Load capacity and failure analysis of alternative repair materials for fibrous plaster wads

A

alpha plaster

AG

RE Aramid Gel™

B

beta plaster

H

hessian scrim

NW

no wire present

Q

quadaxial fabric

W

galvanised steel wire present

Fibrous plaster was patented in the UK in 1856 by the Frenchman Alexander Desachy, and the patent was subsequently acquired in 1863 by George Jackson and Sons (Millar, 1897). By the end of the century, fibrous plaster became the material of choice for ceilings and other ornamental features such as columns and cornices in grand public and private buildings in the Victorian and Edwardian years, for example in theatres, civic buildings, town halls, and grand private residences (Stewart et al., 2019; Awang Ngah et al., 2020). Figures 1(a) and 1(b) illustrate fine examples of early twentieth-century theatre fibrous plaster ceilings.

Fibrous plaster is a composite material that consists of a matrix of gypsum plaster (which is also known as ‘plaster of Paris’ due to the large deposits of gypsum in the Montmartre area of Paris, France). Reinforcement was traditionally provided by hessian scrim (Ireland, 2020), a fibrous weave derived from the jute plant, part of the bast group (Garside and Wyeth, 2006), and native to India, with Dundee in Scotland established as a centre of UK production (Masrani et al., 2020). It should be noted, however, that hessian fibre scrim was not the only reinforcement material historically used, as the authors have discovered cotton fibres to have been employed in the Royal Opera House, London and alternative fibres have been used in other countries, for example, sisal in fibrous plaster sheets in Australasia (St John and Kelly, 1975). Historic fibrous plaster ceiling panels could be as little as 5 –6 mm in thickness (Brookes, 2021a), where two reinforcing layers of hessian fibre scrim were typically incorporated within the plaster matrix. Timber laths, usually at 0.5 m spacings, provided additional reinforcement for ceiling panels. Although these are lightweight, other decorative fibrous plaster casts were commonly affixed to them, some being extremely heavy.

Fibrous plaster panels were sometimes fixed to walls or timber joists using screws or nails, though many fibrous plaster ceilings were suspended from structural steel or timber roof beams, or ceiling beams, by fibrous plaster wadding ties (often shortened to simply ‘wads’) (Dams et al., 2024). Wads traditionally consist of hessian fibre scrim soaked in gypsum plaster, attached to timber laths in the panels. Plaster and scrim were sometimes wrapped around a steel wire, encasing the wire. The steel wire was looped over the beam element at the top and then fed through a drilled hole in a reinforcing timber lath within the ceiling panel, at the base of the wad. This formed a loop of steel which was then twisted centrally about its own axis to form a ‘looped-twisted’ wire (Dams et al., 2024). A theatre roof space, for example, can contain hundreds of wads in a variety of dimensions and orientations, observationally ranging from several centimetres long up to approximately a metre in length. Figure 1(c) illustrates modern wads using traditional hessian scrim, and Figure 1(f) shows an example of a longer wad. Unblemished, lighter plaster indicates modern replacement wads while darker, discoloured plaster are original or aged historical wads (Figure 1(e)). A schematic diagram showing wads wrapped around structural elements and through the reinforcing timber laths present in the fibrous plaster ceiling is illustrated in Figure 1(g).

Fibrous plaster industry professionals have noted that in most historic wads, steel wire is not present (Stewart et al., 2019). The steel wire may have been added in specific locations to help secure panels in place, or historically, the plaster-soaked hessian was considered to be sufficient to support the ceiling’s weight. It is doubtful that mechanical calculations were used to identify the necessary spacing of wads to support the ceiling. Fibrous plaster may degrade over long time periods due to gradual moisture-induced degradation of plaster of Paris, weakening of hessian scrim under load, or biodeterioration of hessian (Maundrill et al., 2023). Once micro-cracking occurs in wads, the hessian scrim is exposed to humidity allowing attack from fungal spores. When degraded, hessian fibres lose their integrity, and the tensile capacity of the fibres is then severely reduced. As soon as one wad fails, load is redistributed to adjacent ones (Brookes, 2021a). If adjacent wads subsequently fail, an area of the ceiling can no longer be supported, and partial collapse occurs by means of a domino effect.

There have been failures of historic fibrous plaster ceilings, occasionally recorded in the press from the post-Second World War period. The most famous recent occurrence was the partial collapse of the auditorium ceiling in the Apollo Theatre, London, in 2013 (BBC, 2013). The collapse happened during a well-attended performance and over 80 people were injured, many seriously, requiring hospitalisation. There have also been other recent failures such as in the Savoy Theatre ballroom, London, in 2019 (France, 2019). These emphasise the importance of surveying, repairing, and maintaining historic fibrous plaster ceilings from both a safety and cultural perspective. Following the Apollo Theatre collapse, the Association of British Theatre Technicians issued new robust guidance for the ongoing survey, repair, and maintenance of such ceilings for specialist fibrous plaster companies, structural engineers, other conservation professionals, venue owners, operators, managers, and other stakeholders (Association of British Theatre Technicians, 2015). A baseline survey for a fibrous plaster ceiling is recommended to be conducted by both a competent fibrous plaster practitioner and a structural engineer (Brookes et al., 2020). Repairing historical fibrous plaster ceilings and wads can be challenging in some cases due to the constricted nature of roof spaces, which can be very difficult to access (Brookes, 2021b).

There are differing approaches to repairing or replacing aged historic wads, which can use either traditional or modern materials. Using traditional materials is not necessarily the most appropriate option for long-term conservation (Ireland, 2022), given their possible failure from environmental conditions described above. Approaches to conservation can include a new wad consisting of traditional beta plaster, hessian scrim, and a galvanised steel wire being placed adjacent to a degraded historic wad. Fibrous scrim can be soaked in a traditional beta plaster, or a stronger modern alpha plaster and applied in layers on top of historic material, or modern synthetic materials such as RE Aramid Gel™ can also be sprayed onto historic fibrous plaster material in a building roof space (Dams et al., 2023).

Regarding modern replacement practice, wads are recommended to contain a galvanised steel wire with plaster-soaked hessian scrim wrapped around, enclosing the wire (Ireland, 2022). Stainless steel wire may also be used as an alternative option, but galvanised wire is a popular choice due to its ductile properties. The elongation capacity of galvanised wire in a newly manufactured wad has previously been demonstrated by the authors (Dams et al., 2024; Dams et al., 2026) using traditional beta plaster and hessian materials. A typical arrangement of fibrous plaster wads supporting a ceiling is to specify four wads per square metre of ceiling; this is an established rule of thumb in industrial practice.

There are modern alternative options for both binding materials and reinforcing fabrics. A leading modern alternative binding material option to traditional gypsum beta plaster is the stronger alpha gypsum, developed in the 1930s. Alpha plaster is still calcium sulfate dihydrate, therefore chemically similar to beta plaster. But the difference with alpha plaster is the method of calcining, which involves the gypsum hemihydrate being heat-treated in a modern steam autoclave, rather than dry calcining in a traditional kiln. The autoclave heat treatment means that alpha plaster has a denser crystalline structure and is a mechanically stronger material (Awang Ngah et al., 2020).

Another approach is spraying HPCP RE Aramid Gel^TM to the surface of historic fibrous plaster. The gel contains DuPont^TM Kevlar® fibres as reinforcement. Invented in 2010 by Rod Stewart of Historic Plaster Conservation Services, the material is patented and used in North America, spraying the material in the fresh state onto historic material (Stewart, 2022). Glass fibres are a leading reinforcing fabric alternative to using traditional hessian or plant-based fibrous reinforcement – a quadaxial glass fibre mat can be used with orthogonal fibre strands and further fibre strands in both diagonal directions (Jesmonite Ltd, 2023). Quadaxial fabric has been used with acrylic-modified plaster as a modern alternative repair option for historic fibrous plaster ceilings (Ireland, 2020). Figure 1(d) depicts modern wads using quadaxial fabric.

This study conducted a range of tests within laboratory environments on newly manufactured fibrous plaster wads using both traditional materials and with a focus on modern alternative repair material options. The authors have previously demonstrated that newly manufactured traditional beta plaster wads of ≈45 mm dia. with hessian scrim were able to support a tensile loading of ≈3 kN (≈300 kg) with a galvanised steel wire included, and ≈1.5 kN (≈150 kg) without a wire included (Dams et al., 2024). The load capacity of wads containing alternative combinations of modern repair material options tested in this study can be compared to these previous values using traditional materials. Tensile tests were conducted on wad specimens in a bespoke test rig to examine loading capabilities and mechanisms of failure. There is extensive knowledge and practical experience devoted to fibrous plaster within a small and specialist fibrous plaster repair industry. However, to date, little research has been conducted in a scientific or laboratory setting to establish the properties of fibrous plaster wads (Ireland, 2022). The implications of the results for the fibrous plaster industry and how they may inform professional practice and the conservation of historic fibrous plaster wads and ceilings moving forward are evaluated. Modern repair materials are assessed, including quadaxial glass fibres (substituted in place of traditional hessian scrim), stronger alpha plaster, and HPCP RE Aramid Gel™.

This paper represents the first time the tensile properties of these modern, alternative materials have been tested and evaluated using large-scale wad elements representative of real-world in situ wads with a region of interest (ROI) in excess of 120 mm (Figure 2(a)). The ensuing failure mechanisms of the composite wad element are also discussed.

Fibrous plaster wad specimens were manufactured in the laboratory using both traditional materials (beta plaster reinforced with hessian scrim) and alternative modern repair materials, both with and without wires.

Table 1 summarises the fibrous and matrix materials used in the tensile wad tests along with the known, published properties of the individual fibre and plaster materials. Alpha and beta plasters were mixed with water in accordance with the manufacturers’ instructions. Siniat Prestia Classic and Saint-Gobain formula Crystacal® ‘R’ were respectively the beta and alpha plaster used for the manufacturing of wad specimens in this study. Hessian (jute plant fibre) woven scrim reinforcement had a mesh size of 5 mm $×$ 5–10 mm (typically 7 mm) and a weight of 102 g/m². Quadaxial glass fibres, consisting of fibre yarns in 0°/90° and +45°/−45° orientations, had an orthogonal mesh size of 8 ± 1 mm $×$ 8 ± 1 mm and a weight of 133 g/m². Plaster and fibre materials were sourced from Industrial Plasters, Chippenham, UK. RE (‘reinforcing’) Aramid Gel™, an acrylic resin which contains DuPont^TM Kevlar^® fibres, was sourced from Historic Plaster Conservation Services, Ottawa, Canada.

Wad test specimens were made both with and without galvanised steel wires of 1.25 mm dia. (also obtained from Industrial Plasters, Chippenham, UK) and the range of six sample groups is detailed in Table 2, with eight specimens manufactured within each sample group. Groups were assigned names based on the following coding system: Matrix plaster material used – fibre type used – looped-twisted wire either present or absent – sprayed RE Aramid Gel™ presence. The following lettering denotes the constituents: B = beta plaster, H = hessian scrim, A = alpha plaster, Q = quadaxial fabric, W = galvanised steel wire present, NW = no wire present, and AG = RE Aramid Gel™. Therefore, as an example, B-H-NW-AG denotes beta plaster with hessian scrim no wire present and RE Aramid Gel™ sprayed onto the specimen. The authors have previously investigated continuous fibre mat glass fibres as an option for wads (Dams et al., 2024) in specimens of equivalent dimensions, therefore they were not investigated in sample groups within this study. As noted in the introduction, the authors have also previously tested wads with a traditional beta plaster matrix and hessian scrim reinforcement in specimens of the same dimensions (Dams et al., 2024), which can be taken as reference and comparison values for this study.

Each wad specimen was individually hand-made reflecting the in situ methodology of roof space wad manufacture. Specimens were manufactured as consistently as possible in the laboratory environment, however moderate dimension variation was inherently inevitable. Figure 2(a) illustrates the dimensions of manufactured wad specimens. The procedure for the manufacture of specimens is detailed below:

1. A galvanised 1.25 mm dia. steel wire was cut to length and looped around two lengths of plastic piping with a 40 mm external diameter placed around and held in position by lengths of timber (Figure 2(b)). A releasing agent was applied to the external surfaces of the plastic pipes to facilitate the removal of the finished hydrated wad specimens once complete. The pipes formed a circular ‘eye’ at either end of the wad.

2. The two ends of the steel wire were joined and intertwined with manual twisting to form a loop (Figure 2(c)).

3. The steel wire was twisted about the central axis using a 5 mm dia. metal rod (Figure 2(d)). Twisting is commonly used in modern industrial practice (note: the steel wire was omitted for sample groups B-H-NW-AG and B-Q-NW).

4. Plaster was mixed with water in accordance with the ratios contained in Table 1, which resulted in 700 g of beta plaster to 500 g of water and 900 g of alpha plaster to 315 g of water.

5. Once the plaster mix had begun to increase in viscosity, a 1000 mm long, 450 mm wide piece of hessian or quadaxial scrim was soaked in plaster and wrapped around the steel wire, or just the plastic pipes if the wire was not present (Figures 2(e) and 2(f)) until the scrim enclosed the steel wire and was secure around the pipes.

6. An extra length of plaster-soaked scrim was wrapped around specimens at both ends (Figure 2(f)). Bulking up the wad end regions ensured that failure mechanisms occurred in the central gauge length of the wad. Failure at the wad-ceiling topside interface has previously been investigated by the authors using adhesion ‘pull-off’ tests (Dams et al., 2023).

7. Completed wad specimens were then conditioned in the laboratory for a period of 2 weeks, at a temperature of 20°C ± 2°C and relative humidity of 50% ± 5% to ensure wads were fully dry prior to tensile testing (Figure 2(h)). Wad specimens were weighed prior to testing. Although every effort was made to ensure adherence to the dimensions shown in Figure 2(a), there was inherently minor variation in the thickness of the test specimens along both the gauge length and the ‘bulked’ end regions; again, this variation in dimension reflects real-world in-situ industrial practice.

Wad specimens were placed into a bespoke testing rig as depicted in Figure 2(g). A length of 40 mm external diameter hollow steel bar was inserted into the wad eyes at either end to distribute the load evenly within the eye, plus a 15 mm solid external diameter steel rod was inserted through the hollow bar to secure test specimens (illustrated in the schematic diagram in Figure 2(i)). Figure 2(j) contains images of the hessian and quadaxial fabric mats.

Displacement-controlled tests were conducted using a Dartec Universal Testing Machine equipped with a 100 kN load cell. A cross-head speed of 2 mm/min was applied to all wad specimens until a 10 mm displacement was reached, at which point the loading rate was increased to 10 mm/min until a maximum displacement of 40 mm had been reached, at which point the test was terminated. A small pre-load (0.04 ± 0.02 kN) was applied once wad specimens had been manoeuvred into position to verify correct attachment prior to commencement of full loading. If the test specimen broke into two pieces prior to the maximum 40 mm displacement, the test was terminated at that point. Load–displacement profiles were recorded for each specimen and because of the inherent variation in sample dimension, tests are presented in load (kN) versus displacement (mm) form rather than stress versus strain. Selected specimens were also subjected to X-ray computer tomography (XRT) scans before and after testing. XRT scans were undertaken using a Nikon XT H 225 ST model device and conducted using a 65 kV source, an exposure rate of 1.5 s and 50 μA X-ray beam output. The resulting scans were imaged and analysed using Avizo software.

This study primarily uses force and displacement for results for two core reasons. The authors have previously published results for wads of equivalent dimensions which used force–displacement (Dams et al., 2024) and the results in this study can be directly compared to the values achieved by traditional materials, which may act as a benchmark for the modern alternative materials investigated in this study. Secondly, each and every wad in the real world is a prototype and these wad specimens were made to represent industrial practice. The same quantity of materials were used for each specimen, but each specimen was handmade and therefore there will inherently be an extent of minor variation in dimensions along the length of the hand-made specimen, and moderate variation in dimensions from one specimen to another, which was a further justification for the large sample set size of eight from which meaningful coefficient of variations (CoVs) could be obtained. However, to give an indication of the stresses involved and the tensile strength in MPa of the specimens, the results section also presents the conversion of maximum forces into stresses assuming a uniform diameter along the ROI of 45 mm and cross-sectional area (CSA) of 1590 mm². Furthermore, in another study by the authors, notably smaller tensile specimens were sanded on all four sides in the central ROI, which provided a precise calculation of CSA from which stress-strain output could be calculated and provided a guide for the tensile strength of the sample groups (the reader is referred to Dams et al., 2026).

This study focuses on the tensile capacities of wads and the rig is designed for that purpose. There are other loading considerations and the authors have also published studies that examine flexural strength (Awang Ngah et al., 2020), compressive strength, pull-off/adhesion (Dams et al., 2023), dynamic loading (Dams et al., 2026), and moisture and fungal attack (Maundrill et al., 2023). Large panel elements which were tested in Dams et al., 2026 could be examined further looking at shear failure between wad and the ceiling and within wads themselves along with pull-through of wires in a future phase of experimentation.

HPCP RE (‘Reinforcing’) Aramid Gel^TM containing DuPont^TM Kevlar^® fibres was manually sprayed to a thickness of approximately 1.5 mm onto wad test specimens using a Newborn professional pneumatic series model 710AL-30 spray gun with a ¼ gallon lightweight aluminium cartridge, connected to a compressed air supply set to 85 psi (depicted in Figure 3(a)). All specimens sprayed with HPCP RE Aramid Gel^TM were left to dry completely over a period of 2 weeks within a laboratory environment before tests commenced in temperatures of 20°C ± 2°C and a relative humidity of 50% ± 5%. Figure 3(b) illustrates freshly sprayed wad specimens in the laboratory.

The results of the tensile wad tests can be observed in Figure 4 which features the force–displacement profiles of all specimens tested within each sample group (Figures 4(a)–4(f)), and typical force–displacement profiles for each sample group (Figures 4(g)–4(i)). The mean maximum loads and specimen weights recorded for each sample group, along with the standard deviations (error bars) and CoV (diamond markers) are displayed in bar chart format in Figures 5(a) and 5(b), respectively. Figure 5(c) divides the mean maximum force recorded for each sample group by the mean density for each sample group to show loading capacity in relation to material weight. Error bars represent the maximum force divided by the minimum weight and the minimum force divided by the maximum weight. A basic traditional hessian scrim and beta plaster wad length previously tested by the authors (Dams et al., 2024), can typically support a load of ≈1.5 kN without a wire, increasing up to ≈3 kN with a looped-twisted wire present, which serves as a benchmark for comparison. The following key outcomes can be deduced from Figures 4 and 5, with salient characteristics of the force–displacement profiles indicated in Figure 5(d):

• A multi-stage composite action can be observed with firstly the plaster matrix experiencing microcracks and then crack propagation continues to increase, followed by the second phase of fibrous reinforcement elongating and if present, the steel wire also elongating, possessing tensile capacity and a further degree of load capacity in comparison to wireless wads.

• Results demonstrate that quadaxial fabric can sustain a tensile loading of ≈2 kN (≈200 kg) through to the maximum 40 mm displacement of the tests when used in conjunction with beta plaster. This tensile capacity is not in evidence when quadaxial fabric is used in conjunction with alpha plaster, with fibres unable to elongate within the stiffer matrix and fibre breakage more extensive during the tests with less unruptured fibres left to carry tensile loading once the plaster matrix has cracked (and there has been a wire breakage within specimens containing a steel wire).

• Considering the benchmark beta/hessian values above, spraying beta plaster and hessian fibre wads with RE Aramid Gel™ increases the maximum loading capabilities of the wads with the gel encasing the wad in a cured state – mean highest forces recorded being above 4 kN (≈400 kg) with a wire and 2 kN (≈200 kg) without a wire, both instances being an ≈33% increase in load capacity.

• Using quadaxial fabric instead of traditional hessian scrim with beta plaster increases the maximum loading capacity of the wads, typically to approximately 5 kN (≈500 kg) without a steel wire and up to ≈8 kN (≈800 kg) with a steel wire present. This corresponds to a 233% increase without a wire (3.33 times higher) and 167% with a wire (2.67 times higher).

• Using quadaxial fabric and a wire, maximum loads with a beta plaster matrix are close to maximum loads using alpha plaster, with only an ≈7% increase in mean maximum loading when using alpha plaster.

• The inclusion of a steel wire increases the mean maximum loading achieved in relation to wire-less wads with the same plaster matrix and fibrous reinforcement – with an ≈50% increase in load capacity for beta plaster and quadaxial fabric, and ≈73% increase for beta plaster, hessian scrim, and RE Aramid Gel™.

• The use of the beta plaster/quadaxial-fibre/steel wire combination (B-Q-W) resulted in the highest mean tensile force capacity per gram of material weight used recorded in this study.

• In a typical load profile, wads containing the same fibre with a beta plaster matrix display higher loading capacity following the failure of the plaster than wads using alpha plaster. This is most pronounced in wads with quadaxial fabric, where the use of alpha plaster increases the maximum loading achieved prior to failure of the plaster matrix, but the beta plaster specimens have a greater loading capacity as the specimens continue to elongate with increased load application. This is up to ≈900% (or ten times) higher, with a post-plaster failure (at approximately 15 mm displacement) continuing load-bearing capacity of ≈2 kN with beta and ≈0.2 kN with alpha (up to a displacement of 40 mm).

• There is variability in the results, but as can be observed in the standard deviations in Figure 5 and the CoV values in Table 3, the CoV was typically below 10% with only two exceptions in mean maximum force in sample groups B-H-NW-AG and A-Q-W, although these remained below 20%. The most consistent mean maximum force was with hessian fibres and a wire, with both B-H-W-AG and A-H-W displaying only 5% and high consistency. Group B-Q-W displayed a mean maximum force almost as high as group A-Q-W, but with a notably lower CoV, suggesting greater consistency, which is something to bear in mind for conservation practice. Results suggest that performance would be more consistent when using beta plaster with quadaxial fibres rather than alpha plaster – this is further supported by the use of beta plaster allowing the quadaxial fibres to fully utilise tensile properties, providing ductility. The presence of a wire is also clearly beneficial with consistency in results if the specimen is sprayed in RE Aramid Gel™. With mean specimen weight, all coefficients of variation were below 10% (only group B-H-W-AG slightly exceeding at 10.97%) – this high consistency was not surprising due to the fact that although each test specimen was hand-made and would inherently feature a small extent of dimensional variation, the quantities of materials used were the same. The variation that did exist can be accounted for by the quantity of freshly applied plaster applied to the fibres dripping off the test specimen while fresh (although this was kept to a minimum due to a ‘creamy’ consistency in the fresh plaster during application and immediately preceding rapid curing.

• Table 4 assumes a consistent diameter of 45 mm along the ROI and a CSA of 1590 mm² to give an indication of stress values for the maximum forces obtained in the sample groups during the tests. Stress values range from 1.59 MPa for sample group B-H-NW-AG to 5.52 MPa for A-Q-W. The table provides a useful indication of the tensile strength of the wad sample groups. For comparison purposes, and using the same CSA assumption, the beta plaster-hessian scrim benchmark values of ≈1.5 kN without a wire and ≈3 kN with a looped-twisted wire would result in tensile strengths of 1.10 and 2.39 MPa, respectively.

Modes of failure were evaluated further from images taken during tests. Figure 6 illustrates the failure of a typical specimen from sample groups B-H-W-AG (top) and B-H-NW-AG. The cured RE Aramid Gel™ product is a malleable and flexible material, which stretches and possesses the capacity for elongation. The tensile strength of Kevlar fibres is very high, as shown with the published value in Table 1, and it is clear in these tests that the fibres impart elongation capability into the cured gel coating. Uneven horizontal cracks first appear in the gel, although plaster cracks were also occurring, which were not readily visible under the concealment of the gel. The gel ultimately splits into two, which typically occurs at less than 20 mm displacement. There remains capacity within the fibres and wire to elongate once the plaster matrix has failed, but typically steel wires failed when displacements reached 30 mm.

Figure 7 shows failing specimens from the sample groups B-Q-W (top) and B-Q-NW. In contrast to wads with beta plaster and hessian scrim, crack formation and propagation were in the vertical direction, rather than in the horizontal direction, as was the case with hessian scrim, with the plaster ultimately spalling around the stronger and more resistant-to-rupture quadaxial fabric; this was externally observable by a displacement of 10 mm with a wire and 5 mm without. Beta plaster and quadaxial fabric proved to be a ductile combination with notable load-bearing capacity retained often up to the maximum displacement limit of the tests, 40 mm. Even without a steel wire present, the quadaxial fabric retained load-bearing capacities of typically ≈2 kN as they were able to elongate as the beta plaster spalled around them. It was also notable that steel wires, if present, did not entirely break into two in six of the eight specimens.

Figure 8 depicts the failure of specimens from sample groups A-H-W (top) and A-Q-W. While it was clear that the alpha plaster was stronger and denser than the beta plaster, and resulted in a higher maximum loading capacity, it resulted in less ductile composite specimens. Typically, cracking occurred in a broadly horizontal, and sometimes in an ultimately diagonal direction, which suggests there may be shear failure processes also occurring in the plaster matrix – cracking was observable within the first 5 mm displacement. This behaviour was closer to the horizontal cracking of traditional beta plaster and hessian fibre wads, as opposed to the vertical cracks and spalling observed with using quadaxial fabric with beta plaster. Rather than a multitude of micro-cracks propagating in these sample groups, one main crack formed, ultimately resulting in a relatively clean break of the specimen; this happened with both hessian and quadaxial fabric. With the quadaxial fabric, the stronger alpha plaster did not spall around the elongating fibres, but instead, the strong plaster constricted, or constrained the fibres – fibres wanted to elongate, but were unable to do so. Steel wires equally could not elongate and every specimen in these sample groups featured breakage of the steel wire, with breakage in the A-H-W group happening prior to a 20 mm displacement being attained.

Using alpha plaster increases initial loading capacity but detracts from the ability of a wad to elongate and continue to bear tensile loading as it continually deforms. Applying this to an in situ fibrous plaster ceiling wad, the ability to retain continuing tensile loading capacity while deforming prior to failure is useful and important in practice. It can provide a visual indication during surveillance and inspection that something is not right with the wad and potentially plaster micro-cracking or fibre degradation has occurred (even if not immediately visible to the naked eye) in a roof or ceiling space, which may have reduced visibility and limited lighting.

In line with the previously published values in Table 1, in which the compressive strength of the alpha plaster is 55 MPa compared to 13 MPa for the beta plaster, the results in this study demonstrate that alpha plaster is notably stronger than traditional beta plaster and therefore while chemical properties may be similar, mechanical properties and behaviour are not.

While alpha plaster offers initial strength, the results show that it does not allow the continuing tensile contribution of the fibrous reinforcement or steel wire to be fully realised to their maximum potential under increased loading and extended deformation, and breakage is sudden and less gradual with larger displacements using alpha plaster. In addition, once one fibre bundle breaks, this places additional load-bearing burdens on surrounding fibres, causing them to also break resulting in an unzipping or domino effect of multiple fibre bundles failing until complete specimen failure occurs.

The tests have confirmed that in line with the previously published values regarding the tensile strengths of the fibres in Table 1, the quadaxial fabric is stronger in tension than hessian scrim. Using the tensile strength values in Table 1 of 1700 MPa for quadaxial and 700 MPa for hessian scrim, along with the g/m² values of 133 for quadaxial and 102 for hessian, we can calculate that the force per unit weight of material of quadaxial fibres is nearly twice as high as hessian scrim (1.28 $×$ 10⁷ N/g for quadaxial compared to 6.86 $×$ 10⁶ N/g for hessian scrim). Quadaxial fabric performed well in tests and was also able to maintain loading capacity while elongating. B-Q-W, in which the quadaxial fibres were able to utilise their tensile capacity while the surrounding plaster spalled, achieved approximately twice the mean maximum force per mean weight of specimen as shown in Figure 5(c).

The decision to use quadaxial fabric instead of hessian scrim however may ultimately be an economic one, informed by methods of practice and frequency of surveillance and maintenance. Quadaxial glass fabric is a considerably more expensive material than traditional woven hessian scrim, but the results of this study suggest that from a mechanical behaviour perspective, it is a suitable potential alternative material to hessian scrim. As with hessian scrim, the use of quadaxial fabric results in a ductile composite wad specimen when used with beta plaster and a wire. With quadaxial fabric, there is the consideration of it being more resistant to moisture and biological/fungal attack over a long time period. However, any degradation consideration may be mitigated by regular inspection and maintenance, and effective replacement of traditional beta-hessian wads – new ‘like for like’ beta-hessian wads, or quadaxial fabric wads, can be added adjacent to any visually ageing historical wads in roof spaces during routine frequent inspection and repair.

As previously discussed by the authors Dams et al. (2024), there is a significant redundancy in traditional beta plaster and hessian wads applied according to the traditional rule of thumb of four wads per square metre, even more so with the inclusion of a wire. A square metre of fibrous plaster ceiling with timber laths, which may be as little as 5 mm thick in places, would typically be expected to weigh in the region of 10–16 kg. Therefore, with the traditional four wads per square metre rule, each wad would have been expected to take a typical loading of approximately 4 kg (this would be greater if a particular or individual feature was also supported). Considering the load capacities recorded by newly manufactured wads, this shows a huge factor of safety already for traditional materials using hessian scrim, which would only be increased further using quadaxial fabric.

Maintaining a large redundancy can be reasoned to be a sensible course of action for aged, historic fibrous plaster ceilings, which are often well over a century old. There remains the risk of unpredictable long-term degradation of the materials due to moisture and fungal-induced biodeterioration, which can significantly compromise the mechanical properties of the material (Maundrill et al., 2023). There is also a range of potential dynamic loading scenarios to which the ceiling could be subjected, such as live loads due to dropping lines for lighting rigs and speaker trusses, building movement due to ground displacements or seismic activity, liquid water roof leaks, or a dropped object or misplaced foot. Dynamic loading scenarios for historic fibrous plaster ceilings have previously been investigated by the authors in laboratory and in situ experimentation (Dams et al., 2026).

Figure 9 contains a selection of XRT images of specimens. Steel wires are highlighted in green and sliced images contained proof that steel wire breakage in specimens was confined to the centre of the wires where they had been twisted prior to the application of plaster-soaked fibre scrim. The end ‘looped’ parts of the wire are intact, as are the upper and lower lengths where the bulking of the specimens begins. An image of a steel wire still intact at the location of twisting following a test is also shown in Figure 9(f), which contrasts with the wire breakage at the twisting location in Figure 9(e).

In the XRT images, fibrous reinforcement is seen as voids along both scrim axes, whether as ‘dots’ in the cross-section or as elongated voids. Looped-twisted wires have been previously demonstrated by the authors to be a suitable method for use with wads (Dams et al., 2024) and results in this study further emphasise that the twisting of the wire is a sensible and suitable practice, enabling fibres to be fully and uniformly wrapped around and encasing the wire. However, it should be noted that care should be taken not to over-tighten the wire when twisting, and therefore risk inducing strain-hardening effects, making complete breakage of the steel wire at high deformations more likely to occur.

This paper has investigated modern alternative repair materials for fibrous plaster wads with large-scale specimens representing real-world practice for the first time. Fibrous plaster wads are crucial elements holding up fibrous plaster ceilings in historic and heritage buildings such as theatres. Traditional wads consist of a steel wire, natural hessian fibres, and gypsum plaster and this study looks at modern alternative repair options, complementing the use of traditional materials.

The long-term performance of fibrous plaster wads is influenced by the intrinsic material properties of the composite system. The ductility provided by fibres is critical to the stability of a ceiling. It enables the wad to redistribute stresses that may be potentially caused by such factors as routine human occupancy and actions resulting in impacts or minor deflections of steel or timber supporting members, movement of the building envelope, and thermal cycling. Plaster is a brittle material and ductile fibrous and steel wire reinforcement that remains flexible over decades can accommodate minor strains and displacements without compromising structural integrity, thereby extending the service life of the ceiling’s suspension system. Adhesion – both between the wad and its ceiling substrate, and within the plaster matrix of the specimen itself – can also affect long-term performance. Degradation processes such as fibre decay or surface contamination can reduce bond strength and lead to progressive failure modes. In conservation settings, any consolidation or repair strategy must ensure compatibility of modern alternative repair materials with the historic substrate to avoid creating differential stresses. Quadaxial fabric has been demonstrated to be a sympathetic and viable alternative to traditional hessian scrim. Gypsum plaster is inherently hygroscopic and moisture can additionally promote biological decay of organic fibres such as hessian scrim, resulting in a reduced tensile capacity and ductility over time (Maundrill et al., 2023).

Regular ongoing surveillance and maintenance of wads in historic roofs is an important component of conservation practice. Future inspection and maintenance may be influenced by material choice. RE Aramid Gel™ increases strength and possesses ductility; a question would be as to how future inspection of a historic wad or ceiling element sprayed in the substance would be inspected. The material is malleable and ductile following curing and does peel away from a surface; peeling away and reapplying may work for a single wad, but a wider area of material application may be a consideration for future inspection strategies. In line with industry practice, any historic substrate would need to be cleaned with accumulated dirt and mould hoovered and the substrate keyed to promote adhesion between the newly applied material and the historic existing material. For a full investigation into adhesion/pull-off tests and application over a wider ceiling area, featuring the materials in this study, the reader is referred to Dams et al. (2023). Strategy again may be informed by material choice and wad inspection has an advantage in comparison to a wider ceiling area in that a wad could be augmented with new material, or a new hessian scrim (or quadaxial fabric) wad can be added and placed adjacent to an aged wad. This may be a logistically preferable strategy, especially if a historic wad has previously been covered or sprayed with newer material previously and the historic wad is not immediately visually accessible.

A bespoke tensile rig designed by the authors applied tensile loading to samples and both observation and XRT complemented visual observation to assess the failure mechanisms. It was discovered that quadaxial fibres performed well in tensile tests while modern alpha plaster, though stronger than traditional beta plaster, provided a more brittle specimen with fibres and wires unable to elongate and more prone to sudden breakage. The paper contributes towards a scientific quantification and understanding of the tensile behaviour of fibrous plaster wads to complement existing empirical knowledge and help preserve heritage buildings.

Understanding how material properties are affected by real-world building conditions and occupational practice enables conservators to formulate interventions that strategically add new wads or stabilise existing wads, maintaining the structural and cultural integrity of the ceiling and preventing future problems that might create larger repair scenarios, requiring greater quantities of repair materials. In sustainability terms, the reduction of the quantity of raw materials required is also important as energy is required to calcine raw gypsum, and reductions in the quantities of gypsum plaster materials used ultimately help conserve the environment and mitigate the carbon footprint. Sustainability considerations also apply to fibres, with quadaxial glass fabric differing from hessian scrim in that quadaxial glass is an engineered composite textile produced through processes requiring energy, including glass melting at high temperatures and mechanical alignment of multiple fibre orientations. The fabric performs well in offering tensile and multidirectional reinforcement properties in addition to durability and moisture-resistance advantages, but also results in embodied carbon and end-of-life recycling/disposal considerations in contrast to hessian scrim, which is from a renewable and biodegradable agricultural plant.

This study has quantified the properties and analysed the failure mechanisms of full-scale, real-world representative fibrous plaster wads subjected to tensile tests. Fibrous plaster sample groups for test specimens featured both traditional beta plaster with hessian fibre reinforcement, along with modern potential replacement material options alpha plaster, quadaxial glass fibres, and RE Aramid Gel™.

Wads tests revealed that quadaxial fabric, when used in place of traditional hessian scrim, significantly increased the tensile loading capabilities of wadding ties, both with and without steel wires included, increasing mean maximum loading capacity from ≈3 to 3.5 kN (using hessian scrim and a wire) to ≈8 kN using beta plaster and a looped-twisted steel wire. The use of stronger alpha plaster instead of traditional beta plaster increased the initial maximum tensile loading capacity, but reduced the ability of the wad to deform, with fibres and steel wire unable to elongate, which resulted in a less ductile composite element. Spraying beta plaster and hessian wad specimens with RE Aramid Gel™ also increased the tensile loading capacities of wads to ≈4 kN with a wire and ≈2 kN without a wire. Assuming a consistent CSA of 1590 mm² to give an indication of maximum stress values, these ranged from 1.59 MPa (beta-hessian-no wire-aramid gel) to 5.15 MPa for beta-quadaxial fabric-wire and 5.52 MPa for alpha-quadaxial fabric-wire wads. For comparison, equivalent beta-hessian wads would result in tensile strengths of 1.10 MPa (without a wire) and 2.39 MPa (with).

A typical failure mechanism of wads is initial micro-cracking of the plaster, followed by the elongation and rupture of the fibres, along with the elongation and potential breakage of the steel wire occurring at high displacements. The ability of fibres and wires to elongate within a stronger and stiffer alpha plaster matrix was impeded and both fibres and wires were more likely to break in lower displacements than when using beta plaster. Quadaxial fabric demonstrated the ability to withstand ≈2 kN (200 kg) of tensile loading up to the maximum 40 mm displacement of the tests when used in conjunction with beta plaster; this was observed both in specimens without a wire and also with a wire, once a breakage in the wire had occurred. This capacity was not observed when used in conjunction with alpha plaster.

This study provides a significant contribution to the understanding of fibrous plaster wads and the modern repair material options, which continue to play a crucial part in architectural and cultural heritage internationally. Quantification and understanding of material properties will strengthen conservation industrial practice to ensure fibrous plaster ceilings in public and private buildings remain safe for the public and are preserved for future generations.

Barrie Dams: writing – original draft, writing – review & editing, conceptualisation, methodology, validation, investigation, formal analysis, visualisation. Mansi Chavan: investigation, formal analysis, visualisation. Martin P. Ansell: conceptualisation, writing – review & editing, supervision, methodology. John Stewart: conceptualisation, writing – review & editing. Marion Harney: writing – review & editing, supervision. Richard J. Ball: conceptualisation, writing – review & editing, supervision, funding acquisition, project administration, methodology.

The authors gratefully acknowledge the funding of the Leverhulme Trust, grant number RPG-2021-147 and the support of Historic England. The authors gratefully extend thanks to the following: Robin Harrison, Russell Hempstead and the whole team at Hayles and Howe Ornamental Plasterwork and Scagliola, Bristol, UK; Gary Buckley, Jon Riley and the team at Locker and Riley Artisans in Plaster, South Woodham Ferrers, Chelmsford, UK; William Bazeley, Martin Naidu, and Neil Price (Department of Architecture and Civil Engineering, University of Bath, UK); Florence Richardson (Department of Mechanical Engineering, University of Bath, UK); Rod Stewart and Eric Stewart (Historic Plaster Conservation Products, Ontario, Canada); Claire Appleby (Theatres Trust, London, UK); and Robin Townley (Association of British Theatre Technicians, London, UK). The data for supporting this manuscript are available from the dataset for the results of fibrous plaster tests, University of Bath Research Data Archive, 10.15125/BATH-01383.