Causes of Failure in External Render & What to do About it.
When I started writing this blog I originally assumed that would be a two part blog but as it transpires it will need to be in three parts since the scope of issues under discussion is so broad. In part three I will discuss specification and application of renders in more detail. In part one of this article I discussed the items that should be checked on site when investigation the failure of external cementitious render and before I examine potential remedial works I’d first like to discuss modes of failure.
Render can fail in a number of ways and failure need not be isolated to one particular mode of failure. It is not unusual to see more than one failure mode, particularly if you are dealing with more than one elevation of a building.
Failure modes may fall into the following categories:
Shelling or Debonding
In this particular image the render was thinly applied in one coat and the render had debonded from the whole front gable. A few very light hammer blows were all that was required for the render to fall away from the building. In this particular case the render was serving a weatherproofing function as the underlying brickwork was roughly constructed and poorly pointed. The building was suffering from rainwater ingress due to complete failure of the external render system.
When shelling or debonding occurs, the render becomes structurally unsound due to not being adequately bonded to the underlying substrate, this happens for a number of reasons.
- High suction in substrate not recognized and addressed during installation.
- Sulphate attack: Ettringite formation between render and masonry blows render bond with wall.
- Freeze/thaw action: water gets in behind render and freezes at the render/wall interface. Since water expands as it freezes the resultant hydraulic action blows the bond between the render and the substrate.
- Differential movement/expansion: Render characteristics not matched to the underlying substrate and substrate moves or expands and contracts at vastly different rates to the render. It is impossible to maintain a durable bond under these circumstances.
- Substrate poorly keyed to in preparation for the render and fully reliant on an adhesive bond, whereas a mechanical keyed bond is also required.
The render exhibits signs of cracking that will allow rainwater ingress through the cracks and consequently a risk of penetrating damp. The render may crack for a number of reasons.
- Shrinkage cracks due to render being applied in poor weather conditions.
- Chemical action: Sulphate or chloride attack.
- Differential movement/expansion: Render characteristics not matched to the underlying substrate and substrate moves or expands and contracts at vastly different rates to the render or junction detailing with incompatible materials fails to account fore differing rates of expansion.
4. Hardness and subsequent inflexibility. Cement render is essentially a large inflexible sheet that cannot accommodate movement over large areas.
5. Junction detail failure. There is a frequent failure to adequately seal junction details with a suitable mastic or sealant. Rainwater then enters at these junctions where it can freeze and cause debonding and cracking.
6. Structural cracking. If the building is affected by subsidence or other structural or impact damage then it is inevitable that this damage will be mirrored to a greater or lesser degree in the render. Render can be repaired or replaced so long as you are satisfied that the underlying cause of failure has been addressed.
Consequences of Cracking
We have something of a chicken and egg situation when it comes to cracking. Is the cracking the primary cause of render failure or has the cracking resulted from from the primary cause of failure? Either way, cracking is not just an aesthetic consideration and cracks will form a direct moisture pathway for rainwater ingress behind the render system. Once there moisture can cause direct penetrating damp and further cracking via hydraulic freeze thaw action and subsequent debonding. This process is self perpetuating and the trick is to establish whether cracked render is recoverable or whether it should be written off. This can be a very subjective assessment but it needn’t be if a cost benefit analysis is carried out to establish repair versus renewal costs. Of course, you should really only consider repair if you are satisfied that the render is compatible and correctly specified in the first place.
Surface erosion is a fairly uncommon form of failure when dealing with cementitious renders. The accompanying image shows severe erosion in an external render system that was incorrectly specified to a London Docklands commercial property that was converted for residential use. The building was circa 200 years old and constructed with lime mortar, so to replicate the breathability and the required degree of softness the installer blended an extremely weak render mix comprised almost entirely of sand with very little Portland cement added. The surface was highly friable as a result and was literally being washed away by rainfall. It was also generally saturated at depth, which was causing secondary erosion through hydraulic freeze/thaw action. The system was also causing a number of problems with internal penetrating damp. Of course what should have happened here is that the installer should have specified a lime render system but wrongly assumed that limes characteristics could be replicated in a weak OPC render mix. The entire system had to be removed and replaced with lime render.
Old and historic buildings may be constructed of softer gauged brickwork and lime mortar, they are meant to breathe and will go through seasonal wet/dry cycles as they manage moisture; this is precisely what they are meant to do and applying hard cement renders will completely interfere with this process and cause a number of unintended consequences. If buildings predate the Victorian era and originally had render applied this is likely to be a weak natural hydraulic lime or a non=hydraulic lime system; both of which allow the building to breathe and are soft enough to accommodate some small movement in the underlying substrate. Lime renders of this sort even have the ability to self heal where small cracks occur. If original render fails, which of course it will over time, then it should be replaced on a ‘like for like’ basis. If you are using a non-conservation specialist then watch them like a hawk wherever lime renders are specified because they sometimes like to sneak a bit or portland cement into the mix in the mistaken belief that this will improve the mix. In fact it will virtually nullify any benefits that would have been gained from using the correct lime mix.
In this image we were dealing with a very old property in Leicestershire that was suffering from a number of issues with internal penetrating damp. The render was applied in an Ashlar finish but was severely cracked on all elevations and testing a small area at the corner of the building confirmed our worst suspicions that a very hard OPC render mix had been used that was completely incompatible with this building. It is one of those occasions were you hope that a very poor bond has been achieved with the underlying substrate but of course it was firmly bonded and incredibly difficult to remove.
Failure to Replicate Underlying Movement Joints
It stands to reason that if the underlying substrate has inbuilt movement joints then these need to be replicated directly above in the render coat. If the render coat cannot accommodate and mirror the underlying substrate movement then it will crack. The form of cracking seen is regular horizontal or vertical cracks often seen at regular centres.
The attached image shows one of our previous investigations, a commercial high rise block in London that was renovated and converted for residential use. The render was applied for purely aesthetic reasons but failed within 12 months of being applied. The situation can be rescued by retrofitting movement joints but given access costs this will prove to be a costly oversight.
Inadequate or Absence of Sealant to Critical Junction Details
This is without doubt of the most common causes of external render failure we see. Once newly applied render is complete then it is critical that junction details are sealed with a high quality mastic and that will mean attending to window and door frames, boiler flues, soffit to wall junctions, pipe and cable penetrations and basically anything else that forms a junction with the finished render. If this is not done then the render system will take in rainwater from pretty much day one and premature failure is guaranteed.
We are even seeing million pound external wall insulation schemes to high rise blocks that have had no sealant applied whatsoever to any of the critical building junctions and why would you risk having a £1m scheme fail all for the want of some sealant? I was shocked recently to find that a bricklayer working on one of my projects had never used a sealant gun and when I did ask him to seal around windows etc, he made a complete hash of it and it all had to be done again. Plasterers & renderers often see sealant application as work to be completed by others though rarely will they make a point of asking the client to ensure that everything is sealed once the render is dry.
Even choosing the correct sealant is fraught with pitfalls and I recently noticed that a client of mine specified low modulus silicone for everything and whilst low modulus silicone is the best choice for sealing UPVc windows and doors, there are better products applicable to the wide range of situations that you will encounter. I will no doubt write a blog on sealants in the not too distant future because the scope is too broad to include here. Even where sealant is applied it is not uncommon to see adhesive failure of the sealant due to wrong material selection or simply that it’s been applied to a poorly prepared or dirty surface.
Improper Flashing Installation at Critical Junction Details
Those who apply cementitious renders are often confused as to how flashing details should be dealt with. Where flashing details pre-exist for low level roofing then it is best to stop the render short of those flashing details and provide a bell cast drip detail running parallel and just above those flashings. In this image we can see where original stepped flashings were removed so that the render could be extended down to meet the roofline. Once the render was dry a channel was cut into the render and an apron, as opposed to the correct stepped flashing, was then installed. Lead has a high coefficient of expansion and therefore can crack render. Moreover when you consider that lead needs to be pegged to firmly secure it in place then how do you secure the lead when any attempts at pegging it would undoubtedly lead to cracking of the render?
External cementitious render can come under chemical attack, the most common of which is sulphate attack. This occurs when the tricalcium aluminate present in ordinary portland cement reacts with any sulphates present to form ettringite. Ettringite formation is an expansive reaction so it can cause cracking, bulging or debonding in the render. Sulphates may be present for a number of reasons, from traffic pollution, contamination of the render mix or most likely the sulphates are present in the underlying masonry. The reaction is expedited where permanent or intermittent saturation of the render occurs so failure to deal quickly with water ingress can lead to sulphate attack writing off the render system.
In part 3, I’ll deal with correct specification of cementitious renders.