Biocarbon Production From Hungarian Sunflower Shells

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

J. Anal. Appl.

Pyrolysis 79 (2007) 86–90


www.elsevier.com/locate/jaap

Biocarbon production from Hungarian sunflower shells


Michael Jerry Antal Jr.*, Samuel Robert Wade, Teppei Nunoura
Hawaii Natural Energy Institute, School of Ocean and Earth Science and Technology, University of Hawaii at Manoa, Honolulu, HI, USA
Received 12 June 2006; accepted 16 September 2006
Available online 13 November 2006

Abstract
The production of biodiesel and bioethanol fuels is often accompanied by the co-production of lignocellulosic ‘‘waste’’ byproducts. For
example, the oil from sunflower seeds (that can be used to manufacture biodiesel fuel) has as ‘‘waste’’ co-products the sunflower shell and its stem.
In this paper we examine the production of biocarbon (charcoal) from sunflower shells produced in Hungary. The conversion of these shells to
biocarbon via the University of Hawaii’s Flash CarbonizationTM process is efficient and very quick. An increase in reaction pressure from 1.27 to
2.17 MPa causes a reduction in reaction time that results in a significant increase in charcoal production and profitability. These findings justify
further study of the benefits of charcoal production at pressures above 2 MPa.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Flash CarbonizationTM process; Sunflower shell; Charcoal; Biocarbon; Biodiesel; Pyrolysis; Pressure

1. Introduction The FC process [3–5] involves the ignition and control of a


flash fire at about 1 MPa within a packed bed of biomass. Heat,
Once again, the rapidly escalating price of transportation released by the fire, triggers the transformation of biomass into
fuels has engendered intense interest in the production of carbon with yields that can reach the thermochemical
biodiesel and other liquid fuels (e.g. ethanol) from renewable equilibrium ‘‘limit’’ in less than 30 min of reaction time.
biomass. Biodiesel is an expensive fuel because it is Typically, the FC process realizes charcoal yields in the range of
manufactured from vegetable oils. Fortunately, the plants that 30–40%, fixed-carbon yields (see below) from 28 to 31%, and
produce vegetable oils often also produce copious amounts of energy conversion efficiencies of biomass to charcoal from 55 to
lignocellulosic byproducts. For example, in addition to 66%. In many cases between 59 and 63% of the carbon in the
sunflower oil the sunflower plant produces shells, a blossom, biomass feedstock is transformed into carbon in the charcoal.
and a stem. Anyone, who has seen fields of sunflowers in bloom About two-third of the feedstock’s moist mass is converted into a
in Hungary, appreciates the abundance of lignocellulosic combustible gas at elevated pressure. This gas can be burned to
biomass that accompanies the precious sunflower seeds. It is generate heat for drying the moist biomass feedstock.
often erroneously assumed that these lignocellulosic bypro- For example, corn stover (including cobs) is the most plentiful
ducts have no value. In a visionary paper Varhegyi et al. [1] agricultural residue in the USA. In a previous paper [3] we
described the production of biocarbon (i.e. charcoal) from the showed that the cob is a particularly attractive feedstock for
sunflower stem. In this paper we describe the yields and carbonization: at 1.2 MPa it ignited after 2 min of heating, and
properties of biocarbons that can be derived from the sunflower airflow to the FC reactor was halted after 18 min. All
shell by use of the University of Hawaii’s Flash Carboniza- temperatures in the bed reached 500 8C before the run was
tionTM (FC) process. Biocarbons enjoy many lucrative markets. terminated. In this run 63% of the cob’s carbon was converted to
A review of technical information concerning their properties charcoal fixed-carbon, the cob charcoal product retained 60% of
and production is available [2]. the energy content of the cob feedstock, and 71% of the cob’s
moist mass was gasified. The fixed-carbon yield of this run was
100% of the thermochemical equilibrium limiting value.
In this paper we detail the effects of pressure and air delivery
* Corresponding author. Tel.: +1 808 956 7267; fax: +1 808 956 2336. on the performance metrics (reaction time, charcoal and fixed-
E-mail address: mantal@hawaii.edu (M.J. Antal Jr.). carbon yields) of the FC process with Hungarian sunflower
0165-2370/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jaap.2006.09.005
M.J. Antal Jr. et al. / J. Anal. Appl. Pyrolysis 79 (2007) 86–90 87

shell feedstock. These metrics enable a comparison of the created a shallow bed of carbon at the bottom of the canister.
economic benefits of operation at low (1.27 MPa) versus Then air was released from the vessel to establish the desired
moderate (2.17 MPa) pressure. This comparison indicates that operating pressure, and pressurized air was delivered to the top
higher operating pressures will improve the economics of of the reactor while combustion gas was vented from the
charcoal production from sunflower shells. bottom. The flame front moved up the bed, against the flow of
air, triggering the conversion of biomass into carbon. When
2. Experimental sufficient air had been delivered to assure carbonization of the
bed, the airflow was halted, the pressure vessel was
Sunflower seed shells (hulls) grown and milled in Hungary depressurized and permitted to cool. Subsequently, the charcoal
were obtained from the Hungaro-Sol Co. Mahogany wood in the bed was divided into three sections (one section of
chips (a readily available waste product of cabinet manu- mahogany charcoal and two sections of sunflower shell
facture in Hawaii) were provided by the Carbon Diversion charcoal), equilibrated in the open air, weighed, and a
Corp. and used as tinder. The shells were carbonized as representative carbon sub-sample was taken from each section
received without treatment of any sort, whereas wood chips for proximate analysis according to ASTM D1762-84. More
that had been exposed to rain were oven-dried before details concerning these procedures are described in previous
carbonization. publications [3–5].
In a typical FC run [3–5] a measured amount of feedstock
was placed inside a cylindrical canister (see Fig. 1). Then the 3. Results and discussion
canister was loaded into a vertical pressure vessel that was
subsequently pressurized at 1.27 MPa (170 psig) or 2.17 MPa Table 1 displays the elemental analysis of a grab sample of
(300 psig) by air. Electric power was delivered to an electric the sunflower shells employed in this work, and an analysis of a
heater at the bottom of the autoclave. Ignition occurred after a grab sample of the same batch of sunflower shells that was
few minutes and the heater was turned off. Gas was released reported in a previous publication [6]. The agreement of the two
from the bottom of the pressure vessel to maintain the pressure analyses is good. We remark that the sunflower shells have
below the vessel’s limit while the temperature of the bottom of about the same oxygen content as corncob [3] (another
the bed rose to a maximum value. This starved-air ignition step interesting agricultural residue), but the shell’s carbon content

Fig. 1. Schematic diagram of the laboratory-scale Flash CarbonizationTM reactor. A, air accumulator; C, compressor; CI, canister with insulation; DDV, downdraft
valve; H, electric heater; IV, isolation valve; MMV, micrometer valve; O, oxygen analyzer; PAV, pneumatically activated valve; PG, pressure gauge; PRV, pressure
relief valve; PS, power source; PT, pressure transducer; PV, pressure vessel; R, regulator; SRD, safety rapture disk; TC, Thermocouple; UDV, updraft valve; VAT,
variable auto transformer; WT, water trap.
88 M.J. Antal Jr. et al. / J. Anal. Appl. Pyrolysis 79 (2007) 86–90

Table 1
Sunflower shell analyses
ID mca (%) Ashb (%) Ultimate analysisc (%) SABRd (kg/kg)
C H O N S Ash
e
SF1 11.22 1.35 50.48 5.82 43.33 0.49 0.11 1.60 5.9
SF2 f NA NA 50.37 5.62 42.64 0.33 0.05 1.62 NA
a
Dry basis moisture content using ASTM E 1756-95.
b
ASTM E 1755-95.
c
Measured by Huffman Laboratories Inc., USA.
d
Stoichiometric air/biomass ratio.
e
Value for experiment 060324; 10.94% for experiment 060217.
f
Values taken from Antal et al. [6].

is higher than cob, whereas its hydrogen, nitrogen, sulfur, and (100 % feed ash)}, where % fC is the percentage fixed-
ash contents are lower. Table 1 also displays the stoichiometric carbon content of the charcoal, and % feed ash is the percentage
air–biomass ratio (i.e. the amount of air required to completely ash content of the feed [3,4]. The fixed-carbon yield represents
combust the dry biomass). Because of its high carbon content, the efficiency realized by the pyrolytic conversion of ash-free
this value (5.9 kg/kg) is somewhat larger than that of corncob organic matter in the feedstock into a relatively pure, ash-free
(5.14 kg/kg). carbon. For example, with corncob feedstock the FC process
Thermochemical equilibrium calculations [3,6] based on the realizes the theoretical fixed-carbon yield (28%) derived from
elemental analysis of the biomass feedstock indicate that thermochemical equilibrium calculations as discussed above.
carbon is a preferred product of biomass pyrolysis at moderate Note that the percentage volatile matter content of a charcoal is
temperatures, with byproducts of carbon dioxide, water, defined by % VM = 100 % fC % char ash, where % char
methane, and traces of carbon monoxide. Depending upon ash is the percentage ash content of the charcoal.
its elemental composition, the ‘‘theoretical yield’’ of carbon Table 2 displays the results of three FC runs with sunflower
from biomass (as determined by the thermochemical equili- shell feed and mahogany wood tinder. Many other runs were
brium calculations) ranges from about 27 to 46 wt% (on a dry conducted as part of this study, but these runs employed either
weight basis) [6]. In the case of sunflower shells the theoretical too little or too much air delivery; consequently they did not
carbon yield is 36% [6], whereas for corncob the theoretical realize optimum fixed-carbon yields. Runs SF170a and SF300
yield is 28% [3]. realized nearly identical fixed-carbon yields (28.7 and 28.0%,
The pyrolytic conversion of biomass to charcoal does not respectively) and produced charcoals with agreeably low VM
produce a pure carbon; consequently to assess the efficiency of contents (<18.1% for the sunflower shell charcoal). These
a carbonization process we employ metrics that have practical fixed-carbon yields were nearly identical to that obtained in our
value and can be compared with a theoretical prediction of the earlier work (involving a purely pyrolytic production of
yield realized when thermochemical equilibrium is established. charcoal at elevated pressure) and do not seem to be susceptible
First we define the charcoal yield: ychar = mchar/mbio, where to further improvement. Recall that the FC process realizes
mchar the dry mass of product charcoal and mbio is the dry mass 100% of the theoretical fixed-carbon yield from corncob. In this
of the feedstock. Unfortunately, this measure of the carboniza- case the fixed-carbon yield (28.7%) is only about 80% of the
tion efficiency is intrinsically vague because the chemical theoretical limit for sunflower shells (36%). At present we have
composition of charcoal is not defined. A more meaningful no good explanation for the relatively low fixed-carbon yield
metric is given by the fixed-carbon yield yfC = ychar  {% fC/ from sunflower shells as compared to corncob feedstock.

Table 2
Results of the Flash CarbonizationTM treatment of Hungarian sunflower shells (SF) with Mahogany wood tinder
ID Section Proximate analysis (%) ychara (%) yfCa (%)
VM fC Ash
SF170a SF top 7.88 86.7 5.4 34.0 28.7
060412 SF middle 16.5 80.2 3.4
Mahogany 11.0 87.6 1.4 31.3 27.6
SF170b SF top 37.9 58.3 3.9 31.4 21.7
060217 SF middle 10.3 84.0 5.8
Mahogany 10.8 87.8 1.3 29.3 25.9
SF300 SF top 8.42 85.3 6.3 34.2 28.0
060324 SF middle 18.1 78.0 3.9
Mahogany 18.5 80.5 1.1 31.2 25.3
a
For sunflower the listed yield represents both the combined value of the top and middle sections.
M.J. Antal Jr. et al. / J. Anal. Appl. Pyrolysis 79 (2007) 86–90 89

Table 3
Summary of reaction conditions
ID Feedstock dry basis (kg) Pressure (MPa) Delivered aira Total airb Superficial velocitye Flame velocityf
(mm/s) (mm/s)
ABRc (kg/kg) ld ABR (kg/kg) l
SF170a 1.01 1.27 1.41 0.238 1.47 0.248 13.0 0.83
SF170b 1.01 1.27 0.972 0.164 1.03 0.174 15.0 1.44
SF300 1.01 2.17 0.851 0.144 0.952 0.161 8.1 1.51
a
Following ignition.
b
Initial amount in reactor + delivered.
c
Air–biomass ratio.
d
l = ABR/SABR.
e
Superficial velocity of the airflow at 298 K and reactor pressure.
f
Upward velocity of the flaming pyrolysis reaction through the packed bed of biomass.

Data for run SF170b are included to illustrate the effect of feedstock (see Table 2 and the discussion above). From this result
VM content on fixed-carbon yield. A few small lumps of we conclude that at a given pressure there is a limit to the speed at
partially pyrolyzed feed were discovered in the product of this which the flaming pyrolysis reaction can move while fully
run. These lumps were the cause of the high VM content of the carbonizing the feedstock. Values of the superficial velocity in
top section. Note the dramatic effect of the high VM content of Table 3 offer an explanation for this limiting behavior. At
this run (relative to that of SF170a) on its fixed-carbon yield. 2.17 MPa the superficial velocity of the air was much less than
Table 3 offers a summary of the reaction conditions that at 1.27 MPa. Consequently, the tarry vapors produced by the
associated with the three experiments. The focus of these pyrolysis reactions had more time to react via secondary
experiments was the effect of pressure on the fixed-carbon reactions before they exited the reactor. In addition the increased
yield, the reaction time, and the air demand. Experiments pressure increased the intensity of heat release by the combustion
SF170a and SF300 realized nearly identical fixed-carbon reactions. These effects of elevated pressure enabled a higher
yields, and we believe that this yield represents the limiting flame velocity with complete carbonization of the feed.
value that can be achieved by the FC process. The flame The benefits of increased pressure are not obtained for free.
velocity is the average speed at which the flaming pyrolysis Air compressors with an outlet pressure of 1.27 MPa (170 psig)
reaction moves upward through the packed bed of biomass, are much less costly than those with an outlet pressure of
converting the biomass into charcoal. Conditions that realize a 2.17 MPa (300 psig). One goal of this work was to learn if more
higher flame velocity require less reaction time and result in a research at pressures above 1.27 MPa is justifiable. To answer
higher throughput. Note that an increase in pressure from 1.27 this question we made a simple economic assessment of the two
to 2.17 MPa nearly doubled the flame velocity (0.83 mm/s in cases that was intended for comparison purposes (only). Cost
SF170a versus 1.51 mm/s in SF300) within the reactor. In estimates were based on actual costs incurred in the fabrication
addition to the improvement in flame velocity, Table 3 shows and setup of a FC Demonstration reactor on the campus of the
that an increase in pressure also decreased the air demand of the University of Hawaii. Important aspects of the estimate include
process. Both the amount of air delivered following ignition, the size of the reactor (1.73 m D  2.74 m H), the capital
and the total air (including that needed to pressurize the reactor) investment (US$ 270,000 exclusive of the air compressor), the
were significantly decreased by the increase in pressure from cost of the feedstock (US$ 30/tonne), the selling price of fixed-
1.27 to 2.17 MPa. carbon in the charcoal product (US$ 437/tonne of fixed-
The superficial velocity [7] displayed in Table 3 is the speed at carbon), the salary of the two operators (US$ 29.93/h) and the
which air moves downward through the top of the canister (above operating schedule (24 h/day, 7 days/week, 50 weeks/year).
the bed of biomass) at 298 K and the reactor pressure following The cost of an appropriately sized 1.27 MPa air compressor is
ignition of the flash fire. Experiment SF170b was an attempt to about US$ 30,000, whereas the cost of a 2.17 MPa air
increase the flame velocity at 1.27 MPa by an increase in the rate compressor is about US$ 120,000. Table 4 displays a succinct
of air delivery; hence its superficial velocity. The experiment was summary of the results. Because of the higher flame velocity the
a partial success: the flame velocity increased from 0.83 to 2.17 MPa system is able to process 26 canisters/day with an
1.44 mm/s – only slightly below the value of the flame velocity at output of 8.4 tonnes/day of fixed-carbon, whereas the 1.27 MPa
2.17 MPa. Unfortunately, at this ‘‘high speed’’ the flaming system processes 18 canisters/day with an output of 6.1 tonnes/
pyrolysis reaction was unable to completely carbonize the day of fixed-carbon. The increased output of product charcoal

Table 4
Economic comparison for operation at 1.27 MPa (170 psig) vs. 2.17 MPa (300 psig)
Pressure (MPa) yfC Batches/day Production (tonnes/day) Capital investment (US$) Payback time (year)
1.27 0.287 18 6.1 300,000 3.7
2.17 0.280 26 8.4 390,000 1.3
90 M.J. Antal Jr. et al. / J. Anal. Appl. Pyrolysis 79 (2007) 86–90

results in a decrease in payback time (for the capital sunflower shell feedstock, and Richard Cox of the UH Office of
investment) from 3.7 to 1.3 years. We emphasize that this Technology Transfer and Economic Development for his
economic calculation is highly dependent upon the availability continued support. One of us (Antal) thanks Dr. Gabor
of electric power from a grid, the cost of land and labor, and Varhegyi, Dr. Emma Jakab and their co-workers with the
other factors. The exact value of the payback time is not a focus Institute of Materials and Environmental Chemistry of the
of this discussion. The surprising conclusion is that the increase Hungarian Academy of Sciences for collaboration during the
in flame speed by operation at 2.17 MPa results in increased past 20 years. This collaboration has resulted in the publication
production of charcoal that dramatically improves the profit- of 24 co-authored papers.
ability of the process relative to its operation at 1.27 MPa.
Further research on the performance of the FC process at References
elevated moderate pressures (e.g. 2 MPa) is well justified.
[1] G. Varhegyi, E. Jakab, F. Till, T. Szekely, Energy Fuels 3 (1989) 755.
Acknowledgments [2] M.J. Antal, M.G. Gronli, Ind. Eng. Chem. Res. 42 (2003) 1919.
[3] M.J. Antal, K. Mochidzuki, L.S. Paredes, Ind. Eng. Chem. Res. 42 (2003)
3690.
We thank following sponsors of our work for their support: [4] T. Nunoura, S.R. Wade, J.P. Bourke, M.J. Antal, Ind. Eng. Chem. Res. 45
The Consortium on Plant Biotechnology Research (CPBR), (2006) 585.
Carbon Diversion Corp. (Michael Lurvey); the Coral Industries [5] S.R. Wade, T. Nunoura, M.J. Antal Jr., Ind. Eng. Chem. Res. 45 (2006)
3512.
Endowment of the University of Hawaii Foundation, and the
[6] M.J. Antal, S.G. Allen, X. Dai, B. Shimizu, M.S. Tam, M.G. Gronli, Ind.
University of Hawaii. We also thank Lloyd Paredes (Hawaii Eng. Chem. Res. 39 (2000) 4024.
Natural Energy Institute) for his assistance with the experi- [7] T.B. Reed, R. Walt, S. Ellis, A. Das, S. Deutch, Fourth Biomass Conference
mental work, Dr. Gabor Varhegyi for obtaining and shipping the of the Americas, Oakland, CA, 1999.

You might also like