5954
Ind. Eng. Chem. Res. 2007, 46, 5954-5967
Do All Carbonized Charcoals Have the Same Chemical Structure? 2. A Model of the Chemical Structure of Carbonized Charcoal† Jared Bourke,‡ Merilyn Manley-Harris,‡ Chihiro Fushimi,§ Kiyoshi Dowaki,⊥ Teppei Nunoura,# and Michael Jerry Antal, Jr.*,| Chemistry Department, UniVersity of Waikato, PriVate Bag 3105, Hamilton, New Zealand, and the Hawaii Natural Energy Institute, School of Ocean and Earth Science and Technology, UniVersity of Hawaii at Manoa, Honolulu, Hawaii 96822
Charcoals and carbonized charcoals (i.e., biocarbons) were prepared from a wide variety of biomass substrates, including pure sugars containing five- and six-membered rings with furanose and pyranose configurations, lignin, agricultural residues (corncob and nut shells), and a hard wood. These biocarbons were subject to proximate and elemental analysis, gas sorption analysis, and analysis by inductively coupled plasma mass spectroscopy (ICP-MS), scanning electron microscopy (SEM), X-ray diffraction (XRD), electron spin resonance (ESR), 13C cross-polarization magic-angle spinning (CPMAS) NMR, and matrix-assisted, laser desorption ionization coupled with time-of-flight mass spectroscopy (MALDI-TOF MS). All the carbonized charcoals contained oxygen heteroatoms, had high surface areas, and were excellent conductors of electricity. Doping the biocarbon with boron or phosphorus resulted in a slight improvement in its electrical conductivity. The XRD analysis indicated that the carbonized charcoals possess an aromaticity of about 71% that results from graphite crystallites with an average size of about 20 Å. The NMR analysis confirmed the highly aromatic content of the carbonized charcoals. The ESR signals indicated two major types of carbon-centered organic radicals. MALDI-TOF spectra of the charcoals and carbonized charcoals greatly differed from those of synthetic graphite. The biocarbons contained readily desorbed discrete ions with m/z values of 317, 429, 453, 465, 685, and 701. These findings were employed to develop a model for the structure of carbonized charcoal that is consistent with the biocarbon’s oxygen content, microporosity and surface area, electrical conductivity, radical content, and its MALDI-TOF spectra. Introduction Carbonsthe basis of all organic lifesis a truly remarkable element: in one allotropic form, it is the hardest material known, and in another, it is among the softest. One of our greatest evolutionary advantages, the ability to control and harness fire, has since led to the ubiquitous use of the highly carbonaceous solid material known as charcoal. Charcoal is defined as the residue of solid nonagglomerated organic matter, of vegetable or animal origin, that results from carbonization by heat in the absence of air at a temperature above 300 °C.1 The physical properties of charcoal are unique: the material can be a good (or poor) electrical conductor, have a high (or low) specific surface area, and can have variable densities depending on the initial feedstock and the highest temperature employed in the carbonization procedure.2 Charcoals that have endured heattreatment temperatures (HTT) above 800 °C are referred to as “carbonized charcoals”. We employ the word “biocarbon” to refer to both charcoals and carbonized charcoals. † Dedicated to Geoffrey N. Richards, recipient of the ACS Anselme Payen Award and respected mentor. * To whom correspondence should be addressed. E-mail: mantal@ hawaii.edu. ‡ University of Waikato. § University of Hawaii at Manoa. Current address: Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguroku, Tokyo 153-8505, Japan. ⊥ University of Hawaii at Manoa. Current address: Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan. # University of Hawaii at Manoa. Current address: Environmental Science Center, University of Tokyo, Kashiwanoha 5-1-5, Kashiwa, Chiba 277-8581, Japan. | University of Hawaii at Manoa. Mailing Address: 1680 East-West Road, Honolulu, HI 96822.
It is recognized that knowledge of the structure of charcoal at the atomic level is essential to the understanding and prediction of its valuable physical and chemical properties. A number of authors have acknowledged this, none more prominent than Rosalind Franklin3 who performed pioneering powder X-ray diffraction (XRD) work in the 1950s and provided the first structural models of two distinct carbon types: “graphitizing” carbon and “nongraphitizing” carbon. Franklin concluded that the structure of such carbons depended not only on the temperature of preparation but also on the nature of the starting material. Nongraphitizing carbons, resulting from oxygen-rich or hydrogen-poor substances, were described as being comprised of randomly orientated individual graphitic units with extensive cross-linkages, whereas graphitizing carbons were described as being composed of very nearly parallel structural units held together with a small number of weak cross-linkages. Early XRD analysis of various carbons resulted in charcoal being classified as an amorphous form of carbon, “the ultimate form of the graphite series”.4 Following Franklin’s compelling XRD work, additional charcoal models have since been proposed. Byrne and Marsh5 proposed a model comprised of sp2 and sp3 carbon atoms bonded in five-, six-, and seven-membered rings. An attempt was also made to illustrate the possible porosity of such a carbonaceous structure. In a review on the structure of nongraphitizing carbons, Harris6 provided high-resolution electron microscopy (HREM) evidence that nongraphitizing carbons, specifically sucrose carbon, may have a microstructure related to that of fullerenes. HREM analysis of carbonized sucrose indicated the presence of closed nanoparticles. These nanoparticles were interpreted as fullerenelike elements similar to particles that can be produced by arc evaporation in a fullerene generator. The model of a nongraphitizing carbon proposed by Harris6 consists
10.1021/ie070415u CCC: $37.00 © 2007 American Chemical Society Published on Web 08/03/2007
Ind. Eng. Chem. Res., Vol. 46, No. 18, 2007 5955
of discrete fragments of randomly curved carbon sheets containing pentagons and heptagons. To date, none of the proposed models fully explain the unique physical and chemical properties of biocarbons. For example, in the case of Franklin’s model, the supposed cross-links present in nongraphitizing carbons would have to be extremely strong to inhibit graphitization at temperatures in excess of 3000 °C. Byrne and Marsh’s sp2 and sp3 bonded carbon model breaks down due to the well-known fact that sp3 bonded carbon is unstable at high temperatures; diamond converts to graphite at 1700 °C. Carbons comprised of fullerene or randomly curved carbon sheets also seem unlikely, as C60 is thermodynamically less stable than graphite and diamond, decomposing into amorphous carbon when heated above 727 °C.7 The decomposition of fullerene is more rapid in the presence of oxygen.8 Previous carbon models also do not explain various observed chemical and physical properties such as the presence of free radicals. Detection of electron spin resonance in charred materials was observed over half a century ago by Ingram and Bennett.9 Also, charcoals, and to a lesser extent carbonized charcoals, contain significant amounts of oxygen: 7-20 wt % in high volatile matter (VM) chars and 2-6 wt % in low VM carbons.2 Despite this analytical evidence and the additional supportive evidence summarized by Marsh et al.10 with respect to the possible oxygen functionalities associated with biocarbons, no comprehensive biocarbon model to date incorporates the presence of the heteroatom. It is recognized that the nature and extent of carbon-oxygen surface complexes are affected by specific surface area, particle size, and ash content, as well as temperature and degree of carbonization.10 Szyman´ski et al.11 assembled a table that shows thermal decomposition temperature ranges associated with various individual carbon surface oxides. Carboxylic groups decompose to CO2 at the lowest temperature region (100-400 °C), closely followed by carboxylic anhydrides and lactones (427-657 °C). The most thermally stable carbonoxygen groups are pyrone12,13 structures (900-1200 °C) followed by ethers, carbonylic and quinonic groups, and phenolic and hydroquinonic groups. We also note that charcoal is a very reactive material: it is classified under the Code of Federal Regulations (49 CFR Ch I 172.101) as “spontaneously combustible material” due to its remarkable chemisorption properties. The oxygen content of a charcoal can in part be derived from its adsorption of oxygen and moisture from the air to which it is exposed. This research has two principal foci. The major focus is the experimental identification of the most important chemical and physical properties of biocarbons that should be included in a structural model. A wide variety of biocarbons were manufactured by the flash carbonization (FC) process14,15and analyzed using techniques including: proximate and elemental analysis, gas sorption analysis, and analysis by inductively coupled plasma mass spectroscopy (ICP-MS), scanning electron microscopy (SEM), X-ray diffraction (XRD), electron spin resonance (ESR), and 13C cross-polarization magic-angle spinning (CPMAS) NMR. Electrical resistivity, density, and surface area measurements were also made for each biocarbon. In addition, the relatively new technique of matrix-assisted, laser desorption ionization coupled with time-of-flight mass spectroscopy (MALDI-TOF MS) was investigated as a possible new tool for obtaining structural related information. A minor focus of this paper is the introduction of an exotic mineral into a select charcoal substrate with the specific purpose of augmenting the biocarbon’s electrical properties. Two methods of exotic mineral
incorporation were conducted: biomass loading via the interaction of a water soluble impurity with a desired carbohydratebased feedstock and the direct solid phase interaction of an impurity salt with a pure carbohydrate source. In both methods, subsequent to impurity loading, the doped feedstocks were converted into charcoal via the FC process. Combinations of dopant impurity and feedstock investigated included boron and corncob, boron and sucrose, and phosphorus and sucrose. Apparatus and Experimental Procedures Charcoal Preparation. The FC process was employed to convert biomass feedstocks D-glucose, D-fructose, sucrose, inulin (a small polysaccharide containing typically 35 fructofuranose monomers for every 1 glucopyranose monomer), Kraft lignin, and corncob into charcoals. The purpose was to synthesize different biocarbons with different properties from these different biomass feedstocks. For example D-fructose is a simple monomeric ketose sugar, which in its crystalline form consists of pyranose rings but when present in inulin or sucrose is in the furanose ring form. The aldose D-glucose is in the pyranose form in virtually all natural occurrences. It should be noted that the oxonium ion of a ketose sugar differs from that derived from an aldose sugar. It is also recognized that the pure sugars proceed through a distinct melt phase during pyrolysis. A study of sugar composition in water with respect to temperature16 indicated increased amounts of the furanose conformation at higher temperatures. The thermodynamically favored isomerization of pyranose units to furanose units is therefore anticipated to take place for the pure sugars during carbonization. Consequently, we anticipate that charcoals derived from such feedstocks would consist primarily of condensed furanone rings and their fragments. In light of this, inulin was selected on the basis that it is a polyfuranose at room temperature and consequently offers alternative possibilities in terms of its pyrolysis chemistry from simple pyranose based feedstocks. Biomass contains lignin, a complex three-dimensional aromatic structure composed of phenyl propane units. It therefore seems plausible that a charcoal derived from this particular feedstock would consist of condensed six-membered rings analogous to graphite. A charcoal derived from corncob was prepared to illustrate the accepted differences in pyrolysis chemistry associated with individual biomass components vs the whole substrate. Corncob is composed of 26.3 wt % cellulose, 25.2 wt % hemicellulose, and 16.3 wt % lignin.17 It might be anticipated that a charcoal derived from corncob would consist of both condensed sixmembered rings in combination with condensed pyranone/ furanone rings and their fragments. Charcoals were also derived from oak wood, kukui nutshell and macadamia nutshell. A small stainless steel beaker positioned at the top of a packed biomass bed within the FC canister was employed to contain biomass feedstocks that proceeded through a prominent liquid phase during pyrolysis. Charcoals and carbons derived from such feedstocks, including D-glucose, D-fructose, sucrose, inulin, and Kraft lignin are referred to herein as “melt” charcoals or carbons. The Cowboy Charcoal Co. supplied the oak wood, used in their commercial manufacture of charcoal. The corncob was obtained from the Waimanalo Research Farm on the island of Oahu, Hawaii. Pure chemicals (ca. 98 to 99 wt %) inulin, D-glucose, and D-fructose were obtained from Alfa Aesar; sucrose was purchased off-the-shelf as C & H pure cane sugar. Kraft lignin (Indulin AT) was obtained from MeadWestvaco and was described as a purified form of Kraft pine lignin having an ash content of about 3 wt % (dry-basis, DB). Preparation of Boron-Doped Corncob. The inorganic mineral content contained within corncob biomass was partially
5956
Ind. Eng. Chem. Res., Vol. 46, No. 18, 2007
Table 1. Percentage Ash Content of Biomass Feed and Proximate Analysis Results of 20-40 Mesh (425-850 µm) Charcoals
feed 130404f
corncob demin.h corncob 060804 BDi corncob 031104 kukui nutshell 020904 Macshellk 180304 oak wood 090704 sucrose 131204 β-D-fructose 180205 Kraft lignin 220305 inulin 120405 R-D-glucose 250405 BD sucrose 160505 PDl sucrose 180505
charcoal proximate analysis (%)c
% asha content in raw feed
sectionb
VMd
fCe
ash
1.63 ( 0.56 ( 0.04
middle middle
26.0 19.8
72.0 79.3
2.0 0.9
1.93 ( 0.06 1.45j
top middle
21.1 5.2
76.0 92.7
2.9 2.1
0.47 0.27 0.03 0.09 1.98 0.10 0.08 2.44 2.06
middle middle top top top top top top top
7.6 22.3 1.8 6.8 5.7 2.5 5.1 4.7 4.3
91.7 77.1 98.2 93.1 90.4 96.7 94.7 85.6 90.4
0.7 0.6 0.0 0.1 3.9 0.8 0.2 9.7 5.2
0.51g
a Dry basissASTM E1755-95. b Refers to the position of the charcoal within the lab-scale flash carbonization canister. c Dry basissASTM D176284 (reapproved 1990). d Volatile matter (VM). e Fixed-carbon (fC). f Number corresponds to date produced ddmmyy. g ( mean and standard deviation of nine samples. h Citric acid demineralized. i BDsboron doped. j Value obtained from ref 17. k Macadamia nutshell. l PDsphosphorus doped.
removed via a hot citric acid treatment.18 Demineralized cob (500 g) was reacted with boric acid solution (4 L, 0.8 M 120 min). Excess boric acid was removed from the cob substrate via the percolation of ample amounts of deionized water. The doped cob was subsequently dried at room temperature prior to treatment by the FC process. Successful doping was indicated by an increase in the percentage ash content of both the feedstock and the respective charcoal (see Table 1). Preparation of Boron- and Phosphorus-Doped Sucrose. Feedstock and additive were dry mixed with a standard mortar and pestle. Boron-doped sucrose was prepared by mixing boric acid with sucrose at a weight ratio of 1:100 (boron to sucrose); phosphorus-doped sucrose was prepared by mixing monobasic ammonium phosphate with sucrose at a weight ratio of 1:100. The doped feedstocks were treated by the FC process. Successful doping was signaled by an increase in feedstock and subsequent charcoal percentage ash contents. Carbonization Procedure. Charcoals were carbonized within a closed porcelain crucible in a muffle furnace (Barnstead Thermolyne FB1215M). In a typical carbonization experiment, 5-10 g of 20-40 Mesh (425-850 µm) charcoal was added to a Coors crucible, lidded, and heat-treated according to the volatile matter (VM) analysis employed in ASTM method D1762-84.19 The ASTM method was modified only by the length of time for which the charcoal-containing crucible was positioned at the back of the furnace; in this case, 30 min instead of 6. Brunauer-Emmett-Teller (BET) Surface Area and Total Pore-Volume Analyses. Biocarbons were analyzed using an automatic gas sorption analyzer (Quantachrome Autosorb-1). Charcoal samples were outgassed under vacuum (212 °C, 4 h) prior to nitrogen adsorption at liquid nitrogen temperature (∼ -196 °C). The BET method was employed to determine the specific surface area from a limited linear region of the adsorption isotherm, P/P0 ) 0.01-0.1. Total pore volumes were determined from a single adsorption point at a relative pressure close to unity (P/P0 > 0.995). To validate the method, a surface area determination was performed on a commercial sample of Barnebey and Sutcliffe (B&S) coconut shell activated carbon and a low surface area aluminum oxide standard. Results (1119
and 89 m2/g, respectively) were close to the accepted values (1106 and 97 m2/g). Electrical Resistivity and Density. The electrical resistivities and apparent densities of the various carbonized charcoal samples were determined by a two-probe packed-bed technique at room temperature. The apparatus and methodology used was identical to that employed by Mochidzuki et al.2 The electrical resistivity measurements were performed on a series of 20-40 Mesh (425-850 µm) charcoal samples that were carbonized at 950 °C for a time period of 30 min. This method was employed due to previous research in this lab indicating that heat treatment times greater than 30 min did not significantly affect the electrical resistivity. The 20-40 Mesh particle size was selected to enable comparison with previous work. SEM. Electron micrographs were obtained using a Hitachi S-4100 field emission scanning electron microscope equipped with an X-ray analyzer. Powdered samples were mounted on aluminum studs using carbon tape and examined at three magnifications; 136×, 595×, and 3060×. ESR Analyses. ESR measurements were performed on a Varian E104A X-Band spectrometer operating at 9 GHz. Biocarbon samples were typically analyzed using a scan time period of 4 min and a time constant of 0.25 s. Biocarbon powders were placed in 4 mm o.d. borosilicate tubes and examined at 25 °C. ESR spectra were calibrated against the stable radical diphenylpicrylhydrazyl (DPPH) at g ) 2.0036. XRD Analyses. XRD measurements were performed using a Philips X’Pert MPD system equipped with a Cu monochromator and a MiniProp detector. The generator was set at 45 kV and 40 mA, programmable divergence slit (PDS), at 0.1 mm, and programmable receiving slit (PRS), at 0.2 mm. All scans were run over the 2θ range of 1.5-70°, using a step size of 0.03°, and a scan speed of 1.5 s/step. Collected data was subsequently treated using X’Pert HighScore software. Treatment steps involved: (1) a 20 point fast Fourier transform smooth, (2) KR2 stripping utilizing the Rachinger technique, (3) peak location by means of a minimum second-derivative peak finder technique, and (4) profile fitting via application of a pseudo-Voigt profile function to the measured data. A Scherrer calculator, which formed part of the software, was employed to determine crystallite size (τ) and lattice strain (). 13C CPMAS NMR Analyses. NMR measurements were performed on a Bruker DRX 200 spectrometer equipped with a 7 mm doubly tuned (H/X) solids probe and ZrO rotors with Teflon caps. 13C spectra were obtained at an ∼35 kHz decoupling field. A 5 µs 1H prep pulse, 2.5 s recycle delay, and 5 ms contact time were employed for all charcoals and carbons. MALDI-TOF MS Analyses. Mass spectra were measured on a Bruker Autoflex II MALDI-TOF/TOF mass spectrometer equipped with a nitrogen laser (λ ) 337 nm) operating at a frequency of 16.7 Hz. Generated ions were accelerated at 20 kV, and the detector voltage was 1.5 kV. In order to avoid detector saturation, ions less than or equal to 40 Da were suppressed. A pulsed ion extraction rate of 100 ns was employed for enhanced spectral resolution. Biocarbon samples were suspended in toluene and subsequently spotted onto a polished steel target plate. Spectra were acquired primarily in linear mode and in both positive and negative mode. The mass spectrometer was calibrated prior to measurement using a fullerene (C60) standard. A minimum of 300 shots was summed. Precursor ions collected for laser induced forward transfer (LIFT) application were accelerated to 6 kV and selected in a timed ion gate. In
Ind. Eng. Chem. Res., Vol. 46, No. 18, 2007 5957 Table 2. Ash Composition of Oak Wood, Demineralized Corncob, and Top Corncob Charcoal Samples (Dry Basis, Measured by Huffman Laboratories, Inc., U.S.A.) LOIa (%) Al2O3 (%) CaO (%) Fe2O3 (%) MgO (%) MnO (%) P2O5 (%) K2O (%) SiO2 (%) Na2O (%) TiO2 (%) SO3 (%) oak wood charcoal ash demin. corncob charcoal ash top corncob charcoal ash a
19.62
0.19
48.39
0.49
2.68
1.23
0.96
11.76
0.89
0.86
0.01
3.09
0.88
1.39
0.77
3.48
0.94
0.04
2.60
8.80
76.14
1.40
0.24
1.63
17.70
0.28
1.34
0.96
5.59
0.08
6.15
31.35
27.44
4.70
0.03
2.44
Loss on ignition; determined in air at 750 °C for 8 h.
Table 3. Proximate Analysis Results of 20-40 Mesh (425-850 µm) Carbons proximate analysis (%)c feed
sectiona
corncob corncob demin. corncob BD corncob kukui nutshell sucrose β-D-fructose β-D-fructose R-D-glucose Inulin Kraft lignin Kraft lignind BD sucrose PD sucrose
middle middle middle top middle top top top top top top top top top
HTTb
(°C)
950 1050 950 950 950 950 950 1050 950 950 950 1050 950 950
VM
fC
ash
3.2 2.5 1.7 1.5 2.6 1.4 1.7 1.6 1.6 1.4 2.7 2.8 3.2 2.7
94.3 91.4 97.0 95.0 95.0 98.5 98.1 98.3 98.3 98.1 93.2 93.1 86.6 92.1
2.5 6.1 1.3 3.5 2.4 0.1 0.2 0.1 0.1 0.5 4.1 4.1 10.2 5.2
a Refers to the position of the charcoal within the lab-scale flash carbonization canister. b Heat treatment temperature. c Dry basissASTM D1762-84 (reapproved 1990). d Particle size 212-425 µm.
Table 4. Elemental Analyses of Charcoal Samples Carbonized at 950 °C (Dry Basis, Measured by Huffman Laboratories, Inc., U.S.A.)
oak wood middle corncob bottom corncob top corncob β-D-fructose Kraft lignin R-D-glucose inulin sucrose
C (wt %)
H (wt %)
O (wt %)
N (wt %)
S (wt %)
ash (wt %)
96.19 94.01 92.92 92.99 96.62 91.31 96.33 96.10 96.70
0.65 0.71 0.64 0.72 0.75 0.68 0.74 0.71 0.76
1.54 1.92 2.19 2.25 2.24 4.62 2.53 2.61 2.17
0.20 0.56 0.64 0.54 0.13 0.51 0.11 0.13 0.12
