Paints should do exactly what they say on the tin. But in the celebrated case of Bath Spa, they didn't - and the result was a public disaster for everyone involved. We report on what went wrong

Bath Spa was to have added the final touch to one of Europe’s most beautiful cities. It was to have been completed in 2002 and it was to have been a model of inspiring architecture and fine craftsmanship. Instead it became a demonstration of just how wrong things can go when things go wrong. What’s more, it still isn’t finished, and it’s at least £27m over budget. And the remarkable thing is that the whole disaster was caused by a tin of paint.

The row centres on a product called RIW Toughseal that was used to waterproof and finish the pools. This product is a solvent-free, polymer-modified epoxide coating supplied in two containers that are mixed together and then applied to the surface to be waterproofed. Simple really. Yet it is Toughseal that famously did not bond to the substrate successfully when applied as a finish to the Spa’s concrete pools. After much anguish, the Toughseal was blasted off and replaced by a different product at a cost of some £4m. But problems have surfaced elsewhere – with Toughseal taking centre stage yet again. It is implicated in leaks from the steam room and from the rooftop pool area. Tiles have been lifted, cracks discovered in the screed and the Toughseal membrane below – causing a further almighty rows. Whatever the truth, the project has been further delayed, its earliest possible opening time is not until the summer, and its original proposed cost of £13m has soared to more like £40m. So what is wrong with Bath Spa? Is it the design? Is it the paint? Or is it the workmanship?

In defence of architect Grimshaw and contractor Mowlem, waterproofing this type of building is not easy. Peter Caplehorn, technical director of architect Scott Brownrigg, has some sympathy for them: “Swimming pools are aggressive environments,” he says. “Water and high humidity are things you generally keep out of a building and, to be honest, problems are not unusual.” He adds that the project is ambitious, involving spa facilities on several levels built within a heritage site location and involving several listed buildings. He says Scott Brownrigg always goes for a belt-and-braces solution if it is accountable for any future problems with projects involving internal water.

RIW’s product literature clearly states that Toughseal is suitable for waterproofing swimming pools; and once it became apparent that it was not going to stay put at Bath, the workmanship was called into question. Sir Nicholas Grimshaw himself let it be known that there were “issues of workmanship”, and the report by STATS – a materials specialist commissioned by Bath and North East Somerset council (BANES) – also highlighted workmanship as an issue. It said the substrate had been inadequately prepared, the paint applied in overly cold conditions and in the wrong way and that it had been contaminated. Mowlem defended itself by insisting it merely had the specified paint applied in strict accordance with the maker’s instructions. Mowlem questions the specification and points out that the STATS report states that Toughseal was always likely to discolour, be inflexible and prone to shrinkage.

Swimming pools are aggressive environments … Problems are not unusual

Peter Caplehorn, technical director of Scott Brownrigg

So should this type of product have been specified in the first place? Current thinking on specifying swimming pools offers some clues. Steve Nelson, of pool provider Certikin, says: “We hardly sell any painted pools any more. The domestic market is dominated by pools that use liners, and commercial pools tend to be tiled.” Alan Powers, of industrial and marine paint specialist Leigh’s Paints, says: “The traditional pool paint for years was chlorinated rubbers. It was very effective, but went out of fashion – partly because of its high solvent content.” He adds: “We do sell a waterproof epoxy paint, but like most paint its success depends on the substrate. Problems generally arise because the water content of the concrete is too high or the surface is too powdery. We would not recommend our product for swimming pools.” Of course Powers is not talking about Toughseal, and RIW states that: “Toughseal will bond exceptionally well to concrete that has been fully prepared in accordance with recognised good practice.”

Tiles and liners, in fact, were considered as replacement finishes at Bath, but were deemed unsuitable for the size and shape of the pools concerned. Other options rejected by consultants included exposed concrete because it causes injury if you fall on it; coloured render, for the same reason; and acrylic paint because it was insufficiently proven. Reapplying Toughseal was rejected as “an unquantifiable risk”. In the end the solution that has been found is the two-part epoxy paint Krautoxin. Produced in Germany by Krautol, and applied to the pools by Dutch specialist Bovitec, this was one of the pricier options. It has stuck.

However, the problems are not over. Cracks have appeared under tiles in the steam room and the rooftop pool. Here, Toughseal has been applied as a waterproofing layer between a water-heated screed and a finish of stone tiles (see “Final design as implemented”, overleaf). E E Movement has apparently caused the screed to crack, and the Toughseal with it, allowing water to penetrate the floor. Mowlem says: “The problem is that the designed membrane needs to have sufficient elasticity to cope with any movement in the screed. RIW Toughseal was not the right product choice for these design conditions.” Mowlem is not blaming RIW for this problem, saying: “RIW’s product data sheets have never claimed that Toughseal has the required elasticity.” And RIW itself states: “Generally, coatings based on epoxy resins may bridge ‘static’ hairline cracks but should not be used to seal or bridge dynamic cracks.”

The design was changed from a perfectly good and workable one to the failed one to save money

Mowlem

If this problem is not down to the waterproof layer, is there a problem with the screed? Some suggest that the change from a conventional steel-reinforced screed to a fibre-reinforced product called Isocrete may be to blame for the larger-than-expected cracks, which are up to 2.5 mm wide. But Mowlem says the screed would have cracked either way – possibly more badly with steel mesh reinforcement – and has experts lined up to defend that position. Mowlem adds that the Toughseal has, significantly, also been found to be cracked in places where the screed is not.

So is it the design or the management of the design which is at fault? Feelings on this subject run high. Mowlem says the design of the floors was changed by agreement between BANES and Grimshaw, “from a perfectly good and workable design – with a suitable membrane on waterproof concrete and drainage to take the water away [see “Original design before changes”], to the failed design, in an attempt to save money.” A BANES spokesperson comments: “The decision to move the membrane to beneath the tiles was taken as part of a value engineering exercise. Mowlem was part of that exercise and agreed to the change.” But Mowlem denies this, insisting the change was undertaken without its involvement by Grimshaw and Arup, the project engineer: “The assertion by BANES that we agreed to moving the waterproof membrane is utter nonsense and can only have been designed to mislead.” And what about Grimshaw? It says the practice prefers not to comment, “given the likelihood of legal proceedings”.

Meanwhile, Mowlem is trying to sort out the problem and is discussing two remedial options with the client. The first would effectively replace the Toughseal layer with a layer of a product called Aquafin-2K/M – a waterproofing slurry described by manufacturer Schomburg as “a two-component flexible cementitious coating”. A Mowlem insider comments: “Aquafin is proven for use on pools, and is very flexible. It has a rubbery feel and could easily cope with movement in the screed.” The other option is to go back to the original design with the “egg-crate” drainage system (see “Original design”).

If any further design inspiration is required, the construction team might do worse than pop around the corner to the Roman baths. The underfloor heating system there is rather impressive, and the main pool’s lining, comprising thick sheets of lead, still holds water after 1700 years.