Gas-Phase Formaldehyde Adsorption Isotherm Studies on Activated

ARTICLE pubs.acs.org/est

Gas-Phase Formaldehyde Adsorption Isotherm Studies on Activated Carbon: Correlations of Adsorption Capacity to Surface Functional Group Density Ellison M. Carter,† Lynn E. Katz,*,† Gerald E. Speitel, Jr.,† and David Ramirez‡ †

Department of Civil, Architectural, and Environmental Engineering, University of Texas at Austin, Cockrell School of Engineering, 1 University Station C1786, Austin, Texas 78712-0284, United States ‡ Department of Environmental Engineering, Texas A & M University at Kingsville, Dotterweich College of Engineering, 700 University Boulevard MSC 213, Kingsville, Texas 78363, United States

bS Supporting Information ABSTRACT: Formaldehyde (HCHO) adsorption isotherms were developed for the first time on three activated carbons representing one activated carbon fiber (ACF) cloth, one all-purpose granular activated carbon (GAC), and one GAC commercially promoted for gas-phase HCHO removal. The three activated carbons were evaluated for HCHO removal in the low-ppmv range and for water vapor adsorption from relative pressures of 0.1
0.9 at 26 °C where, according to the IUPAC isotherm classification system, the adsorption isotherms observed exhibited Type V behavior. A Type V adsorption isotherm model recently proposed by Qi and LeVan (Q
L) was selected to model the observed adsorption behavior because it reduces to a finite, nonzero limit at low partial pressures and it describes the entire range of adsorption considered in this study. The Q
L model was applied to a polar organic adsorbate to fit HCHO adsorption isotherms for the three activated carbons. The physical and chemical characteristics of the activated carbon surfaces were characterized using nitrogen adsorption isotherms, X-ray photoelectron spectroscopy (XPS), and Boehm titrations. At low concentrations, HCHO adsorption capacity was most strongly related to the density of basic surface functional groups (SFGs), while water vapor adsorption was most strongly influenced by the density of acidic SFGs.

’ INTRODUCTION AND BACKGROUND Poor indoor air quality in the workplace, public buildings, and residential dwellings has the potential to adversely impact human health. Volatile organic compounds (VOCs) are common contributors to indoor air pollution, and among them formaldehyde (HCHO) is one of the most studied indoor environmental pollutants as a consequence of its ubiquity indoors and its established human health significance. In addition to being classified as a known human carcinogen,1 HCHO is a sensory and respiratory irritant, and prolonged HCHO exposure has been associated with reduced pulmonary function and asthma,2 especially in children.3 Characteristic residential sources of HCHO include emissions from building materials, such as composite wood products, insulation, carpet and laminate flooring, and consumer products, such as disinfectants, personal care products, fabrics, and furniture. Also, HCHO may be present as a result of incomplete combustion associated with gas stoves or wood-burning fireplaces and as a byproduct of ozone reactions with volatile unsaturated hydrocarbons.4,5 The primary pathway for human exposure to HCHO is inhalation in residential and occupational settings. In the United States, HCHO concentrations in residential, site-built dwellings r 2011 American Chemical Society

average 20
50 ppbv and range as high as several hundred ppbv.6,7 These values exceed the chronic reference exposure limit (16 ppbv) for HCHO set by the National Institute for Occupational Safety and Health.8 In manufactured (prefabricated) housing or temporary housing units, such as those provided by the United States Federal Emergency Management Agency to families who were displaced by hurricanes Katrina and Rita in 2005, baseline sampling of HCHO concentrations by the Agency for Toxic Substances and Disease Registry averaged 1000 ppbv and ranged from 100—3066 ppbv.9 In occupational environments, over two million US workers are exposed to HCHO concentrations that can range from 100 ppbv to 10s of ppmv, depending on the industry.1 As a result of continued awareness of and concern about the health impacts of chronic exposure to HCHO, abundant information is available regarding HCHO indoor sources, concentrations, and exposures. However, few Received: December 21, 2010 Accepted: June 10, 2011 Revised: May 30, 2011 Published: July 07, 2011 6498

dx.doi.org/10.1021/es104286d | Environ. Sci. Technol. 2011, 45, 6498–6503

Environmental Science & Technology studies have evaluated the use of control technologies for the removal of HCHO in indoor environments. Air quality control technologies regularly use microporous activated carbon adsorbents to remove VOCs from contaminated gas streams. In addition to specific surface area, pore size distribution, and surface functional groups (SFGs) characteristic of a given adsorbent, adsorbate chemical properties, such as polarity, strongly influence the adsorption process. In a recent study, Yao et al.10 focused on predicting the adsorption behavior of 18 select VOCs to activated carbon fiber (ACF) cloth using the Dubinin
Radushkevich (D
R) and Freundlich models. Although Yao et al. included HCHO among their 18 primary VOCs for evaluation, no assessment of its adsorption behavior was made citing that the D
R model is not appropriate for polar compounds with dipole moments greater than 2 D (μHCHO = 2.33 D). Only a limited number of investigations have been conducted to evaluate gas-phase HCHO adsorption on activated carbon.11,12 Furthermore, these studies have only looked at single HCHO concentrations, ranging from 2.3 to 74 000 ppmv. Therefore, this work represents the first adsorption isotherm studies of HCHO on activated carbons.

’ MATERIALS AND METHODS Activated Carbons. The three commercially available activated carbons used in this research included two granular activated carbons (GAC1 and GACF) derived from bituminous coal and one ACF cloth woven from rayon-derived fibers. Upon receipt of samples from the manufacturer (Calgon Carbon Corporation, Pittsburgh, PA), all three activated carbons were dried at 105 °C for 24 h and stored in a desiccator under vacuum until testing. GAC1 (BPL 6  16 mesh) is an all-purpose GAC. GACF (Formasorb 4  8 mesh) is marketed for gas-phase removal of HCHO. ACF (ZoreflexWoven ACC, FM10) is a multiuse woven ACF. Physical and Chemical Adsorbent Surface Characterization. Surface area, total pore volume, and micropore volume of activated carbon samples were characterized, as received, by nitrogen adsorption isotherms at 77 K (ASAP 2010, Micromeritics, Atlanta, GA) applying Micromeritics’ Density Function Theory (DFT) software assuming the original DFT model with slit shape pore geometry. Surface chemical composition was investigated and quantified using X-ray photoelectron spectroscopy (XPS) (AXIS Ultra DLD Kratos Analytical, Spring Valley, NY) and Boehm titration techniques. Details of sample preparation and the procedures for each of these analyses are described in the Supporting Information (S1
3). Adsorption Isotherms. Adsorption isotherms were measured using a combined gas generation system and a gravimetric balance (SGA-100 Symmetrical Vapor Sorption Gravimetric Analyzer, VTI Corporation/TA Instruments, Hialeah, FL). Experiments involved passing a stream of gas with a known adsorbate concentration at a flow rate of 500 mL/min over an adsorbent sample suspended from the gravimetric balance. Initial adsorbent sample mass varied from 11 to 22 mg of activated carbon. Samples were individually loaded on weigh pans in the aluminum block sample chamber of the sorption analyzer and dried at 90 °C for at least 1 h to desorb any volatile contaminants that may have adsorbed during handling. Samples were subsequently cooled to 26 °C, at which time the initial sample mass was recorded. A constant temperature bath maintained the sample at 26 °C for the remainder of the experimental run.

ARTICLE

Throughout analysis, adsorption equilibrium was defined on a gravimetric basis as less than 0.001% weight change within a span of 5 min. Sample mass and gas stream temperature were monitored during adsorption using a Cahn D-200 microbalance with a sensitivity of 0.1 μg and a 100 Ohm Resistance Platinum Thermometer, respectively. Sample mass increase was assumed to be due to adsorption of the adsorbate, and thus, adsorption capacity was determined as a change in sample mass from the initial mass divided by the initial sample mass. HCHO concentrations ranged from 0 to 32.5 ppmv, supplied by a gas tank containing a stock concentration of HCHO of 36.5 ppmv in a balance of nitrogen (Air Liquide America, Scott Specialty Gas, Plumsteadville, PA). The formaldehyde gas tank was a specialty order from Scott Specialty Gas of Air Liquide America Specialty Gases LLC, which was certified to contain gasphase formaldehyde at the concentration prepared, 36.5 ppmv, for one year. The tank was completely used up within five months of receiving it, well within the recommended time limit. Single component isotherm studies of water vapor were also conducted, and water vapor relative pressures ranged from 0.1 to 0.9 at 26 °C, recorded as relative humidity (RH). RH was determined using a chilled-mirror dew point analyzer (DewMaster II, EdgeTech Corporation, Marlborough, MA) after passing a stream of dry ultrahigh purity (UHP) nitrogen gas through a canister housed within the sorption analyzer that was filled via syringe with deionized water. RH increments were attained by combining controlled flow of the humidified gas stream and dry UHP nitrogen. Flow rates of the adsorbate stream, the humidified stream, and dry UHP N2 stream were controlled by mass flow controllers (GFC-17A, Aalborg, Orangeburg, NY) via the Flow System Software (VTI Corporation/TA Instruments, Hialeah, FL). Equilibrium Isotherm Modeling. HCHO adsorption isotherms have not been previously investigated; therefore, no model has yet been proposed to fit and describe its adsorption to activated carbon. Moderately hydrophobic organic compounds tend to exhibit adsorption behavior on activated carbon classified as Type I behavior according to the IUPAC isotherm classification system, characterized by an adsorption curve concave to the abscissa, rising steeply at low partial pressures and approaching a limiting value as saturation pressure is reached.13 Adsorption behavior of this type is often described by the Langmuir and Freundlich models. However, in these studies HCHO exhibited a sigmoidal adsorption curve characteristic of Type V isotherms. Type V isotherms develop for adsorbates whose adsorbate
adsorbent interactions are weak relative to adsorbate
adsorbate interactions such that adsorption is low at low partial pressures but increases at higher partial pressures once capillary condensation takes place.13 Adsorption at low partial pressures may be enhanced on porous adsorbents and/or in the presence of SFGs that interact favorably with the adsorbate. Though Type V isotherms are uncommon, many mathematical models have been proposed to describe the most wellknown adsorbate to exemplify this adsorption behavior, water vapor. Two recent reviews14,15 detailed a large set of semiempirical and theoretical models, including the Dubinin
Radushkevich (D
R), Dubinin
Serpinsky (D
S), and Do
Do (D
D) families of models, in addition to several others, including the Qi
LeVan (Q
L), Qi
Hay
Rood (QHR), Mahle (M), and Talu
Meunier (T
M) models that have been more recently proposed. The D
R type models are not appropriate for this work because, as stated previously, they cannot be reliably applied 6499

dx.doi.org/10.1021/es104286d |Environ. Sci. Technol. 2011, 45, 6498–6503

Environmental Science & Technology

ARTICLE

Figure 1. HCHO adsorption isotherms for ACF (]), GAC1 (0), and GACF (4). Solid lines are fitted adsorption isotherms using the Q
L model.

to adsorbates with dipole moments greater than 2 D. The D
S type models are also inadequate because they do not provide a good fit of the adsorption isotherm over the full range of partial pressures investigated. The D
D models were not used in this analysis because the assumption that the adsorbate forms surface clusters could not be verified for HCHO. Ultimately, the Q
L model16 was selected to model the HCHO and water vapor adsorption behavior observed in this study because it met three important criteria: (1) reduction to a finite, nonzero limit at low partial pressures, (2) description of the adsorption isotherm over the entire range of adsorption considered, and (3) successful fit of both HCHO and water vapor adsorption isotherms. The Q
L model describes adsorption according the following expression, explicit in pressure: p¼

n ξ0 + ξ1 n + ξ2 n2 + ξ3 n3

ð1Þ

where p is the partial pressure, n is the surface loading, and the ξi parameters are unique to an adsorbent
adsorbate pair and reflect the influence of reaction equilibrium interactions between the adsorbate and surface adsorption sites, oxygen complexation, and pore structure. In particular, ξ0 represents the adsorbate
adsorbent partition coefficient at low enough relative pressures to be within the Henry’s Law region.

’ RESULTS AND DISCUSSION Adsorption Isotherm Results. HCHO adsorption isotherms for the three selected activated carbons are presented in Figure 1. The Type V adsorption behavior exhibited by HCHO was unusual for gas-phase organic compounds on activated carbon. Unlike moderately hydrophobic organic compounds that have been tested previously on activated carbon and display Type I adsorption behavior, HCHO is a very polar organic compound and should not be expected to behave in the same way. Comparison of the different activated carbons suggested that the relative affinity of the HCHO for each adsorbent was concentration dependent. GACF exceeded the adsorption capacity of ACF and GAC1 for HCHO concentrations below 7.3 ppmv. Between 7.3 ppmv and 14.6 ppmv, two crossover points were observed. First, above 7.3 ppmv, ACF adsorption capacity

Figure 2. Water vapor adsorption isotherms for ACF (]), GAC1 (0), and GACF (4). Solid lines are fitted adsorption isotherms using the Q
L model.

for HCHO exceeded that of GACF. Subsequently, above 14.6 ppmv, GAC1 adsorption capacity also exceeded that of GACF. Water vapor adsorption isotherms for the three activated carbons are presented in Figure 2. As anticipated, water vapor exhibited Type V adsorption behavior on each of these three activated carbons, and the maximum adsorption capacities for GAC1 and ACF were similar to those reported elsewhere.14,17 Comparison of the different activated carbons suggested that the relative affinity for water vapor for each adsorbent was dependent on relative pressure. Below P/P0 = 0.4, ACF and GACF exhibited similar adsorption capacities, both exceeding the performance of GAC1. Similar to the trends observed for HCHO adsorption capacity, two crossover points arise in Figure 2 between P/P0 = 0.4 and P/P0 = 0.7. First, above P/P0 = 0.4, ACF adsorption capacity for water vapor exceeded that of GACF. Above P/P0 = 0.6, GAC1 adsorption capacity exceeded GACF adsorption capacity as well. With respect to maximum adsorption capacity, the performance of these three activated carbon types varied from 282.76 to 422.90 mg/g, suggesting differences in pore volume or accessibility among the activated carbon surfaces. Adsorption Isotherm Modeling. The Q
L model was used to fit the sigmoidal HCHO and water vapor adsorption isotherms using four parameters. Model parameters were determined by minimization of the sum of the squared difference between the experimental and the fitted relative pressures. The parameters used to fit adsorption data to the Q
L isotherm model and the residual sum of squares (RSS) are summarized for each adsorbate
adsorbent pair in Table 1. For comparison, next to the H2O adsorption—GAC1 column, ξi values determined by Qi and LeVan16 are listed for the same activated carbon, labeled BPL. A complete summary of experimental and modeled adsorption data for adsorbate
adsorbent pairs can be found in Supporting Information (S4, Tables S4-1, S4-2). According to Qi and LeVan, the first parameter of their model, ξ0, should reflect the affinity of the adsorbent surface for the adsorbate. Thus, increasing adsorption capacity for either HCHO or water vapor on activated carbon should be associated with increasing values of the respective ξ0 parameter. This expectation is substantiated through analysis of the relationship 6500

dx.doi.org/10.1021/es104286d |Environ. Sci. Technol. 2011, 45, 6498–6503

Environmental Science & Technology

ARTICLE

Table 1. Qi
LeVan Adsorption Isotherm Model Parameters for H2O and HCHO H2O adsorption ξn units [mol/kg-kPa]

ACF

GACF

Qi et al. (2005) GAC1

HCHO Adsorption

BPL

ACF

GACF

GAC1

ξ0

1.4034

4.6862

1.2499

1.780

0.39676

0.57783

0.075603

ξ1

2.9113

1.2295

1.7171

0.137


0.066525


0.10537


0.031122

ξ2


0.079951

0.38009

0.017685

0.046

0.029568

0.042734

0.015191

ξ3


2.8730  10
5


0.036187


2.8185  10
3


0.002


1.7444  10
3


3.6020  10
3


1.2651  10
3

RSS

0.0147

0.0153

0.0153

69.12

0.3688

69.97

Figure 3. Evaluation of relationship between HCHO adsorption at 3 ppmv (0), 7 ppmv (4), and 15 ppmv (]) and the corresponding Q
L model parameter, ξ0. Correlation coefficients are reported beside each trend line.

between water vapor and HCHO adsorption capacity and these adsorbates’ respective ξ0 parameters. Figure 3 illustrates this relationship for HCHO. Furthermore, the relationship between adsorption capacity and ξ0 is concentration-dependent. High values of ξ0-H2O were positively correlated with high water vapor adsorption capacity at low relative pressures (P/P0 < 0.6), but this association was not evident at higher relative pressures (P/P0 > 0.6). Similarly for HCHO adsorption (Figure 3), high values of ξ0-HCHO were associated with high adsorption capacity at low HCHO concentrations ([HCHO] < 15 ppmv) but not at high HCHO concentrations ([HCHO] > 15 ppmv). Surface Characterization of Activated Carbon. To determine whether relationships between physical/chemical characteristics of the different activated carbons and HCHO adsorption/water vapor adsorption existed, correlations between surface chemical and physical data and adsorption capacities of these two adsorbates were investigated. Chemical and physical characteristics are summarized in Table 2. Nitrogen adsorption isotherms for all three activated carbon samples exhibited characteristic Type I behavior, as classified by the IUPAC classification system for adsorption isotherms. The results reported for specific surface area, total pore volume, and micropore volume indicated that these three activated carbons share similar physical characteristics. Likewise, the minimum pore diameters for ACF, GAC1, and GACF were 0.53, 0.50, and 0.53 nm, respectively. Therefore, the assumption that size exclusion would have a negligible impact on the adsorption of either HCHO or water vapor is supported. The specific surface area of ACF and GAC1

agreed reasonably well with reported specific surface area measurements for the same materials by Sullivan et al.14 The surface area of GACF had not been reported previously, but its specific surface area and pore volume is anticipated to be similar to GAC1 because these two carbons are derived from the same material, bituminous coal, and undergo the same initial activation processes. XPS analysis indicated that, in addition to carbon and oxygen, the activated carbons contained trace amounts of nitrogen (N), and GACF contained minimal potassium (K) and iodide (I). After carbon, oxygen was the most abundant element in all three activated carbon samples, and the following trend in oxygen content was observed: GACF > ACF > GAC1. XPS results for ACF indicated oxygen content to be 6.7% (atomic), which is slightly higher oxygen content than the 5.4% reported previously14 for this activated carbon but not at all unreasonable for activated carbon, in general, and may be attributable to manufacturing variation. In addition to XPS analysis, the chemical characteristics of the three activated carbons were probed using Boehm titrations. Boehm titration results presented in Table 2 characterize total acidic and total basic SFGs and confirm the results reported for XPS analysis. The sum of these two values represents the total density of SFGs, for which GACF had the highest density while GAC1 had the lowest. This trend was consistent with XPS analysis that showed GACF had the highest atomic percent noncarbon content and GAC1 had the lowest. The total acidic functional groups were further differentiated by titration analysis according to three types: strong carboxylic acids, weak carboxylic acids, and phenolic groups and carbonyls. Examples of these various SFGs are included in Supporting Information (S5, Figure S5-1). Of the three carbons, acidic SFGs were most abundant on GACF, and over 95% of acidic contributions came from weak carboxylic acid groups and hydroxyl groups. Total acidic SFGs were nearly as abundant on ACF as for GACF, and, similarly, weak carboxylic acid groups and hydroxyl groups contributed much more (98%) to the total surface acidity than strong carboxylic acid groups. In contrast, GAC1 received a proportionately higher contribution to total surface acidity from strong and weak carboxylic acids than either of the other two activated carbons. With respect to basicity, in addition to surface oxides, such as chromenes, pyrones, and quinones, delocalized π electrons in the basal planes of the activated carbon contribute significantly to carbon surface basicity.18 The following trend in surface basicity was observed: GAC1 > ACF > GACF. Correlation of Adsorption with Surface Functional Groups. Water vapor uptake by activated carbon has been associated with increased density of oxygen-containing SFGs, particularly acidic ones.19 Based on surface acidity, water vapor uptake should increase in the following order: GAC1 < ACF < GACF. Figure 4 shows that the strength of this relationship was good at low P/P0 (0.3, 0.4, 0.5), but at higher P/P0 values, no significant trend was observed. 6501

dx.doi.org/10.1021/es104286d |Environ. Sci. Technol. 2011, 45, 6498–6503

Environmental Science & Technology

ARTICLE

Table 2. Physical and Chemical Characteristics of Activated Carbon Adsorbents physical characterization

chemical characterization atomic percent

a

sample carbon

surface area [m2/g]

total pore volume [cm3/g]

micropore volume [cm3/g]

min pore width [nm]

ACF GAC1 GACF

1084 869 860

0.41 0.34 0.34

0.38 0.33 0.32

0.53 0.50 0.53

C 1s

O 1s

N 1s

K 2p

I 3d

92.5 6.7 0.77 nda nd 94.0 5.8 0.23 nd nd 91.8 7.2 0.26 0.47 0.38

total basic groups [μeq/m2]

total acidic groups [μeq/m2]

strong carboxylic groups [μeq/m2]

0.79 0.58 1.10

0.83 0.28 0.89

0.01 0.09 0.03

weak carboxylic groups and lactones [μeq/m2]

phenolic groups and carbonyls [μeq/m2]

0.22 0.18 0.59

0.60 0.01 0.26

Not detected.

Figure 4. Evaluation of relationship between water vapor adsorption at P/P0 = 0.3 (0), P/P0 = 0.4 (4), P/P0 = 0.5 (]), and P/P0 = 0.7 (O) and surface concentration of total acid groups [μeq/m2]. Correlation coefficients are reported beside each trend line.

The positive correlations between total acidic groups and water vapor adsorption capacity at lower relative pressures are consistent with the observation that ξ0 values correlated positively with increased water vapor adsorption at low relative pressures. This trend substantiated the expectation that the parameter ξ0 represents adsorbate
adsorbent affinity due to surface chemical properties. Among the few studies of HCHO adsorption on activated carbon, two have suggested that increased nitrogen content is associated with increased HCHO adsorption.12,20 Although no mechanism was proposed to explain this observation, Matsuo et al.20 suggested that amino groups behave as Schiff bases during the adsorption process. For the activated carbons used in this study, Boehm titration results demonstrated that basic character increased in the following order: GAC1 < ACF < GACF. Based on surface basicity, HCHO uptake should increase in the following order: GAC1 < ACF < GACF. Figure 5 shows that the strength of this correlation was good for low concentrations (3 ppmv, 7 ppmv). However, at higher HCHO concentrations this relationship was not evident. As HCHO concentrations approach 15 ppmv, multilayer adsorption is expected to commence, at which point intermolecular interactions between the activated carbon surface and HCHO would no longer be the driving force behind subsequent HCHO adsorption. These observations suggest that adsorbate surfaces with more basic character have increased potential for HCHO removal than

Figure 5. Evaluation of correlation between HCHO adsorption at 3 ppmv (0,9), 7 ppmv (4,2), and 11 ppmv (],[) and surface concentration of acidic (closed symbols) or basic (open symbols) groups [μeq/m2]. Correlation coefficients are reported beside each trend line. Solid lines (—) correspond to basic SFG density. Dashed lines (- - -) correspond to acidic SFG density.

surfaces with less basic character, and this potential may be greatest at low HCHO concentrations. The previous studies that have demonstrated a positive association between HCHO adsorption and activated carbon nitrogen content ruled out the influence of oxygen-containing SFGs on the basis of no observed association between HCHO adsorption capacity and total oxygen content. However, HCHO may behave as both a Lewis acid and a Lewis base21 and, therefore, potentially interacts with both basic and acidic SFGs on activated carbon. Thus, relationships between HCHO adsorption capacity at low concentrations and total acidic SFGs were also investigated (Figure 5). While the correlation coefficients are not as high as for total basic SFGs, a trend of increasing HCHO adsorption capacity with increasing density of total acidic SFGs was observed at the two lowest concentrations. Water vapor and HCHO adsorption capacity on these different activated carbons may be related through their dependence on total density of acidic SFGs. At low relative pressures for water vapor and low concentrations for HCHO, the first value of the Q
L model, ξ0, already demonstrated a positive relationship with increasing adsorption capacity. Thus, the relationship between ξ0-H2O and ξ0-HCHO was considered, and the two parameters were found to have a correlation coefficient of 82%. This relationship suggests that 6502

dx.doi.org/10.1021/es104286d |Environ. Sci. Technol. 2011, 45, 6498–6503

Environmental Science & Technology water vapor adsorption capacity may be an indicator of relative HCHO adsorption capacity on different activated carbons. This study quantified the capacity of several activated carbons for gas-phase HCHO removal over a range of equilibrium HCHO concentrations. HCHO isotherms exhibited adsorption behavior similar to that of water vapor on microporous adsorbents and were successfully fit using a model previously developed for water vapor adsorption. At low concentrations, total density of basic SFGs demonstrated the strongest relationship with HCHO removal. Modifications of activated carbon surfaces to incorporate more basic properties would be instrumental for investigation of the impact of adsorbent surface basicity on HCHO removal at low concentrations. At the same time, total density of acidic SFGs also demonstrated a positive relationship with increasing HCHO adsorption capacity. Although this relationship was not as strong as was observed for total density of basic SFGs, it led to the observation that for low relative pressures, water vapor adsorption capacity can provide an indication of relative performance of different activated carbons with respect to their expected capacity to remove HCHO to the extent that acidic functional groups influence HCHO adsorption capacity. Adsorption capacity for water vapor, however, will not be an indicator of the impact that basic surface functional groups may have on HCHO adsorption capacity. Given that water vapor adsorption capacity is much simpler to measure on activated carbon than HCHO adsorption capacity, this relationship could facilitate the development of adsorbent-based, gas-phase air treatment strategies for indoor environments. Such treatment strategies would be valuable for the reduction of HCHO concentrations in residential homes, apartments, manufactured housing, and temporary housing units, where the need to reduce personal exposure to HCHO is most acute.

’ ASSOCIATED CONTENT

bS

Supporting Information. Experimental procedures, measured data, and representative images. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: (512) 471-4244; e-mail: [email protected].

’ ACKNOWLEDGMENT Special thanks to Vidyasagar Sunkavalli and Donald Marek (TAMUK, Texas) for assistance with gravimetric adsorption experiments and to Hugo Celio for assistance with XPS analysis. E.M.C. was funded by a National Science Foundation (NSF) IGERT program (Award DGE 0549428).

ARTICLE

(4) Cooke, T. F. Indoor air pollutants: A literature review. Rev. Environ. Health 1991, 9, 137–160. (5) Weschler, C. J.; Hodgson, A. T.; Wooley, J. D. Indoor chemistry: Ozone, volatile organic compounds, and carpets. Environ. Sci. Technol. 1992, 26 (12), 2371–2377. (6) Liu, W.; Zhang, J.; Korn, L. R.; Zhang, L.; Weisel, C. P.; Turpin, B.; Morandi, M.; Stock, T.; Colome, S. Predicting personal exposure to airborne carbonyls using residential measurements and time/activity data. Atmos. Environ. 2007, 41, 5280–5288. (7) Hodgson, A. T.; Levin, H. Volatile Organic Compounds in Indoor Air: A Review of Concentrations Measured in North America since 1990; LBNL51715; Lawrence Berkeley National Laboratory: Berkeley, CA, 2003. (8) National Institute for Occupational Safety and Health (NIOSH). Pocket Guide to Chemical Hazards; United States Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention: Cincinnati, OH, 2009. (9) Agency for Toxic Substances and Disease Registry (ATSDR). An update and revision of ATSDR’s February 2007 health consultation: formaldehyde sampling of FEMA temporary-housing trailers ; Baton Rouge, LA, 2007. (10) Yao, M.; Zhang, Q.; Hand, D. W.; Perram, D. L. Investigation of the treatability of the primary indoor volatile organic compounds on activated carbon fiber clothes at typical indoor concentrations. J. Air Waste Manage. 2009, 59, 882–890. (11) Boonamnuayvitaya, V.; Sae-ung, S.; Tanthapanichakoon, W. Preparation of activated carbons from coffee residue for the adsorption of HCHO. Sep. Purif. Technol. 2005, 42, 159–168. (12) Song, Y.; Qiao, W.; Yoon, S.; Mochida, I.; Guo, Q.; Liu, L. Removal of HCHO at low concentration using various activated carbon fibers. J. Appl. Polym. Sci. 2007, 106, 2151–2157. (13) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area, and Porosity, 2nd ed.; Academic Press: London, NY, 1982. (14) Sullivan, P. D.; Stone, B. R.; Hashisho, Z.; Rood, M. J. Water adsorption with hysteresis effect on microporous activated carbon fabrics. Adsorption 2007, 13, 173–189. (15) Furmaniak, S.; Gauden, P. A.; Terzyk, A. P.; Rychlicki, G. Water adsorption on carbons—critical review of the most popular analytical approaches. Adv. Colloid Interface Sci. 2008, 137, 82–143. (16) Qi, N.; LeVan, M. D. Coadsorption of organic compounds and water vapor on BPL activated carbon. 5. Methyl ethyl ketone, methyl isobutyl ketone, toluene, and modeling. Ind. Eng. Chem. Res. 2005, 44, 3733–3741. (17) Rudisill, E. N.; Hacskaylo, J. J.; LeVan, M. D. Coadsorption of hydrocarbons and water on BPL activated carbon. Ind. Eng. Chem. Res. 1992, 31, 1122–1130. (18) Boehm, H. P. Surface oxides on carbon and their analysis: a critical assessment. Carbon 2002, 40, 145–149. (19) Salame, I. I.; Bandosz, T. J. Study of water adsorption on activated carbons with different degrees of surface oxidation. J. Colloid Interface Sci. 1999, 210, 367–374. (20) Matsuo, Y.; Nishino, Y.; Fukutsuka, T.; Sugie, Y. Removal of formaldehyde from gas phase by silylated graphite oxide containing amino groups. Carbon 2008, 46, 1162–1163. (21) Rivelino, R. Lewis acid-base interactions in weakly bound formaldehyde complexes with CO2, HCN, and FCN: considerations on the cooperative H-bonding effects. J. Phys. Chem. A 2008, 112 (2), 161–165.

’ REFERENCES (1) Monographs on the Evaluation of Carcinogenic Risks to Humans: HCHO; World Health Organization, International Agency for Research on Cancer, 2006; 88 pp. (2) Thompson, C. M.; Subramaniam, R. P.; Grafstr€om, R. C. Mechanistic and dose considerations for supporting adverse pulmonary physiology in response to formaldehyde. Toxicol. Appl. Pharmacol. 2008, 233, 355–359. (3) McGwin, G., Jr.; Lienert, J.; Kennedy, J. L., Jr. Formaldehyde exposure and asthma in children: a systemic review. Environ. Health Perspect. 2010, 118 (3), 313–317. 6503

dx.doi.org/10.1021/es104286d |Environ. Sci. Technol. 2011, 45, 6498–6503