View Online / Journal Homepage Energy & Environmental Science

Accepted Manuscript

oue3 Nme 2 2010 | Number 12 | Volume 3 This is an Accepted Manuscript, which has been through the RSC Publishing peer Energy& review process and has been accepted for publication. To our referees: Environmental Science Accepted Manuscripts are published online shortly after acceptance, which is prior

www.rsc.org/ees Volume 3 | Number 12 | December 2010 | Pages 1813–2020 Energy & EnvironmentalEnergy Science to technical editing, formatting and proof reading. This free service from RSC

Dank u wel kiitos takk fyrir Publishing allows authors to make their results available to the community, in Downloaded by Carnegie Mellon University on 12 July 2012 aitäh děkuji D’akujem Благодаря citable form, before publication of the edited article. This Accepted Manuscript will Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A

Спасибо Thank you Tak be replaced by the edited and formatted Advance Article as soon as this is available. grazie Takk Tack 唔該 Danke To cite this manuscript please use its permanent Digital Object Identifier (DOI®), Merci gracias Ευχαριστω which is identical for all formats of publication. どうもありがとうございます。 More information about Accepted Manuscripts can be found in the

As a result of your commitment and support, RSC journals have a reputation for Information for Authors. the highest quality content. Your expertise as a referee is invaluable – thank you.

ae 1813–2020 Pages Please note that technical editing may introduce minor changes to the text and/or ISSN 1754-5692 www.rsc.org/publishing Registered Charity Number 207890 COVER ARTICLE REVIEW Drain et al. Hofmann and Schellnhuber graphics contained in the manuscript submitted by the author(s) which may alter Commercially viable porphyrinoid Ocean acidifi cation: a millennial dyes for solar cells challenge 1754-5692(2010)3:12;1-G content, and that the standard Terms & Conditions and the ethical guidelines that apply to the journal are still applicable. In no event shall the RSC be held responsible for any errors or omissions in these Accepted Manuscript manuscripts or any consequences arising from the use of any information contained in them.

www.rsc.org/ees Registered Charity Number 207890 Page 1 of 13Energy & EnvironmentalEnergy & Environmental Science Dynamic Article Links ►

Science View Online

Cite this: DOI: 10.1039/c0xx00000x Article www.rsc.org/xxxxxx

Molten catalytic coal gasification with in situ carbon and sulphur capture

Nicholas Siefert,a Dushyant Shekhawat b Shawn Litster c and David Berry b

Received (in XXX, XXX) Xth XXXXXXXXX 20XX, Accepted Xth XXXXXXXXX 20XX DOI: 10.1039/b000000x

5 A molten catalytic process has been demonstrated for converting coal into a synthesis gas consisting of roughly 20% methane and 80% hydrogen using

alkali hydroxides as both catalysts and in situ CO 2 capture agents. Baselines studies were also conducted using no catalyst, weak capture agents (CaSiO 3) and strong in situ capture agent for acid gases (CaO). While a similar gas composition can be achieved using CaO rather than alkali hydroxides, the rate of

syngas production is greater when using molten alkali hydroxides than when using CaO as the in situ capture agent for acid gases, such as HCl, H 2S and

CO 2. Parametric studies were conducted to understand the effects of temperature, pressure, catalyst composition, steam flow rate and the ratio of coal to

10 alkali hydroxide on the performance of the molten catalytic gasifier in terms of kinetics and syngas composition. To measure the amount and the rate of coal conversion, we have developed a method for quantifying the coal conversion as the reduction charge remaining, which is related to the chemical oxygen demand remaining in the coal. At temperatures between 800 oC and 900 oC, we measured first-order steam-coal gasification rates using sub- -1 bituminous coal of 2 hr in a fixed bed reactor while capturing significant quantities of both H 2S and CO 2, and while also generating 20% methane plus ethane in the syngas on a dry volume basis.

15

Broadened Perspective Gasification of solid fuels with in situ capture of carbon dioxide, hydrogen sulphide, and hydrogen chloride provides a simple means of converting coal or municipal solid waste into a gaseous fuel that is capable of being oxidized to generate electricity such that it c an meet current and possible future regulations regarding both acid and greenhouse gas emissions. The gaseous fuel can be converted into synthetic natural gas (SNG ). It can also be sent to

Downloaded by Carnegie Mellon University on 12 July 2012 either the combustor of a gas turbine or to the anode of a high temperature fuel cell. Recent system studies 1-3 indicate that the system efficiencies of Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A advanced fuel cell power plants can approach 60% while capturing >95% carbon dioxide, minimizing water consumption, and lowering the cost of electricity compared to pulverized coal combustion with carbon capture and sequestration (PCC-CCS) and integrated gasification combined cycl e with CCS (IGCC-CCS) power plants. One of the keys to achieving this target is the production of a high methane synthesis gas (>20% dry vol. basis) in a coal gasifier that consumes little to no oxygen. This gasifier can operate without oxygen because the endothermic steam gasification reactions can be offset by the exothermic carbon dioxide capture and methanation reactions. Operating without oxygen enhances the cold gas efficiency of the gasifier and also reduces the cooling load within the solid oxide fu el cell. By accomplishing what normally requires multiple distinct reactors and by eliminating the need for pure oxygen, a molten catalytic coal or waste gasifier with in situ carbon and sulphur capture has the potential to lower the cost of generating electricity and SNG while producing minimal to near zero emissions of acid and greenhouse gases.

an economically viable manner and building power plants that 1. Introduction co-generate fuels/chemicals during times of low electricity demand are pressing goals for research in the energy industry. It Over the last decade, coal-fired power plants have generated appears that one of the ways to achieve both of these goals in an approximately 40% of total electricity across the globe.4 In the 30 economically viable power plant is through the use of a catalytic 20 United States, coal power plants supply approximately 40-50% of gasifier that turns coal or waste into a methane-rich syngas. The ManuscriptEnergy & Environmental Accepted Science the demand for electricity 5 and municipal solid waste power 6 methane-rich syngas then could be converted into pipeline quality plants supply approximately 0.3% of electricity demand. Current natural gas, could be combusted on site in a gas turbine, or could pulverized coal power plants generate acid gases (NO , SO , and x x directly generate electricity within a solid oxide fuel cell. In the weakly CO 2), greenhouse gases (CO 2), and entrained particulates. 35 fuel cell case, several research groups have shown that one can 25 Reducing the pollution emitted by coal and waste power plants in

This journal is © The Royal Society of Chemistry 2012 Energy & Environmental Science , [year], [vol] , 00–00 | 1 Energy & Environmental Science Page 2 of 13 View Online achieve overall system efficiencies near or above 60% on a gasification catalyst is calcium oxide. Calcium oxide acts as both 1-3 higher heating value basis. Methane in the syngas is crucial to 60 a gasification catalyst and as a capture agent for acid gases, such

achieving high system efficiency because internal reforming of as H 2S and CO 2. Gasification processes using lime include the the methane into hydrogen can reduce the parasitic loading that CaO acceptor process by Consol Energy,19 which is similar to 5 would be required to maintain the temperature of the solid oxide chemical looping fluidized bed gasifiers such as the HyPr- fuel cell.7, 8 RING.20 One advantage of calcium oxide as a catalyst is that it Conventional entrained flow gasifiers, such as Shell-Texaco 65 remains in the solid phase and does not strongly interact with gasifier, operate around 1400 oC, produce little methane and are ceramic or steel walls of the gasifier. Though, the fact that lime virtually free of any other hydrocarbons.9 From an equilibrium and limestone are solids at typical gasifier temperatures is part of 10 thermodynamics point of view, the temperature of the gasifier the reason that they are not as good catalysts as alkali catalysts. must be lowered in order to increase the composition of methane To achieve similar rates of reaction, often transition metal in the syngas. However, in order to maintain sufficient kinetics at 70 catalysts are used in combination with calcium oxide in order to these reduced temperatures, and hence to minimize gasifier costs, achieve required reaction rates.21 a catalyst is required.10 Potassium carbonate, hydroxide, and One way of achieving fast kinetics is to operate with a molten 15 sulphide were found to catalyze both the steam-carbon reaction bed of catalysts, such as molten alkali carbonates or hydroxides,

(C+H 2O ↔ CO+H 2) and the methanation reaction (CO + 3H 2 ↔ which provide for high mass and heat transfer. Previous research 11-13 CH 4 + H 2O). The following are the main gasification 75 into molten bed gasification includes research by Kellogg and reactions and capture reactions inside of gasifier when the Rockwell International 22-25 in the 1970s as well as recent research catalyst is alkali hydroxides, calcium carbonate, or calcium by Tokyo Institute and Nagoya University.26, 27 At the time that 20 silicate. Also included are the changes in the Gibbs free energy most of the original research on catalytic gasification was

between reactants and products for each reaction at 1000K. conducted, reducing CO 2 emission was not a driving force for the 80 research, and it has not been until relatively recently that research 1000K C(s) + H 2O(g) ↔ CO(g) + H2(g) G = -8 kJ/mol has been published on molten beds of alkali hydroxides for in situ 1000K 28 CO(g) + H 2O(g) ↔ CO 2(g) + H2(g) G = -3 kJ/mol CO 2 capture. Kamo et al. used alkali hydroxides to capture CO 2 1000K 25 CO(g) + 3H 2 (g) ↔ CH4(g) + H2O (g) G = +27 kJ/mol and HCl while producing relatively pure hydrogen from a 1000K CO 2(g) + 2KOH(l) ↔K2CO 3(l) + H 2O(g) G = -220 kJ/mol chlorine-containing waste plastic. The molten bed designed by 1000K 28 CO 2(g) + CaO(s) ↔ CaCO 3(s) G = -23 kJ/mol 85 Kamo et al. is ideal for generating hydrogen because the 1000K CO 2(g) + CaSiO 3(s) ↔ CaCO 3(s) + SiO 2(s) G = +68 kJ/mol pressure is near atmospheric and there is an excess of alkali hydroxides. By using an excess of alkali hydroxides, Kamo et 14 15 28 30 Yeboah et al . and Sheth et al. have studied the effect of al. were able to maintain a molten bed, even though the gasifier different catalysts on the gasification rate, and found that temperature was below the melting point of sodium and mixtures of alkali species (lithium, sodium, potassium, and 90 potassium carbonate. rubidium) are most effective when the anion is a weak acid, such The goal of this research is to operate a molten alkali gasifier 2- - - as CO 3 or OH , as opposed to a strong acid, such Cl . A notable so as to obtain high concentrations of hydrocarbons and high -

Downloaded by Carnegie Mellon University on 12 July 2012 35 exception to this general rule is the anion, NO 3 , which is a strong kinetic rates of gasification. The gasifier was operated at o o

Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A acid, but also is an oxidant and therefore not stable under gasifier moderate temperatures (600 C – 900 C) and elevated pressures conditions. 95 (2.1 MPa). In addition, on a per carbon basis, less alkali The methods of the Exxon catalytic coal gasification process 16 hydroxides and water vapour were sent to the gasifier than in the were to operate at lower temperatures than conventional gasifiers, molten alkali hydroxide gasifier operated by Kamo et al. 28 . This

40 to impregnate coal with 20%wt K2CO 3 as a catalyst, and to react means that we expect most of the alkali hydroxide to convert to the coal with steam and recycled syngas (H 2, CO) in a single- alkali carbonates. In order to maintain the bed in a molten state, stage, fluidized-bed gasifier. The goal was to produce a syngas 100 we used an equimolar mixture of LiOH, NaOH, and KOH. For

with high methane and CO 2 concentrations. The gasifier operated example, the melting point of a near equal mixture of lithium, at roughly 700 oC and 3.5 MPa.17 It should be noted that 20%wt sodium and potassium carbonate is roughly 400 oC,14 which

45 K2CO 3 represents approximately 5% potassium compared to the means that the molten alkali species should remain molten even carbon in the coal on an elemental basis, and we will present after capturing carbon dioxide. As in the work by Kamo et al.,28

results using potassium carbonate at this same ratio. CO 2 was 105 the alkali hydroxides not only catalyze the steam-coal gasification captured downstream of the gasifier, and methane was separated reaction; they can also capture acid gases, such as HCl, H 2S, and cryogenically from the remaining H 2 and CO, which was recycled CO 2, inside of the gasifier. The alkali hydroxides can be 50 back to the catalytic gasifier. Calcium hydroxide was used in a regenerated from these alkali chlorides, sulphides and carbonates digestion process to yield about 90% catalyst recovery from the outside of the gasifier using cation-selective polymer 29-31 char/ash. The overall carbon conversion was roughly 90%, 110 membranes, such as Nafion®. Figure 1 shows an example

producing mainly methane and CO 2 as end products. To the best process flow diagram for a commercial molten catalytic reactor. of our knowledge, the Exxon catalytic coal gasification process The molten catalytic gasifier operated here represents a way of ManuscriptEnergy & Environmental Accepted Science 55 did not reached commercial scale, and was only demonstrated at converting coal or municipal solid wastes into a syngas with the the pilot-plant scale.16 A similar process is currently being potential to generate synthetic natural gas or electricity with 18 commercialized by GreatPoint Energy. 115 minimal to near zero emissions of acid gases, greenhouse gases, Instead of using alkali carbonates, another well studied coal and particulates.

2 | Journal Name , [year], [vol] , 00–00 This journal is © The Royal Society of Chemistry [year] Page 3 of 13 Energy & Environmental Science View Online 30 that holds the mixture of catalyst and coal. A tube carrying steam is placed inside of the ceramic crucible, such that the steam has to Product Molten Slurry pass through a molten bed of coal and catalyst. A thermocouple is Gas Catalytic H (g) Pump placed inside of the molten bed, which is used to control the 2 Gasifier CH 4(g) Coal furnace. Ash, Na 2S, 35 NaCl, and H2O (l) The syngas exits the reactor and liquids are condensed before Na 2CO 3 going through a pressure controller. Species that can condense from the syngas include water and tars, and depending on the pH Solid Ash Removal Filter of the liquids in the condenser, species such as ammonia, hydrogen sulphide and carbon dioxide can also condense out 40 before the back pressure controller. All experiments were semi- Ion-Selective CO 2(g), NaOH (aq) continuous, meaning that there was a set amount of coal and H2S (g/aq), Electrodialysis HCl(aq) Membranes catalyst loaded into the reactor and then there was a continuous flow of steam into the reactor. So, by the end of each experiment, the coal is mostly consumed, and the main gas specie in the pre- 45 quenched syngas is just the water vapour being continuously Fig. 1. Example process flow diagram of the molten catalytic gasifier with catalyst and carbon dioxide regeneration step included. added. Our experiments include several systematic parametric studies, including those varying: 1) catalyst type, 2) reactor temperature, 2. Experimental Set Up 3) catalyst-to-coal ratio, 4) steam-to-coal ratio, and 5) reactor 50 pressure. The catalysts used were the following: 1) an equimolar 2.1 Materials mixture of LiOH : NaOH : KOH, 2) K 2CO 3, 3) CaO, and 4) The chemicals used as catalysts were obtained from Alfa Aesar CaSiO 3. In all cases, dry coal and catalyst were mixed and then 5 (Ward Hill, MA) and the coal samples were obtained from the loaded into the reactor. While dry mixing of coal and catalyst Argonne National Laboratory Premium Coal Bank.32 We used 5 normally leads to lower gasification rates than if the coal and g samples of a high volatile bituminous coal (Pittsburgh#8 100 55 catalyst are mixed in an aqueous solution, we mixed the coal and mesh) as well as sub-bituminous coal (Wyodak-Anderson 100 catalyst dry because the eutectic alkali catalyst will be molten at mesh). Table 1 lists the coal composition on a weight basis as gasification temperatures,15 and this should provide intimate 10 well as on a molar, ash-free basis. In most of the experiments, contact between coal and catalyst. We used an equal mixture of fresh coal was used; however, we also present results for alkali hydroxides, rather than just sodium or potassium pyrolyzed Pittsburgh#8 coal. In that case, de-volatilization of coal 60 hydroxide, so that the melting point of alkali carbonates was less was performed at 700 oC for 4 hours in an argon atmosphere. than the temperature within of the gasifier. Since calcium oxide After de-volatilization, the coal was ground and sieved to 20 and calcium carbonate will not melt at the temperatures studied 15 mesh size. The de-volatilized coal was then mixed with the here, dry mixing of the coal and catalyst means that the

Downloaded by Carnegie Mellon University on 12 July 2012 catalyst, and de-volatilized a second time under an argon gasification rates here with lime will likely be less than if the coal o

Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A atmosphere at 700 C before steam was introduced into the 65 and catalyst were mixed in an aqueous solution and then dried. reactor. The de-volatilization was done to determine the syngas composition excluding pyrolysis gases, and hence to quantify the 20 non-pyrolysis methane production as a function of pressure.

Table 1: Mass composition (includes ash) and elemental composition (excludes ash) for the coal samples.

C H O N S Ash Coal [% wt] [% wt] [% wt] [% wt] [% wt] [% wt] Pittsburgh#8 75 5 8 1 2 9 Wyodak-Anderson 69 5 17 1 0.6 7.4 C H O N S [% mol] [% mol] [% mol] [% mol] [% mol] Pittsburgh#8 53 42 4 0.6 0.5 Wyodak-Anderson 48 42 9 0.6 0.2

2.2 Reaction Studies Fig. 2. Molten catalytic reactor design. Steam and coal react in a bed of molten alkali salts. Isometric and cutaway views . Furnace not shown. A pressure vessel was constructed that could withstand relatively high temperatures, high pressures, and alkaline The range of values and the standard values for operating ManuscriptEnergy & Environmental Accepted Science 25 conditions. The reactor system and associated equipment are conditions are listed in Table 2. An alkali to carbon ratio of 1 shown in Figure 2. The reactor system was placed inside of a means that the total number of moles for alkali species is equal to Series 3110 Tube Furnace by Applied Test Systems (Butler, PA). 70 the number of moles of carbon in the coal. Since equi-molar The main piece of equipment is a pressure vessel constructed out amounts of Li, Na, and K were used, this means the ratio of of Incoloy® 800HT, and with a ceramic crucible at the bottom

This journal is © The Royal Society of Chemistry [year] Journal Name , [year], [vol] , 00–00 | 3 Energy & Environmental Science Page 4 of 13 View Online elements on a mole-basis was C:Li:Na:K = 3:1:1:1, for most 55 To convert the gas composition measured by the mass cases examined here. It should be noted that a water flow rate of spectrometer into the production-averaged gas composition, the

0.05 mL/min corresponds to a flow rate of 0.56 mol H 2O/mol following steps were taken: 1) we calculated the flow rate versus C/hr, which means that each hour roughly half of a mole of water time of each gas species using its composition versus time, 2) we 5 is added to the reactor for every mole of original carbon in the integrated the flow rate for each species between the first data reactor. This molar flow rate of water vapour (62 sccm) into the 60 until 60% conversion of the coal, and 3) we divided these total reactor was six times larger than the flow rate of argon (10 sccm). flow of each gas by the total flow of all syngas species to The diffusivity of water vapour in the reactor was over three determine the production-averaged dry-gas composition between orders of magnitude larger than the gasification reaction rates, 0% and 60% conversion. At each value of temperature, pressure, 10 and hence the water vapour composition within the reactor should catalyst-to-coal ratio, we ran three experiments and calculated the be spatially constant, even though it will be changing with time as 65 mean gas composition from the three realizations. The following amount of coal remaining decreases with time. section explains how we determined the cumulative coal conversion. Table 2: Range of values and standard values of process variables 15 2.4 Product analysis: Reaction kinetics via reduction charge Range of values Standard value balance Pressure [MPa] 0.1 — 2.1 2.1 70 It is not possible to do an elemental balance for this reactor o Temperature [ C] 600 — 900 700 because (a) there was carbon and sulphur capture in the molten Alkali to carbon ratio [-] 0 2 1 — mixture and (b) water vapour was condensed before the mass Water flow rate [mL/min] 0 — 0.1 0.05 spectrometer. However, the species that were captured or could

not be measured, such as CO 2 and water, have no chemical 33 2.3 Product analysis: Gas composition and flow rate 75 oxygen demand. In order to measure the amount of coal After liquids were condensed and the pressure was reduced, conversion and hence the rate of coal conversion, we generalize the concept of chemical oxygen demand into the concept of 20 the syngas composition was characterized with a mass spectrometer (Pfeiffer OmniStar, Asslar, Germany). reduction charge remaining, where ‘reduction charge’ is defined Occasionally, the syngas was sampled just before the pressure as the number of electrons that can be generated at the anode of a 80 hypothetical fuel cell when oxygen is used as the oxidant in the controller and analyzed using gas chromatography to validate - results from the mass spectrometer. The composition of water cathode. The reduction charge, which is given in units of [mol e ], is equal to four times the chemical oxygen demand (COD) when 25 vapour in the syngas could not be measured because a condenser was placed before both the mass spectrometer and sample port for the COD is given in units of [mol O 2]. The concept of chemical the gas chromatograph. The mass spectrometer was calibrated oxygen demand in units of [g O 2 / L] is often used in wastewater using a series of known binary gas mixtures as well as a gas 85 treatment plants to estimate the amount of combustible material in wastewater;34 however, the concept of chemical oxygen mixture with the following composition: 60% H 2, 20% CH 4, 10% demand is rarely used in research papers on coal/biomass 30 Ar, 5%CO 2, 2% CO, 2% N 2, 1% C 2H6, and 0.5% C 2H4. The gasification or fuel reforming. While Berguerand et al. 35 used the Downloaded by Carnegie Mellon University on 12 July 2012 results from the mass spectrometer were also validated by steam- concept of ‘oxygen demand’ to estimate carbon conversion in a Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A gasifying graphite. Steam gasification experiments are a useful 90 chemical looping combustor, most gasification or fuel reforming verification of the calibration of the mass spectrometer because of 36 the limited number of overall reactions that are possible. For studies measure coal conversion either through a mass balance or carbon balance. 37 Instead of using a mass balance or a carbon 35 example, the composition of hydrogen in the syngas when steam- gasifying graphite is the following: balance, coal conversion can be measured by measuring the flow rate of those gas species exiting the reactor and assigning values - 95 to each gas species according to the mol e of charge that can be 2 ∙ 2 ∙ 3.5 ∙ (1) generated at the anode of a hypothetical fuel cell, i.e. by

measuring the chemical oxygen demand of the syngas leaving the 40 Where is the flow rate of species, i. This equation was verified reactor. o during baseline testing by steam-gasifying graphite at 900 C, 2.1 Reduction charge remaining provides a better definition for MPa, and measuring an average dry syngas composition of 64% 100 coal conversion than conversion of coal by mass because there is H2, 31% CO 2, 4% CO, 0.6% CH 4, and 0.2% C 2H6. no one-to-one correspondence between the mass of the elements The flow rate of syngas was determined by adding in a known in coal and their reduction charge. For example, light species like 45 amount of argon throughout the experiment. We chose 10 sccm hydrogen can volatilize from coal without a significant change in of argon as the flow rate of known inert so that the vessel volume mass of the coal, but can cause a significant measurable change in was never more than 20% inert gases. This way, there was never 105 the chemical oxygen demand of the remaining coal. Whereas a significant difference in syngas partial pressure when operating heavy species like carbon dioxide can volatilize from coal with a at the same overall pressure if no inert gases were present. The measurable change in the mass of the coal, but without a ManuscriptEnergy & Environmental Accepted Science 50 values of pressure given throughout this work are the total significant change in the chemical oxygen demand of the pressure of gases in the reactor, including water vapour and remaining coal. In our case, we measure coal conversion by argon. We chose to use argon, rather than nitrogen, as the known 110 measuring those species exiting the reactor that have a reduction

inert gas because there are no overlaps in the mass spectrometer charge, such as H 2, CO, CH 4, C2H4, and C2H6. Also, since there is analysis between argon and any of the syngas species of interest. no oxygen in these experiments, there is conservation in

4 | Journal Name , [year], [vol] , 00–00 This journal is © The Royal Society of Chemistry [year] Page 5 of 13 Energy & Environmental Science View Online reduction charge in our experiments, i.e. the total reduction 55 contains 0.29 mol of C, 0.25 mol of H, 0.05 mol of O, 0.004 mol charge of the exiting syngas is exactly equal to the reduction of N, and 0.001 mol of S. This yields a value of 1.31 mol e - of charge of the original coal. Though, it should be noted that the reduction charge for every 5 g of Wyodak-Anderson coal. concept of reduction charge balance is generic enough that it To determine the percentage of coal remaining as a function of 5 could account for oxygen addition into a gasifier, a combustor, or time, we first converted the flow rate of each gas species from - a fuel reformer. 60 units of sccm to mol e by using its assigned value of reduction A reduction charge was assigned to each of the following gas charge. We then integrated the flow rate of reduction charge

species measured by the mass spectrometer: H 2 = 2, CO = 2, CO 2 leaving the reactor from t=0 to t=t for all values of t. The = 0, CH 4 = 8, and C2H6 = 14. If molecular oxygen had been normalized reduction charge remaining (NRCR) of the coal as a 10 added, it would have been assigned a value of O 2 = -4. For each function of time for our batch reactor is defined as follows: element in the coal, the following values are assigned: C = +4, H 65 = +1, O = -2, N = 0, and S = -2. The catalysts were assigned an 2 ∙ 2 ∙ 8 ∙ 14 ∙ overall reduction charge of zero because Li,Na,K = +1, Ca = +2, 1 4 ∙ 1 ∙ 2 ∙ 0 ∙ 2 ∙ and OH = -1. The values listed above for coal, catalyst, and gases (4) 15 are the values for the number of electrons generated or consumed on an electrode of a hypothetical fuel cell. For example, graphite The NRCR is also equal to the normalized chemical oxygen (C) can generate 4 electrons; anthracene (C H ) can generate 66 14 10 demand remaining (NCODR), which can be defined as follows: electrons; methanol (CH OH) can generate 6 electrons; and 3 ammonia (NH 3) can generate 3 electrons. As mentioned earlier, ½ ½ 2 3.5 20 the value of reduction charge is equal to 4 times the value of the NCODRt 1 chemical oxygen demand. Note that we have assigned a reduction ¼ ½ 0 ∙ ½

charge value of -2 for sulphur; however, the correct value should 70 (5) actually be +6 because the end redox state of sulphur in the

environment is SO 3(g) and H 2SO 4(a). However, inside a gasifier, In order to determine the first order rate constant, we plotted 25 the final redox state of sulphur is -2 (i.e. Na 2S, H 2S). Sulphur is the NRCR as a function of time and fit an exponential curve to the last of the main coal species oxidized under partial oxidation. the data up to 60% conversion, i.e. when NRCR equals 40%. For this reason, we have decided to assign a reduction charge Note that, if the rate constant were purely first order throughout value of -2 to sulphur. This means that the capture of CO 2(g), 75 the experiment, then we would not have to specify a rate constant COS(g) or H 2S(g) has no effect on the reduction charge (i.e. to 60% conversion. However, as will be seen in the results 30 chemical oxygen demand) of the syngas, and hence has no effect section, the rate constant for up to 60% conversion is larger than on the calculation of coal conversion. Since sulphur is only a the rate constant through the entire run because of the significant minor constituent of coal, using a value of –2 rather than +6 has amount of volatilization early in the experiment. We chose the an error that is on the order of 1% by the end of the experiment. 80 value of 60% conversion because this was the largest value of A general conservation equation can be written for the conversion reached by all data sets. While some data sets with

Downloaded by Carnegie Mellon University on 12 July 2012 35 reduction charge: alkali hydroxides reached conversion values as high as 99%

Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A conversion, many of the data sets without catalyst reached 60% conversion only after 300 minutes of operation. ∑ ∙ ∑ ∙ (2)

85 3. Results Where RCR is the reduction charge remaining, is the molar reduction charge, and is the flow rate of species entering or 3.1 Gasification with and without catalyst 40 exiting the system. For our system, this equation reduces to the We first present the dry gas flow rate exiting the reactor as a following: function of time. Figure 3 shows the flow rates of the major coal

gasification product when the reactor pressure and temperature o 90 were 2.1 MPa and 800 C and the gasifier was operated either 2 ∙ 2 ∙ 8 ∙ 14 ∙ with (a) or without (b) alkali hydroxide catalysts. There was a (3) considerable increase in the reaction rates with catalyst, as seen by the significant increase in the production of hydrogen and Note that the reduction charge of the system is equal to the methane when catalysts are included with the coal. The 45 reduction charge of coal because the alkali hydroxides and any 95 production-averaged, dry gas composition in Figure 3 was gas species captured by the catalyst have a net reduction charge approximately 80% H 2 and 20% CH 4. This gas composition is of zero. Also, there is no reduction charge flowing into the only possible if there is significant capture of CO2 inside the system because the only inputs are water vapour and argon. gasifier. In fact, we determined that approximately 80% of the Using the elemental analysis presented earlier in Table 1, each ManuscriptEnergy & Environmental Accepted Science alkali hydroxide species capture CO 2 inside the gasifier and 50 5 g sample of Pittsburgh #8 coal contains 0.31 mol of C, 0.25 mol 100 convert into alkali carbonates. To the best of the authors’ of H, 0.025 mol of O, 0.005 mol of N, and 0.003 mol of S. By knowledge, this is the first documented case of using a molten assigning a reduction charge value, as listed above, to each - mixture of alkali hydroxides to capture CO 2 inside of a coal element in the coal, we obtain a value of 1.43 mol e for every 5 g gasifier ( in-situ carbon capture). of Pittsburgh#8 coal. Each 5 g sample of Wyodak-Anderson coal

This journal is © The Royal Society of Chemistry [year] Journal Name , [year], [vol] , 00–00 | 5 Energy & Environmental Science Page 6 of 13 View Online

(a) (a)

35 (b) (b) Fig. 4. (a) Flow rate of the syngas components from a coal- Fig. 3. Flow rate of the syngas components from experiments at catalyst-steam experiment at 900 oC and 2.1 MPa, using 5 g of o 5 800 C and 2.1 MPa using 5 g of fresh Pittsburgh#8, Water flow fresh Pittsburgh#8, 2.7 g of LiOH, 4.5 g of NaOH, and 6.3 g of rate of 0.05 mL/min, (a) with catalyst [2.7 g LiOH, 4.5 g NaOH, KOH. Water flow rate was 0.05 mL/min. (b) Normalized and 6.3 g KOH] and (b) without catalyst. Note the different x- 40 reduction charge remaining of the coal versus time for the same axis and y-axis scales. Time = 0 is the time at which the reactor experiment. reaches 2.1 MPa. Downloaded by Carnegie Mellon University on 12 July 2012 10 Figure 4 (b) shows the percent remaining of the coal as a Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A Figure 4 (a) shows the exhaust gas flow rate from gasification function of time for the same conditions as in Figure 4 (a). Note of 5 g of Pittsburgh#8 coal (100 mesh), 2.7 g of LiOH, 4.5 g of 45 that the data in Fig. 4 (b) starts at 90% coal remaining because o NaOH, and 6.3 g of KOH at 900 C and 2.1 MPa. At t=0, the gas 10% of the coal is consumed during the pressurization of the exits the reactor after reaching a pressure of 2.1 MPa. There was reactor. One can see that the fit through the data between 90% 15 an overall decrease in the total flow rate of syngas with time and 40% remaining is nearly exponential, i.e. a first-order rate of because the amount of coal in the reactor was slowly consumed. reaction. For our system, a first-order rate of reaction means The methane and ethane concentration decreased rapidly for three 50 either that the rate limiting reaction is only a function the amount main reasons: (1) the amount of coal available for reaction of coal or that the other species in the rate limiting reaction are decreased with time, (2) pyrolysis reactions occurred mostly at not varying significantly, such as alkali cations. An exponential 20 the beginning of the experiment, and (3) the concentration of fit through the data between NRCR equals 40% and 10% was steam inside of the reactor increased with time because less steam also nearly exponential, and as expected, the rate constant from was being consumed by the coal, and hence the amount of steam 55 40% to 10% was less than the rate constant between 90% and available to reform any methane and ethane increased with time. 40% because the coals studied here contain significant quantities The production of carbon dioxide increased at the beginning of of species that volatilize early in the experiment. It is important to 25 the experiment, and then decreased slowly over the rest of the measure the gas composition with these pyrolysis gases (CH 4, experiment. This trend is due to a combination of effects: (1) the C2H6, C 2H4) because there would be significant amounts of decreasing amount of coal remaining in the reactor and (2) the 60 pyrolysis gases generated even in a commercial scale molten

decreasing amount of hydroxides to capture CO 2 as the catalytic gasifier. ManuscriptEnergy & Environmental Accepted Science hydroxides convert to carbonates.

30 3.2 Effect of catalyst type

Tables 3 and 4 show the production-averaged, syngas

6 | Journal Name , [year], [vol] , 00–00 This journal is © The Royal Society of Chemistry [year] Page 7 of 13 Energy & Environmental Science View Online composition from the molten catalytic gasifier as a function of the 3.3 Effect of temperature type of catalyst for Pittsburgh#8 coal and for Wyodak-Anderson 45 Tables 5 and 6 show the production-averaged, syngas coal, respectively. As described in the experimental set up composition from the molten catalytic gasifier as a function of the section, production-averaged means that the total amount of reactor temperature when the pressure was held at 2.1 MPa and 5 x’ species ‘ produced during the experiment was divided by the the coal was fresh Pittsburgh#8 and Wyodak-Anderson, total amount of syngas during the experiment. In both tables, the respectively. At all temperatures, there is a significant amount of ratio of alkali-to-carbon was 1:1 and the ratio of alkali earth metal 50 capture of H 2S inside of the gasifier because the production- to carbon was 0.5:1. Therefore, there was an equal carbon dioxide averaged composition of H 2S without catalyst addition is on the capture capability when comparing the alkali hydroxides with the order of 700 ppm for Wyodak-Anderson and on the order of 5000 10 alkali earth catalysts because two moles of alkali hydroxides are ppm for Pittsburgh#8 coal. These values without catalysts are required to capture one mole of carbon dioxide, whereas only one significantly greater than the 0-10 ppm and 0-150 ppm, mole of alkali earth metal oxide is required to capture one mole 55 respectively, measured with alkali hydroxide catalysts. of carbon dioxide. As seen in Table 5 , the highest average methane composition We can see significant capture of both carbon dioxide and achieved using Pittsburgh#8 coal was 22% while at the same time 15 hydrogen sulphide when using either alkali hydroxides or calcium producing 2.6% ethane + ethylene. Typically, the ratio of ethane oxide by comparing the amount of carbon species in the syngas to ethylene measured by the mass spectrometer was 2, but we with the case of no catalyst. There is negligible CO 2 or H 2S 60 report the sum of these two species because of the significant capture for the case of calcium silicate. The first-order rate overlap between these species in the mass spectrometer used in constant was largest for alkali hydroxide catalysts, and this was this study. Hence, we use the symbol ‘C 2H4-6’ to represent the 20 most evident when using Wyodak-Anderson coal. However, in sum of the ethane plus ethylene composition in the syngas. On a the case of alkali hydroxide catalysts, the combined hydrocarbon combustion enthalpy basis, the syngas from the Pittsburgh#8 coal composition of the syngas was lower than when using the other o 65 at 800 C with 1:1 alkali to carbon ratio was as follows: H 2 = catalysts or no catalyst. This suggests that the alkali catalyst can 43%, CH 4 = 44%, C 2H4-6 = 11%, CO = 2%. So, while there is a lower the activation for the methane reforming. As mentioned significant amount of hydrogen in the syngas on a molar basis, 25 earlier, potassium hydroxide/carbonate has been shown to lower the hydrogen and methane represent nearly the same combustion the activation barrier energy of the methanation reaction [15-16], enthalpy. The highest average CH 4 composition achieved using and hence, can lower the activation barrier energy of the reverse 70 Wyodak-Anderson coal was 17% while at the same time reaction, i.e. methane reforming. One possible reason why there producing 1.9% C 2H4-6. On a combustion enthalpy basis, the was a lower methane composition in the syngas when operating syngas from the Wyodak-Anderson coal at 800 oC with 1:1 alkali 30 with catalysts compared to when operating without catalysts, as to carbon ratio is as follows: H 2 = 46%, CH 4 = 37%, C 2H4-6 = seen in Tables 3 and 4, is that the catalysts lower the activation 11%, CO = 6%. energy barrier for the methanation and methane reforming pair of 75 The amount of CO 2, and to a lesser extent the amount of H2S reactions. Since the methane and ethane composition due to coal and CO, increased with increasing temperatures, whereas the CH 4 volatilization is often greater than if the syngas were in chemical o and C2H4-6 composition were a maximum at 800 C for

Downloaded by Carnegie Mellon University on 12 July 2012 35 equilibrium, the catalysts have the unintended effect of actually Pittsburgh#8 coal. The maximum of the CH 4 and C 2H4-6 Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A decreasing the methane and ethane composition in the syngas. At composition at 800 oC is due to a combination of effects: these temperatures, this can be avoided by operating at higher 80 temperature, steam composition, and pyrolysis reactions. For pressures than studied here. example, at higher temperatures, there was less steam available to

Table 3: Syngas composition and rate of reaction versus catalyst type to reform the methane and ethane into CO and H 2 because of the 60% conversion. 100 mesh Pittsburgh#8 coal, 2.1 MPa, 700 oC. faster kinetic rates of steam-coal gasification. This, along with increased amount of methane pyrolysis, can offset the decrease in 85 equilibrium methane composition that would be due solely to increased temperatures if the water vapour composition were constant.

While the greatest amount of CH 4 and C 2H4-6 occurred at 800 oC for Pittsburgh#8 and 700 oC for Wyodak-Anderson, the o 90 largest rate of total syngas formation occurred at 900 C for both coal types. We graph the rate constants shown in Tables 5 and 6 Table 4: Syngas composition and rate of reaction versus catalyst type to o for experiments using fresh Pittsburgh#8 coal at a pressure of 2.1 40 60% conversion. 100 mesh Wyodak-Anderson coal, 2.1 MPa, 700 C. MPa in Figure 5, which plots the first order rate constant (from NRCR = 90% to NRCR = 40%) versus the inverse of the 95 temperature. For all catalysts, the clear, and expected, trend was

that the rate constant increased as the temperature increased. The ManuscriptEnergy & Environmental Accepted Science use of calcium oxide or alkali hydroxide as a catalyst also significantly increased the rate of reaction compared with the cases with no catalyst. For comparison, when steam-gasifying o 100 graphite at 900 C and 2.1 MPa, the first order rate constant was only 0.014 hr -1, which is just under two orders of magnitude

This journal is © The Royal Society of Chemistry [year] Journal Name , [year], [vol] , 00–00 | 7 Energy & Environmental Science Page 8 of 13 View Online slower than the rate of steam-gasification of Pittsburgh#8 coal at a similar pressure, temperature and size of particles. While the general trend from the activation barrier energies makes sense, it should be noted that these values are the apparent Table 5: Syngas composition and rate of reaction versus temperature to activation barrier energies and should not be interpreted as the 60% conversion. Pittsburgh#8, 1:1 alkali to carbon ratio, 2.1 MPa. 35 intrinsic activation barrier energies for the rate limiting reactions.

Standard It would be challenging to obtain one single activation energy

Temperature H2 CH 4 C2H4-6 CO 2 COH2S Rate Deviation barrier in a fuel as complex as coal, and therefore, the values of Rate calculated above may represent an average activation energy [oC] [%] [%] [%] [%] [%] [ppm] [1/hr] [1/hr] barrier that lumps together the multiple reactions occurring in 900 65 20 2.6 7.1 5.4 28 1.4 +/- 0.4 40 parallel and in series near the surface of the coal. For example, 800 66 22 2.6 5.0 3.6 148 1.1 +/- 0.4 700 74 16 2.9 2.9 3.3 13 0.7 +/- 0.2 the activation barrier energy calculated in the alkali hydroxide o 600 82 13 2.8 1.3 1.3 0 0.13 +/- 0.02 case does not include the data at 600 C, which shows a significant decrease in reaction rate, much larger than implied by Table 6: Syngas composition and rate of reaction versus temperature to the activation barrier energy as suggested for data between 700 oC 5 60% conversion. Wyodak coal, 1:1 alkali to carbon ratio, 2.1 MPa o 45 and 900 C. Therefore, the activation barrier energies reported

Standard above are not applicable outside of this temperature range, and in

Temperature H2 CH 4 C2H4-6 CO 2 COH2S Rate Deviation addition, these values may be unique to the fixed-bed gasifier of Rate design used in these experiments if the chemical reactions were o [ C] %vol %vol %vol %vol %vol [ppm] [1/hr] [1/hr] mass transport limited. 900 73 17 1.9 2.1 6.8 10 2.4 +/- 0.6 800 70 16 2.3 4.5 7.7 4 2.0 +/- 0.4 700 73 17 2.3 0.5 8.0 2 1.9 +/- 0.2 50 Table 7: Effective activation barrier energy in kJ/mol for both coal type and different type of catalyst. The lime case is a ratio of CaO to carbon of 0.5:1, and alkali hydroxide case is a ratio of alkali to carbon of 1:1. 10 By assuming that the gasification reactions follow an Effective Activation Alkali Arrhenius rate equation, we estimated an activation energy barrier No Catalyst Lime for the following cases: without catalyst, with lime catalyst Barrier Energy [kJ/mol] Hydroxide 62 ±13 52 ±15 37 ±12 having a 0.5:1 Ca:C ratio, and with alkali hydroxide catalyst Pittsburgh#8 Wyodak-Anderson 55 ±11 15 ±8 12 ±6 mixtures having Li:Na:K:C ratios of 0.33:0.33:0.33:1. These 15 values for both Pittsburgh#8 and Wyodak-Anderson coals are summarized in Table 7. The trend of activation barrier energies was expected because (a) a singly charged cation can ion- 55 3.4 Effect of catalyst to coal ratio exchange into more acid groups in coal than a doubly charged Tables 8 and 9 show the syngas composition and the reaction cation and (b) the alkali catalyst is in the molten state whereas the rate as a function of the amount of alkali catalyst used. We also 20 alkali earth catalyst remains solid. There is large uncertainty in include experimental data using potassium carbonate as the

Downloaded by Carnegie Mellon University on 12 July 2012 the calculated values of activation energy barrier because the catalyst instead of the mixture of alkali hydroxides. Two values

Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A uncertainty at each temperature propagates into the uncertainty of 60 of catalyst to carbon ratio were used with potassium carbonate. the energy barrier. The 4.6% potassium carbonate is a similar catalyst to carbon ratio as used by Exxon in their catalytic fluidized bed gasifier. The other value was chosen to be similar to 0.25 alkali hydroxide to carbon ratio, so as to highlight the point that potassium carbonate 65 can be used to capture sulphur species, but it can’t be used for in situ capture of carbon dioxide. In the Pittsburgh#8 case, the alkali-to-carbon ratio was varied between 0 and 2, and in the Wyodak-Anderson case, the alkali-to-carbon ratio was varied between 0 and 1. In both cases, there is an increase in carbon 70 dioxide and hydrogen sulphide capture with increasing amounts of alkali hydroxide catalyst. When the alkali-to-carbon ratio was equal to one for both coal types, the alkali hydroxide catalyst captured 99.7% of the sulphur that would have been released if there was not capture agent. Capture of acid gases leads to greater 75 concentration of hydrogen in the syngas because of the increased amount of water vapour available for gasification. There is also a

25 general trend of increasing reaction rate with increasing catalyst Fig. 5. Rate constant versus inverse temperature for various ManuscriptEnergy & Environmental Accepted Science catalysts, Pittsburgh#8 coal, 2.1 MPa, to 60% coal conversion. amount; however, in the Pittsburgh#8 case, there seems to be a Catalysts were added at an equal capability to capture CO 2. limit to which increasing the catalyst amount can increase the Ratio of alkali hydroxide to carbon = 1:1, and ratio of CaO to 80 reaction rate. In the Pittsburgh#8 case, there is a trend of carbon = 0.5:1. Error bars represent the standard deviation of the decreasing methane and ethane composition with increased 30 gasification rate for the three runs at each data point. catalyst addition, but this trend is largely absent in the Wyodak-

8 | Journal Name , [year], [vol] , 00–00 This journal is © The Royal Society of Chemistry [year] Page 9 of 13 Energy & Environmental Science View Online Anderson data. This is probably due to the increase in water vapour entering the reactor normalized by the original gasification rate, and hence the decrease in average water vapour 30 moles of carbon in the coal. The clear trend was an increase in the composition in the syngas. Too much water vapour in the gasifier amount of hydrogen and a decrease in the amount of methane in will cause reforming of hydrocarbons, which is to be avoided if the syngas with an increase in the rate of water injection to the 5 the goal of the gasifier is to produce a high energy content gasifier. In this case, the water vapour was reforming methane syngas. From the data in these tables, it appears that potassium and ethane into hydrogen within the gasifier, but this did not

carbonate and alkali hydroxide mixtures are equally effective at 35 cause an increase in the total amount of CO and CO 2, implying catalyzing the steam-coal gasification reaction, as well as there was an increase in the amount of CO 2 capture with capturing hydrogen sulphide. The main difference is the increased levels of water vapour. Quite interestingly, the first-

10 capability for hydroxides to capture CO 2 in the gasifier. In section order rate constant [1/hr] was often larger than the normalized 4.2, we will compare the experimental data and the chemical water flow rate [1/hr]. This is partially due to the fact that equilibrium simulations in Tables 8 and 9. 40 volatilization reactions do not require water vapour, but it also implies that the water required for steam-coal gasification Table 8: Syngas composition and rate of reaction versus catalyst reaction is being supplied by the hydroxide species in the alkali o amount. Pittsburgh#8 coal to 60% conversion, 2.1 MPa, 700 C. hydroxide. For example, in the case of zero water addition, there

(a) Experimental Data is significant coal consumption without any water addition, other Standard 45 than water that evolves from an alkali hydroxide species when Ratio of Alkali Metal in H2 CH 4 C2H4-6 CO 2 CO H2S Rate Deviation Catalyst to Carbon in Coal [%] [%] [%] [%] [%] [ppm] [1/hr] of Rate alkali hydroxide species either (a) capture acid gases, (b) cation [1/hr] exchanges with a proton on an acid group in the coal, or (c) 2.0 (Eutectic Hydroxide) 78 14 2.2 1.0 4.3 8 0.6 +/- 0.2 1.0 (Eutectic Hydroxide) 73 16 2.8 4.0 3.5 13 0.6 +/- 0.2 convert into an alkali oxide (such as Na 2O). Within the standard 0.5 (Eutectic Hydroxide) 63 20 3.5 6.3 6.5 119 0.6 +/- 0.2 deviation in the data, the results in Table 10 suggest that the 0.25 (Eutectic Hydroxide) 61 23 3.4 10 2.7 106 0.6 +/- 0.2 0.28 (Potassium Carbonate) 36 22 3.5 21 18 35 0.68 +/- 0.06 50 amount of water vapour does not have a strong effect on the rate 0.125 (Eutectic Hydroxide) 42 22 3.7 17 15 108 0.54 +/- 0.10 of reaction, though it should be noted that water vapour would be 0.046 (Potassium Carbonate) 41 19 2.8 26 12 689 0.32 +/- 0.04 0.0 (No Catalyst) 36 27 5.1 17 15 5625 0.19 +/- 0.05 required to achieve 100% conversion of the coal.

(b) Chemical Equilibrium Simulations Table 10 : Syngas composition and rate of reaction versus water flow rate. Pittsburgh#8 coal to 60% conversion, 700 oC, 1:1 alkali to carbon ratio . Ratio of Alkali Metal in CO H2 CH 4 C2H4-6 CO 2 Water flow rate is given in units of the mol/mol C/hr of water vapour Catalyst to Carbon in Coal [%] [%] [%] [%] [%] entering the reactor normalized by the original carbon in the coal. 2.0 (Eutectic Hydroxide) 99.9 0.06 0.0000 0.01 0.01 1.0 (Eutectic Hydroxide) 97 2 0.0000 1 0 Standard 0.5 (Eutectic Hydroxide) 72 13 0.0004 10 4 Water Flow Rate H2 CH 4 C2H4-6 CO 2 CO H2S Rate Deviation of Rate 0.25 (Eutectic Hydroxide) 61 13 0.0004 18 7 [mol H 2O / mol C / hr] %vol %vol %vol %vol %vol [ppm] [1/hr] [1/hr] 0.28 (Potassium Carbonate) 51 13 0.0006 26 9 0.125 (Eutectic Hydroxide) 59 11 0.0003 22 7 1.12 79 14 2.2 3.1 1.5 11 0.4 +/- 0.1 0.046 (Potassium Carbonate) 60 7 0.0001 26 7 0.56 74 16 2.9 2.9 3.3 13 0.7 +/- 0.2 0.28 69 19 4.0 3.5 4.0 20 0.4 +/- 0.1 0.0 (No Catalyst) 59 7 0.0001 26 7 0.0 65 25 3.5 1.5 5.2 11 0.6 +/- 0.2 Downloaded by Carnegie Mellon University on 12 July 2012

Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A 15 Table 9: Syngas composition and rate of reaction versus catalyst amount. 3.6 Effect of pressure Wyodak-Anderson coal to 60% conversion, 700 oC, 2.1 MPa. (a) Experimental Data 55 In this set of experiments, both Pittsburgh#8 and Wyodak- Anderson coals were de-volatilized before use in the reactor Ratio of Alkali Metal in CO Rate Standard H2 CH 4 C2H4-6 CO 2 H2S because we wanted to rule out any change in the amount of Catalyst to Carbon in Coal [%] [%] [%] [%] [%] [ppm] [1/hr] Deviation of Rate [1/hr] volatile gases as a function of pressure. At 700 oC and at 2 MPa, 1.0 (Eutectic Hydroxide) 73 16 2.3 0.8 7.9 2 1.9 +/- 0.2 0.5 (Eutectic Hydroxide) 69 14 2.2 8.2 7.3 2 1.6 +/- 0.3 both coals released 30% of their chemical oxygen demand after 0.25 (Eutectic Hydroxide) 55 11 2.3 23 9.3 148 0.35 +/- 0.05 60 de-volatilization for five hours. On a molar basis, the pyrolysis 0.28 (Potassium Carbonate) 47 13 1.6 30 9.0 112 0.34 +/- 0.05 gas composition was 41% CH , 7% H , 5% CO , 36% CO, 9% 20 0.0 (No Catalyst) 38 20 3.0 28 12 722 0.21 +/- 0.05 4 2 2 C2H4-6 and 2% H 2S for the Pittsburgh#8 coal and was 26% CH 4, (b) Chemical Equilibrium Simulations 23% H 2, 23% CO 2, 22% CO, 6% C 2H4-6 and 0.4% H 2S for the Ratio of Alkali Metal in CO H2 CH 4 C2H4-6 CO 2 Wyodak-Anderson coal. As far as the original chemical oxygen Catalyst to Carbon in Coal [%] [%] [%] [%] [%] 65 demand of the coal, this represents: 18% CH , 0.8% H , 4.0% 1.0 (Eutectic Hydroxide) 79 20 0.0008 0.5 0.8 4 2 0.5 (Eutectic Hydroxide) 75 18 0.0007 4.1 2.7 CO, and 7.0% C 2H4-6 for Pittsburgh#8 and 12% CH 4, 2.6% H 2, 0.25 (Eutectic Hydroxide) 64 9.3 0.0002 20 6.5 2.4% CO, and 4.7% C 2H4-6 for Wyodak-Anderson. 0.28 (Potassium Carbonate) 55 10 0.0003 27 8.4 Table 11 shows the syngas composition, as well as the kinetic 0.0 (No Catalyst) 62 4.6 0.0001 27 6.5 rates of steam-coal gasification, as a function of pressure of the o 70 vessel, when the temperature was held constant at 700 C. As ManuscriptEnergy & Environmental Accepted Science 25 3.5 Effect of steam to coal ratio expected, there was an increase in the methane concentration as the pressure was increased. In addition, increased pressure also Table 10 shows the syngas composition and the first-order increased the reaction rate constants. reaction rate constant as a function of the ratio of steam to carbon

in the coal. The water flow rate is given in units of the mol/hr of

This journal is © The Royal Society of Chemistry [year] Journal Name , [year], [vol] , 00–00 | 9 Energy & Environmental Science Page 10 of 13 View Online Table 11 : Syngas composition and rate of reaction versus pressure. Pre- Where is equal to the mass of the capture agents added o devolatized Pittsburgh#8 coal to 60% conversion, 700 C, 1:1 alkali to to the reactor in [g], and is equal to the enthalpy released carbon ratio. in CO 2 capture [kJ/g]. Since the MCGE in this case is nearly Standard 100%, this means that the gasifier could be operated without Pressure H2 CH 4 C2H4-6 CO 2 CO H2S Rate Deviation of 45 [MPa] [%] [%] [%] [%] [%] [ppm] [1/hr] Rate [1/hr] oxygen and without heat addition, as long as heat transfer from 0.1 80 1.8 1.2 11 5.6 23 0.42 +/- 0.10 the gasifier to the surroundings can be minimized. Given the 0.4 81 2.9 1.6 13 2.4 N/A 0.39 +/- 0.04 success with in situ H2S and CO 2 capture in a gasifier that does 1.0 87 8.1 2.4 1.5 2.0 4 0.59 +/- 0.28 not require oxygen addition, there is the need for the development 2.1 80 13 2.6 3.5 2.3 4 0.64 +/- 0.08 of scalable techniques for regenerating alkali hydroxide catalysts 50 from the alkali sulphide and carbonate. This is a focus of our continuing and future work. 4. Discussion & Analysis 4.1 Cold gas efficiency 4.2 Comparison with chemical equilibrium simulations We first address the general question: what is the efficiency in In this section, we compare the experimental results with 5 converting the coal into syngas using a molten hydroxide 55 thermodynamic simulations using HSC Chemistry (Outotec catalytic gasifier? Since the flow rate data presented in Figure 4 Solutions, Espoo, Finland), which calculates the chemical (900 oC, 2.1 MPa, 1:1 alkali-to-carbon) was collected up to ~1% equilibrium gas composition as a function of temperature, coal reduction charge remaining (NRCR), we calculated the cold- pressure, and inlet species. By comparing the experimental results gas efficiency of this reactor at this temperature and pressure to with simulation, we were able to make some general conclusions 10 be 115%. This means that the chemical enthalpy of the syngas 60 about the degree to which the output syngas was in chemical leaving the molten hydroxide gasifier is greater than the chemical equilibrium. In HSC Chemistry, we modelled Wyodak-Anderson enthalpy in the original coal, even though the total reduction (CH 0.86 O0.18 ) and Pittsburgh#8 (CH 0.8 O0.08 ) using combinations of charge of the syngas was nearly exactly equal to the original anthracene (C 14 H10 ), benzyl benzoate (C 14 H12O2.0), and water reduction charge of the coal. The equation used to calculate the vapour in order to obtain the same ratio of C:H:O as the coal. We 15 cold gas efficiency is the following: 65 did not include sulphur or nitrogen in the coal in these HSC Chemistry simulations; however, for some simulations listed below we included the alumina-silicates in the coal in order to 242 283 803 1428 CGE estimate the amount of catalyst deactivated by the ash in the coal. ∙ When nitrogen and sulphur can be ignored, the equilibrium (6) 70 composition of gases in a hypothetical molten catalytic gasifier is only a function of pressure, temperature, H/C, O/C, and X/C Where CGE is the cold gas efficiency, is equal to the mass (where H, C, O, and X are the amounts of hydrogen, carbon, of the coal added to the reactor in [g], is equal to the higher oxygen and alkali elements entering the gasifier, respectively.) 20 heating value of coal in [kJ/g] at standard conditions, and where o

Downloaded by Carnegie Mellon University on 12 July 2012 Using the data from Figure 4 (900 C, 2.1 MPa), the average is the flow rate in [mol/s] of the syngas species that can be 75 composition to 99% conversion was 63% H , 14% CH , 3% Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A oxidized. The reason that the cold-gas efficiency is greater than 2 4 C H , 10% CO, and 10% CO on a dry molar basis. However, 100% is that steam-coal gasification reactions are endothermic. 2 4-6 2 the predicted equilibrium gas composition using HSC Chemistry Some of thermal energy required to offset this endothermic was 74% H 2, 4% CH 4, 0% C 2H4-6, 16% CO and 5% CO 2 on a dry 25 reaction is supplied by the exothermic CO capture reaction 2 molar basis. As expected, a chemical equilibrium solver will inside of the gasifier. For example, at 1000K, the CO 2 capture by 80 drastically underestimate the composition of C + hydrocarbons in sodium hydroxide releases -174 kJ/ mol CO : 2 2 the syngas because volatilization is a reaction that occurs far from

chemical equilibrium. These hydrocarbon de-volalitization CO (g)+2KOH(l) ↔ K CO (l) + H O(g) H1000K = -179 kJ / mol 2 2 3 2 reactions are highly dependent on the surface properties of the 30 (7) 38 coal. In the experimental reactor, many of the C2+ hydrocarbons 85 that volatilize from the coal exit the reactor before they can If one divides the total chemical combustion enthalpy in the chemically react, even when there is a significant amount of exit syngas by the combustion enthalpy in the coal plus the alkali catalyst. The methane and ethane reforming reactions appear to be kinetically limited, even though there are significant enthalpy of CO 2 capture, one obtains a modified cold gas 35 efficiency (MCGE) of 97%, i.e. 97% of the chemical enthalpy in quantities of alkali hydroxides in the reactor. Similar results have the coal and hydroxides is converted into chemical enthalpy of 90 been found in fixed-bed reactors, in which two-bed, partial- the exiting syngas. The equation used for the MCGE is the equilibrium models often have to be created in order to model the 39 following: gas composition leaving the reactor. Energy & Environmental Science Accepted ManuscriptEnergy & Environmental Accepted Science In addition, it appears the CO 2 capture reactions are kinetically limited. For example, from the experimental results in Figure 4, 242 283 803 1428 MCGE 95 there was a total of 0.19 mol of C that exited the reactor either as ∙ ∙ CO 2, CO, CH 4 or C 2H4-6. So, of the original 0.31 mol of C in the 40 (8) coal, 0.12 mol of C was captured by the molten alkali hydroxides. Since there was 0.31 mol of alkali hydroxides added to the

10 | Journal Name , [year], [vol] , 00–00 This journal is © The Royal Society of Chemistry [year] Page 11 of 13 Energy & Environmental Science View Online reactor, there could have been 0.155 mol of C captured. This material is deficient in oxygen to be composed only of carbon

means that there is incomplete capture of CO 2 by the alkali 55 and alkali carbonates. However, the ratio of alkali:carbon:oxygen hydroxides. We calculated a capture efficiency using the listed above could be formed from a combination of carbon, following equation: alkali carbonate, and alkali oxide. 5 Using the ‘with steam’ results from Fig. 4 out to ~100% conversion and assuming water gas shift equilibrium, we were 60 able to determine that the material remaining in the reactor had virtually no hydrogen had an alkali:carbon:oxygen ratio of 1.8 : ½ ∙ 1.0 : 3.0, which is similar to the elemental ratio of alkali (9) carbonate to within the uncertainty in measurements. As expected, if there is sufficient water vapour in the reactor, the Using the data from Figure 4 (900 oC, 2.1 MPa), we measure a 65 material remaining in the reactor will be alkali carbonates. If capture efficiency of 80%. This should be compared with there is insufficient water vapour, the ‘no steam’ results in Table estimates from HSC Chemistry, which predict roughly 90% 10 suggest that the material remaining in the reactor will be a 10 capture efficiency. The reason that HSC Chemistry predicts only combination of unconverted carbon, alkali carbonate and alkali 90% capture efficiency is that (a) species in the coal (such as oxide. alumina-silicates and sulphur) will react with the alkali 70 Alkali carbonates and hydroxides are more stable than alkali hydroxides, and (b) there will still be small amounts of alkali oxides at these temperatures, given that the change in Gibbs free hydroxides in chemical equilibrium. Therefore, the fact that the energy is positive for each of the individual alkali hydroxide 15 experimental capture efficiency is less than the capture efficiency decomposition reactions. predicted by chemical equilibrium suggests that the capture

reaction is kinetically limited. 1000K 75 2 LiOH(l) ↔ Li 2O(l) + H2O(g) G = + 89 kJ/mol We now present HSC Chemistry results as a function of the 1000K 2 NaOH (l) ↔ Na 2O(l) + H2O(g) G = +107 kJ/mol catalyst to carbon ratio so that we can compare with the 1000K 2 KOH (l) ↔ K2O(l) + H2O(g) G = + 292 kJ/mol 20 experimental results as a function of the catalyst to carbon ratio.

Table 8 (b) and Table 9 (b) list the simulated equilibrium gas The large positive values of the alkali hydroxide compositions using same amount of coal, alkali hydroxide and 80 decomposition reactions suggest that the alkali hydroxides will water vapour is in the experiments in Table 8 (a) and Table 9 (a). not convert to bulk alkali oxides species. Though, it is likely that As expected, there are significant differences between the alkali oxides species can form on surface of the coal surface or in 25 experimental results and the chemical equilibrium simulations. the molten bed of alkali hydroxides and carbonates. We were By comparing these tables, we can make the following unable to model interactions of alkali oxides with surfaces or in conclusions. (1) The methane and ethane compositions were 85 the molten mixtures using HSC Chemistry. much higher than would be expected if the gasifier were at 40 Prior research into catalytic gasification by Wood et al. and chemical equilibrium. This means that most of the methane and Wigmans et al. 41 concluded that carbon will develop a surface 30 ethane products were volatilization products of the coal that did

Downloaded by Carnegie Mellon University on 12 July 2012 with a sub-stoichiometric amount of oxides and with an excess of not react before leaving the reactor. (2) There was less CO 2 12

Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A the alkali metal. Mims and Pabst concluded that the active capture measured experimentally than would be expected if the 90 catalyst sites are complexes similar to alkali phenolate. The ‘no gasifier were at chemical equilibrium. This suggests that the CO 2 steam’ results in Table 10 suggest that gasification of coal can capture reaction was kinetically limited. proceed in the absence of water vapour or carbon dioxide because 35 alkali hydroxides can convert to alkali oxide species either on the 4.3 Coal gasification without steam addition surface of the remaining coal or in the molten bed. 95 It is important to emphasize the interesting result in Table 10 of coal conversion without any water, just alkali hydroxides. This 5. Conclusions experiment, like all other experiments, was repeated 3 times, and 40 in each case, there was greater than 60% coal conversion, i.e. less We demonstrated a process for turning coal into a syngas than 40% reduction charge remaining. In fact, the rate of syngas consisting of roughly 20% methane and 80% hydrogen using production was three times larger than the production of syngas alkali hydroxides as both catalysts and in situ capture agents of

from the Pittsburgh#8 coal without catalyst but with steam. As 100 acid gases, such as CO 2 and H 2S. We measured significant discussed in section 3.6, only 30% of the chemical oxygen capture of both H 2S and CO 2 by the alkali hydroxides throughout o o 45 demand was released during pyrolysis of coal in the absence of the temperature range studied, 600 C to 900 C. As expected, the alkali hydroxides. To achieve 60% conversion of the coal and to hydrocarbon composition increased with pressure; the kinetic

have the gas composition listed in Table 10, the material inside of rates increased with temperature; and the amount of CO 2 capture

reactor after the experiment would contain 40% of the chemical 105 increased with an increasing ratio of alkali hydroxide to carbon in ManuscriptEnergy & Environmental Accepted Science oxygen demand of the original coal, contain virtually no the coal. We compared the experimental results with chemical 50 hydrogen, and have an alkali:carbon:oxygen ratio of 1.0:0.7:1.0. equilibrium simulation from HSC Chemistry. By comparing This ratio of alkali:carbon:oxygen implies that the material results, we can make the following conclusions: (1) The methane remaining in the reactor is not solely composed of some and ethane composition was much higher than would be expected combination of unconverted carbon and alkali carbonates. The 110 if the gasifier were at chemical equilibrium, implying that

This journal is © The Royal Society of Chemistry [year] Journal Name , [year], [vol] , 00–00 | 11 Energy & Environmental Science Page 12 of 13 View Online hydrocarbon reforming was kinetically limited; (2) There was 16. R. L. Hirsch, J. E. Gallagher, R. R. Lessard and R. D. Wesselhoft,

less CO 2 capture experimentally than would be expected if the Science, 1982, 215, 121-127. gasifier were at chemical equilibrium, implying that CO 2 capture 17. N. C. Nahas, F uel, 1983, 62, 239-241.

was kinetically limited. In addition, we demonstrated 60% 60 18. http://www.greatpointenergy.com/ , accessed June 21, 2012 5 conversion of Pittsburgh#8 coal when there was an equal molar 19. G. P. Curran, C. E. Fink and E. Gorin, A dv Chem Ser, 1967, 141-&. amount of alkali hydroxide and carbon in the coal, but with no 20. K. Kuramoto, K. Ohtomo, K. Suzuki, S. Fujimoto, S. Shibano, K. o addition of water into the system. At temperatures between 800 C Matsuoka, Y. Suzuki, H. Hatano, O. Yamada, L. Shi-Ying, M. o and 900 C, we measured first-order steam-coal gasification rates Harada, K. Morishita and T. Takarada, I nd Eng Chem Res, of 2 hr -1 in a fixed bed reactor while capturing significant 65 2004, 43, 7989-7995. 10 quantities of both H S and CO , and while also generating 20% 2 2 21. J. Fermoso, F. Rubiera and D. Chen, E nerg Environ Sci, 2012. methane plus ethane in the syngas on a dry molar basis. The 22. A. E. Cover, Schreine.Wc and Skaperda.Gt, A bstr Pap Am Chem S, overall conclusion is that molten hydroxide gasification of coal is 1971, 14-&. a promising route for generating a methane-rich syngas that can 23. A. E. Cover, Schreine.Wc and Skaperda.Gt, C hem Eng Prog, 1973, be used in solid oxide fuel cells or in gas turbines to generate 70 69, 31-36. 15 electricity with minimal to near zero emissions of acid and greenhouse gases. 24. C. A. Kumar, L. D. Fraley and S. E. Handman, A bstr Pap Am Chem S, 1975, 170, 39-39. 25. C. A. Trilling, A bstr Pap Am Chem S, 1977, 173, 17-17. 6. Acknowledgements 26. G. Jin, H. Iwaki, N. Arai and K. Kitagawa, E nergy, 2005, 3 0, 1192- 75 1203. We thank the National Energy Technology Laboratory for their 27. J. Matsunami, S. Yoshida, Y. Oku, O. Yokota, Y. Tamaura and M. 20 support of this research. In particular, we thank Tristan McQuain, Kitamura, E nergy, 2000, 25, 71-79. Jack Ferrel, Richard Bergen, Randy Barnes, and William Grimes 28. T. Kamo, K. Takaoka, J. Otomo and H. Takahashi, F uel, 2006, 8 5, for their expertise during the operation of the molten catalytic 1052-1059. gasifier. 80 29. H. Nagasawa, A. Yamasaki, A. Iizuka, K. Kumagai and Y.

Yanagisawa, A iche J, 2009, 55, 3286-3293. 30. Y. Tanaka, I on Exchange Membrane Electrodialysis: Fundamentals, 25 Notes and references Desalination, Separation, Nova Science Pub Inc, 2010. a National Energy Technology Laboratory, US Dept of Energy, 31. T. W. Xu and C. H. Huang, A iche J, 2008, 54, 3147-3159. Pittsburgh, PA USA; Tel: (412) 386-4404; [email protected] 85 32. K. S. Vorres, E nerg Fuel, 1990, 4, 420-426. b National Energy Technology Laboratory, U.S. Dept of Energy, Morgantown, WV, USA. 33. A. M. Pisarevsky, I. P. Polozova and P. M. Hockridge, R uss J Appl c 30 Department of Mechanical Engineering, Carnegie Mellon University, Chem+, 2005, 78, 101-107. Pittsburgh, PA, USA. 34. D. Orhon and E. U. Cokgor, J Chem Technol Biot, 1997, 68, 283-

293.

Downloaded by Carnegie Mellon University on 12 July 2012 1. K. Gerdes, E. Grol, D. Keairns and R. Newby, DOE/NETL- 90 35. N. Berguerand and A. Lyngfelt, F uel, 2008, 87, 2713-2726. Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A 35 2009/1361, 2009. 36. J. Brix, P. A. Jensen and A. D. Jensen, F uel, 2010, 89, 3373-3380. 2. E. Grol and J. Wimer, DOE/NETL- 40/080609, 2009. 37. D. G. Roberts, A. Y. Ilyushechkin and D. J. Harris, Fuel Process 3. A. Audenaert, L. De Boeck, S. De Cleyn, S. Lizin and J. F. Adam, Technol, 2012, 94, 86-93. Renew Energ , 2010, 35 , 2674-2682. 38. J. F. Espinal, F. Mondragon and T. N. Truong, Carbon, 2005, 43, 4. IEA, "Electricity Information 2011," ISBN 978-92-64-101913. 95 1820-1827. 40 5. U.S. Department of Energy, Energy Information Administration, 39. M. L. Hobbs, P. T. Radulovic and L. D. Smoot, Fuel, 1992, 71, 1177- International Energy Outlook 2006, DOE/EIA-0484 (2006). 1194. 6. http://www.epa.gov/cleanenergy/energy-and-you/affect/municipal- 40. B. J. Wood, R. H. Fleming and H. Wise, F uel, 1984, 63, 1600-1603. sw.html , accessed June 21, 2012. 41. T. Wigmans, R. Elfring and J. A. Moulijn, C arbon, 1983, 21, 1-12. 7. P. K. Cheekatamarla, C. M. Finnerty and J. Cai, Int J Hydrogen 100 45 Energ , 2008, 33 , 1853-1858. 8. S. D. Park, J. M. Vohs and R. J. Gorte, Nature , 2000, 404 , 265-267. 9. EPA/540/R-94/514 July 1995 http://www.p2pays.org/ref/09/08676.pdf , accessed June 21, 2012 10. B. J. Wood and K. M. Sancier, C atal Rev, 1984, 26, 233-279.

50 11. D. W. Mckee, F uel, 1983, 62, 170-175. 12. C. A. Mims and J. K. Pabst, F uel, 1983, 62, 176-179. 13. T. Wigmans, R. Elfring and J. A. Moulijn, C arbon, 1983, 2 1, 1-12. ManuscriptEnergy & Environmental Accepted Science 14. Y. D. Yeboah, Y. Xu, A. Sheth, A. Godavarty and P. K. Agrawal, Carbon, 2003, 41, 203-214.

55 15. A. Sheth, Y. D. Yeboah, A. Godavarty, Y. Xu and P. K. Agrawal, Fuel, 2003, 82, 305-317.

12 | Journal Name , [year], [vol] , 00–00 This journal is © The Royal Society of Chemistry [year] Page 13 of 13 Energy & Environmental Science

View Online

Coal and molten alkali hydroxides were gasified in the presence of water vapour, where the alkali hydroxides capture both carbon dioxide and hydrogen sulphide, yielding a syngas from this batch reactor that was mostly hydrogen and methane.

Downloaded by Carnegie Mellon University on 12 July 2012 Published on 02 July 2012 http://pubs.rsc.org | doi:10.1039/C2EE21989A Energy & Environmental Science Accepted ManuscriptEnergy & Environmental Accepted Science