Chapter 18: Obtaining a BBA certificate for Reynobond 55 PE

The application to the BBA: March to August 2006

¶18.1 By the mid-2000s, it had become commercially essential for Arconic to obtain a BBA certificate to enable it to sell Reynobond in this country. Arconic's relationship with the BBA had begun in 1987, when the BBA issued a certificate for Reynolux, a solid aluminium cladding panel manufactured by the company under its former name of Alcoa Architectural Products S.A.S.⁴⁷⁴⁷ Initially issued in 1987, reissued in 2009 {BBA00008143}; Wehrle {MET00053190/42} page 42, paragraph 146. In March 2004 Arconic approached the BBA to certify its Reynobond product,⁴⁸⁴⁸ BBA project number S3/35071 {MET00053158_P13/137-138}; {MET00053158_P13/139-140}; {MET00053158_P13/141-144}. but it did not pursue the process for reasons of cost.⁴⁹⁴⁹ {MET00053158_P13/155}.

¶18.2 At a meeting in Luton on 21 March 2006, Colin Southgate discussed the UK marketing strategy for various products, including Reynobond, with Didier Scheidecker, Sales Manager for Reynobond. In his report of the meeting⁵⁰⁵⁰ {MET00053158_P13/162-165}. Mr Scheidecker noted that an increasing number of enquiries related to public and private housing developments which he thought could represent 50% of the UK market in 2006.⁵¹⁵¹ {MET00053158_P13/164} Item 5 under "BBA Approval". A BBA certificate was always required for such projects. He therefore asked whether Arconic could obtain BBA certification based on the approval by CSTB of its cassette version. (Since the test by CSTB of Reynobond 55 PE in cassette form had been a complete failure, that may have been a mistake for the riveted form.)

¶18.3 It seems clear, therefore, that Arconic had recognised that there was a significant potential market for Reynobond 55 PE in the UK and was intending to exploit it. It is also clear that its senior management appreciated that a BBA certificate was required in order to do so.

¶18.4 Arconic did not pursue its application to the BBA until 2 August 2006 when Mr Wehrle wrote again. He used the same form as on the previous occasion (wrongly dated 9 March 2004) and provided very similar information.⁵²⁵² {MET00053158_P13/166}. The form invited Arconic to provide any available test data relating to the product and pointed out that such information could result in a reduction in the cost of the certificate and the "duration" of the contract (probably meaning the time required to produce it).⁵³⁵³ {MET00053158_P13/173}.

¶18.5 When he completed the form, Mr Wehrle said that Reynobond was available with two cores, PE (unmodified polyethylene) and FR (fire retardant).⁵⁴⁵⁴ {MET00053158_P13/171} Section 2.1. He also referred to the tests carried out under BS 476 Parts 6 and 7 and referred to the Warrington Fire Centre reports.⁵⁵⁵⁵ {MET00053158_P13/169} Section 1.3. He attached to the application two test reports,⁵⁶⁵⁶ {MET00053158_P13/173} Section 3.1. each of which related to the fire retardant version of the product.⁵⁷⁵⁷ Warrington Fire Test Reports 132316 and 132317 {MET00053158_P13/173} Section 3.1. Mr Wehrle also cited two CSTB Technical Assessments on Reynobond 55 PE, one in cassette form and the other in rivet form. They supported structural and system performance but were not relevant to any fire performance criteria recognised in the UK.

¶18.6 Mr Wehrle completed the application to the BBA about a week after he received Colin Southgate's note of 26 July 2006⁵⁸⁵⁸ {MET00064988/19}. and his meeting with Gérard Sonntag the following day.⁵⁹⁵⁹ {MET00064988/20}. He must therefore have known that Arconic did not have sufficient data to make claims about the reaction to fire of all variants and forms of Reynobond 55 and that the information he was providing to the BBA was incomplete and misleading insofar as it suggested that it applied equally to the product in its cassette form.

Negotiations between Arconic and the BBA

¶18.7 On 23 August 2006 the BBA formally offered to assess Reynobond 55,⁶⁰⁶⁰ BBA project number S3/41014 {MET00053158_P14/100-101}; {MET00053158_P14/102-103}; {MET00053158_P14/104-105}. but despite its view that that a BBA certificate was increasingly important for UK sales,⁶¹⁶¹ {MET00053158_P14/115}; {MET00053158_P14/131}. Arconic did not immediately accept the offer. Following negotiations, the BBA agreed to reduce the cost of the certificate and to accept the Technical Assessments issued by CSTB as the basis for it. (The BBA and CSTB were both members of the European Union for Technical Approval in Construction ('UEAtc'), a group of certification bodies whose assessments were generally carried out to a consistent set of standards, so that data produced by one member was acceptable to all others.)⁶²⁶² Gregorian {BBA00011096/2} page 2, paragraphs 8-9; Gregorian {Day105/27:6}-{Day105/28:3}. However, because the fire performance analyses in the Technical Assessments issued by CSTB were directed to the French national testing system rather than the British Standards applicable under the UK testing regime, the BBA also required test reports from Warrington Fire Research on the product's fire performance. In his note of the meeting in November 2006 Mr Southgate recorded that the BBA would need certificates from Warrington Fire relating to the "PE Cores". It therefore appears that the BBA had asked for fire test data relating to Reynobond with both unmodified polyethylene and fire-resistant cores.⁶³⁶³ {MET00053158_P14/115}.

¶18.8 At a meeting between Arconic and the BBA at the latter's premises on 2 November 2006, Colin Southgate of Arconic suggested that it might be better for the certificate to apply to Reynobond as a material rather than the rainscreen system as a whole, since (in the words of his note) that would enable a "cross connection to be put together".⁶⁴⁶⁴ {MET00053158_P14/115}. The meaning of that part of his note is not entirely clear, but he seems to have been suggesting that the BBA should be asked to certify Reynobond 55 PE generally, without regard to the differences in fire performance between the rivet and cassette forms. By "cross- connection" we think that Mr Southgate had in mind that Arconic could use a certificate that did not refer to the method of fixing in relation to both the riveted and cassette forms without distinction, and in the event, the BBA eventually agreed to provide a certificate for Reynobond without linking it to any particular fixing system. It was envisaged that different fixing systems would simply be mentioned in the certificate.⁶⁵⁶⁵ {MET00053158_P14/131}. It is fair to point out that at that stage the BBA had yet to embark on an assessment of the product.

¶18.9 Warrington Fire Research had not tested Reynobond 55 PE panels in accordance with BS 476 Parts 6 or 7. It had tested only Reynobond FR in accordance with those methods, and those were the only British standard tests to which Arconic had referred in its application form.⁶⁶⁶⁶ {MET00053158_P13/173}. See penultimate row in table of test data. Arconic possessed no relevant European standard test data in relation to Reynobond 55 PE in cassette form, apart from Test 5B carried out in December 2004, in which the product performed so disastrously that the test had to be stopped.⁶⁷⁶⁷ {ARC00000536}.

¶18.10 There is no indication that Arconic disclosed to the BBA that Reynobond 55 PE in cassette form had performed so badly or that it had reason to believe that the fire performance of the product in cassette form was significantly worse than in riveted form. On the contrary, Arconic succeeded in persuading the BBA that the certificate could properly relate to Reynobond as a generic product without distinguishing between the different fixing systems. In doing so, it deliberately and dishonestly concealed from the BBA the fact that, as Test 5B had shown, its reaction to fire was significantly affected by the fixing system adopted.

¶18.11 At that time, some of those at the BBA knew that altering the shape of a product could materially affect its fire performance. John Albon, then Section Head of a different department at the BBA, understood that particularly well,⁶⁸⁶⁸ Albon {Day109/143:18}-{Day109/144:9}. although he did not become responsible for the technical content of certificates until 2009.⁶⁹⁶⁹ In 2006 John Albon was Section Head of the Materials department, but he came to be Technical Manager responsible for the technical content and correctness of all BBA certificates in 2009 and later Head of Approvals for Construction Products, Albon {BBA00010723/3} page 3, paragraphs 11-12. However, at the time of the negotiations with Arconic in 2006 the BBA agreed to assess Reynobond as a generic product in the form in which it left the factory, even though it knew that it would have to undergo fabrication to some extent before it could be used for the purpose for which it had been manufactured.⁷⁰⁷⁰ Gregorian {Day105/58:3-20}. At that stage no thought appears to have been given by the BBA to the fact that it might perform in different ways depending on the nature of the fabrication and the system of fixing adopted, even though separate test evidence was provided on the structural integrity of the cassette and rivet versions.⁷¹⁷¹ {MET00053158_P13/173}; Gregorian {Day105/59:3-20}. Nor does anyone appear to have attached any importance to the fact that it was not possible to test the product's reaction to fire in an unfabricated form in accordance with EN 13823, which would be necessary to obtain a European standard classification.

¶18.12 Mr Albon said that it was for the client to determine precisely what would be assessed as part of the certification process and that the BBA would not resist that unless it had good reason to do so.⁷²⁷² Albon {Day109/127:25}-{Day109/128:10}; {Day109/128:24-}-{Day109:129:2}; See also assessment flowchart which includes a step of "Suitability of product and use considered" {BBA00008042/17}. Hamo Gregorian, a Project Manager in the BBA's engineering systems department who managed the initial assessment of Reynobond 55 and attended the meetings with Arconic in November 2006 and February 2007, told us that the BBA's policy was generally to approve a product, such as a cladding panel, rather than a system, because the BBA could exercise more control over the quality of a product than a system.⁷³⁷³ Gregorian {Day105/48:5-18}; {Day105/57:16-23}. Mr Gregorian's evidence, which we accept, was that assessment of fire performance for certification was not his area of expertise and that all such matters were checked by a fire expert or agreed in consultation with the BBA's Technical Manager, Brian Haynes, who had relevant experience and knowledge.⁷⁴⁷⁴ Gregorian {BBA00011096/3} page 3, paragraph 10; {Day105/12:20-25}; {Day105/23:9-19}; {Day105/69:4-16}. At all events, the potential effect of the different configurations on the fire performance of the product was overlooked by the BBA at that early stage in the certification process.

Arconic's contract with the BBA: March 2007

¶18.13 On 26 February 2007 Arconic submitted a fresh application to the BBA for certification of Reynobond 55 that was largely based on the form used in August 2006.⁷⁵⁷⁵ {BBA00008042/13-35}. In section 3 of the form Arconic was expected to set out all data relevant to the product and its use produced during the previous three years. In relation to fire performance it referred, as it had on the previous occasion, only to the results of the tests carried out by Warrington Fire in accordance with BS 476-6 and 476-7 on Reynobond 55 with a fire-resistant core.⁷⁶⁷⁶ Tests numbered 132316 and 132317 {BBA00008042/29}. It made no mention of the results of any test on Reynobond with an unmodified polyethylene core. Nor did it refer to the results of any fire test which showed that there might be a difference in performance between the cassette and riveted forms of the product. However, the tests on Reynobond with a polyethylene core, in both riveted and cassette forms, that had been carried out in December 2004 fell within the three-year period contemplated by the form, with reports issued by CSTB in January 2005.⁷⁷⁷⁷ {ARC00000535}; {ARC00000358}; {ARC00000536}. Those omissions can only have been deliberate.

¶18.14 On 23 March 2007 the BBA entered into a contract with Arconic to provide certification for the product.⁷⁸⁷⁸ BBA project number S3/41014 {BBA00008042/2-5}. In the final version of the contract handwritten amendments were made to the name of the product substituting "Reynobond Architecture" for the original reference to "Reynobond 55". The description of the product was also changed from the original "Aluminium/polyethylene composite wall cladding system" to "Aluminium/polyethylene composite wall cladding panels" (our emphasis). That amendment reflected the agreement with Arconic that the certificate would refer to the product without identifying the system of fixing to be used.⁷⁹⁷⁹ Gregorian {Day105/64:2-21}.

¶18.15 Under the BBA's standard terms and conditions Arconic was obliged to disclose to the BBA full particulars of the product, including particulars of any test data already available and the test procedures used to obtain them.⁸⁰⁸⁰ {BBA00008042/3} clause 7(a). Arconic was also obliged to notify the BBA immediately of any change in the particulars supplied to it and any new or additional information it obtained concerning the product or its suitability for use.⁸¹⁸¹ {BBA00008042/3} clause 7(g).

¶18.16 Under its standard terms the BBA had the power to withdraw or suspend any certificate if Arconic breached the contract terms, if there were any change in the technical specification of Reynobond, if any information became available that was not at the disposal of the BBA at the time the certificate was issued or if there were any change in its description, including its physical or chemical composition.⁸²⁸² BBA Standard terms clause 13 (a)-(c) {BBA00008042/4}.

Arconic's further provision of test data to the BBA: May to November 2007

¶18.17 On 15 May 2007 Hamo Gregorian asked Claude Wehrle to provide reaction to fire data for the standard PE panel, by which he meant the version with an unmodified polyethylene core, and warned him that a French national classification (as used in the CSTB Technical Assessments) was not recognised in the UK.⁸³⁸³ Gregorian {Day105/95:5-10}. Mr Gregorian said that if such evidence was not available Arconic would need to arrange for further testing to be carried out. Mr Gregorian noted that fire test reports for the product with a fire-resistant core had been submitted with the application.⁸⁴⁸⁴ Mr Gregorian's reference to FR product tests in this correspondence is likely to be to Warrington's BS 476-6 and BS476-7 testing reports from September 2003, numbers 132316 {BBA00008042/163-175} and 132317 {BBA00008042/177-191} which were cited in each of Arconic's applications to the BBA and probably appended to them. However, Mr Gregorian did not ask for separate test evidence relating to the cassette form of the standard product and appears to have been content to accept the results of tests on the riveted form as applicable generally, without making it clear that they had been obtained only from tests on that particular form of the product. That was despite the fact that the classification report issued by CSTB stated that the European classification B-s2, d0 applied only to the riveted form.⁸⁵⁸⁵ {ARC00000358/4}.

¶18.18 Mr Gregorian told us that he regarded the manner of fixing as irrelevant because the BBA was assessing only the product, by which he meant the product as it emerged from the factory.⁸⁶⁸⁶ Gregorian {Day105/98:20}-{Day105/99:2}. However, that was a fundamentally unsound approach, and not consistent with what the certificate stated. The BBA should have appreciated not only that the fire performance of the product could be tested only in one or other of the fixing configurations and might differ depending on which was adopted, but also that the proposed certificate would be making statements about fire performance which covered both fixing systems. In our view the BBA should not have been willing to give a certificate for the product in its unfabricated form because, as it was aware, that did not represent the condition in which it was designed to be used. As a result, a certificate that did not take into account the method of fixing was likely to be misleading in some, if not all, cases. The BBA should have been aware when it entered into the contract that the fire performance of the product might be affected by the method of fixing adopted, but if it did not, it should have become aware of that by the time it issued the certificate.

¶18.19 On 25 May 2007 Mr Wehrle sent Mr Gregorian a copy of the results of Test 5A.⁸⁷⁸⁷ {MET00053158_P15/90}. He did not provide the BBA with any test data relating to Reynobond 55 PE in cassette form, let alone the results of Test 5B.

¶18.20 Mr Wehrle sought to justify the decision not to provide the BBA with test data relating to Reynobond 55 PE in cassette form on the grounds that Test 5B related to a single sample and was not a valid classification report.⁸⁸⁸⁸ Wehrle {MET00053190/48} page 48, paragraph 177. However, Arconic's contract with the BBA required it to disclose particulars of any test data available.⁸⁹⁸⁹ BBA Standard terms clause 7(a) {BBA00008042/3}. Mr Wehrle also suggested that Arconic relied on the BBA to ask for any additional information it required,⁹⁰⁹⁰ Wehrle {MET00053190/17} page 17, paragraph 59. but that is no justification for ignoring the terms of the contract. In any event, we regard Mr Wehrle's suggestion as disingenuous. First, in his email of 15 May 2007 Mr Gregorian had actually asked for reaction to fire data for the standard PE panel, without distinction between riveted and cassette forms. Secondly, Arconic had not told the BBA that Arconic had tested Reynobond 55 PE in cassette form and it follows that the BBA was not aware that any such data existed. Mr Wehrle could not have understood Mr Gregorian's request as limited to test data for the riveted form. Thirdly, Mr Wehrle described the results of Test 5A that he attached to the email he sent Mr Gregorian on 25 May 2007 as "the fire reaction certificate for our product Reynobond PE", implying that there was only one test result. That was deliberately misleading. The reality is that Mr Wehrle withheld Test 5B from the BBA, despite its obvious importance both to the BBA and to the UK market, because it showed that the fire performance of the product when used in cassette form was very significantly worse than when used in the riveted form covered by Test 5A.

¶18.21 If Mr Wehrle had given evidence in person he would have had an opportunity to respond to the suggestion that he deliberately and dishonestly misled the BBA into thinking that Arconic held no results of any test of the product in cassette form. As it is, we have not had the benefit of hearing his response. Nonetheless, the evidence we have seen, which includes his statement, leaves us in no doubt that that is what he did. Our conclusion is reinforced by the fact that in none of its lengthy written or oral statements to the Inquiry did Arconic seek to suggest otherwise.

¶18.22 On 29 November 2007, Mr Gregorian asked Mr Wehrle to provide test data in respect of the fire performance of the back face of Reynobond 55 PE panels.⁹¹⁹¹ {MET00053158_P16/134}. In response, Mr Wehrle sent Mr Gregorian a report from CSTB classifying Reynolux (an aluminium sheet coated with a polyester finishing coat) under EN 13501 as Class A1.⁹²⁹² {BBA00008042/139}. He justified that on the basis that similar coatings were applied to the front and back skins of Reynobond 55 PE.⁹³⁹³ {MET00053158_P16/136}. It is not clear why Mr Wehrle sought to support an application for a BBA certificate concerning Reynobond 55 PE, an ACM panel product, with test data relating to an entirely different product which was not ACM. It is possible that Arconic thought that, because BS 476-6 is the method for testing the fire propagation index of the surface of a product, there was a useful analogy between the test on Duragloss-coated aluminium and Reynobond 55 PE, or that it simply left it to the BBA to ask for further information. It is equally possible that that was another step in the deliberate concealment from the BBA of relevant fire performance test data with a view to obtaining a useful certificate from the BBA for Reynobond PE 55. We cannot reach a firm conclusion on that point, given the possibility of an innocent explanation, but it is another matter on which Mr Wehrle's assistance would have been helpful if he had chosen to give evidence.

Drafting and issue of the BBA certificate: October 2007 to January 2008

¶18.23 By September 2007, Arconic, and Mr Wehrle himself, were well aware that many in the construction industry were seriously concerned about the fire safety of ACM with an unmodified polyethylene core. That knowledge ought to have led Arconic to ensure that the text of the BBA certificate was clear, accurate and complete.

¶18.24 Several drafts of the certificate were produced before it was issued. On 22 October 2007, Mr Gregorian provided Mr Wehrle with what was described as a "working draft" and invited Arconic to comment on its contents.⁹⁴⁹⁴ {MET00053158_P15/188}. He specifically asked Arconic to provide any information that was missing but would be helpful to the user and any proposed amendments to information that Arconic considered to be inaccurate. Claude Wehrle reviewed the certificate at least three times and asked Hamo Gregorian to include some specific wording.⁹⁵⁹⁵ Claude Wehrle reviewed drafts of the certificate in October 2007 {BBA00008042/535}; November 2007 {BBA00008042/515-517} and December 2007 {BBA00008042/465}. On each occasion, Mr Gregorian accepted Mr Wehrle's proposed text.⁹⁶⁹⁶ Hamo Gregorian accepted Arconic's comments in November 2007 {BBA00008042/529} and {BBA00008042/511} and in December 2007 {BBA00008042/459-61}, {MET00053158_P16/149}.

¶18.25 On 7 January 2008, shortly before the certificate was issued, Mr Wehrle also circulated a draft to Colin Southgate and Deborah French, Arconic's UK sales representatives for Reynobond at the time.⁹⁷⁹⁷ {MET00053158_P16/167}; {MET00053158_P16/165}. Mr Southgate proposed further text which was also incorporated into the certificate.⁹⁸⁹⁸ {MET00053158_P16/167}; {BBA00008042/463}.

¶18.26 The BBA certificate was issued on 14 January 2008. We consider its contents in detail in Chapter 19. For present purposes it is enough to note two features: first, that it declared that "the panels may be regarded as having a Class 0 surface in England and Wales";⁹⁹⁹⁹ {BBA00000047/3}. secondly, that it expressly covered the product when used in both forms, riveted and cassette, without distinction or qualification.¹⁰⁰¹⁰⁰ {BBA00000047/2} Figure 1.