Seasonal Variability in the Composition of Particulate Matter and the Microclimate in Cultural Heritage Areas

: This study is the ﬁrst attempt to decipher the e ﬀ ect of particulate matter (PM) composition on people’s health and on historic sites, in correlation with the daily and seasonal microclimate monitoring of the indoor and outdoor areas of the Roman Mosaic Ediﬁce museum (the maritime port of Constanta, Romania). More speciﬁcally, the increase of metal concentrations in particulate matter during the summer of 2018 and spring of 2019 in the museum under investigation could possibly be associated with the microclimates of both seasons, with coastal factors, as well as with the anthropic activities speciﬁc to the port of Constanta. FTIR and inductively coupled plasma mass spectroscopy (ICP-MS) techniques, used for the investigation of PM 2.5–10 samples, revealed high concentrations of Fe, Al-rich, and soluble particles inside the investigated museum area. In this respect, the chemical measurements of the PM 2.5–10 masses highlighted high concentrations of heavy metals (i.e., Al, Fe, Zn, Mn, and Pb) and low concentrations of trace metals (i.e., Cr, Ni, Cu, and Cd). Statistical analysis showed that the chemical compositions of the particulate matter in the indoor and outdoor areas of the Roman Mosaic Ediﬁce were inﬂuenced by microclimatic conditions, mainly temperature and relative humidity (RH). A potential health risk for tourists is the thermal and humid conditions, alongside the toxic components of the particulate matter. This research seeks to provide solutions for improving the environmental conditions inside the Roman Mosaic Ediﬁce and to o ﬀ er useful suggestions concerning health promotion and the protection of museum exhibits against possible future deterioration. correlation was found between Mn and Cu concentrations, temperature, and AT. A strong relationship was found between Cr, Fe, Ni, Zn, Cd, and Pb concentrations, temperature, and AT.

Sea, 179 km from the Bosphorus Strait, there are several museums located in old buildings with natural ventilation, low heating and natural luminosity, high humidity, and often with a moldy smell. The safety of the museum environment in relation to the inside air quality, with its direct consequences on exhibited collections and on visitor's health, involves the development of new strategies for indoor/outdoor air pollution restrictions and of new rehabilitation procedures for old buildings.
The Roman Mosaic Edifice is located near the Museum of National History and Archaeology and was discovered in 1959, being part of the ancient town of Tomis (present-day Constanta) [29]. Further research discovered that the monument was built in the fourth century and was gradually expanded until around the sixth century, when the construction work stopped. In its times of glory, the edifice represented the largest building of its kind in the whole Roman Empire and served as a link between the port and the ancient city, the place where it conducted its trade and secured the storage of goods. Originally, the building spanned three of the four terraces of the ancient Tomis harbor waterfront [29]. After the fall of the Tomis fortress in the sixth century, the building fell into decay. However, the high-quality mosaic of the floor has been very well preserved. At present, from a total area of 2000 square meters of mosaic pavement, only a portion of about 850 square meters has been preserved [27]. The floor is of a unique beauty and consists of two distinct parts: a framework, which borders the room perimeter, and the vegetal and geometric patterns of the mosaic itself, which is made up of pebbles of different colors including white, red, yellow, green, and black [29]. Unfortunately, the premises hosting the Roman Mosaic Edifice are improper for storage and visiting purposes. This place is built on a metal structure with large windows, natural ventilation, very high temperatures and humidity during the warmer season, and no heating during the rainy and cold seasons. It is located about two kilometers away from the sea port of Constanta, the main port in the Black Sea, the most important one in Romania, and the fourth most important in Europe.
These environmental conditions (indoor and outdoor) can significantly influence the degradation processes of the materials and components of the Roman Mosaic. The air temperature and relative humidity are key variables of research in the field of environmental protection. Both parameters are hard to monitor, especially at a national scale, due to spatial heterogeneity. The temperature is expected to increase in the course of this century, and extreme values (53.7 • C) were recorded in 2017. The variations in the minimum and maximum temperatures recorded are the main factors that impact the degradation process of the original materials. It is well-known that organic and inorganic materials are extremely sensitive to thermo-hygrometric cycles, especially to shorter ones (i.e., daily cycles, with repeated dilatations/contractions), because they generate steep gradients starting from the outer surface of the object and giving rise to inner tensions, which result in dimensional variations and may lead to irreversible changes in the chemical composition of the original materials.
A potential health risk for visitors is the thermal and humid conditions, together with the toxic components of the particulate matter. Due to climate change and the potential changes in ambient temperatures and RH, it is essential that we seek a better understanding of the indoor thermal conditions and of the air quality in cultural heritage areas.
This research is the first attempt to decipher the effect of PM composition on public health and historical artefacts, in correlation with the monitoring of the daily and seasonal microclimate inside and outside the Roman Mosaic Edifice area. The aim of this study was to investigate the indoor and outdoor conditions of the Roman Mosaic Edifice by several parameters: temperature, apparent temperature (AT), relative humidity (RH), and particulate matter composition. The novelty of this study is highlighted by the multitude of reported data (indoor and outdoor) with direct correlations between climatic factors and PM chemical composition, which has a direct influence on artwork in the museum and on visitor's health.
This research seeks to provide solutions for improving the environmental conditions inside the Roman Mosaic Edifice area and identify useful ways to promote the health safety of visitors and to protect the museum exhibits against possible future deterioration.

Methodology
This research presents the results of an in-field experimental campaign carried out by means of noninvasive measuring instruments, such as a PCE FWS20 weather station and a TE-Wilbur low-volume air particulate matter sampler, equipped with polytetrafluoroethylene (PTFE) filters (d = 0.45 µm, Φ = 47 mm) for particle sampling. These filters were chosen so that a relatively clean IR absorbance spectrum would be obtained [5,24]. The monitoring campaign was undertaken during four seasons-the summer, autumn, and winter of 2018 and the spring of 2019. The measurements and samplings were performed for 24 h/d, 7 d/week, for every month in the risk period (very warm months and rainy, cold months), also taking into account the meteorological predictions provided by the National Meteorological Administration (INMH). A total of 84 filter samples were collected for the purpose of chemical composition examination.
2.1. Temperature, Relative Humidity, and Apparent Temperature Investigations Temperature (T) and relative humidity (RH) measurements were performed by using the PCE FWS20 weather station, with an accuracy of ±0.5% from −25 to 70 • C for temperature and ±2.5% from 11% to 90% for RH. The temperature and RH measurements were used in order to determine the indoor and outdoor thermal conditions. The apparent temperature, the temperature felt by people, was calculated according to the equation [30]: where T is the mean temperature, and T d is the dewpoint temperature measured together with the temperature values.

PM 2.5-10 Sample Analysis
Molecular investigation of the functional groups of inorganic and organic compounds deposited on filters was performed by attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) (Bruker, Wetzlar, Germany), using a Vertex 80v spectrometer (Bruker, Wetzlar, Germany) equipped with diamond ATR crystal accessory for highly refractive index bulk samples, as well as by a Hyperion microscope (Bruker, Wetzlar, Germany). ATR-FTIR spectroscopy has limited applications in quantitative research, since it has a penetration depth of only a few microns, but for qualitative investigation it could be a suitable technique. The ATR-FTIR method did not require a special preparation of samples. The blank was handled exactly as each sample filter from the pre-scan until the final analysis was completed. The samples were chosen based on their black color and thickness (~2 mm) and were placed into 47 mm Petri dishes. The "thick" spectra (recorded from seven samples/season of PM collected on filter) were analyzed in the range of 4000-400 cm −1 (being representative for what can generally be obtained from PM 2.5-10 particles) by non-destructive transmission FTIR spectroscopy.
The concentrations of Al, Cr, Mn, Fe, Ni, Cu, Zn, Cd, and Pb in PM 2.5-10 samples were determined by inductively coupled plasma mass spectroscopy (ICP-MS), using an iCAP™ Qc device (Thermo Scientific, Bremen, Germany). The optimal instrumental parameters and detection limits for the elements analyzed are presented in Table 1. The samples were digested in HNO 3 on a hot plate by using a TOPwave microwave-assisted pressure system (Analytik Jena, Munich, Germany). All chemical reagents were of analytical grade. The measurements were performed in triplicate. The standard reference material (i.e., NIST SRM 1648a, Urban Particulate Matter) was used to verify the accuracy and traceability of the method. The relative standard deviation (RSD) of the standard was 0.36%, the RSD of the samples was 1.2-2.4%, and the recovery of the elements ranged between 92.5% and 104.8%. Table 1. Inductively coupled plasma mass spectroscopy (ICP-MS) instrumental parameters and detection limits for determined metals. Table 2. Descriptive statistics of temperature, relative humidity (RH) and apparent temperature (AT) from the Roman Mosaic Edifice area comparative to reported data [35].      Regarding the RH, high values were recorded in the autumn of 2018 inside the Roman Mosaic Edifice area, where the maximum recommended value for human comfort was exceeded. High outdoor humidity and low outdoor temperatures were measured in the autumn season, and in that case the air felt as chilly as in winter [36].
The mean thermal humidity index (MTHI) was calculated for each season using the Formula (2) [37]: where T and RH are the mean values of temperature and RH. According to the MTHI obtained, the thermal comfort status of the visitor was recorded in the autumn and winter seasons, both indoors (57. 33    Regarding the RH, high values were recorded in the autumn of 2018 inside the Roman Mosaic Edifice area, where the maximum recommended value for human comfort was exceeded. High outdoor humidity and low outdoor temperatures were measured in the autumn season, and in that case the air felt as chilly as in winter [36]. The mean thermal humidity index (MTHI) was calculated for each season using the Formula (2) [37]: where T and RH are the mean values of temperature and RH. According to the MTHI obtained, the thermal comfort status of the visitor was recorded in the autumn and winter seasons, both indoors (57.33 and 51.57) and outdoors (38.78 and 52.40) of the Roman Mosaic Edifice. For the summer and spring seasons, a high thermal comfort status was registered for both the indoor (77.77 and 72.65) and outdoor areas (76.74 and 71.73) of the Roman Mosaic Edifice. No thermal discomfort conditions were recorded in any season. The relationship between outdoor and indoor temperature is shown in Figure 2. Through a linear regression model, a strong correlation was detected between the average outdoor and indoor temperatures (r = 0.90, b = 0.69).

PM2.5-10 Metal Composition
The results of the descriptive statistics analysis of Al, Cr, Mn, Fe, Ni, Cu, Zn, Cd, and Pb concentrations in PM2.5-10 collected from the indoor and outdoor area of the Roman Mosaic Edifice are presented in Table 3 and Table 4.

PM 2.5-10 Metal Composition
The results of the descriptive statistics analysis of Al, Cr, Mn, Fe, Ni, Cu, Zn, Cd, and Pb concentrations in PM 2.5-10 collected from the indoor and outdoor area of the Roman Mosaic Edifice are presented in Tables 3 and 4.   The order of the metals present, corresponding to their abundance in the PM 2.5-10 samples from the indoor area of the Roman Mosaic Edifice, was: Cd < Cr < Mn < Al < Pb < Ni < Zn < Cu < Fe in the summer season; Cd < Cr < Pb < Cu < Ni < Al < Mn < Zn < Fe in the autumn and winter seasons; and Cd < Cr < Al < Pb < Ni < Cu < Cu < Fe < Mn < Zn in the spring season. The order of the metals present, in terms of their abundance in the PM 2.5-10 samples from the outdoor area of the Roman Mosaic Edifice was: Cd< Cr < Al < Pb < Cu < Ni < Mn < Zn < Fe in the summer and spring seasons and Cd < Cr < Pb < Ni < Cu < Al < Mn < Zn < Fe in the autumn and winter seasons. It can be noticed Atmosphere 2019, 10, 595 9 of 17 that in the autumn season, the order of the metals was the same in both the indoor and outdoor areas. A high value of the coefficient of variation (CV), defined as the ratio of the standard deviation to the mean values, was found for Zn (78.73) and Fe (72.84) in the indoor samples collected in the spring seasons. There was a moderate variation in the concentration of Cr (25% < CV < 50%) in the indoor samples collected in all the seasons and in the outdoor samples collected in the summer season as well. There was also a moderate variation in the concentration of Fe (28.27%) in the indoor samples collected in the summer season. Except for the abovementioned situations, a weak variability (CV < 25%) was found for all the metals.
From data reported in Tables 3 and 4, the indoor and outdoor (I/O) ratio was calculated, and the obtained results are presented more clearly in the following graph ( Figure 3). In this respect, two types of observations can be reported: the first group of metals (i.e., Al, Cu, Cd, and Pb) showed high ratios in summer and spring seasons, while the second group of metals (i.e., Cr, Mn, Fe, and Zn) showed high ratios in autumn and winter seasons. This can be explained by the high level of metal pollution in warm seasons, while in cold seasons, the metal content in PM (outdoor) decreased. The Ni was a singular case, and the I/O ratio was constant throughout the entire monitoring process.
The order of the metals present, corresponding to their abundance in the PM2.5-10 samples from the indoor area of the Roman Mosaic Edifice, was: Cd < Cr < Mn < Al < Pb < Ni < Zn < Cu < Fe in the summer season; Cd < Cr < Pb < Cu < Ni < Al < Mn < Zn < Fe in the autumn and winter seasons; and Cd < Cr < Al < Pb < Ni < Cu < Cu < Fe < Mn < Zn in the spring season. The order of the metals present, in terms of their abundance in the PM2.5-10 samples from the outdoor area of the Roman Mosaic Edifice was: Cd< Cr < Al < Pb < Cu < Ni < Mn < Zn < Fe in the summer and spring seasons and Cd < Cr < Pb < Ni < Cu < Al < Mn < Zn < Fe in the autumn and winter seasons. It can be noticed that in the autumn season, the order of the metals was the same in both the indoor and outdoor areas. A high value of the coefficient of variation (CV), defined as the ratio of the standard deviation to the mean values, was found for Zn (78.73) and Fe (72.84) in the indoor samples collected in the spring seasons. There was a moderate variation in the concentration of Cr (25% < CV < 50%) in the indoor samples collected in all the seasons and in the outdoor samples collected in the summer season as well. There was also a moderate variation in the concentration of Fe (28.27%) in the indoor samples collected in the summer season. Except for the abovementioned situations, a weak variability (CV < 25%) was found for all the metals.
From data reported in Tables 3 and 4, the indoor and outdoor (I/O) ratio was calculated, and the obtained results are presented more clearly in the following graph ( Figure 3). In this respect, two types of observations can be reported: the first group of metals (i.e., Al, Cu, Cd, and Pb) showed high ratios in summer and spring seasons, while the second group of metals (i.e., Cr, Mn, Fe, and Zn) showed high ratios in autumn and winter seasons. This can be explained by the high level of metal pollution in warm seasons, while in cold seasons, the metal content in PM (outdoor) decreased. The Ni was a singular case, and the I/O ratio was constant throughout the entire monitoring process.  High concentrations of Pb were found in the samples collected in the summer and spring seasons from inside and outside the Roman Mosaic Edifice, which exceeded the maximum admitted value of Pb concentration in the air, 0.7 mg/kg (Figure 4), according to the Romanian legislation. This can be explained by the fact that the maritime traffic became very heavy in the spring and summer seasons. To sum up, the elements present evidenced high four-season average concentrations (i.e., exceeded only for Pb), thus indicating that the composition of particles was strongly influenced by anthropogenic activities and sea salt.
Pearson correlation analyses for the metal concentration in PM 2.5-10 samples as well as indoor and outdoor temperature, RH, and AT were carried out. The results of the correlation analysis are shown in Tables 5 and 6. Concerning the indoor parameters, a strong relationship was found between Cr, Ni, Cu, Cd, and Pb concentrations, temperature, and AT. Also, a high correlation was found between Al concentrations, temperature, and AT. Concerning the outdoor parameters, a weak correlation was found between Al concentrations, temperature, and AT. Also, a high correlation was found between Mn and Cu concentrations, temperature, and AT. A strong relationship was found between Cr, Fe, Ni, Zn, Cd, and Pb concentrations, temperature, and AT. value of Pb concentration in the air, 0.7 mg/kg (Figure 4), according to the Romanian legislation. This can be explained by the fact that the maritime traffic became very heavy in the spring and summer seasons. To sum up, the elements present evidenced high four-season average concentrations (i.e., exceeded only for Pb), thus indicating that the composition of particles was strongly influenced by anthropogenic activities and sea salt. Pearson correlation analyses for the metal concentration in PM2.5-10 samples as well as indoor and outdoor temperature, RH, and AT were carried out. The results of the correlation analysis are shown in Table 5 and Table 6. Concerning the indoor parameters, a strong relationship was found between Cr, Ni, Cu, Cd, and Pb concentrations, temperature, and AT. Also, a high correlation was found between Al concentrations, temperature, and AT. Concerning the outdoor parameters, a weak correlation was found between Al concentrations, temperature, and AT. Also, a high correlation was found between Mn and Cu concentrations, temperature, and AT. A strong relationship was found between Cr, Fe, Ni, Zn, Cd, and Pb concentrations, temperature, and AT.  Significance level, p: * < 0.05; ** < 0.01. Previous research highlights the fact that the elements mostly concentrated in the accumulation mode are S, As (with chemical speciation), Se, Ag, Cd, Tl, and Pb, while the elements having multimode distributions are Be, Na, K, Cr, Mn, Co, Ni, Cu, Zn, Ga, Mo, Sn, and Sb [5,23,24]. Previous studies have highlighted the relationship between metal compositions and functional groups from PM 2.5-10 in different sizes [5,23,24,26,[38][39][40]. In this respect, the presence of sulfate, carbonate, ammonium, and nitrate groups, as well as of organic functional groups such as aliphatic carbons, carbonyls, and organic nitrates in PM 2.5-10 samples, collected during a complete seasonal cycle was identified by the non-destructive ATR-FTIR technique, as shown by the data presented in Tables 7 and 8. Based on FTIR spectra, the molecular characteristics in PM 2.5-10 were examined as well as the changes in chemical composition under the influence of temperature, RH, and different oxidizing processes. FTIR spectra were acquired rapidly and non-destructively from PTFE filters, which are commonly used for gravimetric mass analysis in regulatory monitoring. After the correction for the background spectrum was made, all the spectra analyzed showed weak and medium vibrational frequencies (Tables 7 and 8 (Table 8). The medium or strong peaks around 400 cm −1 and 777-797 cm −1 were attributed to Si-O from quartz, which was present in all samples. FTIR also identified several organic functional groups, although specific organic molecules could not be identified. The broad bands in the region 3357-3367 cm −1 were assigned to OH-stretching mainly from water. In conclusion, the ions distribution in PM 2.5-10 highlighted several main peaks for NO 3 − , CO 3 2− , SO 4 2− , SiO 4 4− , and NH 4 + groups as well as for organic carbon (i.e., amines, carbonyl compounds) as a major part of the particle mass. In addition, the PM 2.5-10 mass composition highlighted the fact that ions and organic compounds constituted a major part of the particulate matter, while metals and other substances constituted the remaining particle mass. Statistical analysis showed that the chemical composition of particulate matter examined in the indoor and outdoor areas of the Roman Mosaic Edifice was influenced by microclimatic conditions, mainly temperature and RH.

Conclusions
The final results allowed an estimation of indoor and outdoor air quality, from the point of view of the PM chemical composition, thus giving insight into the health risks for visitors and within the Roman Mosaic Edifice museum space hosted in buildings with natural ventilation. The particulate matter analyses showed variability related to indoor microclimate conditions as well as to outdoor, coastal anthropic activities. The FTIR and ICP-MS techniques, used for the investigation of PM 2.5-10 samples, revealed high concentrations of Fe, Al-rich, and soluble particles inside the investigated museum area. The high values of the measured RH in outdoor areas (99% in the monitoring process in the autumn of 2018 and spring of 2019, and over 50% in the indoor area in all seasons), correlated with the temperature and influenced the chemical composition of PM 2.5-10 samples. A strong relationship was found between Cr, Ni, Cu, Cd, and Pb concentrations, temperature, and AT inside the Roman Mosaic Edifice area. On the other hand, a high correlation was found between Al concentrations, temperature, and AT. A low correlation was observed between Al concentrations, temperature, and AT in the outdoor area of the Roman Mosaic Edifice as well as a high correlation between Mn and Cu concentrations, temperature, and AT. In this respect, a strong relationship was remarked between Cr, Fe, Ni, Zn, Cd, and Pb concentrations, temperature, and AT inside the investigated area. The order of the metals analyzed in terms of their abundance in the cold seasons (i.e., autumn and winter, when the temperatures ranged between 0 • C and 10 • C), inside and outside the investigated area, were found to be the same. However, the rise in temperature led to a change in the order of the metals inside the Roman Mosaic Edifice area. The data obtained indicated that as the temperature increases (i.e., in the summer and spring seasons), the Pb concentrations both inside and outside the investigated area are much higher than expected, mainly because of the anthropic activities conducted in the port of Constanta. In conclusion a confined outdoor environment may not be suitable for the conservation of original heritage materials, depending on the climatic region. Several solutions will be proposed in the future, at the end of the project, in order to reduce the impact of the external climatic risk and the consequences of the thermo-hygrometric variations inside the museum, which may have harmful effects on the historical materials and the visitors' health.