This invention relates to the welding of packaging containers. More particularly, it relates to the welding of materials such as lacquer coated sheet metal and metal/polymer laminates for forming into can bodies. It has recently become popular for laminates of polymer and metal to be used in the manufacture of can bodies and ends. This is desirable for recycling purposes since conventional lacquers produce organic volatiles during application and toxic fumes when burnt. Such laminates comprise a metal substrate of tinplate which is protected by a thin chrome layer in order to prevent oxidation of the tin and laminated with a polymer layer on either or both sides. Alternatively, the laminate may comprise a metal substrate of tin free steel, onto which a polymer layer is laminated.

One known method of lamination comprises thermally bonding the polymer to the metal substrate. In particular where a polymer is applied to both sides of the metal substrate, the polymer is provided with a thin bond layer, for example obtained by coextrusion. Thus a composite of coextruded polyethylene terphthalate (PET) and a copolyester bond layer may be thermally bonded to one side of the metal substrate and, simultaneously, a composite layer of polypropylene and a bond layer of maleic anhydride graft modified polypropylene thermally bonded to the opposite side. The bond formed by softening the intermediate bond layers during lamination is exceptionally strong.

In the formation of three piece can bodies, the laminate or lacquered sheet is first cut into blanks.

Each blank must have uncoated side margins so that, when

it is formed into a cylinder, overlapping edges can be welded together to form the can body.

The complete removal of the polymer layer in a metal/polymer laminate (or of a lacquer coating from a metal substrate) is required if the metal is to be weldable. This is necessary for example for resistance welding in order that the contact resistance be low enough. If there is any contamination due to polymer or lacquer remaining on the metal, or if the metal is damaged in any way, the weld cannot be completed successfully.

Polymeric material is particularly difficult to remove successfully from metal/polymer laminates in order to expose the underlying metal substrate for welding. Furthermore, where the polymer has been thermally bonded to the metal via a bond layer, as described above, this bond layer is exceptionally difficult to remove completely without damage to the underlying substrate. A system for the removal of a polymer coating from sheet metal for the manufacture of cans is described in DE-4313871 (Krupp Maschmentechnik GmbH) . In that patent application, the coated metal as cut into strips and the edges of the strips are then stripped of polymer by passing under a laser beam before cutting into can blanks.

Although the edge cleaning by laser according to the prior art seems to produce visually acceptable margins, in practice we have found that welds made following such cleaning of lacquered sheet metal or of metal/polymer laminates were inadequately bonded. The degree of bonding obtained using such methods was typical of a cold weld. This indicates that the prior art systems give incomplete

edge cleaning which is unacceptable for welding purposes. This would arise if there is any contamination of the surface of the metal, even if the remaining polymer or lacquer is very thin, for example if it is the bond layer used in thermal lamination. A satisfactory weld would also only be possible if there has been no damage to the underlying metal substrate, ie the tin layer where the substrate is tinplate.

According to the present invention, there is provided a method of forming a tubular can body of material comprising a metal substrate laminated with a polymer layer, the method comprising: stripping margins of the polymer layer from the metal substrate by applying a pulsed laser beam having a pulse duration of less than lOμs and a peak power output of at least 0.1 megawatt to remove all the polymer along the margin without damage to the metal; forming a blank of the material into a tube by overlapping margins of metal substrate; and welding the margins of metal together to form a tubular can body, the weld having an extrusion width of from 20 to 60 % of the width of the overlap.

During the welding process, softened metal is squeezed out, or "extruded", from the overlap by the pressure applied. If there is little or no extrusion, fusion will not have occurred. If the extrusion is less than 20%, then the weld is likely to be cold. Preferably, this extrusion is smooth and consistent along the whole length of the weld. Where the extrusion is too great, the weld is likely to be brittle and may crack.

It is preferred that welding can be carried out with a minimum welding latitude of 200A, more usually at least 250A.

The lasers used in the present invention are of much higher peak power output than are CO,- continuous wave

(CW) lasers and other lasers which have been used in the prior art edge cleaning methods. For example, a CO CW laser typically may have a peak power output of the order of 0.001 MW and a pulse duration of 200 μs . It was believed that it would not be possible to use lasers having higher peak powers without damaging the tin layer on the metal substrate, thereby rendering it unsuitable for welding. Furthermore, known cleaning methods using lasers having lower peak powers were generally acceptable for most purposes which do not have the requirement of a very high standard of cleaning such as is needed for welding to a commercially acceptable quality.

In one embodiment, the pulsed laser may be a TEA (transversely excited atmospheric) CO _. - laser, which has a wavelength of 9.0 to 11.0 μm, and the preferred pulse duration of the laser beam may be between 0.1 and 10 μs. Alternatively, the pulsed laser may be an excimer laser, which has a wavelength in the ultraviolet range, with a preferred pulse duration of about 25 ns . The laser beam of the TEA CO.-. laser may have a fluence of between 0.5 and 10 Jmm ^": , typically between 0.8 and 8 Jmπf ² . Where the fluence is too low, incomplete removal of the coating may arise. Where the fluence is too high, there is a risk that the metal substrate may become damaged. Either situation results in unacceptable material conditions for welding.

Laminates which may be welded using the method of the present invention include laminates of metal and polymers such as pigmented PET, clear and white polypropylene. In addition, coatings of lacquers or metals, such as the chrome passivation layer used on tin free steel (electro-chrome coated steel) or the thicker chrome coating on tin free steel may be removed prior to welding in the same manner.

Preferred embodiments of the invention will now be described, by way of example, with reference to the drawings, in which:

Figure 1 is a side section of typical tinplate substrate; Figure 2 is a schematic side section of tinplate coated with polymer; Figure 3 is a schematic transverse side section of a fully acceptable weld;

Figure 4 is a graph showing a typical pulse shape for an excimer laser; Figure 5 is a graph showing a typical pulse shape for a TEA C0 ₂ laser;

Figure 6 is a schematic diagram of a system for removal of a coating using a TEA laser; and Figure 7 is an enlarged schematic view of a laser interaction area on the workpiece. Figure 1 shows a schematic side section of "E 2.8" tinplate 10. The tinplate substrate comprises a steel base layer 12 which is 0.2 mm thick, plated with a tin- iron alloy layer 14 and a tin layer 16 which are respectively 90x10 ^" " mm and 300x10 ^" " mm. Finally the tin layer is coated with a thin layer of chromium 18

(0.3x10 ^B mm) on which is a layer of chromium oxide 20 (0.8xl0 ^"6 mm) and finally a protective oil film 22(4xl0 ^" ' ⁵ mm) .

A polymer 26 is usually bonded to the tinplate of figure 1 by a bonding layer 24 as shown in figure 2. The removal of this bonding layer and not just the polymer has been found to be critical in the formation of a good weld. The chromium 18, 20 and oil 22 layers on the tinplate are sufficiently thin that they are removable by vaporisation during the welding process provided that complete cleaning of both polymer and bonding layer is carried out. Any bonding layer which remains after edge cleaning has been found by the present inventors to be a major problem in the welding of samples which might have appeared visually to be "edge cleaned". This is particularly a problem when white polypropylene or white PET is edge cleaned since the bond layer is clear. Removal of the white layer can therefore indicate on inspection with the naked eye that the margin is clear although in practice the clear bond layer may not have been successfully removed.

Another problem that arises in prior art cleaning methods is that even if the polymer 26 and its bonding layer 24 have been removed, the tin layer itself may be affected detrimentally. Typically, this is caused by overheating of the tin, which results in depletion of free tin coating and growth of the tin-iron alloy layer. This damage can also lead to incomplete weld formation. Since this depletion was already a recognised problem when using low peak power output lasers, an increase in this peak power by using different lasers was not

considered because further damage to the free tin coating would be expected.

A satisfactory weld is assessed by means of its "welding latitude". This is the range of current which when applied to form a weld will give a weld which is neither too cold nor having excessive splash. The minimum current setting at one end of the range will just avoid a cold weld with little extrusion, whereas the maximum setting just avoids one which is too hot and brittle. The welding latitude is the difference between these two current settings. A cold weld can also be easily ripped apart and splash arises due to excessive extrusion of molten metal as a consequence of excessive heat generated during the welding process. For example, two layers of the tinplate of figure 1 can be welded with a 0.7kA welding latitude.

Figure 3 shows a transverse side section of a fully acceptable weld. The welded region has been compressed by the application of pressure during welding to a thickness of approximately 1.4t, which has led to the extrusion of softened metal from the joint. In the figure, the upper extrusion b corresponds to about 40% of the overlap width a, whereas the lower extrusion c is about 25% of the overlap width a. A difference in the amount of extrusion is often desirable for aesthetic reasons, such as a greater extrusion being preferred within the container, although for the weld to be fully acceptable the extrusion must be at least 20° and not more than 60% of the overlap width. Pulse shapes and durations are shown schematically for excimer and TEA CO; pulsed lasers in figures 4 and 5 respectively. It can be seen from figures 4a and 5a that

both the excimer and the TEA laser produce very short pulses with high peak power. Typically, the pulse duration "a" for the excimer is 25ns and for the TEA laser is 3 to 5μs . In the schematic graphs of figures 4b and 5b, the pulse duration "a" has been magnified to show the different waveforms more clearly.

The selection of an excimer laser in the removal of coatings to create a weld margin does give satisfactory results but there is a risk at such high peak powers that the tin layer may be overheated and consequently affect welding latitudes. Furthermore, the processing speeds are much higher for TEA CO- lasers than for excimer lasers and the operating costs of TEA C0 lasers are much lower than for excimer lasers. Consequently, the preferred embodiment discussed below uses a TEA C0> laser.

A system for removal of a coating using a TEA CO? laser is shown in figure 6. A laser 30 emits a laser beam 32 of 5μs in pulse duration and corresponding to the waveform of figure 4. The ideal beam quality is obtained when the energy distribution across the beam is completely uniform, ie it looks like a "top hat". Preferably, this is achieved by making adjustments within the laser. Alternatively, a rectangular aperture 34 may be used to remove the edges of the beam, leaving the "top hat" profile but this is not a preferred option since it reduces available energy. The beam area is controlled by laser optics (here represented by convex lens 36) to the desired rectangular interaction area. The workpiece sample 38 is mounted on a movable frame to enable the laser beam to track along the workpiece at the desired line speed.

A close-up of the laser interaction area 40 is shown in figure 7. From this figure it can be seen that the laser optics gives an interaction area of constant, selected width "w" and length "1". The length of the interaction area dictates the achievable cleaning speed using a beam pulsing at a given rate with a particular power. This length is ideally between 5 and 500 mm. The laser optics are therefore selected in order to achieve an interaction area having a length within this desired range.

A preferred weld margin for a resistance weld is typically between 1 and 3 mm, although smaller margins may be weldable, with a pre-weld overlap of 0.4 to 0.7 mm. For production of welded can bodies from a laminate of sheet metal having a polymer layer thermally bonded on one side, this interaction area is moved along the strip (typically by movement of the can blank) at a line speed of up to 60 to 70 m/min. A second laser may be used to strip coating from the opposite edge of the can blank so that opposing edges are cleaned of coating before the blank is formed into the body shape with the edges overlapping and passed to a welder such as a resistance welder to form the can body side seam. Alternatively, the laser beam may be split into two beams for stripping both edges using a single laser source. It is necessary to clean margins on both edges where a resistance welder is used so as to avoid contact between the electrodes and the coating, particularly where this is a polymer or other insulating material. It is also essential that there is no polymer in the overlap.

Where the metal substrate has a coating on both sides, such as a polymer/metal/polymer laminate, a margin of coating needs to be removed from opposing edges and from both faces of the blank. Although it is possible to split a single laser beam into four in order to achieve this, provided that the laser output is such that the fluence is sufficient to strip coating effectively at the required speed, a pair of lasers or four independent lasers will probably be needed.

Comparative Example 1

Normal and lacquered body blanks were edge cleaned using a C0 ₂ continuous wave (CW) laser. The laser cleaned blank could not be welded to a commercially acceptable quality. Examination of the blank revealed the following defects:

- only a narrow stripe of 1mm width had been affected;

- small deposits of lacquer debris had accumulated along the burred edge of the blank;

- 0.5μm thick lacquer remained as a surface layer on the blank;

- some lacquer exfoliation had arisen.

Clearly the incorporation of the organic residue in the weld prevents welds of acceptable quality with any extrusion from being obtained.

Comparative Example 2

Blanks of tinplate coated on one side with 25μm white PET and 40μm polypropylene on the other side were edge cleaned using a CW CO-- laser. Each blank was cut

into sample sections. One set of blank samples received a single pass at an average power of 300W and a second set received two passes at 300W. Difficulty was experienced in welding the blanks and analysis showed polymer residue in the weld margin.

Examination of the samples revealed the following defects:

- on all samples, a 5μm layer of polymer on the whole of the margins on both sides of the steel; - burr of up to 30?. increase in metal thickness at the edge of the plate, causing a pool of polymer at the plate edge;

- residual polymer on the white PET side was clear except for a small area adjacent to the unaffected film where the pigmented film tapered out;

- residue still remained even with the blank samples which had a second pass of the laser.

It was believed that the clear residue on the white PET side was the bonding layer. Since this residue was clear, it appeared on exmination with the naked eye that the margin was clean. However, the difficulty in obtaining a satisfactory weld and examination under microscope revealed the presence of polymer residue.

Example 1

A laminate comprising 25μm white PET ("25WPET"), 0.19 mm single reduced E 2.0/2.0 tinplate and 40μm clear polypropylene ("40PP") was cut into can blanks. Four 3 mm margins were produced by laser edge cleaning using a TEA C0 ₂ pulsed laser having a pulse duration of 5μs. The laser fluence was 5.3 Jmπf' ^{^} for a line speed of 1 mms ^"1 .

The blanks were formed into can bodies with the PET side external. Side seams were welded using a resistance welder with a current potentiometer setting of 4.15 k-A. Unlaminated control samples were welded using the same potentiometer setting. Each weld was radiographed for examination and the radiographs showed a range of weld conditions and variability from can to can.

Metallurgical transverse and longitudinal sections were then taken from one of each of the welds (laminated/unlaminated) to assess the weld quality. The results are summarised in table 1, items 4 and 5.

Example 2

A laminate comprising 40μm clear polypropylene ("40PP")and 0.19 mm single reduced E 2.0/2.0 tinplate was cut into can blanks and edge cleaned using the system of example 1. The blanks were formed into can bodies with the polypropylene side internal. Side seams were welded using a resistance welder with current potentiometer settings of 3.8 and 4.1 kA. Unlaminated control samples were welded using the same potentiometer settings. Each weld was radiographed for examination and as in example 1, these radiographs showed a range of weld conditions and variability from can to can. Metallurgical transverse and longitudinal sections were taken of all the welds to assess the weld quality. These results are also summarised in table 1, items 1-3.

The welds of both example 1 and example 2 exhibited variable pre-weld overlaps although the level of variation within individual samples was generally acceptable. However, this variation can give rise to significant variation in weld quality at similar welder

settings. The variability from can to can shown by the radiographs also makes direct comparison difficult. However, no evidence of undesirable weld features was attributable either to the laminating or to the edge cleaning process. Fully acceptable welds from laminated can blanks which have been edge cleaned using a TEA C0- pulsed laser were produced using the higher current potentiometer settings.

Example 3

Samples of laminates of tinplate and polypropylene (pp) were welded following removal of a welding margin using a TEA C0 ₂ laser at differing fluence levels. The welding latitudes for these samples are provided in table 2. The welder used was a Soudronic Fbb 5501 throughout. It can be seen that for the samples of white pp/tinplate laminate, a larger welding latitude was available where the fluence level was lower. At high fluence levels, the latitude reduces due to damage to the free tin layer. In addition, the welding latitude increases if the welding speed is reduced.

Table 1

Current Ind i cat.ed Weld Weld

Sample se ting pre-weld thickness quality

(kA) overlap (mm)

(mm)

1 40PP internal 3.8 0.43 - 1.46- acceptable 0.56 ] .51

40PP internal 4.1 unclear 1.45 - acceptable 1.6

3 unlaminated as 2 0.57 - 1.55 - acceptabl e 0.65 1.65

4 25WPET/40PP 4.15 0.58 - 1.54 - good 0.67 1.63

5 unlaminated as 4 0.55 - 1.57 - good 0.63 1.59

Table 2

Laminate Welding Welding Fluence latitude speed/force level

(kA)

0.19 E2.0 tinplate 42m/min (control) 0.85 40kgf

0.19 E2.0 tinplate 42m/min 40μm clear pp 0.77 40kgf

0.19 E2.0 tinplate 53m/min 40μm white pp 0.55 4 kgf minimum

0.19 E2.0 tinplate 53rn/min 40μm white pp 0.42 45kgl mid

0.19 E2.0 tinplate 53m/min 40μm white pp 0.32 45kgf high

SUBSTITUTE SHEET (RUU 26)