The EU Servowood Project – predicting the service life of exterior wood coatings

The external use of wood in buildings is widespread across Europe where it is valued for both structural and aesthetic reasons. Wood is a renewable resource, with 90% of wood consumed in Europe being grown in European Forests.

To enhance the appearance and extend the life-time and use of wood joinery (windows, doors, cladding etc.) it is typically coated with paint, varnish or stain. The two latter are often preferred since they enhance the appearance; however their transparent nature does not protect from the damaging effects of UV light as efficiently as opaque paints. During its lifetime coated wood needs maintenance which is costly and needs to be kept to a minimum. There is a need in Europe to maximise the use of timber as a sustainable resource and in common with all materials coated wood must extend its service life. However this raises a major challenge. The best wood coatings already last 7-10 years, and reports of even 28 years have been recorded. Coatings manufacturers constantly have to re-formulate their products in response to environmental legislation and such changes can easily affect durability. Products must be launched before they have been fully tested for long periods under all conditions which introduces commercial risk.

This leads to the strategic objectives of the Servowood project which is to develop and establish European Standards that will facilitate the prediction of service life for exterior wood coatings across different climatic zones, and to significantly improve the capability of short term laboratory tests, including “accelerated weathering” to predict behaviour under field conditions.

Jon Graystone (PRA) and Hugh Williams (BCF) have answered questions we set them to explain the project in greater detail:

Who is involved in the project?

In order to help better determine the life of exterior coated wood the three year Servowood project was agreed by the EC under its R4A (Research for SME Associations) funding scheme. The project is run by a consortium made up of SME Associations namely the European Council of the Paint, Printing Inks and Artists Colours Industry – (CEPE), Danmarks Farve og Limindustri, Federation of the European Building and Joinery Associations – (FEMIB), Cluster Construcción Sostenible (Spain), British Coatings Federation – (BCF) and British Woodworking Federation – (BWF).

These Associations request the necessary technical work through Research and Development Organisations (RTO’s) namely – (Institut Technologique FCBA, Holzforschung Austria, Paint Research Association (PRA), CATAS S.p.A. and Swiss Federal Laboratories for Materials Science and Technology – EMPA) to carry out the work with the help of a group of SME companies (George Barnsdale & Sons Ltd, Drywood Coatings B.V, Pausch Messtechnik GmbH and Overfladeteknik Maleteknisk Rådgivning ApS) – a total of 15 partners. Importantly several of the consortium are also members of CEN TC 139/WG2 – the group responsible for developing exterior coating standards for wood.

When did Servowood start and how much does it cost?

The project (FP7 Project 606576) started in January 2014 and will run until January 2017, although the final report will not be completed until a few months after. The EC funding for the project is over EUR 2.6 million with those involved committed to add over EUR 1 million from their own budgets; therefore the funding available is over EUR 3.6 million for this project.

What is service life prediction and its methodology?

Service life is often defined as a time period before replacement or repair (i.e. maintenance) is required; thus for many purposes Durability and Service Life are equivalent. Predicting service life requires defining an end-of-life condition and then relating this to the factors that have caused the progressive degradation leading to failure which in the case of Servowood are the various elements of weathering. End-of-life conditions are often described as a ‘limit state’ which in the case of an electric light bulb or a fuse can be unequivocal. This is not the case with coating failure where properties such as gloss, colour, cracking, flaking etc. change more progressively and at different relative rates. Times to failure have a large distribution of possibilities which are strongly effected by macro and micro climates and building orientation.

All Service Life Prediction (S-L-P) methods require data, and in the case of exterior weathering this may be a major factor in the product development cycle. One way of generating the necessary data is to have an extensive outdoor exposure programme preferably under service conditions. This is expensive and does not give fast feedback, particularly for new technologies and where the desired lifetime is a long one. Thus to meet the objective of shortening development cycles other approaches are used such as:-

• Artificial accelerated weathering
• Intensified natural weathering
• Extrapolation from early results
• Mechanistic studies

Interpreting the results requires a methodology which will influence both the design of the experiments and the tools of analysis. Martin and others have drawn a distinction between ‘Descriptive’, ‘Scientific’ and ‘Reliability’ methodologies. All of these combine field and laboratory exposure. ‘Descriptive’ methodologies place emphasis on simulated weathering and correlation statistics. In the ‘Scientific’ approach analytical tools are used to study mechanistic aspects and photo-degradation kinetics. ‘Reliability’ theory is well established in many different areas of service life prediction and can be either probabilistic or deterministic.

The Servowood project will concentrate on better understanding of the relationship between so called accelerated weathering and natural weathering using the established test methods EN 927-3 and EN 927-6 as a basis for comparison. The work is expected to improve the precision of these tests and develop further tests that can predict service life more accurately.

How will Servowood differ from previously reported studies?

An early deliverable of the project was to carry out an extensive survey of published work and relate this to the objectives of the Servowood project. To this end a database has been created which is available to the consortium. It is considered essential that the project should build upon what has gone before. A key element of Servowood is to develop a dose-response modelwhich will quantify the behaviour of laboratory and field exposure, treating both as equally valid. It is of course understood that it is field exposure (natural weathering) that is of prime importance to the end-user. However much published work takes a test cabinet, with fluorescent or xenon arc irradiation source, and adjusts the conditions including wetness and temperature so that the response aligns as closely as possible with natural weathering.

Conclusions are often expressed in terms of a correlation coefficient determined by regression analysis. Although relatively good correlations can sometimes be obtained they do not necessarily demonstrate a cause-effect relationship and it is frequently found that the rank-order of performance (which coating performs best, second best etc.) changes between laboratory and natural weathering, and indeed between weathering sites. The literature shows a great deal of scepticism about the ability to predict the effect on coatings of natural weathering from laboratory exposure. Furthermore natural weathering is something of a moving target as the conditions continually change. By treating the two sets of behaviours separately, rather than trying to force a correlation Servowood will try to establish the parameters on which service life prediction should be based. It will also provide information on to what extent exposure conditions can be increased (i.e. accelerated) before entering the realms of un-natural effects. (The classical illustration of this problem is to consider the outcome of incubating or boiling a fertile egg!).

So what experiments are being carried out to develop the model?

The experimental plan involves a number of elements including:-

• Preparation of test panels for both laboratory and field exposure
• Selection and preparation of coatings
• Characterisation of the panels at the macro and micro level before, during and after exposure
• Data-logging of conditions during exposure
• Systematic variation of the exposure conditions

Overall this has required the preparation of thousands of test panels which must be exposed in the correct sequence by the RTO partners.

Coating systems have been selected that are known from previous work to have a significant and differing effect on performance. Formulations have been prepared with different level and type of pigmentation (and therefore colour), different binder chemistries (e.g. alkyd and acrylic), systematic variation of the binder polymer to give different mechanical properties, presence or otherwise of UV absorbers. Coatings have been exposed at different film thicknesses, since this is known to have a major effect on performance, and on different wood species including pine, spruce, larch (flat sawn and quarter sawn), oak and meranti. The data generated will be interesting in its own right; however, the prime objective is to improve the capability of accelerating tests and field exposure to give the same predictions and the selection will provide useful data for a wide range of combinations.

How will you be able vary the exposure conditions in a systematic way?

For laboratory exposure this is relatively straightforward. Commercial “Weatherometers” are increasingly sophisticated in their options for control. There are several well-known configurations available which may use fluorescent UV lamps combined with spray or condensation cycles. Other options are to use a xenon arc system which can also be combined with water spray. Temperature can also be controlled but there are some constraints on temperature control due to the heat of the lamps. It has also been found in the past that when wood panels are subject to spray cycles they become progressively drier. The project will expose selections of coating/substrate combinations, as described above, to systematic variation of irradiance, time of wetness and temperature. This will build up a large body of data to establish the dose response relationships.

In the case of natural weathering the experimenter has less control over the conditions. Climate differs to a degree within Europe and the natural weathering test panels are exposed at the five sites of the RTO partners. It is important to have accurate and frequent data logging at the test sites so that differences in the dosage can be recorded and related to the coating system response. It is normal practise in Northern Europe to expose test panels at 45 °facing South since this will maximise the average solar irradiance.(Strictly speaking panels should be exposed at the latitude angle but by convention it is 45 °). Using a single exposure angle is a compromise and it is common experience to see significant differences in coating degradation on the North and South faces of a building, and for example between vertical and horizontal surfaces. The Servowood project will take advantage of this situation by exposing a selection of the test coatings on a multi-faceted exposure rack (MFER) with nine faces. (N, S, E, W at 45 °and 90 °, plus one horizontal face). Thus each test system exposed will generate nine sets of data and a more accurate dose-response relationship. Furthermore the exposure racks can be fitted with a heater, a water spray or a combination of both. Thus the range of weathering conditions is extended whilst still using natural sunlight.

An important point to make about the dosage variations that the natural and accelerated weathering regimes create is that they will be moderated by their interaction with the coated substrate. For example a white and brown panel will reach a different temperature side by side and it is this temperature rather than the ambient temperature which must be used in modelling. The same situation applies to the effect of solar radiation on binder degradation which require an absorption band to cause damage. It is thus an important aspect of the project to determine the ‘effective dose’ as opposed to the applied dose in accounting for the response of the coating system.

What are the major perceived benefits of the project?

The project will contribute to our understanding of the factors than cause coating degradation and also to defining the parameters that can be used for Life Cycle Analysis. There are a number of specific benefits that are of interest to the SME’s in the consortium and would be of value to the wood coating supply chain. These are best summarised in the bullet points below, which illustrate that coating manufacturers, specifiers, architects and building managers all stand to gain from this project.

• Improved Precision for durability standards
• Greater Confidence in Guarantees, Warranties, Accreditation for long term performance (reduced risk)
• Clearer guidelines on maintenance scheduling
• Understanding of how coating systems will perform in different locations (climatic zones)
• Account for within and between species influence on service life
• Protect market share of coated wood products (through longer life products)
• Increase the confidence of consumers in the performance of external wood products
• Enhance sustainability of external coated wood products (through longer life products)
• Speed up development of new products (e.g. in response to legislation)