Steam gasification of plant biomass using molten carbonate salts

Energy 49 (2013) 211e217

Contents lists available at SciVerse ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Steam gasification of plant biomass using molten carbonate salts Brandon J. Hathaway, Masanori Honda, David B. Kittelson, Jane H. Davidson* Department of Mechanical Engineering, University of Minnesota, 111 Church St SE, Minneapolis, MN 55455, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 March 2012 Received in revised form 30 October 2012 Accepted 1 November 2012 Available online 21 December 2012

This paper explores the use of molten alkali-carbonate salts as a reaction and heat transfer medium for steam gasification of plant biomass with the objectives of enhanced heat transfer, faster kinetics, and increased thermal capacitance compared to gasification in an inert gas. The intended application is a solar process in which concentrated solar radiation is the sole source of heat to drive the endothermic production of synthesis gas. The benefits of gasification in a molten ternary blend of lithium, potassium, and sodium carbonate salts is demonstrated for cellulose, switchgrass, a blend of perennial plants, and corn stover through measurements of reaction rate and product composition in an electrically heated reactor. The feedstocks are gasified with steam at 1200 K in argon and in the molten salt. The use of molten salt increases the total useful syngas production by up to 25%, and increases the reactivity index by as much as 490%. Secondary products, in the form of condensable tar, are reduced by 77%. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Gasification Molten salt Thermochemistry Biomass Pyrolysis Solar

1. Introduction One of the potential options for sustainable, low carbon fuels to replace fossil fuels is biofuel from cellulosic biomass sources. Low energy input biomass, such as a blend of perennial native plant species, offers the potential to produce life cycle carbon neutral-tonegative fuels [1]. Switchgrass, a monoculture energy crop, and corn stover, an agricultural residue are also candidate feedstocks for fuel production. The use of concentrated solar thermal energy to drive the steam gasification of plant biomass can nearly double fuel production per unit biomass while storing intermittent solar energy as a high quality, carbon neutral synthesis gas [2]. Gasification in general encompasses any endothermic reaction at temperatures greater than 1000 K that breaks down organic materials into gaseous species such as H2, CO, CH4, and CO2. Typically an oxygen carrier such as H2O, CO2, or O2 is added to react with non-volatile carbon species. Steam gasification of biomass encompasses two coupled chemical reactions: pyrolysis and carbon, or char, gasification. The ideal products of steam gasification are a blend of H2 and CO, referred to as synthesis gas, which can be combusted directly, or used as a precursor to generate liquid fuels, synthetic natural gas, or other products produced from hydrocarbons [3]. Pyrolysis is a rapid thermal dissociation reaction

* Corresponding author. Tel.: þ1 612 626 9850; fax: þ1 612 625 6069. E-mail address: [email protected] (J.H. Davidson). 0360-5442/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.energy.2012.11.006

in which biopolymers including cellulose, hemicellulose, and lignin are decomposed at temperatures from 550 to 1300 K. With sufficiently fast heating rates, on the order of several hundred degrees Kelvin per second, flash pyrolysis yields primarily gaseous products and a solid carbon char. Considering a general carbohydrate feedstock of molar hydrogen-carbon ratio x and molar oxygen-carbon ratio y, the idealized flash pyrolysis reaction is

x CHx Oy;ðsÞ /y$COðgÞ þ $H2;ðgÞ þ ð1  yÞ$CðsÞ 2

(1)

While the elemental composition of most biomass feedstocks and cellulose are similar, the plant biomass differs structurally. Specifically, plant matter contains hemicellulose and lignin, which are branched and cross-linked heteropolymers structurally distinct from the straight-chain homopolymer cellulose. Both biopolymers have the potential to reduce feedstock reactivity and increase production of secondary products including char and tars due to the increased thermal stability of lignin and hemicellulose [4]. Carbon or char gasification is slower than pyrolysis and requires the introduction of a gasifying agent such as steam or carbon dioxide. The gasifying agent is reduced in the presence of carbon at high temperatures producing hydrogen or carbon monoxide for steam and carbon dioxide, respectively. The net gasification reaction of steam with carbon at 1200 K is

CðsÞ þ H2 OðgÞ 4COðgÞ þ H2;ðgÞ

Dhg ¼ 144

kJ mol

(2)

212

B.J. Hathaway et al. / Energy 49 (2013) 211e217

The conventional approach to biomass gasification involves carrying out reactions (1) and (2) with oxygen beyond that required for stoichiometric char gasification, resulting in partial combustion of the feedstock or product gas to supply the required energy for complete gasification. In this mode of operation, typically referred to as autothermal, complex combustion reactions with many intermediate species are involved and the final products include byproducts of combustion such as ammonia polycyclics or increased carbon dioxide. Supplying sufficient energy to achieve autothermal operation requires that 20e40% of the feedstock be combusted within the gasifier to generate the heat required to drive the gasification and pyrolysis reactions. Improved biomass gasification processes utilize allothermal conditions, where heat is supplied by carrying out combustion of a portion of the feedstock or product gas in an external reactor. This approach avoids dilution of the product stream and the need for an oxygen facility. An alternative to burning the feedstock or product gas is the use of concentrated solar energy as the source of process heat. Not only are the products of combustion eliminated and the cost of an oxygen plant avoided, but all of the available feedstock is conserved for upgrading into synthesis gas. Moreover, the solar energy is stored in chemical form. In a prior study [2], justification was provided for utilizing a molten blend carbonate salts as a reaction medium in order to create a uniform temperature field within a solar reactor, to retain the solid and condensable byproducts including ash, as well as to convert the tarry byproducts into useful product gas. The benefits of the molten salt approach were verified in an electric furnace for cellulose and activated carbon feedstocks. A two fold increase in the rate of pyrolysis and a tenfold increase in the rate of gasification compared to pyrolysis and gasification in an inert gaseous environment were demonstrated. Given the relatively slow rate of carbon gasification reactions compared to pyrolysis, increased char production will inherently result in longer overall conversion times for plant biomass. Thus, the molten salt’s ability to catalyze the gasification of carbon is anticipated to yield greater improvements in the overall conversion rates for plant biomass than those observed with pure cellulose. In addition to the potential for high char production, plant biomass also contains moderate levels of ash, consisting primarily of calcium, potassium, and magnesium oxides. Chemical thermodynamics predicts that, in the presence of molten salt, the ash may form liquid phase compounds with the salt melt, thus avoiding formation of solid slag within the gasifier [5]. To test these hypotheses, the present study paper considers steam gasification of representative biomass feedstocks in a eutectic blend of molten of lithium, potassium, and sodium carbonate salts at 1200 K. 2. Experimental approach Biomass gasification experiments are carried out for each feedstock in a batch reactor heated by an electric furnace. Measurements of product gas yield rate, composition, and total yield of both gaseous and secondary products are presented and compared between runs with and without salt present. The data are used to quantify the influence of molten salt on the reaction kinetics, product gas composition, and production of tar. The crop biomass materials were obtained from test plots managed by the University of Minnesota. The blend of perennial plants consists of 32 different species native to the state of Minnesota. Both switchgrass (Panicum Virgatum) and the perennial blend are selected for study due to their widespread adoptability, moderate to high yields, and ability to grow on marginal lands and thus avoid competition with traditional food production agriculture

[1]. Corn stover is an agricultural residue consisting of corn stalks and cobs left on the field after harvest. Cellulose is used as a control against which the performance of these candidate feedstocks is compared. Microcrystalline cellulose was obtained from ARCOS Organics. Each feedstock material is characterized according to the standard test method ASTM-E 870 for proximate and ultimate (elemental) analyses [6]. The plant feedstock is ground to a particle size of less than 450 mm using a blade grinder and sieve, while the cellulose powder is utilized as-received with a reported average particle size of 50 mm. Each feedstock, including the cellulose powder, is compressed into cylindrical tablets with 8 mm diameter and 4 mm thickness to ensure uniformity between runs. A pyrolysis reaction model from a previous study predicts that the rate of pyrolysis is unaffected by particle size within the tablet across the range covered in this work [2]. The tablets are dried in a vacuum oven at 390 K and 0.2 atm to 4% wt moisture, determined by measurement of the mass loss during drying. Based on an estimated average true density of 1.45 g/cm3 for cellulose [7] and the other biopolymers, the porosity of the tablets is 15% for cellulose and 40% for the plant biomass. The salt is an eutectic blend of lithium, potassium, and sodium carbonate. The composition and properties of the carbonate salt blend are listed in Table 1. This blend is selected because its melting point is 670 K, as compared to 970 K without lithium carbonate, corrosion of stainless steel surfaces is reduced when lithium is present [8], and there is prior evidence of lithium carbonate influencing gasification performance beyond that observed for any single or binary carbonate salt blend without lithium [9]. Carbonate salts are preferable to chlorides or fluorides to avoid the production of chlorine or fluorine gas as well as their low corrosivity when compared to hydroxide or nitrate salts [8e10]. A schematic of the facility used to obtain kinetic data is shown in Fig. 1. The reactor, shown in greater detail in Fig. 2, is a 64 mm diameter by 178 mm long cylindrical alumina crucible which is positioned vertically within a stainless steel enclosure and sealed with a flange containing connections for feedstock delivery, gas extraction, and temperature sensors. Heating is provided by a 2.5 kW crucible furnace capable of temperatures up to 1573 K. The furnace power level is controlled to achieve the desired set point based on the temperature indicated by two type K thermocouples located beneath the salt level. The stainless steel tubing carrying the steam and argon is heated by a flexible heater up to 500 K to prevent condensation of water. The feedstock tablets are fed through the inner of two concentric tubes which make up the feedstock/reactant delivery assembly to the bottom of the reactor using a plunger. Steam and argon are fed through the annulus of the same assembly to the bottom of the reactor. During experiments, the reactor is supplied with a 200 std mL/min reactant flow containing a molar steam concentration of 60% with the balance argon. The absolute pressure within the reactor remains constant at w120 kPa for the given flow conditions. The product gases are vented at atmospheric pressure, thus the operating pressure is a result of the flow resistance of the downstream tubing and particulate filter. The openings of the

Table 1 Composition and properties of ternary eutectic alkali carbonate salt blend at 1200 K. Composition [%wt]

Thermal conductivity Specific heat capacity Melting point Density

32% Li2CO3 33% Na2CO3 35% K2CO3 0.75 [W/m-K] 1842 [J/kg-K] 670 [K] 1680 [kg/m3]

B.J. Hathaway et al. / Energy 49 (2013) 211e217

213

Flow Controllers Pressure Feedstock

Ar

Cooling Water

H2 O Boiler

Mass Spectrometer

Product Gas Quench

Product Gas Resistance Reactor Furnace

Condensate Particulate Trap Filter

Fig. 1. Diagram of the system used to carry out the steam gasification of biomass.

concentric gas and feedstock delivery tubes are located mid-depth (30 mm from the surface) of the salt melt, and 20 mm beneath a perforated stainless steel plate (the plate being located 10 mm from the surface) that ensures the feedstock is submerged during reactions with salt present. Product gases leaving the reactor are passed first through a condenser held at 274 K, followed by a liquid trap and a HEPA filter to capture the secondary products such as tar, char, and ash for measurement. Immediately after exiting the reactor, the product gas stream is diluted with argon at 7000 std mL/min. This stream dilutes and cools the product gas to prevent secondary reactions, speeds transit time to the mass spectrometer and (along with the argon stream into the reactor) provides a reference flow from which the absolute product gas yield rates are calculated. A portion of the product gas stream is sampled by an Inficion Transpector CPM mass spectrometer/residual gas analyzer to determine the product composition while the remainder is vented into a fume hood. The mass spectrometer is configured for detection of the common gasification products of CO, H2, CO2, CH4, C2H2, and C2H4 as well as Ar and H2O. The signals from the mass spectrometer as well as the gas flow rates, pressure and temperature are recorded every 0.875 s by an automated computer data acquisition system.

Feedstock Inlet

1 Thermocouples 2 Argon, Steam Inlet

Product Gas Outlet

Prior to each experiment, an air purge is carried out by performing four cycles from 67 to 237 kPa (absolute), using a vacuum pump and the pressurized argon supply, to ensure all ambient air is removed from the system. Following the purge, a leak test is carried out by confirming the difference in flow rate of the inlet and outlet gas is less than 1% of the inlet flow. For each run the reactor is heated initially with only argon flowing through the inlet until the measured temperature near the bottom of the reactor crucible reaches the desired set point, at which point steam flow is initiated. After steam delivery is stabilized (w1e2 min), the feedstock is delivered to the bottom of the reactor to undergo gasification. The feedstock is stored initially in a pre-staging volume attached to the reactor and equalized to the reactor pressure, allowing delivery despite the slightly positive pressure of the reactor. Each run continues until the mass spectrometer readings for hydrogen and carbon monoxide are back down to their pre-delivery background levels (within measurement accuracy), typically 20 min for cellulose and 30 min for the other feeds. All data are acquired at a target temperature of 1200 K. The experiment is repeated four times for each feedstock with and without salt. Table 2 lists the instruments used to obtain key measurements along with their calibrated uncertainties. Table 3 lists the test matrix and conditions for the experiments. The primary data collected from each experimental run are the mass of feedstock supplied, transient reactor temperatures, mass of secondary products, such as tar, present at the end of the run, inert and reactant gas flow rates, and mass spectrometer signal intensities for the product gas species. The data are processed to obtain molar flow rates of product gases as well as the extent of carbon conversion. The total carbon-normalized gas production rate is the sum of the molar flow rates of H2, CO, CO2, and CH4 divided by the moles of carbon delivered as feedstock. The extent of carbon conversion is calculated from the measured molar flows by integrating the amount of carbon present in the product gases from the start of the reaction to time ‘t’ according to

Concentric Inlet Tubes 178 mm Table 2 Instruments used for analysis and their uncertainty.

Melt Surface Perforated Plate 60 mm

64 mm Fig. 2. Cross-section of the reactor assembly.

Measured Parameter

Instrument

Uncertainty

Molar fraction

Inficon Transpector CPM Mass Spectrometer MKS Mass Flow Controller 1174 MKS Mass Flow Controller 330A KQXL Type K Thermocouple PX309 Pressure Transducer Sartorius GD503 Balance

þ/0.7% to þ/2.6% (Varies by gas) þ/1% FSR þ/1% FSR þ/9 K þ/5 kPa þ/0.1 mg

Argon flow Steam flow Reactor temperature Reactor pressure Filter and residue mass

B.J. Hathaway et al. / Energy 49 (2013) 211e217

Table 3 Test matrix covering the gasification runs used in this study.

Table 4 Ultimate and proximate analyses of dry feedstock.

Number Feedstock of runs

Reactor Reactor Reactor gas Inlet steam temperature environment flow rate concentration (K) (sccm)

4 4

1200 1200

Gaseous Gaseous

200 200

60% 60%

1200 1200 1200 1200

Gaseous Gaseous Molten salt Molten salt

200 200 200 200

60% 60% 60% 60%

1200 1200

Molten salt Molten salt

200 200

60% 60%

4 4 4 4 4 4

Cellulose Perennial blend Corn stover Switchgrass Cellulose Perennial blend Corn stover Switchgrass

Zt XC ðtÞ ¼

  n_ CO þ n_ CO2 þ n_ CH4 þ 2n_ C2 Hx dt

0

n0C

(3)

where n_ l is the molar flow rate of the gas species i, and n0C is the initial number of moles of carbon in the feedstock. During the course of each experiment with salt present, the salt melt releases carbon dioxide due to thermal dissociation and interaction with steam, converting a portion of the carbonate salt into a hydroxide salt according to the reaction

M2 CO3 þ H2 O42MOH þ CO2

(4)

where M is the alkali metal cation associated with the carbonate or hydroxide anion [1]. To adjust the product gas production for the carbon dioxide released from the carbonate melt, the yield rate of carbon dioxide, modeled according to the method described in [2], is subtracted from the observed products. The mass balance of carbon based on the product gas composition and the measured mass of any residual char and tar was closed to within 5% for all experiments. Following each run, the salt melt is exposed to a 200 std mL/min stream of 60% carbon dioxide in argon to reform the carbonate salt through the reverse of Eq. (4). The rates of feedstock conversion are compared with and without salt by a commonly used measure termed the reactivity index. This index is defined as the inverse of twice the time needed to reach 50% carbon conversion and gives an estimate of the overall rate for complete conversion of the reactant material.

I ¼ ð2$t50% Þ1

(5)

3. Results The composition of the plant materials and the cellulose are listed in Table 4. An elemental analysis is used to determine total amount of carbon delivered to the system. The proximate analysis confirms the increased amount of fixed carbon produced during pyrolysis of the biomass feedstock compared to pure cellulose under ASTM-E 870 conditions [9]. The reported value of fixed carbon for the plant biomass feedstock is 18e19%wt as compared to 13%wt for cellulose. Unlike the pure cellulose, the plant biomass also contains ash at concentrations of 3.66%e5.26%wt. With the elemental analysis known, the ideal equilibrium compositions for the stoichiometric gasification of the feedstocks are determined from 300 K to 1300 K using Gibbs free energy minimization [5]. The equilibrium composition of the gas phase products for the perennial blend is shown in Fig. 3 and is representative of the other candidate feedstocks due to their similar elemental compositions.

Ultimate analysis C [%wt] H [%wt] O [%wt] N [%wt] S [%wt] Proximate analysis Ash [%wt] Volatiles [%wt] Fixed carbon [%wt]

Cellulose

32 species

Corn stover

Switchgrass

44.16 6.37 49.44 0.02 0.01

48.21 5.83 41.62 0.64 0.04

45.701 5.70 42.36 0.92 0.06

47.04 6.04 41.15 0.84 0.09

<0.01 87.10 12.90

3.66 77.36 18.98

5.26 75.03 19.71

4.84 76.10 18.98

Complete carbon conversion is favored at temperatures above 1175 K, guiding the selection of 1200 K as a reactor temperature for the present work. Though equilibrium predicts only hydrogen and carbon monoxide as products, the rapidly released pyrolysis volatiles often do not achieve equilibrium producing methane and traces of acetylene and ethane in addition to condensable tars. Additionally, the equilibrium predictions are based on a stoichiometric amount of steam, however the reactions were carried out under excess steam conditions. The total gas production rate from the reactor during the first minute of gasification of each feedstock is shown in Fig. 4 for both gaseous and molten salt environments. The reported gas production rates have a maximum propagated uncertainty of 0.07 mol/ min/mol-carbon primarily due to the uncertainty in the reference flow rate of argon. Considering the data for runs without salt (dashed lines), the peak gas production rate during pyrolysis is higher with cellulose (Fig. 4a) than for any of the other feedstocks. The maximum gas production rate for cellulose is 3.2 mol/mol-C/min while the other feedstocks produce gas rates less than 1.7 mol/mol-C/min with corn stover (Fig. 4c) having the slowest rate of gas production of all the feeds. This difference in gas production rate is due to both slower pyrolysis of the more complex and thermally stable biopolymers as well as the lower volatile content of the complex feeds (as listed in the proximate analysis in Table 4). Next considering the effect of molten salt, the data show that in all cases molten salt increases the gas production rate by a statistically significant degree. The maximum gas productions rate for cellulose is 4.0 mol/mol-C/min (Fig. 4a), while maximum rates for the complex feeds are 2.9e4.1 mol/mol-C/min. Additionally, the peak rates occur with salt present, around 10 s after feed delivery compared to 15 s without salt. Compared to without salt, the differences between the rate of gas production from cellulose and the complex feeds are much less significant when molten salt is present. Given the heat transfer rate limited nature of pyrolysis,

Equilbrium Composition [mol]

214

1.2 1.0

H2O

H2

0.8 0.6

CO CH4

CO2

0.4 0.2 0.0

300

500

700 900 Temperature [K]

1100

1300

Fig. 3. Equilibrium product composition for the stoichiometric gasification of blended perennial species from 300 K to 1250 K, normalized to 1 mol of carbon.

B.J. Hathaway et al. / Energy 49 (2013) 211e217

a

b Gas Production Rate [mol/mol-C/min]

5 Gas Production Rate [mol/mol-C/min]

215

Salt

4 3

No Salt

2 1 0 0

10

20

30

40

50

5 4

Salt

3

No Salt

2 1 0 0

60

10

20

Time [sec]

d

5 4

Salt

3 No Salt

2 1

0 0

10

20

30

40

50

50

60

5

Gas Production Rate [mol/mol-C/min]

Gas Production Rate [mol/mol-C/min]

c

30 40 Time [sec]

4

Salt

3

No Salt

2 1 0 0

60

10

20

Time [sec]

30 40 Time [sec]

50

60

Fig. 4. Molar gas production rates for the combined pyrolysis and steam gasification of (a) cellulose (b) perennial blend of grasses (c) corn stover and (d) switchgrass at 1200 K with and without salt.

feedstock builds up on the unreacted char and inhibits gasification without salt present. This behavior is suggested based on an observation, unique to the switchgrass feedstock, of grey-black spheres which remained in the reactor following gasification runs. Because these spheres contained carbonaceous material and because they were not observed in the runs with salt present, we suspect that they contained unreacted material that was able to be fully reacted with salt present, giving switchgrass the largest relative production increase. The total product gas produced from cellulose is less than that for the plant biomass due to the higher ratio of carbon to oxygen for the plant biomass, as seen in Table 5. This behavior is consistent with reactions (1) and (2); as y approaches zero, total hydrogen and carbon monoxide yield is maximized. Another notable but less important feature is the large H2/CO ratio when salt is not present due to the unreacted steam in the reactor favoring the forward water-gas shift reaction

CO þ H2 O4H2 þ CO2

1.75

Reactivity Index

[min -1]

the comparative results demonstrate the benefit of enhanced heat transfer to the feedstock offered by the molten salt. After approximately 50 s with salt and 80 s without salt, pyrolysis is complete and the gas production rate slows as steam gasification of the remaining char becomes the dominant reaction. Although not shown in these plots, char gasification requires on average another 20 min for the cellulose runs and 30 min for the other feeds to reach completion. The first 60 s of the reaction are presented because the majority of the product gas (70e94%) is produced and the effect of salt is most pronounced during the initial pyrolysis reaction. This yield estimate is based on the “volatile fraction” reported for each feedstock which predicts well the pyrolysis only yield. Direct measurement of the yield during pyrolysis alone is difficult due to confounding contributions from the simultaneous gasification reactions. To quantify the average reaction rates, the reactivity index for each feedstock is shown in Fig. 5. The molten salt increases the reactivity index of cellulose from 1.2 min1e1.6 min1. For the plant biomasses, the molten salt increases the reactivity indices from 0.5, 0.2, and 0.4 min1 to 1.3, 1.4, and 1.4 min1 for the perennial blend, corn stover, and switchgrass feedstocks respectively. The largest relative gain in reactivity index is for corn stover, which has a 600% faster reactivity index in the presence of molten salt. When salt is present, the reactivity indices from gasification of the plant feedstocks are comparable to that of cellulose. The average product gas yield per unit mole of carbon in the feedstock delivered is shown in Fig. 6 and listed in Table 5. With molten salt, the useful syngas (CO þ H2) production increases for all of the feedsock considered. For cellulose (Fig. 6a) the gain is a 9% increase, from 0.94 mol/mol-C without salt to 1.02 mol/mol-C of useful syngas with salt. For the perennial blend (Fig. 6b) and corn stover (Fig. 6c) feeds, a 20% increase in useful syngas is observed with salt. Switchgrass shows the largest relative increase in useful syngas production at 30%. This increase is primarily due to reduced production of secondary products such as tar as well as reduced net production of non-useful carbon dioxide gas. The larger gain for switchgrass may be due to the fact that the ash in the switchgrass

(6)

No Salt

Salt

Perennial Blend

Corn Stover

1.50 1.25 1.00 0.75 0.50 0.25 0.00 Celllulose

Switchgrass

Fig. 5. Reactivity index for gasification of each feedstock with and without salt.

216

B.J. Hathaway et al. / Energy 49 (2013) 211e217

b 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

No Salt

H

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

CO

No Salt

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

CH

C Hx

H

CO

CO

Salt

CO

CH

C Hx

CH

C Hx

d

Salt

CO

No Salt

H

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Gas Yield [mol/mol-C]

Gas Yield [mol/mol-C]

c

CO

Salt

Gas Yield [mol/mol-C]

Gas Yield [mol/mol-C]

a

CH

C Hx

No Salt

H

CO

Salt

CO

Fig. 6. Molar gas yield for the combined pyrolysis and steam gasification of (a) cellulose (b) perennial blend of grasses (c) corn stover and (d) switchgrass at 1200 K with and without salt.

Table 5 The difference in syngas production from gasification with and without molten salt. Feedstock

Increase in syngas production with molten salt (%)

Cellulose Perennial blend Corn stover Switchgrass

5.8 15.9 15.9 25.7

Without Salt 0.6 LHV [MJ/mol-C]

which results in less carbon monoxide and more hydrogen in the product gas. With salt present, this ratio is closer to the near unity H2/CO ratio predicted by the equilibrium analysis as some of the unreacted steam interacts with the salt producing carbon dioxide which counteracts the forcing effect of steam on the water-gas shift reaction. The change in production of methane and other hydrocarbons when salt is present is not statistically significant. To compare the energetic yields of the process, Fig. 7 shows the mass normalized lower heating value (LHV) of both the product gas with and without molten salt as well as the feedstock itself. As expected, the LHV of the gas from the molten salt environment is greater than that without the salt, with a 3% increase in energy content for cellulose, 10% increase for the perennial blend, 14% increase for corn stover and 30% for the switchgrass. These increases are due primarily to the increased overall gas yield previously discussed; however, there is also a small boost in energetic content with salt present due to carbon monoxide being favored over hydrogen and carbon monoxide having a higher molar LHV than hydrogen. Secondary products in the form of condensable liquid tar or solid unreacted char are collected in and downstream of the reactor and on the HEPA filter. Without molten salt, much of the collected material is an oily tar, which represents about 4.2%wt of the total feedstock mass. For the switchgrass, 5.4% of the mass remains in the reactor in the form of a mixture of ash and the unreacted char spheres mentioned previously. Based on visual observation in an optical microscope (100 magnification), ash appears to have

With Salt Feedstock

0.5 0.4 0.3 0.2 0.1 0 Celllulose

Perennial Blend

Corn Stover

Switchgrass

Fig. 7. Molar lower heating value (LHV) of the product gas and the feedstock.

formed a slag barrier that prevented the remaining char from reacting with the steam. For the perennial blend and corn stover, the ash forms a solid slag at the bottom of the crucible with no visible carbon retained within. This slag could not be removed with solvents or mild detergent, but was removed after heating the crucible to 1200 K with salt present, confirming our earlier prediction that the ash may not remain solid when salt is present. With molten salt, there are no noticeable oily tar deposits and no unreacted char remaining in the reactor. Tar collected on the HEPA filter was reduced in the presence of molten salt to only 0.8% of the total mass delivered into the reactor. The ash present in the feed material enters into solution or forms secondary products with the melt thus no slag was found in the reactor following runs with salt. The presence of ash within the melt is not expected to affect the reaction behavior based on results of testing of a pilot scale molten salt coal gasifier where ash in quantities up to 20% of the salt mass were maintained during continuous operation [11]. 4. Conclusion The primary objective of the present study is to quantify the ability of molten carbonate salts to enhance the overall conversion of various cellulosic feedstocks into synthesis gas via the process of

B.J. Hathaway et al. / Energy 49 (2013) 211e217

gasification. A comparison of gas production rates and total yields is presented along with an analysis of the secondary products formed during the gasification process. The results demonstrate that the presence of molten salt greatly enhances the rate and extent of conversion of switchgrass, blended perennials, and corn stover as gasification feedstocks. The overall rate of gasification increases by up to 600% for the plant biomass, allowing more structurally complex feedstock to be consumed at rates comparable to that of simple cellulose. The rates of conversion of the candidate feedstocks are also much more uniform in a molten salt environment. In addition to faster reactions, the useful syngas yield for plant biomass is increased by up to 30% while overall energetic yield per unit feedstock (based on the LHV of product gas) are increased by up to 22%. These production gains are primarily due to reduced secondary product formation. Specifically, liquid tar deposits on the filter are reduced by 77% along with the elimination of ash or unreacted ash/carbon material in the reactor following runs. These are all indications that complete conversion of plant biomass to synthesis gas is improved with molten salt. Expanding on the previous work [2] which considered only ideal cellulose or carbon as feedstocks, the present study gives further evidence that the benefits for the use of molten carbonate salts as a reaction media for solar gasification remain valid for more complex candidate feedstock materials. Additionally, the results point to the ability of a solar gasification process using molten salts to handle a wide variety of feedstock without major process adjustments required for changing feedstocks, a powerful benefit for flexibility of larger scale gasification operations. Acknowledgments This project is funded by the University of Minnesota Initiative for Renewable Energy and the Environment (IREE). The authors wish to thank Professor David Tilman, director of the Cedar Creek Ecosystem Science Reserve at the University of Minnesota, for his assistance in obtaining the feedstock material utilized in this paper. Nomenclature

Dh I m

molar specific enthalpy of reaction [kJ/mol] reactivity index [1/h], [1/min] mass [g], [kg]

P R T t V_ X

217

absolute pressure [atm] universal gas constant, 8.314 [J/mol-K] temperature [K] time [sec], [min] volumetric flow rate [m3/s], [std L/min] extent of reaction [e]

Subscripts 0 initial condition value C carbon CO carbon monoxide carbon dioxide CO2 methane CH4 acetylene or ethylene C2Hx (g) gaseous phase g pertaining to the carbon gasification reaction p pertaining to the pyrolysis reaction (s) solid phase std standard flow conditions of 1 [atm], 273 [K] x molar hydrogen to carbon ratio y molar oxygen to carbon ratio References [1] Tilman D, Hill J, Lehman C. Carbon-negative biofuels from low-input highdiversity grassland biomass. Science 2006;314(5805):1598e600. [2] Hathaway BJ, Davidson JH, Kittelson DB. Solar gasification of biomass: kinetics of pyrolysis and steam gasification in molten salt. Journal of Solar Energy Engineering 2011:021011-1e021011-9. [3] Klass DL. Encyclopedia of energy. San Diego: Elsevier; 2004. p. 193e6. [4] Keshwani DR. Biomass to renewable energy processes. Boca Raton, FL, USA: CRC Press; 2010. p. 7e40. [5] HSC chemistry 7.0. Pori, Finland: Outotec Research Oy; 2011. [6] Standard test methods for analysis of wood fuels. American Society for Testing and Materials; 1982. ASTM E870. [7] Sun C. True density of microcrystalline cellulose. Journal of Pharmaceutical Sciences 2005;94(10):2132e4. [8] Coyle RT, Thomas TM, Schissel P. The corrosion of materials in molten alkali carbonate salt at 900  C. Solar Energy Research Institute; 1985. [9] Jin G, Iwaki H, Arai N. Study on the gasification of wastepaper/carbon dioxide catalyzed by molten carbonate salts. Energy 2005;30(7):1192e203. [10] Yoshida S, Matsunami J, Hosokawa Y. Coal/CO2 gasification system using molten carbonate salt for solar/fossil energy hybridization. Energy and Fuels 1999;13(5):961e4. [11] Trilling CA, Martin LW. 500-MW combined-cycle power plant fuelled with low-Btu gas produced by the Rock gas molten salt coal gasification process. In: American power conference, Chicago, IL; 1979. p. 364.