Maintenance of Painted Steel-sheet Roofs on Historical Buildings in Sweden

ABSTRACT

This work presents a model and tools with which stakeholders and others involved in the conservation and maintenance process to make informed, evidence-based decisions for the treatment of painted steel-sheet roofs on historical buildings. The aim is to bridge the communication gap between research and practice and to provide useful tools for balancing technical, economic, environmental and historical values. The presented research is based on research and field experiences from practice in Sweden, and focuses on the different stages of building conservation: anamnesis, diagnosis, therapy and control/monitoring. The results are exemplified by systematic descriptions of material characteristics, anticorrosive treatment procedures and quality control checkpoints. A material and method matrix is proposed that can be used in diagnosis, planning and documentation of general procedures, or which can be refined for specific object procedures. It could be used for comparing cases, for suggesting areas of further work or for setting minimum levels for documentation. The matrix can be connected to different quality standards of conservation. It may favour the communication and systematic assessment of working procedures for different substrates, and the further development of best practices.

KEYWORDS:

• Anticorrosive paints
• guidelines
• historical buildings
• maintenance
• quality control
• roofing
• steel-sheet

1. Introduction

Roofing protects a building as a whole and is thus a strategic element to consider for preservation. Steel-sheet roofing is particularly vulnerable to atmospheric corrosion and is therefore coated by metallisations or anticorrosive paint-systems. The painted surfaces still suffer degradation from electromagnetic radiation, moisture, contamination and temperature variations (Berdahl et al. 2008; Sørensen et al. 2009).

Painted steel-sheet roofs have been used for hundreds of years in Sweden and are common on historical buildings (see Figure 1 for an example). The Swedish Church and National Property Board estimate their annual need for painting maintenance of steel-sheet roofings on listed properties to be at least 600,000 m² at a cost of about 12 million euros (Johansson 2012). Over the last few decades, stakeholders in Sweden have experienced costly problems with the maintenance of painted steel-sheet roofing, caused by poor choice or incompatible combinations, of materials and/or methods.

Figure 1. A hot-rolled steel-sheet roofing from the late nineteenth century. Photo by authors.

The steel-sheets and paint-systems available for renovation and maintenance today are extensive, but not always compatible with the materials used for historical roofs. Complete information about the composition or qualities of modern products in relation to their utility on historical roofs are rarely available. Pretreatment methods could be suitable for one type of steel-sheets, but may cause damages to other types of metal roofings.

The research presented in this work derives from need to produce guidelines to make better conservation decisions (research reports and publications by Källbom 2014, 2018). The trustees and planners have difficulties in translating scientific evidence into practice, and are reluctant to deviate from contractor’s or paint suppliers´ recommendations in the treatment and safeguarding of authentic material and traditional methods. Data collected from the research project interviews, case studies and analysing of documentation reports show that it is common to totally remove older paint layers before renovation and to replace old sheets with modern cold-rolled sheets. The arguments for extensive rather than minimal intervention cite contractors’ guarantees (they prefer to totally remove older paint layers) and high costs for scaffolding and project start-ups. The result may be durable in a technical sense but destructive for the roof’s heritage and a sustainable circular economy.

There is a need to connect scientific evidence on corrosion and material durability to the different types of historical steel-sheets, paints and working methods. A historical building and its steel-sheet roofing can be a source of tangible and intangible knowledge, such as aspects of craftsmanship, material history and characteristics, and evidence for scientific development.

Conservation and minimal intervention reflect the cultural heritage sectors’ principles for conservations with impact on cultural heritage (ICOMOS 1994, 2003, 2019) and the sustainable development goals for a circular economy (UN 2015). In practice, it is important to consider the range of historical materials and methods and to balance the maintenance work with technical, economic, environmental and cultural historical values.

This article presents a model and tools with which to guide the stakeholders in making relevant, systematic evidence-based decisions about the maintenance and restoration of steel-sheet roofs on historical buildings. The aim is to bridge the communication gap between research and practice and to provide a useful tool to balance technical, economic, environmental, and historical values. The focus of the article is on tools for how to determine the properties of materials, make relevant working procedures and how to control the results. The study focuses on common methods that are used for historical buildings as elicited by the stakeholder interviews and case studies. The aim is not to introduce new methods (such as thermal spraying, etc.), but to use existing methods in better ways.

2. Research gap and methodology

The costs for corrosion damages are high on a global scale and therefore the mechanisms and effects of atmospheric corrosion on metals have been thoroughly analysed and described (Cramer and Covino 2003; Kallias, Imam, and Chryssanthopoulos 2017; Leygraf et al. 2016; Morcillo et al. 2019a, 2019b; Odnevall 1994; Tidblad 2013). Research describes the impact of corrosion on historical metals, and also on synthesising instructions for conservation of metal artifacts, sculptures and architectural elements (Dillmann et al. 2013; Faltermeier 2014; Godfraind, Pender, and Martin 2012; Watkinson 2013). In contrast, the research on historical steel-sheet roofing is scarce and the instructions are often brief and generalising. Steel-sheet roofing is a common sub-section in academic monographic works on historical building conservation, but the complexity of the historical steel-sheets, anticorrosive paints, and working procedures has not yet been thoroughly investigated (Godfraind, Pender, and Martin 2012, 2013; Look and Waite 1980; Peterson 1968; RAÄ 1979; Sweetser 2004). Nonetheless, many materials are available for the treatment of steel roofings, and are used in practice.

Modern anticorrosive paint-systems consist mainly of epoxies and polyurethanes, and their properties and mechanisms of degradation have been the objects of extensive research (Bayliss and Deacon 2002; Brock, Groteklaes, and  Mischke 2002; Forsgren 2006; Sørensen et al. 2009; Tator 2015a; 2015b; Talbert 2008; van Eijsberg 2015). Outdoor painting failure and defect analysis are described by Weldon (2009); Fragata et al. (2006); Tator (2015a, 2015b); Greenfield and Scantlebury (2000). Amongst several general factors causing degradation, electromagnetic radiation (especially UV-light), temperature variations and penetrating moisture causing accelerated brittleness or loss of adherence, have high impacts on the lifetime of paint-systems. Synergetic effects cause oxidation of the binders, embrittlement, and corrosion on the substrate. Despite extensive research on anticorrosive paints, the use of linseed oil (a common authentic paint binder) is only mentioned as a minor ingredient in different types of alkyds or acrylate today (Araujo et al. 2010; Behzadnasab et al. 2017; Kalita et al. 2018). The conclusion reached from the above-mentioned references is that much paint research on treatments of metals does not recognise the value the importance of preserving authentic paints. There is existing research on historical anticorrosive painting systems, but is limited in terms of the complexity of conservation (Dillmann et al. 2013; Godfraind, Pender, and Martin 2012, 2013; Watkinson 2013). Research from the previous century on the outdoor use of linseed oil paints is still viable (Burgener and Carter 1950, 1953; Hudson and Stanners 1955; IVA 1935, 1961; Mayne 1970). An exceptional contribution is made by Karlsdotter Lyckman (2005). With her study of historical oil paints in architecture and conservation. Similar descriptions of how the linseed oil paints were made redundant by the introduction of modern paints at some point around the time of WWII have also been published by Standeven (2011).

The main research gap in the literature is the intersection between studies of historical steel-sheet roofing, the painting system used on the roofing, and the associated craft involved in roof treatments. However, the problem cannot solely be described as a research-to-practice problem. Deficiencies in the practice of interpreting and assimilating evidence, are part of a more complex reality. Conservation practice reflects conditions, such as economy, capability and cultural significance (Balazs and Morello-Frosch 2013; Dillon et al. 2014). This complex mission has been referred to as “management of change” (Fielden and Jokilehto 1998). Today, there is extensive and beneficial research on management and maintenance models adapted for historical buildings (Achig-Balarezo et al. 2017; Boniotti et al. 2019; Balen and Vandesande 2018; Balen and Vandesande 2019). Still, mainstream facility management and contemporary frameworks favour substitution and renovation before repair and continuous maintenance, and may sacrifice cultural significance as secondary to durability or short-term profit (Dann and Cantell 2005; Orbaşl 2017). The aim of management when dealing with ordinary buildings is to “retain the continuity of function” while the management of historical buildings also requires attention to the associated cultural significance. We consider historical roofings to be sources of knowledge of intangible aspects of craftsmanship and material history. Wrought or hot-rolled, for example, cannot be replaced, since the production methods which would have been utilised to create them are not in use today. The same can be said for plane sheets; the hot-dipping galvanising that was performed for each individual sheet (RAÄ 1979; Sahlin 1934). These materials are not only historically significant, but they also have a long life-span and continuity in function. Preservation of historical materials, including authentic paints, and the knowledge and skills associated with them may also be a vital contribution for sustainable development and a maintenance-based circular economy.

Aside from the state-of-the-art survey, the empirical material for the research consists of the assessment of 75 case studies of anticorrosive treatments of steel-sheet roofings of listed historical buildings, of which 15 have been presented by Källbom (2018). The analysis is performed with respect to the conservation process, the materials and methods used, and documentation. The cases highlight real-life situations and problem, and review different stakeholders’ perspectives. The age of the buildings’ spans from the early eighteenth century to the mid-twentieth century, which is representative among the group of listed buildings. The types of buildings studied are those where steel-sheet roofs have been common: churches, public houses, manors and residential buildings and technical buildings. The documentation from the cases has been analysed together with 55 interviews of stakeholders. In addition, five workshops were held for a smaller group of experts, each with different roles, to discuss the results of the study and development of good practices. In addition to the case studies, interviews and instructions from steel-sheet and paint-system suppliers have been critically reviewed and evaluated. Twenty collected steel-sheets, manufactured from 1787 to 2017, have been characterised metallographically (Nilsson and Pettersson 2017). The research delimits the use of materials and treatments that are common and recommended for historical buildings, as confirmed by the stakeholders and presented in the cases. New methods, like thermal spraying or paint-systems with binders such as epoxies and polyurethanes, are therefore excluded.

The structure of the article follows the building conservation process, where the general steps anamnesis, diagnosis, therapy and control are used (ICOMOS 2003; Van Balen and Vandesande 2018). The initial stakeholder analysis is outlined in accordance with the stakeholders’ general selective criteria for technical, environmental, economic, and historical qualities concerning steel-sheet roofing.

3. The stakeholders’ perspective on painting maintenance of historical steel-sheet roofings

The involvement of the stakeholders is recognised as fundamental in the conservation of historical buildings (Della Torre 2013; Van Roy 2018). The first step in the research on historical steel-sheet roofs, therefore, was to establish a dialogue with stakeholders, including material producers, trustees, consultants, entrepreneurs, and craft practitioners. The stakeholders may have different professional scopes and different roles in the conservation process. Their experience and input have been collected through interviews connected to the case studies and workshops, and are comprehended below (Källbom 2018).

3.1. Technical qualities

The painting system should provide an effective electrochemical isolation, adhere well to the substrate, penetrate laps and crevices and produce an adhering, dense, hydrophobic, physical barrier able to withstand atmospheric attacks (moisture with dissolved gases, mechanical impact from hail and snow, UV exposure and photochemical degradation of binders or pigments, heat exposure) and movements of the substrate – without cracking or peeling. Electrochemically inhibitive and barrier-forming primers are beneficial. The total thickness of the painting system should be sufficient to create a barrier that may depend on the chemical aggressiveness of the geographical location. The painting system must be highly reproducible, and robust enough to endure the short Nordic painting season. The painting systems should be conducive to maintenance in a reasonable way, without the need for a complete renovation. Degradation in the form of erosion is preferable to flaking that can expose the metal substrate. In terms of technical values, the existing steel-sheet roofings, and the wrought and hot-rolled steel-sheets are thicker and may withstand atmospheric corrosion better than thinner sheets. Wrough and hot-rolled sheets are also individually hot-dipped, which produces thicker zinc layers of high quality.

3.2. Environmental qualities

The painting system should not produce unacceptable emissions of metal ions into the environment, nor pose a health risk during the treatments. Maintenance measures should have as few environmental impacts as possible; leftover paint, remainders, and packaging material should be collected and recycled. Renewable raw materials should be prioritised. Use of azo- or phthalo pigments should be prohibited. The use of volatile organic compounds (VOCs) should be minimised or excluded. Environmental values include the conservation of existing steel-sheets, and local repairing is preferred before exchange according to ICOMOS (1994, 2003).

3.3. Economical qualities

The painting system should provide long maintenance intervals to avoid the necessity for expensive scaffolding. Alternatively, short continuous touch-up intervals could be economical, depending on roofing type and accessibility. Continuous maintenance and preventive actions are beneficial for the economy and for the paint-system lifetime, as well as for the substrate. A conscientious and strategic approach is essential for economic success and sustainability. Economical values also include the conservation of existing steel-sheets, and local repair is preferred before exchange according to ICOMOS (1994, 2003).

3.4. Historical qualities

Historical qualities must be evaluated in the case of each particular building. The painting system may not be allowed to jeopardise the future or values of the original steel-sheets or paint work. The existing materials themselves are historical sources that may also hold information of scientific value. Historical buildings should be maintained using knowledge of the original material and craft. A great effort should be placed on the aesthetic appearance, such as to use historical colours, preferably with original pigments and binders. Sheets should only be replaced when necessary. It is important to maintain high standards in craftsmanship, material authenticity and documentation for older sheet types than for modern ones.

4. Anamnesis: the history of painted steel-sheet roofings in Sweden

The anamnesis, or case history, is particular to the historical building; nevertheless, the unique biography of the building often relates to a more general history of the building materials. At this stage, the aim is to diagnose the condition of the roofing and plan for relevant treatments. Laboratory analysis of materials and surfaces can provide in-depth information, that may complement in situ methods that are highly desirable by professionals in the conservation. Ocular examination and measurements combined with historical knowledge are, in most cases, sufficient to identify the type and quality of the steel-sheet roofing. Visual examinations and simple empirical tests may be used to determine the type of binder in existing paint. It is also possible to use rapid laboratory methods, such as Fourier Transformation Infrared Spectroscopy, to determine the paint type.

4.1. Characteristics of metallised or non-metallised steel-sheets for roofing

In Sweden, ferrous roofing was exclusive and rare until late in the eighteenth century. The earliest steel-sheet roofing was wrought (and later hot-rolled), and known as black sheet, because of the black magnetite spinel surfaces that resulted from the manufacturing processes (RAÄ 1979). The wrought sheets were made of iron from small blast furnaces and reduced in hearths. The formats were small due to the character of the production method. The sheets could occasionally be metallised with hot-dipped zinc or tin for cathodic or anodic protection of the ferrous material.

In the early nineteenth century, hot-rolled plates were used, with the hot-dip zinc processes introduced in the 1860s-1870s substantially improving corrosion resistance. Corrugated hot-rolled steel was produced in Sweden from 1884. Hot-rolled and individually hot-dipped steel-sheets were common in Sweden from around 1850 to the 1960s. After the 1960s, cold-rolled sheets with thinner metallisation became normative (Källbom 2018). These are still used today with or without factory pre-coatings. Typical characteristics of sheet roofing for historical buildings are shown in Table 1.

Table 1. Typical characteristics of low-carbon ferrous sheet roofing for historical buildings

Sheet characteristics	1700~1850	1850~1920	1920~1960	1960 –
Format	45x59 cm	60x120 cm	60x ≤120 cm	∞ - length of roof
Thickness	1–1,1 mm	0,9–0,7 mm	0,9–0,7 mm	0,7–0,6 mm
Production method	Wrought	Hot-rolled	Hot or cold-rolled	Cold-rolled
Laps	Single	Single	Double	Double
Metallisation	None or individually hot-dipped sheet	None or individually hot-dipped sheet	Individually hot-dipped sheet	Electroplated or continuous hot-dipped coating of coil

The assessment of projects and interviews with restorers shows that roofing with wrought black plates is extremely rare today. Cases where it still exists (usually churches, manors or fortifications) are typically governed by restrictions and controls. It is also notable that these old wrought roofs have usually resisted corrosion and deterioration due to the larger thickness of the sheets, the high quality of the metal (the almost pure iron, resulting in ferritic microstructure instead of a duplex structure), and the characteristics of anisotropic bands of certain slags (Nilsson and Pettersson 2017).

Until recently, metallised steel-sheet roofings were allowed to weather before painting to obtain a chemically stable zinc patina, i.e. an insoluble zinc carbonate hydroxide (Odnevall 1994). This also provided a rougher surface profile that was more suitable for painting. Around 2010, the chromated zinc transportation layer was changed by the steel manufacturers, from hexavalent chromium to the more stable trivalent chromate, due to environmental concerns. The chromate type exchange prolongs the time required for patina formation on new sheets. Today, painting is also performed on electrochemically active zinc.

Eroded paint and corrosion pittings may be the cause of restoration and replacement of steel-sheet roofings to modern steel-sheets. Early twentieth-century corrugated steel-sheets and the hot-rolled steel-sheets manufactured up until the 1960s are next to be replaced, often on the basis of incomplete anamnesis and diagnosis.

4.2. Anticorrosive paint types in Sweden

Historically, anticorrosive paints were produced with linseed oil as a binder. The constituents, refining procedures, and qualities of linseed oils have varied over time. Until about the 1920s, purified linseed oil was heated and held to temperatures close to boiling point (c. 280°C) with a small addition of litharge (Karlsdotter Lyckman 2005). The procedure of heating close to boiling point and the addition of basic pigments is vital to the quality of the paint. The paint films is high gloss, density, strength and ductility, as well as superior penetration and adhering properties – all appropriate features for paintings with good weathering resistance. Sometimes the paint could be diluted with balm oil, which also interacted in the auto-oxidative polymerisation process when the paint dried. Balm oil was expensive and thus rarely used before the early 20th century. The use of solvents was not recommended for anticorrosive treatments since this may degrade the paint films faster (Paulson 1937). Linseed oil must be painted in several thin layers to allow oxidation and polymerisation through the entire layer quicker, rather than a few thicker layers.

Standoils were produced in the 1930s by boiling linseed oil under a vacuum (Karlsdotter Lyckman 2005). Chinese tung oil also came into use, regularly used as one-quarter tung oil and three-quarters of linseed oil (IVA 1935). The tung oil has a high capacity for chemical crosslinking when drying, providing higher ductility combined with hardness, resistance against atmospheric degradation and the ability to withstand both acidic and alkaline conditions. Armour paint (pansarfärg) was used as a primer and/or topcoat for roofing (IVA 1935). The pigments used in the paint were metallic aluminium flakes, micaceous iron oxide, and carbon black in binders of one-quarter tung oils.

Standard procedures gradually changed during the 1920s and 1930s. The linseed oils were heated up to 130–150°C, in combination with air-blowing and the addition of manganese compound driers for reduced drying times (Karlsdotter Lyckman 2005). Air-blown oils are hydrophilic, i.e. the paint binder attracts water when deteriorating and the driers induce accelerated cracking in paint layers. These treatments reduced the quality and lifetime of oils and paint films. After WWII, alkyd paints became more common as an alternative to linseed oil paints. The alkyds require VOCs, resulting in more porous paint layers which are more susceptible to moisture penetration and increased risk of under-film corrosion and embrittlement (van Eijsberg 2015). Furthermore, the modern painting types that came into use were not fully compatible with the historical materials. During the building conservation revival of the 1980s in Sweden, linseed oil paints went into production again (Karlsdotter Lyckman 2005). However, the quality of modern linseed oils was not comparable to the older types and many paints also required VOCs. The modern linseed oil paints were manufactured from raw linseed oil or oils that were heated up to 130–140°C (Karlsdotter Lyckman 2005). Zinc oxide is now added instead of toxic lead white (for formation of water-resistant soaps) and increased film quality of linseed oil painting systems. The results from the research project show that linseed oil paints containing standoils, without VOCs, are still used for steel-sheet roofings and have anti-corrosive qualities. Also, alkyds and polymeric primers (styrene acrylate or vinyl) are used. There are varied many opinions concerning whether it is better to use an original type of paint or a modern substitute. There is a need to balance the function of the anticorrosive treatments with the desire to use compatible and environmentally friendly treatments and historically accurate (non-toxic) paints.

Linseed oil paints have often been rejected in modern management, but the case studies, interviews and experiences in our research project show that this material has often been misinterpreted and has qualities as anti-corrosive painting (Källbom 2018). Case studies on armour linseed standoil paints have shown in reality to have a life-time of about 70 years (Reuterswärd 2014). Linseed oil alkyds may have a lifespan three to four times longer than epoxies on constructions such as steel bridges (Kronberg 2008). Sixteen years of field exposure on steel-sheet roofings has shown that so-called long oil (length) alkyds and linseed stand oil paints are comparable in terms of their lifespan (Törnblom 2009).

5. Diagnosis: the nature and causes of decay

Basic questions in a paint degradation analysis are (according to Koleske 2012): What are the existing paint failure mechanisms? What type of painting system do we have? What desired property has failed? How were the paintings applied to the substrate? What evidence does the actual situation represent?

According to our experiences, visual examination and format or thickness measurements combined with historical knowledge are, in most cases, sufficient to identify the type and quality of the steel-sheet roofing (Table 2). Sampling for metallographic analysis is usually not performed because this means that the steel-sheet roofing construction must be opened in laps. Deformation hardening in the steel makes it difficult to reclose opened laps without cracking. Destructive testing should thus be avoided. Diagnosing a steel-sheet roofing requires guidances to determine the causes of a state of degradation (see, for instance, Berdahl et al. 2008; Brock and Groteklaes 2000; van Eijsberg 2015). The case studies show that metallisation type is identified by visual appearance and layer thickness. Large grains, spangles, are typical features of hot-dipped metallisations, and could be 50 µm or more after 120 years of atmospheric exposure. Thicknesses could be measured by magnetic induction, since zinc layers are not magnetic.

Table 2. Attributes of existing paint layers

Paints	Structure, density, thicknesses	Aging characteristics
Linseed oil paints	Quite even, dense, thin layers. Check for brush strokes. Thick layers/flows have wrinkled appearance. Usually 50–120 µm.	Chalks and erodes evenly, may crack in fine patterns on the north side. Colour change is likely to have a brighter appearance due to chalking.
Alkyd paints	Quite even, dense. Check for brush strokes. Thick layers/flows are usually unwrinkled. Usually 50–200 µm.	Similar to linseed oil but usually glossier for a longer time. May be embrittled, crack and flake. The colour change is a brighter appearance due to chalking.
Styrene acrylate and acrylate, latex paints	Pored. Check with a pocket lens. Often applied with a roller. Thick layers/flows are unwrinkled. Usually 100–300 µm.	May hold gloss well or may become matte, depending on types. May flake or erode. Colour change is likely to vary depending on type.
PVC paints	Porous or even, depending on the application method. Quite thick. Thick layers/flows are unwrinkled. Usually 100–300 µm.	Chalks. May flake after a long time, especially where the sun is exposed due to migration of plasticiser. Colour change has a brighter appearance due to chalking.
Asphalt, bitumen layers	Porous, thick layers. Thick layers/flows are unwrinkled. Usually 100–500 µm.	Thermo-plastic. Cracking. The sheet may corrode heavily under the layers. Thick porous layers, smells of asphalt when heated. Usually gets a greyer appearance.

Some of the most important degrading factors of substrates and paint layers are as follows (Berdahl et al. 2008; Sørensen et al. 2009; Talbert 2008; van Eijsberg 2015). Embrittlement of paint layers is due to electromagnetic (ultra-violet) radiation resulting in cracking and flaking. Common causes are inhomogeneities in the paint, ageing by UV exposure (photo-oxidation), and moisture (hydrolysis). Leaching or erosion of paint layers is caused by precipitation, fouling, deposition, and being damp for a long time. High wind velocities accelerate degradation. The results are lack of binder, chalking/eroding and matte surfaces. Mechanical impacts, like scratches, hits and cuts on paint surface, often in connection to installations, may cause corrosion of the metal substrates. Corrosion of the substrate is a chemical degradation that occurs when the painting system is penetrated by moisture due to inhomogeneities, ageing or osmosis. Gases and salts may dissolve in the moisture, leading to increased corrosion rates. Deposition, excrements and microorganism degradation may result in under-film corrosion on zinc metallisation or steel substrate. Fatigue of the substrate is caused by repeated temperature and form variations when large format sheets expand and contract. The results are cracking in metallic substrate and/or paint layers.

Identification of existing paints determines the alternatives for anticorrosive treatments. Different combinations of paint and substrate may be unsuitable or incompatible due to, for instance, differing expansion coefficients, density, and degradation mechanisms (van Eijsberg 2015). Generally, existing layers should be repainted with similar types of paint. The stakeholders and craft practitioners’ experience is that oil-based paints can usually be overpainted with polymeric types, but not automatically vice versa. Whether or not it is possible to overpaint one type with another depends on the density, ductility and adherence of the substrate/old paints and the new paint (Källbom 2018). If the existing layers lack adhesion, the new layers will not adhere, leading to flaking, cracking, and material loss.

Different types of paint deteriorate with different attributes (Table 2). According to the stakeholders’ and craft practitioners’ experiences, identification of existing paint types determine the alternatives and performance of repainting, as the different combinations of paint and substrate may be unsuitable or incompatible – i.e., differing expansion coefficients, density, and degradation mechanisms. Visual examinations and simple empirical tests may also be used to determine the type of binder in existing paint. Figure 2 presents a flow chart that could be used as a guiding tool in the diagnosis of various existing paint layers. If a certain chemical dissolves a paint layer, the flow chart reveals the associated paint type. If the layers do not dissolve, the flow is followed to the next type of chemical.

Figure 2. A guiding flow chart for identifying paint type binders (orange boxes). The diagnostic operations (chemicals, grey boxes) could be used as indicators together with visual characteristics. This figure is based on information from several paint producers, such as Alcro and Akzo Nobel, with minor modifications.

Adhesion and thicknesses of existing paint layers should be measured and evaluated (Reuterswärd 2014). Old layers may be beneficial for increasing the total thickness but may also be detrimental due to the risk of stresses, causing flaking. If polymeric coatings contain plasticisers, this may cause non-drying or softening of oil paint layers, and could be prevented by using certain barrier paints (Selander et al. 2007).

6. Therapy: a guidance for formulation of best practice procedures

The best practice for repairing sheets and metallisations is beyond the scope of this article. The quality and life-time of painting systems are affected by many factors and are difficult to compare in practice (van Eijsberg 2015). Adhesion to substrate and interlayers, painting thicknesses, painting defects, and environmental exposure are examples of factors which have great impact on performance and lifetime. Field exposure tests need evaluations for at least ten years (Reuterswärd 2014; Törnblom 2009). During this period, formulas of painting systems may have changed. This causes a time lag, making it difficult to evaluate procedures and performance. Laboratory tests are accelerated but difficult to translate to real conditions. The results from this study show that technical documentations for historical buildings are almost always insufficient and do not provide data for long-term evaluations.

A basic decision for maintenance of steel-sheet roofings is whether to maintain the roofing with new painting or to completely remove existing paint layers. The ISO standard 4628–1 describes different levels of degradation of painted surfaces and may be useful when determining the needs for maintenance. Various criteria also exist for determining the needs for complete, partial or local renewal of painting systems or just repainting the topcoat (see Bayliss and Deacon 2002; Brock, Groteklaes, and  Mischke 2002; van Eijsberg 2015). The criteria are based on adhesion strength between paint and substrate, estimation of paint loss over surfaces, types of damages, and so on.

Historically, steel-sheet roofings were maintained after one or 2 years initially, and then every fifth or tenth year (Karlsdotter Lyckman 2005). Today, maintenance is performed over intervals of 10–20 years (Källbom 2018). With long intervals, the painting system becomes highly deteriorated. As a consequence, the repainting is often performed with complete paint removal and extensive pretreatments. This also means a potential risk for local corrosion damage and pittings in the substrate. With shorter intervals, the maintenance actions are usually limited to cleaning and renewal of the top coat, and although the expenditures are more frequent, the total costs over time are usually lower and reduce the risk of substrate damages (Källbom 2018). Linseed oil paints have the benefits of gradual erosion, making them appropriate for intermittent maintenance and touch-up.

An anticorrosive treatment comprises three comprehensive steps:

1. Pretreatment and cleaning; 2. Application of a barrier-forming primer, preferentially electrochemically inhibitive and/or passivating; 3. Painting, with intermediate and (usually) top layers.

The pretreatments are detrimental for the quality, because they could affect the adhesion between paint and substrate by factors such as cleanliness and surface profile (Bayliss and Deacon 2002; Brock, Groteklaes, and  Mischke 2002; van Eijsberg 2015). Inadequate pretreatment could also affect the risk of under-film corrosion due to impurities or unstable surface compounds. Because there is no universally applicable pretreatment method for historical buildings, several methods may be combined in order to remove different types of impurities such as corrosion products, grease, salts, soaps, etc., and with respect to the substrate. The methods should be based on adequate knowledge derived from expert observations, sampling, and testing. An important instruction is that painting must be performed on solid, clean, and compatible surfaces with a certain surface profile for good adhering and anticorrosive action.

As with all painting treatments, it is the substrate that determines the choice of paint type and painting procedure. Therefore, the comprehensive procedures should be based on the values of substrate material, metallisation, and paint-systems but also the goal of the intervention. They should involve adequately skilled and experienced personnel. The desired thickness of the paint-system depends on the corrosivity class, which is determined by geographical locations, as described by ISO 12944–2 (1998) and ISO 9223 (2012), (Table 3). C1 and C2 represent rural locations (most common in Sweden), C3 and C4 represent urban or industrial environments. The C5 class corresponds to marine environments. If the surface profile is rough, extra priming should be used (IVA 1961). One brush stroke yields a dry film thickness (DFT) of about 40–50 μm, making it possible to build up a total paint/painting thickness of about 120 μm for three or four layers of paint (Hudson and Stanners 1955). Manual paint film thickness could be highly variable, and it is important to control thicknesses and variations. It is always important to follow the specifications of the paint manufacturers, in order to avoid warranty responsibilities and to achieve the efficiency of the paint-system.

Table 3. General guidelines for anticorrosive painting in different corrosivity classes and for total paint-system thicknesses of linseed standoil paint for metallised steel surfaces (ISO 9223-2 (2012);, IVA 1961)

Corrosivity class	General painting guideline	Minimum total DFT (Dry Film Thickness)
C1	One primer + one top layer	100 ± 20 μm
C2	Two primers + one top layer	140 ± 20 μm
C3	Two primers + two top layers	180 ± 20 μm
C4-C5	Usually 4–5 layers of special paints.	According to suppliers’ recommendations

In order to get an overview of the complex challenge of diagnosing existing roofing and planning for best practice treatments, a matrix of pre-existing materials and alternative treatments is compiled (Table 4). In Figure 3 and Table 5 an example is shown on how the matrix can be used for formulating a general working procedure. The matrix can be used as a tool in diagnosis, planning and documentation of general procedures, or it can be refined for object-specific procedures. It could also be used for comparing different/similar types of cases, areas of further work or for setting minimum levels for documentation. The matrix can be connected to different quality standards of conservation. It may favour the communication and systematic assessment of working procedures, and the further development of best practices. A large number of pretreatment processes are possible, which we cannot review here. Paint application methods used in practice are by brush, by spraying or by roller, and sometimes those methods in combination, but this is not further explained here.

Table 4. A material and method matrix that can be used as a tool for formulating general working procedures

Category	Examples
1.Substrate type	1A: Wrought plane sheet 1B: Hot-rolled plane sheet 1C: Hot-rolled corrugated or other profiled sheet 1D: Cold-rolled plane sheet, etc.
2. Metallisation type	2A: No metallisation, missing 2B: Tin, hot-dip metallisation, 1700-ca 1800 2C: Zinc, hot-dip metallisation, 1800–1900 2D: Zinc, hot-dip metallisation, 1900–1960 s 2E: Zinc, electroplated, 1920–1960s 2F: Zinc, hot-dip/electroplated, c. 1970–2006 2G: Zinc, hot-dip/electroplated, chromated, 2006– 2H: Zinc, hot-dip/electroplated, raw or lubricated, c. 2015– 2I: Aluminium-zinc alloy metallisation
3. Existing type of paint layers/coatings	3A: Factory coated Organosol 3B: Factory coated Plastisol 3C: Factory coated PVDF (fluropolymers) 3D: Factory coated polyester, acrylate, polyurethane 3E: Factory coated, unknown type 3F: In situ painted linseed oil paint 3G: In situ painted alkyd paint 3H: In situ painted polymeric paint, specify 3I: In situ applied paint, unknown type 3J: In situ applied asphalt/bitumen, coal tar 3K: In situ applied kautschuk, etc.
4.Pre-treatments	4A: Manual scraping, cleaning 4B: Manual grinding, cleaning 4C: Manual (gentle) steel brushing, cleaning 4D-I: Different types of dry and wet blasting 4J: High pressure washing, combined with (alkaline) emulsifying detergent, post washing (pw), 200–300 bar 4K: Hand wash with (alkaline) emulsifying detergent, neutralising, pw 4L: Paint stripper, neutralising, pw 4M: Pickling, neutralising, pw 4N: Scrub pad (or 240 grit sandpaper), pw 4O: Atmospheric corrosion (i.e. zinc patina formation and coarsened surface profile)
5. Pre-treatment oils and paints	5A: Linseed oils 5B: Tung Oil mixtures (with linseed oils) 5C: Linseed oil alkyds 5D: Other oils 5E: Plasticiser stoppers
6.Primers	A: Red lead base, film-forming (f.f) linseed oil paint B: Red lead base, alkyd (long oil length) paint C: Micaceous iron base, f.f linseed oil or linseed oil alkyd (long oil length) paint D: Red natural hematite base, f.f linseed or linseed oil alkyd (long oil length) paint E: Zinc phosphates base, f.f linseed or long or alkyd linseed oil paint F: Zinc phosphates base, polymeric solvent borne paint G: Complex (aluminium poly) phosphates bases, f.f linseed standoil paint H: Armor paint, MIO + metallic aluminium base in f.f quarter tung oil paint I: Metallic aluminium flakes base, f.f linseed standoil or quarter tung oil paint J: Graphite base f.f linseed standoil or quarter tung oil paint K: Other paints All except A, B, F may be used also as intermediate/top coats. C, H, I, J as primers only on special occasions.
7.Pre-primed sheet	A: Polyester, etc.
8.Intermediate or top paints. In-situ paint types. See also 6.	A: Modern linseed oil paint B: Linseed standoil paint, also with. addition of tung oil C: Alkyd, linseed oil type (long oil length) D: Alkyd, other type than 8C, long oil length, E: Others

Table 5. The same example as in Figure 3, but described as a working procedure.^a

Substrate 1A + 2A, 3 F
Original small format wrought non metallised steel sheet roofing from 1771 in corrosion zone C2, example of procedure. The former paint has partly good adherence.
Step	Specification	Dry film thickness
Pretreat- ments	4A, 4B, 4C manual methods combined with 4J High pressure blasting ca 200 bar hot water with emulsifying detergent, to St 2. + 5B Tung oil mixture (with linseed oils) applied in waterways, crevices	-
Primer	6A Red lead in film forming linseed oil or 6G Complex aluminium phosphates in linseed standoil paint. Applied undiluted with brush maximum 24 hours after surface cleaning, in two layers.	Each layer 40–60 μm
Location paints	8B Linseed standoil paints or 6G Complex aluminium phosphates in linseed standoil paints, in wanted colour. Applied undiluted with a brush after each layer is completely dry + one day extra, in two layers.	Each layer 40–60 μm
Total dry film/coating thickness		160–230 μm

^aThe sheet needs to be cleaned from contaminations and non-adhering paint layers with gentle manual methods. This is combined with quite gentle high-pressure water blasting of approximately 200 bars (with addition of emulsifying detergents in order to dissolve grease). Shortly after cleaning, the dry waterways and crevices are applied once or repeatedly with penetrating and water repellent linseed tung oil mixture. The primer is applied undiluted within 24 hours after cleaning, in order to prevent re-corrosion, and is brushed in two layers. A inhibitive and passivating linseed oil primer should be used since the substrate is not metallised. If red-lead is rejected due to environmental concerns, a primer with complex phosphates could be used instead. After complete drying, location paints are applied by brush in two layers. The drying time between the layers is crucial for the drying and quality of the anticorrosive system, since the oxidation and polymerisation of the underlying layers are strictly retarded when covered by a new layer. When the current top layer is completely dried, one extra day is recommended before the next layer is applied. The recommended location paints are linseed stand oil paints due to their combination of ductility and strength. The location paints could also contain complex phosphates, since these pigments are almost transparent and do not affect the colour of top layer pigmentation. The thickness of each layer is controlled, and also the total thickness of the paint-system that should be in the interval of 160–230 µm in total.

Figure 3. An example of treatments for an original eighteenth-century wrought steel-sheet roof with existing linseed oil paint. The matrix shows possible combinations of materials and treatments for this substrate. On the horizontal axis, categories for diagnosis of existing conditions. On the vertical axis are the categories for treatments.

7. Control checkpoints and monitoring

Table 6 contains the recommended main checkpoints for quality control in the process of maintaining steel-sheet roofings on historical buildings, with reference to applicable standards (Källbom 2018). These checkpoints usually require not only supervision from a building conservator but also from a technical controller. Management and monitoring once the maintenance treatment is controlled and finished, involves, for instance, different local touch-up actions, control of adhesion and layer thicknesses, and cleaning of surfaces and waterways, etc.

Table 6. Recommended checkpoints for quality control of anticorrosive treatments

8. Discussion and concluding remarks

We have to take care of our historical buildings today and cannot wait for all scientific evidence before we practice painting treatments of steel-sheet roofings. Advocating the benefits of traditional materials and methods has become difficult in a time of increasingly standardised and rationalised building processes, and the historical materials have also changed. Thus, building conservation needs to develop best practices and appropriate ways to describe and communicate how to proceed in anamnesis, diagnosis, treatment, and control in the cultural context. The need for informed decisions is also pointed out in the principles of the ICOMOS charter from 2003.

This article provides a guiding model for anamnesis, diagnosis, treatment and control/monitoring of painted steel-sheets roofings on historical buildings. Based on a survey of current practices, the research has compiled selective criteria for technical, environmental, economic, and historical quality concerning steel-sheet roofings. The matrix is a tool for specifying combinations of relevant materials and methods for the working procedure, and provides a structure and communication platform for the actors in the conservation process. Identifying the most useful standards for quality control will extend awareness and measurable results will increase in both the short term and the long term. On the basis of the results, we propose a model for the circular nature of maintenance (Figure 4). The main steps in the process are marked, and the control activities are common at several stages.

Figure 4. A model for the repetitive nature of maintenance of steel-sheet roofs.

Anticorrosive treatments of ferrous heritage constitute a large and complex field. Of course, the gap cannot be closed solely by this research. The tools provided in the conservation process and the matrix system to assist with organising appropriate working procedure, could provide for more systematic, reliable and comparable long-term results. We still need research and experts, and mutually conscious dialogue in the conservation processes in order to reduce the gap between research and practice. From our standpoint, it is important that craft practitioners are also considered as experts, who must be included in the dialogue on the processes of anamnesis, diagnosis, therapy and control/monitoring.

There is also a need for face-to-face instruction and knowledge transfer as part of the continuous, intangible professional development. The most important actors in the conservation process are probably the traditional architectural painters and paint-makers who are close to the source of knowledge. They are continuously engaged in the conservation process and make decisions that directly impact the final results. The evidence-based research starts from broadly defined information from experts, case studies and systematic reviews, which is then narrowed down depending on the objects, their cultural significance, their status and specifications on best practice depending on materials and procedures. This research gives scopes for developing a standard in the CEN/TC 346 programme. Similar working processes and tools could also be used in the are of pine tar on wooden roofings, and conservation or surface treatments of inorganic substrates such as stone or mortars for historic buildings.

Acknowledgments

We thank the Swedish Church, The National Heritage Board, and The National Transportation Board, for the funding of this work.

Disclosure statement

The Authors declare that there is no conflict of interest.

Additional information

Funding

This work was supported by the The Swedish National Heritage Board; Swedish Church; The Swedish National Transportation Board.