:2018 at HV0.1 reveal a hardness dip of 8–12 HV in poorly consolidated seams due to aligned oxide fragments and coarse recrystallised grains. The seam line is best examined by electron backscatter diffraction; the poorly bonded seam shows a 50 µm layer of refined grains adjacent to unhealed boundaries, whereas a sound seam has 80–120 µm elongated grains across the interface. Weld chamber depth is a key tooling variable: increasing chamber depth from 15 mm to 35 mm can raise the local welding pressure by up to 40 MPa for the same extrusion ratio and ram speed. However, excessive chamber depth increases billet skin retention and dead metal volume, requiring more aggressive nitride treatment of the die components. The die correction procedure must therefore be documented with pressure transducer values, not solely with dimensional measurements. Seam consolidation is also affected by vacuum-assisted die purging in the first billet of a new run; a 10 min purge at slow ram speed before production reduces gas porosity along the seam. This procedure is applied to replacement profiles with wall thickness below 2 mm because seam porosity of 0.3 mm diameter is sufficient to initiate cracking in thin sections. Extrusion ratio alone can be misleading because a high ratio increases strain but can also generate excessive die friction and localised temperature rises that lower the alloy flow stress and promote liquation. The processing window for seam-critical extrusions is therefore defined by pressure, temperature, and time, not by a single ratio value.

During forensic examination of failed replacement profiles, failure mechanisms are categorised by the location of the seam relative to the installed load path. In one documented case, a 6063-T6 hollow transom profile cracked longitudinally at the seam after 3 years in a coastal facade. The profile had a wall thickness of 2.0 mm, a visible seam line at the inner web, and the failure originated at a linear stringer of aluminium oxide approximately 4 mm long. Scanning electron microscopy with energy-dispersive X-ray spectroscopy showed oxygen enrichment of 8 wt% at the initiation site, against 0.8 wt% in the adjacent matrix. The crack propagated in mode I through the outer flange under wind-induced positive pressure cycles. In another case, a replacement mullion fabricated from roll-formed and gas tungsten arc welded 5052-H32 sheet failed along the fusion boundary during manual lifting because the weld underbead had 0.6 mm lack of penetration. The lower weld toe exhibited a 3 mm fatigue crack that had initiated at a surface porosity cluster. These cases demonstrate that seam failure analysis must differentiate between extrusion longitudinal seams and field-fabricated butt or fillet weld seams, because the two populations require distinct fractographic and process diagnostics. For extrusion seams, the fracture surface often shows alternate bands of oxide and ductile tearing, while welded seams show gas pores, intermetallic networks, or undercut geometry. Metallographic preparation per ASTM E3 and grain size measurement per ASTM E112 are used to document the extent of recrystallisation at the seam. If the grain size at the seam is more than twice that of the surrounding material, stress concentrations at grain boundaries contribute to early crack initiation. The seam is also examined by differential interference contrast microscopy after etching with 0.5% hydrofluoric acid to reveal the oxide film distribution. The presence of a continuous oxide film longer than 250 µm is correlated with seam tensile strength reduction of 20–40% across production lots. Replacement profile lots that contain continuous oxide stringers are not accepted for structural applications under EN 1090-3 because the seam cannot reliably transmit design tensile stress, and field repair by welding over an extrusion seam is prohibited unless the entire seam line is removed by machining to sound metal.

When replacement profiles are subjected to reverse bending fatigue

Under reverse bending fatigue, inactive seam discontinuities convert into through-wall cracks. A replacement aluminium profile installed in a facade subjected to wind pressure oscillations from vortex shedding may experience a local bending stress range at the seam of 55–75 MPa depending on profile geometry and mullion spacing. The fatigue endurance limit for defect-free 6063-T6 profiles is typically above 70 MPa at 10⁷ cycles under R=-1, but a seam oxide line longer than 0.5 mm can reduce the fatigue strength by 30–50%. The application of a conservative notch factor of 1.8 to seam defects below 1 mm is necessary when using finite element assessment. The evaluation protocol requires rotating or four-point bending fatigue tests on full-profile sections with the seam positioned in the tension zone. If full-profile fatigue data is unavailable, coupon testing per ISO 1099:2017 or ASTM E466 is performed with the seam included in the gauge section. In replacement applications where the profile is subjected to pedestrian live load or vehicular loading, cyclic loading with a minimum load ratio R=0.1 is performed for 2 million cycles. The seam is then inspected with liquid penetrant and macro-etch. Cracks exceeding 1 mm in depth or 5 mm in length trigger rejection. The strain field at the seam is measured using strain gauges with a gauge length of 0.6 mm located at the seam line and 20 mm adjacent. If localisation at the seam exceeds 1.5 times the nominal strain, the profile is judged unsuitable for fatigue-critical service. Since seam fatigue data for replacement profiles is rarely available, qualification testing is performed on each new profile family. The testing programme includes a first series at constant amplitude to establish the S-N curve, and a second series with variable amplitude representing wind gust spectrum. A reduction in fatigue limit below 45 MPa at 2 million cycles for 6063-T6 is considered unacceptable for facades in exposure category III under EN 1999-1-1. The seam side is additionally inspected for fretting damage at bearing supports, because fretting-induced shear along the seam can cause oxide disruption and create a crevice that accelerates crack nucleation. Bearing connectors are torqued to the lower end of the specified range to avoid local compressive yield in hollow profiles with seam lines near the web, and steel pins are fitted with isolating polymer sleeves 0.5 mm thick.

Before installation of replacement profiles, non-destructive testing (NDT) begins with a close visual inspection of the seam using a 10× loupe and a light source directed at 20° to the surface. For hollow profiles, ultrasonic testing with a 5 MHz 6 mm crystal diameter transducer and a 70° angle-beam probe is performed along the longitudinal axis, using a 3 mm side-drilled hole as the reference reflector. The acceptance level is based on EN ISO 11666:2018, with further evaluation per EN ISO 17640:2018. For anodised or powder-coated surfaces, the coating must be locally removed or a velocity correction factor applied; a 0.2 mm thick powder layer can attenuate the signal by 4–7 dB depending on resin fillers. Eddy current testing is limited to near-surface discontinuities to approximately 2 mm depth at 100 kHz. Macroetch inspection of sacrificial lengths is required at the rate of one per 1000 m of profile, using a 10% sodium hydroxide solution at 60–70°C for 10 min. The specimen is inspected at 10× magnification for linear seam opening, oxide line discontinuity, or crack branching. For seam continuity in porthole die extrusions, a flattening test on a ring section of 50 mm length is performed; the seam is placed at the centroid of maximum compressive strain and checked for re-opening. In addition, full-section tensile testing is performed on a 600 kN universal testing machine per EN 755-2, with a strain rate of 0.00025 s⁻¹ through yield and 0.0067 s⁻¹ after yield. Elongation is measured with a 50 mm gauge length. If the fracture path follows the seam for more than 80% of the fracture surface, the lot is quarantined. Seam efficiency is computed as the ratio of tensile strength across the seam to tensile strength of a solid profile of the same alloy and temper; a seam efficiency below 0.85 for hollow extrusions indicates incomplete consolidation. The destructive test results are compared with the extrusion record, including billet temperature, container temperature, ram speed, breakout pressure, and die exit temperature. A production lot is released only when the seam efficiency exceeds 0.85, no single indication exceeds the ultrasonic acceptance level, and the macroetch specimen shows no continuous oxide line longer than 5 mm in any 500 mm segment.

Corrosion-assisted seam splitting in architecturally exposed aluminium retrofit sections

In architecturally exposed aluminium retrofit sections, corrosion-assisted seam splitting involves preferential attack of the seam line when the profile is in a wet, chloride-laden environment with crevice geometry. The longitudinal seam in 6060 and 6063 alloys contains oxide fragments and sometimes magnesium-rich zones that behave as micro-crevices. Under cyclic salt spray testing per ISO 9227:2022 for 1000 h, sections with seam defects showed pitting depths of 0.3–0.8 mm along the seam, compared to 0.1–0.2 mm on the nominal surface. When coupled with steel fasteners without isolating washers, galvanic cell current density at the seam can increase from below 0.1 µA/cm² to above 1 µA/cm², accelerating intergranular attack. The problem is intensified in replacement profiles because old steel contact points may have left iron contamination on the aluminium surface. Aluminium replacement profiles installed against weathering steel must be separated by 0.5–1.0 mm thick EPDM or polyamide isolators; direct contact is prohibited by EN 1090-3. During failure analysis, the presence of chloride ions is confirmed by extraction in boiling ultrapure water and ion chromatography with a reporting limit of 0.1 mg/L. Stress corrosion cracking in seam-bearing 6000-series profiles is rare but can occur when the copper content exceeds 0.10 wt% and the material is peak-aged; the threshold stress in the transverse direction is approximately 75% of the Rp0.2 value. Seam defects lower the threshold further because the oxide line acts as a pre-existing crack or crevice. Test programmes for coastal replacement profiles therefore include a pre-exposure fatigue test: the profile is exposed to 500 h of neutral salt spray before cyclic loading, and the fatigue limit is compared with that of unexposed control specimens. A reduction of more than 20% after exposure indicates that the seam will be the controlling degradation path. Surface treatments are not fully protective if the seam has a surface-connected oxide stringer, because the anodising electrolyte penetrates the seam oxide and can produce localised soft coating. The anodised layer thickness over a defective seam may be 10–15 µm compared with 20–25 µm on the adjacent surface, a difference detectable by eddy current thickness measurement. Where replacement profiles are powder coated, acetic acid-curing silicone sealants are not used in contact with seam defects because acidic by-products etch the seam oxide and enlarge surface imperfections. Only neutral or polyurethane-based adhesives with documented compatibility to aluminium are permitted for seam contact.

Following destructive evaluation and field inspection, remedial validation of seam-bearing aluminium replacement profiles uses a gated production trial. Before batch release, the extruder records billet temperature, container temperature, ram speed, breakout pressure, and die exit temperature for every billet. For profiles with wall thickness below 2.5 mm and a seam line within 60° of the neutral axis, destructive evaluation is carried out on the first three profiles of each shift. The measured process data are not considered sufficient on their own; only the combination of destructive test acceptance and process traceability permits release. If the press record shows a pressure drop below the defined threshold for more than 10 s or a billet temperature excursion outside ±10°C, the affected profiles are segregated for additional testing. On-site welded seams in replacement profiles are qualified before installation by a procedure qualification record under EN ISO 15614-2 and welder qualification under ISO 9606-2. The operator qualification is valid only for the specific alloy and thickness range used; a repair weld made on 6060 is not automatically valid for 6082. For field-fabricated replacement splices, the joint is welded only after removal of the anodised layer to a distance of 20 mm from the weld edge, using a stainless steel wire brush dedicated to aluminium. Pre-cleaning is performed with a solvent degreaser that leaves no conductive residue, and the joint is dried before arc ignition. If ambient relative humidity exceeds 70%, welding is stopped unless the joint is locally shielded and preheated to 50–80°C. The weld zone is inspected visually after each pass; cracks, porosity clusters, and oxide inclusions larger than 0.5 mm are removed by grinding and re-welded. The completed replacement profile assembly is then subjected to a full-scale load test according to the structural specification, with deflection and strain measured at the seam location. The load is applied in increments of 10% of the design load, held for 60 s at each step, and released to check for permanent set. A permanent set greater than 0.2% of the span or a strain localisation factor greater than 1.5 at the seam triggers rejection. The final inspection includes a direct surface magnetic particle test for steel parts and liquid penetrant testing for aluminium welded seams, followed by application of a chromate-free conversion coating to restore corrosion resistance. For profiles that are powder coated after replacement, the coating must be applied over a chromium-free conversion layer within 4 h of surface preparation to prevent re-oxidation of exposed seam surfaces. These operational boundaries ensure that seam-bearing aluminium replacement profiles are not judged solely on the basis of average mechanical properties but through the combined evidence of process records, destructive test data, and non-destructive inspection.