Nature-Based Solution for Reducing CO 2 Levels in Museum Environments: A Phytoremediation Study for the Leonardo da Vinci’s “Last Supper”

: This work investigates the possibility of implementing a nature-based solution (NBS) based on the photosynthetic process of Laurus nobilis L. (common laurel), for reducing peak CO 2 concentrations in an air-tight museum environment, namely the Refectory of the Santa Maria delle Grazie Church (Milan, Italy), home of Leonardo da Vinci’s painting “Last Supper”. The phytoremediation potential of laurel plants was evaluated at CO 2 (cid:27) 1000 ppm under controlled environmental conditions. Furthermore, light-saturated net assimilation (Pn max ) was measured at two CO 2 concentrations (380 and 1000 ppm) during the growing season. Steady-state gas exchanges were not a ﬀ ected by elevated CO 2 in the short-term, while Pn max was signiﬁcantly increased, also showing higher values in spring and autumn, and a reduction during summer. Our estimated CO 2 removal rates indicate that, in order to control visitors’ respiratory CO 2 emissions in view of an increase in visitor numbers in the Refectory, a possible NBS in the form of an external greenhouse, connected to the HVAC system of the museum, should allocate from 58 to 112 young laurel plants, depending on their seasonal phytoremediation capacity. These results, although preliminary, allow to hypothesize the possibility of controlling CO 2 indoors through a combination of traditional air-cleaning systems and a properly designed NBS, thus increasing the sustainability of air-tight museum environments.


Introduction
Indoor air quality has become a fundamental issue worldwide, following the increase in time (80-90%) that people spend inside private or public buildings, in particular during winter [1]. Outdoor air can be rich in compounds harmful for people's health, mainly arising from the combustion of fossil fuels and industrial emissions, including carbon monoxide (CO), nitrogen oxides (NO X ), tropospheric ozone (O 3 ), Sulphur dioxide (SO 2 ), volatile organic compounds (VOCs) and particulate matter (PM). Added to indoor-generated air pollutants, these result in high indoor air pollution levels, that can be frequently higher than those recorded outside [2,3]. Among the important molecules necessary to monitor in indoor environments, there is also carbon dioxide (CO 2 ). The use of fossil fuels and land areas" at the visitors' entrance and exit doors. The internal atmosphere is constantly maintained in slight overpressure to prevent the diffusion of pollutants from the surrounding urban atmosphere, and a HVAC system with several air filters is installed to abate pollutants in the incoming air [42,43], as well as to control temperature (maintained at 24-25 • C and 20-22 • C during summer and winter, respectively) and relative humidity (maintained at 50%) [44]. To ensure optimal conditions for the painting, the HVAC system strongly limits the air exchange rate from the outdoor (around 0.67 h −1 ); therefore, in order to prevent the increase of CO 2 levels indoor, the number of visitors and the length of the visit are kept strictly limited (30 people for 15 min, from 8:00 AM to 17:00 PM for 6 days/week) [44]. In spite of this limitation, however, diurnal CO 2 levels in the Refectory room frequently reach the threshold concentration of 1000 ppm [44]. Considering that the museum has more than 300,000 unsatisfied visit requests every year, and the admittance of at least part of such visitors would cause the exceedance of the 1000 ppm CO 2 threshold, an appropriate control of indoor CO 2 is needed, in order to guarantee the wellbeing of visitors and workers, in view of a possible increase in the number of visitors, thus allowing a greater enjoyment of the "Last Supper".
The typical Mediterranean evergreen tree Laurus nobilis L. (common laurel) was considered for the phytoremediation process, due its ability to assimilate CO 2 through the whole year, and because of its symbolic importance in art [45]. Such aspects make this species particularly suitable for the realization of a NBS in this important museum environment. Our specific objectives were: i) estimating the capacity of L. nobilis to reduce indoor CO 2 levels in an air-tight, controlled environment; ii) quantifying the light-saturated maximum net assimilation (Pn max ) of L. nobilis at different CO 2 partial pressures. Such data are needed in order to optimize the design of a possible NBS to control CO 2 in the Refectory of the Santa Maria delle Grazie Church, in view of a possible increase in the number of visitors.

Plant Material and Experimental Design
Laurus nobilis plants, two years old, coming from the forest nursery of the Aurunci Regional Park (Southern Latium, Italy), were used as experimental material. The plants were grown in the Experimental Garden of the Department of Environmental Biology, Sapienza University of Rome (Italy), inside 7 l pots, filled with garden soil supplemented with a slow releasing fertilizer (Nitrophoska Blue, 12-12-17 and microelements), and watered daily to prevent the occurrence of stress conditions. Two experimental activities were carried out. Firstly, the response of gas exchanges to the maximum CO 2 values measured in the Refectory during the visiting hours ( 1000 ppm, Table 1), has been investigated though a CO 2 fumigation experiment, carried out under controlled environmental conditions. Furthermore, the photosynthetic response of L. nobilis to varying light intensity (Pn vs. PAR response curves) has been measured under different CO 2 partial pressures (ambient, 380ppm, and 1000ppm) during spring, summer, and autumn 2017.

CO 2 Fumigation Experiment in Controlled Environmental Conditions
The fumigation experiment was carried out at the Department of Environmental Biology, inside two closed "walk-in" chambers (2.5 × 3.9 × 3.0 m), one used as control and one for CO 2 fumigation. Air temperature was maintained at 27 ± 1.2 • C, and 20 ± 1.5 • C (mean ± s.d.) during day and night, respectively. Relative humidity (RH) was 59 ± 4%, and vapor pressure deficit (VPD) was 1.4 ± 0.3 kPa. In each chamber, a photosynthetically active radiation (PAR) of approximately 700 µmol m −2 s −1 was provided for 12 h per day by 6 metal halide lamps (1000 W, Philips HPI-T). Microclimatic conditions were monitored at 6-min intervals, and did not differ significantly between chambers. Plants were watered at full capacity over the whole duration of the experiment.
Prior to the start of the fumigation, 30 plants were left to acclimate for 5 days to the chamber conditions. At the end of this acclimation period, they were grouped into two experimental sets: 15 plants were fumigated with CO 2 at 988.7 ± 37.1 ppm h for 4 h per day (09:00-13:00), for 5 days (set "F"), while the other 15 were employed as control (not-fumigated, set "C"). In the fumigation chamber, pure CO 2 was supplied from a CO 2 cylinder into the air-mixing, for dilution with ambient air. The control chamber was instead supplied with ambient air only (CO 2 was kept at 380 ± 10 ppm). CO 2 concentrations inside the chambers were monitored continuously through an infra-red gas analyzer (IRGA, model CIRAS 2, PP Systems, Amesbury, MA, USA, employed in "analyser only" mode, without the foliar cuvette).
Steady-state gas exchanges were measured both continuously, during CO 2 fumigation, and as "spot" measurements, carried out at the end of the fumigations. Both measurements were carried out on fully developed leaves placed at the top of the plants.
Continuous measurements were carried out on a different L. nobilis individual for each fumigation day (5 plants in total), in order to assess the instantaneous photosynthetic capacity of the species under high CO 2 levels. Each leaf was inserted in the foliar cuvette and left acclimatizing for 15 min prior to the start of the fumigation. Gas exchange parameters were recorded automatically every half minute by the CIRAS2 for the whole duration of the fumigation treatment, under the environmental conditions and CO 2 levels reached in the chamber. About 400 measurements were recorded for each day of fumigation (DOF).
At the end of each fumigation (between 14:00 and 16:00), "spot" gas exchange measurements were also carried out on both F and C plants, in order to assess plant functional response to the high CO 2 treatment. One leaf per plant was measured, for a total of 15 measurements for sampling date and experimental set. During these spot measurements, CO 2 in the leaf cuvette was maintained at normal atmospheric concentration ( 380 ppm).

Photosynthetic Light Response at Different CO 2 Concentrations
The CIRAS 2 instrument was also used to measure the response of photosynthesis to changing light intensity (Pn vs. PAR response curves), under different CO 2 partial pressures. Such measurements were carried out in the laboratory between 09:00 and 14:00. Cuvette environment was maintained at 60% RH and 25 • C T, while photosynthetic active radiation was progressively reduced step-wise from the maximum value of 1200 µmol photons m −2 s −1 up to 0, by using the CIRAS 2 built-in light emitting diodes. Each intensity level was maintained for 5-7 min to allow photosynthetic response to stabilize before taking the measurements (6-10 records for each light step). For each plant, Pn/PAR response curves were measured at two CO 2 concentrations: 380 ppm (ambient value) and 1000 ppm. CO 2 concentration in the leaf cuvette was controlled using the CIRAS 2 inner CO 2 cartridge. From the analysis of the Pn/PAR curves, the following photosynthetic traits were derived: Pn max (µmolCO 2 m −2 s −1 ), the maximum photosynthesis at saturating light intensity, which indicates the instantaneous CO 2 removal performance on the basis of leaf area; Φ (µmolCO 2 µmol photons −1 ), photosynthesis quantum yield, which indicates the amount of CO 2 fixated for each absorbed photon [46]. The quantum yield was calculated empirically, as the slope of the linear part of the Pn/PAR curve (between 50 and 250 µmol photons m −2 s-1). The number of points considered was assessed for each single curve according to the "best fit" method, considering a minimum of four data points [47].
The measurements were carried out in four different moments of the vegetative season 2017: Spring (between 15 and 19 of May), early Summer (from 20 June to 7 July), late Summer (from 19 to 25 July) and in Autumn (from 9 to 13 October). For each period, 3-4 plants were measured.

Estimate of Laurus nobilis Capacity to Remove Indoor CO 2
The instantaneous CO 2 removal performance on the basis of leaf area, derived from Pn/PAR response curves (Pn max, µmolCO 2 m −2 s −1 ), was used to develop preliminary baseline data for the planning and design of a pilot plantscape installation, targeted at reducing CO 2 levels in the Refectory of Santa Maria delle Grazie Church. The Refectory is a large rectangular hall measuring 36.65 × 8.85 m, and with a height of 10.3 m, isolated from the outside with a system of air-tight triple doors. Therefore, it is possible to calculate the total volume of air indoors, which is approximately equal to 3340 m 3 . Assuming air as an ideal gas at standard temperature and pressure, it is possible to roughly estimate the molar amount of CO 2 contained in the Refectory as follows: where V R is the volume of air indoor (3340 m 3 ), X is the maximum daily CO 2 concentration in ppm measured in the Refectory (996.18 ppm, Table 1), and 0.022414 m 3 mol −1 is the molar volume of air.
Assuming an average of 12h of photosynthetic activity, the daily CO 2 removal by 1 m 2 leaf area of L. nobilis is equal to: The residual CO 2 concentrations in the Refectory is then easily derived as the difference between the initial molar concentration Equation (1), and the CO 2 subtracted after 1 day of photosynthetic activity by 1 m 2 of leaf area Equation (2). By reverting Equation (1), this value can be expressed again in ppm, and used for a preliminary dimensioning of the total leaf surface area needed to be allocated in a NBS, in order to satisfy the need to increase the daily number of visitors allowed in the Refectory without exceeding the recommended 1000 ppm CO 2 threshold. Specifically, a potential increase of the daily visitors' number of about one third was evaluated.

Statistical Analysis
Continuous gas exchange measurements were analyzed by a One-way Analysis of Variance (ANOVA). In order to determine significant differences in gas exchange parameters between control and CO 2 -fumigated plants for each day of fumigation, a T-test at p < 0.05 was applied. The parameters derived from the Pn/PAR response curves were analyzed with the two-way ANOVA, considering CO 2 concentration and month of measurement as factors. Two-way ANOVA was followed by post hoc Neuman-Keuls test at p < 0.05 when necessary. Data reported in figures and tables are expressed as mean ± standard deviation. All analyses were performed by using the software STATISTICA 7.0 (Statsoft, OK, USA). Figure 1 shows the hourly trend of continuous gas exchanges, recorded on laurel plants during the CO 2 fumigation experiment. There was no significant difference (p > 0.05) in gas exchanges between fumigation days, thus only hourly values averaged over the whole fumigation period are shown. Net photosynthesis (Pn, Figure 1a), stomatal conductance (gs, Figure 1b  "Spot" gas exchange measurements recorded after each fumigation are shown in Figure 2. It is interesting to highlight that net photosynthesis was significantly reduced (p < 0.01) in CO 2 fumigated plants only during the second day of fumigation (DOF 2, Figure 2a), while stomatal conductance did not differ significantly between control and fumigated plants through the whole experiment (Figure 2b). The Ci/Ca ratio showed a slight, but significant (p < 0.05), increase in fumigated plants at DOF 2, concurrently with the Pn reduction (Figure 2c).  Figure 3 shows the average Pn/PAR curves, recorded on L. nobilis plants in different months from May to October, at two CO 2 concentration (380 ppm and 1000 ppm, Figure 4a,b, respectively). For both CO 2 concentrations and during all measuring dates, maximum net assimilation Pn max was reached at 1000 µmol photons m −2 s −1 , that therefore corresponds to the saturating light intensity for L. nobilis.  No marked seasonal trend is evident in the Pn/PAR curves measured at [CO 2 ] 380 ppm (Figure 3a). On the contrary, the curves measured at 1000 ppm clearly show a different seasonal response to changing light intensity.

Estimate of Laurus nobilis Capacity to Remove Indoor CO 2 from the Refectory of Santa Maria delle Grazie Church
Applying Equation (1), the maximum amount of CO 2 contained within the Refectory can be estimated as equal to 148.44 mol. Then, the residual CO 2 in the Refectory after 12h of photosynthetic activity operated by 1 m 2 leaf area of L. nobilis was derived. In this regard, in Equation (2), the maximum and minimum Pn max recorded during the Pn/PAR curves at 1000 ppm CO 2 in May ( 23 µmolCO 2 m −2 s -1) and July (12 µmolCO 2 m −2 s −1 ), respectively (Figure 4b), were considered, obtaining maximum and minimum daily CO 2 removal of 0.99 and 0.52 mol CO 2 m −2 day −1 , respectively. Accordingly, the residual CO 2 in the Refectory after the photosynthetic assimilation ranged from a minimum of 147.45 mol CO 2, to a maximum of 147.93 mol CO 2 , corresponding to maximum and minimum Pn max values, respectively. Such values can be reconverted into ppm by reverting Equation (1), thus deriving the maximum (6.67 ppm) and minimum (3.48 ppm) daily potential improvement of CO 2 concentration operated by 1 m 2 of leaf surface of L. nobilis.
These values can be used to estimate the potential total leaf surface area that should be allocated in a NBS, in order to satisfy the need to increase the daily number of visitors allowed in the Refectory without exceeding the recommended 1000 ppm CO 2 threshold. Considering that, due to the current number of visitors, maximum indoor CO 2 levels in the Refectory increase 600 ppm over the ambient value (i.e., from 400 ppm to 1000 ppm), a reduction of one third such value ( 200 ppm) would lower maximum CO 2 to 800 ppm, thus allowing a potential increase of the daily number of visitors of about one third (i.e., from 30 to 40 people each 15 min round). The leaf surface area needed for such reduction would therefore range from a minimum of 29 to a maximum of 56 m 2 . Since potted laurel plants of about 1.5 m of height have an average total leaf area of 0.5 m 2 , it is possible to estimate that the approximate number of plants that should be necessary to accommodate within the NBS in order to obtain such ideal CO 2 reduction would range between a minimum of 58 to a maximum of 112 plants.

Discussion
Plant phytoremediation potential for indoor CO 2 is underpinned by the photosynthetic process, therefore measuring gas exchanges under high [CO 2 ] is fundamental for planning a NBS targeted at reducing indoor CO 2 levels [34,37,48]. The gas exchange rates recorded in our experiment are in the range of those reported for laurel plants in natural environments (2.8 ≤ Pn ≤ 5 µmolCO 2 m −2 s −1 and 20 ≤ gs ≤ 55 mmolH 2 O m −2 s −1 , [49]). Such low values of photosynthetic traits are a common characteristic of evergreen sclerophyllous woody species, representing a physiological adaptation to the stressful conditions typical of the Mediterranean environment [50,51]. Indeed, Maatallah et al. [52], who investigated the drought effects on L. nobilis seedlings, found a maximum net photosynthesis of 3.8 µmol CO 2 m −2 s −1 and a minimum of 0.5 µmol CO 2 m −2 s −1 on well-watered and drought stressed plants, respectively.
Besides a transient reduction of net assimilation occurring during the first day of fumigation, no significant effect of high CO 2 on steady-state gas exchanges was evident. Continuous measurements recorded during the fumigations showed only slightly higher Pn and lower gs values, coherently with what reported for other Mediterranean evergreen tree species under high [CO 2 ] [53]. Photosynthetic quantum yield (Φ) and light-saturated net photosynthesis (Pn max ) were instead significantly increased by high CO 2 during the whole growing season. There is a large amount of literature about the effects of high CO 2 on plants (see the reviews by Hättenschwiler et al. [54], and Ainsworth and Long [55]), highlighting that elevated CO 2 increases carboxylation and reduces photorespiration, particularly in those functional groups, such as trees, whose photosynthesis is Rubisco-limited also at elevated [CO 2 ] [56]. Supporting evidence of stimulated CO 2 assimilation at high [CO 2 ] on Mediterranean evergreen trees was also provided under chamber conditions [57,58], as well as in natural CO 2 springs [59]. In the long-term, however, acclimation to high CO 2 could occur, determining a down-regulation of photosynthesis caused by the reduction of Rubisco activity or content, the decrease in stomatal aperture and density, or the occurrence of nutrient imbalances [59][60][61]. Such downward acclimation must be taken into account when planning a NBS targeted at reducing indoor CO 2 , since it could negatively affect plant phytoremediation potential, thus making necessary to continuously replace plant material with an additional economic cost. It is, however, worth emphasizing that there are few evidences of photosynthetic acclimation to CO 2 occurring in trees, particularly in young or juvenile individuals [55,62], thus making young potted trees a more sustainable choice for an indoor NBS, if compared to herbaceous species. Furthermore, it is unlikely that such photosynthetic acclimation would occur in the NBS considered in this study, since CO 2 level within the Refectory of Santa Maria delle Grazie Church varies with the flux of visitors, dropping to values as low as 300 ppm during the museum closing hours (Table 1).
Besides acclimation, the potential stomatal sink strength for CO 2 is often limited by the suboptimal growth conditions typically found in indoor environments [36,39,48]. Among these, low light quality and intensities, as well as the lack of an appropriate light-dark cycle (photoperiod), have been recognized as the main sources of limitation for the physiological activity of indoor plants [63]. In our case, light saturation of photosynthesis in L. nobilis occurred at PAR 1000 µmol m −2 s −1 , similarly to what highlighted for other woody species in interiorscapes [31]. Interestingly, our results also highlight that, at high [CO 2 ], both Φ and Pn max showed a marked seasonal trend, with lower values during summer in respect to spring and autumn. The seasonal dynamics of photosynthesis in Mediterranean evergreen species has been extensively studied, mainly in relation with drought or high temperature stress occurring during the summer period [50,64,65]. In our case, since water availability was not limited through the whole experiment, we can argue that both ambient air temperatures (Supplementary Materials Figure S1) and plant phenology, played a role in determining the observed seasonal trend, ultimately affecting the CO 2 phytoremediation potential on a seasonal basis [66].
The abovementioned factors should be taken into account when transferring the removal rates estimated in laboratory experiments to the planning of a NBS, particularly in relation to a challenging museum environment such as the Refectory of Santa Maria delle Grazie Church. In this case, the most feasible way to realize a NBS would be to connect the HVAC system of the Refectory to an external greenhouse, in which plants could be grown under controlled environmental conditions, in order to guarantee optimal irrigation, fertilization, temperature and irradiance levels (PAR 1000 µmol m −2 s −1 ) to sustain maximum CO 2 removal in the museum. The CO 2 -rich air recirculating from the museum would be filtered by plants in the greenhouse, and then streamed back in the Refectory, after an additional filtration from the HVAC to control water vapor, VOCs and PM that could be emitted by plants and visitors as well ( Figure 5). However, our estimated CO 2 removal rates per unit leaf area point out that, for lowering [CO 2 ] in the Refectory to 800 ppm to allow a potential one third increase in the number of visitors, the external greenhouse should allocate from a minimum of 58 to a maximum of 112 laurel plants of 1.5 m of height. This would require a minimum greenhouse surface area of 30 m 2 . Based on these results, the use of plants as a standalone means to control indoor CO 2 in the Refectory appears impractical, mainly because of the large volume of the Refectory itself, that requires a massive phytoremediation effort. This is in agreement with previous studies, underlining that botanical air filtration systems, such as green walls or potted plants, can be self-sufficient for passive air phytoremediation only in the case of small rooms with few occupants [34,36,67]. Therefore, in the specific case of the Refectory, the use of physicochemical CO 2 filtration systems cannot be fully avoided. Despite the abovementioned limitations, the NSB based on plant phytoremediation can be evaluated as an additional installation to improve the efficiency of traditional filtration systems, to reduce CO 2 levels while releasing an equimolecular amount of oxygen in such a challenging museum environment. In particular, phytoremediation efficiency will be higher during periods of maximum plant physiological activity, such as spring and autumn, thus increasing the efficiency of CO 2 filtration. In this sense, our results support previous researches advocating the possibility to combine traditional air-cleaning technologies with properly designed NBS, also with the help of smart sensor technologies [39], as a promising solution for improving indoor air quality, opening the road for the possible optimization of energy use and maintenance cost of the HVAC systems also in air-tight museum environments.

Conclusions
Our work represents the first quantitative investigation aimed at providing science-based evidence and knowledge for planning a NBS in an air-tight museum environment, where high CO 2 levels limit the visitors' fruition of an important work of art such as the Leonardo da Vinci's "Last Supper". Such a NBS could improve the efficiency of traditional air cleaning technologies, to control [CO 2 ] in the considered museum in view of a potentially increased number of visitors.
There is, however, the need of further research, in order to develop optimum combined systems for such pioneering installations, targeted at solving an important societal issue such as indoor air pollution. For instance, besides photosynthetic acclimation, daytime respiratory CO 2 emissions from the soil substrate, including microbial and root respiration, should be taken into account when calculating the CO 2 removal rates within the greenhouse. The fertilization and irrigation schedules needed to sustain the maximum CO 2 removal rate, as well as the replacement cycles of plant material, should be also optimized. Furthermore, the costs of energy consumption for potential nature-based or traditional air-cleaning solutions, as well as for a combination of both, should be quantified and compared, in order to determine the best choice in terms of both cost effectiveness and energy efficiency. It is also necessary to test further plant species, with higher CO 2 removal rates than L. nobilis, as well as other, more efficient botanical air filtration systems, such as for example green walls, which reduce the space needed to allocate the plant material and, at the same time, limit the soil substrate and its CO 2 emissions. In this regard, the system phytoremediation capacity towards multiple indoor air pollutants, such as PM, NOx, and O 3 , should be also considered and quantified for a proper cost/benefit evaluation. Finally, it must be underlined that our modelling approach represents a simplification of the issue of building ventilation, since the amount of CO 2 in the indoor air is a dynamic function of different parameters, such as Tair, RH, as well as the number of air exchanges, CO 2 concentration outdoors, and the number of people producing CO 2 allowed indoor. Therefore, a CO 2 dynamic model should be implemented and tested for planning the specific case-study.
In spite of this, our data represent the starting point for the development of such a NBS, opening the road to the possibility of their implementation in other indoor environments, such as schools, hospitals, or public offices, through the employment of the natural and low cost photosynthetic process, rather than relying on "grey" technologies only. The "Cenacle", that preserves Leonardo's genius, can thus become a model of sustainability, and pave the way to the development of innovative solutions for the future.