1. Introduction
The protection of cultural heritage is one important priority for our society. It is a significant resource for economic growth in terms of both services offered via tourism and cultural exchange. In Italy alone, a report by the Italian National Statistical Institute [
1] indicates that 4265 museums and similar institutions, public and private, were open or partially open in 2020, comprising 3337 museums, 295 archaeological areas, and 633 monuments or monumental complexes. There are also many structures that are not open to the public; these include cultural heritages that require appropriate protection [
1]. Air pollution, is a main hazard for human health, the local and global environment, and materials, also poses a risk for the conservation of cultural heritage. Indeed, air pollution affects cultural heritage materials exhibited both outdoors and indoors, causing irreversible damages to the art-works [
2].
The major polluting species that may be found inside museums and pose a risk to cultural materials are sulphur dioxide, nitrogen dioxide, nitrogen oxide, ozone, reduced sulphur gases such as hydrogen sulphide, and particulate matter [
3]. To effectively protect art-works preserved in museums, it is therefore appropriate to develop monitoring activities with the aim of assessing the nature of contaminants and their concentration levels. This information, coupled with the evaluation of physical parameters such as temperature and relative humidity, can guide researchers toward more effectively preventing potential damage to art-works caused by exposure to these elements. Although no standard has been issued for the most common air pollutants of interest for cultural heritage conservation, some limits have been reported [
2]. They range from practically zero to a few parts per billion. Therefore, assessing pollution, especially in indoor environments (museums, etc.) requires techniques that are sufficiently sensitive and reliable.
To fulfil this requirement, many measurement campaigns have been conducted in museums using analytical systems with different complexities [
4,
5,
6,
7,
8]. However, these monitoring approaches have two limitations. First, techniques can be very expensive and require highly specialized personnel, leading to excessive expenses; in addition, they cannot be applied simultaneously in several rooms inside museums where air contaminants may show different concentrations. Second, most of these activities are usually conducted over a relatively short period of time. This does not allow data to be acquired over a wide time horizon, which may be relevant for fully understanding the evolution of air pollution [
9,
10]. Although the most important museums can afford and perform sophisticated campaigns for the characterization of atmospheric pollutants, the number of sites that require protection and monitoring is very high. Therefore, the complexity and the cost of monitoring excludes many important cultural sites from a proper evaluation of risks posed by air pollution.
A direct solicitation to develop a monitoring approach that is very simple, low-cost, and easily adaptable methodology that can be used simultaneously in several sites is discussed here. This approach offers significant opportunity for widespread applications as it is compatible with reduced budgets and can provide important and preliminary information for possible and desirable in-depth monitoring and control of atmospheric contaminants. The approach pursued in our strategy is very similar to that established by the EU Directive on ambient air quality 2024/2881 (European Union, 2024) [
11], a revision of the former EU Directive 2008/50/CE, Art. 9, comma 3a, which allows indicative measurements or modelling applications to provide sufficient information for the assessment of air quality with regard to limit values, target values, critical levels, alert thresholds and information thresholds, as well as adequate information for the public, in addition to the information provided by the sampling points for fixed measurements. Indicative measurements provide data quality objectives that are less strict than those required for fixed measurements, which can then be used for preliminary assessment. This implies the use of less sophisticated systems than those required by the Directive for the areas in which monitoring with certified and an approved system is mandatory [
12]. The preliminary assessment then provides data on which the control bodies may plan the final monitoring network in compliance with the Directive.
According to this approach, a simple system based on the use of passive samplers with the aim of monitoring the most important species that cause damage to cultural heritage was adopted. The system is based on the use of Analyst type passive samplers that have been developed for the measurement of acid gases [
13] and nitrogen dioxide [
14].
The use of passive sampler Analyst for monitoring the MANN (the National Archaeological Museum of Naples) museum is very effective, as it can be exposed for a long period due to its large capacity. This allows data to be averaged over a long period of time (in this case: 1 month), reducing the cost of long-term monitoring. The paper primarily addresses acidic species, which play an important role in the preservation of stone art-works. This is particularly true in museums housed in historical buildings, in which treating air is difficult for technical reasons. This paper contains important information about nitrous acid. As is well known, this acid is generated by the interaction of nitrogen dioxide with the surface, leaving nitric acid. This acid is highly corrosive. The method used for monitoring is designed to limit the cross-interference between the measurements of NO2 and HONO. This is quite evident from the data treatment. This issue is particularly important in indoor pollution. Now, experiments are underway to test this finding in the laboratory. The generation of a pure HONO gas mixture is the next objective needed to demonstrate our findings. Nitric acid is another important monitored species. As with HONO, the interaction of NO2 with the surfaces generates nitric acid that remains on the surface. Our data demonstrate that the source for this species is the dissociation of nitrate aerosols. Also, this result was obtained by data treatment, including those provided by the environmental dataloggers. Passive samplers used through the monitoring campaign consist of quartz and carbon paper active surfaces alkalinized with sodium carbonate and glycerin. The quartz active surface irreversibly adsorbs acids, while carbon paper collects nitrogen dioxide that is converted into nitrite (NO2− anion). Ions on both surfaces are then extracted with bi-distilled water and analyzed by Ion Chromatography (IC).
This configuration, integrated with an environmental datalogger for the measurement of temperature and relative humidity, have been used for a one year of monitoring in five rooms at one of the most important Italian museums: MANN. This demonstration project provided useful data related to the presence of pollutants and also gave important and sometimes unexpected results about the reliability of the suggested technique, especially when intended for the evaluation of nitrogen dioxide and nitrous acid.
The first reason we chose not to measure O
3 is the low levels of O
3 concentration in indoor museum environments. In fact, the literature [
5] shows that the levels of O
3 concentration (0.94 μg/m
3 is average concentration recorded in Capodimonte’s museum, which is another museum located in Naples) are below the legal limit (which is 2 μg/m
3, according to literature [
5] and references cited therein] and come essentially from external pollution. Internal sources of O
3 could be attributable to the organic material surfaces and wooden components; secondary reactions can occur on these which contribute to the production of O
3 in indoor environments. In the case of MANN museum, there are fewer organic materials and wood surfaces because the sculptures are mostly in marble, so internal sources of O
3 are negligible.
The second reason that O
3 was not considered in this study is its negligible contribution to the damage and risk equation/R, formulated in the literature for stones (especially sandstones) as reported by [
15] and very well described by
Equation (S1) as shown on Supplementary Information Section, File S1. In this equation, the damage function R depends linearly on the concentration of acids (mainly nitric acid/HNO
3 concentration/µg m
−3, by a coefficient equal to 0.078 × [HNO
3] × RH
60, where RH
60 is the measured relative humidity when it is 1 if RH > 60%, otherwise 0); 0.054 Rain [H
+] and on the particulate matter (firstly PM
10), where this latter component exhibits the smallest coefficient (as 0.0258·[PM
10]) between that of nitric acid and acid rain, respectively. This equation model/R [
15] does not consider the damaging effect caused by Ozone at all, showing that it is negligible (compared to the main effects provoked by acids) for art-works in museums. Furthermore, the contribution of O
3 to the damage of metal surfaces is described by another equation, as reported in the literature [
16]. However, even here, the coefficient of proportionality of O
3 is significantly lower (0.20 × [O
3]) than that of hydrogen sulphide (45.30 × [H
2S]); sulphur dioxide (3.90 × [SO
2]); nitrogen dioxide (1.46 × [NO
2]); hydrochloric acid (4.81 × [HCl]); relative humidity (1.04 × RH); and temperature (0.79 T), towards metallic surfaces (as Ag, Cu, Zn, etc.; see Ref [
16], and also
the Equation (S2) and (S3) on Supplementary Information Section, File S1).
Also, several other pollutants, such as VOCs (Volatile Organic Compounds) are not described/included into R equation for stones [
15] for the metallic damage equation [
16] or the soiling equation [
17] for PM10. For this purpose, and according to these literature (concerning the negligible damage effects of VOCs on art-works), the VOC pollutants have been not measured during this work.
Always considering this R damage equation [
15], it is clear that the contribution of particulate matter (with respect to Ozone) is present but always at lower levels if compared with those reproduced for acids and acid rain (which have the greatest impact on cultural heritage). Therefore, the particulate matter has also not been measured in this study, but future studies may foresee monitoring campaigns that also include sampling of the particulate matter and the estimation in the R function of its contribution according to the proportionality coefficient equal to 0.0258·[PM
10], which is lower than that of inorganic acids and acid rains [
15].
Furthermore, it is worth adding that particulate matter contributes to the damage of historical surfaces more with the “Soiling effect” than with the “Etching effect”; this latter mainly related to the acids, described above by the R equation [
15]. For the soiling effect, there is a “Square Root Law”, which explains this darkening effect, as cited in the literature [
17], and also highlighted by
the Equation (S4) on Supplementary Information Section, File S1. The square root function is mathematically slower as a law that correlates certain quantities, including the concentration of pollutants, compared to a linear function (as with the R equation for etching effects provoked by the gaseous acid pollutants, reported previously), and this implies that the action of a “soiling” particulate matter will be secondary to a corrosion effect (mainly provoked by HNO
3 and acid water/rain). This means that corrosion/etching phenomena is far superior to the typical soiling effect of the particulate matter for marble sculptures and cultural heritage.
It should be added that the contribution of PM
10 to the damage equations towards metal surfaces is negligible, as shown by the equation in the literature [
16]. Data of particulate matter would be interesting if they included the chemical speciation.
Finally, the reasons above and those supported by the literature [
15,
16,
17] consistently justify the choice adopted in this work not to measure O
3, VOCs, and PM
10. We also did not consider the data concerning relative humidity (RH) because it is negligible according to the R equation damages (mainly due to the acid etching effects [
15]), metal surface equation damages [
16] and soiling events (for PM
10 solid matter component) that were reported above).
We believe that the results presented in this paper and other information can be used to predict the evolution of pollutants in the museum, especially according to the previous damage equations (cited above [
15,
16,
17]) and briefly reported on in the
Supplementary Information Section, File S1. This is certainly true over a long-term time basis (monthly). Briefly, we can easily model the daily evolution of nitrogen dioxide according to synoptic meteorological conditions (as reported later in the full text). By knowing the air exchange rate, we may apply Equation (2) (in the full text) to estimate the indoor concentration. Thus, it is possible to evaluate damage through the deposition velocity of acidic species (which are responsible for most of the etching and oxidation events on stones and metallic surfaces, respectively). We preliminarily estimated the damage. However, these data require further evaluation and will be published as soon as they are complete.
2. Material and Methods
2.1. Monitoring Sites and Monitoring Program
The National Archaeological Museum of Naples (MANN) is an Italian museum extending over an exhibition area of 12,650 m
2, and it is considered to be one of the most important archaeological museums in the world. The main exhibitions include Roman sculptures from the Farnese collection, a Pompeian collection that includes many finds from the area of Pompeii and other locations near the Vesuvius volcano [
18], and an important Egyptian collection (
https://www.electa.it/en/product/guide-to-the-egyptian-collection-in-the-mann/, accessed on 8 October 2016). The museum is located in the central area of Naples in a historical building with nearby streets characterized by very heavy traffic, causing significant atmospheric pollution episodes. A monitoring station near the museum consistently measures high levels of pollutants, especially nitrogen dioxide (
https://www.arpacampania.it/bollettini, accessed on 16 April 2025). The museum is extended over three floors and a basement.
Figure 1 shows the museum site maps of the ground and underground floors, as well as the monitoring site locations.
In this demonstration project, the locations were mostly placed on the ground floor as that level is the most exposed to air pollutants. It includes showrooms and deposits that comprise statues and marbles which are very sensitive to air pollutants of acidic nature [
19,
20]. Site E is still located on the ground floor but in a room (restoration laboratories) near the central garden, which is directly exposed to external air during working periods. Site A is underground, where many stone and marble art-works that are not yet being exhibited to the public are stored.
Measurements were carried out in these sites for 9 periods, shown in
Table 1.
As mentioned before, the average exposure of the samplers for the first eight periods was approximately one month. The ninth period had duration of approximately four months. The last period was extended for a longer time since it is known that the passive samplers used are characterized by a high capacity, and therefore, the results were used to evaluate their suitability for a longer period of time. Auto-consistency tests have been proved in [
14]. The overall sampling period lasted approximately one year.
2.2. Passive Samplers and Exposition Shield
Analyst
® passive samplers have long been used for the monitoring of air pollutants [
13]. The basic design of this kind of passive sampler is shown in
Figure 2.
It is characterized by extensive laboratory evaluation and has high capacity, allowing sampling over extended periods of time (up to several months). The design also includes inlet diffusive screens made of steel and polyester to reduce air turbulence, which is known to affect the diffusion mass flow. For this monitoring campaign, a three-place exposition shield was used. Two supports were used to accommodate two passive samplers which monitored acids and nitrogen dioxide. The third support accommodates a data logger for temperature/relative humidity monitoring (see
Section 2.6). The shield includes a front steel net that protects the passive samplers from dust and insects and provides a first dumping of air turbulence. This is furtherly reduced by the inlet screens on the passive samplers.
The shields were fixed to the walls inside the rooms where monitoring was performed. They were placed at a height of 3 m in rooms B, C, and D, while in rooms A and E, they were operating at approximately 2 m. The shields are characterized by an insignificant visual impact, which is essential for exposition rooms. An example of the exposition in room C is given in
Figure 3.
2.3. Passive Sampler for Acids
Acids are directly adsorbed onto the active surface of quartz filters (Whatman QM-A) coated with an alkaline solution (1.8% Na
2CO
3 and 1.8% glycerin in water/ethanol 60/40). This solution is also used for the preparation of passive samplers intended for NO
2 (see
Section 2.4). Such an active surface is able to collect inorganic and organic acids with high efficiency and retention. After the sampling step, passive samplers were extracted with water and analyzed by ion-chromatography (IC). The apparent flow rates used for the calculations of concentration from IC data are shown in
Table 2.
These data were used to calculate the gas phase concentration by the concentration of relevant ions extracted from the filter. Since laboratory and field blanks approach zero, the minimum detectable amount of most acids is less than 0.2 ppb for a monthly sampling time. For nitrous and nitric acids, this amount corresponds to less than 0.4 µg/m
3 and 0.6 µg/m
3, respectively. Both the Field and Laboratory control sample values are summarized on
Table 3.
Table 3.
Limit of Detection (L.O.D.) * for field and laboratory sample controls.
Table 3.
Limit of Detection (L.O.D.) * for field and laboratory sample controls.
Pollutants | L.O.D./µg m−3 Laboratory Control | L.O.D./µg m−3 Field Control | Sampling Time |
---|
Nitric Acid HNO3 | 0.22 ± 0.01 | 0.60 ± 0.02 | Monthly |
Hydrogen Chloride HCl | 0.20 ± 0.02 | 0.53 ± 0.01 | Monthly |
Nitrous Acid HNO2 | 0.08 ± 0.005 | 0.40 ± 0.01 | Monthly |
Formic Acid | 0.14 ± 0.01 | 0.50 ± 0.03 | Monthly |
Acetic Acid | 0.12 ± 0.005 | 0.49 ± 0.005 | Monthly |
Table 4.
Analytical parameters quantified in IC analytical method LOD (the Detection Limit) is quantified by measuring and reading 10 blanks (and is reported as the average value); LOQ (limit of quantification) is the first calibration line point.
Table 4.
Analytical parameters quantified in IC analytical method LOD (the Detection Limit) is quantified by measuring and reading 10 blanks (and is reported as the average value); LOQ (limit of quantification) is the first calibration line point.
Ions | LOD (ppm) | LOQ (ppm) | Φ (mL min−1) | LOD (µg/m3) | LOQ (µg/m3) |
---|
Chloride/Cl− (from HCl) | 0.01 | 0.1 | 8.1 | 0.14 | 1.48 |
Nitrite/NO2− (from NO2) | 0.02 | 0.2 | 12.3 | 0.18 | 1.8 |
Nitrite/NO2− (from HNO2) | 0.02 | 0.2 | 9.1 | 0.25 | 2.5 |
Nitrate//NO3− (from HNO3) | 0.02 | 0.5 | 10.5 | 0.22 | 5.5 |
Sulphate/SO42− (from SO2) | 0.045 | 0.5 | 9.9 | 0.52 | 5.2 |
Earlier experimental data [
13] confirmed that the reproducibility of this type of passive sampler is within 10–15%. This good result is also due to the positive effects of the inlet steel screens that smooth out any air turbulence at the inlet.
Data from passive samplers intended for acids also include organic acids. However, formic acid was not measured because it was emitted by the plastic shields. Acetic acid was always found below the minimum detectable concentration.
One important point to be considered concerns the possibility that nitrogen dioxide interferes with the substrate of the active surface intended for the collection of acids, yielding nitrite and nitrate ions according to the following reaction:
This reaction generates nitrite and nitrate ions that are retained on the filter, then interfere with the same ions formed by nitrous and nitric acid, respectively. Such interference can be very high if the concentration of nitrogen dioxide largely exceeds those of the two acids.
Reaction occurs on all surfaces, and its extent depends on many variables, such as the type of surface, pH, and relative humidity. Thus, the extent to which nitrogen dioxide is converted into nitric and nitrous acid is difficult to estimate a priori. However, as will be shown in the
Section 4, this reaction is practically not occurring on the active surfaces of passive samplers; thus, the devices can measure acidic species with sufficient reliability without the interference of nitrogen dioxide. The starting point is that, in a case where the reaction (Equation (1)) is occurring on the filter, the resulting nitrite and nitrate ions ware irreversibly adsorbed on the surface, yielding a 1:1 ratio. Experimental values of this ratio are never close to 1:1.
Quality control was carried out by analyzing field blanks and replicates of passive samplers used in the monitoring campaign. For each monitoring period, one sampling site was supplemented by a pair of replicates and one field blank, i.e., a passive sampler not exposed. The variability found between replicates was on average less than 6–7%
It is worth observing that the monitoring campaign required just 10 passive samplers per period and then 10 analyses per month. This, in turn, positively reflects the low cost of the simultaneous monitoring activity for the selected five rooms.
2.4. Passive Sampler for Nitrogen Dioxide
Most passive samplers intended for monitoring nitrogen dioxide use of triethanolamine (TEA) as an active surface. TEA converts nitrogen dioxide into nitrite ions that can be measured by ion chromatography. Although a number of experimental campaigns have been carried out [
21], data shows an evident poor reliability in terms of accuracy and precision. The reason for deviations may be that TEA is not a perfect sink for nitrogen dioxide, as was previously demonstrated by using a diffusion denuder [
22]. Moreover, TEA collects nitrous acid. This species is present in indoor environments at high concentrations, often higher than those experienced outside and having the same order of magnitude as nitrogen dioxide levels [
23]. Since nitrous acid is also converted into nitrite ion, it may cause a 100% interference. The effect of nitrous acid has been largely overlooked since most applications of nitrogen dioxide passive samplers were carried out in ambient atmosphere, where the presence of nitrous acid was negligible.
The sampler used for collecting nitrogen dioxide in this campaign was based on an active surface prepared by impregnating a 20-mm carbon paper disk (Envint srl, Montopoli di Sabina, Italy) with the solution described in
Section 2.3. This procedure follows the method developed in [
24]. As shown in previous studies, alkaline carbon directly converts nitrogen dioxide into nitrite ions that can be analyzed by IC. The conversion of NO
2 into nitrite ion was also confirmed by Raman spectroscopy [
25]. The apparent flow rate for this type of passive sampler for NO
2 is 12.3 ± 0.7 mL min
−1.
After sampling, the filter disc is removed from the passive sampler, extracted with 5 mL of water for approximately 1 h, and analyzed by ion chromatography (Dionex ICS-1000, Thermo Scientific, Waltham, MA, USA). The procedure is such that, for an exposition of 1 month, the minimum detectable concentration is less than 0.4 µg/m3.
Unfortunately, this kind of passive sampler for nitrogen dioxide was not fully exploited in the field. An interesting report [
21] concluded that it can be used for the monitoring of nitrogen dioxide without further details.
The carbon filter paper treatment ensures that the blanks are practically zero, and zero is the field blank. The stability of nitrites on the filter is good, even in the presence of a huge amount of ozone, which could oxidize nitrites into nitrates, causing negative deviations. However, the alkaline carbon paper also adsorbs many other acidic substances. Nitrous acid (HONO) is very likely to be included, yielding nitrites and potentially interfering 100% with the measurement of nitrogen dioxide.
The problem of HONO interference has been discussed in several papers, especially those in which triethanolamine was used [
26], but it was considered unimportant because the ambient concentrations of HONO are much lower than those of NO
2. Unfortunately, for indoor environments, this is not always the case. Some devices were also developed to reduce this interference. A paper that specifically addressed this problem [
27] described a device based on three-layer active source. This device is not simple to prepare and use, so its use has been limited to a few applications in archives and libraries.
Our measurements in MANN demonstrate that HONO is present at concentrations comparable to, and often larger than that of NO2. As will be shown later, data demonstrate that the interference of HONO with the measurement of NO2 is not significant, and that the passive samplers for nitrogen dioxide are free from interference due to nitrous acid. This is one of the major technical findings for this monitoring campaign.
2.5. Analysis of the Samples
Analysis of the samples was performed by ion chromatography (IC). The IC apparatus is the model ICS 1000 instrument, having a Dionex Ion pack 12° as the chromatographic column. This IC apparatus is equipped with the suppressor aers 500 carbonate 4 mm. The chromatographic separation is carried out isocratically (and not in gradient mode). After exposure, the passive samplers were extracted with 5-mL water and analyzed using an ion chromatograph Dionex ICS-1000. The analytical procedure ensures a limit of detection (LOD) and a limit of quantification (LOQ) in ppm, shown in
Table 4.
Both passive samplers were analyzed using the same analytical procedure. Corrections for blanks were not significant, and the reproducibility of collocated samplers showed a standard deviation of 7%.
It is worth mentioning that, by using the apparent flow rate of pollutants into the passive sampler and for an exposition time of one month (30 days), it is possible to convert LOD and LOQ in ppm into concentrations in µg/m3. From the same table, it is possible to observe that, for most of the species of interest, the minimum detectable concentration was well below 1 µg/m3 while reliable quantifications are above a few µg/m3. The analysis of both passive samplers used in this monitoring campaign used the same procedure. This also reduced the effort required to improve the quality of the data and, clearly, reduces the cost of the monitoring campaign.
2.6. Data Logger
Data loggers used for this campaign are based on the Honeywell HumidIcon™ Digital Humidity/Temperature sensor HIH8120 (Honeywell, Industrial Automation 2425 South 21st Street, Phoenix, AZ 85034, USA).
This sensor ensures an accuracy level of ±2.0%RH (Relative Humidity) and a temperature accuracy level of ±0.5 °C; thus, it is very suitable for indoor measurements. The sensor is within a small plastic box that fits the clips used to hold the passive samplers, and can then be placed into the same exposition shield. The data logger is battery operated and programmed to select the sampling frequency and to set up the internal clock. The sampling frequency can be selected from 1 to 60 min. For the specific application, the sampling frequency was 15 min, along which the dataloggers collect one measurement per minute and then average the data. Thus, the data logger records 4 average temperature and humidity data points every hour. At the end of the sampling period programmed for passive samplers, the data loggers were removed and connected to a personal computer from which an Excel file is extracted. After sampling and reading the output files, the data-loggers can be used repeatedly.
2.7. Outdoor Pollution Data
Monitoring of atmospheric pollution in the city of Naples is continuously carried out by the technical authority in charge, in this case, the Regional Agency for Environmental Protection of Campania ARPAC, (
https://www.arpacampania.it/, accessed on 23 December 2014).
The Agency manages a monitoring network in compliance with European Union Directives. Data are regularly published at (
https://www.arpacampania.it/bollettini, accessed on 16 April 2025). Our interest was in the station “Museo” which is located near the MANN museum. This station provides an hourly average of several pollutants, including nitrogen dioxide. Meteorological data were provided by the station “Osservatorio” in Naples.
4. Discussion
This section is dedicated to a discussion of the results and has been organized into sub-paragraphs, each concerning the individual measured species. Emphasis is placed on nitrogen-containing compounds because they are in relatively high concentrations, causing exposed art-work to rapidly deteriorate.
Dataloggers are essential for monitoring the evolution of temperature and relative humidity in the exposition rooms. However, they are also valuable because they provide data to estimate the evolution of pollutants entering the museum (
Ci) from outside (
Co). The indoor/outdoor concentration ratio can be expressed by the following Equation (2):
where
Air-exchange rate, usually expressed as h−1, i.e., exchanges per hours.
Volume of the room (m3).
Surface area of the room (m2).
Deposition velocity of the pollutant (mass of the pollutant deposed on a unit surface per unit time).
By knowing the deposition velocity of an individual pollutant and the geometrical characteristics of the exposition rooms (
V and
S), it is possible to estimate the concentration ratio (indoor to outdoor) if the air exchange rates
a is known. As is well known, many pollutants show a time trend that can be followed using Equation (2). This is, for instance, the case of ozone that shows a time trend defined by a vertical process occurring in the atmosphere, causing the fumigation of this pollutant from the free troposphere to the ground (see
Figure 7).
In order to estimate
Ci, it is necessary to know the term
a, air exchange rate. This can be achieved using various techniques [
37,
38,
39]. However, most of these experimental methods are expensive, time consuming, and not simple to implement. According to our main goal, which is to obtain preliminary information on the museum atmospheric environment with simple methods that can be reproduced in many locations, we attempted to estimate the air exchange rate by the difference in indoor–outdoor temperatures. In fact, the exposition rooms of MANN are not conditioned. The size of the rooms is such that the visitor effect can be neglected, and the low surface-to-volume ratio is such that the thermal effects of the walls can also be neglected.
In this hypothesis, the indoor air temperature
Ti can be expressed as a function of the outdoor temperature
To, according to Equation (3):
where
dTi is the increase (or decrease) in indoor temperature caused by ventilation of a volume of air
dV.
To and
Ti are the outdoor and indoor temperatures, respectively. Unfortunately, this equation cannot be integrated directly because the variables involved in it are both time dependent. The indoor temperature in an unconditioned room, such as those of the MANN museum, will be just the low-pass-filtered outdoor temperature. Therefore, in a steady state, the indoor temperature will be the same as the outdoor temperature. However, the thermal capacity of the surfaces and the poor mixing of outdoor air are such that strong deviations from this simple model can be expected.
Our simple approach starts by assuming that the time trend of indoor temperature follows the external temperature, and that the maximum value of room temperature can be used in Equation (3). It is therefore possible to estimate, with the said uncertainties, the value of dV, and then the value of a.
Several calculations conducted throughout the monitoring periods, showed that the term dV/V is about 0.02, with little difference between the exposition rooms and monitoring periods. For instance, in room B (dimensions 36 × 18 × 7 m), the volume is about 4500 m3, while the total surface area is 2000 m2. The ratio S/V = 2000/4500 = 0.44 m−1.
For the term deposition velocity, it is possible to determine the average deposition velocity of all surfaces in a room from Equation (3), as well as the so-called ‘
surface removal rate’
Vd (S/V). This rate is directly comparable to the air exchange rate. The deposition velocity of a particular pollutant varies within different material types and under different conditions, such as changing relative humidity [
40]. A study of museum buildings showed a surface removal rate of 0.4 h
−1 for a large and open gallery [
41]. A surface removal rate of 0.4 h
−1 means a deposition velocity of approximately 1 m h
−1.
When adapting these data to Equation (3) and using the value of deposition velocity ranging from 1 to 0.1 h−1, then the estimated Ci/Co ranges from 0.043 to 0.31. The ratio between the NO2 concentrations found in the outdoor and indoor ranges is between 0.5 and 0.3, which fits the expected concentration ratios calculated by our simple model.
4.1. Nitrogen Dioxide (NO2)
The average concentrations of the NO
2 pollutant reported in
Table 4 appear to be relatively constant, except for the data from room A. This means that the exposure to this pollutant in rooms B, C, and E are approximately the same. This could not be consistent with the observation that data from site E should be higher because this room is more exposed to external air. However, it should be considered that the room is mostly affected by external air during the workday; as shown in
Figure 6 and
Figure 7, external air pollution is modulated in such a way that during the workday, the concentration of pollutants (except ozone) in the external air is low.
The concentration of nitrogen dioxide in the external air is well over the standard fixed by the European Union for the protection of public health, at a yearly average of 20 μg/m
3 and at a daily average of 50 μg/m
3 [
11,
42].
The ratio between the concentrations found in the outdoor and indoor range was between 2 and 3.5. This result is consistent with similar data found in different museums and other indoor environments where no internal sources of NO
x were present [
43]. The highest ratios between outdoor and indoor data were found for periods 4 to 7, i.e., in autumn–winter, when most of the atmospheric stability processes are likely to occur and when intrusion from external air is limited by windows and doors closed during this period of the year.
The relatively high levels of air pollutants found inside the museum are associated with the evolution of pollution in the outside environment. As mentioned before, the area around the MANN museum is characterized by intense emissions, especially by vehicles, which cause high concentrations of primary NOx pollutants. These, in turn, when irradiated by sunlight, generate many photochemical pollutants, leading to high concentrations of nitrogen dioxide (as seen in
Figure 5).
In outdoor air, Nitrogen Dioxide concentrations start to increase at approximately 6–7 AM, reaching a level of more than 100 µg/m3. In the late morning through the early afternoon, the concentration drops to low levels, increasing once again in the late afternoon, when it reaches values of more than 120 µg/m3. This behavior is very similar to that observed in many locations and can be easily explained by considering the time development of the boundary layer. Overnight, the layer is quite stable, and in the early morning, primary pollution increases because of intense traffic. Solar irradiation in the morning provides heating of the ground, resulting in turbulence that causes ground air to mix with air in the free troposphere. At this time, ozone concentrations increase, whereas nitrogen dioxide concentration decreases because of mixing with relatively clean air advected from the free troposphere. In the late afternoon, solar radiation decreases, and a ground-based mixed layer is again established.
In this residual layer, the reactions of radicals are very active, oxidizing nitrogen oxides into nitrogen dioxide. Overnight, these reactions are quenched, and the decrease in emission decreases the nitrogen dioxide concentration to low values. In conclusion, two peaks were experienced: one in the morning and the other in the afternoon, with a minimum in the early afternoon. In most cases, the daily maxima occur in the evening, whereas in some cases, they are recorded in the morning. Such general behavior is reproduced throughout the year, although in summer, due to the more intense solar radiation, the window for clean air advected from the free troposphere is much wider and can be as large as 7 h, instead of 3–4 h in winter.
4.2. Sulphur Dioxide (SO2)
Sulphur dioxide has long been considered an important pollutant in museums.
Figure 9 shoes that the concentration levels are in the range 1–3 µg/m
3 and do not show a definite trend. The average concentration at site A, similar to NO
2, was less than the average of the other four sampling sites. This is a clear indication that sulphur dioxide is coming from the air outside the museum. Sulphur dioxide is fairly correlated with nitrogen dioxide (R
squared = 0.77), demonstrating that both of them are coming from intrusion by external air.
Sulphur dioxide levels measured by the public Naples monitoring network are approaching values near the minimum detectable concentration. Therefore, no direct comparison between the two sets of data is possible.
4.3. Hydrogen Chloride (HCl)
Acids measured using the suggested technique include hydrogen chloride and nitric acid. The latter is reported in
Figure 10, which shows the average values found in the data from locations B through D. Data from site A do not show a definite difference from those from other sites. For HCl, (
Figure 11) the observed concentrations are about 1 µg/m
3. This amount is very close to the minimum detectable concentration and is derived from chloride ions extracted from the quartz alkaline passive sampler. As is well known, contamination may play an important role in the analysis of chlorides; thus, the reported data are not of sufficient reliability for further discussion. However, the data show that HCl does not reach significantly high values; therefore, it does not imply an excessive risk.
4.4. Nitric Acid (HNO3)
Nitric acid shows relatively high concentrations of up to 6 µg/m
3. This species is generated by the reaction of nitrogen dioxide with OH
• radicals (according to Equation (4)):
Under ambient conditions characterized by intense photochemical activity, the amount of both NO
2 and OH
• radicals in the atmosphere are expected to be high; therefore, the formation of nitric acid is highly probable. It is worth mentioning that nitric acid in ambient air reacts immediately with ammonia, generating ammonium nitrate in particulate matter according to the reversible reaction (Equation (5)):
From
Figure 10, it appears that HNO
3, as expected, shows the highest concentrations during sampling carried out in summer. During this period of the year, the temperature and humidity are such that ammonium nitrate (NH
4NO
3) may dissociate back into HNO
3, NH
3, and evaporate [
44]. Thus, nitric acid inside the museum may be generated by particulate matter entering the museum from outside. Ammonium nitrate may migrate inside the rooms where thermodynamic conditions for dissociation may be present, i.e., low humidity and high temperature. In order to show some data about the occurrence of thermodynamic conditions leading to the formation of nitric acid from particulate nitrate, we used data from the environmental data loggers. By adapting the formulas already developed in reference [
44] for the calculation of the deliquescence humidity (RHD) and the dissociation constant K
n [
45] expressed in ppb
2, it is possible to estimate the concentrations of nitric acid:
The evaluation of RHD (Deliquescence Relative Humidity) is the premise for applying relationships/Equations (6) and (7), according to the literature [
46]. If the ambient humidity is higher than the RDH, ammonium nitrate is in the deliquescent liquid phase, and thus, it is not dissociated. From the data concerning T and RH obtained using data loggers installed in the sampling rooms,
Table 6 can be compiled. This shows that the calculated concentration of HNO
3 in equilibrium with ammonia is approximately 2 ppb in winter, reaching 8 ppb in summer. Because the range of observed concentrations is between approximately 6 and less than 2 µg/m
3, these data are consistent with the calculated [HNO
3]
EquationTherefore, the dissociation of ammonium nitrate particulate matter may bring a relatively high amount of this acid inside the museum rooms. It is worth stressing that particulate nitrate does not need to be formed locally because it can be transported over long distances [
47].
The evaluation of nitric acid merits additional comments. This species was evaluated using nitrate ions extracted from an alkaline quartz passive sampler. Simultaneously, the alkalinized carbon paper sampler intended for NO2 was also in operation; therefore, it could be expected that the nitrates found in both samplers are the same. However, the behavior of the samplers with respect to nitric acid was completely different. The average values in the quartz filter show a mean value of 2.14 against 1.94 from the carbon paper. The standard deviation of the first set was 2.14, whereas 1.11 was the value observed for the second. This means that nitrate on carbon filter does not show significant variation. For instance, there were no significant differences between the data collected in room A and those collected in other rooms.
One hypothesis could be the oxidation of collected nitrites by ozone. However, this is in contrast with experimental data that exclude the collected nitrites from being oxidized to nitrate by ozone [
24] and, in addition, the amount of indoor ozone is expected to be very low. In another paper, it was shown that the adsorption of peroxy-organic nitrates on alkaline carbon surfaces also yields nitrates [
48]. Because the concentration of these species is very low, at least compared with that of nitric acid, this source is to be considered not significant. Thus, the presence of nitrates on the carbon paper sampler has not yet been explained and merits further investigation.
4.5. Nitrous Acid (HONO)
Nitrous acid is a weak acid (K
a = 4.0 10
−4) that, at least in principle, moderately impacts the conservation of art-works. However, its presence in the indoor environment is clear evidence of the occurrence of reaction (Equation (1)) that, with nitrous acid, also generates nitric acid, a strong and oxidant acid that has a definite impact on the conservation of exposed materials. In addition, HONO is important for human health because it may generate nitroso-amines [
49]. Therefore, investigations into the presence of HONO in the MANN museum are considered of high interest. The sources of indoor HONO, ref. [
50] can be essentially caused by:
- −
The reaction (Equation (1)) occurring on the surfaces followed by HONO desorption, or
- −
The intrusion of HONO from the external atmosphere.
The extent of reaction (1) depends on the surface, ambient temperature, and relative humidity. Additionally, the reaction is favored by the surface-to-volume ratio (S/V), which, in ambient air, depends on the height of the mixed layer. After the conversion of NO
2, HNO
3 is retained on the surfaces, and HONO can be desorbed into the atmosphere. During atmospheric stability conditions, where the surface-to-volume ratio of the boundary layer is high, the formation of HONO leads to noticeable concentrations in ambient air that can reach several pbb
s [
51].
Consequently, during the early mornings, relevant amounts of HONO may be formed, and thus, photolysis into OH• radicals may occur. If high concentrations of nitrogen oxides are present, high amounts of nitrogen dioxide may also be generated by reaction radicals. This important aspect should be properly considered in future monitoring initiatives.
Using the same reaction mechanism (Equation (1)), we can postulate the presence of HONO inside museum rooms. Since the formation of HONO depends upon the surfaces and the S/V ratio [
51,
52], it is first useful to look into these ratios for the rooms selected as monitoring sites. The amount of HONO, as shown in
Figure 3 and
Figure 4, ranges from a few through 15 µg/m
3 and is only weakly correlated with NO
2 and showing concentrations ranging from 10 to 25 µg/m
3. The amount of HONO decreased during the last period of monitoring, exhibiting maximum values during the first three periods. This is probably due to an increase in ambient temperature causing a more efficient desorption of HONO formed on surfaces.
If we consider HONO found in room A, characterized by a higher S/V ratio, then the amount of HONO is similar to that of NO
2 and sometimes higher, as shown in
Figure 4. This confirms the hypothesis that the HONO found in the museum can be attributed mostly to the indoor formation through Equation (1) on the museum surfaces.
An important point to consider is the possible interference of HONO in the measurement of NO
2. In principle, alkaline carbon paper can quantitatively collect nitrogen dioxide and nitrous acid. However, the active surface of the passive sampler intended for collecting acids ultimately collected HONO. In this hypothesis, the true concentration of nitrogen dioxide [NO
2]
0 is a value lying between [NO
2]
1 and [NO
2]
2, where the former is the value from the carbon paper filter and the second is the difference between the value [NO
2]
1 and the concentration of HONO derived from the quartz alkaline filter:
Data relevant to rooms B, C, and D show that the term [NO
2]
2 is constantly less than [NO
2]
1. However, data from room A, which is characterized by a higher [HONO]/[NO
2] ratio, show that the difference is often less than zero. This means that the interference of the HONO is very low. Data of the nitrite/nitrate molar ratios in acid passive samplers are reported in
Table 7.
As expected, the mean molar ratio is always greater than 1, especially in room A, demonstrating that nitrites and nitrates in the acid filter are not significantly generated by the disproportion of nitrogen dioxide.
This result (nitrous acid collected on the alkaline filter and not on the carbon filter) is not fully unexpected because several acid species, such as formic acid, are not collected by the alkalinized carbon paper filter. As mentioned before, formic acid was not detected in this study because it was emitted by the plastic shield used and then collected in high amounts by the quartz alkaline filter. Both passive samplers were exposed in the same container; however, no formate anion (HCO2−) were found in the carbon paper filter. To provide some context, the equivalent formic acid concentrations found in the glass fiber filter were in the range 60–80 µg/m3, while the amount in the alkaline carbon paper was always zero.
This means that this acid is not adsorbed by the carbon filter. Activated carbon is very effective at removing formic and acetic acids [
53,
54]. The removal of HONO may follow the same basic mechanism. This hypothesis is now being thoroughly tested through direct laboratory and experimental tests.
5. Conclusions
The goal of a simple low-cost monitoring activity and preliminary assessment was fully achieved at the MANN museum in Naples. In five rooms, it was possible to monitor many air pollutants, most of which are crucial for the conservation of exposed art-work objects. Coupling the observations with passive samplers and environmental dataloggers, a complete picture of indoor environment in the museum can be achieved with minimum resources.
The results show that pollutants found in the museum indoor environment often exceeded the maximum recommended concentrations for protecting art-work objects. Nitrogen dioxide ranged from 10 to 25 µg/m3, i.e., from 2 to 3.5 less than the amount found outdoors, and sulphur dioxide was found in the range 1–3 µg/m3. The source of nitrogen dioxide is the intense vehicular traffic in the streets around the building hosting the museum, which releases nitrogen oxides that are converted to nitrogen dioxide.
In addition to nitrogen dioxide, high concentrations of nitrous acid were found. This species is generated on surfaces by the reaction of nitrogen dioxide with water; in fact, the highest amounts were found in rooms where the surface-to-volume ratio is the highest. We were able to demonstrate that the use of alkali impregnated carbon paper for the monitoring of nitrogen dioxide is not significantly affected by the presence of HONO. Nitrous acid is, in turn, efficiently measured with the alkali-impregnated quartz filter.
The presence of HONO is of great concern because its formation also implies the formation of nitric acid, an oxidant and corrosive species. HONO may also come directly from ambient air. In fact, the nitrogen dioxide peaks are associated with very shallow mixed layers, where the formation of HONO is highly probable. Here, in ambient air, the ratio of the surface to the volume of ambient air is small; thus, the formation is highly enhanced. Moreover, the presence of HONO in the atmosphere triggers early morning photochemical processes, increasing the amount of nitrogen dioxide in the atmosphere. The formation of HONO can be partially controlled using appropriate materials applied to the walls.
Data on gas-phase nitric acid showed maximum concentrations during summer sampling. Because the formation in the ambient atmosphere is the direct oxidation of nitrogen dioxide by hydroxyl radicals, nitric acid can be transported from ambient air to the museum either as nitric acid or as ammonium nitrate. The concentration of this species can be as high as 6 µg/m3. This corresponds to a noticeable mass flow rate for deposition on surfaces.
The sources of nitric acid inside the museum could also be due to the dissociation of particulate ammonium nitrate. Such a reaction, according to the indoor thermodynamic conditions, is highly probable to explain the excess of nitric acid in the exposition rooms. For this reason, a campaign about the content of ammonium nitrate in the gas phase is solicited. Sulphur dioxide data are consistent with the average data found outdoors. However, the concentration of this species appears to be low, compared to nitrogen dioxide.
More attention should be paid to the presence of other sulphur species (H2S, H2SO4, etc.) in the museum. The experiments carried out inside the MANN offer very clear ideas about the temporal and spatial distribution of the pollutants and experiments that are recommended to clarify the many aspects discussed in this paper.
The simplicity and low cost involved in this activity, compared with the reliability and amount of gained information, make this campaign an example that can be easily replaced in other museums, where such information is not yet available. Our preliminary assessment could be a good example of how to address the problem of indoor museum contamination by external pollution and to evaluate, even as an early approximation, the risk of exposed art-work objects in thousands of museums and sites where art-work objects are preserved, but where environmental pollution is not considered.
The full evaluation of pollutants is recommended after this preliminary assessment. Priority should be given to particulate matter that should be monitored not only in terms of number or mass concentration but also in terms of basic chemical composition, especially nitrate and ammonium. In addition, it is recommended to include ozone and hydrogen sulphide in future monitoring campaigns, which are additional important pollutants for the conservation of cultural heritage. These species are also efficiently monitored by passive samplers, maintaining the simplicity and low cost, which were the basic objectives of this project.
Finally, it is permissible to have neglected the measurements of O
3, PM10 and PM2.5, VOCs, and the graphs of relative humidity/RH (which have not been reported, unlike those of temperature plot), in accordance with the premise/introduction of the work which considers the contributions of these pollutants in museums to be negligible to the damage equations cited in references [
15,
16,
17] and reproduced in
the Supplementary Information Section of this work.