Science of the Total Environment 473–474 (2014) 308–316

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Single-solute and bi-solute sorption of phenanthrene and dibutyl phthalate by plant- and manure-derived biochars Jie Jin a, Ke Sun a,⁎, Fengchang Wu b, Bo Gao c, Ziying Wang a, Mingjie Kang a, Yingcheng Bai c, Ye Zhao a, Xitao Liu a, Baoshan Xing d a

State Key Laboratory of Water Environment Simulation, School of Environment, Beijing Normal University, Beijing 100875, China State Key Laboratory of Environmental Criteria and Risk Assessment, Chinese Research Academy of Environmental Sciences, Beijing 100012, China State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, China d Stockbridge School of Agriculture, University of Massachusetts, Amherst, MA 01003, USA b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Sorption of DBP and PHE in single and binary systems was studied. • PHE and DBP sorption by the biochars was influenced by their surface polarity. • PHE could increase DBP adsorption on biochars. • Increased PHE adsorption by Cd2 + depends on the amount of sorbed Cd2 + by biochars.

a r t i c l e

i n f o

Article history: Received 7 October 2013 Received in revised form 7 December 2013 Accepted 7 December 2013 Available online 27 December 2013 Keywords: Biochar Bi-solute sorption Dibutyl phthalate (DBP) Phenanthrene (PHE) Surface polarity Cd2 +

a b s t r a c t The spatial arrangement of biochar and the exact underlying interaction mechanisms of biochar and hydrophobic organic compounds both remain largely unknown. The sorption of dibutyl phthalate (DBP) and phenanthrene (PHE) to plant- and manure-derived biochars in both single- and bi-solute systems was investigated. The significant positive relation between surface polarity and ash content suggests that minerals benefit the external distribution of polar groups on particle surfaces. PHE and DBP sorption by the biochars was regulated by their surface polarity. The PHE generally displayed a pronounced enhancement of DBP sorption, likely resulting from the formation of biochar–PHE–DBP complexes, suggesting that DBP and PHE had different sorption sites on the biochars. The enhancement of Cd2+ (a soft Lewis acid) on DBP sorption implied that π–π interactions should not dominate DBP sorption by biochars. The influence of Cd2+ on PHE sorption by biochars would depend on the balance between suppressive sorption by Cd2 +\PHE bonding and enhanced sorption by Cd2+complexed functionalities, and the amounts of Cd2+ adsorbed by biochars determined the relative role of increased sorption by Cd2+ in the overall PHE sorption. © 2013 Elsevier B.V. All rights reserved.

⁎ Corresponding author. Tel./fax: +86 10 58807493. E-mail address: [email protected] (K. Sun). 0048-9697/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.scitotenv.2013.12.033

J. Jin et al. / Science of the Total Environment 473–474 (2014) 308–316

1. Introduction Biochar is a charcoal-like material produced by the thermochemical pyrolysis of biomass materials. Biochar has great potential to sequester carbon in soil, improve soil fertility, and remediate groundwater and soil contamination (Barrow, 2012; Laird, 2008). In addition to these promising applications, much of the attention towards biochars is directed to their physicochemical diversity and strong affinity for organic pollutants (Cao et al., 2009; Keiluweit et al., 2010). There is a general consensus that the physicochemical heterogeneity of biochars is controlled by the sources of biomass feedstocks (e.g., wood, crop residues, and poultry litter) and combustion conditions (Keiluweit et al., 2010; Sun et al., 2012; Zheng et al., 2013a, 2013b). It is well established that the progressive thermal alteration of biomass across a charring temperature gradient yields biochars with a dynamic molecular structure from predominantly amorphous carbon to turbostratic crystallites (Keiluweit et al., 2010). At the same time, stacks of graphene (which are polycyclic aromatic) grow into larger sheets, enabling biochars to acquire the ability to engage in π–π electron donor–acceptor (EDA) interactions. The aromatic sheet is expected to be polarizable and may act as an electron donor, acceptor, or possibly show amphoteric properties towards sorbates. Biochars generated at low or intermediate energy input levels are expected to retain a significant number of electron-withdrawing functional groups and may be bipolar; aromatic rings (π-acceptor) in the center of a given sheet are electron deficient, and carbon rings closer to the edges (π-donor) are left electron rich (Sun et al., 2012; Zhu and Pignatello, 2005). Nevertheless, the ways in which aromatic compounds interact with the biochar aromatic core and edge functionality is not clear. Moreover, research on molecular variations among different biochar categories, differences in their organic sorption, and associated underlying mechanisms is still in its infancy. The sorption of hydrophobic organic compounds (HOCs) to biochars has been demonstrated to be strong and nonlinear (Sun et al., 2012; Wang and Xing, 2007). Previous studies attributed the adsorption of aromatic contaminants on chars to π–π EDA interactions, pore-filling processes, hydrophobic driving forces, and H-bonding mechanisms (Sun et al., 2012; Wang and Xing, 2007; Zhu et al., 2005). Furthermore, the bonding strengths of EDA interactions have been demonstrated to be larger than the H-bonding and hydrophobic effects (Keiluweit and Kleber, 2009). However, the contribution of π–π EDA interactions to the overall interactions has not yet been confirmed. The elucidation of major molecular-level interactions that control the sorption of aromatic compounds and the influence of solution-phase compositions (such as coexisting organic compounds and ions) on sorption by biochar properties is of considerable theoretical and practical importance. Phthalate acid esters (PAEs) are plasticizers that are widely used to induce flexibility and workability in polymeric materials. They are prevalent in various media, including surface water, sediments, soil and food (Lin et al., 2003; Zhu et al., 2006). Previous studies demonstrated that PAEs could interfere with normal hormone-regulated processes in humans and wildlife (Liu et al., 2009), and they are categorized as endocrine disruption compounds (EDCs) and priority pollutants by the US EPA and European Union. Because of their widespread presence and adverse health effects, PAE contamination and behavior have become major concerns. Although the sorption of other HOCs such as polycyclic aromatic hydrocarbons (PAHs) (Wang and Xing, 2007; Zhu and Pignatello, 2005) has been extensively investigated, few studies have been performed to illustrate the sorption mechanism of PAEs by biochars. PAHs can behave as donors towards π-acceptor sites in biochars, and PAEs may act as acceptors because of their ester functional group (Sun et al., 2012; Zhu and Pignatello, 2005). Taken together with prior reports indicating that the aromatic sheets in chars were capable of engaging in EDA interactions with both π-donors and π-acceptors, we hypothesize that if π–π EDA interactions control PAH and PAE sorption onto biochars, they should have no or negligible influence on mutual

309

sorption in a bi-solute sorption system. Additionally, the hypothesis can be further tested by investigating the effects of Cd2 + addition. It has been reported that Cd2 + is a soft Lewis acid and can occupy electron-rich sites on biochar via cation\π bonding (Harvey et al., 2011). Thus, it could be expected that the addition of Cd2 + should lessen the PAEs sorption by the biochars via competition of the same electron-rich sites of the biochar and have no or negligible effect on the PAHs sorption due to no competition of sorption sites of the biochars. Therefore, phenanthrene (PHE) and dibutyl phthalate (DBP) were chosen as a representative PAH and PAE. We investigated the sorption of PHE and DBP by five biochars produced from feedstock sources including grass, wood and animal manure in single-solute and bi-solute systems. Cd2+ was introduced as a competitor to elucidate its effect on the sorption of PHE and DBP. The major objectives were as follows: 1) to examine property variations among different biochar categories and their impacts on the sorption of PHE and DBP; and 2) to discuss the contribution of π–π EDA interactions to the overall sorption of PHE and DBP onto biochars. 2. Methods 2.1. Sorbents and sorbates Biochar preparation was described elsewhere (Chen et al., 2008). In brief, the five biochars used in this study were produced from grass (soybean, rice, and cotton) straw, wood dust and swine manure at 450 °C for 2 h after being charred at 200 °C for 2 h in a closed container under N2 conditions. The biochars were further demineralized with 0.1 M HCl. The residues were subsequently separated from the supernatant by centrifugation, washed with deionized (DI) water, freezedried, milled to pass a 0.25-mm sieve and stored for further use. Here, the biochars were classified as WB (wood biochar), GB (grass biochar) and AB (animal waste biochar) according to the feedstock sources. For example, soybean, rice, cotton, wood dust, and swine manure were referred to as GSB, GRB, GCB, WWB, and ASB, respectively. DBP (99 + %) and PHE (98 + %) were purchased from Dr. Ehrenstorfer GmbH (Augsburg, Germany) and Sigma-Aldrich Chemical Co., respectively. The hexadecane–water partition coefficient (log KHW) of PHE and DBP are 4.74 and 3.81, respectively (Abraham et al., 1994; Wang et al., 2010). Cd2 + was applied as a nitrate salt in this study. 2.2. Biochar characterization The physicochemical properties of biochar were determined through elemental analysis, bulk and surface functional group characterization, crystallization (X-ray diffraction, XRD), and CO2 surface area (SA-CO2). For elemental analysis, the C, H, N and O contents were measured with an Elementar Vario EL Elemental Analyzer. Ash contents were measured by heating samples at 750 °C for 4 h. Surface chemistry was evaluated with an FTIR spectrophotometer and an X-ray photoelectron spectroscope (XPS). The FTIR spectrum was collected on a Nexus 670 FTIR spectrophotometer (Thermo Nicolet Corporation, US) using KBr pellets (1–2 mg of sample dispersed in 100 mg of KBr) from 4000 to 400 cm−1 with a resolution of 4 cm−1. XPS was performed on a Thermo Scientific ESCALAB 250 XPS with a Kratos Axis Ultra electron spectrometer using monochromatic Al Kα radiation operated at 225 W. The C1s binding energy levels were assigned as follows: 284.9 eV to C\C, 286.5 eV to C\O, 287.9 eV to C_O, and 289.4 eV to COO. To obtain information on the chemical composition of biochars, their solid-state cross-polarization magic angle-spinning 13 C nuclear magnetic resonance (13C NMR) spectra were obtained using a Bruker Avance 300 NMR spectrometer (Karlsruhe, Germany) operating at a 13C frequency of 75 MHz and a magic angle spinning rate of 12 kHz. XRD patterns were obtained on an X'Pert PRO MPD

310

J. Jin et al. / Science of the Total Environment 473–474 (2014) 308–316

diffractometer (PANalytical, Holland) with Cu Kα radiation at 40 mA and 40 kV. For each sample, a scan time of 19.69 s per 0.033° step was used for 2θ in the 5–80°. The SA-CO2 with a CO2 isotherm at 273 K was calculated by using nonlocal density functional theory (NLDFT) and a grand canonical Monte Carlo simulation (GCMC) (Braida et al., 2003). 2.3. Sorption experiments Batch sorption experiments were conducted to obtain the sorption isotherms of DBP and PHE by five biochars in 15-mL or 40-mL glass vials. The background solution contained 0.01 M CaCl2 to maintain a constant ionic strength and 200 mg L−1 NaN3 to inhibit biological activity. The amount of sorbents was selected to yield 20–80% uptake in the initial aqueous phase concentrations of sorbate (100–10,000 μg L−1 for DBP and 2–1100 μg L−1for PHE), which were obtained by diluting the stock solution with background solution. The methanol content of the test solutions was kept below 0.05% of the volume to minimize the cosolvent effect. All vials were filled with the solution up to the minimum headspace and sealed with Teflon-lined screw caps. On the basis of the preliminary tests (Fig. S1), samples were equilibrated for 7 days (DBP) and 10 days (PHE), respectively, by a shaker at room temperature (23 °C). All the vials were then placed upright for 24 h, and the supernatant was sampled for HPLC analysis. All experiments including the blanks were run in duplicate. In addition, the control samples showed that the loss of DBP and PHE in both systems was negligible. Hence, sorbate uptake by biochar was calculated by mass balance. Single-point Cd2 + sorption experiments were performed with biochars at an initial concentration of 20 mg L− 1, reflecting the environmental contamination level (Ajmal et al., 2003). The pH of the Cd2+ solutions was adjusted to 4.5 with HNO3 or NaOH. Preliminary experiments showed that the sorption equilibrium was reached within 24 h. Procedures for the competitive experiments were the same as those for single-solute sorption experiments with the exception of the following: 1) mixtures of primary solute as well as competing solute were injected into the suspensions; 2) single-point PHE and DBP sorption experiments were performed with the initial PHE concentration of 50 μg L − 1 and the initial DBP concentration of 100 μg L−1 with various concentrations of competitor (DBP or PHE); and 3) 20 mg L−1 Cd2+ was introduced into the various concentrations of PHE and DBP as their competitor. Independent tests with crystalline PHE and DBP were conducted to demonstrate that they did not affect one another's solubility. 2.4. Solubility enhancement experiments To test the hypothesis of cation–π bonding between PHE and Cd2+, the aqueous solubility of PHE was measured in an aqueous solution of Cd(NO3)2 · 4H2O. Vials containing PHE and Cd 2 + solution were shaken in an orbital shaker at 25 ± 1 °C for at least 3 days. An aliquot of the water phase was withdrawn carefully to analyze PHE concentration. The saturated solubility of Cd(NO3)2 · 4H2O in chloroform was also measured in mixtures of PHE. Vials containing mixtures of Cd(NO3)2 · 4H2O powder and PHE in chloroform were covered with aluminum foil to shield them from the light, and they were shaken in an orbital shaker at 25 ± 1 °C for 7 days. Four milliliters of solution was carefully withdrawn and mixed with 10 mL of DI water for Cd2+ extraction. 2.5. Detection of solutes The concentrations of DBP and PHE were determined by HPLC (Dionex Ultimate 3000, reversed phase C18, 250 mm × 4.6 mm × 5 μm, Supelco, PA, USA). DBP was detected by using a UV detector and the detection wavelength was set at 228 nm. The mobile phase was a mixture of 80:20 (v:v) acetonitrile and DI water and the flow rate

was 1 mL min− 1. Low concentrations of PHE were analyzed with a fluorescence detector at 250 nm (excitation wavelength) and 364 nm (emission wavelength) and the UV detector was set to 250 nm to determine concentrations higher than 50 μg L−1. The mobile phase consisted of 90% methanol and 10% water at a flow rate of 0.8 mL min−1. In binary systems, an acetonitrile and water mixture at a volume ratio of 65% to 35% was used at a flow rate of 1 mL min−1 as the mobile HPLC phase. The Cd2 + concentrations were measured by ICP-AES (SPECTRO Company, Germany). 2.6. Data analysis The DBP and PHE sorption data were fitted to a linearized Freundlich model (FM) with the following expression: logqe ¼ logK F þ n logC e where qe (μg g−1) and Ce (μg L−1) are the DBP or PHE solid-phase and liquid-phase concentrations, respectively. KF ((μg g− 1)/(μg L−1)n) is the sorption affinity, and n is the nonlinearity index. The fitting was processed by SigmaPlot 10.0, and the statistical analysis (Pearson correlation analysis and t-test) was performed using SPSS 18.0. 3. Results and discussion 3.1. Biochar characteristics The five biochars varied substantially in their elemental composition (Table 1). It was noted that their organic carbon (OC) contents decreased in the following order: WB N GB N AB. Moreover, their H/C ratios exhibited the reverse sequence. Lower H/C ratios for WWB relative to other biochar samples indicated that WWB was more carbonized and highly thermally altered and that it was an unsaturated material. In addition, the highest H/C ratio of ASB (Table 1) implied that a substantial fraction of original organic residues was reserved (Chun et al., 2004). The O/C and (O + N)/C ratios of ASB were also the highest, reflecting that the ASB had considerable amounts of polar and hydrophilic functional groups. The ash content of the investigated biochars varied by their source of biomass feedstock, and ASB had the highest ash content with a value of 50.9%, followed by GRB (26.2%). The surface elemental composition and functional groups of these biochars according to XPS data are presented in Table 1. The surface polarity ((O + N)/C) values of biochars decreased in the following order: AB N GB N WB (Table 1). Furthermore, the surface polarities of the tested biochars were higher than the corresponding bulk values, with the ratios of bulk polarity to surface polarity ranging from 0.55 to 0.87 (Table 1). In particular, the surface (O + N)/C ratios of GRB and ASB were approximately twice their corresponding bulk ratios (Table 1), indicating that the majority of hydrophobic moieties in GRB and ASB were located in the interior. This orientation left the hydrophilic functionalities facing outside, which differed from a previous report indicating that the surface polarity of humic substances (HSs) from a peat soil was much lower than the corresponding bulk values (Wang et al., 2011). This difference could be attributed to the fact that surface functional groups in HSs likely experienced geological alterations and were modified during the process. The biochars investigated in the present study were produced by charring for a short time and were made of relatively “fresh organic matter” containing abundant polar moieties on the surface. The obvious differences between the surface and bulk composition of these biochars indicated that the composition within their spatial arrangement was heterogeneous. Furthermore, the surface polarity of the biochars was closely associated with their ash contents (Fig. 1a); however, no significant correlation was observed between their bulk polarity and ash contents. Thus, biochar minerals may benefit the concentration of functional groups on the external

J. Jin et al. / Science of the Total Environment 473–474 (2014) 308–316

311

Table 1 Bulk and surface elemental composition, surface functional groups, surface area (SA-CO2) and pore volume of biochars. Samples

GCB GSB GRB WWB ASB

Bulk element composition

Surface area and pore volume

C (%)

H (%)

N (%)

O (%)

N/C

H/C

O/C

(O + N)/C

Ash (%)

SA-CO2a (m2 g−1)

Pore volumea (cm3 g−1)

Pore sizea (nm)

71.6 70.8 57.9 75.9 33.7

3.89 3.92 3.31 3.66 2.55

1.17 0.98 0.83 0.05 2.57

13.3 15.6 11.8 16.7 10.2

0.014 0.012 0.012 0.001 0.065

0.65 0.66 0.69 0.58 0.91

0.14 0.16 0.15 0.16 0.23

0.15 0.18 0.16 0.17 0.29

10.1 8.7 26.2 3.7 50.9

367.1 402.2 293.4 601.6 162.0

0.10 0.12 0.09 0.17 0.05

4.58 4.79 4.18 4.18 4.79

Surface functionalities and element composition

GCB GSB GRB WWB ASB a b c

Total C (%)

C\C (%)

C\O (%)

C_O (%)

COOH (%)

O-containing groups (%)

O (%)

N (%)

Si (%)

Ca (%)

Surface O/C

Surface (O + N)/C

B(O + N)/C / S(O + N)/Cc

79.40 78.26 63.38 74.58 48.48

69.35 84.58 89.49 89.62 75.71

21.89 1.76 2.73 0.68 21.23

4.63 8.03 5.23 7.10 0.57

4.13 5.63 2.54 2.60 2.50

30.7 15.4 10.5 10.4 24.3

18.8 19.8 21.5 18.8 25.7

1.79 1.98 3.14 UL 4.55

ULb UL 11.96 6.58 12.32

UL UL UL UL 8.97

0.18 0.19 0.25 0.19 0.40

0.20 0.21 0.30 0.19 0.48

0.78 0.84 0.55 0.87 0.61

Calculated from nonlocal density functional theory (NLDFT) and grand canonical Monte Carlo simulation (GCMC) using CO2 adsorption. Under limitation of detection. Ratio of bulk to surface polarity. GCB (cotton); GSB (soybean); GRB (rice); WWB (wood dust); and ASB (swine waste).

surface of biochars during production. It was previously reported that the soil minerals were mainly covered by organic matter (OM) (Mikutta et al., 2009) and that the polarity of biochars was independent of their ash contents (Keiluweit et al., 2010). Therefore, it could be concluded that the functional groups of OM mainly contributed to the surface polarity of the investigated biochars because OM was likely easier to scatter and was extensively exposed to the surface minerals when biochars contained higher amounts of minerals. The spatial arrangement of the biochar polarity will influence their sorption of HOCs. The functional groups of biochars tested in this study were examined by using FTIR between 4000 and 500 cm− 1 (Fig. S2). The absorption of aliphatic C\H stretching (2925, 1430 and 1380 cm− 1) was attributed to heat-resistant aliphatic structures in biopolymers (Chen et al., 2005). The carbonyl/carboxyl C_O (~ 1700 cm− 1) and aromatic C_C (~ 1600 cm− 1) stretching vibrations were the most prominent features of the FTIR spectra, indicating the formation of aromatic components. Aromatic C was further confirmed by the appearance of bands between 885 and 750 cm−1 (the aromatic C\H out-ofplane deformations), which also showed a large degree of condensation (Guo and Bustin, 1998). In addition, the band intensity at 1030 cm−1 in the ASB reflected that it contained a lot of undecomposed cellulosic and ligneous C, which was consistent with its relatively high H/C ratio. To better understand the biochar functional groups, the 13C NMR spectra and integrated data of biochars are shown in Fig. S2 and listed in Table S1. The biochars were enriched in aryl and O-aryl C at 93–165 ppm, and they were poor in aliphatic C (0–93 ppm) and carboxylic or carbonylic C (165–220 ppm), indicating that these biochars were mainly characterized by highly aromatic (aryl-dominated) structures, which was consistent with the FTIR spectra. Additionally, the low field shoulder at 152 ppm could be assigned to noncarbonized aromatic C in residual lignin (Sun et al., 2011). The aliphatic C of the biochars was ascribed to paraffinic C (0–45 ppm) (Table S1), supporting the appearance of absorption at 2925, 1430 and 1380 cm−1, as shown in the FTIR spectra (Fig. S2). Thus, it could be concluded that aliphatic structures associated with the biochars should be heat resistant under the charring condition used here. The XRD patterns of the biochars are shown in Fig. S2. Sharp, strong peaks in the ASB indicated that the inorganic components of ASB were mainly composed of quartz. These remarkable ASB peaks were consistent with the highest Si and ash contents in the swine biochars (Table 1). The elevated broad peaks between 2θ = 14–30° in wood and grass biochars were likely attributed to organic C (Cao and Harris, 2010). Furthermore, the formation of graphene sheets within turbostratic crystallites in the biochars was indicated by d-spacings

between 0.392 and 0.372 nm and between 0.209 and 0.207 nm, which were assigned to hkl 002 (Kercher and Nagle, 2003). Based on the classification of biochars according their XRD and FTIR characterization by Keiluweit et al. (2010), the biochar in this study would likely belong to composite biochar, which was made of turbostratic crystallites embedded in a low-density amorphous phase. The cumulative SA-CO 2 and pore volume of biochars are listed in Table 1 with their ranges of 162.0–601.6 m 2 g − 1 and 0.05– 0.17 cm 3 g − 1 , respectively. ASB exhibited a relatively smaller SA when compared with other biochars. The negative correlation between the SA and ash content and the positive relationship between the SA and OC content of biochars were observed for selected biochars (Fig. 1b and c), which indicated that the SA of biochars could represent the SA of OM rather than the SA of minerals within the biochars. 3.2. Single-solute sorption results The single-solute sorption isotherms of PHE and DBP onto biochars are displayed in Figs. S3 and S4, respectively. The FM fitted to the sorption data quite well with R2 N 0.975. The coefficients KF and n of the FM were different among different biochars for a given solute. All sorption isotherms were nonlinear with n b 1, indicating the heterogeneous energy distribution of sorption sites on the biochars. It should be noted that PHE exhibited more nonlinear sorption isotherms for any given sorbent than DBP (n: 0.41–0.54 for PHE and 0.72–0.83 for DBP) (Tables 2 and 3). OC-normalized concentrationspecific sorption coefficients (log KOC) were calculated at a low solute concentration (Ce = 0.005 Cs) (Tables 2 and 3). The ASB demonstrated the highest log KOC value of both chemicals. In addition, any sorbent tested in this study exhibited higher PHE sorption capacity than that of DBP, which was consistent with their different hydrophobicity values (K HW ). To screen out hydrophobic effects, KOC values were normalized by KHW (KOC / KHW) in this study (Tables 2 and 3). After KHW normalization, PHE still exhibited relatively higher KOC / KHW values than DBP for any given sorbent (Tables 2 and 3), suggesting that other specific sorbate–sorbent interactions, such as H-bonding and π–π interactions, may operate in addition to the hydrophobic effect. The biochars had a high aromaticity (≥ 86.8%); thus, π–π bonding was proposed as one of the major adsorption mechanisms. The spectral information and elemental composition presented here (Table 1 and Fig. S2) supported the formation of a large aromatic sheet and the preservation of electron withdrawal (O- or N-containing functional groups) by functional groups within the biochars. It was previously proposed that the π-donor solutes interact with π-acceptor

312

J. Jin et al. / Science of the Total Environment 473–474 (2014) 308–316

Polarity index (O+N)/C

a

0.5 0.4

Bulk polarity Surface polarity

0.4 0.4

R2=0.984, p