Influence of Molecular Structure and Adsorbent Properties on Sorption

Article pubs.acs.org/est

Influence of Molecular Structure and Adsorbent Properties on Sorption of Organic Compounds to a Temperature Series of Wood Chars Charisma Lattao,† Xiaoyan Cao,‡ Jingdong Mao,‡ Klaus Schmidt-Rohr,§ and Joseph J. Pignatello*,† †

Department of Environmental Sciences, The Connecticut Agricultural Experiment Station, 123 Huntington Street, P.O. Box 1106, New Haven, Connecticut 06504-1106, United States ‡ Department of Chemistry and Biochemistry, Old Dominion University, 4541 Hampton Boulevard, Norfolk, Virginia 23529, United States § Department of Chemistry, Iowa State University, Ames, Iowa 50011, United States S Supporting Information *

ABSTRACT: Chars from wildfires and soil amendments (biochars) are strong adsorbents that can impact the fate of organic compounds in soil, yet the effects of solute and adsorbent properties on sorption are poorly understood. We studied sorption of benzene, naphthalene, and 1,4dinitrobenzene from water to a series of wood chars made anaerobically at different heat treatment temperatures (HTT) from 300 to 700 °C, and to graphite as a nonporous, unfunctionalized reference adsorbent. Peak suppression in the NMR spectrum by sorption of the paramagnetic relaxation probe TEMPO indicated that only a small fraction of char C atoms lie near sorption sites. Sorption intensity for all solutes maximized with the 500 °C char, but failed to trend regularly with N2 or CO2 surface area, micropore volume, mesopore volume, H/ C ratio, O/C ratio, aromatic fused ring size, or HTT. A model relating sorption intensity to a weighted sum of microporosity and mesoporosity was more successful. Sorption isotherm linearity declined progressively with carbonization of the char. Application of a thermodynamic model incorporating solvent−water and char−graphite partition coefficients permitted for the first time quantification of steric (size exclusion in pores) and π−π electron donor−acceptor (EDA) free energy contributions, relative to benzene. Steric hindrance for naphthalene increases exponentially from 9 to 16 kJ/mol (∼ 1.6−2.9 log units of sorption coefficient) with the fraction of porosity in small micropores. π−π EDA interactions of dinitrobenzene contribute −17 to −19 kJ/mol (3−3.4 log units of sorption coefficient) to sorption on graphite, but less on chars. π−π EDA interaction of naphthalene on graphite is small (−2 to 2 kJ/mol). The results show that sorption is a complex function of char properties and solute molecular structure, and not very predictable on the basis of readily determined char properties.

■

INTRODUCTION Biomass chars formed in vegetation fires or produced intentionally for use as soil amendments (biochar)1 are reputed to be strong adsorbents of organic compounds. Sorption to chars is typically more nonlinear and greater per unit mass than sorption to the nonpyrogenic natural organic matter of soils and sediments.2−4 Consequently, wildfire chars may play an important role in the fate of pollutants,5,6 and biochar amendments may have undesirable or desirable effects, such as reducing efficacy and mobility of soil pesticides,7 or reducing the mobility and bioavailability of hazardous organic compounds.8,9 With heating, biomass evolves from the biopolymer phase to an amorphous phase of thermally altered biomolecules, to a composite phase of emerging clusters of graphene sheets randomly mixed with the altered biomolecule phase, and finally to an amorphous state consisting of large, irregularly-spaced © 2014 American Chemical Society

and poorly-stacked graphene sheets, possibly interconnected with flexible linkers.10 The products of lignin-rich and ligninpoor feedstocks differ in crystallinity, condensed aromaticity, and surface/pore dimensions.10−12 Previous 13C NMR studies of the maple wood char series used in the present study reveal a transition from primarily ligno-cellulosic structure at 300 °C to predominantly polyaromatic structure starting at 350 °C.13 Char properties that may influence sorption include surface area, pore size distribution, polar group identity and content, surface charge (for ionic solutes), and others. Such properties can vary greatly with feedstock, heating rate, final heat treatment temperature (HTT) and duration, surrounding Received: Revised: Accepted: Published: 4790

November 20, 2013 March 25, 2014 April 7, 2014 April 23, 2014 dx.doi.org/10.1021/es405096q | Environ. Sci. Technol. 2014, 48, 4790−4798

Environmental Science & Technology

Article

biochars.1 The solutes benzene (BEN), naphthalene (NAP), and 1,4-dinitrobenzene (DNB) were selected as model compounds to reveal systematic effects of molecular structure (specifically, size and π-donor/π-acceptor character) and char properties on sorption intensity and isotherm linearity. BEN and NAP are known soil contaminants originating mainly from petroleum fuel spills; whereas DNB is not used commercially but is closely related structurally to 1,3-dinitrobenzene, a precursor of explosive compounds. The NMR relaxation probe, 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO) was employed to assess matrix penetration of sorbing molecules. We were especially interested in how much of the char C lies in proximity to sorption sites with evolution of char structure; how sorption intensity and linearity vary with char physical− chemical properties and surface/pore characteristics; how steric effects vary with pore size distribution; and how π−π EDA interactions vary with fused ring C cluster size. To aid interpretation we employed a thermodynamic model that uses data on sorption to nonporous graphite and partitioning into nonpolar solvents to calculate free energies associated with π−π EDA interactions and steric effects.

gases, and other factors. Critical solute properties governing sorption include “hydrophobicity”, polarity, hydrogen bonding capability, steric size/shape, π-donor−acceptor properties, and charge. The contribution of individual driving forces, and the ways in which char and solute properties interrelate to determine sorption intensity and concentration dependence, are poorly understood. Char polar group content influences sorption through retention of water, which competes for sorption space14,15 and hydrogen bonding sites.16 Evidence is accumulating that π−π electron donor−acceptor (EDA) interactions can occur between aromatic π-acceptor rings of the solute having strong electron-withdrawing groups, and π electron-rich regions of the polyaromatic surfaces of graphite, chars,14,17,18 and carbon nanotubes.19 The π−π EDA force is governed mainly by quadrupolar interactions between ring systems. The quadrupole arises from the electronic distribution in the π−σ systems of the interacting rings. The strength of the π−π EDA force varies as the quadrupole vectors of the interacting species differ in magnitude and direction; thus, the strongest interactions are between a π-donor and a π-acceptor and the weakest between two π-donor species. The contribution of π−π EDA interactions to sorption may depend on available polyaromatic surface area of the adsorbent, but this hypothesis has not been tested. Sorption of nonionic aromatic compounds to chars and soots seems to trend with surface area, but surface area alone is poorly predictive.4,20−22 One problem is that N2 adsorption at 77 K, the most common technique for measuring surface area, suffers from kinetic diffusion limitation for N2 in small micropores due to matrix inflexibility, leading to artificially low surface area for some chars.23−25 Because pore filling is an important sorption process in rigid porous materials,3,20,26 pore size distribution complicates the relationship between surface area and sorption because the filling of micropores (< 2 nm) involves more points of contact than the filling of mesopores (2−50 nm). In addition, pore-filling is subject to size exclusion that depends on the critical diameter of the solute and the pore diameter. Previous studies of single char samples have identified steric effects attributable to size exclusion.18,26,27 For example, adsorption of aromatic compounds to a wood char relative to nonporous graphite declined with the number of substituents or fused ring size.18,27 Studies relating sorption to surface area and pore size distribution are lacking, and few studies have addressed the effects of feedstock and pyrolysis conditions. Sorption of hydrophobic compounds to chars derived from lignin-poor biomass was attributed to surface adsorption combined with a minor partitioning component for chars made at low HTT.21,28 Grass char showed lower affinity for benzene and toluene than wood char at the same HTT, ascribed to the lower surface area and higher mineral content of the former.29 Phenanthrene sorption to field and laboratory wood chars was greatest for chars with the highest N2 surface areas, yet the correlation across a range of HTTs was weak.22 Sorption nonlinearity increased with HTT, however.22,29 This study focuses on the adsorbent properties of a series of maple wood chars produced at 300−700 °C toward three nonionic aromatic solutes. Maple is a hardwood and thus its ligno-cellulosic structure is representative of angiosperms combusted in vegetation fires, as well as hardwood waste materials, such as sawdust and forest litter, that are sometimes used as feedstocks for biochar. The employed temperature range is also typical of the range experienced in vegetation fires and used to produce most laboratory and commercial

■

EXPERIMENTAL SECTION Materials. Graphite (< 20 μm particles; 99.99% C; no detectable O), benzene (BEN, 99+ %), and naphthalene (NAP, 99+ %) were purchased from Aldrich, and 1,4-dinitrobenzene (DNB, 99.5%) was obtained from Chem Service. Note that benzene or hexadecane used as a liquid partitioning phase will be notated “liq-ben” or “liq-hex”. Table S1 in the Supporting Information (SI) lists selected physical constants and properties of the solutes. Chars were prepared as previously described13 from maple wood shavings by heating in a flow of N2 at 25 °C min−1 to the desired HTT between 300 and 700 °C and held there for 2 h. The chars are designated c300, c350, etc. Characterization of chars was reported previously13 and selected results are reproduced in the SI including elemental and ash composition (Table S2), functional group composition by 13C NMR (Table S3), and surface and pore dimensions (Table S4). NMR Spectroscopic Determination of the Fraction of Char Carbon Near Sorption Sites. TEMPO (2,2,6,6tetramethylpiperidine-1-oxyl; Aldrich) was sorbed to the chars by three successive incubation periods of 1 d, 1 d, and 6 d, with a solution of 8000 mg L−1 TEMPO in water (solubility, 11 270 ± 80 mg L−1). Samples were freeze-dried and subjected to cross-polarization/total suppression of sidebands (CP/TOSS) NMR spectroscopy in a manner similar to the previous study.30 Details of sorption and NMR analysis are provided in the SI. Sorption. Vials containing known amounts of char (15−50 mg), water (10−250 mL) containing 200 mg L−1 NaN3 as bioinhibitor, and test solutes in methanol carrier ( NAP > BEN. The isotherms were normalized for hydrophobic effects based on the liquid benzene−water or hexadecane−water partition coefficients (SI Figure S5b). After so doing, the BEN isotherm and, to a lesser extent, the NAP isotherm approach the Hydrophobic Effects Isotherm, whereas the DNB isotherm lies far above the Hydrophobic Effects Isotherm, suggesting DNB has much greater affinity for graphite than expected based on its hydrophobicity. The enhanced affinity of DNB cannot be due to dipolar or Hbonding forces because graphite lacks polar functionality and DNB is an exceedingly weak base. The enhanced sorption of DNB is thus consistent with π−π EDA interactions with the graphite basal plane.17,18 The formation of π−π EDA bonds between DNB and aromatic donors is supported by the appearance of chargetransfer bands in the UV/visible spectra consistent with such interactions. Figure 4 shows absorbance tailing of DNB into the visible region between ∼425 and 500 nm in benzene but not in hexadecane. This absorbance may be assigned to a chargetransfer band due to weak π−π EDA bonding between DNB and benzene.34 A charge-transfer band appears also in dilute solutions of DNB and pyrene in both hexadecane and benzene. The formation constant for complexation between a given πacceptor and a series of π-donors increases with fused ring size (naphthalene → pyrene) due to increasing donor polarizability.18,35,36 Consistent with such trend, DNB does not show a charge-transfer band with naphthalene or phenanthrene

(6)

Sorption in a “partition” domain composed of uncarbonized lignocellulose material is also possible, especially for lower HTT chars, but unquantifiable from our data. Nevertheless, we regarded it as negligible ( 50 nm). Although macropores do exist and may control water movement to the smaller pores, their contribution to total equilibrium sorption is likely to be small. Considering the data in total, the model of eq 6 is moderately predictive (Figure 3). The residuals are normally distributed

Figure 3. Predictive ability of the model that assumes a weighted distribution between micropores and mesopores on a volumetric basis: Kch‑C,r = a · microporosity + b · mesoporosity.

and reasonable in size for most points (mean, 0.30 log units). However, the r2 and adjusted r2 values for the individual solute−Kch‑C,r data sets (ranging from 0.66−0.93 and from 0.32−0.65, respectively; SI Table S8) are poor. In two cases, negative coefficients were obtained (a for NAP at Cr = 0.001, and b for DNB at Cr = 0.01), which is unrealistic. In addition, the coefficient of mesoporosity is relatively less important at the higher of the two Cr, which seems counterintuitive. Thus, this study offers only tepid support to models such as eq 6 based on pore size distribution. It is possible that some of the volume probed by CO2 and N2 may be unavailable to the organic

Figure 4. Difference spectra showing charge-transfer absorbance of p-dinitrobenzene π−π complexes in solution. The spectrum of the cosolute in each case is subtracted out. (a) 0.02 M π-donors in hexadecane (S = liq-hex) saturated with p-dinitrobenzene; (b) 0.02 M pyrene and 0.01 M dinitrobenzene in benzene (S = liq-ben). Spectra obtained on a Shimadzu UV-2600. 4794

dx.doi.org/10.1021/es405096q | Environ. Sci. Technol. 2014, 48, 4790−4798

Environmental Science & Technology

Article

in dilute solution in hexadecane (Figure 4), whereas it does so with pyrene. The charge-transfer band does appear when DNB alone is dissolved in benzene by virtue of benzene’s high concentration. The formation of π−π EDA bonds between DNB and aromatic donors is also supported by solvent−water partitioning behavior. The octanol−water partition coefficient (log Kow) is 1.49 for DNB, 2.15 for BEN, and 3.30 for NAP (SI Table S1). DNB favors liquid benzene over octanol by 1.15 log units, much more so than does NAP (by 0.81 units) or BEN (by 0.55 units). Likewise, DNB favors liquid benzene over hexadecane (by 1.94 log units) much more so than does NAP (by 0.70 units) or BEN (by 0.55 units). Taken together, these trends support π−π EDA interactions between DNB and benzene, in which DNB serves as a strong π-acceptor and benzene as a weak π-donor. Returning attention to the hydrophobic effects normalized isotherms in SI Figure S5b, the gap between the DNB isotherm and the Hydrophobic Effects Isotherm on graphite is smaller employing benzene rather than hexadecane as the reference solvent. This is because liquid benzene in the solvent shell of DNB forms weak π−π EDA bonds with DNB that “compete” virtually with interactions between DNB and π-donor sites on graphite. (The isotherm of NAP is much closer to the Hydrophobic Effects Isotherm, but is affected by benzene in an analogous manner as DNB. This suggests that naphthalene may interact, albeit weakly, with π-acceptor sites along graphite step edges (where layers overlap) which are thought to be positively polarized,37 and that benzene competes virtually with NAP for these sites.) Comparison of Sorption Behavior on the Chars. Sorption of the test compounds on the chars is strongly nonlinear; N follows the order: BEN > NAP > DNB (SI Table S5). DNB characteristically sorbs more strongly than NAP and BEN (SI Figure S3), although there is some isotherm crossover at high concentration for chars produced at 500 °C or greater. The hydrophobic effects-normalized isotherms in SI Figure S6 show a similar trend, and also reveal that the gap between the DNB and BEN isotherms is smaller when benzene rather than hexadecane is the reference solvent. As on graphite, these results suggest π−π EDA interaction of DNB with the polyaromatic surfaces of char. As chars are porous and the molecules are of different sizes, however, it is first necessary to take into account steric effects. The critical molecular diameter of NAP is larger than those of BEN and DNB, which are comparable (SI Table S1). Application of a Thermodynamic Model to Quantify π−π EDA and Steric Contributions. Mechanistic interpretation of sorption on char and graphite is aided by a previously developed model18 that partitions sorption energetically into different driving forces: (a) hydrophobic effects (hyd), assigned to forces involved in disruption of water cohesion, plus dispersion interactions of i with the surface; (b) dipole−dipole and hydrogen bonding interactions (dip); (c) π−π EDA interactions (π−π); and (d) steric hindrance to sorption (ster). The governing equation is derived in the SI in suitable form for this paper. It expresses the Gibbs free energy of sorption (notation, Δsbt‑W) of compound i relative to BEN (notation, Δi‑BEN), as

Δi ‐ BENΔsbt ‐ W G ≡ (Gsbt − G W )i − (Gsbt − G W )BEN e,dip = aS,sbtΔi ‐ BEN (ΔS − W G − ΔS − W G π − π ) + Δi ‐ BENGsbt π−π ster + Δi ‐ BENΔsbt ‐ W G + Δi ‐ BENΔsbt ‐ W G

(7)

where ΔS−W G is the free energy of solvent−water partitioning (S = liq-hex or liq-ben), which is linearly correlated with Δsbt‑W Ghyd‑S with a slope of aS,sbt; and where the superscript (e) denotes the excess free energy compared to the standard state of i. Equation 7 is identical to SI eq S7. The SI further includes equations that follow from eq 7 for quantifying steric and π−π EDA contributions, both relative to BEN. The π−π EDA contribution to sorption on graphite is given by eq 8 Δi ‐ BEN Δgr ‐ W G π − π = −RT ln

K gri ‐ W,r K grBEN ‐ W,r

+ aS,grRT ln

i KSW BEN KSW

+ {aS,gr Δi ‐ BENΔS − W G π − π }S = liq ‐ ben (8)

The term in braces applies only to S = liquid benzene, since hexadecane has no π system. Figure 5a shows Δi‑BEN Δgr‑W Gπ−π as a function of sorbed concentration, normalized by the hypothetical Freundlich maximum sorbed concentration to facilitate comparison between NAP and DNB. The following observations can be made. First, the value of Δi‑BEN Δgr‑W Gπ−π when S = liq-hex ranges from −17 to −19 kJ/mol for DNB, but is close to zero for NAP (∼ −2 to +2 kJ/mol). This indicates that (relative to BEN) the surface is π-donating toward the strong π-acceptor DNB, but does not interact strongly with the weak π-donor NAP. Recalling that 1 kcal/mol increase in free energy corresponds to −(RT log K)/2.3 units of increase in the equilibrium sorption constant, the range of values for DNB represents a contribution of π−π EDA forces to overall sorption of 3.0−3.4 log units of Kgr‑W. Second, Δi‑BEN Δgr‑W Gπ−π becomes less negative with concentration. This may indicate that π-donating sites on the surface are distributed in energy, with the strongest filled first. Third, Δi‑BEN Δgr‑W Gπ−π is more negative for S = liq-hex than S = liq-ben, especially for DNB. Because water has no π system, the term in braces in eq 8 represents the difference in π−π EDA free energy between i··· liq-ben and BEN···liq-ben. For DNB, this term is expected to be negative because π−π EDA interactions for DNB···liq-ben are more favorable than for BEN···liq-ben. Comparing the S = liq-hex and S = liq-ben cases, the term in braces in eq 8 corresponds to ∼ −9 kJ/mol. Thus, liq-ben in its role as reference solvent “dampens” DNB···graphite π−π EDA interactions by engaging (virtually) in π−π EDA interactions with DNB. The Δi‑BEN Δgr‑W Gπ−π for NAP is much less affected by choice of reference solvent (< 2 kJ/mol) because both NAP···liq-ben and BEN···liq-ben π−π EDA forces are weak, as the molecules are not opposite in quadrupolarity in either case. The calculated and measured quadrupole moments of BEN and NAP are both negative (more negative in the case of NAP), whereas the calculated quadrupole moment of DNB is positive due to the electron withdrawing nitro groups.34 It is possible to quantify steric effects by selecting graphite as a reference adsorbent. SI Figure S7 shows char−graphite isotherms constructed from char−water and graphite−water isotherms and normalized to the CO2−SA. Note that the char− graphite isotherms of BEN are displaced upward of the 1:1 4795

dx.doi.org/10.1021/es405096q | Environ. Sci. Technol. 2014, 48, 4790−4798

Environmental Science & Technology

Article

exclusion of NAP relative to BEN. The DNB isotherm is also downward-displaced of the BEN isotherm, but the gap is roughly the same at all HTTs. DNB and BEN have the same critical diameter (SI Table S1), so this behavior may signify electronic rather than steric effects. The steric contribution to sorption of NAP by char relative to graphite and relative to BEN is quantified by ΔNAP ‐ BEN Δch ‐ gr Gster = −RT ln

NAP Kch ‐ gr BEN Kch ‐ gr

+ ΔaS,ch ‐ grRT ln

NAP KSW BEN KSW

+ {ΔaS,ch ‐ gr ΔNAP ‐ BEN ΔS ‐ W G π − π }S = liq ‐ ben

(9)

The steric contribution of NAP sorption to the chars (Figure 5b) ranges from +9 to +16 kJ/mol. These values represent a substantial impact on NAP sorptionranging 1.6−2.9 log units reduction in Kchar at 20 °C from what it otherwise would be. The term in braces represents the difference between NAP and BEN in π−π EDA free energy with S. The near-independence of ΔNAP‑BEN Δch‑gr Gster(NAP) on S implies that the term in braces must be small, consistent with the expectation that NAP···BEN and BEN···BEN π−π EDA interactions are both small, as the pairs in each case lack opposing quadrupolarity. The steric contribution to NAP sorption increases with the fraction of pore volume present in small micropores (3.5−7.2 Å) relative to the total micropore−mesopore volume (3.5−500 Å). Thus, size exclusion becomes more severe as sorption volume becomes dominated by small micropores. Steric contribution to sorption of DNB on the chars cannot be separated from the dipolar contribution by the present approach. The sum of the π−π EDA and dipolar contributions to DNB sorption by char relative to graphite is given by dip π−π ΔDNB ‐ BEN (Gch + Gch − Ggrπ − π )

= −RT ln

DNB Kch ‐ gr BEN Kch ‐ gr

+ ΔaS,ch ‐ grRT ln

DNB KSW BEN KSW

+ {ΔaS,ch ‐ gr ΔDNB ‐ BEN ΔS ‐ W G π − π }S = liq ‐ ben

(10)

Equation 10 assumes that steric contributions for DNB and BEN are similar. This assumption is reasonable because DNB and BEN have the same critical diameter (4.3 Å), and because char−graphite distribution coefficients of benzene with 0, 1, or 2 substituents were the same within a factor of 2 and showed no trend.27 Estimates of the π−π EDA and dipolar contributions to DNB sorption by char relative to graphite obtained from eq 10 are plotted as a function of fused ring size in Figure 5c, with HTT indicated above the points. The values of eq 10 for S = liq-hex are slightly positive (1−3 kJ/mol), indicating that the combined π−π EDA/dipolar interactions of DNB on char are less favorable than the π−π EDA interaction alone on graphite. This, in turn, implies that π−π EDA interactions of DNB are inherently weakeror, more likely, that fewer sorbed molecules engage in themon char than on graphite. For S = liq-ben the values of eq 9 are ∼6 kJ/mol more positive than for S = liq-hex suggesting liquid benzene is a better virtual competitor for π−π EDA sites on char than on graphite. Finally, the combined π−π EDA/dipolar contribution of DNB is nearly invariant with fused ring size. Dipolar forces should

Figure 5. Gibbs free energy contributions, all relative to benzene. (a) π−π EDA component for p-dinitrobenzene and naphthalene on graphite; (b) steric component of naphthalene sorption to chars; (c) π−π EDA plus dipolar component for p-dinitrobenzene sorption to chars.

qch:qgr line; that is, KBEN ch‑gr > 1 for all of the chars. This could mean that CO2 SA is not a perfect yardstick for comparing graphite and char surfaces with respect to organic compound adsorption, or that graphite is an imperfect model. Using BEN as the reference solute, as we do later, makes this irrelevant. SI Figure S7 reveals that the char−graphite isotherms of NAP are displaced downward relative to the corresponding isotherm of BEN, with the gap increasing along the char temperature series (except between c400 and c500). This trend may reflect size 4796

dx.doi.org/10.1021/es405096q | Environ. Sci. Technol. 2014, 48, 4790−4798

Environmental Science & Technology

Article

(6) Lohmann, R.; MacFarlane, J. K.; Gschwend, P. M. Importance of black barbon to sorption of native PAHs, PCBs, and PCDDs in Boston and New York harbor sediments. Environ. Sci. Technol. 2005, 39, 141− 148. (7) Martin, S. M.; Kookana, R. S.; Zwieten, L. V.; Krull, E. Marked changes in herbicide sorption-desorption upon ageing of biochars in soil. J. Hazard. Mater. 2012, 231−232, 70−78. (8) Beesley, L.; Moreno-Jiménez, E.; Gomez-Eyles, J. L. Effects of biochar and greenwaste compost amendments on mobility, bioavailability and toxicity of inorganic and organic contaminants in a multielement polluted soil. Environ. Pollut. 2010, 158, 2282−2287. (9) Hale, S. E.; Hanley, K.; Lehmann, J.; Zimmerman, A. R.; Cornelissen, G. Effects of chemical, bological, and physical aging as well as soil addition on the sorption of pyrene to activated carbon and biochar. Environ. Sci. Technol. 2011, 46, 2479−2480. (10) Keiluweit, M.; Nico, P. S.; Johnson, M. G.; Kleber, M. Dynamic molecular structure of plant biomass-derived black carbon (Biochar). Environ. Sci. Technol. 2010, 44, 1247−1253. (11) McBeath, A. V.; Smernik, R. J. Variation in the degree of aromatic condensation of chars. Org. Geochem. 2009, 40, 1161−1168. (12) Brewer, C. E.; Schmidt-Rohr, K.; Satrio, J. A.; Brown, R. C. Characterization of biochar from fast pyrolysis and gasification systems. Environ. Prog. Sustainable Energy 2009, 28, 386−396. (13) Cao, X.; Pignatello, J. J.; Li, Y.; Lattao, C.; Chappell, M. A.; Chen, N.; Miller, L. F.; Mao, J. Characterization of wood chars produced at different temperatures using advanced solid-state 13C NMR spectroscopic techniques. Energy Fuels 2012, 26, 5983−5991. (14) Zhu, D.; Kwon, S.; Pignatello, J. J. Adsorption of single-ring organic compounds to wood charcoals prepared under different thermochemical conditions. Environ. Sci. Technol. 2005, 39, 3990− 3998. (15) Müller, E. A.; Hung, F. R. Adsorption of water vapor-methane mixtures on activated carbons. Langmuir 2000, 16, 5418−5424. (16) Li, X.; Pignatello, J. J.; Wang, Y.; Xing, B. New insight into the mechanism of adsorption of ionizable compounds on carbon nanotubes. Environ. Sci. Technol. 2013, 47, 8334−8341. (17) Sander, M.; Pignatello, J. J. Characterization of charcoal adsorption sites for aromatic compounds: Insights drawn from singlesolute and bi-solute competitive experiments. Environ. Sci. Technol. 2005, 39, 1606−1615. (18) Zhu, D.; Pignatello, J. J. Characterization of aromatic compound sorptive interactions with black carbon (charcoal) assisted by graphite as a model. Environ. Sci. Technol. 2005, 39, 2033−2041. (19) Chen, W.; Duan, L.; Zhu, D. Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environ. Sci. Technol. 2007, 41, 8295−8300. (20) Wang, X.; Xing, B. Sorption of organic contaminants by biopolymer-derived chars. Environ. Sci. Technol. 2007, 41, 8342−8348. (21) Chun, Y.; Sheng, G.; Chiou, C. T.; Xing, B. Compositions and sorptive properties of crop residue-derived chars. Environ. Sci. Technol. 2004, 38, 4649−4655. (22) James, G.; Sabatini, D. A.; Chiou, C. T.; Rutherford, D.; Scott, A. C.; Karapanagioti, H. K. Evaluating phenanthrene sorption on various wood chars. Water Res. 2005, 39, 549−558. (23) Kwon, S.; Pignatello, J. J. Effect of natural organic substances on the surface and adsorptive properties of environmental black carbon (char): Pseudo pore blockage by model lipid components and its implications for N2-probed surface properties of natural sorbents. Environ. Sci. Technol. 2005, 39, 7932−7939. (24) Pignatello, J. J.; Kwon, S.; Lu, Y. Effect of natural organic substances on the surface and adsorptive properties of environmental black carbon (Char): Attenuation of surface activity by humic and fulvic acids. Environ. Sci. Technol. 2006, 40, 7757−7763. (25) Zimmerman, A. R. Abiotic and microbial oxidation of laboratory-produced black carbon (biochar). Environ. Sci. Technol. 2010, 44, 1295−1301. (26) Nguyen, T. H.; Cho, H.-H.; Poster, D. L.; Ball, W. P. Evidence for a pore-filling mechanism in the adsorption of aromatic hydro-

decrease with ring size or HTT (less O functionality), while π−π EDA forces should increase (more and greater polarizability of the polyaromatic surface), consistent with the observed result. We have seen that sorption depends in a complex and not easily predictable manner on routinely measured char properties, and that sorption can be profoundly affected by steric effects and certain electronic forces, namely π−π EDA forces, if such forces are applicable to the solute. We have also seen that only a small fraction of the char is available for interaction with solute molecules. This work has contributed insight into the sorption of some organic compounds on some chars toward the goal of a general predictive model. This insight can be helpful in understanding and modeling the behavior of contaminants in soil or sediment containing black carbon substances. It can also be used to tailor biochars for maximum contaminant binding efficiency in remediation applications, or, alternatively, for minimum binding efficiency in cases where beneficial species chemical signaling is disrupted or when agrochemical efficacy is jeopardized as a consequence of strong adsorption.

■

ASSOCIATED CONTENT

S Supporting Information *

Details on analytical methodology, benzene−water partition coefficient measurement, NMR spectroscopic analysis, and derivation of a thermodynamic model; tables on solute and sorbent properties and sorption model parameters; figures on surface area and pore volume as a function of HTT, char and graphite and char−graphite sorption isotherms, and sorption intensity and linearity versus porosity, fused ring cluster size, atomic ratios and N2−BET surface area. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; phone: 203-974-8518. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (CBET 0853682 and 0853950). REFERENCES

(1) Lehmann, J.; Joseph, S. Biochar for environmental management: An introduction. In Biochar for Environmental Management: Science and Technology; Lehmann, J., Joseph, S., Eds.; Earthscan: London, 2009; pp 1−10. (2) Cornelissen, G.; Gustafsson, Ö . J. Sorption of phenanthrene to environmental black carbon in sediment with and without organic matter and native sorbates. Environ. Sci. Technol. 2004, 38, 148−155. (3) Jonker, M. T. O.; Koelmans, A. A. Sorption of polycyclic aromatic hydrocarbons and polychlorinated biphenyls to soot and soot-like materials in the aqueous environment: Mechanistic considerations. Environ. Sci. Technol. 2002, 36, 3725−3734. (4) Bucheli, T. D.; Gufstafsson, O-R. J. Quantification of the sootwater distribution coefficient of PAHs provides mechanistic basis for enhanced sorption observations. Environ. Sci. Technol. 2000, 34, 5144− 5151. (5) Cornelissen, G.; Gustafsson, Ö .; Bucheli, T. D.; Jonker, M. O.; Koelmans, A. A.; Van Noort, P. C. M. Extensive sorption of organic compounds to black carbon, coal, and kerogen in sediments and soils: Mechanisms and consequences for distribution, bioaccumulation, and biodegradation. Environ. Sci. Technol. 2005, 39, 6881−6895. 4797

dx.doi.org/10.1021/es405096q | Environ. Sci. Technol. 2014, 48, 4790−4798

Environmental Science & Technology

Article

carbons to a natural wood char. Environ. Sci. Technol. 2007, 41, 1212− 1217. (27) Pignatello, J. J., Adsorption of organic compounds by black carbon from aqueuous solution. In Molecular Environmental Soil Science; Xu, J.-M., Sparks, D. L., Eds.; Springer: New York, 2013; pp 359−385. (28) Chen, Z.; Chen, B.; Zhou, D.; Chen, W. Bisolute sorption and thermodynamic behavior of organic pollutants to biomass-derived biochars at two pyrolytic temperatures. Environ. Sci. Technol. 2012, 46, 12476−12483. (29) Bornemann, L. C.; Kookana, R. S.; Welp, G. Differential sorption behaviour of aromatic hydrocarbons on charcoals prepared at different temperatures from grass and wood. Chemosphere 2007, 67, 1033−1042. (30) Lattao, C.; Cao, X.; Li, Y.; Mao, J.; Schmidt-Rohr, K.; Chappell, M. A.; Miller, M. F.; dela Cruz, A. L.; Pignatello, J. J. Sorption selectivity in natural organic matter studied with nitroxyl paramagnetic relaxation probes. Environ. Sci. Technol. 2012, 46, 12814−12822. (31) Bourke, J.; Manley-Harris, M.; Fushimi, C.; Dowaki, K.; Nunoura, T.; Antal, M. J., Jr. Do all carbonized charcoals have the same chemical structure? 2. A model of the chemical structure of carbonized charcoal. Ind. Eng. Chem. Res. 2007, 46, 5954−5967. (32) Urbanski, T.; Benbenek, S.; Bedynski, S.; Wasilewski, A. Free radicals in charcoal and the combustion of compositions containing charcoal. Explosivstoffe 1970, 1, 9−11. (33) Mackay, A.; Gschwend, P. M. Sorption of monoaromatic hydrocarbons to wood. Environ. Sci. Technol. 2000, 34, 839−845. (34) Heard, G. L.; Boyd, R. J. Calculation of quadrupole moments of polycyclic aromatic hydrocarbons: Applications to chromatography. J. Phys. Chem. 1997, 101, 5374−5377. (35) Zhu, D.; Hyun, S.; Pignatello, J. J.; Lee, L. S. Evidence for pi-pi electron donor-acceptor interactions between pi-donor aromatic compounds and pi-acceptor sites in soil organic matter. Environ. Sci. Technol. 2004, 38, 4361−4368. (36) Lide, D. R., Ed. Handbook of Chemistry and Physics, 85th ed.; CRC Press: Boca Raton, FL, 2004. (37) McDermott, M. T.; McCreery, R. L. Scanning tunneling microscopy of ordered graphite and glassy carbon surfaces: Electronic control of quinone adsorption. Langmuir 1994, 10, 4307−4314.

4798

dx.doi.org/10.1021/es405096q | Environ. Sci. Technol. 2014, 48, 4790−4798