COURT OF APPEAL FOR ONTARIO
DATE: 2003-05-28 DOCKET: C36834, C36835
BEFORE: CARTHY, CHARRON AND ROSENBERG JJ.A.
B E T W E E N:
ORLANDO CORPORATION Plaintiff/Appellant (Respondent by way of cross-appeal)
- and -
BOTHWELL-ACCURATE CO. LTD. And DOMTAR INC. Defendants/Respondent (Appellant by way of cross-appeal)
A N D B E T W E E N: (C36835)
ORLANDO CORPORATION Plaintiff/Appellant (Respondent by way of cross-appeal)
- and -
DUFFERIN ROOFING LIMITED And DOMTAR INC. Defendants/Respondent (Appellant by way of cross-appeal)
COUNSEL: Martin Teplitsky, Q.C. and Bradley Teplitsky, for the appellant, respondent by way of cross-appeal John W. Adams, Q.C., Kate Broer and Jason Annibale, for the respondent Domtar Inc., appellant by way of cross-appeal
HEARD: March 19, 2003
On appeal from the judgment of Justice Lee K. Ferrier of the Superior Court of Justice dated May 23, 2001, reported at [2001] O.J. No. 1946.
CARTHY J.A.:
[1] There are two appeals before us from judgments in two actions that were tried together in which the plaintiff, Orlando, sought damages from the defendant, Domtar, for the cost of replacing corroded steel roofs on two warehouse buildings. The claims against the co-defendants who installed the roofs were dropped prior to trial. After a lengthy trial, Ferrier J. found both parties to be negligent, and assigned 90% of the fault to the plaintiff. Accordingly, he awarded Orlando $53,551.60 in respect of what I will call the Coast building and $106,615.50 in respect of what I will call the Cooper building. The plaintiff appeals and the defendant cross-appeals against the judgments, each claiming essentially that the other should bear the full blame and, in the case of the appellant, that Domtar should at least bear a larger percentage of responsibility.
[2] The trial judge analyzed the evidence from a lengthy trial in meticulous and comprehensive reasons. I will attempt no more than a summary of the facts necessary to an understanding of the issues on appeal.
[3] Orlando is a developer of properties and buildings. It owns over 300 buildings, most of which are industrial and located in southern Ontario. It does its own architectural and structural design work. We are concerned with two flat-roofed warehouses constructed by Orlando in or about 1990.
[4] A typical flat-roofed building is covered by a ribbed steel deck, on top of which an insulation material is affixed. Other materials, which are irrelevant to issues in these actions, are layered on top of the insulation. The Ontario Building Code, R.R.O., Reg. 61, ss. 5.2 and 5.3 requires, in specified circumstances, a vapour barrier and an air barrier. A combined vapour-air barrier is commonly referred to as a “vapour barrier”. The vapour barrier is placed between the steel deck and the insulation. It might typically comprise a sheet of polyethylene that is sealed at all edges and perforations. Its main function is to prevent condensation on the steel deck as warm humid air inside the building is cooled by outside air.
[5] For at least ten years prior to 1990, Orlando had used fibreglass insulation. It was not its practice to install a vapour barrier over the warehouse portion of its buildings, notwithstanding the Ontario Building Code requirement.
[6] When the Cooper and Coast buildings were designed, Orlando decided to use an insulation product known as phenolic foam. Phenolic foam was manufactured and sold by Domtar, and claimed to have a higher “R value” than fibreglass.[^1] It contains a small amount of each of formic and sulfonic acids, both of which are corrosive if leached from the foam by moisture.
[7] The Canadian General Standards Board, a national standards-writing organization, published a standard for phenolic foam insulation boards that contained a note reading:
8.5 Corrosion – Phenolic foams may contain some corrosive compounds. When it is anticipated that the foam will be in direct contact with metal, the supplier shall provide the proper installation procedure.
[8] Domtar was aware that Orlando did not use vapour barriers in the warehouse portions of its buildings and did not inform Orlando of the acidic content of its product. Its sales brochure refers to compliance with the Canadian General Standards Board requirements, but makes no specific reference to corrosive compounds. The sales brochure reference to vapour barriers reads:
Vapour Retarder:
Responsibility for the inclusion of a vapour retarder within the roofing assembly rests with the design authority.
A vapour retarder performs the vital function of preventing moisture vapour within the interior atmosphere of the building from entering the roof insulation and roof assembly. Moisture entering the roofing assembly can cause insulation to lose its thermal performance and contribute to poor membrane performance.
Factors when selecting a vapour retarder should include interior relative humidity, climatic conditions, and dew point.
Thus, while the Standards Board note refers to corrosion, the brochure refers to degradation of the R value.
[9] The reader will at this point in the narrative have guessed that in the warehouse portion of each building condensation occurred above the steel deck, that acid, particularly formic acid, was dissolved in the water, and that sufficient corrosion of the steel decks ensued that, for safety reasons alone, both roofs needed replacement well before their life expectancy.
[10] No issue is taken with the preceding facts by the parties to the appeal. The contention of each of the appellant and cross-appellant is that the trial judge erred in the application of these facts and associated expert opinions in reaching his conclusion. Each says that the other is totally at fault for the damage. Alternatively, Orlando claims that the apportionment was wrong.
[11] The essence of the trial judge’s reasoned conclusion is that Orlando was at fault for failure to use a vapour barrier and Domtar was at fault for failure to warn of the corrosive potential in the use of its phenolic foam without a vapour barrier. He found the predominant cause of the corrosion of the steel deck was the water vapour and that the formic acid was a minor contributor to, perhaps only an accelerator of, the corrosion that was being caused by the water. Thus, he allocated 90 per cent responsibility to Orlando and 10 per cent to Domtar.
[12] It is my conclusion that the trial judge’s findings were justified and should not be disturbed.
Analysis of arguments
[13] Orlando says that there was no foreseeability on its part that the failure to use a vapour barrier would cause corrosion damage. Domtar did not inform it of the presence of acids or inform it of the warning note in the Canadian General Standards Board specification. Despite Domtar’s awareness of the corrosion potential, its brochure recommended a vapour barrier to protect the R value of the insulation, not the steel deck. Domtar knew Orlando did not use vapour barriers in its warehouse roofs. In effect, Domtar was dealing with a pre-existing condition. Why, it is argued, should Orlando look behind the brochure? It is also noted that Domtar made a claim in an action in the United States against its supplier based on the corrosive potential of the insulation. Additionally, subsequent to its sale to Orlando, Domtar made the use of a vapour barrier mandatory in every phenolic foam installation.
[14] This appears to be a formidable argument in favour of Orlando, but there are omissions from it which were persuasive to the trial judge.
[15] To begin, the trial judge was clearly influenced by the failure of Orlando to call witnesses to explain why vapour barriers were not used. He states:
[123] No one involved in the design of these buildings was called as a witness by the plaintiff, yet the plaintiff utilized the services of its own architects and professional engineers in the design of these buildings. Consequently, counsel were left with limited opportunity to cross-examine the plaintiff’s witnesses about Orlando’s design decision to omit a vapour barrier.
[124] In 1995, Orlando changed its design policy concerning vapour barriers and thereafter, they were installed in the warehouse or industrial part of its buildings. The reason given by Mr. Turner was that it would give the plaintiff greater flexibility in the uses to which a building could be put, and they could build “on spec” without having to secure a tenant in advance. However, it is to be noted that the plaintiff did not produce one single document on the subject, be it a memorandum explaining the need for a change in policy or a direction to those concerned that there had been a change in policy. Nor did those responsible for this alleged change in policy testify. The defence argues that an inference may be drawn from this. Counsel suggests that by 1995, having cut corners and saved money by omitting vapour barriers, it had become clear to the plaintiff that prudent design required their installation.
[16] This suggestion of cost saving is supported by Orlando’s failure to even consider vapour barriers for any preventative purpose. The trial judge found:
[125] As part of its sales approach, Domtar gave Orlando its “Roofing Systems” brochure. More will be said about it below, but it included the following about the need for vapour barriers:
Most conventionally insulated roofing assemblies require a vapour retarder. It performs the vital function of preventing moisture vapour within the interior atmosphere of the building from entering the roof insulation and roof assembly …
Responsibility for the inclusion of a vapour retarder within the roofing assembly rests with the design authority. The decision-making process should include calculation of interior relative humidity, climatic conditions and dew point.
[126] The same information was provided in Domtar’s “Rx Roof Insulation” brochure. Mr. Turner of Orlando acknowledged that the types of calculations referred to in the brochure were not done for the Coast and Cooper buildings. He also agreed that they were not difficult calculations.
[128] The plaintiff’s witness, Richard Canon, an expert in roofing and structural engineering, acknowledged that a roof designer takes condensation into consideration, calculating the extent to which there will be condensation in the roof, the concern being potential corrosion of the fasteners and the deck, and ridging of and damage to the membrane.
[129] I am satisfied that the plaintiff gave no thought to the need for vapour barriers in these roofs.
[17] The next logical step in the analysis is to determine if this casual approach to the potential for condensation of water could lead to the damage without the presence of formic acid in the insulation. The trial judge made a positive finding on the evidence before him that galvanized fasteners, which could not have been affected by acid, were corroded. He stated:
Fastener Corrosion
[141] There were fastener plates taken from the Coast roof that were corroded to the point of perforation (the plates had three or four times the amount of galvanization than that on the deck). Others were significantly corroded. It was also demonstrated that several fasteners (screws) were more heavily corroded in the area where they passed through the fibreboard than where they passed through the foam. It was also established in evidence that there were instances where the fasteners and plates were significantly corroded, but where the underlying deck was not corroded. The defence argues that this evidence (acknowledged by Dr. Slater) shows that corrosion mechanisms were at work which had nothing to do with the insulation. I agree.
[142] The defence argues that the plaintiff’s theories cannot explain how acids from the insulation could corrode fasteners above the fibreboard and yet not corrode the deck below. Dr. Slater’s theory of the “off-gassing” of formic acid could not possibly account for this. Fibreboard has been used on roofs for decades as an insulation. It has been placed directly against steel decking. However, nobody maintains that fibreboard is responsible for corroding fasteners or steel decks. Any such problem would be attributed, rightly, to the presence of water. I agree with this analysis.
[143] Mr. Canon, for the plaintiff, acknowledged in cross-examination that fasteners which he was shown had been stripped of their epoxy protection by severe corrosion, and this was observable in the part of the fastener shank that passed through the fibreboard and was above the foam. He acknowledged that this was attributable to the presence of water.
[18] Further, the evidence was clear that there was no corrosion over the office portions of the Cooper and Coastal buildings where vapour barriers were installed.
[19] The trial judge identifies other mechanisms for corrosion by water in urban atmospheric conditions:
[105] The defence submits that it would be highly unreasonable to ascribe the blame for corrosion to a low level of formates because they somehow make the water highly conductive. There was no evidence of that. However, there was evidence that there are other substances that could increase the conductivity of water on the deck. The Maxxam Laboratory Report shows the presence of fluoride, chloride and orthophosphates and the Ontario Power Generation report shows ascetic acid and chlorides. According to Dr. Rudin, these substances do not come from the insulation. It must be presumed that there are other substances in the corrosion products that have not been analyzed. Again, the onus is on the plaintiff to do this.
[106] Dr. Scepanovic testified about atmospheric corrosion. She said that in urban and industrial atmospheres, the mere absorption of carbon dioxide results in a pH of 4 or 4.5 in any water and that there is enough conductivity from that alone to cause corrosion. She said that hydrogen sulfide is “all over the place”.
[20] Moreover, subsequent to the events in issue, there was evidence of corrosion to other Domtar roofs where the insulation was fibreglass. The trial judge recounts:
[122] Domtar, through its own inquiries, discovered that in 1998 the plaintiff had replaced approximately 30,000 sq. ft. of the roof on its Norampac building in Mississauga. The building was about ten years old when the roof had to be removed and the deck scraped and painted. The cause was corrosion due to excessive condensation. The insulation on the Norampac building had been fibreglass and there had been no vapour barrier. Mr. Turner of Orlando acknowledged that there were other roofs that Orlando had to repair because of condensation problems with insulation other than phenolic foam.
[21] Ferrier J. summarizes his view as to the standard of care in the industry in these terms:
Standard of Care
[134] The plaintiff argues that the standard of care in the industry in 1990 and 1991 was to omit a vapour barrier. I disagree. While it was established that 95% of warehouses in the Toronto area did not contain vapour barriers, the literature, industry publications, building codes, and other evidence in this case make it clear that at that time, the design authority had to consider whether a vapour barrier was necessary. When doing so, the designers had to take into account not only the design elements of the building, but also the atmospheric conditions, the dew point calculations and, significantly, the use to which the building would be put. Orlando did none of this.
[22] Ferrier J.’s conclusion on the effect of moisture in the roofs was:
[321] The presence of moisture in the roof assemblies was by far the greatest contributor to the corrosion in each roof. Very large amounts of water were present in Coast over a period of ten years. The effect of the condensation was the most significant factor in the corrosion process.
[322] While the effect in Cooper appears to be of a significantly lesser degree, here too moisture and condensation were the major problems.
[323] It is clear that where a vapour barrier was installed with the insulation, there was no problem – witness the office roofs.
[324] The Coast warehouse had no vapour barrier, was heavily humidified and had significant air leakage. It was heavily corroded.
[325] The Cooper warehouse had no vapour barrier, was generally well ventilated and had obvious signs of condensation, but where the foam boards lay flat against the deck, there was virtually no corrosion. Rather, there was corrosion only at the joints where condensation had occurred.
[326] I conclude on the evidence that had Orlando installed a proper air/vapour barrier in each warehouse roof, as required by the Ontario Building Code, there would have been no corrosion.
[23] In dealing with Domtar’s liability, the trial found that it failed in its duty to warn that, if not installed with a vapour barrier, its product had the capacity to contribute to corrosion of the steel deck. Formates were found on both steel roofs, indicating that formic acid had been involved in the corrosion process. Thus, the failure to warn was causally connected to the damage. He also found that, by omitting any warning, Domtar had made a negligent misrepresentation as to the suitability of phenolic foam for use over a steel deck without a vapour barrier. He further found that Domtar knew or ought to have known, by testing and analysis, of the corrosive potential in the presence of moisture.
[24] Ferrier J. then concludes:
[421] Furthermore, as I have indicated, I accept Dr. Rudin’s analysis of the corrosive power of the foam. I agree with his opinion, without hesitation, that the foam had a very limited ability to contribute to the corrosion. However, as I have indicated, it is my view that even though the contribution was slight, the attack on the galvaneal protective cover significantly reduced the deck’s capacity to withstand the corrosive effect of moisture.
[423] In addition, Dr. Rudin’s opinion, which I accept, was that if all of the formic acid leached out of the foam, it would consume only 80 pounds of steel in a 1,000,000 pound deck.
[424] The defendant argues that if Domtar is found to have contributed to the damages, its contribution is de minimis and it raises the defence of de minimis non curat lex. I disagree that the contribution was de minimis. The deterioration of the deck’s galvaneal protective cover, caused in part by the acid in the defendant’s product, contributed materially to the corrosion of the steel. However, moisture was by far the major player in the corrosion and the fault for that is the plaintiff’s.
[25] The principal thrust of the appellant’s argument is that it was not foreseeable to it that the failure to use a vapour barrier could cause corrosion. Domtar sold its product with knowledge of its acidic content and with knowledge that Orlando did not use vapour barriers except over office premises. On Orlando’s successful prior experience with fibreglass how could it have foreseen the problem? It was pointed out that the trial judge nowhere refers to Orlando foreseeing the problem.
[26] A beginning to the answer to this argument is found in the trial judge’s observation at para. 420 of his reasons:
[420] I am satisfied that the reason the plaintiff had not had many instances of corrosion using fibreglass insulation was that the fibreglass was air and water permeable, allowing for drying of the roof assembly. The plaintiff failed to consider the effect of a rigid board foam insulation on the drying process within the roof assembly.
[27] The essence of the reasons, taken as a whole, is that moisture caused the damage, modestly accelerated by the acid. It was Orlando’s obligation to design the buildings appropriately and it knew of the potential for moisture damage, as evidenced by the use of vapour barriers over the offices. There was less humidity in the warehouse areas, but apparently there was still enough humidity to cause moisture damage. If water caused the damage, then it was appropriate for the trial judge to look to Orlando for an innocent explanation. The absence of such an explanation was properly noted. The inference the trial judge drew was that Orlando failed to consider the switch from a permeable to an impermeable insulative material, an implicit finding that it was negligent in doing so. I have no difficulty in reading a finding of foreseeability of consequential damage into that failure to consider permeability. Once it is acknowledged that moisture was the predominant cause of the damage, and was a competent cause of the damage without the involvement of acid, the party who produced that moisture may justifiably be expected to explain that the presence of moisture was not due to its negligence.
[28] In Bow Valley Husky (Bermuda) Ltd. v. Saint John Shipbuilding Ltd., [1997] 3 S.C.R. 1210 (S.C.C.) McLachlin J. set out the definition of contributory negligence at p. 1254 in these terms:
I accept the defendants’ submissions. The test for contributory negligence was summarized by Denning L.J. in Jones v. Livox Quarries Ld., [1952] 2 Q.B. 608 (C.A.), at p. 615:
Although contributory negligence does not depend on a duty of care, it does depend on foreseeability. Just as actionable negligence requires the foreseeability of harm to others, so contributory negligence requires the foreseeability of harm to oneself. A person is guilty of contributory negligence if he ought reasonably to have foreseen that, if he did not act as a reasonable, prudent man, he might be hurt himself; and in his reckonings he must take into account the possibility of others being careless.
[29] In my view, the evidence amply supports the conclusion that Orlando was not reasonable in simply assuming that an impermeable insulation board could be safely used without a vapour barrier in the face of all of the warnings (if it needed a warning) as to the effect of moisture on steel. It took no account of the possibility that Domtar might be careless. That is the essence of the trial judge’s reasoning, and I agree with his conclusion that Orlando must share responsibility.
[30] Orlando’s further argument is that if it must share blame the trial judge used the wrong approach to apportion that blame. He should have looked at degrees of fault, not the extent of the damage caused.
[31] Section 3 of the Negligence Act, R.S.O. 1990, c. N.1 reads:
In any action for damages that is founded upon the fault or negligence of the defendant if fault or negligence is found on the part of the plaintiff that contributed to the damages, the court shall apportion the damages in proportion to the degree of fault or negligence found against the parties respectively.
[32] Orlando says that the fault of Domtar in selling a product known to contain acids to a company known not to use vapour barriers, and putatively to assure a sale, with no warning, is far more egregious than any fault to be ascribed to Orlando. It was thus wrong to conclude that, since the impact of the acid was minor, the damages should be at the same level. The greater fault should be the measure of apportionment and the trial judge never gave this consideration.
[33] This argument has a superficial attraction until thought is given to a fuller description of Domtar’s fault or negligence. It was negligent in so conducting itself that if Orlando failed to use a vapour barrier it was foreseeable that its product could make a minor contribution to damage to the steel deck. By paying a minor portion of the loss, its contribution is commensurate with its negligence. It is to be remembered that there was very little formic acid content in the phenolic foam, and corrosion occurred that could not have been caused by acid in the most highly galvanized fasteners. I, therefore, put aside that ground of appeal.
[34] A final ground of appeal was to urge that the allocation of contribution to damages should be different as between the two buildings. There was evidence that at typical temperatures condensation only occurs at or above 45 degrees humidity. While the Coast building ranged around 45 degrees, the Cooper building was maintained well below that level. Thus, Orlando says, the influence of acid must have been higher in the building with less condensing moisture.
[35] The trial judge notes that the laying of the concrete floor in the Cooper building, and the consequent release of moisture as the concrete curing, may have lead to the corrosion. He also finds that both roofs required replacement for safety reasons. He then assessed damages and allocates responsibility without comment on this argument – no doubt because it was not put to him.
[36] This court should not speculate on fresh arguments dependant upon factual findings we cannot make, and I would reject this ground of appeal. In any event, it seems of dubious merit on the finding of the trial judge that condensation was produced by the curing of the concrete.
Cross-appeal
[37] The cross-appeal contests the trial judge’s finding that Domtar had a duty to warn of the acidic content of its product, citing Hollis v. Dow Corning Corp., [1995] 4 S.C.R. 634 for the proposition that the duty is restricted to inherently dangerous products, which, Domtar says, phenolic foam is not.
[38] This is a redundant argument in light of Ferrier J.’s finding of negligence independently of the warning issue. It is also answered by Domtar’s own factum in response to the appeal, reading from para. 64: “The test for contributory negligence is no different in a duty to warn case than in other cases. The task of the Court remains to determine the causes of any damage and to apportion liability amongst those who contributed to the damage.” A duty to warn may be only one facet of a general finding of negligence and does not change the formula for allocating responsibility. I would therefore dismiss the cross-appeal.
Conclusion
[39] I would dismiss the appeal and cross-appeal with costs on a partial indemnity scale to the respondent for 90 per cent of an appropriate assessment, reduced by reason of its lack of success on the cross-appeal. At the time of argument, the trial judge had not awarded trial costs. He has now done so but has left quantum open to be spoken to. We are told that there will be an appeal from that disposition. From a reading of the reasons of Ferrier J. as to costs, I see no need that any such appeal be before this panel, nor any reason to defer fixing costs of the present appeal. The respondent may have two weeks to submit an appropriate bill of costs with supporting material and the appellant may have ten days thereafter to make submissions, followed by reply from the respondent within four days.
Released: May 28, 2003 “JJC” “J.J. Carthy J.A.”
“I agree Louise Charron J.A.”
“I agree M. Rosenberg J.A.”
[^1]: “R value” is a term used to describe the measured insulating capacity of any given material.