Article pubs.acs.org/est
Removal of Formaldehyde from Air Using Functionalized Silica Supports Abdunaser M. Ewlad-Ahmed,† Michael A. Morris,‡ Siddharth V. Patwardhan,*,§ and Lorraine T. Gibson*,† †
Department of Pure and Applied Chemistry, WestCHEM, University of Strathclyde, 295 Cathedral Street, Glasgow, U.K., G1 1XL Department of Chemistry, University College Cork, Cork, Ireland § Department of Chemical and Process Engineering, University of Strathclyde, 75 Montrose Street, Glasgow, U.K., G1 1XJ ‡
ABSTRACT: This paper demonstrates the use of functionalized meso-silica materials (MCM-41 or SBA-15) as adsorbents for formaldehyde (H2CO) vapor from contaminated air. Additionally new green nanosilica (GNs) materials were prepared via a bioinspired synthesis route and were assessed for removal of H2CO from contaminated indoor air. These exciting new materials were prepared via rapid, 15 min, environmentally friendly synthesis routes avoiding any secondary pollution. They provided an excellent platform for functionalization and extraction of H2CO demonstrating similar performance to the conventional meso-silica materials. To the authors’ knowledge this is the first reported practical application of this material type. Prior to trapping, all materials were functionalized with amino-propyl groups which led to chemisorption of H2CO; removing it permanently from air. No retention of H2CO was achieved with nonfunctionalized material and it was observed that best extraction performance required a dynamic adsorption setup when compared to passive application. These results demonstrate the first application of GNs as potential adsorbents and functionalized meso-silica for use in remediation of air pollution in indoor air.
■
sprayed onto glass furnace fiber filters.14 However this was only reported to be a short-term solution. Passive removal of H2CO using porous adsorbents of metals oxides has been achieved in living environments, but only for short times before concentrations elevated to initial levels.13−15 Tenax TA has been reported as the best evaluated sorbent for monitoring general volatile organic compounds (VOCs) in indoor air due to its high thermal stability (up to 350 °C).16 Carbon sorbents have also been used to quantitatively trap a wide range of VOCs in indoor air. However both Tenax TA and activated carbon are more suited to the C3 to C12 range of VOCs,17,18 not for H2CO. In addition carbon sorbents are known to be susceptible to pore clogging and are difficult to regenerate.19 Recently mesoporous silica, such as SBA-15 or MCM-41, has been used to remove VOCs from air.20 However, most studies focused on the calculation of adsorption capacity experiments using adsorption isotherms or thermal methods of analyses. Few studies have examined their practical application for VOCs,19,21−25 and not much is known about the ability of mesoporous materials to remove H2CO vapor. Of the few studies that have been published26−28 limited data was given to support the sampling and extraction methods used. A new and exciting group of silica materials that are gaining attention due to the ease of application, environmental friendly synthesis routes,29 and rapid preparation times are bioinspired silica
INTRODUCTION Formaldehyde (H2CO) vapor is given increased attention as an indoor air pollutant due to the potential impact on human health and comfort when present at high concentration (900 μg m−3).1 As H2CO is classified as a Group 1 probable human carcinogen1,2 it is arguably the most widely studied aldehyde. A long-term exposure to lower levels of H2CO vapor has been shown to cause irritation of the eyes and nose3 and H2CO vapor has been related to an increased risk of respiratory disease.4 Despite being toxic and allergenic, H2CO is still widely used in industrial processes to produce building materials and wood products; primarily due to its low cost and high reactivity. Its incorporation into wood products such as plywood, particleboard, chipboard, and fiberboard (as urea-formaldehyde or phenol-formaldehyde resin binders) provides sources of emission for H2CO vapor in indoor air.5−7 In addition, other emissive products include varnishes, paints,8,9 and combustion sources such as wood-burning and tobacco smoking,10 which can generate significant indoor H2CO levels. As the main source of H2CO is from materials and products typically used for construction or furniture, its concentration indoors is usually 2−10 times higher than that found outdoors.11,12 When H2CO levels increase in the indoor air environment, steps need to be taken to protect occupants from ill health effects. Eriksson et al.13 attempted to remove gaseous H2CO from homes using potassium permanganate, activated carbon, and aluminum oxide. Although successful, the adsorption capacities of the materials were limited due to their unspecific reactions with other air pollutants. H2CO levels have been reduced by polymeric amines and polyethylenehydrazine © 2012 American Chemical Society
Received: Revised: Accepted: Published: 13354
September 27, 2012 November 16, 2012 November 26, 2012 November 26, 2012 dx.doi.org/10.1021/es303886q | Environ. Sci. Technol. 2012, 46, 13354−13360
Environmental Science & Technology
Article
was achieved by placing the sample below the straight through beam and raising it until the beam of intensity was attenuated by 50%. Ideally, when the samples are rotated both clockwise and anticlockwise, the straight through beam intensity should be lowered equally. These experimental conditions are required if the low angle data are to be quantified properly. Nitrogen adsorption−desorption isotherms were measured using a Micromeritics ASAP 2420 surface area and porosity analyzer after sample degassing for 6 h at 200 °C. The Brunauer− Emmett−Teller (BET) surface areas were calculated using experimental points at a relative pressure (P/P0) of 0.05−0.25. The total pore volume was calculated from the N2 amount adsorbed at the P/P0 of 0.99, and the average pore size distribution of the materials was calculated using the Barret− Joyner−Halenda (BJH) model from a 30-point BET surface area plot. MWD-MCM-41 and SBA-15 exhibited a type IV adsorption isotherm typical of mesoporous solids. Desorption isotherms were used to calculate the pore diameters. Infrared spectra were obtained in the 4000−400 cm−1 region with a resolution of 4 cm−1, by accumulating 32 scans using an attenuated total reflectance (Diamond with ZnSe lens reflection ATR plate) Fourier transform infrared (ABB, MB3000 FTIR) spectrometer. Elemental analysis was carried out using an Exeter Analytical CE440 elemental function to provide the functional group (carbon and nitrogen) and hydrogen content of the studied materials. 2.3. Amine Functionalization of Materials. Functionalization of sorbents (MWD-MCM-41, SBA-15, or GN) was achieved by postsynthesis grafting as reported by Lim et al.34 In this work, 3-aminopropyltrimethoxysilane was used as the functional organic amine group to produce NH2-MWD-MCM41, NH2-SBA-15, or NH2-GN. Five g of each material was preheated to 120 °C for 2 h and immersed in 50 cm3 of dry toluene (Sigma-Aldrich) in a 250 cm3 flask before addition of 10 cm3 of 3-aminopropyltrimethoxysilane (97%, Fluka). Finally, the mixture was refluxed for 4 h. Products were filtered, washed with 100 cm3 of ethanol (Sigma-Aldrich), and dried in an oven at 80 °C for 1 h. 2.4. Preparation of Tubes Loaded with Mesoporous Sorbent. Each sorbent was packed into a stainless steel or glass tube (90 mm × 5 mm i.d.) for VOC or H2CO collection, respectively. Prior to loading, sorbents were pressed into a disc (using a KBr press), crushed, sieved (to 60−80 mesh), and conditioned at 120 °C for 2 h. Each tube contained 100 mg of sorbent mixed with 1.5 g of glass beads (750−1000 μm) held between two plugs of glass wool. The loaded sampling tubes were then connected to a N2 cylinder via a mass flow controller, and N2 was used to bed the adsorbent into the sampling tube using a flow rate of 100 cm3 min−1 for 5 min. 2.5. Dynamic Atmospheric Chamber Setup. To generate a stream of air contaminated with H2CO a dynamic chamber reported by Idris et al.25 was used which contained temperature-controlled ovens, an air compressor, two air streams, and a sampling chamber (20 dm3). An air compressor (Jun Air 122-50) was used to provide two separate air streams. The first flow was passed through a humidification system (controlled by a mass flow controller at 200 cm3 min−1) to provide the sampling chamber with high relative humidity (HRH) air at 80% RH. The second flow stream entered the temperature-controlled oven (Kin-Tek Laboratories Inc. model 491 M-B), and passed over a permeation device (containing solid paraformaldehyde) at an accurately measured flow rate of 200 cm3 min−1. At 50 °C and 200 cm3 min−1, an average (n =
materials. In recent years there has been an exciting growth in biologically inspired silica synthesis research,30 with in vitro experiments leading to the development of bioinspired green nanomaterials (GNs). With such materials it is possible to control GN properties such as surface area and particle size.31 The focus of most publications so far has been directed toward different synthesis routes for GNs and to the authors' knowledge there are no published reports on practical applications of GNs let alone H2CO removal from a contaminated air stream. This research article provides systematic data on the results of mesoporous silica materials (MCM-41 or SBA-15) for the removal of H2CO vapor from air. Furthermore this paper reports the first practical application of GNs in the removal of H2CO and it is suggested that GNs provide excellent adsorption platforms. Adsorption was assessed in dynamic and passive mode to assess different remediation techniques for contaminated indoor air environments, or for use in museum cabinets where H2CO vapor is measured at high concentration. Finally, it is recognized that H2CO will never exist in indoor air on its own and so the adsorption performance of each material was examined in the presence of other selected volatile organic compounds (VOCs).
2. EXPERIMENTAL SECTION 2.1. Synthesis of MCM-41, SBA-15, and GNs. MCM-41 was prepared by microwave digestion (MWD) using the method reported by Idris et al.;32 this material will be referred to as MWD-MCM-41 and it will be shown below that the method used provides materials that differ in their structure compared to conventional MCM-41 prepared by calcination. SBA-15 was prepared by the method reported by Idris et al.25The bioinspired silica, GN, was synthesized using slight modification to a method reported previously31,33 and described briefly below. Sodium metasilicate (Na2SiO3·5H2O, Fisher Scientific) was used as a silica source at 30 mM, whereas tetraethylenepentamine (TEPA, C8H23N5, ACROS Organics) was used as a bioinspired additive at a Si:N molar ratio of 1:1. To provide solution A, 6.5 g of sodium metasilicate was dissolved in 500 cm3 of deionized water in a 1000 cm3 polyethylene bottle. A second solution (B) was prepared using1.14 g of TEPA dissolved in 400 cm3 of deionized water in a 500 cm3 polyethylene bottle. Then 65 cm3 of 1 M hydrochloric acid (HCl, Fisher Scientific) was added to solution A to produce an acidified solution (solution C) which was shaken for 1 min before adding solution B. The pH of the final mixture was immediately adjusted to pH 7 ± 0.1 by adding a few drops of 1 M HCl, and rapid precipitation of silica was visible within 1−2 min. The resultant as-synthesized precipitate was collected after 15 min (hereafter referred to as “as-GN”) by filtration, washing three times with deionized water, and drying in air at room temperature before calcination at 500 °C for 5 h, to produce GN. 2.2. Characterization and Analysis. X-ray diffraction (XRD) profiles were recorded on a Philips X’Pert Diffractometer, equipped with a Cu Kα radiation source and accelerator detector. Incident and reflected Stoller slits of 0.2° were used with a programmable divergent slit (a constant 10 mm sample footprint). A sample block was placed directly above the center of the sample at about 0.3 mm from the substrate plane. This reduced low angle scatter allowing more accurate low angle operation. Care was also taken to place the sample surface at the axis of rotation of the goniometer. This 13355
dx.doi.org/10.1021/es303886q | Environ. Sci. Technol. 2012, 46, 13354−13360
Environmental Science & Technology
Article
4) emission rate of 541 ng min−1 was calculated for H2CO with a relative standard deviation (RSD) value of 4.9% over a period of 28 days. The theoretical concentration of H2CO generated in the atmospheric chamber was 1.4 mg m−3; the chamber was equilibrated for 48 h before sampling. 2.6. Determination and Detection of H2CO Vapor or VOCs in Air. An air stream (SKC pumps were used to remove air at a flow rate of at 100 cm3 min−1) was pulled from the contaminated chamber for 60 min and sampled using Sep-Pak Plus C18 cartridges (Waters) impregnated with 0.8 cm3 of a 2,4-dinitrophenylhydrazine (2,4-DNPH) trapping solution. The trapping solution (10 mM 2,4-DNPH) was prepared by dissolving approximately 100 mg of doubly recrystallized 2,4DNPH (Aldrich, 97%) in a solution that contained 49.5 cm3 of acetonitrile (Rathburn) and 0.5 cm3 of orthophosphoric acid (BDH Laboratory supplier, 85%). The solution was slightly heated (approximately 40 °C) to promote dissolution of 2,4DNPH. H2CO (F) vapor trapped as formaldehyde-dinitrophenylhydrazone (F-DNPH) was eluted from cartridges using 3 cm3 of acetonitrile (Fisher Chemical). The extracted derivative was filtered (anotop IC 20 μm filters, Whatman) prior to determination of F-DNPH by a Varian 2510 liquid chromatograph (LC) using a 10 cm C8 Hypersil column (3 mm, 5 μm). The eluent used contained water/acetonitrile mix (70:30, % v/ v) at 1.5 cm3 min−1 and the LC was coupled to a Varian 2550 UV−vis detector set to 360 nm with a detector range of 0.32. The H2CO masses (wtnosorb) trapped by the DNPH sampling cartridges were quantitatively determined by external calibration using F-DNPH (Supelco, 99.95%) standard solutions in acetonitrile. A typical calibration result of concentration (0−60 μg) versus peak area had a regression line of y = 162865x + 27841 with a correlation coefficient of r2 = 0.998. To assess the efficiency of adsorbents for H2CO extraction, sorbent tubes were placed immediately before the hydrazine trapping cartridge. Air was drawn through the sorbent tubes and cartridge at a rate of 100 cm3 min−1 for 60 min and the mass (wtsorb) trapped was determined as given above. The mass losses (difference between wtnosorb and wtsorb) measured by hydrazine sampling cartridges were attributed to the adsorbent which had been placed in the contaminated air stream; all experiments were repeated. In control experiments (n = 4) an average mass (wtnosorb) of 8.12 μg was trapped by hydrazine cartridges. Sorbents were placed in-line and experiments were repeated (n = 3). Sorbents were removed from the line and the wtnosorb value was determined. Sorbent efficiencies were calculated as shown in the equation below: Extraction Efficiency =
and DCB were calculated as 13643, 7221, 4939, and 3130 ng min−1 over a 28 day period, giving theoretical VOC concentrations of 34.05, 18.02, 12.38, and 7.81 mg m−3, respectively. All pollutant sources were replenished after a 28 day period. 2.7. Passive Extraction of H2CO Vapor Using Silica Sorbents. Scavenging efficiencies were also determined in passive extraction mode. H2CO vapor at 1.4 mg m−3 was generated inside an environmental chamber. Prior to sorbent addition 500 cm3 air was removed from the chamber and the mass of H2CO present in the air sample was measured at 1.2 μg. Sorbents (100 mg) were placed inside the chamber and left to passively adsorb H2CO vapor. Subsequent decreases in H2CO concentration, which were attributed to sorbent extraction, were determined by analyses of further 500 cm3 aliquots of air. Concentrations were recalculated after each aliquot of air was removed from the chamber to correct for dilution. Concentration measurements were taken at 5, 10, 15, 20, 25, 30, and 35 min after sorbent addition.
3. RESULTS AND DISCUSSION 3.1. Adsorbents Characterization. XRD patterns of MWD-MCM-41, SBA-15, and GN prior to functionalization are shown in Figure 1. The diffraction patterns of MWD-
Figure 1. SAS-XRD patterns of MWD-MCM-41, SBA-15, and GN. Dashed lines identify peaks of interest (see text for details).
MCM-41 and SBA-15 demonstrate the typical (1 0 0) diffraction peak at a 2θ° angle of 2.17 and 0.925 with an intensity of approximately 6000 and 16 700, respectively; with weaker reflections for SBA-15 assignable to (1 1 0) and (2 0 0). Interestingly the XRD data confirmed that the pore structure in MWD-MCM-41 was retained even after the harsh microwave digestion procedure used. The GN XRD pattern indicated that this new bioinspired material was amorphous with no measurable crystalline phases. Nitrogen adsorption isotherms provided further data on the textural properties of the prepared materials (data are summarized in Table 1 and Figure 2). Surface areas of 760 and 644 m2 g−1 were measured for MWD-MCM-41 and SBA15, which is in agreement with data commonly reported in the literature (for example 813 m2 g−1,32 and 678 m2 g−1,35 for MCM-41 and SBA-15, respectively). A lower surface area of 268 m2 g−1 was measured for GN, which was also comparable with values reported in the literature.33 The BJH pore size distribution plot for the amorphous GN was unusual and did not permit calculation of an average pore size, yet the pore volume was calculated to be 0.25 cm3 g−1; indeed the pore system for this material needs further examination to be explained in detail. Typical type IV isotherms were measured for MWD-MCM41 and SBA-15 confirming their mesoporosity, while the isotherm for GN was of type II indicating a material with
Wt sorb × 100 Wt nosorb
To examine the H2CO extraction performance of each sorbent in the presence of other air pollutants, a second dynamic atmospheric chamber was used to generate air contaminated with toluene, ethylbenzene (EB), cumene, and dichlorobenzene (DCB). Four GC autosampler vials (with silicon seal tops) were used to store 1 cm3 each of toluene (Fisher Scientific), EB (Sigma-Aldrich), cumene (ACROS Organics), and DCB (Aldrich Chemical). The GC vials were pierced with a GC syringe needle (18 gauge) to provide a controlled permeation source, and the pierced vials were placed in the temperature-controlled oven. A constant rate of air, 200 cm3 min−1, passed over the VOCs sources held at 25 °C. Under these conditions, the emission rates of toluene, EB, cumene, 13356
dx.doi.org/10.1021/es303886q | Environ. Sci. Technol. 2012, 46, 13354−13360
Environmental Science & Technology
Article
with the low surface area (and as yet indeterminable pore system) the material was successfully grafted with amine groups with greater surface coverage (L0 = 1.69) than was achieved for SBA-15. 3.2. Assessment of Sorbents for H2CO Extraction in Dynamic Mode. Before determination of sorbent extraction efficiencies, H2CO masses collected by 2,4-DNPH cartridges were determined using a flow rate of 100 cm3 min−1 and sampling times of 10, 20, 30, 40, 50, or 60 min. Using the measured H2CO emission rate, theoretical H2CO masses trapped by cartridges were calculated and compared to measured data. The results obtained demonstrated that the atmospheric chamber provided accurate concentrations of H2CO vapor and that the masses trapped increased linearly with increased sampling time (y = 34245x + 62882, r2 = 0.9866). To assess sorbent extraction efficiencies, a sampling time of 60 min was chosen. Repeated sampling of the chamber with the chosen conditions provided an average mass of 8.12 μg with an RSD value of 4.6%. 3.3. Sorbents Extraction Efficiencies When Used in Active Mode. With sorbents in-line it was shown that H2CO breakthrough (defined as 5% of the expected mass collected onto the hydrazine cartridge without sorbent present) occurred rapidly for nonfunctionalized mesoporous sorbents MWDMCM-41, SBA-15, or GN with breakthrough values of 31, 33 or 24%, respectively (see Figure 4). Interestingly the mesomaterials with significantly higher surface areas performed no better than GN. In stark contrast when materials were grafted to incorporate amine functionality sorbent breakthrough was not observed after the 60 min sampling time (6 dm3 of contaminated air was sampled at 1.4 mg m−3). The sampling volume of contaminated air was increased to 12, 18, or 24 dm3; still no breakthrough was observed indicating that 100 mg of NH2- sorbents could be used to clean H2CO from 24 dm3 of air even if present at elevated concentration (1.4 mg m−3). Again it is significant to note that the NH2-GN material performed as well as the aminated MCM-41 or SBA-15 despite having a surface area that was approximately 3 times lower. To test the adsorbate−adsorbent interaction and ensure H 2 CO was permanently removed from the air (by chemisorption), all sorbents were back flushed with air for 60 min at 100 cm3 min−1 (see Figure 5). NH2-MWD-MCM-41, NH2-SBA-15, or NH2-GN loaded with H2CO bound the target strongly through chemisorption, resulting in its permanent removal from air. Unmodified MWD-MCM-41 and SBA-15 materials released 45 and 65%, respectively, of the small masses of H2CO originally trapped. 3.4. Air Polluted with H2CO and VOCs. To prove the potential utility of the NH2-sorbents as H2CO environmental scavengers the efficacy of the materials was assessed for H2CO
Table 1. Physical Characterization Data for MWD-MCM-41, SBA-15, and GN sample MWD-MCM41 SBA-15 GN
BET surface area (m2 g−1)a
pore size (nm)b
pore volume (cm3 g−1)c
760
6.74
0.99
644 268
6.93 N/Ad
0.82 0.25
a
Calculated by BET model from sorption data in relative pressure range from 0.05- to 0.25. bCalculated by BJH model from the adsorption branches of isotherm. cCalculated from N2 amount adsorbed at a relative pressure P/P0 of 0.99. dN/A: no clear size seen, only broad line, see Figure 2.
nonporous or macroporous surface area with narrow slit-like pores (H4). The average pore size for MWD-MCM-41 was 6.74 nm, which was significantly higher than conventionally prepared materials at 2.53 nm reported by Idris et al.32 It was surmised that the high pressures induced in the microwave resulted in pore wall tear-down leading to a MCM-41 material with slightly lower surface area but higher average pore size and a wider pore distribution range. Such properties have not been evident in MCM-41 materials without alteration of the chemistry, for example by the addition of swelling agents which have been known to produce materials with an increased pore size;36 but notably, these materials still possessed a narrow pore size distribution. The combination of large average pore size and wide pore size distribution is a unique feature of the MCM-41 framework and occurs only after MWD. The ATR-FTIR data (Figure 3) confirmed the formation of a silica framework as evident from peaks i−iii at ∼1100 cm−1 (νas Si−O−Si), 950 cm−1 (νs Si−OH) and 800 cm−1 (νs Si−O−Si); Figure 3a. Further, the appearance of peaks at ∼1560 cm−1 and 1480−1490 cm−1 (v and vi in Figure 3b) represent νs N−H and νs C−N, respectively, and confirm successful amine-functionalization. A band measured between 1635 and 1645 cm−1 (band iv in Figure 3b) was assigned to a water bending mode and it was present in all samples before and after functionalization. The spectroscopic results were supported by elemental analysis (see Table 2). The percentage of nitrogen measured was used to estimate the degree of surface functionalization (L0) using the equation reported by Idris et al.32 The calculated L0 values (Table 2) demonstrated that a higher degree of functionalization (L0 = 1.99) was achieved for MWD-MCM-41 compared to SBA-15 (L0 = 1.58), despite the similar pore size and surface area. Hence the difference can only be attributed to the wide pore size distribution present in MWD-MCM-41 making a more successful platform for chelate co-ordination. The result obtained for GN was significant as it demonstrated that even
Figure 2. Nitrogen adsorption isotherms (a) and BJH pore size distribution patterns (b) of MWD-MCM-41, SBA-15, and GN. 13357
dx.doi.org/10.1021/es303886q | Environ. Sci. Technol. 2012, 46, 13354−13360
Environmental Science & Technology
Article
Figure 3. FT-IR spectra of MWD-MCM-41 (1), NH2MWD-MCM-41 (2), SBA-15 (3), NH2-SBA-15 (4), GN (5), and NH2-GN (6) showing silica formation (a) and amine functionalization (b). Important peaks are highlighted and are as follows: (i) ∼1040 and 1120 cm−1 νas Si−O−Si; (ii) ∼950 cm−1 νs Si−OH; (iii) 800 cm−1 νs Si−O−Si; (iv) 1635−1645 cm−1 νs bending vibration of adsorbed water; (v) ∼1560 cm−1 νs N−H; (vi) 1480− 1490 cm−1 νs C−N.
presence of VOCs in the flowing stream was shown to have no competitive detrimental effect on the masses of H2CO trapped by NH2-sorbents (Figure 6). To further validate this result
Table 2. Elemental Analysis Data for Silica Samples sample
%C
%H
%N
MWD-MCM-41 NH2-MWD-MCM-41 SBA-15 NH2-SBA-15 GN NH2-GN
trace/nil 8.2 trace/nil 6.9 trace/nil 5.82
0.52 2.32 trace/nil 1.66 0.36 1.11
trace/nil 2.79 trace/nil 2.21 trace/nil 2.36
L0 (mmol/g)a 1.99 1.58 1.69
a
Functionalization degree (L0 millimoles of nitrogen per gram of functionalized silica)
Figure 6. Effect of VOC on H2CO removal. Left axis shows selective removal of H2CO (% breakthrough) from a mixture of H2CO and VOC. Right axis shows displacement of preadsorbed H2CO by VOC (% H2CO recovered). Error bars indicate the % relative standard deviation value (n = 3).
pollutant masses passing through all sorbent materials were increased by a factor of 3 (sampling time was increased to 180 min); no H2CO or VOC breakthrough was observed for NH2based sorbents. The competitive nature of NH2-sorbents for H2CO in the presence of VOCs was also assessed by first loading sampling tubes with a known mass of H2CO (approximately 8 μg) and passing contaminated air (containing 34.05, 18.02, 12.38, or 7.81, mg m−3 of toluene, EB, cumene, or DBC, respectively) through the sorbent tube. Little to no displacement (
