https://orcid.org/0000-0002-3524-3833
Abstract
Successfully detecting and treating mould on cultural heritage to a ‘safe’ level is an important concern due to the variety of remediation approaches, substrates and health hazards posed by such fungi. In conservation, rapid adenosine bioluminescent swab testing has been used in two main applications: to identify if fungi are present, and to attempt to quantify if a remediated object is ‘clean enough’ for regular use or storage. A literature review across the food hygiene, healthcare and conservation sectors was combined with a simple lab experiment comparing RLU (Relative Light Unit) values obtained using a Kikkoman PD-30 lumitester with LuciPac Pen swabs on various surfaces against low magnification, high magnification and SEM imaging of the same samples. Results show it is impossible to establish numerical benchmarks for ‘clean enough’ and furthermore confirm that the devices are not ideal for diagnosing what is and is not mould. However, tracking the efficacy of remediation processes with the devices was successful when three sample areas were compared: a visually clean area before treatment, a visually mould damaged/dirty area before treatment, and the latter after a mould remediation treatment. This suggests that rapid adenosine bioluminescent swab testing can provide supporting evidence to conservators to make more informed decisions about how effective cleaning processes are for a particular artefact substrate.
Résumé
« Comment évaluons-nous les niveaux de moisissures? Test de détection rapide de bioluminescence à l’adénosine par écouvillonnage en conservation »
Détecter et traiter avec succès les moisissures sur le patrimoine culturel à un niveau « sûr » est une préoccupation importante en raison de la variété des approches d’assainissement, des substrats et des risques posés par ces champignons pour la santé. Dans le domaine de la conservation, les dépistages rapides par écouvillon bioluminescent à l’adénosine ont été utilisés dans deux applications principales: pour identifier si des champignons sont présents et pour tenter de quantifier si un objet assaini est « suffisamment propre » pour une utilisation ou un stockage normal. Une revue de la littérature dans les secteurs de l'hygiène alimentaire, de la santé et de la conservation a été associée à une simple expérience en laboratoire en comparant les valeurs RLU (Relative Light Unit) obtenues à l'aide d'un luminomètre Lumitester® PD-30 Kikkomen avec des écouvillons de LuciPac® Pen sur diverses surfaces avec un faible grossissement, un fort grossissement et une imagerie SEM des mêmes échantillons. Les résultats montrent qu’il est impossible d’établir des références numériques pour « suffisamment propre » et confirment en outre que les appareils ne sont pas idéaux pour diagnostiquer ce qui est ou non de la moisissure. Cependant, le suivi de l'efficacité des processus de décontamination avec les dispositifs a été couronné de succès lorsque trois zones d'échantillon ont été comparées: une zone visuellement propre avant le traitement, une zone visuellement endommagée par les moisissures/salie avant le traitement et cette dernière après un traitement de décontamination des moisissures. Cela suggère que les tests rapides par écouvillon bioluminescent à l'adénosine peuvent fournir des preuves aux restaurateurs pour prendre des décisions plus éclairées sur l'efficacité des processus de nettoyage pour un substrat d'artefact particulier.
Zusammenfassung
„Wie bewerten wir Schimmelbefall? Prüfung der Parameter von biolumineszenten Adenosin-Schnelltupfern in der Konservierung“
Die erfolgreiche Erkennung und Behandlung von Schimmelpilzbefall an Kulturgütern auf einem ‘sicheren’ Niveau ist ein wichtiges Anliegen, da es eine Vielzahl von Sanierungsansätzen, Substraten und Gesundheitsgefahren gibt, die von diesen Pilzen ausgehen. In der Konservierung wurde der Adenosin-Biolumineszenz-Schnelltest mit Tupfern vor allem in zwei Bereichen eingesetzt: um festzustellen, ob Pilze vorhanden sind, und um zu versuchen, zu quantifizieren, ob ein saniertes Objekt ‘sauber genug’ für die regelmäßige Nutzung oder Lagerung ist. Eine Literaturrecherche in den Bereichen Lebensmittelhygiene, Gesundheitswesen und Restaurierung/Konservierung wurde mit einem einfachen Laborexperiment kombiniert, bei dem RLU-Werte (Relative Light Unit), die mit einem Kikkoman PD-30 Lumitester mit LuciPac-Stifttupfern auf verschiedenen Oberflächen ermittelt wurden, mit niedrigen und hohen Vergrößerungen sowie REM-Aufnahmen derselben Proben verglichen wurden. Die Ergebnisse zeigen, dass es unmöglich ist, numerische Richtwerte für ‘sauber genug’ festzulegen, und bestätigen außerdem, dass die Geräte nicht ideal sind, um zu diagnostizieren, was Schimmel ist und was nicht. Die Verfolgung der Wirksamkeit von Sanierungsprozessen mit den Geräten war jedoch erfolgreich, wenn drei Probenbereiche verglichen wurden: ein visuell sauberer Bereich vor der Behandlung, ein visuell schimmelgeschädigter/verschmutzter Bereich vor der Behandlung und letzterer nach einer Schimmelsanierungsbehandlung. Dies deutet darauf hin, dass schnelle Adenosin-Biolumineszenz-Tupfer-Tests Restauratoren dabei unterstützen können, fundiertere Entscheidungen darüber zu treffen, wie effektiv Reinigungsprozesse für ein bestimmtes Artefaktsubstrat sind.
Resumen
“¿Cómo evaluamos los niveles de moho? Comprobando los parámetros usados en las pruebas rápidas con hisopos bioluminiscentes de adenosina en conservación”
Detectar y tratar con éxito el moho en el patrimonio cultural hasta un nivel ‘seguro’ es una preocupación importante debido a los diversos enfoques de remediación, sustratos y peligros para la salud que plantean estos hongos. En el ámbito de la conservación, las pruebas rápidas con hisopos bioluminiscentes de adenosina se han utilizado en dos aplicaciones principales: para identificar la presencia de hongos y para intentar cuantificar si un objeto tratado está ‘suficientemente limpio’ para su uso o almacenamiento habituales. Se combinó una revisión bibliográfica de los sectores de la higiene alimentaria, la asistencia médica y la conservación, con un simple experimento de laboratorio en el que se compararon los valores de RLU (Relative Light Unit, unidad de luz relativa), obtenidos utilizando un luminiscente Kikkomen PD-30 con hisopos LuciPac en varias superficies con imágenes de las mismas muestras de bajo aumento, alto aumento y SEM. Los resultados muestran que es imposible establecer puntos de referencia numéricos para determinar si está ‘suficientemente limpio’ y, además, confirman que los dispositivos no son ideales para diagnosticar qué es moho y qué no es moho. Sin embargo, fue efectivo el seguimiento de la eficacia de los procesos de remediación con los dispositivos cuando se compararon tres zonas de muestra: una zona visualmente limpia antes del tratamiento, una zona visualmente dañada/sucia por el moho antes del tratamiento y la última después del tratamiento de corrección del moho. Esto sugiere que las pruebas rápidas con hisopos bioluminiscentes de adenosina puede ofrecer pruebas que apoyen a los conservadores, y así, poder tomar decisiones más informadas sobre la eficacia de los procesos de limpieza para el sustrato de un objecto determinado.
摘要
“如何评估霉菌水平？用于保护的快速腺苷生物荧光拭子参数测试”
由于修复方法、基底和真菌对健康的危害多种多样，因此成功检测文化遗产上的霉菌，并将其处理到‘安全’水平是一个重要问题。在文物保护中，快速腺苷生物荧光拭子测试主要应用在两个方面：确定真菌是否存在，以及量化经过治疗处理的文物是否‘足够干净’，以便正常使用或存放。本文对食品卫生、医疗保健和保存修复领域的文献进行了回顾，并结合简单的实验室实验，将借助Kikkoman PD-30荧光检测仪和LuciPac检测棒取得的不同表面的RLU（相对光单元）数值，与相同样品的低倍放大率、高倍放大率和扫描电镜成像进行比较。结果表明无法建立‘足够干净’的数值基准，且进一步证实了这些设备并非诊断霉菌的理想工具。不过，在对三个样本区域进行比较时，使用这些装置跟踪修复的效果很成功：处理前的视觉清洁区域、处理前的视觉霉菌损坏或污脏区域以及处理后的霉菌修复区域。这表明，快速腺苷生物荧光拭子测试可为文物保护人员提供支持性证据，使他们能够就特定文物基底的清洁有效性做出更明智的决定。
Introduction
The ability to detect and remediate mould to a ‘safe’ level for users of cultural heritage objects is an important concern for collection caretakers. The variety of historic and contemporary approaches to remediation, as well as the growing body of research on the health hazards posed by fungi on heritage surfaces have not always provided straightforward methods to determine the level of risk a mould outbreak poses, or how best to determine if treatment methods are successful in ameliorating risks.Footnote1 To this end, rapid bioluminescent swab testers—or ATP monitoring devices—have become more frequently used in conservation in recent years.
Rapid swab testing (ATP monitoring) was developed and used primarily in the food industry to help verify cleaning routines of high-touch or food preparation surfaces,Footnote2 although it is also used in the healthcareFootnote3 and wastewater management fields, albeit with the latter using a slightly different swab configuration.Footnote4 The devices work by the detection of trace amounts of adenosine triphosphate (ATP) in samples collected by swabbing surfaces. As ATP is found in all living cells and decreases after cell death or lysis—decaying to adenosine diphosphate (ADP) and adenosine monophosphate (AMP)—it is used in these devices as a marker for organic contamination of a surface.Footnote5
Rapid swab tests are conducted with a disposable swab, a reagent test tube kit and a luminometer device. Generally, a sterile swab is removed from a test tube, swabbed over a surface and placed back into the tube. The tube is then shaken to mix the reagents it contains with the sample, which is then inserted into a luminometer device to take a reading. The by-product created from the reaction of the swabbed organic material and proprietary enzyme-based reagents is a small output of light—bioluminescence. The luminometer reads the light intensity emitted from the reaction and gives a numerical result in what are called Relative Light Units (RLUs). This output is a relative indication of the level of cleanliness of a surface, i.e. the higher RLU number equals more contamination of a surface by organic contaminants or deposits. Correspondingly, the larger the area of a given surface that is swabbed, the larger the quantity of contaminants that can be picked up, and the larger the resulting RLU reading will be.
The majority of swab testers test only for ATP, although one series of devices by Kikkoman (the Lumitesters PD-20, PD-30 or Smart) used in combination with their swabs can test for ATP+AMP (LuciPac Pen) or ATP+ADP+AMP (LuciPac A3 Surface swabs).
Rapid swab testers are an attractive tool for in-situ cleanliness monitoring as each individual test process is relatively quick to complete (usually marketed as 10–30s per test) and there is minimal training required to use the devices. As such, rapid bioluminescent swab testers have started to find use in museums and archives as a method for detecting mould, which is a generally common microbiological contaminant found in collections.
Although not an exhaustive list of every instance where luminometers have been used for heritage purposes, reviews uses of ATP in the heritage context as known to the authors, and where these focus on exploring the parameters of device use for mould or for other health-related benchmarking. Of the 13 swab test users identified, the majority used the technology for mould assessment, with only two users identifying analysis of other organic contaminants. Over half of the instances in used Kikkoman branded luminometers and ATP+AMP LuciPac Pen swabs. As ATP luminometry is relatively new in the conservation context there is as yet no field-wide standardisation, as is suggested by the various methods for swabbing and interpreting the results listed in . In heritage, the rapid swab testers most commonly appear in applications seeking to identify if mould is present or quantifying if an object is ‘clean enough’ for use after mould remediation.
Table 1 Cultural heritage instances of ATP luminometer use for mould or health-related benchmarking.
Reviewing the manuals produced by ATP monitoring manufacturers noted to be of use by conservators, the devices generally have similar (though not identical) methods for the swabbing technique, shaking time, baselines RLUs and swabbing surface areas. compares suggested sampling techniques of the branded devices.
Table 2 Luminometer and swab specifications.
Individual manufacturers will also suggest different thresholds for what RLUs are considered pass (adequately cleaned), caution and fail (unclean) for surface types such as ‘rough’ vs. ‘smooth’, and non-planar surfaces such as tap rims and hands. Most ATP monitoring system manuals indicate that it is recommended that RLU pass/fail limits are adjusted to the individual needs of the users or space.
There is no current standard for Relative Light Unit measurements so any one brand’s ratio of ATP to RLU will differ to another’s. As an example, Hygiena’s website states that ‘all ATP systems measure femtomoles of ATP (1×10–15 moles). Hygiena keeps it simple: 1 femtomole of ATP equals 1 RLU. On Neogen’s system, 1 femtomole of ATP equals approximately 10 RLU’.Footnote6 Hygiena supplies a conversion tool to convert RLU values from various systems (they are the only manufacturer to offer this). However, as they note, the RLUs on their comparison table are only suggested levels and are based on ‘experience with customer conversions’.Footnote7
Comparative studies of devices from the food science or hygiene field show that devices from different brands will vary in the following factors:Footnote8
linearity —the ability for the device to accurately measure varying amounts (dilutions) of pure ATP;
repeatability —the ability for the device to consistently measure a specific amount of ATP;
sensitivity —the lowest amount of ATP detectable by the device; and
lowest level of quantitation—the ability for the devices to accurately and consistently measure low levels of ATP.
Two experimental studies in the literature report on their use of >1000 data points from known quantities of both purified ATP, and/or bacteria/yeasts and dilute organic matter which were tested with devices from various manufacturers, with differing results on the performance of luminometers across these four factors.Footnote9 Of the seven devices tested in these experiments, the three brands that generally emerged as higher ranked on metrics like precision (linearity and repeatability) and low level detection (sensitivity and lowest level of quantitation) were Hygiena, 3M and Kikkoman.
The results of these two studies that have the most consequence for use of ATP monitors in cultural heritage settings relate to the ability for the devices to repeatedly measure low levels of organic residues. This is of particular importance if the devices are used as a method to ‘pass’ or ‘fail’ the cleaning of mould-affected artefacts as they test if there are still organic residues on the surface of the artefact after cleaning. Both studies show that there are inconsistencies in repeatability with all ATP monitoring devices (see note 12), and while the majority of results are within a median range of error, all the devices tested occasionally showed unexpectedly high or low RLU results.Footnote10
While it is not expected that an in-field device should be absolutely precise or accurate, Greg Whiteley et al. suggest that the data recorded indicate a ‘regularity of outlier [swab results] both above and below the interquartile ranges of samples’.Footnote11 This is reiterated in the report by Brian Kupski et al., the ‘Silliker Report’, where it is noted that ‘the CV [coefficient of variance] % values increased as the limits of detection were approached. This is expected because the closer to the detection limit, there is much less ATP to measure and there is more variability in the measurement’.Footnote12 As explained in another article from 2016 by Whiteley with different colleagues, ‘in studies using a single sampling point where the CV is above 0.4 [40%], there is a 20% chance that any reading could be wrong by a factor of two’.Footnote13
Results from the two studies suggest that it is not reliable to use only a single swab as the possibility of random errors (randomly high or low readings) may constitute a false positive or false negative RLU. Other studies in the heritage field include one carried out by the Tokyo Research Institute for Cultural Properties (Tobunken) which came to a similar conclusion and suggests that ‘because measurements can vary widely, it is believed that the RLU values need to be recognised as approximate figures and be understood in terms of 10 to the power of 1, power of 2, power of 3, and so on’.Footnote14 To reliably carry out ATP monitoring, it was also suggested by Whiteley and colleagues in 2016 that a method of sampling where two or more swabs are taken would improve reliability of the sampling as it helps prevent false positive or negative results.Footnote15
Beyond the limits of the technology, in adapting ATP monitors for use in conservation, other factors should be taken into account. A review of the literature suggests that changes to sampling techniques and differences in the substrates found across cultural heritage will affect the results of ATP monitoring.
The recovery of trace amounts of adenosine triphosphate (ATP) from swabbed surfaces is influenced by many factors, including the physical nature of the surface swabbed, whether the swab is pre-moistened, as well as the swab material and shape, among others. These factors have been summarised by Gilbert Shama and Danish Malik (in 2013) and Ralph Moon (in 2021).Footnote16 Of these, one of the main issues that has consequences for conservation is that many of the swabs are packaged pre-moistened by manufacturers. This is a consideration when selecting a device as it may be preferable to choose one that allows for dry swabbing as this avoids introducing moisture onto sensitive hygroscopic surfaces (such as paper or leather) and minimises risks of tidelining or other undesired effects. However, in healthcare experiments on bacteria, dry swabbing was demonstrated to reduce the swabbing efficiency—the ability of the swab to pick up surface contaminants.Footnote17 The choice of dry or wet swabbing should be noted in sampling methodologies, and kept consistent for any comparative purposes. With wet swabbing there can also be a question around the nature of the cleaning agent used as if it is non-oxidising like a QAC (quaternary ammonium compound) it can interact with the firefly luciferase used in the luminometer. Because the authors are focussed on dry swabbing and proposing a comparative swab method, the potential for quenching or enhancing effects that some cleaning agents may have is less of an issue in this study.
Another factor to note is that the porous, often fragile substrates of artefacts are qualitatively different from the surfaces ATP monitors were devised for. For example, the Kikkoman PD-30 manual gives benchmark values for plastic cutting boards, knives and pots to mention some, but no porous materials are given reference values.Footnote18 While base reference values are often given for smooth plastic or stainless steel surfaces, artefact surfaces are unlikely to be as cleanable as these. It is also necessary to be aware that the swabbing efficiency of textured anisotropic surfaces are likely to produce lower RLU values due to the swab not reaching below the top surfaces of the artefact.Footnote19
The reference values given by manufacturers are also based on a particular sample size. While many manufacturers suggest approximately a 100cm2 sample area (10×10cm) (see ) this is difficult to achieve on three-dimensional surfaces, particularly where mould may not be all over the surfaces, or on smaller scale items. In the healthcare field a standard sampling area of 10cm2 (2×5cm) has been suggested,Footnote20 albeit in conservation, sample sizes have tended to be smaller (see ). Some authors have scaled their RLU results (through multiplication or division) to create equivalent surface area results as needed.Footnote21
ATP monitoring devices are designed to be used to assess cleaning mainly for high-touch environments and food service surfaces to indicate the presence of recently deposited organic matter between regular sanitation or cleaning procedures. For their use in heritage where artefacts are often not subjected to routine cleaning, organic surface contamination is likely to be older with declining viability or declining detectable ATP. It has been shown that ATP deposits—from microorganisms, mixed organic deposits and purified ATP—dried on a surface can persist for longer than 29 days and that ATP signals can occur even when microorganisms are non-viable.Footnote22 While ATP signals can persist, the decision by some conservators to use ATP+AMP swabs produced by Kikkoman is based on the assumption that the higher sensitivity of ATP+AMP swabs will aid in detecting any form of adenosines, including non-viable mould.Footnote23 Though there is contradictory information from competing manufacturers about the sensitivity of ATP swabs compared with ATP+AMP or ATP+ADP+AMP swabs,Footnote24 in one conservation experiment comparing results from an ATP-only device against similar samples with an ATP+AMP device, the ATP+AMP swabs were shown to produce higher RLU results on similar locations,Footnote25 perhaps supporting the view that ATP+AMP or ATP+ADP+AMP-based devices may be more useful to heritage. This is particularly advantageous because the end-goal of mould remediation cleaning processes is to remove as much residue as possible, regardless of its viability.Footnote26
One possibility for resolving the uncertainties about standardised reference RLUs is to create bespoke reference values. It is possible to consider borrowing techniques from the water damage recovery industry where the moisture content of a dry area of a material in a similar environment to a waterlogged example of the material is used as a standard moisture content value.Footnote27 Applying this technique for ATP monitoring, comparing a clean non-mouldy area to a mouldy area, could help define a mould remediation cleanliness standard for a particular substrate. This also may help reduce some of the difficulty with assessing RLU results of differently sized sample areas, as the same sized samples can be taken on the same substrate. This also allows for adapting swab sampling techniques to the requirements of the substrate. For instance, while many substrates permit gently dragging a swab across the surface in multiple directions, other more brittle or delicate surfaces may require sampling by very gently rolling.
In the literature reviewed to determine ‘how clean is clean enough’, guidelines on mould remediation in indoor spaces generally recommend that visual assessment or surface sampling—such as wiping with a cloth or tape lift sampling to visually determine if mould fragments are still on the surface—are adequate for small mould outbreaks.Footnote28 Since most of remediation guidelines are aimed at non-heritage buildings, a common solution to visible mould damage on porous surfaces (such as gypsum wall board) is to remove and discard the material entirely.Footnote29 Specific remediation guidelines for the retention of severely damaged porous surfaces (such as cultural heritage artefacts) were not found.
The conclusions drawn here by the authors after the literature review are as follows.
It is not recommended to use the ATP monitoring devices as a method to determine whether or not a surface deposit is mould, or to identify if mould is present. The ATP monitoring devices are not mould-specific devices and will read ATP from other organic sources such that a high RLU reading does not necessarily equate to mould. Swabs collect material only from the surface, and therefore cannot be used to assess fungal matter embedded in porous substrates such that a low RLU reading does not mean that mould is not present.
In reviewing the literature from health and mould remediation guidelines, it is apparent that the ATP monitoring devices cannot be used to establish whether an artefact is ‘clean enough’ after mould remediation. This is in part because there is currently no established guidance on what constitutes a ‘safe’ level of fungal exposure.Footnote30 While there is general agreement that exposure to mould is hazardous to human health, it is difficult to establish a standard below which levels can be considered safe. This is in part due to the difficulties of measuring and quantifying exposure from specific sources, as well as the large number of symptoms and health outcomes, which vary widely by individual.Footnote31 Minimising exposure by cleaning damaged artefacts and undertaking preventive measures against further contamination continues to be the most effective way to minimise the risks associated with exposure.Footnote32
Experimental
Simple conservation lab experiments were designed to test the practical utility of the swabs in a conservation mould remediation context. Visual judgements easily made by conservators, such as selecting regions of heavy mould, light mould, very light mould and visually clean areas, as well as common imaging techniques such as micrographs in low magnification, high magnification and SEM imaging of the same samples, were all compared with swab results in Relative Light Units from an ATP+AMP luminometer to determine if the luminometers provided valuable added information.
1 Luminometer
Experiments employed a Kikkoman™ PD-30 Lumitester™36 with LuciPac™ Pen swabs,Footnote33 which detect both ATP and AMP.Footnote34 The sensitivity of this system to produce a signal from both ATP and AMP aligns with the goals of mould remediation, where surface cleaning aims to reduce as much of the possible surface biological load, regardless of whether the mould structures are viable or non-viable. As per the manufacturer’s recommendations, the swabs are stored in the refrigerator to extend the lifespan of their enzymatic reagent packets. Prior to testing, the swabs need to be removed from cold storage and allowed to come to room temperature (see note 34).
2 Swab technique
A consistent swab technique () was adapted from the manufacturer and by others as reported in the literature to improve repeatability and consistency in collecting samples from as similar a total surface area as possible while being practical for application to heritage artifacts.Footnote35 This technique compensates for anisotropic surfaces by completing four sampling passes, two along each axis. Gently dragging the swabs in an easy to reproduce pattern across the sampling area was found to be the easiest action to repeat.Footnote36 Maintaining a low angle between swab handle and sample, as well as rotating the swab itself by 90 degrees after each sampling pass to expose a ‘clean’ swab surface for the next pass maximises swab surface area contact with the sample area.Footnote37 It’s important to note that clean masks were used for each sample. Although Kikkoman recommends larger sample sizes (approximately 100cm) conservators need to be practical with testing protocols needing to accommodate hard-to-reach areas (e.g. mould in the gutter of a book) and relatively finite areas (e.g. mould only found on the spine of a book or on smaller artefacts in general), while also limiting the risk of damage posed by swabbing fragile surfaces. As seen in , sample areas in this study were delineated with custom 4-mil polyester (PET) film masks with 3×3cm apertures (9cm2). This is similar in size to previous conservation applications as discussed with colleagues elsewhere.Footnote38
Fig. 1 Diagram of swab mask and standard swabbing technique used for RLU sampling.
To collect a sample, the LuciPac Pen sterile swab was removed from the sleeve containing the reagents and swabbed dry across the sample surface as in . After collecting the sample, the swabs were directly inserted back into their reagent sleeve, the two foil diaphragms punctured, and the swab/reagent package shaken for 30s taking care to ensure thorough and consistent mixing of reagents and sample before inserting into the luminometer to read the resulting RLUs.Footnote39
3 Sample substrates
To test the utility of the swabs for tracking the course of a mould remediation treatment, two didactic books exhibiting mould growth were selected (), providing a number of test sites on the same types of paper. Both books are modern case-bound volumes commonly found in general collections. Although the mould on the two books is different—they are from different source locations and so became mouldy in different environments—the experiment is designed to look at the total reduction of mould, not compare starting levels of mould. In both cases, the books had been dried sufficiently to stop active mould growth. The books were selected due to the contrasting properties of their textblock papers: the paper of the first book from 1974 is robust and remains intact even in the areas of mould growth,Footnote40 while the paper of the second 1924 book is fragile and friable, particularly in the areas of mould growth.Footnote41 These papers from each will be described respectively as ‘intact’ and ‘fragile’ in the results and discussion sections.
Fig. 2 Two didactic mouldy books. On the right, the 1974 paper is quite robust and intact, despite mould growth; on the left the 1924 paper is fragile and friable, particularly in areas of mould growth.
Four different types of areas were sampled on each of the two paper types:
Heavy (H) mould: areas characterised by significant surface mould, easily visible to the naked eye. It is expected that surface cleaning can visibly reduce a significant portion of these surface deposits, although it is expected that embedded mould structures and staining will remain.
Light (L) mould: mould structures are still easily visible to the naked eye, but only a small change in visual appearance is expected at most with surface cleaning. Most of the visible mould is due to staining or mould structures embedded in the paper substrate.
Very light (VL) mould: areas where perceptible mould staining is present, but where there are little to no surface mould structures perceptible to the naked eye. Little change in appearance is expected with surface cleaning.
Visually clean (VC): areas of the paper that appear unaffected by mould damage. These areas provide background RLU values for a given substrate.
Surface cleaning was performed using latex-free polyurethane (PU) cosmetic sponges gently moving in multiple directions across the sample surfaces to lift as much soiling from the sample areas as possible. This type of surface cleaning was selected as it is a common method for cleaning paper substrates. It provided the additional experimental benefit that a clean sponge could be used for each sample area, ensuring there was no cross contamination between samples. The mould remediation was carried out in a Class 1 biological safety cabinet and the conservator was wearing nitrile gloves.
For each paper type, six H, L and VL and four VC locations were recorded before surface cleaning treatment (BT) and again in the same location after surface cleaning treatment (AT). A separate set of BT/AT data was collected on the fragile paper to document what happens when the swabbing action breaks through the substrate surface—either on the BT, the AT or on both samples. These samples will be described as ‘surface breakthrough’ in the results and discussion sections. The sampling areas selected for these tests would have visually been described as ‘very light’. A total of 100 data points were collected from the 3×3cm sample areas swabbed.
It should be noted that while swabbing itself represents a cleaning action, the treatment goal in this experiment was to lift as much biological surface contamination off each sample as possible using surface cleaning. As such, the cleaning action of the swab testing itself to record the BT sample simply advances this cleanliness goal—whether surface contamination was removed by swabbing or by the sponge, it was reduced in a consistent manner in each case prior to recording the AT sample.
4 Microscopy and SEM imaging
To complement the luminometer data and help visualise the level of surface cleaning recorded with the RLU values, selected sample locations of each category (H, L, VL, VC) were imaged before and after surface cleaning with a Dino-Lite Edge AM4515ZT digital microscope at 12.5× magnification (called ‘low magnification’ in the results). Adjacent to these same selected locations, paired before and after cleaning 5mm diameter samples were also taken from areas exhibiting similar levels of mould growth using a Japanese screw hole punch (). High magnification photomicrographs of these paper samples were recorded by incident light microscopy using a Leica DM4 P microscope. The images were recorded with a 10× objective for a total magnification of 100×. The same samples were then placed on polished carbon planchets using double-sided carbon tape. The samples were then coated for 2 minutes at 20mA with gold-palladium using a Cressington 108 Auto sputter coater. Scanning electron microscopy (SEM) imaging was performed using a Hitachi SU3500 SEM operated at an accelerating voltage of 20kV in high vacuum mode.
Fig. 3 Example of sample locations for microscopy and SEM before (left) and after sampling with a Japanese screw punch (right). The black dashed square delineates the ATP/AMP sampling area. The white circles were identified as visually similar quantities of mould growth using a USB microscope. The area inside the black square was cleaned—the left circular sample was then taken to be ‘AT’; the right circular area outside the ATP/AMP sampling area was not cleaned, and represents the BT mould levels for imaging purposes.
Results
RLU results are visualised as box and whisker plots in . Box and whisker plots are useful to visualise how diverse datasets skew. The box shows where the central 50% of the data sit, with the median data point marked as a bisecting horizontal line. The average of the dataset is identified by an x, and the whiskers extend to the data maximum or minimum, within a range up to 1.5 times the length of the box (also called the interquartile range). Any data points outside that range would be identified as outliers;Footnote42 in this case, we do not have any data outliers. The H, L, VL and VC areas do group, but with broad areas of overlap in the RLU results. The BT groupings are more clear for the intact paper (which is both easier to clean and easier to sample), but the L and VL sample groups show virtually identical average values and central 50% of the data for fragile paper, which also shows a significant overlap with the intact paper. All groups except SB show a substantial reduction in RLUs, with a tightening of quartile and whisker spreads AT. These results help show the utility of the RLU devices for aiding a conservator’s visual assessment. For example, the L and VL samples are visually differentiable, but a conservator might easily underestimate the amount of biological matter present on a VL sample given how much the VL samples overlap with the visually L and H BT samples according to their RLU values. also demonstrates that SB sample data are fundamentally not useful for decision-making as there is far too much overlap in datasets BT and AT.
Fig. 4 Biological surface load (RLUs) on two types of paper before (left) and after surface cleaning (right). Sampling areas are categorised by visual quantity of surface mould before surface cleaning: high, low, very low and visually clean. For the fragile paper, those locations where sampling broke through the paper surface are separated into the category of surface breakthrough.
The individual RLU datapoints are tabulated in Appendix 1 and plotted in as BT and AT values. The BT/AT data pairs have been grouped into sets by paper type (fragile or intact) and by level of visual mould growth (VC, VL, L and H). Each dataset has then been ordered from lowest to highest BT RLU value to easily visualise the wide range of BT values. Both the table and the graph also document whether surface breakthrough (SB) occurred only on a BT or AT sample, or on both a paired AT and BT sample using a bolded line around the bar graph data point or an asterisk in the table. SB-1 broke through for the BT reading only, SB-2 through SB-4 broke through on the AT reading only, and SB-5 and SB-6 broke through for both BT and AT readings. The sample pairs in demonstrate several things: that the VC areas had low RLU values both before and after cleaning and, that while BT readings range widely (and are not necessarily grouped by visual characterisation of mould growth), the AT readings show a consistent reduction in surface contamination with surface cleaning remediation. All AT readings are much lower than their BT counterparts, with the exception of an F-VC-2, where the values are both 30 or less RLU (and likely near the detection limit of the device). The other exception is the SB dataset which demonstrates that if there is only breakthrough on the BT reading, with no breakthrough on the AT reading, the usual trend in reduction of biological surface load with cleaning can be seen (re. SB-1). However, if SB occurs AT or on both samples (SB-2 to SB-6), the data are much more chaotic, and become difficult to reliably interpret.
Fig. 5 RLU values recorded before and after surface cleaning treatment (BT and AT) on intact (I) and fragile (F) paper for visually clean (VC), very low (VL), low (L) and high (VL) visual regions of mould growth. The areas of surface breakthrough (SB) during sampling of the fragile paper are grouped to the right with dark outlines indicating whether the SB occurred BT, AT or during both readings. Each data grouping is sorted from lowest to highest BT RLU value.
The microscopy and SEM results for the intact paper are found in . Note that the low magnification BT and AT images are from the same location; the BT and AT image locations in high magnification and SEM are taken from locations showing similar levels of mould BT; the sampling required for high magnification and the coating required for SEM analysis does not permit the cleaning of the BT sample after imaging. The same samples were used for high magnification and SEM image sets. The images in confirm that surface cleaning, while useful to reduce mould structures on the surface, is unable to remove embedded mould. This is echoed in the RLU data in : no mould-contaminated sample set had RLU values that read as low as the VC sample set. In the imaging, mould structures seen as dark staining under low magnification, as red-brown hyphae and spore structures contrasted against creamy white paper fibres under high magnification, can be seen in SEM as the recognisable morphology of hyphae and spores as white structures contrasting with dark grey paper fibres.Footnote43
Fig. 6 Representative imagery of the intact paper sample in sample areas categorised as high (H), low (L), very low (VL) and visually clean (V) before and after surface cleaning. Imaged under low magnification (images are 3cm across), high magnification (scale bar=100µm) and SEM imaging (scale bar=50µm).
The low magnification images are of a magnification that is readily available in a conservation lab using a stereomicroscope or digital microscope. They document the kinds of reduction in staining for the H, L and VL sample sets that are visible to the naked eye during cleaning, but also show the remaining embedded mould staining that is expected of porous substrates after surface cleaning each sample grouping. The high magnification images help contextualise how the mould structures look on the surfaces of paper. Before treatment (BT) the mould structures are often above or obscuring the substrate surface. After treatment (AT) the embedded structures that remain are seen as ‘staining’ rather than obscuring the substrate.
The SEM images confirm a lack of mould structures in areas denoted as visually clean and the presence of mould structures in areas visually assessed as being ‘very low’ in surface mould. Coupled with the data in , this demonstrates that low RLU readings occur where surface mould structures were not present (VC samples) or have been substantially reduced by surface cleaning (AT samples relative to BT samples). They also confirm why the SB samples produce chaotic data given SB swabs have the possibility to access embedded mould structures that are not removable by the surface cleaning techniques employed.
Discussion
In the design of the experiment, the areas of mould on the two papers were grouped visually in zones of H, L, VL and VC. This visual grouping, while subjective, was included as this is the information that a conservator performing mould remediation usually has access to, that is, a qualitative evaluation of how much mould is on the surface and the visual feedback of how much that surface contamination has been reduced after surface cleaning. Separating the RLU data using the same visual feedback categories permits a comparison of how much added value swab data might provide to a conservator during the course of cleaning. Although relatively inexpensive compared to other types of analyses, the swabs are not without cost, and do generate waste. In order to determine their utility in conservation, it is necessary to assess whether they provide added value above and beyond simply showing that a surface is cleaner after remediation.
With the exception of the surface breakthrough data, both and show that surface cleaning of all areas of the paper using a combination of the swab sampling and sponge eraser does produce a reduction in the biological load on the surface. helps to confirm that this is due to a reduction in surface mould structures. This reduction is particularly evident in the change of scale in the inset after treatment. With only two exceptions, the cleaned surfaces read below 300 RLUs. While it could be tempting to use this to define a threshold level of what counts as clean for the papers in this study, this is problematic on several levels.
For example, beyond the lack of clear public health definitions of what is ‘clean enough’, the wide spread of the datasets suggests that establishing a threshold value isn’t advisable. Looking closely at graphical data in and the numerical data in Appendix 1, nine of 42 BT data points (not including VC areas) also fall below 300RLU (19%). However, shows that all of the non-surface breakthrough values could be cleaned, i.e. their RLU values are less after cleaning than before. As arguably a goal of mould remediation is to expose the user of an artefact to as low a quantity of mould structures as possible, the ability to further reduce the RLU values with surface cleaning—even in areas that start with a low BT RLU value—demonstrates that ‘clean enough’ cannot be tied to an RLU value.
On the other hand, the high magnification and SEM images of the L and VL sample areas in reinforce how the RLU readings can add value in the course of conservation cleaning. In the case of the L samples, the BT readings are low as most of the mould structures are embedded (so the swab does not detect them). Still, some surface mould structures can be removed with cleaning, decreasing the AT reading. The VL samples in have only a spot or two of surface mould, very close to the printed text. As such, they were not very observable with the naked eye, but could read quite high due to their localised concentration. During the course of a mould remediation it is easy for a conservator to feel confident that cleaning has had an effect in highly mouldy areas. It can be more challenging to judge when cleaning has had any effect in areas where visible structures are less noticeable, or where discolouration (due to embedded mould structures or the metabolic by-products of mould) are the predominantly observable type of mould damage. Using RLU swabbing to confirm successful cleaning in these areas could be useful for a conservator’s workflow, particularly if repetition fatigue sets in during long mould remediation projects.
Turning to the question of the effects of surface breakthrough on data interpretation, the images in show embedded mould structures that are present after surface cleaning—with some that are fully embedded, gathered around features in the paper fibre topography, or caught in interstices between fibres. Some of these residues are likely due to the nature of the chosen surface remediation technique using sponges as there are limits to the conformability of sponge erasers to the surface, so small crevices and interstices are unlikely to be completely cleaned. However, the remaining embedded mould structures also illustrate why surface breakthrough during sampling has such a strong influence on RLU data: if the swabs sample below the paper surface, they are accessing more biological matter. Looking at and the data table in Appendix 1, a BT SB sample (such as SB38, which was taken in a VL region) can read very high, or an AT SB sample (such as SB49) can almost double its corresponding BT reading. Looking at SB47 provides a case where, although the surface was penetrated during sampling, the RLU value remained low (RLU=107). In this case, there may not have been embedded mould structures present to sample. Attentiveness to whether a sample records only the surface or results in surface breakthrough is necessary to interpret data results.
Comparing data points to imaging results indicates that RLU values where the surface remains intact after swabbing generally correlate with the visual assessment of the level of surface mould. From the box and whisker diagrams in , the results, although scattered, do group according to these visual assessment categories. The areas assessed with a visually H level of mould show generally higher RLU values, while areas assessed as VC showed consistently lower values. The differences between low and very low RLU values are the least significant as they show a broad overlap in the range of RLU values, although either the median or mean values are lower for the VL samples. The high magnification and SEM imaging in demonstrate that the L and VL areas can be very similar in nature, but that even a small spot of mould (such as those seen in the BT VL SEM) can have a strong influence on the RLU readings. All high magnification and SEM micrographs indicate a reduction of surface mould after cleaning. However, the low magnification images show little change for the L and VL areas. It is in these zones where the use of the RLU values to confirm that cleaning has been achieved could have the biggest value-added impact for the conservator.
The SEM images confirm that it is not possible to use the ATP monitoring devices in a quantitative way to indicate that an item is free from mould. As a surface monitoring device, it can only describe the surface. It is also unlikely that the devices can be used in a field-standardised manner to indicate that an item is ‘safe’ after mould remediation. Despite manufacturer designs to use the devices in a ‘pass’ or ‘fail’ mode, as discussed in the literature, there is no current standard benchmark for acceptable contamination levels for porous artefacts. However, the VC reference areas designated in this study did provide a useful comparative anchor for the BT and AT results. Because RLU values are not mould specific, a VC sample area provides information on what a non-mouldy area tests like on a given substrate. This can provide a reference point for designing conservation remediation targets with AT values closer to a VC value being considered ideal, although, as the data in this study show, hard to achieve.
That some of the BT L and VL RLU values are in the same range as for the VC samples also reinforces the necessity to sample in multiple locations. These L and VL values could either be instrumental outliers, or simply represent instances where mould structures are not robustly present on the surface of the artefact despite being present below the surface. That shows a consistent decrease in RLU values from BT to AT RLU values suggests the latter scenario. Ultimately, comparing BT and AT values relative to VC values can confirm that a reduction in surface mould residues has been carried out during remediation.
Conclusion
Using ATP monitors can provide supporting evidence for conservators to make more informed decisions about how effective cleaning processes are for a particular artefact substrate. Testing may be particularly useful in instances where there is minimal visual change after cleaning and so confirmation of successful remediation is desired. It can also be useful for the periodic assessment of treatment efficacy to maintain quality control, safeguarding against drift in repetitive work.
Challenges of using the devices in cultural heritage contexts relate to accuracy issues, particularly given how the devices are adapted for use. Taking swab samples from large areas is likely to damage artefacts (an undesirable result), particularly for fragile surfaces. In addition to posing a risk to artefacts, the damage caused by surface breakthroughs on fragile artefacts is also likely to affect RLU results in a manner that makes interpretation challenging. Conversely, taking smaller sample areas equates to recording lower RLU values, which have a higher likelihood of instrumental error.
In practice, mitigating the repeatability and accuracy issues of rapid adenosine bioluminescent testing would help produce usable results. Establishing a background level for each artefact by taking multiple samples from both visually clean and visually mould damaged locations provides the user with a sense of what results may be expected. Taking multiple samples will minimise misinterpretations due to instrumental limits of detection. Awareness is needed that fragile sample surfaces can affect results by inadvertently sampling below the substrate surface.Footnote44
ATP monitoring cannot be considered a reliable technique for determining whether mould is present on an artefact as it is not a mould-specific device. It also cannot be used on a pass/caution/fail basis to quantify if a remediation treatment is required or ‘complete’, as there are no standard values below which mould levels can be considered safe.
Overall, ATP luminometer monitoring can be thought of as an indication of risk due to the amount of organic load that can be swabbed from a particular surface. Knowledge of the limitations of luminometers, having carefully designed protocols for sampling and having an understanding of the goals of mould remediation can make the swab testers a useful addition to a cultural heritage toolkit.
Future work
Reviewing current health standards and the tools and techniques used in mould remediation resulted in the authors thinking about how conservators can determine ‘how clean is clean enough?’ when there is no absolute value for what is clean. The authors are developing three conceptual tools to try and address this question. ‘Thresholds of Cleanliness’ is a conceptual framework to guide thinking in how far to take a mould remediation treatment. Based on other risk assessment charts, the ‘Mould & Health Risk Diagram’ maps scenarios for safe access to mould damaged or mould remediated artefacts. Finally, a ‘Mould Remediation Decision Tree’ is a decision-making flowchart that helps sequence the most pertinent steps in the remediation process. Provisional versions of the tools are currently released under a Creative Commons licence,Footnote45 and will be presented in a future paper.
Finally, it would be interesting to run an image-based ATP study to compare the efficacy of other mould remediation techniques other than sponge cleaning such as HEPA vacuum and brush cleaning.
EngMoore_Maitland-Footnotes_FINAL.docx
Download MS Word (28.2 KB)EngMoore_Maitland-FigureCaptions-rev.docx
Download MS Word (15.4 KB)Acknowledgements
We would like to express gratitude to our colleagues for conversations and guidance throughout the research. Thank you to Kamila Bladek (CCI) for the photomicrograph and SEM images recorded in the paper; Tom Strang (CCI) for much early guidance; Anne-Stephanie Etienne (CCI) for collaborating with early samples of mould on wood; Kate Helwig (CCI); Christine McNair (CCI) for support during the project; Amanda Gould and Emily Lin (Canadian Museum of History) for material samples and the opportunity to test other substrates; Rosaleen Hill, Emy Kim and the students of the Masters of Art Conservation programme (Queen’s University) for the opportunity to test other substrates; other ATP luminometer users and members of the informal ATP tester user group who were willing to share knowledge and experience with the devices.
Additional information
Notes on contributors
Tiffany Eng Moore
Tiffany Eng Moore is a book and archival conservator who runs a private company, TEM Conservation in Ottawa, Canada. Tiffany has worked on projects for institutions such as the Supreme Court of Canada Library, Ottawa City Archives and private clients in the national capital region (2018–present). She is a founding member of the Canadian Conservation Consortium, also based out of Ottawa. She provides workshops and training for conservation and book arts, and has taught at the Algonquin College for the Applied Museum Studies Program. Tiffany graduated from West Dean College, UK with a Masters in Conservation of Books and Library Materials in 2015. She completed a post-graduate fellowship at the Canadian Conservation Institute, and has worked at York Explore Library and Archives in North Yorkshire, UK and the House of Lords Parliamentary Archives in London.
Crystal Maitland
Crystal Maitland is the Senior Conservator for Works of Art on Paper at the Canadian Conservation Institute. Her ongoing research areas include stain reduction techniques, implications of and innovations for aqueous treatments for paper objects, treatment possibilities for iron- and copper-containing media on paper, and mould remediation for porous collection materials. Previously, she was the Paper Conservator for the Sheridan Libraries and Museums of Johns Hopkins University. She completed fellowships and internships at Johns Hopkins University, the Queen’s University Archives, the Museum of Anthropology at UBC, the Vancouver Art Gallery, the City of Vancouver Archives and Library and Archives Canada. Crystal graduated from Queen’s University with a Masters in Art Conservation, focussing on paper conservation, and an Honours Bachelors Degree in Chemistry.
Notes
1 In this paper we follow the terminology convention used in CCI Technical Bulletin 26, Endnote 1: ‘Fungi, members of the kingdom Fungi, are organisms that feed on organic matter. They include moulds, yeast, mushrooms and toadstools. Fungal growth is commonly called mould. In this Technical Bulletin, the terms fungi and mould are used interchangeably’. Sherry Guild, Maureen MacDonald, and Tom Strang, ‘Technical Bulletin No. 26—Mould Prevention and Collection Recovery: Guidelines for Heritage Collections’, Canadian Conservation Institute Technical Bulletins, revised edition (2019).
2 Jonathan H. Sogin et al., ‘Implementation of ATP and Microbial Indicator Testing for Hygiene Monitoring in a Tofu Production Facility Improves Product Quality and Hygienic Conditions of Food Contact Surfaces: A Case Study’, Applied and Environmental Microbiology 87, no. 5 (2021); 3M Food Safety News, ‘How Clean Is It? What ATP Can Tell You about the Safety of Your Food Handling Areas’, 2020, https://food-safety-news.3m.com/fsn/how-clean-is-it-what-atp-can-tell-you-about-the-safety-of-your-food-handling-areas/ (accessed 10 December 2023); C.A. Davidson et al., ‘Evaluation of Two Methods for Monitoring Surface Cleanliness—ATP Bioluminescence and Traditional Hygiene Swabbing’, Luminescence 14, no. 1 (1999): 33–8.
3 Nicola Nante et al., ‘Effectiveness of ATP Bioluminescence to Assess Hospital Cleaning: A Review’, Journal of Preventive Medicine and Hygiene 52 (2017): 177–83.
4 Katherine E. Greenstein and Eric C. Wert, ‘Using Rapid Quantification of Adenosine Triphosphate (ATP) as an Indicator for Early Detection and Treatment of Cyanobacterial Blooms’, Water Research 154 (2019): 171–9.
5 A search of ATP monitoring devices includes brands such as BioControl, Charm Science, Hygiena, Neogen, 3M and Kikkoman among others. Overviews or user manuals are often available through the manufacturer websites.
6 Hygiena LLC, ‘Hygiena Comparison to Neogen Luminometers’, Commercial, (n/d), https://www.hygiena.com/food-safety-solutions/atp-monitoring/hygiena-comparison-to-neogen-luminometers/ (accessed 10 December 2023).
7 Hygiena LLC, ‘Hygiena Comparison to Neogen Luminometers’.
8 See, for example, Brian Kupski, Erdogan Ceylan, and Cynthia Steward, ‘Performance Evaluation of Various ATP Detecting Units’, Food Science Center Report, Silliker Group Corporation, 2010, https://www.keydiagnostics.com.au/images/PDF/new/systemSURE%20Plus/Rep_Hygiena_ATP%20Instrument%20Evaluation_RPN13922_rev012110.pdf (accessed 10 December 2023); Navid Omidbakhsh, Faraz Ahmadpour, and Nicole Kenny, ‘How Reliable Are ATP Bioluminescence Meters in Assessing Decontamination of Environmental Surfaces in Healthcare Settings?’, PLoS ONE 9, no. 6 (2014): e99951; Carmen Sciortino and R. Allen Giles, ‘Validation and Comparison of Three Adenosine Triphosphate Luminometers for Monitoring Hospital Surface Sanitization: A Rosetta Stone for Adenosine Triphosphate Testing’, American Journal of Infection Control 40, no. 8 (2012): 233–9; Greg S. Whiteley et al., ‘The Perennial Problem of Variability in Adenosine Triphosphate (ATP) Tests for Hygiene Monitoring Within Healthcare Settings’, Infection Control & Hospital Epidemiology 36, no. 6 (2015): 658–63.
9 Whiteley et al., ‘The Perennial Problem of Variability in Adenosine Triphosphate (ATP) Tests’; Kupski, Ceylan, and Steward, ‘Performance Evaluation of Various ATP Detecting Unit’.
10 Whiteley et al., ‘The Perennial Problem of Variability in Adenosine Triphosphate (ATP) Tests’, 660; Kupski, Ceylan, and Steward, ‘Performance Evaluation of Various ATP Detecting Unit’, 10.
11 Whiteley et al., ‘The Perennial Problem of Variability in Adenosine Triphosphate (ATP) Tests’, 660, para. 6.
12 In the experiments, all devices tested by Whiteley et al. showed Coefficient of Variances (CV) higher than 35%, while experiments by Kupski et al. showed CV values varying from 9% to 123%, depending on the device and the ATP sources measured, and that these values increased near the detection limits. See Whiteley et al., ‘The Perennial Problem of Variability in Adenosine Triphosphate (ATP) Tests’; Kupski, Ceylan, and Steward, ‘Performance Evaluation of Various ATP Detecting Units’, 10, para. 5.
13 Greg S. Whiteley, Trevor O. Glasbey, and Paul P. Fahey, ‘A Suggested Sampling Algorithm for Use with ATP Testing in Cleanliness and Measurement’, Infection, Disease & Health 21, no. 4 (2016): 171, para. 1.
14 Rika Kigawa and Yoshinori Sato, ‘文化財展示収蔵施設等のＡＴＰ調査における留意点の検討’[‘Study of Considerations Regarding the Application of the ATP Tests at Cultural Property Display and Conservation Facilities, Etc.’], unofficial translation by the Translation Bureau, Government of Canada (Tokyo: Tokyo Research Institute for Cultural Properties, 2016), section 4.1.
15 In Whiteley, Glasbey, and Fahey, ‘A Suggested Sampling Algorithm’ the authors propose a robust four-step sampling algorithm for use in healthcare environments. The paper reiterates a previous ‘Letter to the Editor’ in the same journal by Whiteley and other colleagues which also focusses on ATP testing in duplicate or triplicate: Greg S. Whiteley, Chris Derry, and Trevor Glasbey, ‘Sampling Plans for Use of Rapid Adenosine Triphosphate (ATP) Monitoring Must Overcome Variability or Suffer Statistical Invalidity’, Infection Control & Hospital Epidemiology 36, no. 2 (2015): 236–7.
16 Gilbert Shama and Danish J. Malik, ‘The Uses and Abuses of Rapid Bioluminescence-Based ATP Assays’, International Journal of Hygiene and Environmental Health 216, no. 2 (2012): 115–2; Ralph E. Moon, ‘ATP Testing: Use and Misuse in the Restoration Industry’, The Journal of Cleaning Science 3, no. 1 (2021): 10–8.
17 Timothy F. Landers, Armando Hoet, and Thomas E. Wittum, ‘Swab Type, Moistening, and Preenrichment for Staphylococcus Aureus on Environmental Surfaces’, Journal of Clinical Microbiology 48, no. 6 (2016): 2235–6.
18 Kikkoman Corp., ‘ATP+AMP Surface Hygiene Monitoring: Application Manual for Food and Beverage Industries’, 2014, https://lagotec.de/public/PD30_EN_short.pdf (accessed 10 December 2023).
19 See Tomoko Shimoda et al., ‘ATP Bioluminescence Values are Significantly Different Depending upon Material Surface Properties of the Sampling Location in Hospitals’, BMC Research Notes 8 (2015): 807; Shama and Malik, ‘The Uses and Abuses of Rapid Bioluminescence-Based ATP Assays’, 121, for further discussion on the effect of porous textured surfaces in ATP monitoring.
20 Whiteley, Glasbey, and Fahey, ‘A Suggested Sampling Algorithm’.
21 Cf. for example, Christina Meier and Karin Petersen, Schimmelpilze auf Papier: ein Handbuch für Restauratoren; biologische Grundlagen, Erkennung, Behandlung und Prävention (Tönning Lübeck Marburg: Der Andere Verl, 2006), Appendix 1: Einführung in die ATP/AMP-Messung—Eine Messmethode zur Erkennung der Sauberkeit—Beprobung an einem Textilmusterbuch [Introduction to ATP/AMP Measurement—A Measurement Method for Detecting Cleanliness—Sampling on a Textile Pattern Book]; Kigawa and Sato, ‘Study of Considerations Regarding the Application of the ATP Tests’.
22 Michelle Alfa, Nancy Olson, and Brenda-Lee Murray, ‘Adenosine Tri-Phosphate (ATP)-Based Cleaning Monitoring in Health Care: How Rapidly Does Environmental ATP Deteriorate?’, Journal of Hospital Infection 59, no. 65 (2015): 59–65.
23 Cf. Meier and Petersen, Schimmelpilze auf Papier (2006), Appendix 2: Biolumineszenz—Biomonitoring bei Schimmelpilzbefall und Dekontamination [Bioluminescence—Biomonitoring for Mold Infestation and Decontamination for Discussion on the Use of ATP–AMP Monitoring Devices].
24 See, for example, Mikio Bakke and Shigeya Suzuki, ‘Development of Novel Hygiene Monitoring System Based on the Detection of Total Adenylate (ATP+ADP+AMP)’, Journal of Food Protection 81, no. 5 (2018): 729–37; Hygiena LLC, ‘Measuring More, Meaning Less: Kikkoman A3 System Doesn’t Pass’, Hygiena, n/d, https://cdn.brandfolder.io/KA71VJV5/at/fgstmfct5gx45p2k5vm2hr4/EnSURE-Touch-vs-Kikkoman.pdf (accessed 10 December 2023).
25 Kigawa and Sato, ‘Study of Considerations Regarding the Application of the ATP Tests’, Table 2.2.1.
26 Cf. Guild, MacDonald, and Strang, ‘Technical Bulletin No. 26’, section 1.1, 2.2, 2.4.
27 Cf. Brandon Burton, Kevin Fisher, and Bob Lintzenich, Water Damage Restoration Coursebook (Burlington, WA: Legend Brands, 2010).
28 See, for example, Chrystal Palaty and Mona Shum, Mould Remediation Recommendations—Revised (Vancouver, BC: National Collaborating Centre for Environmental Health (NCCEH), 2014); National Institute of Building Sciences, ‘UFGS 02 85 00.00 20 Mold Remediation | WBDG—Whole Building Design Guide’ (Washington, DC: National Institute of Building Sciences, 2018); Hollace S. Bailey, Fungal Contamination: A Manual For Investigation, Remediation And Control (Jupiter, FL: Building Environment Consultants, Inc. (BECi), 2005); Robert C. Brandys and Gail M. Brandys, Post-Remediation Testing and Verification for Mold and Bacteria, 5th edn (Las Vegas: Oehcs Publications Inc., 2014).
29 See, for example, Donald M. Weekes, Philip R. Morey, and Pierre Auger, ‘Update of Canadian and International Mold Guidelines and Standards’, in The 6th International Scientific Conference on Bioaerosols, Fungi, Bacteria, Mycotoxins in Indoor and Outdoor Environments and Human Health, ed. Eckardt Johanning (Bioaerosols, Saratoga Springs, NY: Fungal Research Group Foundation, Inc, 2011), 347–58; Palaty and Shum, Mould Remediation Recommendations; National Institute of Building Sciences, ‘UFGS 02 85 00.00 20 Mold Remediation’; Health Canada, ‘Guide to Addressing Moisture and Mould in Your Home’, 2020, https://www.canada.ca/en/health-canada/services/publications/healthy-living/addressing-moisture-mould-your-home.html (accessed 10 December 2023); Occupational Safety and Health Administration (OSHA), ‘A Brief Guide to Mold in the Workplace’ (Washington, DC: US Department of Labor, 2013), https://www.osha.gov/publications/shib101003 (accessed 10 December 2023).
30 See, for example, Health Canada, ‘Mould’, 2012, https://www.canada.ca/en/health-canada/services/air-quality/indoor-air-contaminants/reduce-humidity-moisture-mould.html (accessed 21 March 2023); WHO Guidelines for Indoor Air Quality: Dampness and Mould, ed. Elisabeth Heseltine and Jerome Rosen (Copenhagen: World Health Organization, 2009); Guild, MacDonald, and Strang, ‘Technical Bulletin No. 26’; ‘Mould Guidelines for the Canadian Construction Industry’ (Ottawa: Canadian Construction Association, 2018), https://www.cca-acc.com/wp-content/uploads/2019/02/Mould-guidelines2018.pdf (accessed 10 December 2023); United States Environmental Protection Agency, ‘A Brief Guide to Mold, Moisture, and Your Home’, 2012, https://www.epa.gov/mold/brief-guide-mold-moisture-and-your-home (accessed 10 December 2023); US Department of Health & Human Services—Centers for Disease Control and Prevention, ‘Basic Facts about Mold and Dampness’, 2022, https://www.cdc.gov/mold/faqs.htm (accessed 10 December 2023).
31 WHO Guidelines for Indoor Air Quality; Hugo Paiva de Carvalho et al., ‘Fungal Contamination of Paintings and Wooden Sculptures inside the Storage Room of a Museum: Are Current Norms and Reference Values Adequate?’, Journal of Cultural Heritage 34 (2018): 268–76; Yasemin D. Aktas et al., ‘Indoor Mould Testing and Benchmarking: A Public Report’ (UK: UK Centre for Moisture in Buildings, 2018); Bradley Prezant, Donald M. Weekes, and David J. Miller, eds, Recognition, Evaluation & Control of Indoor Mold (Falls Church, VA: American Industrial Hygiene Association, 2008).
32 Guild, MacDonald, and Strang, ‘Technical Bulletin No. 26’, Section 2.5.
33 This device was selected as being a type of luminometer in use in cultural heritage contexts. See for instances of international use. A users group has been piloted in Canada with the Canadian Conservation Institute, and three other institutions, the Canadian Centre for Architecture, Ingenium (Canada Agriculture and Food Museum, Canada Aviation and Space Museum and the Canada Science and Technology Museum), and the Canadian Museum of History.
34 NB. Kikkoman currently uses LuciPac Pens to A3 swabs sensitive to ATP, ADP and AMP, which our Canadian supplier has yet to stock as an option.
35 Cf. Kikkoman Corp., ‘Kikkoman ATP Test—Precautions for Use’, Kikkoman Biochemifa Company Test Kit, n/d, https://biochemifa.kikkoman.com/e/kit/method/atp-test-20/ (accessed 12 December 2023).
36 Cf. Whiteley, Glasbey, and Fahey, ‘A Suggested Sampling Algorithm’; Kikkoman Corp. ‘Kikkoman ATP Test—Precautions for Use’; please also refer to other heritage users listed in .
37 As rolling is gentler than dragging, a modified rolling technique using a similar pattern to that outlined in is suggested for more delicate surface substrates that cannot tolerate the friction of a dragged cotton swab. Care is needed during rolling to keep the swab from skittering across surfaces, as the irregular shape of the swab handle makes it surprisingly challenging to rotate consistently while also traveling across the sampling area in a repeatable manner—hence the preference for dragging the swab when the sample surface is sufficiently robust. It is also easier to maintain a consistent pressure while dragging than while rolling. For consistency, all samples recorded in this article were collected by dragging; in sampling actual artifacts a rolling sampling method might have avoided sample breakthrough on the more fragile substrate.
38 A discussion of swabbing techniques using the Kikkoman system took place between the authors and Yashinori Sato of the Tokyo Research Institute for Cultural Properties in September 2018; Christina Meier Wolff, ‘How Clean Is My Object? The Bioluminescence Measurement (ATP/AMP)—Method, Practice and Opportunities’, (12–16 October 2015—XIIIth IADA Congress Berlin), 83, https://iada-home.org/wp-content/uploads/2021/06/Berlin_2015_Abstracts_Thursday.pdf (accessed 20 December 2023).
39 It is noted by Malalanirina Rakotonirainy et al. that commercial reagent kits are often developed for bacteria and therefore are not optimised for the use on fungal cells. Malalanirina Rakotonirainy et al., ‘Detection of Fungi and Control of Disinfection by ATP-Bioluminescence Assay’, AICCM Bulletin 28, no. 1 (2003): 16–22. Meier and Peterson—who used the same Kikkoman device as in this study—noted that it is possible to shake the swab in the water chamber first for 30s, followed by breaking the seal to the reagent chamber and shaking for another 30s in order to better penetrate the fungal cell walls and react with the enzyme. Meier and Peterson, Schimmelpilze auf Papier. Upon consultation with the supplier of our ATP monitoring device, we used their suggested technique which is to break both seals but to ensure shaking occurred for 30s in total (email conversation with Ashlee Donaher, Technical Sales Representative, LuminUltra Technologies Ltd, 2018). Table 1 documents other reagent mixing methodologies used in cultural heritage practice.
40 E.E. Williams, Investment Analysis (New Jersey: Prentice-Hall Inc., 1974).
41 A 1924 edition of The Diary of Samuel Pepys released by Harcourt, Brace and Co in New York.
42 See, for example, Valentina Alto, ‘Introduction to Box Plots and How to Interpret Them’, Medium blog, 2021, https://medium.com/analytics-vidhya/introduction-to-box-plots-and-how-to-interpret-them-22464acbcba7 (accessed 10 December 2023).
43 See examples of structural morphologies of mould viewed using SEM in: Flavia Pinzari and Beata Gutarowska, ‘Extreme Colonizers and Rapid Profiteers: The Challenging World of Microorganisms That Attack Paper and Parchment’, in Microorganisms in the Deterioration and Preservation of Cultural Heritage, ed. Edith Joseph (Cham: Springer International Publishing, 2021), 79–113; Nicholas Nastasi et al., ‘Morphology and Quantification of Fungal Growth in Residential Dust and Carpets’, Building and Environment 174 (2020): 106774.
44 Citing some of the same studies as reviewed in this article, Ralph Moon in Moon, ‘ATP Testing’, came to similar conclusions on the limitations of the device. Furthermore, in the context of residential usage Moon does not recommend the use of ATP monitors to ‘clear’ homes as part of any post-remediation verification process. This is of interest as residential restoration applies similar criteria as conservation remediation: organic material substrates (porous, often cellulosic substrates) and ensuring human safety after biological infestation including mould remediation. He also does not recommend use of the devices for most porous surfaces which diverges from the recommendations in this article as the authors consider ATP monitoring devices as a useful supporting tool when it is paired with an appropriate methodology.
45 The tools are available for use, comment and collaboration: http://canadianconservationconsortium.ca/en/assess-mould-levels/ (accessed 10 December 2023).
Appendix 1:
RLU data
During swabbing samples were assigned a unique identifying number of 03–52 (samples 01 and 02 were discarded as the sampling methodology shifted after they had already been recorded). Each sample was linked with a qualifier based on the visual quantity of mould sampled from the surface (high=H, low=L, very low=VL, visually clean=VC, surface breakthrough=SB). For the purposes of data visualisation in this paper, sample groups have been sorted from lowest to highest BT RLU value, and re-assigned a sequential number in the following format: the paper type (intact=I; fragile=f), the visual quantity of mould and finally the data sort sample number. For SB samples, an * marks whether the sampling breakthrough occurred on the BT, the AT or on both samples.
Table A1 Sample-by-sample RLU data values.