Air Cleaning Performance of Two Species of Potted Plants and Different Substrates

Potted plants have been reported to uptake VOCs and help ‘cleaning’ the air. This paper presents the results of a laboratory study in which two species of plants (Peace Lily and Boston Fern) and three kinds of substrates (expanded clay, soil and activated carbon) were tested and monitored on their capacity to deplete formaldehyde and CO 2 in a glass chamber. Formaldehyde and CO 2 were selected as indicators to evaluate the bio-ltration ecacy of 28 different test conditions; relative humidity (RH) and temperature (T) were monitored during the experiments. To evaluate the ecacy of every test the Clean Air Delivery Rate (CADR) was calculated. Overall, soil had the best performance in removing formaldehyde (~ 0.07–0.16 m 3 /h), while plants, in particular, were more effective in reducing CO 2 concentrations (Peace lily 0.01m 3 /h) (Boston fern 0.02-0.03m 3 /h). On average, plants (~ 0.03 m 3 /h) were as effective as dry expanded clay (0.02–0.04 m 3 /h) in depleting formaldehyde from the chamber. Regarding air cleaning performance, Boston ferns presented the best performance among the plant species, and the best performing substrate was the soil.


Introduction
Studies have shown that poor Indoor Air Quality (IAQ) affects human health in a long-term exposure (WHO, 2010). In the INDEX project ) several chemicals, their concentration levels and their toxicity information were analysed and evaluated in indoor environments. It was concluded that Volatile Organic Compounds (VOCs), such as benzene, toluene and xylene, together with aldehydes should be considered as priority pollutants regarding their health effects. Several studies related with IAQ have indicated that VOCs are emitted by indoor sources such as building materials, furnishings and cleaning products (Bluyssen et al. 1997; Bluyssen et al. 1996;Brown et al. 1994;Campagnolo et al. 2017; Sofuoglu et al. 2011). In 1998, Yu and Crump published a review on VOC-emissions from newly built houses (Yu and Crump 1998). They stated that building material emissions are the sources of VOCs in the indoor environment, especially most during the rst six months after construction. Among the indoor pollutants, VOCs are ubiquitous and have harmful effects on human health such as asthma, wheezing, allergic rhinitis, and eczema.
VOCs are frequently classi ed according to their boiling point (Bluyssen 2009): very volatile organic compounds (VVOCs), such as formaldehyde; VOCs, such as solvents and terpenes; Semi VOCs (SVOCs), such as pesticides; and Particulate Organic Matter (POM), such as biocides. Regarding IAQ, VOCs and VVOCs are the pollutants most frequently found indoors (Wolkoff 2003). Some of them are toxic and carcinogenic, such as formaldehyde; and in general, exposure to formaldehyde is higher indoors than outdoors (IARC

Sources of formaldehyde
Formaldehyde is released directly into the indoor air from various types of sources. People are exposed to environmental formaldehyde from adhesives, lubricants, wall coverings, rubber, water-based paints, cosmetics, electronic equipment, and glued wood-based products. For instance, formaldehyde is known to be emitted considerably by chipboard, MDF, plywood and other woodbased products containing resins (Bluyssen et al. 1996;Campagnolo et al. 2017). Next to these building materials, formaldehyde is a component of tobacco smoke and of combustion gases from heating stoves and gas appliances. It is used as a disinfectant and as a preservative in biological laboratories. It is also used in the fabric and clothing industry.
Major sources of formaldehyde in non-smoking environments are building materials and consumer products. This applies to new materials and products and can last several months, especially in conditions with high relative humidity (RH) and high indoor temperatures (Haghighat and De Bellis 1998;Knoeppel 1990;Salthammer et al. 2010). Formaldehyde is also one of the main components for resins, which are contained in various products, mainly in wood products. Furthermore, it should be noted that secondary formation of formaldehyde occurs in air through the oxidation of VOCs. However, the in uence of these secondary chemical processes to the ambient and indoor concentrations has still not been fully measured (Kaden 2010).

Health effects of formaldehyde
In general, humans are mainly exposed to formaldehyde through inhalation. Since formaldehyde is soluble in water, it is rapidly absorbed in the respiratory and gastrointestinal tracts and metabolized (WHO 2010). Predominant symptoms of formaldehyde exposure in humans are irritation of the eyes, nose and throat, discomfort, sneezing, coughing, nausea, among others (WHO 2000).
The lowest concentration may cause sensory irritation of the eyes with humans, increasing eye blink frequency and conjunctival redness (WHO 2010).

Formaldehyde guidelines and regulations
In the Netherlands, several formaldehyde measurement studies have been executed specially in homes and at schools, where there were complaints which might have been caused by formaldehyde. Several complaints were connected with a concentration above 120 µg/m 3 . In Dutch schools the highest concentration measured was 2.5 mg/m 3 . In homes, the highest concentrations found were between 0.75 and 1 mg/m 3 (Knoeppel 1990). In 2011, Van Gemert reported that the odour thresholds for formaldehyde can uctuate from 0.03 to 2.2 mg/m 3 (Van Gemert 2011).
WHO 2010 reported that the lowest concentration to cause sensory irritation of the eyes in humans is 0.38 mg/m 3 for four hours.
Besides, a formaldehyde concentration of 0.6 mg/m 3 increases eye blink frequency and conjunctival redness. Regarding the perception of odour of formaldehyde, some individuals reported sensory irritation, and formaldehyde may be perceived at concentrations below 0.1 mg/m 3 . However, this is not considered to be an adverse health effect (Kaden 2010;WHO 2000WHO , 2010).

Effects of plants on formaldehyde removal
It has been well established that potted-plants can help to phytoremediate a diverse range of indoor air pollutants. In particular, a substantial body of literature has demonstrated the ability of the potted-plant system to remove VOCs from the indoor air. These In 2011, Aydogan and Montoya tested the formaldehyde removal e ciency of the root area and aerial parts independently and found that while the aerial parts of plants were capable of VOC removal, removal by the root area occurred at a substantially faster rate (Aydogan and Montoya 2011). Other research has identi ed the potential for the microorganisms existing on and in leaves to remove VOCs (Khaksar 2016;Sandhu et al. 2007). However, most recent research has acknowledged that the mechanisms of removal are mainly located in the substrate, rather than the plant itself (Kim et al. 2008;Orwell et al. 2004;Wood et al. 2002).
Based on the studies mentioned, it is valid to assume that plants together with is substrate can have a positive removal effect on the concentration of formaldehyde in indoor environments. However, the extent to which different plants remove formaldehyde is not well known yet. This paper presents the results of a study on the uptake of formaldehyde and CO 2 from selected potted plants and substrates, with the objective of using the outcome of these experiments to select the best performing plant and substrate for the construction of an indoor plant-based system (biowall).

Experimental setup
The setup, schematically presented in Fig. 1, consisted mainly of a dynamic chamber. The dynamic chamber was made out of glass with an inner diameter of 28 cm, height of 60 cm and volume (V) of 0.033 m 3 . The glass chamber had three air entrances that were sealed during the tests. The gas stream of 300 ppb concentration of formaldehyde was released in the chamber by heating the formaldehyde solution.
The actual formaldehyde concentration was determined by a formaldehyde sensor (DART-sensor 11 mm, calibrated, ppb-level, lower detection limit of < 30 ppb, response time (T90) < 30 s, resolution 10 ppb). Two axial fans were placed into the glass chamber to distribute the air evenly within the chamber. The sensor performed a measurement every minute. During the tests a LED growing lamp was activated (1500 µmolm − 2 s − 1 -1900 µmolm − 2 s − 1 ), and the temperature, relative humidity and CO 2 levels were also monitored. CO 2 levels were monitored with VAISALA CO 2 probe GMP252 (ppm-level). Furthermore, the glass container was sealed with a solvent free, plastic, self-adhesive sealant, kneading material, based on synthetic rubber during the tests.

Chemicals
The formaldehyde solution used for these experiments was: Solution Sigma F8775, 25 ml (36.5-38% formaldehyde in H 2 O). The formaldehyde solution was mixed with demi-water in order to generate 300 ppb within the chamber. The mixture was executed by technicians in the laboratories of the University of Wageningen, as follows: 10 µl formaldehyde + 90 µl demi-water = 100 µl ( nal mixture) 10 µl of the nal mixture generated 300 ppb of formaldehyde, within the chamber.
It is important to report that the formaldehyde solution contained 10-15% of methanol, as stabiliser to prevent polymerisation. The DART-sensor is also sensitive to methanol. So, by introducing formaldehyde, a small amount of methanol was introduced as well. The response of the DART-sensor to this amount of methanol therefore also needed to be tested.

Preparation of the substrates
Three different growth media were chosen for the test: soil, activated carbon and expanded clay. The selected potting soil was composed by peat, green compost, lime and fertilizers. The selection of the substrates was based on previous studies and because they are common substrates available on the market (Aydogan and Montoya 2011; Wolverton et al. 1989). For every type of substrate six tests were executed, three with a dry substrate and three with a wet substrate. The substrates were placed each in a plastic container with a capacity of 1.1 litres (0.0011 m 3 ) with 0.14 m diameter in the upper part, which was the exposed area of the substrate.

Preparation of the plant samples
Two different plant species were tested: Spathiphyllum Wallisii Regel (Common name: Peace Lily) and Nephrolepis exaltata L.
(Common name: Boston Fern) (Fig. 2). Three plants of every species were chosen for the tests and they were selected with similar characteristics of age and size (Peace lily: 0.35m height; Boston fern: 0.30 m height). The plants were selected based on information gathered by previous studies, which demonstrated that the capability of these species in uptake of some VOCs was good (Liu et al. 2007; Wolverton and Wolverton 1993;Wood et al. 2002). And they were also chosen because they can be used in Living Wall Systems (LWSs) and/or green walls, besides, they are commonly used for indoor decoration. The plants were bought in a house-plant shop in the Netherlands and were re-potted 25 days prior the experiments, to minimize the stress of the plant, in a 14 cm diameter plastic pot of 1.1 litre (0.0011 m 3 ) of expanded clay growth medium. The expanded clay was selected as a growth medium for the tests because it is the most common substrate used indoors and it is most suitable to be used in indoor living wall systems. All the plants went through a 30 min acclimatization and adaptation process in the laboratory where they were exposed to similar conditions, in order to minimize the stress of the plants prior the execution of the tests.

Procedure
Two zero-measurement evaluations were performed to establish the conditions of the set-up in the glass container in which the depletion of the formaldehyde took place: one at the beginning of the test series and one at the end. Similarly, two extra zeromeasurement evaluations were performed with a plastic container that had the same characteristics of the containers that were used during every test.
The measurements were executed for 1-1.5 hours until the formaldehyde was depleted or stabilized in the chamber. Gas concentrations were measured in ppb in the case of formaldehyde and in ppm in the case of CO 2 . For further analysis the concentrations of these gases were expressed as micrograms per cubic meter (µg/m 3 ) and milligrams per cubic meter, respectively.
For each test, ~ 368.48 µg/m 3 (~ 300 ppb) of formaldehyde was released in the chamber to generate every time exactly the same condition.
Each set of experiments was conducted three times, in order to evaluate consistently each condition tested (Tables 1 and 2). For each test, the glass container was wiped with a wet paper towel after each measurement. The plastic container with the substrate or plant sample was placed in the centre of the glass chamber. Depending on the height of the plant a stainless-steel base was placed at the bottom (stainless steel is an inert material).
A small plate connected to a heat source was placed in the lower hole and 10 µl of formaldehyde solution was placed on the plate with a pipette. After a drop of formaldehyde solution was placed on the plate, the hole was closed, and the heat source was activated in order to realise the solution in the air. This was the beginning of the test. During the tests with the Boston ferns, it was necessary to inject some CO 2 when the level was lower than ~ 410 ppm (~ 738 mg/m 3 ) which is the global atmospheric CO 2 concentration (average outdoor concentration) (IPCC, 2014; NASA, 2019) and is su cient for the plants to grow although some studies have shown that the optimal CO 2 concentration is around 900 ppm (Zheng et al. 2018).
To calculate the amount of formaldehyde depleted inside of the chamber the following formula was used (Irga et  A portable leaf area meter was used to scan and calculate the leave area of the plant species. Since the three plants of every species had similar characteristics, one plant of every species was selected to be measured (Fig. 3).
Conversions for chemicals in air were made assuming an air pressure of 1 atmosphere and an air temperature of 25 degrees Celsius. The conversion factor was based on the molecular weight of the chemical and is different for each chemical in this case the molecular weight of formaldehyde is 30.031 g/mol and of the carbon dioxide (CO 2 ) is 44 To stablish the statistical signi cance of the results, several Independent T-Tests were executed and the mean values and standard errors (± S.E.) were included. Finally, the one-way analysis of variance (ANOVA) was chosen to determine whether there are any statistically signi cant differences between the means of the tested variables. Additionally, a Pos-Hoc test was also required to con rm where the differences occurred. Based on the nature of this data set, Tukey HSD and the Student-Newman-Keuls were performed to execute a multiple comparison among the groups and to determine homogeneous sets. Tables 3 and 4 present the statistical analysis of the CADR caused by the selected growth media and selected plants.

Results
During the zero measurements of the setup, the sensor indicated the presence of around 30.7 µg/m 3 (25 ppb) of formaldehyde in the system. It is believed that this value was due to the calibration process. The zero measurement tests indicated that the formaldehyde decreased slowly in the chamber (Figs. 3-6), which could be the natural decay of the gas or because it was partially adsorbed by the setup. When the plastic container was placed inside of the chamber the reduction slightly increased, which shows that the formaldehyde was adsorbed by the container. These two values have to be taken in account when analysing the real effect of the substrates and plants regarding formaldehyde depletion (Table 1). Therefore, to calculate the CADR and establish the real aircleansing-impact of every test condition, the natural decay of the chamber (λn = 0.11 h − 1 ) and the decay rate of the plastic container (λp = 0.15 h − 1) were subtracted from the total decay rate (Tables 1 and 2). Figure 4 presents the depletion of formaldehyde when expanded clay was tested, under dry and wet conditions, indicating that wet expanded clay was more effective on depleting formaldehyde than under dry conditions. Among all the conditions tested, soil was the most effective element to reduce formaldehyde in the chamber, especially under wet conditions (Fig. 5). Figure 6 shows that activated carbon under dry conditions was more e cient than under wet conditions in reducing formaldehyde in the chamber.
With regards to CO 2 levels, potted plants seemed to be the only test condition that reduced CO 2 , of which Boston fern was the most effective ( Table 2). While in the case of activated carbon and soil, the levels of CO 2 seemed to increase in the chamber. Table 1 shows that under dry conditions inside of the chamber, the selected soil adsorbed formaldehyde faster than the other substrates, while the performance of the dry expanded clay was the lowest. The wet soil and expanded clay performed better than the dry conditions tested. Furthermore, Table 1 shows that the selected plants together with the substrate did not perform as well as the wet substrates, but, in general, they performed better than the dry substrates with the exception of the dry soil. Regarding leaf area, the selected plants had similar characteristics in size and number of leaves, therefore, for every species one plant was selected and all its leaves were measured. Consequently, it was considered that the area of the other two plants of the selected species were in the same area range. In general, the peace lilies (approx. 0.14 m 2 ) had more leaf area than the Boston ferns (approx. 0.11 m 2 ). Table 3 presents the statistical analysis of the CADR of formaldehyde depletion caused by the selected growth media. It shows that soil has a better performance than the other samples. Regarding the data set of formaldehyde depletion, and once it was established the statistically signi cant differences between the means of the tested variables (P = 0.00) with ANOVA, the differences between the variables were analyzed in Tables 4 and 5. Table 4 presents the statistical difference among the variables. It shows that mainly wet soil has statistical differences with the other analyzed variables. Table 5 indicates three homogeneous subsets among the variables in terms of formaldehyde depletion. Within a subset there is no signi cance different while between subsets there is a signi cant difference. It is clear that Group 3 (wet soil, dry soil, wet expanded clay) is signi cantly different from Group 1 (wet activated carbon, dry activated carbon, dry expanded clay, peace lily, Boston fern).

Discussion
This study provides data for the characterization of the removal of formaldehyde by three different substrates and two different potted plants. Four series of zero measurements were executed to evaluate the setup. Two measurements of these series were executed with a plastic pot to evaluate the effect of this element in the depletion of the formaldehyde inside of the chamber. As expected, once the plastic pot was placed in the chamber the formaldehyde level was lower than the natural decay measured during the zero-measurement evaluation. This value was used to calculate the CADR for every test condition as shown in Tables 1 and 2 Previous studies suggest that the depletion of formaldehyde also occurs due to photosynthesis and metabolism of the plant at daytime (Teiri et al. 2018). A growing light was used during this test to ensure the optimal conditions of the plant.
Studies with potted plants in closed chambers continue to be useful for isolating factors that may enhance removal e ciency and contribute towards the improvement of plant-based systems (e.g. plant species and growth medium). Therefore, it is recommended to use the lessons learned from this study in creating a plant-assisted botanical puri er ("Biowalls" or active green walls), which

Depletion of CO 2
For the evaluation of the reduction of CO 2 levels inside of the chamber, it is important to mention that in general, plants regulate the internal CO 2 concentration through a partial stomatal closure when the CO 2 concentration is too elevated to maintain adequate internal CO 2 and optimize water use e ciency (Van de Geijn and Goudriaan 1996). Stomata are pores on leaf epidermis for both water and CO 2 uctuations that are controlled by two major factors: stomatal behaviour and density (Elliott-Kingston et al. 2016; Wang et al. 2007). The fast speed opening and closing of the stomata can save energy and increase photosynthesis and water use e ciency (Grantz and Assmann 1991). Taking this in account, Table 2 Figure 8 shows that in order to provide the optimal conditions for the plant it was necessary to inject CO 2 inside of the chamber because the concentration was too low for the plants (IPCC 2014; NASA 2019). In each test condition, activated carbon permanently released CO 2 inside of the chamber, which, possibly could be compensated by the uptake of CO 2 by the plants.

Plants vs. growth media
Formaldehyde and CO 2 were used as indicators of the effect of growth media and plants in reducing gaseous pollutants in a controlled environment. Table 1 shows that, in general, growth media were more effective in the depletion of formaldehyde inside of the chamber than the plants. Regarding CO 2 reduction inside of the chamber, as expected, Table 2 shows that plants were more effective than growth media: in most of the cases with only a growth medium present, CO 2 was released instead of reduced inside of the glass chamber. Figure 9 presents the different behaviours of the potted plants regarding these two elements. Even though the leave area of the Boston fern (approx. 0.11 m 2 ) was smaller than the peace lily (approx. 0.14 m 2 ), the Boston ferns reduced the concentration of CO 2 inside of the chamber faster than the peace lilies, which indicates that the stomatal conductance of the Boston fern was higher than the peace lily, opening the hypothesis about the uptake of more gaseous pollutants by the stomata.
Regarding the depletion of formaldehyde, Tables 4 and 5 show that wet soil, dry soil and expanded clay perform similarly and they are more effective than the other variables tested ( Table 3).
As mentioned before formaldehyde is soluble in water (WHO 2010). However, this study shows that high levels of humidity seemed to have no effect on the formaldehyde depletion inside of the chamber as in most of the test conditions the relative humidity level was above 90%. Nonetheless, it is important to mention that in the case of the plants, high humidity levels may affect the depletion of the CO 2 and the formaldehyde inside of the chamber due to the fact that plants close their stomata at high humidity levels  (Tables 1 and 2), therefore, it seemed to have no effect on the formaldehyde and CO 2 depletion, but in general in the presence of wet growth media the depletion of formaldehyde was faster. Regarding the effect of the growth media on the depletion of formaldehyde and CO 2 , it is important to mention that when the substrate (wet or dry) was tested without the plant, the whole surface of the substrate was exposed directly to formaldehyde and CO 2 . However, when the plants were included, the exposed surface of the selected substrate was reduced and the results show that the depletion also was lower, which indicates that the e cacy of the growth media, in some cases, was higher. This effect is produced by the microbial activity in the root zone, where bacteria absorb the gaseous pollutants   ). Therefore, it is necessary to have > 100 plants for every square meter of oor space in order to meet the standards without any additional ventilation system. Therefore, without any extra mechanical ventilation it is necessary an indoor forest to meet the minimum standards for ventilation rates in breathing zones just with plants, however, in real situations less plants will be required taking in account the size of the room and the ventilation system of every case.

Limitations
One of the limitations of this group of tests is the size of the chamber. Even though it has the requirements of a sealed glass container with the necessary inlets, for future research it is recommended to execute the tests in a bigger sealed glass container to prevent or reduce the stress of the plant, avoiding the closure of its stomata and reducing its metabolism.
As mentioned before, plant stress should be minimized, therefore, for future experiments the plant should be placed in the chamber one day prior the execution of the test together with the activated growing light.
A series of tests was performed to evaluate the effect of potted plants on reducing formaldehyde and CO 2 levels in a controlled glass chamber. The outcome of the tests showed some clear advantages and disadvantages of the different test conditions to consider for the design of an indoor plant-based system.
In terms of air 'cleaning' of formaldehyde, the measurements and analysis showed that soil, in general, was most effective in reducing formaldehyde concentrations in the chamber (~ 0.07-0.16 m 3 /h). Plants (~ 0.03 m 3 /h) were as effective as dry expanded clay (0.02-0.04 m 3 /h). Wet and dry soil, wet expanded clay and dry activated carbon performed better than the selected plants in formaldehyde depletion. In this study, it became clear that the substrate is an important ally in reducing gaseous pollutants, such as formaldehyde.
Regarding CO 2 reduction in the chamber, potted plants (Peace lilies − 0.01 m 3 /h) (Boston ferns 0.02-0.03 m 3 /h) were more effective than the other tests. Specially, Boston fern which has a higher stomatal conductance than the peace lily, indicating the possibility of allowing more gaseous pollutants to be absorbed in the long term.
Studies with potted plants in closed chambers showed to be useful for isolating factors that may enhance removal e ciency and contribute towards the improvement of plant-based systems (e.g. plant species and growth medium). However, the impact of one potted plant on the cleaning of the indoor air, was insigni cant. Therefore, several potted plants will be required to improve the IAQ taking in account the speci c characteristics of the place such as, size and the ventilation system.
It must be noted, however, that in this study the formaldehyde was introduced in a glass chamber in which the plant and its substrate were located, hereby surrounding the plant and its substrate with formaldehyde. In a 'normal' indoor environment, usually the source of formaldehyde may not be close to the plant system. For the plant-system to take-up the formaldehyde, the polluted air needs to be transported to the vicinity of the plant. This could be realized, for example, by an active plant-substrate system, in which the contaminated air is forced to go through the plant-leaves and through the substrate-roots. Further research with active plantbased systems on the depletion of formaldehyde and other pollutants, is required. Consent to Publish: We con rm that the manuscript has been submitted solely to this journal and is not published, in press, or submitted elsewhere.
Authors Contributions: All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Tatiana  Competing Interests: We con rm that we do not have any potential or perceived con icts of interest             Depletion of CO2 (mg/m3): for the three Boston Fern (NEPH_1, NEPH_2, and NEPH_3) and for the three Peace Lilies (SPA_1, SPA_2, and SPA_3). Data means ± SE, n=3.