Journal of Horticultural Science & Biotechnology (2002) 77 (1) 120129

Potted-plant/growth media interactions and capacities for removal of volatiles from indoor air
By R. A. WOOD, R. L. ORWELL, J. TARRAN, F. TORPY and M. BURCHETT* Plants and Environmental Quality Group, Faculty of Science, University of Technology, Sydney, Westbourne St., Gore Hill, NSW 2065, Australia (e-mail: Margaret.Burchett@uts.edu.au) (Accepted 20 October 2001)
SUMMARY

Results are presented of an investigation into the capacity of the indoor potted-plant/growth medium microcosm to remove air-borne volatile organic compounds (VOCs) which contaminate the indoor environment, using three plant species, Howea forsteriana (Becc. (Kentia palm), Spathiphyllum wallisii Schott. `Petite' (Peace Lily) and Dracaena deremensis Engl. `Janet Craig'. The selected VOCs were benzene and n-hexane, both common contaminants of indoor air. The ndings provide the rst comprehensive demonstration of the ability of the potted-plant system to act as an integrated biolter in removing these contaminants. Under the test conditions used, it was found that the microorganisms of the growth medium were the ``rapid-response'' agents of VOC removal, the role of the plants apparently being mainly in sustaining the root microorganisms. The use of potted-plants as a sustainable bioltration system to help improve indoor air quality can now be condently promoted. The results are a rst step towards developing varieties of plants and associated microora with enhanced air-cleaning capacities, while continuing to make an important contribution to the aesthetics and psychological comfort of the indoor environment. has become a major health consideration (Abbritti and Muzi, 1995; Krzyzanowski, 1995; American Lung Assoc., 2001). City air is always polluted, mainly from motor vehicles, and inside buildings further chemicals are added, mainly volatile organic compounds (VOCs) (WHO, 1989; Brown et al., 1994; Brown, 1997; Smith, 1997). The VOCs originate from both outdoor sources (mainly from fuel emissions, e.g. benzene), and indoor sources (e.g. from furnishings, machines, solvents, cleaning agents, clothes, cosmetics, e.g. n-hexane). The harmful effects of the chemical mixtures have been recognized as components of ``sick building syndrome'' or ``building-related illness'', particularly in air-conditioned buildings (Burge et al., 1987; Mendell and Smith, 1990; Carpenter, 1998; Brasche et al., 1999; Carrer et al., 1999). Over three hundred VOCs have been detected in indoor air, generally as complex mixtures; although each compound is likely to be in very low concentration, the cocktail can produce additive effects, with symptoms of headache, respiratory problems and loss of concentration (National Occupational Health and Safety Commission (Aust.), 1991; Wolkoff, 1995; Weschler and Shields, 1997). The use of plants to adorn and soften the indoor environment has a long history (Hammer, 1991; Hammer and Wood, 1999), and it is argued that this is an expression of our need to relate directly with nature, a need derived from our evolutionary roots (Kaplan and Kaplan, 1982, 1989; Kellert and Wilson, 1993; Lewis, 1996). The ndings presented here provide a new, very direct justication for their use, namely that they help improve the quality of the indoor atmosphere.
*Author for correspondence.
120

ndoor air in urban environments is in increasing need I of improvement, and since city dwellers often spend over 90% of their time indoors, the quality of indoor air

Plants, microorganisms and environmental cleansing

It is well established that ``outdoor'' plants can absorb many toxic compounds from air, water or soil and detoxify or metabolize them (Schulte-Hostede et al., 1987; Taylor et al., 1991; Sandermann, 1992; Foyer et al., 1994; Giese et al., 1994). Leaf absorption and metabolism of air-borne VOCs such as benzene and toluene have been reported in a number of species, although the results have mainly been derived from experiments in which the VOCs have been applied at much higher concentrations than are likely to be found in reality (Jen et al., 1995; Ugrekhelidze et al., 1997; Collins et al., 2000; Cape et al., 2000). Screening studies have also shown that a number of species of ``indoor'' potted-plants can reduce concentrations of VOCs, as well as dust and other air-borne pollutants (Wolverton et al., 1989; Wolverton Env. Serv., 1991; Wolverton and Wolverton, 1993; Lohr and Pearson-Mims, 1996; Coward et al., 1996). Giese et al. (1994) showed that both whole plants and isolated leaf cells of the spider plant (Chlorophytum comosum) and soybean (Glycine max) could absorb formaldehyde and metabolize it to carbon dioxide. Wolverton and Wolverton (1993) suggested that the microorganisms of the growth medium might be involved in the removal of the air-borne VOCs by indoor potted-plants. Microorganisms are well-known to be effective in the bioremediation of soil contamination with organics, e.g. from oil spills (e.g. Radwan et al., 1998; Siciliano and Germida, 1998a, b; 1999; Li et al., 2000). They have also been used for the treatment of hydrocarbon vapours in ``biolter reactors'', in which the contaminated air is passed through a damp, porous medium which supports active microorganisms (Hodge et al., 1991; Zhou et al., 1998; Yeom and Yoo, 1999). However, their possible role as a component of the

R. A. Wood, R. L. Orwell, J. Tarran, F . Torpy

indoor potted-plant system in removing VOCs has not previously been investigated. Our own earlier testchamber studies (Wood ., 1997; 1999a, b; 2000), using Becc. (Kentia palm) ( ), showed that potted specimens had the capacity to remove several times the maximum allowable Australian occupational exposure levels of benzene and -hexane, two common VOCs (National Occupational Health and Safety Commission, Australia, 1991). Our results also indicated clearly that it was the microorganisms of the potting mix that were the primary agents in accounting for the rates of removal observed. Similar responses were also found when the plants were transferred to hydroponics. The aims of the current study have been to compare the removal capacity of potted with two other internationally top-selling interior potted-plant species, Schott. `Petite' (Peace Lily) ( ), and Engl. `Janet Craig' ( ), and to explore the respective roles of the plants and the potting mix microorganisms in that process. To this end, the capacity for VOC removal was measured with potted specimens of each plant species, under light and dark regimes, and in both a standard potting mix and a standard hydroponic growth medium. Investigations into dose-response relationships, and the relative contributions to the removal process of the plant and the growth media microorganisms, were also investigated. This is the rst research undertaken to examine any of these matters systematically with respect to indoor potted-plant species. However, such information is essential to provide a scientic foundation for the horticultural breeding of indoor potted-plant varieties and improved growth media, to provide an enhanced capacity for VOC removal. The goal is the development of a new, effective, sustainable, exible, portable, bioltration system to help improve the quality of the indoor environment, while it also continues to beautify and soften our urban living/working spaces.
et al Howea forsteriana Arecaceae n H. forsteriana Spathiphyllum wallisii Araceae Dracaena deremensis Dracaenaceae Approach

and

M. Burchett

121
Dra-

Well established 12 month old potted plants of each species were used, all 0.30.4.m tall, in 150.mm pots. Tissue-cultured and cutting-grown were obtained from Wood's Nursery, Kenthurst, NSW, and seedlings from Mountain Range Nursery, Woonoona, NSW. Plants were in a standard, well aerated potting mix, composed of composted hardwood sawdust, composted bark nes, and coarse river sand (2:2:1) (bulk density ;0.6.g ml21; air-lled porosity ca. 30%), with Macracote ``Greenplus'' nine-month fertilizer (12:4.6:10.N:P:K, with trace elements) (Langley Chemicals, Welshpool, Western Australia). Plant morphological characteristics were measured for species comparisons, including total leaf area, using a leaf area meter (Licor LI-3000-A, Nebraska, USA); and fresh weight (fwt) and dry weight (dwt) of shoots, roots and potting mix, using a drying oven at 708C for 24.h.
Plant materials Spathiphyllum forsteriana caena H.

Where fresh sets of plants were to be tested in hydroponic solution, they were rst watered well and gently removed from the potting mix. Roots were then washed with several changes of sterilized (boiled) deionized water to remove visible remains of the potting mix, with no further effort made to sterilize the roots. Plants were then placed in a beaker with 300400.ml of hydroponic medium (4.g l21; `Aquasol'; Hortico, Laverton, Vic., Aust.), also prepared with sterilized, deionized water. The medium was aerated throughout each experiment. Plants were acclimatized to hydroponic conditions for at least 48.h prior to testing. Plantfree medium controls indicated no incidental microbial contamination from the air or apparatus.
Hydroponic conditions

MATERIALS AND METHODS Initial experiments used plants grown in a potting mix, and under continuous light, which is commonly found in public indoor environments (e.g. hotels, shopping malls, ofce blocks, hospitals, airports). The pot-plants were then tested under dark conditions, after which the plants were removed and the potting mix placed back in the test chambers, to establish specically the relative contributions of plant and potting mix to VOC removal rates. Other batches of plants were uprooted from their potting mix, transferred to hydroponic medium, and tested in light and dark conditions as before, and once again being removed to test the residual activity of the hydroponic medium. Since no previous studies into the role of indoor growth media micoorganisms in VOC removal have been carried out, the microbiological experiments reported here have been pilot studies only. Bacterial responses were chosen for investigation, since Prior to each experiment, the inner surfaces of the they were considered to be more likely to be more chambers were swabbed thoroughly with 90% ethanol to effective than the fungi, cyanobacteria or algae (Song clean and remove any traces of previously used test VOCs, and to reduce any microbial contamination. The ., 1986; Leahy and Colwell, 1990).
Test procedure and chemicals et.al

Four replicate ``Perspex'' static test chambers were used, 0.6.3.0.6.3.0.6.m (volume 0.216.m3, 216.l). The chambers had ``Perspex'' lids with stainless steel frames, sealed with adhesive foam rubber tape and held closed by six metal clips. Each chamber was equipped with rubber septa through which the VOC could be introduced and air samples withdrawn; a 0.5.m coil of copper tubing (i.d. 4.mm), through which water circulated from a thermostat bath at 23.6.0.18C; a 2.4.W fan to accelerate equilibration of the atmosphere; and a light box above (with air gap of 5.cm), tted with ve 18.W uorescence tubes designed for optimum plant growth (Wotan L 18/11 Maxilux 2Daylight, Ozram, Germany) (ca. 120.mmol quanta m 2 sec21). When necessary, chambers were darkened by switching off lights and covering with sheets of black plastic. For hydroponic experiments, individual aerators were added, which bubbled a continuous stream of chamber air through the liquid medium in which the plant roots were immersed.
Test chambers

122 walls were then wiped with dry tissues and given 20/30 min of fan-assisted ventilation to remove all traces of enthanol. Test plants (4 replicates, 1 per chamber) were well watered and allowed to drain on a wire rack for 1.h before starting each experiment. A pot plant was then placed in a plastic saucer in each chamber, the lid sealed, and the light box positioned. Benzene was Analar grade (BDH Chemicals Aust. Pty Ltd., Port Fairy, Vic.); -hexane was Mallinkrodt Nanograde (Rhone-Poulenc, Clayton South, Vic.). The required volume of the VOC was injected with a 50.ml syringe through the septum on to a suspended paper tissue to evaporate. For each VOC measurement, 1.0.ml of chamber air was withdrawn through the septum with a gas-tight syringe. Chambers were always sampled in duplicate (i.e. total of eight readings). VOC levels were measured using a gas chromatograph (GC) (Shimadzu GC-8A). GC was calibrated using standard gas samples. Initial doses were 25.ppm benzene (80.mg m23 at 1 atmos, 258C) and 100.ppm (353.mg m23) -hexane, being respectively ve and two times the Australian maximum allowable occupational exposure concentrations (National Occupational Health and Safety Commission, 1991). Chamber VOC concentrations equilibrated in about 1.5.h and lower limits of detection of the GC were found to the 0.2.ppm -hexane and 0.1.ppm benzene. Experimental samplings were carried out at hourly, several hourly or daily intervals as required. Additional injections of VOC were performed as needed for the particular experiment. In longer experiments, plants were watered weekly via a tube through the chamber septum and into the pot. ``Leak'' tests were carried out immediately before and/or after experiments to correct for chamber leakage or any VOC adsorption/absorption on test chamber apparatus. 21 for 25.ppm Losses were recorded of 2.73.8% d benzene and 1.63.0% d21 for 100.ppm -hexane, and results were corrected accordingly. No differences in loss rates were found between light and dark conditions, indicating that light-degradation of VOC was negligible. Leak tests with empty pots and containers, and with hydroponic aerators in operation, showed that VOC emissions from these articles were also neglible. A control test with unused potting mix was also made, using 25 ppm benzene as the VOC.
n n n n Experimental sequence

Potted-plants to remove volatiles

(d) A dark regime was now imposed, and additional top-up injections applied as needed, the responses again being followed at 24.h intervals. (e) A higher VOC dose was then applied, still in the dark, to investigate dose-response reactions. For this, the concentrations were increased from 25 to 50 ppm benzene, and from 100 to 150 ppm -hexane. (f) The plant was removed, and removal rates in the potting mix alone were measured, still in the dark. For this, the potting mix was returned to the chamber with a new, standard dose of the VOC (i.e. 25 ppm benzene or 100 ppm -hexane). The same experimental sequence was used for plants tested in the hydroponic medium.
n n

The sequence of the experimental steps employed with each of the three plant species, with each VOC, was usually as follows: (a) After an initial dose of VOC (25 ppm benzene or 100 ppm -hexane), under light conditions, samples were taken to determine initial rates of removal; (b) Sampling was continued over 2448.h or more, as required, to follow any induction of increased removal rates, which generally occurred (usually within 2.d for benzene, and 35.d for -hexane). (c) The effects on removal rates of one or more topping-up doses to the original concentration were then followed, over successive 24.h intervals.
n n

First, to determine changes in the numbers of culturable bacteria in the potting mix following exposure of the pot-plants to benzene, substrate samples (10.g) were taken from each of four pots, both before and after one of the experiments outlined above. From these, serial suspensions were made in 0.1% w/v sodium pyrophosphate solution, and bacteria enumerated on spread plates of tenth-strength tryptic soy agar (TSA, Oxoid). Control samples were taken from the potting mix of plants placed in chambers for the same length of time but without benzene. No-treatment controls were also used, i.e. plants kept on the laboratory bench in ambient light conditions for the period of the experiment. The plates were incubated at room temperature (23.6.1.58C) for 4.d. Secondly, to determine whether the culturable bacterial community derived from the potting mix was in fact capable of degrading 2benzene in the absence of the plants or substrate, 10 1 (w/w) potting mix suspensions were made, this time from plants that had been shown to remove benzene. The suspensions were used to inoculate tenth-strength tryptic soy broth (TSB) in quadruplicate. The suspension was added to autoclaved, Grade 3 vermiculite as a support medium (50.ml broth to 8.g vermiculite), to increase the effective surface area of the liquid phase of the broth. The samples were aseptically sealed in light-proof, air-tight jars, the lids of which were tted with silicone rubber septa. Jars containing a similar volume of potting mix as the broth/vermiculite suspension were used to compare the benzene-degrading efcacies of each system, and jars containing uninoculated media were used as controls to correct for background loss of benzene. In this experiment, 5 ppm benzene was used, the Australian occupational allowable maximum concentration (National Occupational Health and Safety Commission, 1991), to test whether induction would occur with this lower dose. Benzene was introduced into the jars and its concentration monitored by GC analysis of headspace gas samples taken by syringe through the septa. Treatments were incubated at room temperature (23.6.1.58C) with the sampling pattern as follows: Day 1: benzene added to jars; Day 2: benzene levels monitored and topped up as necessary; Day 3: benzene levels monitored; Day 4: benzene levels monitored and topped up again as necessary; Day 5: benzene levels monitored.
Microbiological tests Dracaena H. forsteriana

R. A. Wood, R. L. Orwell, J. Tarran, F . Torpy

and

M. Burchett

123
1460. 6 96 . 121. 6 8.2 17.6 6 1.0 14. 6 . 13.9 6 2.9 4.48 6 0.7 32. 6 . 3.9 6 . 452 . 6 18 . 0.75 6 0.05
Dracaena

Plant and potting mix characteristics of

Howea, Spathiphyllum 758. 6 89. 42.2 6 3.9 11.1 6 1.1 26. 6 . 12.4 6 1.3 3.34 6 0.2 26. 6 . 3.3 6 . 810 . 6 26. 1.35 6 0.07
Howea

and

Dracaena
Table I

. Data are means

Plant Species Leaf area (cm2) Shoots fwt (g) Shoots dwt (g) Shoot dwt/fwt (%) Roots fwt (g) Roots dwt (g) Root dwt/fwt (%) Shoot dwt/root dwt (ratio) Potting mix dwt (g) Potting mix vol (l)

weight

6 SE (n.=.4). fwt.=.fresh weight; dwt.=.oven-dried

878. 6 40. 85.9 6 5.4 11.3 6 0.5 13. 6 . 22.6 6 4.8 4.27 6 0.8 19. 6 . 2.6 6 . 352 . 6 14 . 0.60 6 0.04
P

Spathiphyllum

Data were corrected for losses of benzene related to the signicant reduction ( >0.05) in rates of removal. This experimental apparatus or broth media. result indicates that, at least with these species and under these test conditions, neither continued absorption into the leaves through the stomata, nor concurrent light Means and standard errors were calculated for all metabolism, is of immediate importance to rates of data. Student's paired t-test was used to assess different removal rates in the same plants at different stages of the experimental sequence, and Student's t-test for unpaired data was used to compare different plant species, or the same species in the different growth media. Single factor analyses of variance (Microsoft Excel 5.0) were used to analyse the microbiological data. In reporting all of the results, differences were regarded as signicant where <0.05. RESULTS Plant characteristics are presented (Table I) to provide size comparisons, and some are used later for alternative comparisons of VOC removal performance. The plants had almost twice the leaf area of the other two species. This difference is reected in the shoot weights as well (which is not surprising since in these species most shoot weight is leaf tissue). had the highest root fresh weight, although this was not reected in the dry weights, which were fairly similar among the species. This is because the roots had a higher water content (succulence) than the other species and hence a greater root volume (thus tting less potting mix per pot).
Data analysis P Plant and potting mix parameters Dracaena Spathiphyllum Spathiphyllum

Rates of VOC removal per chamber (i.e. per potplant) for each species, in either potting mix or hydroponics, are shown in Figures 1 and 2. It can be seen that there were differences among the species in patterns of response to each VOC, nevertheless there were strong and obvious similarities in response, which are considered rst. Following the initial dose of either VOC, with plants in the potting mix, removal rates were initially very slow (stage a), but they then generally accelerated markedly (stage b), i.e. an ``induction'' effect was observed. With benzene, for all species, the induction of more rapid removal rates occurred within 12.d, whereas with hexane 45.d were usually required. With repeated topping-up doses of VOC, the induced activity was either (stage c) maintained or, in some cases, further increased (i.e. a further induction response). Moving to dark conditions (stage d) produced no
General patterns of response to VOCs n

1 Concentrations of benzene (Bz.) in test chambers during experiments with potted (Kentia) (Kt.), (Sp.), and (Dc) plants respectively, in potting mix and in hydroponics, the step increments in concentration corresponding to injections of benzene over stages a to f of the experimental sequence (full description in text). Pmx. and Hyd. indicate experiments in potting mix and hydroponics respectively. L/D refers to transfer from light to dark conditions; PR.=.plant removed. Data are means 6SE (n.=.4).
Fig.

Howea

Spathiphyllum

Dracaena

124

Potted-plants to remove volatiles

2 Concentrations of -hexane (Hx.) in test chambers during experiments with potted (Kentia) (Kt.), (Sp.), and (Dc.) plants respectively, in potting mix and in hydroponics, the step increments in concentration corresponding to injections of -hexane, over stages a to f of the experimental sequence (full description in text). Pmx. and Hyd. indicate experiments in potting mix and hydroponics respectively. L/D refers to transfers from light to dark conditions; PR.=.plant removed. Data are means 6SE (n.=.4).
Fig.

Howea

Spathiphyllum

Dracaena

to 150 ppm. Work is continuing to establish the upper limits of removal capacities of this microcosm. To check more precisely the relative contributions of plant and potting mix to VOC removal rates, using two species, the plants were at this point removed altogether, and the potting mix returned to the (still dark) chambers with a new standard dose of VOC (stage f). The removal continued, although species differences in response were found. In the pots from the VOC continued to disappear at rates statistically no different from those obtained prior to the plant's removal, however, in those used with , the removal rates were signicantly reduced, although not eliminated. When the experiments were continued with topping-up VOC doses for a further 710.d, the activity in the pots from each species was maintained in every case. The activity of the growth medium was further examined, by testing unused potting mix (Figure 3; Table II). Again, signicant benzene removal rates were found. In this case, however, the initial activity was much slower than in the presence of the plants (cf Figures 1, 2), although it increased after 2.d, and showed a further increase with repeated doses of the VOC (in other words, induction also occurred with the unused potting mix). However, the maximum removal rates achieved were signicantly lower than in the presence of plants (2379% less) (Table II). Moreover, by days 8 and 9, the activity fell to about 47% what it had been in Stage 3 (Figure 4, Table II). The VOC removal activity in unused potting mix was not unexpected, in the light of the preceding experiments, since potting mixes contain a microbial complement. However, the lower rates and possible exhaustion of the system found in this trial suggest a need for the plant to establish and sustain the soil microora and their activity. On placing ``soil-less'' plants in hydroponics under conditions where transfer of soil particles was minimized (Figures 1, 2) substantial removal activity was recorded with exposure to benzene, although rates were slower than in potting mix. Removal of -hexane was very slow. Once again, the eventual removal of the plant did not eliminate removal activity. These results again point to the role of microorganisms in the removal process. However, they also indicate that a close relationship
Howea Dracaena n

disappearance of the VOC. This nding points strongly towards the primary role of the substrate microorganisms in the process under this set of ``indoor'' conditions. When a higher VOC dose was applied (50 ppm benzene and 150 ppm -hexane, respectively), with the chambers still in the dark, it can be seen in Figures 1 and 2 that removal rates usually increased further again (stage e). That is, the pot-plant microcosm could respond to, and cope with, the higher dose of either compound. In fact, over this range of concentrations the dose-response relationship appears to follow approximately rst-order kinetics with respect to VOC concentration. That is, the removal rate in most cases almost doubled when the benzene concentration was doubled, from 25 to 50 ppm, and increased by a factor of about 1.5 when the -hexane level was raised from 100
n n

3 Concentrations of VOC in test chambers with unused potting mix exposed to initial and topping-up doses of 25 ppm benzene; ie. following Stages 13 of the experimental sequence. Data are means 6SE (n.=.4).
Fig.

R. A. Wood, R. L. Orwell, J. Tarran, F . Torpy

and

M. Burchett

125
and Initial benzene

Benzene removed activity of virgin potting mix, compared with such activity in the presence of dose of 25 ppm. Data are means SE (n.=.4)

Stage in experimental Virgin potting mix Howea sequence (see text) control experiment Initial 4.6 6 0.4 20.3 6 1.7* Induction with single dose 24.5 6 2.3 51.2 6 7.3* `Steady state' 43.7 6 4.1 55.5 6 3.6 Days 8 and 9 20.4 6 1.4 n.a. Mass of dry potting mix per pot = (995 6 34) g. *Mean signicantly higher than for virgin potting mix control experiment (P n.a. Data not available.

Howea Spathiphyllum Dracaena. Rate of benzene removal (mg m3 d1 kg1 dry potting mix)
Table II

Spathiphyllum

34.5 6 4.0* 36.0 6 10. 171. 6 18* . n.a.

10.0 6 2.3 69.0 6 6.1* 194 .6 49* . n.a.

Dracaena

<0.05).

exists between plants and microorganisms in determin- removing a further full dose. By day 3, the vermiculite ing VOC-removing activity. culture had developed sufciently to remove virtually all of the benzene dose, showing no signicant difference in activity from the potting mix. This pattern was repeated day 4. These results show that this vermiculite/ Using potting mix samples from plants, no on bacterial can potentially account for the whole of signicant differences were found between total cultur- the VOCculture removal observed in the potted-plant system. able bacterial numbers, scored as colony-forming units Work is continuing on the characterization of the (cfu), following exposure to benzene (Table III). microbial species concerned. However, several types of bacteria were observed to either increase or decrease in numbers when the samples were exposed to benzene. In addition, several types of bacteria appeared in the cultures only after exposure to benzene. Thus, although no signicant numerical alterations occurred in the total bacterial community as a The unit of plant use in the horticultural industry is result of benzene exposure, changes occurred in com- generally based on pot size, which gives a readily munity structure, as increases and decreases of different understood representation of the size and appearance taxa. the plant specimen. It is therefore likely that some Results of the comparative assay between potting mix of ``per-plant'' basis will continue to be used in the samples and TSB/vermiculite bacterial cultures from industry, with to their use in VOC removal. , are presented in Figure 4. The potting mix For that reason,respect a comparative summary is presented in showed an immediate high level of benzene removal Table IV of the rates of removal of the two VOCs with capacity, removing virtually the entire dose in 24.h. This all three plant species, on the basis of both plant and was not surprising, as the samples had been induced to a potting mix parameters, at steady-state (i.e. fully more rapid benzene removal in the preceding experi- induced) removal rates (stage c of experimental ment in the plant test chambers. The derived bacterial sequence) (Table II). Among the plant species, there cultures, however, initially removed only a low, though are large differences, as measured by various parasignicant, amount of the benzene (ca. 20% after 1.d), with respect to the removal of any one VOC, being no doubt due to the undeveloped state of growth meters, which again to a direct involvement of the plant in of the bacterial community in the broth. After 2.d, the determining point the VOC capacity of the pottedculture had increased in capacity to remove about 60% plant system. There areremoval also substantial with of the benzene, with the potting mix again completely any one plant species with respect to its differences response to the two VOCs, which suggest that different sets of microorganisms may be involved for each compound. , (i.e. per chamber), for benzene removal, showed about a 50% higher rate than , which in turn had about a 50% higher rate than . However, for -hexane the order is reversed: the removal rate with was some 70% higher than with and with the removal rate was much lower than in the other two species. The same trends in ranking among species are found per (although differences between species on this basis were not necessarily of the same magnitude as per pot-plant). , (which is where most of the direct activity occurs) with benzene, and showed rates about 34 times higher than . 4 Comparison of removal of initial and topping-up doses of 5 ppm With -hexane, rates with were again benzene by potting mix (unbroken line) and TSB/vermiculite bacterial highest, those of were intermediate, and those of cultures developed from the potting mix (broken line). Data are means 6 SE (n.=.4). were lowest.
Effects of benzene on potting mix microorganisms Dracaena Plant-based comparisons of steady-state capacities for VOC removal Howea Per whole plant Dracaena Spathiphyllum Howea n Howea Spathiphyllum Dracaena unit root dry weight Per kg potting mix Spathiphyllum
Fig.

Dra-

caena

Howea

Spathiphyllum

Howea

Dracaena

126
Numbers of bacterial colony-forming units (CFU) g

Potted-plants to remove volatiles

Table III
dry weight potting mix, before and after ve-day treatments with initial and topping-up doses of 5.ppm benzene. Data are means

Treatment Before treatment After chamber treatment with benzene exposure After chamber treatment with no benzene No treatment (kept on lab. bench)

n 24 4 4 4

SE

CFU (106) g1 dry weight potting mix 25 6 71 60 6 31 54 6 10 50 6 15

, benzene removal rates were not signicantly different among the species, and there was only a minimal difference between and . With respect to -hexane, removal on the basis of both leaf area and shoot dry weight was fastest with , intermediate with and much slower with . In summary, on a leaf area basis, the steady-state removal rate responses were: for benzene in potting mix, there were no signicant differences among the species; in all other treatments, signicant and consistent differences were found, namely, for benzene in hydroponics, and for -hexane in either growth medium: > > . Leaf area as a basis of comparison has the justication that it is a standard unit for measuring plant physiological performance. In reference to VOC removal, it also provides a direct indication of the amount of foliage required in the indoor situation to achieve a given rate of removal (presumably by supporting its substrate microorganisms). Overall, the comparisons give a further indication of the importance of plant/microorganism interactions in determining VOC removal capacity.
Per unit leaf area dry weight Howea Spathiphyllum n Howea Spathiphyllum Dracaena

per g shoot

Howea

Spathiphyllum

Dracaena

DISCUSSION The results conrm, rst, that the VOC removal is clearly a biological response, not merely an adsorption/ absorption process, which would tend towards saturation. They also conrm that under these conditions, the microorganisms of the growth medium are the primary agents of rapid VOC removal, and that they can remain active for at least a week without the plant. Although previous studies on indoor plants have speculated that bacteria play a role in VOC removal (Wolverton and Wolverton, 1993), this is the rst denitive demonstration that substrate bacteria are ``rapid response'' agents of this phenomenon in the interior potted-plant microcosm.
Spathiphyllum
and

For benzene, the removal appears to be associated with microorganisms closely associated with the root system, which are hence persistent on transfer to hydroponics. On the other hand, with -hexane, the relevant microorganisms may not be so tightly bound to the roots, and so are more effectively reduced in numbers on washing and placement in hydroponic conditions. These aspects need further investigation, along with differences among different potting and hydroponic media. Bacteria are well known to have the capability to degrade aromatic pollutants (e.g. Leahy and Colwell, 1990; Hutchins, 1991; Edwards and GrbicGalic, 1992; Dia z and Prieto, 2000). The high level of efciency and dose-related removal capacity shown by the bacteria here have not previously been reported. The results are in line with applications of ``outdoor'' plant-and-soil relationships, where plants and/or their associated soil microbial populations are used in phytoremediation/bioremediation of soils contaminated with VOCs and related organics, for instance after oil spills (see, e.g. Radwan ., 1995, 1998; Gunther ., 1996; Siciliano and Germida, 1998a, b; 1999; Frick ., 1999; Li ., 2000; Nemergut, 2000). Although it was found in the current trials that removal of the plant often (though not always) caused little change in the rates of VOC removal, the question remains open as to the plants' reponses to the VOCs that would have been absorbed, mainly via gaseous diffusion through the stomata in the light, and more slowly, but in both light and dark, via the cuticle, by virtue of their lipid solubilities (Ugrekhelidze ., 1997). These compounds cause metabolic responses in plants, which could, in turn, possibly be chemically signalled to the substrate microorganisms via the roots, assisting in a heightened microbial metabolic response that is to some extent species-specic. It is known, for example, that root exudates have a profound effect on the soil microbial community, increasing the total numbers of microorganisms in the rhizosphere by orders of magnin et al et al et al et al et al

Table IV
Rates of VOC removal in ``steady state'' (stage c of experimental sequence), following induction and repeated doses of benzene or

Dracaena Species
Howea

-hexane, by

, in units based on several pot-plant characteristics. Initial and topping up benzene doses of 25 ppm and 100 ppm. Data are means

VOC & Medium Benzene potting mix

n

Spathiphyllum Dracaena H. forsteriana Spathiphyllum

-hexane potting mix 1Signicantly different from 2Signicantly different from

Dracaena Howea P

Spathiphyllum

( <0.05). ( <0.05).
P

Rates of VOC removal mg m3 d1 mg m3 d1 kg1 dry mg m3 d1 m2 potting mix per plant leaf area 40.8 6 2.31 56 6 3.6 537 6 69 60.2 6 5.8 171 6 181 686 6 73 88.2 6 22 195 6 49 606 6 155 306 6 341 378 6 43 4032 6 6451 509 6 1201,2 2040 6 4901,2 179 6 421,2 53 6 12 118 6 26 367 6 82

Howea

-hexane doses of

SE (n.=.4). dwt.=.oven-dried weight

mg m3 d1 g1 dwt shoot 3.7 6 0.4 1 5.4 6 0.56 5.0 6 1.3 27.6 6 4.1 15.9 6 3.81,2 3.0 6 0.7

mg m3 d1 g1 dwt root 12.2 6 1.0 14.1 6 3.1 19.7 6 5.7 92 6 12 42 6 131 1 11.9 6 3.2

R. A. Wood, R. L. Orwell, J. Tarran, F . Torpy

tude over those in the bulk soil (Hawes and Brigham, 1992; Brigham ., 1994). Plants are also known to expend considerable energy nourishing their rhizosphere microorganisms, sometimes secreting up to 45% of their net photosynthetic product from their roots to select and maintain their microorganism community (Kraffcyzk ., 1984; Schwab, 1998). These matters remain to be investigated in indoor plant species and their growth media, and work is continuing on these aspects. In summary, the results indicate that: The potted-plant, i.e. plant-growth-medium microcosm, is capable of reducing or eliminating indoor air concentrations of the two model VOCs. Under these growth conditions, it is normal microorganisms of the growth medium that are the primary ``rapid response'' agents of VOC removal. The system improves on exposure to the chemical, and maintains performance with repeated doses. These increases in rates following initial exposure indicate the induction of one or more biochemical pathways in the substrate bacteria (and possiblyalso the plant) to metabolize the chemical on a sustainable basis, and/or to changes in the microbial community in favour of those which can utilize VOCs. The system can deal with relatively high air-borne concentrations of the VOCs, the removal rate increasing linearly with dose across the range of concentrations used here. (These dose-response relationships are currently under further investigation.) The system can also remove very low residual VOC concentrations, since, if not topped up, concentrations were effectively reduced to zero (Figures 1, 2) The phenomenon appears to be general; similar responses can be expected with any plant species and standard growth medium, and testing of different types of both of these components of the microcosm are continuing. Comparisons among these plant species with each VOC point to different species having different communities and/or relationships with their root-zone microorganisms. Therefore, since indoor air contains a dynamic mixture of gaseous pollutants, it seems that a mixture of potted-plant species is likely to be the most
et al et al

and 127 effective in improving indoor air quality. This is in accord with common aesthetic and design principles used in interior plantscaping. Several European and North American enterprises are already developing the integrated use of plant materials and ``plant-ltration'' boxes, as part of building design or as later installations. In a parallel development, ``biolter reactors'', based upon air ows across compost, peat bed or activated carbon systems, are being designed as part of bioengineering solutions for the removal of air-borne VOCs from industrial processes (Bibeau, 2000; Marek ., 2000; Mohseni and Allen, 2000). The indoor potted-plant/growth medium microcosm appears to offer several advantages over the engineering-based biolter systems, with the rhizospheres providing an effective, more sustainable environment for the adsorption and microbial transformation of organic compounds. Oxygen transfer to rootzone microbial communities is enhanced and gas-phase diffusion in the soil facilitated by the presence of the plants, providing appropriate conditions for compounds to be degraded aerobically. The potted-plant system could avoid some of the limitations of the articial biolter systems, such as absorption saturation, compaction of the medium, and in some cases very slow induction rates (measured in months) (Zhou ., 1998; Yeom and Yoo, 1999). The horticultural development of ornamental and crop species and their microora is never-ending. It should be possible to develop improved indoor plant varieties and optimum growth media/ microorganism complements, to enhanced the capacity for cleaning indoor air, while continuing to beautify the indoor environment. We thank Horticulture Australia (previously the Horticultural Research and Development Corporation, Aust.) and the Nursery Industry Association of NSW (via the Horticultural Stock and Nurseries Act) for funding a large part of this project. Thanks also to the Interior Plantscapers Association of NSW, Mountain Range Nursery and Woods Nursery for their assistance, and at UTS, Sian Munro, and Laboratory Managers Narelle Richardson, Bill Booth and James Phillips, and other members of the Faculty of Science, who have contributed to the progress of this project.
M. Burchett

et al

et al

Abbritti, M. C.

American

nothing else matters

and (1995). Indoor air quality and health effects in ofce buildings. In: . (Maroni, M., Ed.). University of Milano and International Centre for Pesticide Safety, Milan, Italy, Vol. 18595. (2001). . Air Quality. www.lungusa.org/air/.
Muzi, G.

REFERENCES

Brigham, L. A., Nicoll, M. S.

Proceedings of Healthy Buildings

`95, An International Conference on Healthy Buildings in a Mild Climate

Brown, S. K., Sim, M. R., Abramson, M. J.

1,

Lung

Association

When you can't breathe,

Brown, S. K.

Brasche, S., Bullinger, M., Gebhardt, H., Herzog, V., Hornung,

Bibeau, L., Kiared, K., Brzezinski, R., Viel, G.

and (1999). Factors determining different symptom patterns of sick building syndrome results from a multivariate analysis. In: , Edinburgh, UK, 4027. and (2000). Treatment of air polluted with xylenes using a biolter reactor. , 37793.
P., Kruppa, B., Meyer, E., Moreld, M., Schwab, R., Mackensen, S., Winkens, A. Bischof, W.

Burge, S., Hedge, A., Wilson, S., Harris, B. J.

Proceedings

Cape, J. N., Binnie, J., Mackie, N.

of Indoor Air `99, The 8th International Conference on Indoor Air Quality and Climate
5,

Heitz, M.

Carpenter, D. O.

Air, Water and Soil Pollution

118,

and (1994). Plant genes controlling the release of root exudates. , 6171. and (1994). Concentrations of volatile organic compounds in indoor air A review. , 12334. (1997). Volatile organic compounds in indoor air: sources and control. , January/February, 103. and (1987). Sick building syndrome: a study of 4373 ofce workers. , 493504. and (2000). Uptake of volatile organic compounds by grass. , 2000, Brighton, UK. (1998). Human health effects of environmental pollutants: New insights. , 24558.
Stephenson, M. B.

Biotechnology

and Plant Protection

4,

Gray, C. N.

Indoor Air

4,

Chemistry in Australia

Robertson, A.

Annals of Occupational Hygiene

31,

Skiba, U. M.

Proceedings of Third

SETAC World Congress

Environmental Monitoring and Assess-

ment

53,

128

Potted-plants to remove volatiles

Carrer, P., Alcini, D., Cavallo, D., Visigalli, F., Bollini, D. Maroni, M. (1999). Home and workplace complaints

and and

Marek, J., Paca, J.

and

Gerrard, A. M.

(2000). Dynamic responses

of biolters to changes in the operating conditions in the process of removing toluene and xylene from air. Acta Biotechnologica,
20, 1729.

symptoms in ofce workers and correlation with indoor air pollution. In: Proceedings of the 8th International Conference on
Indoor Air Quality and Climate, Edinburgh, UK, 1, 12934.

Mendell, M. J.

and

Smith, A. H.

(1990). Consistent pattern of

Collins, C. D., Bell, J. N. B.

and

Crews, C.

(2000). Benzene

elevated symptoms in air-conditioned ofce buildings: A reanalysis of epidemiological studies. American Journal of Public
Health, 80, 11939.

accumulation in horticultural crops. Chemosphere, 40, 10914.

Coward, M., Ross, D., Cowards, S., Cayless, S.

and

Raw, G.

(1996). Pilot study to assess the impact of green plants on NO 2

levels in homes. Building Research Establishment Note N154/ 96, Watford, UK.

Mohseni, M.
and

and

Allen, D. G.
volatile

(2000). Bioltration of hydrophilic organic compounds.

Chemical

hydrophobic

Engneering Science, 55, 154558.

z, E. Dia

and

Prieto, M. A.
of

(2000). Bacterial promoters triggering pollutants.

Current Opinion in

biodegradation

aromatic

National Occupational Health and Safety Commission (Australia) (1991). Exposure standards for atmospheric contaminants in the occupational environment . Australian Government

Biotechnology, 11, 46775.

Edwards, E. A.

and

Grbic-Galic, D.

(1992). Complete miner-

Printing Service, Canberra, Australia.

alization of benzene by aquifer microorganisms under strictly anaerobic conditions. Applied Environmental Microbiology, 58, 26636.

Nemergut, D. R., Wunch, K. G., Johnson, R. M. and Bennett, J. W. (2000). Benzo(a)pyrene removal by Marasmiellus troyanus
in soil microcosms.
Journal of Industrial Microbiology and Biotechnology, 25, 1169.

Frick, C. M., Farrell, R. E.

of phytoremediation as

and

Germida, J. J.

(1999). Assessment

an

in-situ technique for cleaning oil-

Radwan, S., Sorkhoh, N.

and

El-Nemr, I.

(1995). Oil biodegrada-

contaminated sites. Petroleum Technology Alliance of Canada,

tion around roots. Nature, UK, 376, 302

Calgary.

Foyer, C. H., Descourvieres, P.

Protection mechanism against studied oxygen in

and

Kunert, K. J.
An important
Plant,

(1994). defence
and

Radwan, S. S., Al-Awadhi, H., Sorkhoh, N. A. and El-Nemr, I. M. (1998). Rhizospheric hydrocarbon-utilizing microorganisms
as potential contributors to phytoremediation for the oily Kuwaiti desert. Microbiology Research, 153, 24751.

radicals:

transgenic

plants.

Cell

Environment, 17, 50723.

Sanderman, H. J. R.
17, 824.

(1992). Plant metabolism of Xenobiotics, TIBS

Giese, M., Bauer-Doranth, U., Langerbartels, C. and Sandermann, J. R. H. (1994). Detoxication of formaldehyde by the
spider plant
max

(Chlorophytum L.)

comosum

L.)

and
Plant

by

soybean

Schulte-Hostede, S., Darrall, N. M., Blank, L. W. and Wellburn, A. R. (Eds) (1987). Air Pollution and plant metabolism,
Elsevier, NY, USA.

(Glycine

cell-suspension

cultures.

Physiology,

104, 13019.

Schwab, A. P., Al-Assi, A. A.

Adsorption of naphthalene

and

Banks, M. K.
plant roots.

(1998).
of

Gunther, T., Dornberger, U.

sphere, 33, 20315.

and

Fritsche, W.

onto

Journal

(1996). Effects of

Environmental Quality, 27, 2204.

ryegrass on biodegradation of hydrocarbons in soil. Chemo-

Siciliano, S. D.

and

Germida, J. J.
biochemical and

(1998a).

Mechanisms

of

Hammer, N.
USA.

phytoremediation:

ecological

interactions

(1991). Interior landscape design. McGraw Hill, NY,

between plants and bacteria. Environmental Review, 6, 6579.

Siciliano, S. D.
and

and

Germida, J. J.
ester proles promote

(1998b). Biolog analysis and indicate that pseudomonad the root-

Hammer, N. R.
American

Wood, R. A.
portfolio of

(1999). Interior landscapes: An

green environments.

fatty

acid

methyl that

design

Rockport

inoculants

phytoremediation

alter

Publishers, Massachusetts, USA.

associated microbial community of Bromus biebersteinii. Soil (1992). Impact of root border
Biology and Biochemistry, 30, 171723.

Hawes, M. C.
11948.

and

Brigham, L. A.

cells on microbial populations. Advances in Plant Pathology, 8,

Siciliano, S. D.

and

Germida, J. J.

(1999). Enhanced phytoreme-

diation of chlorobenzoates in rhizosphere soil. Soil Biology and

and

Hodge, D. S., Medina, V. F., Islander, R. L.

(1991). Treatment of hydrocarbon fuel
Environmental Technology, 12, 65562.

Devinny, J. S.
in biolters.

Biochemistry, 31, 299305.

vapors

Smith, L. M.

(1997). Strategy for nationwide reduction of total

exposure to indoor and outdoor polluants. In: Proceedings of

Healthy Buildings/IAQ `97, Global Issues and Regional Solutions. (Woods, J. E., Grimsrud, D. T. and Boschi, N., Eds).

Hutchins, S. R.
carbons by nitrous

(1991). Biodegradation of monoaromatic hydro-

aquifer microorganisms using oxygen, nitrate or as the terminal electron acceptor.

Applied

oxide

Washington DC, USA Vol. 1, 45761.

Environmental Microbiology, 57, 24037.

Song, H. G., Pederson, T. A.

and

and

Bartha, R.

(1986). Hydrocarbon

Jen, M. S., Hoylman, A. M., Edwards, N. T.

14

Walton, B. T.

mineralization in soil: Relative bacterial and fungal contribution. Soil Biology and Biochemistry, 18, 10911.

(1995). Experimental method to measure gaseous uptake of C-toluene by foliage. Environmental and Experimental Bot-

Taylor, G. E., Pitelka, L. F. Ugrekhelidze, D., Korte, F.

and

Clegg, M. T.

(1991). Ecological

any, 35, 38998.

genetics and air pollution, Springer Verlag, NY, USA.

Kaplan, R. Kaplan, R.

and

Kaplan, S.

(1982). Humanscape: Environments for

and

Kvestitadze, G.

(1997). Uptake

people. Ulrich's Books, Ann Arbor, Michigan, USA.

and transformation of benzene and toluene in plant tissues.

Ecotoxicology and Environmental Safety, 37, 249.

and

Kaplan, S.
perspective.

(1989). The experience of nature; A Cambridge University Press, Cam-

psychological

Weschler, C. J.
among indoor 348795.

and air

Shields, H. C.
pollutants.

(1997). Potential reactions

Environment, 31,

bridge, UK.

Atmospheric

Kellert, S. R.

and

Wilson, E.O.
and

(Eds.) (1993). The biophilia

hypothesis. Island Press, Washington, DC, USA.

Wolkoff, P.

(1995).

Volatile

organic

compounds

Sources,

Kraffczyk, I., Trolldenier, G.

root exudates of maize:

Beringer, H.
of

measurements, emissions and the impact on indoor air quality.

Indoor Air. Suppl. No. 3.

(1984). Soluble supply and

Inuence

potassium

rhizosphere microorganisms. Soil Biology and Biochemistry, 16, 31522.

World Health Organisation (Who) (1989).

Organic pollutants, EURO reports and

Indoor air quality: 111.

studies.

WHO,

Leahy, J. G.
30515.

and

Colwell, R. R.
in

Copenhagen, Denmark, 170. (1990). Microbial degradation of

Microbiology Review, 54,

Wolverton, B. C., Johnson, A.

landscape plants for indoor

and
air

Bounds, K.
pollution

(1989). Interior Final

hydrocarbons

the environment.

abatement,

Report, NASA, Stennis Space Centre, MS, USA. (1996).

Green nature/Human nature. University of

Lewis, C. A.

Wolverton Environmental Services Inc.

(1991).

Removal

of

Illinois Press, Chicago, USA.

formaldehyde from sealed experimental chambers, by Azalea,

Li, G., Huang, A., Lerner, D. N.

Enrichment of degrading microbes

and and

Zhang, X.

(2000). of

Poinsettia and Dieffenbachia. Research Report No. WES/100/ 01-91/005, USA. Plants for Clean Air Council, Mitchellville, MD,

bioremediation

petrochemical contaminants in polluted soil. Water Research,

34, 384553.

Wolverton, B. C. Pearson-Mims, C. H.
(1996). Particulate matter microorganisms ammonia from

and

Wolverton, J. D.
Removal indoor of

(1993). Plants and soil xylene

of

Lohr, V. I.

and

formaldehyde,

and
the

accumulation on horizontal surfaces in interiors: Inuence of foliage plants. Atmospheric Environment, 30, 25658.

the

environment.

Journal

Mississippi Academy of Sciences, 38 (2), 1115.

R. A. Wood, R. L. Orwell, J. Tarran, F . Torpy

and

M. Burchett

129
and

Wood, R. A., Orwell, R.

and Burchett, M. D. (1997). Rates of absorption of VOCs by commonly used indoor plants. In:
Proceedings of healthy buildings/IAQ `97, global issues and regional solutions.

Wood, R. A., Orwell, R. L., Burchett, M. D., Tarran, J. Brown, S. K.

(2000). Absorption of organic compounds in

Proceedings of buildings 2000. 6th international healthy

(Woods, J. E., Grimsrud, D. T. and Boschi, N., Eds). Washington, D.C., Vol. 2, 59 64. Wood, R. A., Orwell, R. L. and Burchett, M. D. (1999a). Living plants to improve indoor air quality. In: Towards a new
millennium in people-plant relationships. Contributions from

indoor air by commonly used indoor plants. In:

healthy buildings conference. 2,

(Seppanen, O. and Sateri, J., Eds). Espoo, Finland,

12530. and
Yoo, Y. J.

Yeom, S. H.

(1999). Removal of benzene in a hybrid 2818. and

Yang, S. T.

(Burchett, M. D., Tarran, J. and Wood, R., Eds). Sydney, Australia, 11522. Wood, R. A., Orwell, R. L. and Burchett, M. D. (1999b). Study of absorption of VOCs by commonly used indoor plants. In:
Proceedings of Indoor Air `99, 8th International Conference on Indoor Air Quality and Climate.

the International People Plant Symposium.

reactor.

Process Biochemistry, 34,

Zhou, Q., Huang, Y. L., Tseng, D. H., Smim, H.

(1998). A trickling brous bed reactor for bioltration of benzene in air.

nology, 73, Journal of Chemical Technology and Biotech-

Edinburgh, UK, Vol. 2, 6904.

35968.