Is Elevated Pressure Required To Achieve a High ... - ACS Publications

Jun 2, 2011 - the same post that was held earlier by Lavoisier, released the. Received: ...... There is some irony here; practical equip- ment employi...
0 downloads 0 Views 1MB Size
ARTICLE pubs.acs.org/EF

Is Elevated Pressure Required To Achieve a High Fixed-Carbon Yield of Charcoal from Biomass? Part 1: Round-Robin Results for Three Different Corncob Materials Liang Wang,† Marta Trninic,‡ Øyvind Skreiberg,§ Morten Gronli,† Roland Considine,|| and Michael Jerry Antal, Jr.*,|| †

)

Department of Energy and Process Engineering, Norwegian University of Science and Technology (NTNU), Kolbjørn Hejes vei 1B, NO-7491 Trondheim, Norway ‡ Department of Process Engineering, Faculty of Mechanical Engineering, University of Belgrade, Kraljice Marije 16, 11000 Belgrade, Serbia § SINTEF Energy Research, Sem Saelands vei 11, NO-7465 Trondheim, Norway Hawaii Natural Energy Institute, School of Ocean and Earth Science and Technology, University of Hawaii at Manoa, Honolulu, Hawaii 96822, United States ABSTRACT: Elevated pressure secures the highest fixed-carbon yields of charcoal from corncob. Operating at a pressure of 0.8 MPa, a flash-carbonization reactor realizes fixed-carbon yields that range from 70 to 85% of the theoretical thermochemical equilibrium value from Waimanalo corncob. The fixed-carbon yield is reduced to a range from 68 to 75% of the theoretical value when whole Waimanalo corncobs are carbonized under nitrogen at atmospheric pressure in an electrically heated muffle furnace. The lowest fixed-carbon yields are obtained by the standard proximate analysis procedure for biomass feedstocks; this yield falls in a range from 49 to 54% of the theoretical value. A round-robin study of corncob charcoal and fixed-carbon yields involving three different thermogravimetric analyzers (TGAs) revealed the impact of vapor-phase reactions on the formation of charcoal. Deep crucibles that limit the egress of volatiles from the pyrolyzing solid greatly enhance charcoal and fixed-carbon yields. Likewise, capped crucibles with pinholes increase the charcoal and fixed-carbon yields compared to values obtained from open crucibles. Large corncob particles offer much higher yields than small particles. These findings show that secondary reactions involving vapor-phase species (or nascent vapor-phase species) are at least as influential as primary reactions in the formation of charcoal. Our results offer considerable guidance to industry for its development of efficient biomass carbonization technologies. Size reduction handling of biomass (e.g., tub grinders and chippers), which can be a necessity in the field, significantly reduces the fixed-carbon yield of charcoal. Fluidized-bed and transport reactors, which require small particles and minimize the interaction of pyrolytic volatiles with solid charcoal, cannot realize high yields of charcoal from biomass. When a high yield of corncob charcoal is desired, whole corncobs should be carbonized at elevated pressure. Under these circumstances, carbonization is both efficient and quick.

’ INTRODUCTION Coal combustion is the largest source of carbon dioxide emissions in the U.S.A.1 Alternatives to coal-fired powerplants (e.g., wind, photovoltaics, solar thermal, natural gas, etc.) are now being deployed, but cost-competitive substitutes for coal as a reductant (i.e., coke) are lacking. CO2 emissions from the iron and steel industries represented 16% of energy-related coal CO2 emissions in 2000.2 During that year, coal use was responsible for 8.7 Gt or 37% of global CO2 emissions from fossil fuels. In 2008, CO2 emissions because of coal grew to 12.6 Gt (i.e., 42% of global CO2 emissions).3 This growth of emissions was (in part) due to world crude steel production that increased from 848 Mt in 2000 to 1.3 Gt in 2008.4 Most of the CO2 emissions associated with conventional crude steelmaking result from the reduction process in a blast furnace,5 whereby coke made from hard coal and/or pulverized coal made from steam coal are used to convert iron ore into iron. The substitution of biocarbon (i.e., charcoal) for coal in the iron and steel industry can reduce CO2 emissions6 if the biocarbon is manufactured efficiently from sustainably grown r 2011 American Chemical Society

biomass. This use of biocarbon is not novel; before the dawn of recorded history, mankind employed charcoal to smelt tin for the manufacture of bronze tools,7 and today in Brazil, blast furnaces use charcoal produced from Eucalyptus wood that is cultivated nearby.8 Likewise, the Norwegian ferroalloy industry makes heavy use of charcoal imports from the Pacific.9 Unfortunately, biocarbon is not produced efficiently by conventional technology;1012 consequently, greenhouse gas emissions associated with biocarbon production are unnecessarily large and worrisome.1315 The goal of this work is to learn what reaction conditions offer the highest yields of biocarbon from biomass. Anxiety about the efficient production of charcoal and its resultant properties motivated one of the earliest publications concerned with industrial chemistry research. In 1851, Violette, who was Commissioner of Gunpowder Production in France, the same post that was held earlier by Lavoisier, released the Received: March 23, 2011 Revised: June 2, 2011 Published: June 02, 2011 3251

dx.doi.org/10.1021/ef200450h | Energy Fuels 2011, 25, 3251–3265

Energy & Fuels

ARTICLE

first16 of two important papers16,17 concerning the production and properties of charcoal. Violette observed very high yields of charcoal when carbonization was conducted at high pressure. His observations were confirmed by Palmer18 in 1914 and Bergstrom19,20 in 1915. Almost 7 decades passed before Mok and Antal21,22 reported an increase from 12 to 22% in the yield of charcoal from cellulose when the gas pressure was increased from 0.1 to 2.5 MPa. Their findings concerning improvements in charcoal yields at elevated pressures were corroborated by Blackadder and Rensfelt,23 Richard and Antal,24 and Mok et al.25 Thus, there can be no doubt that elevated pressures enhance charcoal yields. In 1909, Klason and his co-workers used their experimental results to deduce a stoichiometric equation for the pyrolytic production of charcoal from cellulose26 and “wood”27 at 400 °C. For cellulose, they found C6 H10 O5 f 3:75CH0:60 O0:13 + 2:88H2 O + 0:5CO2 + 0:25CO + C1:5 H1:25 O0:38

ð1Þ

where the first product is charcoal with a carbon content of 81.7 wt %. If we employ the usual definition of charcoal yield: ychar = Mchar/Mbio, where Mchar is the dry mass of charcoal and Mbio is the dry mass of the feedstock, then Klason et al. realized a value ychar = 34.0% from cellulose as represented by eq 1. For “wood”, Klason et al. reported a charcoal yield of 36.7% with a carbon content of 68.1%. The findings by Klason et al.26,27 (summarized above) illustrate the fact that charcoal is not a well-defined chemical compound with an explicit chemical formula. To further emphasize this point, we note that Schenkel in his Ph.D. thesis10,12,28 reported a steady increase from 70 to 95% in the carbon content of the charcoal product of beech wood pyrolysis as the pyrolysis temperature increased from 400 to 800 °C. This increase in carbon content was accompanied by a decrease in ychar from 35 to 24%. The notion of charcoal yield as a “moving target” is further illustrated by our own experience. In 1990, Antal et al.29 coauthored a review that listed charcoal yields ranging from 27.9 to 50 wt % together with proximate analyses when available. In 1996, Antal et al.30 speculated that corncob could offer a charcoal yield of 55 wt % and reported an experimental value of the charcoal yield from kukui nut shell of 62.1 wt % with a fixed carbon content of 78.1 wt %. These findings cause us to conclude that, although the value of the charcoal yield is a convenient metric for qualitative discussions, it is meaningless as a quantitative measure of the efficiency of a carbonization process. In our previous work,10,3135 we introduced the fixed-carbon yield yfC as a meaningful metric of carbonization efficiency   % fC yfC ¼ ychar ð2Þ 100  % feed ash where % fC and % feed ash represent the percentage of fixedcarbon contained in the charcoal and the percentage of ash in the feedstock, respectively. Note that the fixed-carbon content of a charcoal approximates the fraction of carbon that is effective as a metallurgical reductant. From a different perspective, the fixedcarbon content approximates the amount of pure carbon that can be obtained by further thermal treatment of the charcoal and yfC serves as a measure of the efficiency of the pyrolysis process in converting biomass into pure carbon. Modern thermochemical equilibrium software (e.g., StanJan36 software or the NASA computer program Chemical Equilibrium

Figure 1. (a) Effects of the temperature on the products of cellulose pyrolysis, following the attainment of thermochemical equilibrium at 1 MPa. (b) Effects of the pressure on the products of cellulose pyrolysis, following the attainment of thermochemical equilibrium at 400 °C.

with Applications37) enables calculations of the equilibrium yields of the products of biomass pyrolysis as a function of the reaction temperature and pressure. Figure 1 displays “theoretical” equilibrium yields of the products of cellulose pyrolysis at 1.0 MPa as a function of the temperature and at 400 °C as a function of the pressure. Although carbon yields increase somewhat at temperatures below 400 °C, the reaction rate at low temperatures is very slow. In Figure 1, the yield of H2 is represented, despite its small value that caused us to omit its display (but not its presence) in our previous publications.10,32 Although the mass fraction yields of H2 are small, hydrogen plays an important role in the equilibrium chemistry, as evidenced by its yield on a mole basis. Figure 1 shows that the maximum yield of carbon from cellulose is 28 wt % at 400 °C (i.e., 62 mol % of cellulose carbon is converted into biocarbon). Note that the pyrolytic conversion of cellulose into carbon is an exothermic reaction with a large increase in entropy when equilibrium is reached; consequently, the carbonization process is irreversible, and pressure has little effect on the yield of carbon at 400 °C, while higher temperatures evoke a small decrease in the carbon yield.32 Ten years ago, we published comparisons for many different biomass feedstocks of experimental values of yfC with theoretical values of yfC obtained using StanJan software and the elemental compositions of the feedstocks.31 The experimental values of yfC 3252

dx.doi.org/10.1021/ef200450h |Energy Fuels 2011, 25, 3251–3265

Energy & Fuels were obtained at both atmospheric and elevated pressures using electrically heated carbonization retorts. In all cases, the values of yfC obtained at elevated pressure exceeded those obtained at atmospheric pressure, and in some cases, the yfC values approached the theoretical, limiting, equilibrium values calculated using StanJan. Three years later, a flash-carbonization (FC) reactor at the University of Hawaii (UH), which effected carbonization by air oxidation at elevated pressure (see below), realized comparably high fixed-carbon yields using various biomass feedstocks. Furthermore, the flame speeds at elevated pressure in the FC reactor were very high; consequently, the time required for carbonization of a bed of feedstock was very short (see below). Together, these findings indicate that charcoal can be produced very efficiently and quickly at elevated pressure in practical equipment. Nevertheless, pressurized equipment is costly to purchase and operate and demands special operational expertise. Also, it is a curious fact that most fundamental studies of biomass pyrolysis at atmospheric pressure have employed small samples or the thermal ablation of a large sample. Could the relatively low fixedcarbon yield obtained at atmospheric pressure be a result of small particles and facile mass transfer of volatiles away from the hot, pyrolyzing solid? This question is not new; in 1991, Hancox et al.38 proposed studies of the systematic effect of a reduction in particle size on pyrolysis chemistry, with the goal of detailing the vaporsolid secondary reactions and their impact on char yields. However, to the best of our knowledge they did not accomplish such studies. Could large particles provide high fixed-carbon yields of charcoal at atmospheric pressure? This question is especially important because tub-grinders are increasingly used by industry to shred waste biomass. Is shredded biomass suitable for carbonization? Similarly, some carbonization processes employ fluidized-bed reactors that require particulate feedstocks. Can fluidized-bed reactors realize high fixed-carbon yields of charcoal? The aim of this paper is to elucidate the effects of particle size on fixed-carbon yields at atmospheric pressure and elevated pressure in FC equipment. We hypothesize that increasing particle size substantially improves fixed-carbon yields. This improvement in yield is a result of increasing heterogeneous interactions between the pyrolytic vapors and the solid charcoal together with its mineral matter, both of which may be catalytic for the formation of charcoal. Also, we hypothesize that elevated pressure enables the carbonization of liquid bio-oil before it can vaporize and escape the solid matrix. In this paper, we examine the validity of these hypotheses using corncob samples sourced from three different locations. Corncob is particularly well-suited for this work because of its widespread availability and complementary results that are now becoming available from other laboratories.3942 Subsequent papers will examine other feedstocks of interest to industry. The use of chemical dehydration agents to produce charcoal is not a focus of our work but was discussed in a previous review.29

’ APPARATUS AND EXPERIMENTAL PROCEDURES Grab samples of corncobs were obtained from Surcin, Belgrade’s municipality in Serbia (Scob; ZP Maize Hybrid, ZP 505), Pioneer HiBred International (Pcob), Oahu, Hawaii, and the Waimanalo farm of the UH College of Tropical Agriculture and Human Resources, Oahu, Hawaii (Wcob). There are two varieties of Wcob: red and white. All of the work described in this paper employed red Wcob, except the FC

ARTICLE

Figure 2. Cross-sectioned view of (a) Pcob and (b) Scob.

Table 1. Specifications of Instruments and Their Crucibles/ Pans crucible/ instruments

crucible/pan crucible geometry

pan number volume (μL)

(d  h, mm)

TA Q600

1

crucible (90)

64

Mettler Toledo TGA 851e

2

crucible (150)

7  4.5

TA Q5000

3

pan (100)

10  1

Figure 3. Crucibles/pans used in pyrolysis experiments. experiment labeled 101014 that used white Wcob. No pretreatments (except for drying and grinding) were applied to the feedstocks to enhance carbon yields. At NTNU, each sample of corncob was prepared for carbonization tests in two different ways: either the cob was ground in a cutting mill mounted with a 1 mm sieve, or using a sharp knife, a thin cross-section was sliced from a whole cob, as shown in Figure 2. This cross-section included representative amounts of pith, woody ring, and chaff material. Also, in some cases, cubic samples were cut from the woody ring of Pcob and Scob cross-sections. The cubic particles weighed 5, 10, 20, and 40 mg, corresponding to cube sizes from 2 to 6 mm. At NTNU, all cob samples were dried in an oven at 105 °C for 24 h prior to carbonization. At the UH, the cobs were used as received, following storage in the open air. Three atmospheric pressure, thermogravimetric analyzers (TGAs) were employed in this work: models TA Q5000 and TA Q600 of TA Instruments and a Mettler Toledo model TGA/SDTA 851e. Table 1 and Figure 3 summarize the geometry and depth of the crucibles and the pans used with each TGA. All TGA runs employed nitrogen (99.999% pure) as purge gas with a flow rate of 100 mL min1. Prior to each experiment, a measured amount of corncob material (5, 10, 20, and 40 mg in single particle or powder form) was loaded into the appropriate crucible/sample pan. Each experiment was initiated with a 30 min purge at room temperature, followed by 30 min of drying at 105 °C. Then, the sample was heated from 105 to 950 °C at a heating rate of 10 °C min1. This temperature program is summarized in Table 2. For some experiments conducted in the TA Q600 and Mettler Toledo model TGA/ SDTA 851e instruments, a lid with a small pinhole was used to cover the crucible with loaded sample. These runs are identified as “closed crucible” experiments. The char yield ychar was calculated by dividing the 3253

dx.doi.org/10.1021/ef200450h |Energy Fuels 2011, 25, 3251–3265

Energy & Fuels

ARTICLE

Table 2. Temperature Programs

Table 3. Proximate Analysis, Heating Value, and Fixed Carbon Yield of Feed Materials

pyrolysis method

step

dynamic

1 2 4

heating rate

isothermal

(min)

(K/min)

temperature (°C)



30 jump

105

 

3 

proximate analyses (wt %)

time

30 10

feed

VM

fC

ash

yfC (wt %)

HHV (MJ/kg)

25

Pcob

79.64

17.75

2.61

18.23

18.87

25 f 105

Wcob

80.32

17.64

2.04

18.01

18.43

Scob

81.08

17.47

1.45

17.73

18.63

105 f 950

final sample mass by the mass measured at the end of the drying period at 105 °C. A temperature controllable muffle furnace (approximately 0.009 m3), a stainless-steel retort (approximately 0.004 m3), and two ceramic crucibles (approximately 200 mL) with lids were employed by NTNU to determine the maximum charcoal and fixed-carbon yields that can be realized at atmospheric pressure from untreated whole corncob samples. Cob samples were placed in each crucible and thereafter covered with a lid. Then, the crucibles were placed in the retort that was covered with a metal lid prior to insertion into the muffle furnace. The retort was purged with nitrogen for 30 min before heating as well as during the run to ensure carbonization in an inert atmosphere. The furnace was heated from room temperature to 950 °C with a heating rate of 5 °C/min. A thermocouple was placed in the retort to monitor the temperature history during the experiment. At NTNU, all corncob feed samples were subjected to proximate analysis according to American Society for Testing and Materials (ASTM) E 871 and 872; however, the ash content of the feeds was determined according to ASTM D 1102. Both NTNU and UH employed the ASTM D 1762-84 procedure for proximate analyses of the charcoal products. However, at UH, the charcoal volatile matter content was determined by preheating the covered crucible with sample for 2 min on the outer ledge of the furnace and 3 min on the edge of the furnace with the furnace door open and then heating the covered crucible with sample at the rear of the furnace for 6 min at 950 °C with the furnace door closed. At NTNU, the covered crucibe with sample was immediately placed at the rear of the furnace and heated for 6 min at 950 °C. The greater overall heating time used by UH suggests that the UH values for volatile matter content may be somewhat larger than those of NTNU. We offer our readers the following explanation for the differences in our procedures for proximate analysis. The NTNU researchers altered the standard procedure because of the practical difficulty in following it exactly. In the standard procedure, the measurement of volatile content requires the muffle furnace to be heated to 950 °C. Then, with the furnace door open, the crucible with the sample is set on the outer ledge of the furnace (300 °C) and then for 3 min on the edge of the furnace (500 °C). Then, the sample is moved to the rear of the furnace for 6 min with the muffle furnace closed. There are two practical difficulties that may cause misleading volatile content measurements: (1) It is hard to ensure the temperatures on the outer ledge and edge of the furnace are 300 and 500 °C. (2) When the furnace door is open, the temperature in the furnace drops quickly and requires a relatively long time to return to 950 °C. This means that, after the crucible with the sample is placed in the back of the furnace with its door closed, during the first several minutes, the temperature in the furnace is lower than 950 °C. It is possible that the temperature is still lower than 950 °C after 6 min. This would mean that the sample is not heated at 950 °C for 6 min. To perform the volatile content measurement more efficiently and to ensure that the sample is heated at 950 °C for 6 min, NTNU places the crucible with the sample directly in the rear of the furnace without preheating on the outer ledge and edge of the furnace.

Elemental analyses of the samples by NTNU were conducted by use of an elemental analyzer (Vario MACRO Elementar) according to standards ASTM E 777 (carbon and hydrogen), ASTM E 778 (nitrogen), and ASTM E 775 (sulfur). The oxygen content was determined by the difference of 100% and the sum of the ash, C, H, N, and S contents. Also, elemental analyses were obtained from two commercial laboratories in the U.S.A. The microstructure and surface topography of the char particles were investigated using a Zeiss Supra-55 variable-pressure field emission scanning electron microscope (LV FE-SEM). Samples were mounted on carbon tape without further preparation and scanned by a SEM. The SEM is equipped with an energy-dispersive X-ray spectroscope (Bruker Quantax) that enables the detection of the elemental compositions of selected spots. At UH, the biomass feed was placed in a canister that was subsequently loaded into the top of a pressure vessel (the FC reactor) that was then pressurized with air to 0.8 MPa (100 psig). Electric heating coils at the bottom of the pressure vessel ignited the lower portion of the biomass. After the specified ignition time, compressed air was delivered to the top of the pressure vessel and flowed through the packed bed of feed to sustain the carbonization process. The pressure within the reactor was continuously monitored and maintained at 0.8 MPa by a valve located downstream of the reactor. After sufficient air was delivered to carbonize the corncob, the airflow was halted and the reactor cooled overnight. The charcoal was removed from the reactor and allowed to equilibrate under a fume hood for 2 days before proximate analysis (i.e., ASTM D 1762-84) was performed. For moisture content determination, the charcoal samples were dried in a Fisher Scientific Isotemp model 282A vacuum oven evacuated below 0.015 MPa (4 in. Hg). The volatile matter and ash analyses were performed using a Thermolyne 1300 muffle furnace.

’ RESULTS Fixed-Carbon Yields. In Table 3, we see that all three cobs enjoy similar proximate analyses. The proximate analysis procedure can be viewed to be a type of carbonization process. From this perspective, we can calculate the fixed-carbon yields that are offered by proximate analysis: 18.2 wt % (Pcob) versus 17.7 wt % (Scob) versus 18.0 wt % (Wcob). In this context, the proximate analysis offers a benchmark value for the fixed-carbon yield that can be obtained at atmospheric pressure. Throughout this paper, we will be comparing fixed-carbon yields obtained under different conditions with the values obtained from the proximate analysis procedure. Table 4 displays ultimate (i.e., dry, elemental) analyses of the cobs used in this work as determined by three different laboratories, together with published analyses of cobs used in our earlier studies. Considering the fact that the Pcob and Wcob both originated in Hawaii, whereas the Scob originated in Europe, their elemental analyses are remarkably similar. Note that both NTNU and Hazen determine oxygen content by difference (i.e., they normalize their analyses to 100%), whereas Huffman does not. Because of the normalization, we lack a metric of the 3254

dx.doi.org/10.1021/ef200450h |Energy Fuels 2011, 25, 3251–3265

Energy & Fuels

ARTICLE

Table 4. Ultimate Analyses of Avicel Cellulose, Waimanalo (Wcob), Pioneer (Pcob), and Serbian (Scob) Corncobs and the Calculated Theoretical Fixed-Carbon Yield yfC ultimate analysisa (wt %) MCb (wt %)

H

O

N

S

ash

total

yfC (wt %)

Avicel cellulose (Hazen)

5.20

44.50

6.02

49.07

0.25

0.01

0.15

100.00

28.7

Avicel cellulose (Huffman)

5.48

44.37

6.15

49.23

0.01

0.00

0.01

99.77

28.4

corncob (2000)c Pcob (NTNU)d

6.40

48.22 46.98

6.20 6.39

42.92 43.38

1.57 0.54

0.13 0.10

3.48 2.61

102.52 100.00

34.1 32.4

Pcob (Hazen)d

7.37

49.66

5.74

41.82

0.44

0.05

2.29

100.00

36.5

Pcob (Huffman)

6.69

46.79

5.80

44.80

0.40

0.04

1.88

99.71

32.8

Wcob (NTNU)d

4.18

47.79

6.37

43.19

0.52

0.09

2.04

100.00

33.1

11.01

48.79

5.72

41.48

0.75

0.08

1.31

98.13

36.1

Wcob (Hazen)d

9.69

50.27

5.58

43.12

0.43

0.00

0.60

100.00

36.4

Wcob (Huffman)

10.64

48.55

5.81

43.23

0.56

0.04

1.61

99.80

34.8

5.18

47.61

6.27

43.89

0.55

0.23

1.45

100.00

32.8

Wcob (2003, Huffman)c

Scob (NTNU)d a

C

Dry mass basis. b Moisture content on a wet mass basis. c From ref 32. d By difference.

Figure 4. Effects of pressure on Waimanalo cob pyrolysis following the attainment of thermochemical equilibrium at 400 °C.

absolute accuracy of the two measurements. Values of C, H, and O listed in Table 4 for the Hazen analysis of Avicel microcrystalline cellulose suggest an accuracy of at least two significant figures, but the Hazen values for N, S, and ash are high (unless the sample was somehow contaminated). The large range in values of the C, H, and O contents reported by the three laboratories for the same cob material was unexpected. When the nitrogen, sulfur, and ash contents of the cobs are neglected, these elemental analyses can be used to calculate the yields of the pyrolysis products as a function of the pressure when thermochemical equilibrium is achieved at 400 °C using StanJan software. Figure 4 displays these yields for the Wcob ultimate analysis as a function of the pressure at 400 °C. Note that the value of the fixed-carbon yield is largely independent of the pressure. Returning to Table 4, we see that the theoretical fixedcarbon yield of Pcob ranges from 32.4 to 36.5 wt %, that the theoretical fixed-carbon yield of Wcob ranges from 33.1 to 36.4 wt %, and that the theoretical fixed-carbon yield of Scob is 32.8 wt % based on only the NTNU analysis. Note the impact of the higher oxygen content (43.23 wt %) in Huffman’s analysis of Wcob on the theoretical value of yfC (34.8 wt %), relative to Huffman’s 2003 value (41.48 wt %) for Wcob and the resultant

Figure 5. Influence of different instruments on one Pcob single particle sample char yield in an open crucible.

value of yfC (36.1 wt %). The oxygen measurements by Huffman displayed in Table 4 were direct measurements, whereas the Hazen and NTNU measurements were by difference. Clearly, the value of the theoretical fixed-carbon yield is quite sensitive to the accuracy of oxygen determination of the feed. We note that, in Table 4, we correct an error in our earlier publication32 of the analysis of Wcob (i.e., the “Wcob (2003, Huffman) values”). In the 2003 laboratory report, Huffman included the moisture content (11.01%) of the sample in the elemental analysis (i.e., its elemental analysis was on a wet basis), but this was not clearly declared in its laboratory report. For this reason, the 2003 values of the H and O contents of the Wcob, which we published under the impression that they were dry basis values,32 were high and the C content was low relative to dry basis values. In our paper, we noted these anomalies but incorrectly ascribed them to natural variations in the cob composition. The low value of C together with the high values of O and H resulted in a low value of the thermochemical equilibrium “theoretical” fixed-carbon yield and caused us to believe that our experimental value of the fixed-carbon yield was equal to the theoretical value. Table 4 presents the corrected, 3255

dx.doi.org/10.1021/ef200450h |Energy Fuels 2011, 25, 3251–3265

Energy & Fuels

Figure 6. Influence of different instruments on one Scob single particle sample char yield in an open crucible.

Figure 7. Influence of different instruments on Pcob powder sample char yield in an open crucible/pan.

dry-weight elemental analysis values for the theoretical fixedcarbon yield of the 2003 data (i.e., 36.1 wt %). When the experimental value is compared to this corrected value, the experimental value obtained by the FC process in 2003 was 78% of the theoretical value (see below). We emphasize that the chief point of Table 4 is the surprising inefficiency of conventional pyrolysis procedures (e.g., proximate analysis); the theoretical fixed-carbon yield values are double the actual fixed-carbon yields obtained by the proximate analysis procedure! This disparity between practice and theory indicates the improvements in yield that can potentially be realized by informed chemical reaction engineering of the carbonization process. The availability of three different TGA instruments enabled us to conduct an internal round-robin study of char and fixedcarbon yields from two of the three cob samples. Figures 5 and 6 display the effects of particle size on char yields at 950 °C as measured by the three instruments. These results represent TGA runs of a single, nominally cubic “particle” with varying

ARTICLE

Figure 8. Influence of different instruments on Scob powder sample char yield in an open crucible/pan.

sizes (to achieve the indicated mass) cut from the woody ring annulus of each cob. Because the composition of the woody ring is not representative of the composition of the whole cob, the char and fixed-carbon yields displayed in these figures cannot be directly compared to those of whole cobs (see below). Nevertheless, the trends displayed in these figures are meaningful. For all of the cobs in all of the instruments, the char yield increases with an increase in particle size. Furthermore, the TA Q600 instrument realizes a significantly higher char yield than the other instruments. As noted previously, the TA Q600 employs a narrow, deep crucible that isolates the sample from the flow of purge gas and thereby enhances secondary reactions. We expand upon this finding below. Figures 7 and 8 display similar results representing the effects of sample size with two of the three cob powders (not cubes) on their respective char yields. In this case, the sample size represents the amount of powder loaded into the open TGA crucible/ pan. In agreement with Figures 5 and 6, higher char yields are obtained from larger samples and the TA Q600 instrument provides the highest char yields. However, in these figures, the char yields from the powders are lower than the comparable yield from single cubes. Because the powder is representative of the composition of the whole cob, whereas the cubes are not, the lower yields may reflect compositional differences, as well as the reduced dimensions of the particles (see below). The char sample remaining in the TGA from these runs was too small to ash. To obtain an estimate of the fixed-carbon yield, we employed the volatile matter and ash contents of charcoals heated to the same final temperature in the N2-purged muffle furnace (see below; Table 6). With this additional information, values of the estimated fixed-carbon yields for the TA Q600 as a function of the sample size with open and closed crucibles are listed in Table 5. All estimated values exceed the comparable yfC obtained by the proximate analysis procedure. In all cases, larger sample sizes offered enhanced charcoal and estimated fixedcarbon yields. In all cases, the closed crucible increased the estimated fixed-carbon yield by about 1520%; nevertheless, even the closed crucible yields are much lower than the theoretical fixed-carbon yield. Note that, in some cases, the crucible/ pan was not able to accommodate the largest sample; consequently, the measurement was not made. 3256

dx.doi.org/10.1021/ef200450h |Energy Fuels 2011, 25, 3251–3265

Energy & Fuels

ARTICLE

Table 5. Charcoal and Fixed-Carbon Yields Realized at Atmospheric Pressure (0.1 MPa) in the TA Q600 Micro-TGA ychar (wt %)a

yfC (wt %)b

open crucible

closed crucible

open crucible

closed crucible

PCobc

20.43

24.51

19.27

23.12

PCobd

21.64

24.87

20.41

23.46

PCobe SCobc

22.80 20.52

25.66 24.46

21.50 19.14

24.20 22.81

SCobd

21.28

25.14

19.84

23.44

SCobe

22.56

25.62

21.04

23.89

PCobf

27.31

25.76

SCobf

27.12

25.29

WCobf

26.47

24.73

Percent of dried feed material. b yfC = charcoal yield  (100  % volatile matter  % char ash)/(100  % feed ash). Here, the volatile matter for char produced at 950 °C and ash content measured from the muffle-furnace-produced charcoal is used. c A total of 5 mg of powder sample. d A total of 10 mg of powder sample. e A total of 20 mg of powder sample. f A 180190 mg thin cross-sectioned sample. a

Figure 9. Particle size distributions of ground Pcob and Scob samples.

Table 6. Charcoal and Fixed-Carbon Yields Realized at Atmospheric Pressure (0.1 MPa) in a N2-Purged Muffle Furnace proximate analysis (wt %) ychar (wt %) yfC (wt %)

VM

FC

ash

Pcob charcoal Wcob charcoal

4.61 4.85

91.85 92.07

3.54 3.08

26.77 26.45

25.25 24.86

Scob charcoal

4.69

91.90

3.41

26.52

24.73

To further explore the effects of the particle size on char yield, we sieved 10 g samples of the ground Pcob and Scob and measured the char yield from each of the sieved samples (eight different particle sizes) using the MT T851e instrument. Figure 9 displays the particle size distributions obtained from the two ground cob samples, while Figure 10 displays the char yields. The results are startling; both cobs provide evidence of nearly identical behavior, with a steady increase in the char yield from 15.2 to 23.5 wt % as the particle size increased from 0.0630.125 to 2.53.0 mm. Clearly, the particle size has a strong effect on the char yield. We remark that smaller (80 wt %) in carbon content, whereas analyses of the balls (e.g., points 1, 9, 10, and 11) were lower in carbon content (