Biomass and Bioenergy 81 (2015) 304e314
Contents lists available at ScienceDirect
Biomass and Bioenergy journal homepage: http://www.elsevier.com/locate/biombioe
Research paper
Enhancing cation exchange capacity of chars through ozonation Matthew Smith a, Su Ha b, James E. Amonette c, Ian Dallmeyer d, Manuel Garcia-Perez a, * a
Department of Biological Systems Engineering, Washington State University, Pullman, WA 99164, USA Gene and Linda Voiland School of Chemical Engineering and Bioengineering, Pullman, WA 99164, USA c Fundamental & Computational Sciences Directorate, Pacific Northwest National Laboratory, Richland, WA 99352, USA d Composite Materials & Engineering Center, Washington State University, Pullman, WA 99164-2262, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 April 2015 Received in revised form 2 July 2015 Accepted 12 July 2015 Available online xxx
The use of ozone to increase the cation exchange capacity (CEC) of two chars produced from pyrolysis of Douglas fir (Pseudotsuga menziessii) and a control bituminous coal activated carbon (AC) is reported. Chars were produced from the wood fraction of Douglas fir (DFWC) and the bark (DFBC) at 500 C using an auger driven reactor with a nitrogen sweep gas under mild vacuum. Five ozone treatment times, ranging from 5 min to 60 min, were investigated. The initial properties of each char were found to differ significantly from the other samples in terms of surface area, proximate composition, and elemental composition. DFWC did not show significant mass loss or temperature variation during ozone treatment; however, after 1 h of oxidation both DFBC and AC samples resulted in 20% and 30% mass loss, respectively, and reactor temperatures in excess of 60 C. Analysis of the pore size distribution of each treatment shows that ozone treatment did not significantly affect small micropores after 30 min of treatment for any material, but did reduce the apparent surface area of mesopores. Increases in carboxylic groups were identified with ozone treatment and found to correlate strongly with changes in measured CEC. The formation of lactone was found to correlate positively with reactor temperature during oxidation. These results indicate that the properties of chars, including surface area, pore structure, and chemical composition, as well as reactor conditions strongly affect the ozone oxidation of chars. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Biochar Pseudotsuga menziessii Ozone Surface acidity Carboxylic groups Lactonic groups
1. Introduction Chars and activated carbon have a long history of use in adsorption processes. The work of Lehmann in the early 2000s on Terra Preta soils from the Amazonian Basin brought renewed focus on the potential use of chars for agronomic purposes [1,2]. Evidence suggests that char-like materials have substantially altered physical and chemical properties in these soils, leading to long lasting carbon storage and improved crop production. Reviews by Glaser et al. [3]. and by Blackwell et al. [4] show generally positive effects of char addition to soil, though some negative effects have been reported. The effectiveness of char remediation treatments were found to be sensitive to the rate of addition [5], soil type, and initial feedstock [6]. Despite a range of studies aimed at investigating the effectiveness of chars in soils, assessment of the physicochemical structure of the char used for treatment is often not explored,
* Corresponding author. Department of Biological Systems Engineering, LJ Smith, Room 205, PO Box 646120, Pullman, WA 99164-6120, USA. E-mail address:
[email protected] (M. Garcia-Perez). http://dx.doi.org/10.1016/j.biombioe.2015.07.012 0961-9534/© 2015 Elsevier Ltd. All rights reserved.
leaving numerous possibilities for the disparity in observed effects. Several physical and chemical properties are known to impact the behavior of char in soils, including surface area, porosity, and the quantity of acid exchange sites on the char in relation to the soil. Higher quantities of acidic exchange sites have been identified in Terra Preta soils as compared to adjacent soils [2]. These sites exist primarily in the form of weakly acidic carbon-oxygen groups that can chemically bind Lewis bases, such as metal or ammonium ions in solution. Based on this principle, oxidized chars and activated carbons have been investigated for the removal of ammonia from gas streams and ammonium from liquid streams [7e12]. Chiang et al. [12] has shown a strong correlation between the quantity of ammonia adsorbed by treated carbons and the concentration of acid groups on the surface. The development of carbon-oxygen groups on a freshly produced char surface can take months to years to develop under ambient conditions [13]. Research groups have investigated various methods to more rapidly oxidize carbon surfaces; mechanisms studied include strong acids [9], oxygen at elevated temperatures [14], and ozone [7,8,10,12e18]. While acid treatment methods are
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
effective, the cost of facilities, materials, and post-treatment of liquid effluents increases material handling costs and complexity. Formation of acid groups through gas phase reactions can limit secondary treatment requirements and can be achieved through the use of either air or ozone. While ozone requires special handling and disposal, reactions can be carried out under ambient conditions, whereas air oxidation requires elevated temperatures to progress at a substantive rate. In addition to studies with activated carbon, the reaction of moderate temperature wood chars with ozone has also been investigated [10,13]. Analysis of the changes in surface chemistry, and the relation of these changes to the cation exchange capacity (CEC) of these chars was not reported. Here the effects of ozone on the surface characteristics are examined through chemical and spectroscopic analysis to investigate how surface properties, including CEC, develop on chars during ozonation.
305
2.2. Ozonation Each feedstock was oxidized under ozone using approximately 3.0 g of sample per run in a 316L stainless steel (SS) packed-bed reactor with an internal diameter of 7.29 cm. Samples were supported on a 150 mm 304 SS screen. All connections and peripherals were constructed of 304 S S, 316 SS, or polytetrafluoroethylene material. Ozone was generated by a corona-discharge type system (Lab-11 ozone generator, Pacific Ozone Technology, Inc., Benecia, CA) using purified oxygen gas (minimum purity 99.5%). Five oxidation times were used for each material ranging from 5 min to 60 min (numerical suffix of 0e60 denotes ozone oxidation time for each sample), with all treatments run in triplicate. All tests were carried out at an average ozone concentration of 70 mg L1 in a 2.0 L min1 oxygen/ozone stream (all gas volumes reported at standard conditions of 298 K and 101.3 KPa). The temperature of the char within the packed bed was measured by a type-K thermocouple (Omega Engineering, Inc., Stamford, CT). The concentration of ozone was measured by an iodine wet chemistry method outlined by Rakness et al. [22].
2. Experiments and methods 2.3. Bulk properties 2.1. Chars Douglas fir (Pseudotsuga menziessii) wood (DFW) and bark (DFB), harvested from the Cascades region of Washington State has been provided by Herman Brothers Logging & Construction (Port Angeles, WA) from standard mill operations. The received chips were ground to below a 2 mm sieve size by hammer milling (model number 400 HD, serial number 2404, Bliss Industries, Inc). Ground samples were stored in polypropylene bags at 4 C in a dark environment. Composition of each sample is reported in Table 1 and was measured in duplicate following standard methods provided by the National Renewable Energy Lab [19,20]. Prior to pyrolysis samples were dried overnight at 105 C. Chars from the DFW and DFB samples (DFWC and DFBC, respectively) were produced in a 1 kg h1 auger-driven reactor at Washington State University following the methodology described elsewhere [21]. Briefly, the reactor's external wall temperature was maintained at 500 C with a particle residence time of approximately 1 min. Nitrogen was used as a carrier gas at 10 L min1 to ensure an oxygen-free atmosphere. Char temperature at the reactor exit was determined to be between 270 and 370 C. A vacuum pump maintained a slight negative pressure of 20 Pa (2 mm H2O) inside the reactor to ensure forward gas flow. Char samples were collected by gravity from the auger line. Visual inspection of the final product did not indicate the presence of residual unconverted biomass in the collected samples. Pyrolysis of the wood fraction resulted in a char yield of 20 mass %, while the char yield of the bark fraction was 24 mass % char on an oven dry basis. The resulting char products were stored in sealed polypropylene bags under ambient conditions in a dark environment until ozone oxidation treatments. A commercial activated carbon (AC) produced from bituminous coal was purchased from Sigma Aldrich (Sigma Aldrich 242268 Lot 33896CJ, manufactured July 2008) and used without modification in the ozone treatment experiments.
Ash content of the initial samples was determined in triplicate using a high-temperature furnace to determine the ash fraction following ASTM method D1102-84 [23]. 1.1 g ± 0.1 g of air-dry material was heated in air at 575 C for 12 h, and the ash fraction determined on an air-dry basis. Volatile matter and fixed carbon were determined in triplicate for each initial and ozone treated sample using thermogravimetric analysis (Mettler Toledo TGA/ SDTA851e) based on methods described in ASTM method E1131-08 [24]. Alumina crucibles were loaded with 6 mge7 mg of sample and heated under nitrogen (minimum purity 99%) flowing at 50 ml min1. The heating program used is as follows: from 25 C to 105 C at 10 K min1 with a 15 min hold, and from 105 C to 650 C at 30 K min1 with a 10 min hold. Mass loss at 105 C was attributed to moisture and highly volatile matter while mass lost between 105 C and 650 C was attributed to volatile matter. The solid fraction remaining is attributed to fixed carbon and ash. The moisture content determined by TGA was used to correct the ash fraction to an oven-dry basis. Elemental analysis for carbon, hydrogen and nitrogen was determined on an oven-dry basis and conducted by the University of Washington Oceanography Department using a Leeman Labs Model CEC440 elemental analyzer. A sample size of 2.5e4.0 mg was used. The equipment was calibrated using acetanilide and caffeine standards. Calibration was checked by a caffeine deviation of less than 1%. Oxygen was determined by difference. Mineral analysis was conducted using an ICP-MS (Agilent 7500cx series) equipped with a double-bypass quartz spray chamber with a quartz nebulizer. To prepare the analysis solutions, 0.1 ge0.2 g of each char and AC were pre-digested using 3.0 ml of concentrated NHO3 and 2.0 ml of 30 mass % H2O2 in a microwave digester (SP-D, CEM Corporation, Matthews, NC). Digestion was carried at out 300 C and 1.72 MPa for 5 min. A 5 min ramp was used to reach digestion conditions. Digest samples were diluted to
Table 1 Composition of initial biomass, all results reported as mass fraction (%). Sample
Hemi-cellulose
Cellulose
Lignin
Extractives
Insoluble ash
Total
DFB DFW
8.8 (0.2) 17.0 (0.1)
18.2 (0.3) 43.2 (0.2)
43.4 (1.3) 29.9 (1.5)
19.5 (0.6) 5.2 (ND)
1.65 (0.35) <0.1 (<0.1)
89.3 (0.8) 95.2 (1.8)
ND: Results from single trial, no deviation could be determined.
306
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
100 ml using deionized (DI) water with a minimum resistance of 18 MU-cm (Barnstead e-pure system model D4641)water and a 100 mg L1 multi-elemental internal standard of 6Li, Sc, Ge, Y, In, Tb and Bi (Accustandard, Inc., New Haven, CT). The ICP-MS was operated between 1580 We1600 W for multi-element analysis and at 1300 W for P analysis. An Ar carrier gas was supplied at 0.9 L min1 for the nebulization gas and 0.25 L min1 for the makeup gas. In addition, P analysis employed a collision and reaction cell operation with a flow of 2 ml min1 He to reduce interference from diatoms. All other parameters were adjusted to obtain good sensitivity for Li6þ, Y, and Tl at a 1 mg L1 concentration. Eight point calibration curves were prepared from a multi-element environmental standard analysis (Accustandard, Inc., New Haven, CT) and ranged from 100 mg L1 to 100 mg L1 for Na, K, Mg, Ca and Fe and from 1 mg L1 to 1 mg L1 for all other elements. All standards and samples contained the same 100 mg L1 internal standard as the samples. A minimum r2 of 0.995 was obtained for all elements reported. Elements not reported did not pass the minimum quantification limit for any sample. 2.4. Gas physisorption analysis Specific surface area was determined by N2 (g) and CO2 (g) adsorption. Approximately 0.10 g of sample was degassed for a minimum of 8 h at 200 C. Higher degas temperatures were avoided to minimize possible volatilization of carboxylic groups. Prior to degassing, samples were washed with methanol via a Soxhlet extraction system for 4 h at rate of 4 siphons per hour to remove adhered polar compounds. These compounds otherwise require degassing temperatures of 300 Ce350 C to remove, and were found to inhibit collection of stable isotherms when lower degassing temperatures were used. Isotherms were collected using a TriStar II plus automatic physisorption analyzer (Micromeritics Instrument Corporation, Norcross, GA) at partial pressure ranges of 0.0001e0.99 and 0.00001e0.03 for N2(g) and CO2(g) respectively. Apparent surface area (SA), micropore volume (Vmi), and average pore width (Wavg) were determined by the DubinineRadushkevich (DR) equation fit to data points between 2 < log (P0 P1)2 <6 for both N2 and CO2 analysis. Average pore width has been calculated from the characteristic energy given by the DR equation through Equation (1) [25]. Total pore volume (VT) was determined from the maximum adsorption quantity at a partial pressure of approximately 0.99 for N2 isotherms. Total mesopore volume (Vme) was determined by subtracting the micropore volume determined by N2 from VT. Pore-size distribution was determined by a nonlocalized density functional theory (NLDFT) method available with the Micromeritics MicroActive™ software.
Lo ðnmÞ ¼ 10:8 ðnm$kJ=molÞ=ðEo 11:4 ðkJ=molÞÞ
determined by plots of pH against the solid mass fraction for each NaNO3 solution (pH 3, 6, 11). The pH of convergence at infinite mass fraction is estimated from these plots and represents the pHPZC [26,27], an error of 0.25 pH has been assumed for the extrapolation. A representative plot is shown in Fig. 1 detailing the convergence behavior the pH for increasing concentration of DFWC-0 in solution. The surface chemistry of select treatments were evaluated by Xray photoelectron spectroscopy (XPS). Spectra were obtained with an AXIS-165 manufactured by Kratos Analytical Inc. (Spring Valley, NY, USA) using an achromatic X-ray radiation of 1253.6 eV (MgKa). A pass energy of 40 eV and spot size of approximately 120 mm were used to acquire all spectra. The spectrometer was calibrated against both the Au 4f7/2 peak at 84.0 eV and the Ag 3d5/2 peak at 368.3 eV. Static charging, when present, was corrected with a neutralizer (flood gun) and the spectral energy calibrated by placing the primary low-energy carbon peak (C1s) at between 284 and 285 eV. Baseline corrections, peak areas and deconvolutions were analyzed using XPSpeak 4.1 [28]. A Shirley type background correction was used in all cases. Gaussian peaks with a full width at half maximum (FWHM) of 2 eV were used to deconvolute the C1s peak with the exception of the first CeC peak which was allowed to vary between 1 eV and 2 eV, and between Gaussian and Lorentzian to best match in peak shape. XPS analysis parameters are listed in Table 2, including binding energy center and relative intensity. The CeC/ CeH peak has been assumed to develop from two distinct peaks, a primary peak composed of sp2 bonded carbon centered at approximately 284.4 and a secondary peak, relating to sp3 oriented carbons as discussed by Jackson and Nuzzo [29] all other parameters are from Moulder et al. [30]. Boehm titration [14] was conducted in triplicate to determine the type and content of acidic groups on the surface of each char. Prior to analysis, samples were washed with 0.1 M HCl. Here a 10 ml:1 g solution: char mixture was shaken at 3.3 Hz for 30 min to remove acid soluble material that may otherwise interfere with the Boehm bases and to fully protonate the acid groups present. Acidwashed (AW suffix) samples were rinsed with DI water until chlorine could not be detected by precipitate formation with a 0.01 M solution of AgNO$3 . Rinsed samples were dried at 105 C for 8e24 h 10 ml of 0.020 mol L1 solutions of NaOH, NaHCO3, and Na2CO3 were combined with 100 mg of each char and AC and agitated for 48 h at 3.3 Hz. Solutions were filtered using #42
(1)
where: Lo is average pore width and Eo is average energy of adsorption from the DR equation. 2.5. Surface chemistry Mass titration, as described by Noh and Schultz [26,27], was used to determine the pH at the point of zero charge (pHPZC). Briefly, three aqueous 0.01 mol L1 NaNO3 solutions were prepared and adjusted to a pH 3, 6 or 11 using NaOH and HNO3. A series of five char solutions were prepared at (0.1e5.0) mass % solids using each of the pH adjusted NaNO3 solutions. Solutions were agitated for 48 h on an orbital shaker at 3.3 Hz and allowed to settle for 6 h. The supernatant was filtered using Whatman #42 paper and the pH of the filtrate determined using a pH probe (S20 SevenEasy™ pH meter, Mettler-Toledo LLC, Columbus, OH). The pHPZC was
Fig. 1. Effect of char concentration in NaNO3 solutions of varying initial pH. The solid gray dashed line represents the projected pH at an infinite char ratio, determined to be the pHPZC, while the dashed gray lines represent a 0.25 unit uncertainty. DFWC0 shown as an example.
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314 Table 2 XPS binding energies and sensitivity factors for elements and chemical structures identified. Element peak chemical structure
Binding energy (eV)
Sensitivity factor
C1s CeC/CeH CeC/CeH (sp3)a CeO C¼O OeC]O piepi* O1s Pbf (doublet) Si2p Si2s Cl2p K2p (doublet) Ca2p (doublet) In3d5/2
284e292 284.4e284.6 285.2e285.4 285.8e286.2 287.2e287.6 288.8e289.2 290.2e290.6 530e535 137&142 99 151 201 294 & 297 347 & 351 444
0.296
a
0.711 6.968 0.283 NA 0.891 1.466 1.634 4.359
Jackson and Nuzzo [29].
Whatman paper. After filtration, 1.0 ml of supernatant was mixed with 2.0 ml of standardized 0.020 mol L1 HCl (3 ml for Na2CO3 solutions), and diluted to 100 ml using DI water. The diluted solutions were back-titrated using freshly standardized 1.0 mmol L1 NaOH to a pH of 4 to minimize potential interference from dissolved CO2. Neutralization was determined by comparison to a blank Boehm solution. A TukeyeKramer pairwise comparison test with a ¼ 0.10 was used to determine if measured changes were significant.
2.6. Cation exchange capacity The CEC of each char and AC was determined in triplicate by BaCl2 exchange followed by compulsive MgSO4 exchange as outlined by Hendershot and Duquette [31]. Prior to CEC determination, 0.5 g of each sample was acid washed with 30 ml of 0.10 mol L1 HCl for one hour and rinsed with DI water to a pH in excess of 5. During BaCl2 saturation the pH was monitored and adjusted to pH to 5.3 ± 0.3 using dilute solutions of NaOH and HCl. Saturated samples were strained and dried at 105 C for 6 h. 100-mg subsamples of each char and AC were mixed with 10 ml of 0.01 mol L1 MgSO4, and agitated for 1.0 h to precipitate adsorbed Ba2þ as BaSO4. The concentration of Mg2þ in each sample solution was determined by atomic adsorption spectroscopy (AAS) (Varian Spectra AA 220) following a standard method [32] modified to use a nitrous-oxide flame. 1.0 mg L1 of KCl was added to each solution as a readily ionizable element to minimize potential matrix interference. Solutions were diluted with 2 mass % nitric acid in DI water to bring samples into the measurement range. The CEC was determined from the quantity of Mg2þ adsorbed by the sample during the precipitation of BaSO4. Significant differences were determined by a TukeyeKramer pairwise comparison at a ¼ 0.1.
307
3. Results and discussion 3.1. Chars studied Table 3 shows the elemental analysis of each sample studied as well as the composition of the ash. DFWC was found to have approximately twice the quantity of hydrogen as the DFBC sample, suggesting a substantially lower degree of aromatic condensation. Oxygen content was found to be only marginally higher for DFWC than DFBC. By comparison the AC sample shows only minor noncarbon impurities beyond the ash fraction, with the carbon structure accounting for 98.2 mass % of the initial mass on a dry ash free (daf) basis. The hydrogen fraction was 0.6 mass %, indicating a highly condensed carbon structure with an atomic C:H ratio greater than 13. The mineral composition of each sample varied substantially. Ca was the major constituent of DFBC ash, approximately 3 mass % of the total char. Mg, K, Al, and Fe were also important constituents, contributing (0.3e0.5) mass % each. AC contained a considerable fraction of Al, approximately 0.8 mass %. Both AC and DFBC contained visible fraction of silicates not soluble in nitric acid, however these have not been quantified. DFWC contains only small quantities of mineral impurities, never more than 0.2 mass% for any element. In addition to these major elements several others were detected in mg kg1 quantities, including: Cr, Ni, Cu, Ba and Pb. Of these, only Ba were found in concentrations above 100 mg kg1 for any sample. Mn was also detected by ICP-MS, however strong interference from 39K16Oþ preclude accurate assessment [33]. The mass of detected minerals are reported on an elemental basis rather than an oxidized basis and are not directly comparable to the ash fraction presented in Table 2. It should be noted that due to the oxidization of minerals during ashing by combustion, the total ash fraction could be somewhat overestimated. The textual properties of each material differed substantially as seen in the adsorption isotherms for N2, given as solid lines in Fig. 2A for AC and Fig. 2B for DFBC and DFWC. The corresponding CO2 isotherms also show significant variance, solid lines in Fig. 2C. DFWC and DFBC show moderate apparent surface area under CO2, 192 and 333 m2 g1 respectively. N2 isotherms however, resulted in no apparent surface area for DFWC and a moderately reduced apparent surface area for DFBC. The narrow average pore diameter determined from CO2 isotherms, 0.82 nm for DFWC and 0.91 nm for DFBC suggest restricted gas diffusion at 77 K. This inhibition was sufficient to prohibit complete equilibration in the 40 h timeframe allowed for each isotherm, as demonstrated by the failure of the N2 desorption isotherm of DFBC to converge with the adsorption isotherm at low partial pressures (see Fig. 2B). Though not fully equilibrated, the shape of the isotherms for DFWC and DFBC reveals that these chars are either non-porous as seen in the case of DFWC (Type III isotherm) or predominantly microporous as in the case of DFBC (Type I isotherm). The increased average pore width determined by N2 adsorption, 1.30 nm, is likely a result of constrained access to the smallest micropores. In contrast to the chars, the surface area of AC is considerably higher when determined by N2 than by CO2, 669 m2 g1 compared to 433 m2 g1. This indicates a
Table 3 Chemical composition of initial chars. Sample
DFWC DFBC AC
Macro elements (g kg1 ±0.1)
Elemental analysis (mass %)
Micro elements (mg kg1 ± 5)
C
H
N
O
Ash
Na
Mg
K
Ca
Al
Fe
P
Cr
Mn
Ni
Cu
Zn
Ba
72.4 64.6 84.0
3.9 1.6 0.5
0.2 0.3 0.7
21.4 15.5 0.3
2.0 17.6 14.1
0.5 0.6 2.8
0.4 4.6 0.8
1.5 4.8 0.3
2.0 29.8 2.0
0.2 5.0 8.2
0.3 3.9 1.9
0.2 1.5 0.1
ND 84 12
113 692 13
ND 12 14
49 69 44
12 83 8
8 385 740
308
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
3.2. Reactions with ozone The effect of ozonation on mass loss and reactor temperature is shown in Fig. 3A and B respectively. Mass loss on exposure, as well as the measured reactor temperature varied considerably between samples, showing distinct reaction rates not only between char and activated carbon, but also between chars of different origin. Minimal effects were noted during treatment of DFWC while treatment of DFBC resulted in a significant increase of the reactor temperature as well as oxidative loss of sample. Changes observed for DFBC were largely similar to those recorded for AC during treatment. The reaction of both AC and DFBC resulted in significant reaction of ozone
Fig. 2. (A) Temperature variation of the packed bed reactor, (B) mass fraction remaining after oxidation, and (C) ozone consumed during reaction. Filled triangles e DFWC, open circles e DFBC, filled circles e AC.
wider pore size distribution with pore entrance widths extending beyond the micropore range (>2 nm), as also indicated by the strong hysteresis observed above partial pressures of 0.4 (see Fig. 2A). Calculated pore size distributions for AC indicate that approximately 60% of the total measured pore volume exists as mesopores. The isotherm of AC also does not plateau near partial pressures of 1, suggesting that the AC sample also has a significant fraction of macropores (entrance width > 50 nm).
Fig. 3. N2 adsorption isotherms for (A) untreated and oxidized AC and (B) untreated and oxidized DFWC and DFBC; respective CO2 isotherms for each sample (C). Note that CO2 isotherms for ACs have been shifted by 20 cm3 g1 to avoid convolution with DFBC curves. Solid lines e untreated, dashed lines e 60 min treatment.
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
in the gas stream, whereas no measurable reaction of ozone with DFWC was seen. Fig. 3C shows the quantity of ozone reacted for AC and DFBC, where an initial drop in reactivity from 1.0 min to 5.0 min of treatment was noted for both carbons. While the apparent reaction rate for DFBC stabilized after 5 min, the reaction rate of AC samples continued to decrease until 10 min of treatment, after which the rate stabilizes. Table 4 shows the effect of oxidation on the proximate and elemental analysis of the samples studied. Ozone oxidation was not found to have a significant effect on the overall H:C ratio of DFWC. A minor increase with treatment can be seen for DFBC, while AC shows a 1.3e1.8 fold increase in the H:C ratio with ozone treatment. The increase of the H:C ratio indicates that either carbon dioxide or carbon monoxide are formed preferentially to water vapor. Significant decreases in the FC/VM ratio with treatment for both DFBC and AC were also identified, suggesting that the condensed aromatic structure was largely degraded during treatment. While accurate studies of the reaction mechanisms and kinetics of ozone with complex macrostructures such as char and activated carbon are limited, the literature pertaining to soots and polyaromatic hydrocarbons provide insights into the behavior of similar polycondensed carbon species with ozone. From studies of the interaction of dust and soot with atmospheric ozone, both the available reactive surface and the mineral support have been found to affect reactivity [34e36]. It is noted that DFWC, which did not show strong reactivity during treatment, had much lower pore volume and mineral content as compared to the other materials studied. While the effect of mineral matter cannot be discounted based on the materials studies, the large variation in mineral speciation between DFBC and AC suggests that the more important factor leading to high reactivity is the presence of large micropores and mesopores in the condensed carbon structure, though further investigation is required to confirm this hypothesis. The rapid initial reaction of ozone observed for both DFBC and AC samples, followed by the relatively stable reaction regime that follows, indicates at least two mechanisms control the reaction rate. The first, most probably, being a rapid adsorption process and the second being either a slower degradation process, a result of diffusion limitations within the micropore structure, or a combination of both effects. Such a mechanism is in agreement with the results of Smith and Chugthai [37]. Their investigation of the reaction of hexane soot with ozone identified a rapid adsorption process during the first three minutes of contact followed by the
309
formation of carboxylic groups on the surface of the soot particles and CO2 from the surface. Other studies have identified that ozone preferentially favor alkene and aromatic bonds through the formation of a primary ozonide followed by the formation of a C*HOO* biradical (Criegee intermediate) [36,38]. This intermediate can then follow one of several reaction pathways to form surface oxides and gas phase compounds. The rates of these reactions likely control the overall reaction rate in the second regime. A variety of potential pathways for the degradation of the Criegee intermediate is discussed by Yu et al. [38]. in relation to the gas phase reaction of ozone with monoterpenes.
3.3. Effects of ozone oxidation on surface morphology The effect of ozone oxidation on the microporous surface area of each sample is summarized graphically in Fig. 4. Because increasing the CEC of a material requires a stable surface for exchange sites to develop, minimizing the loss of surface area during treatment is essential to maintaining high surface acidity. Example isotherms for ozone treated materials are presented as dashed lines in Fig. 2. The surface area and pore volume of AC decreases marginally over the first 30 min of treatment, with a reduction of approximately 7
Fig. 4. Effect of oxidation time on microporous surface area as determined by physisorption of CO2, * denotes anomalous high reading for DFWC-30.
Table 4 Proximate and elemental analysis of treated samples. Sample
Moisture (mass %)
FC/VM (mass ratio)
H:C atomic ratio
Oa:C atomic ratio
AC-0 AC-5 AC-10 AC-20 AC-30 AC-60 DFBC-0 DFBC-5 DFBC-10 DFBC-20 DFBC-30 DFBC-60 DFWC-0 DFWC-5 DFWC-10 DFWC-20 DFWC-30 DFWC-60
2.4 3.6 3.1 3.3 3.6 5.5 2.3 2.2 2.2 2.1 2.1 2 2.8 2.9 3.3 3.3 2.6 2.6
25.3 10.8 9.4 7.5 5.8 2.5 5 4 4.1 3.4 3.6 2.8 2.3 2.2 2.2 2.1 2.1 2
0.14 0.19 0.2 0.18 0.21 0.25 0.29 0.33 0.35 0.33 0.32 0.34 0.64 0.68 0.64 0.63 0.66 0.69
<0.01 0.04 0.05 0.41 0.16 0.29 0.18 0.12 0.15 0.09 0.23 0.13 0.21 0.23 0.22 0.21 0.22 0.25
a
Determined by difference.
310
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
and 4% noted for N2 and CO2 respectively. These losses increased significantly to 24% and 18% when exposure was increased to 60 min. Despite significant changes in overall surface area the average pore diameter did not vary considerably for either probe molecule. CO2 isotherms for both DFWC and DFBC show little change with treatment. Surface area, pore volume and pore width varied by less than 8% from the untreated sample. N2 isotherms were also collected for DFBC and DFWC samples but the total quantity absorbed remained low when compared to CO2 adsorption, indicating minimal formation of mesopores in either sample. Total N2 adsorption decreased continuously with ozone treatment for DFBC samples despite stable micropore structures observed by CO2, suggesting partial collapse of the mesopores. Full results for each carbon are presented in Table 5. NLDFT was applied to the N2 and CO2 isotherms in order to more fully explore the effect of treatment on pore size distribution. Fig. 5 shows the effect of treatment on pore size distribution for AC samples. The pore distribution between 2 and 5 nm, is largely stable, while those distributed between 1 and 2 nm are the primary pores affected by ozone treatment. An initial decrease with treatment is observed in pores of 1 nme1.2 nm and 1.8 nme2.0 nm, accompanied by a sharp increase in sub nanometer pores. With further oxidation, pores of 1.3 nme1.5 nm begin to constrict, as do the sub nanometer pores. Such behavior suggests that chemisorption of ozone or oxygenated species in the micropores is occurring, filling the pore space and resulting in pore constriction. The distribution of pore space in pores of approximately 5 nme50 nm is continuously reduced during treatment, likely due to reaction and stripping of the surface. After 60 min of treatment, the number of all pores below 50 nm was strongly reduced. As smaller pores did not increase in line with the loss of mesopores, it is suspected that macropore formation was the primary origin of loss. Analysis of changes in pore structure further indicates that the initial rapid reaction involves oxidation of the mesopores. The relatively large size of these pores allows the initial reaction to occur under a kinetically controlled regime. As these surfaces are oxidized and become less reactive, diffusion of ozone into the micropores becomes possible, forcing the reaction to proceed under diffusion limitations. Rapid initial reaction within mesopores is a likely cause of the heat generation noted in the treatment of both DFBC and activated carbon (see Fig. 3A).
Fig. 5. Effect of ozone treatment on pore size distribution of activated carbon, nearest plot AC-60, furthest AC-0. Treatment results in continuous decrease of pores of 1e1.5 nm while an initial increase, followed by gradual reduction is observed for those of ~0.5 nm.
3.4. Generation of an oxygenated surface A series of wide scan spectra for ozone treated DFBCs are shown in Fig. 6. An increasing prevalence of Ca can be seen in DFBC samples as the oxidation time is increased. The concentration of Ca and other minerals at the surface is a result of the loss of the carbon matrix due to volatilization. Acid washing was noted to reduce the silica content of AC, and fully remove Ca from DFBC, confirming that many of these species are loosely adhered surface deposits. Analysis also indicates that the rinse procedure used was sufficient to remove all but trace quantities of chlorine from the surface. Interestingly, trace Pb contamination was identified on DFBC samples after acid washing. This contamination is most likely a result of trace levels present in the HCl used, however such contamination was not identified on either AC or DFWC (figures not shown). A summary of the results obtained by XPS is presented in Table 6. These results reveal the generation of carboxylic groups on all samples studied. Example C1s spectra are presented in Fig. 7. Beyond the generation of carboxylic groups, only minimal changes are identified in AC samples with oxidation, Fig. 7A. Moderate increase in other oxygenated groups are seen in DFWC samples however, Fig. 7B. Ozone treatment of DFBC also resulted in
Table 5 Effect of ozone treatment on surface morphology. Sample
AC-0 AC-5 AC-10 AC-20 AC-30 AC-60 DFBC-0 DFBC-5 DFBC-10 DFBC-20 DFBC-30 DFBC-60 DFWC-0 DFWC-5 DFWC-10 DFWC-20 DFWC-30 DFWC-60
CO2 isotherm results
N2 isotherm results
SA (m2)
Vmi (cm3/g)
Wavg (nm)
SA (m2)
VT (cm3/g)
Vmi (cm3/g)
Vme (cm3/g)
Wavg (nm)
433 439 417 419 414 352 333 346 338 336 349 345 192 200 199 197 290 197
0.17 0.18 0.17 0.17 0.17 0.14 0.13 0.14 0.14 0.14 0.14 0.14 0.08 0.08 0.08 0.08 0.12 0.08
1.06 1.04 1.03 1.02 1.03 1.00 0.82 0.80 0.80 0.78 0.79 0.79 0.91 0.90 0.92 0.92 0.86 0.90
669 663 629 623 619 509 314 276 269 256 237 147 <1 <1 <1 <1 <1 <1
0.65 0.63 0.61 0.60 0.61 0.51 0.15 ND 0.14 0.12 0.12 0.09 0.03 0.03 0.03 0.03 0.02 0.02
0.24 0.24 0.22 0.22 0.22 0.18 0.11 0.10 0.10 0.09 0.08 0.05 ND ND ND ND ND ND
0.41 0.39 0.39 0.38 0.39 0.33 0.04 ND 0.04 0.03 0.04 0.04 ND ND ND ND ND ND
1.60 1.70 1.55 1.53 1.61 1.75 1.37 3.78 2.11 2.05 2.30 1.33 ND ND ND ND ND ND
ND: Could not be determined from collected isotherms.
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
311
Fig. 6. XPS wide scan results for DFBC oxidation time increases from 0 to 60 min moving up the graph. Bold series at the top is acid washed DFBC-60. Treatment results in sharp increase in oxygen peak, with simultaneous reduction and broadening of the carbon peak. A strong increase in calcium is noted on the surface after prolonged exposure to ozone, mild acid washing successfully remove this material.
significant increases in all oxygenated groups, with very large increases in the range typically assigned to carbonyl and carboxyl groups, Fig. 5C. The large increases in carboxyl and carbonyl like groups noted for DFBC were found to be related to adsorbed or acid soluble compounds on the surface. Mild acid washing was found to strongly reduce this region of the spectra, Fig. 7D. Subtraction of the spectra collected for DFBC-60-AW from DFBC-60 residual spectra containing 2 broad peaks centred at 289.3 and 292.8 eV, also shown in Fig. 7D. The FWHM of 4 and 3.4 are significantly larger than expected for individual chemical species (approximately 2 eV), suggesting that these peaks are the result of distributions of several related compounds. Even when factoring these adsorbed peaks into the deconvolution, significant increases in carbonyl and carboxyl groups are noted during the oxidation of DFBC. The relatively higher concentration of carbonyl groups on DFBC may be responsible for the Pb adsorption seen in the wide scan analysis. Malik et al. [39] have hypothesized that these groups may be more selective towards Pb due to interactions between the lone electron
pair of the carbonyl groups and the Pb ion's easily polarized electron shell. To verify the validity of the C1s deconvolution calculated O:C ratios were compared with the O:C ratio obtained via the wide scan spectra. Calculations were based on generalized assumptions of origin, namely that peak 3 results predominately from ether type bonds, resulting in a carbon to oxygen ratio of 2:1 (CeOeC), peak 4 results for carbonyl groups with a ratio of 1:1 (C]O) and peak 5 is the result of equal parts carboxyl and lactone with ratios of 1:2 and 1:1 respectively (OeC]O and CeOeC]O), this is summarized in Equation (2). While these assumptions are not exact, they allow for basic comparison and validation of deconvolutions. The residual peaks on DFBC were assumed to contain carbon to oxygen ratios of 2:1 and 1:1 respectively to maintain reasonable agreement with the C:O distribution determined by wide scan analysis. The results of comparison yield generally good agreement between observed and theoretical oxygen content. All results are listed with the wide scan and C1s deconvolution data in Table 4.
Table 6 Elemental composition and carbon structures as determined by XPS. Sample
Wide scan elemental analysis atomic fraction (%) C
O
N
Si
Ca
CeC/CeH
CeC sp3
CeO
C¼O
COO
Wide scan
C1s calc.
AC-0 AC-60 AC-60-AWa DFBC-0 DFBC-60b DFBC-60-AWa,c DFWC-0 DFWC-60 DFWC-60-AWa
89 79 79 90 65 69 83 73 74
7.4 17 18 10 34 29 17 27 26
2.4 ND ND ND ND ND ND ND ND
1.2 3.3 2.2 ND ND ND ND ND ND
ND ND ND 0.5 1.7 ND ND ND ND
69 61 51 71 20 64 44 36 49
15 16 29 13 ND ND 29 20 19
9.8 11 6.7 11 13 12 18 24 16
5.3 6.5 6.4 4.2 6.9 15 5.5 10 9.4
0.8 5.5 6.7 1.9 6.4 9.1 3.9 10 6.6
0.08 0.22 0.23 0.11 0.52 0.42 0.2 0.37 0.35
0.12 0.22 0.22 0.13 0.56 0.37 0.21 0.40 0.28
a
C1s deconvolution atomic fraction (%)
C:O ratios
Less than 1 atomic % chlorine was detected on surface from HCl wash. C1s deconvolution determined by addition of acid washed signal and 2 residual curves at 289.3 (FWHM ¼ 4) and 292.8 eV (FWHM ¼ 3). Residuals contributed 46% and 8% of C1s signal respectively. Species were assumed to have C:O ratios of 2:1 and 1:1 respectively to obtain similar overall C:O ratio as obtained from the wide scan relation. c 0.3% lead detected on surface, present as trace impurity in HCl. b
312
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
Fig. 7. Comparison of untreated and 60 min ozone treated samples; normalized to CeC/CeH peak and corrected for background. Solid line-untreated, dashed line-treated, dotted lines-best fit curves from deconvolution of untreated samples (A) AC (B) DFWC (C) DFBC. (D) Comparison of DFBC-60 before and after acid washing; normalized to the CeC/CeH peak and corrected for background. Bold dashed line e unwashed DFBC-60, solid line e acid washed DFBC-60, gray line e residual from subtraction of acid washed spectra from unwashed spectra. dotted lines Gaussian fit of residual peak: 1st residual at 289.3 FWHM of 4, 2nd residual at 292.8 FWHM of 3.
, O : C¥ðpeak 3Þ*0:5 þ ðpeak 4Þ þ ðpeak 5Þ*1:5
5 X
peak ðnÞ
n¼1
(2) The increase in oxidized groups noted in XPS were found to have a dramatic effect on the point of zero charge (pHPZC) as shown in Table 7. This point describes the pH at which the net charge of the surface of the char is neutral. At more acidic pH the surface becomes positively charged, favoring cation exchange, while at more basic pH the surface is negatively charged, favoring anion exchange. It is important to note however that due to the high heterogeneity of chars and the dispersion of oxygenated groups and minerals some regions will be positively charged and others negatively charged at most intermediate pH values. The formation of acidic groups can be seen to strongly influence the pHPZC for both the AC and DFWC samples. The decrease observed after 5 min is striking, but only minor to moderate changes occur over the course of the remaining hour of treatment (Table 7). In contrast to DFWC and AC, DFBC shows only very minor changes pHPZC despite significant increases in surface acid groups formed. This effect is the result of the accumulated of Ca and other minerals, most likely as oxides or carbonates, providing basic constituents that act as a pH buffer in solution. By contrast both the AC and DFWC samples showed strong decreases in pHpzc with treatment. Boehm titration was performed to quantify the potential acid groups detected by XPS; the results are detailed in Table 7. While moderate errors were observed between replicates, statistically
significant changes can be identified as a result of oxidation. Both DFBC and AC show large increases in lactone groups at treatment times in excess of 10 min. For DFBC, a sharp initial increase in carboxylic acid was observed. Subsequent increases at treatment times longer than 5 min were not significant. Lactone groups, however, were found to increase steadily for treatment times of up to 20 min before approaching a plateau. Both carboxyl and lactone groups developed similarly during oxidation of AC. Here, lactone group form at only a slightly lower rate than carboxyl groups. The delay in lactone development may be a result of the low initial temperature of the reaction chamber during treatment. Comparing recorded surface groups with the surface area from CO2 yields average site densities ranging from 0 to 2 mmol m2 for each group, in excellent agreement with results from Biniak et al. [40] for nitricacid-oxidized activated carbons. A strong positive correlation between reactor temperature and the formation of lactone groups is shown in Fig. 8. These results are in agreement with the work of Alvarez et al. [17] where it was noted that ozone oxidation temperatures of 100 C resulted in an approximately equivalent distribution of carboxyl and lactonic groups. In contrast this same work demonstrated that ozone oxidation at room temperature strongly favored the development of carboxylic groups. Also in agreement with the work of Alvarez et al. [17], phenolic groups were not found to increase significantly with ozone treatment regardless of reactor temperature. Oxidation was found to result in increased cation exchange capacity (CEC) for each material, as shown in Fig. 9A. The CEC of AC was found to increase with ozone treatment, from near zero to
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
313
Table 7 Effect of ozonation time (0e60 min) on the pHPZC and quantity of carboxyl, lactonic, and phenolic groups as identified by Boehm titration. Sample
pHPZC (±0.25)
Functional groups m mol/g (m mol/m2) Carboxyl
DFWC-0 DFWC-5 DFWC-10 DFWC-20 DFWC-30 DFWC-60 DFBC-02 DFBC-52 DFBC-102 DFBC-202 DFBC-302 DFBC-602 AC-0 AC-5 AC-10 AC-20 AC-30 AC-60
4.45 3.72 3.4 3.37 3.35 3.15 7.99 7.37 6.99 6.74 6.48 6.34 8.75 4.75 3.9 3.1 2.5 2.03
a
100 160a,b 170a,b 180a,b 200b 160a,b 260a 390b 400b 390b 450b excluded 110a 210a,b 250b,c 300c,d 350d1 470e
Lactonic calculated (0.52) (0.83) (0.85) (0.91) (0.69) (0.81) (0.78) (1.13) (1.18) (1.16) (1.29) excluded (0.25) (0.48) (0.60) (0.72) (0.85) (1.33)
a
230 200a 240a 240a 220a1 230a 100a 190a 250b,c 340c 220a,b 230a,b 110a 200a,b 240a,b 300b,c 290b,c 370c
(1.20) (1.00) (1.21) (1.22) (0.76) (1.17) (0.30) (0.55) (0.74) (1.01) (0.63) (0.67) (0.25) (0.46) (0.58) (0.72) (0.70) (1.05)
Phenolic calculated a
320 380a 290a 310a 330a1 370a 300a 260a,b 210b 190b 220b 920d 10a 30a 40a 40a* 70a 50a
(1.66) (1.90) (1.45) (1.58) (1.68) (1.86) (0.90) (0.75) (0.62) (0.57) (0.63) (2.67) (0.02) (0.07) (0.10) (0.10) (0.17) (0.14)
Total acid 650a 740a 690a 730a 760a1 760a 660a 830b 860b,c 920d 890c,d
(3.38) (3.73) (3.51) (3.71) (3.13) (3.84) (1.98) (2.43) (2.54) (2.74) (2.55)
230a 440b 530c 640d1 680d 900e
(0.52) (1.01) (1.28) (1.54) (1.72) (2.52)
1: one replicate was excluded from analysis based upon a standard residual in excess of 2.0. 2: Boehm titrations results for DFBC samples are the result of two replicates. aee: values with the same letter show no significant difference by TukeyeKramer pairwise comparison, a ¼ 0.10.
335 mmoleq kg1. Lower values were observed for DFBC and DFWC, with maximums of approximately 200 and 70 mmoleq kg1 respectively. When these results are compared to the increase in carboxylic groups for each sample, the change in CEC was found to correlate very strongly (Fig. 9B). Here, a linear regression analysis, with the origin forced, yields a nearly perfectly linear correlation of 0.98 with an r2 of 0.94. That the increase of CEC is correlated with the increase in carboxylic groups rather than the total number of carboxylic groups detected by titration suggests that some interfering species are still present in the char after acid washing and yield an elevated apparent carboxylic group concentration. This may be explained by the relatively mild washing procedure utilized, which, while effective for the removal of acid soluble compounds, was insufficient to remove alkali soluble compounds which may also interfere with the back titration [41,42]. Despite the associated errors, the strong and consistent trends observed for all samples suggests that any interferences were largely consistent between ozone oxidation treatments and do not affect the final conclusions. Use of a combination of preliminary acid and alkali washing has been shown to remove most sources of contamination equilibration in the Boehm solutions [41]. Use of a combination of N2 sparging
Fig. 8. Correlation between content of Lactone groups and oxidation temperature (r2 ¼ 0.89).
Fig. 9. (a) Effect of ozonation on the CEC of each material (b) regression analysis of CEC compared with changes in carboxyl groups (r2 ¼ 0.93 and slope ¼ 0.99 with the origin forced).
314
M. Smith et al. / Biomass and Bioenergy 81 (2015) 304e314
and barium precipitation of the final solutions prior to analysis, as recommended by Fidel et al. [42], has also been shown to be effect in removing dissolved organic carbon after equilibration with the Boehm bases, removing interferences during back titration. Use of either method in future studies should reduce matrix interference with the Boehm bases. 4. Conclusions The behavior of various carbons under ozone show markedly different trends. DFWC, with by far the lowest ash and mineral matter content and lowest initial surface area was found to have a considerably slower oxidation rate than either DFBC or AC with no meaningful mass loss recorded. DFBC and AC reacted vigorously with ozone, resulting in significant temperature increases during oxidation. This effect is hypothesized to be linked to rapid oxidation within the mesopores and larger micropores of the material. Despite strong variances in oxidation rate, significant increases in oxygenated functional groups were identified by both chemical and spectroscopic means for all materials. These functional groups resulted in a decrease in the pHPZC for all materials, though reduction was tempered for DFBC due to accumulation of ash, exposed during oxidation. Increases in carboxylic groups, as detected by Boehm titration, were noted to correlate strongly with the measured CEC of each material. Formation of lactone groups was limited to DFBC and AC samples, and was found to correlate strongly with increasing reactor temperatures during oxidation. From these results it can be concluded that, to maximize the development of CEC, the generation of predominately carboxylic groups is required. To achieve this, ozone oxidation temperatures should be maintained near room temperature to avoid excessive sample loss and lactone formation. The data collected suggests that variations in either or both surface area and ash content strongly affect oxidation rate, however further studies are required to isolate these effects. Acknowledgments This project was financially supported by Washington State Department of Agriculture through the Appendix A program. Funding was also provided by the Washington State University Agricultural Research Centre through Hatch Project 0701. References [1] J. Lehmann, S. Joseph, Biochar for environmental management: an introduction, in: J. Lehmann, S. Joseph (Eds.), Biochar for Environmental Management: Science and Technology, Earthscan, Washington DC, 2009. €fer, B. Liang, J. Kinyangi, et al., Mo[2] D. Solomon, J. Lehmann, J. Thies, T. Scha lecular signature and sources of biochemical recalcitrance of organic C in Amazonian dark earths, Geochim. Cosmochim. Acta 71 (9) (2007) 2285e2298. [3] B. Glaser, J. Lehmann, W. Zech, Ameliorating physical and chemical properties of highly weathered soils in the tropics with charcoal e a review, Biol. Fertil. Soils 35 (2002) 219e230. [4] P. Blackwell, G. Riethmuller, M. Collins, Biochar application to soil, in: J. Lehmann, S. Joseph (Eds.), Biochar for Environmental Management: Science and Technology, Earthscan, Washington DC, 2009, pp. 207e226. [5] S. Kishimoto, G. Sugiura, Charcoal as a soil conditioner, in: Symposium on Forest Products Research International: Achievements and the Future, 1985, pp. 12e23. Pretoria, Republic of South Africa. [6] J. Streubel, H. Collins, M. Garcia-Perez, J. Tarara, D. Granatstein, C. Kruger, Influence of contrasting biochar types on five soils at increasing rates of application, Soil Sci. Soc. Am. J. 75 (4) (2011). [7] S.-J. Park, S.-Y. Jin, Effect of ozone treatment on ammonia removal of activated carbons, J. Colloid Interface Sci. 286 (1) (2005) 417e419. [8] H. Valdes, M. Sanchez-Polo, J. Rivera-Utrilla, C.A. Zaror, Effect of ozone treatment on surface properties of activated carbon, Langmuir 18 (6) (2002) 2111e2116. [9] C.-C. Huang, H.-S. Li, C.-H. Chen, Effect of surface acidic oxides of activated carbon on adsorption of ammonia, J. Hazard Mater. 159 (2e3) (2008) 523e527.
[10] J.R. Kastner, J. Miller, K.C. Das, Pyrolysis conditions and ozone oxidation effects on ammonia adsorption in biomass generated chars, J. Hazard Mater. 164 (2e3) (2009) 1420e1427. [11] X.-L. Long, H. Cheng, Z.-L. Xin, W.-D. Xiao, W. Li, W.-K. Yuan, Adsorption of ammonia on activated carbon from aqueous solutions, Environ. Prog. 27 (2) (2008) 225e233. [12] H.-L. Chiang, C.P. Huang, P.C. Chiang, The surface characteristics of activated carbon as affected by ozone and alkaline treatment, Chemosphere 47 (3) (2002) 257e265. [13] K. Kawamoto, K. Ishimaru, Y. Imamura, Reactivity of wood charcoal with ozone, J. Wood Sci. 51 (1) (2005) 66e72. [14] H.P. Boehm, E. Diehl, W. Heck, R. Sappok, Surface oxides of carbon, Angew. Chem. Int. Ed. 3 (10) (1964) 669e677. [15] D.B. Mawhinney, J.T. Yates, FTIR study of the oxidation of amorphous carbon by ozone at 300 K e direct COOH formation, Carbon 39 (8) (2001) 1167e1173. [16] J. Jaramillo, P.M. Alvarez, V. Gomez-Serrano, Preparation and ozone-surface modification of activated carbon. Thermal stability of oxygen surface groups, Appl. Surf. Sci. 256 (17) (2010) 5232e5236. n, F.J. Masa, F. Medina, Ozonation of [17] P.M. Alvarez, J.F. García-Araya, F.J. Beltra activated carbons: effect on the adsorption of selected phenolic compounds from aqueous solutions, J. Colloid Interface Sci. 283 (2) (2005) 503e512. [18] O. Belyaeva, T. Krasnova, S. Semenova, Effect of modification of granulated activated carbons with ozone on their properties, Russ. J. Appl. Chem. 84 (4) (2011) 597e601. [19] A. Sluiter, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, Determination of Extractives in Biomass, National Renewable Energy Laboratory, 2008. NREL/TP510e42619. [20] A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, et al., Determination of Structural Carbonhydrates and Lignin in Biomass, National Renewable Energery Laboratory, 2011. NREL/TP-510e42618. [21] S.-S. Liaw, Z. Wang, P. Ndegwa, C. Frear, S. Ha, C.-Z. Li, et al., Effect of pyrolysis temperature on the yield and properties of bio-oils obtained from the auger pyrolysis of douglas fir wood, J. Anal. Appl. Pyrolysis 93 (0) (2012) 52e62. [22] K. Rakness, G. Gordon, B. Langlais, W. Masschelein, N. Matsumoto, Y. Richard, et al., Guideline for measurement of ozone concentration in the process gas from an ozone generator, Ozone Sci. Eng. 18 (3) (1996) 209e229. [23] ASTM, Standard Test Method for Ash in Wood D1102-84, ASTM, American Society for Testing and Materials, West Conshocken, PA, 1984. [24] ASTM, Standard Test Method for Compositional Analysis by Thermogravimetry E1131-08, ASTM, American Society for Testing and Materials, West Conshohocken, PA, 2008. [25] F. Stoeckli, A. Guillot, A.M. Slasli, D. Hugi-Cleary, The comparison of experimental and calculated pore size distributions of activated carbons, Carbon 40 (3) (2002) 383e388. [26] J.S. Noh, J.A. Schwarz, Estimation of the point of zero charge of simple oxides by mass titration, J. Colloid Interface Sci. 130 (1) (1989) 157e164. [27] J.S. Noh, J.A. Schwarz, Effect of HNO3 treatment on the surface acidity of activated carbons, Carbon 28 (5) (1990) 675e682. [28] R. Kwok, XPS Peak Fitting Program for WIN95/98 XPSPEAK Version 4.1, Department of Chemistry, The Chinese University of Hong Kong, 2000. [29] S.T. Jackson, R.G. Nuzzo, Determining hybridization differences for amorphous carbon from the XPS C 1s envelope, Appl. Surf. Sci. 90 (2) (1995) 195e203. [30] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photoelectron Spectroscopy: a Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Physical Electronics, Eden Prairie, Mn, 1995. [31] W.H. Hendershot, M. Duquette, A simple barium chloride method for determining cation exchange capacity and exchangeable cations, Soil Sci. Soc. Am. J. 50 (3) (1986) 605e608. [32] Varian, Flame Atomic Absorption Spectrometry Analytical Methods, Australia Varian Australia Pty. Ltd., Mulgrave Victoria, 1989. [33] T.W. May, R.H. Wiedmeyer, A table of polyatomic interferences in ICP-MS, A. T. Spectrosc. 19 (5) (1998) 150e155. [34] R.C. Sullivan, T. Thornberry, J.P.D. Abbatt, Ozone decomposition kinetics on alumina: effects of ozone partial pressure, relative humidity and repeated oxidation cycles, Atmos. Chem. Phys. 4 (5) (2004) 1301e1310. [35] E. Perraudin, H. Budzinski, E. Villenave, Kinetic study of the reactions of ozone with polycyclic aromatic hydrocarbons adsorbed on atmospheric model particles, J. Atmos. Chem. 56 (1) (2007) 57e82. [36] C.R. Usher, A.E. Michel, V.H. Grassian, Reactions on mineral dust, Chem. Rev. 103 (12) (2003) 4883e4940. [37] D.M. Smith, A.R. Chughtai, Reaction kinetics of ozone at low concentrations with n-hexane soot, J. Geophys Res. 101 (D14) (1996) 19607e19620. [38] J. Yu, D.R. Cocker, R.J. Griffin, R.C. Flagan, J.H. Seinfeld, Gas-phase ozone oxidation of monoterpenes: gaseous and particulate products, J. Atmos. Chem. 34 (2) (1999) 207e258. [39] D.J. Malik, V. Strelko Jr., M. Streat, A.M. Puziy, Characterisation of novel modified active carbons and marine algal biomass for the selective adsorption of lead, Water Res. 36 (6) (2002) 1527e1538. [40] S. Biniak, A. Swiatkowski, M. Pakula, Electrochemical studies of phenomena at active carbon-electrolyte solution interfaces, Chem. Phys. Carbon 27 (-) (2001) 125e225. [41] L. Tsechansky, E.R. Graber, Methodological limitations to determining acidic groups at biochar surfaces via the Boehm titration, Carbon 66 (0) (2014) 730e733. [42] R.B. Fidel, D.A. Laird, M.L. Thompson, Evaluation of modified boehm titration methods for use with biochars, J. Environ. Qual. 42 (6) (2013) 1771e1778.