Comparison between Catalytic Fast Pyrolysis and Catalytic Fast

Subscriber access provided by WEBSTER UNIV

Biofuels and Biomass

Comparison between Catalytic Fast Pyrolysis and Catalytic Fast Hydropyrolysis for the Production of Liquid Fuels in a Fluidized Bed Reactor Devin Scott Chandler, and Fernando Luis P. Resende Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.8b03782 • Publication Date (Web): 05 Mar 2019 Downloaded from http://pubs.acs.org on March 5, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Comparison between Catalytic Fast Pyrolysis and Catalytic Fast Hydropyrolysis for the Production of Liquid Fuels in a Fluidized Bed Reactor Devin S. Chandler, Fernando L.P. Resende* School of Environmental and Forest Sciences, University of Washington, Seattle, WA *Email: [email protected]

Abstract We report results for the conversion of Arundo Donax into liquid transportation fuels with an HZSM-5 catalyst via catalytic fast pyrolysis (CFP, inert atmosphere) and catalytic fast hydropyrolysis (CHP, hydrogen atmosphere), with the goal of comparing the two technologies and understand the effect of hydrogen on the product yields and composition. For CFP, we evaluated the effects of temperature, weight hourly space velocity (WHSV), and pressure. The highest yield of hydrocarbons for CFP at atmospheric pressure was 11.6 C%. The main effects of adding hydrogen (CHP) were a decrease in the yield of coke, an increase in the yield of liquid hydrocarbons, and an increase in gaseous olefins. The introduction of hydrogen also led to a 17 % reduction of coke relative to the coke produced under a helium atmosphere, with reduced pore blockage and surface area loss relative to CFP. Overall, CHP with HZSM-5 at 400C and 35 bar can produce 8.9 gal/ton of liquid hydrocarbons (mostly BTX), as opposed to 6.0 gal/ton from CFP. Keywords: Hydropyrolysis; catalytic; pyrolysis; biomass; biofuels; sustainable; drop-in; hydrocarbons

1 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1) Introduction For over a century, we have primarily fueled our transportation sector with petroleum-based fuels. As petroleum resources are depleted, we will need to explore alternative fuels. One of the main technologies for producing alternative fuels is fast pyrolysis, which leads to a liquid brownish fuel referred to as bio-oil [1–9]. While bio-oil from fast pyrolysis can be used for heating and power applications, it is not a proper liquid fuel for the transportation sector [10]. For this application, the fuel will need to be compatible with our current liquid-fuel infrastructure, cheap, and readily available. Catalytic fast pyrolysis (CFP) and catalytic fast hydropyrolysis (CHP) of biomass can make such a fuel. In CFP, lignocellulosic polymers (cellulose, hemicellulose, and lignin) break down into oxygenated volatiles at temperatures between 300 and 600C in an inert atmosphere. The presence of a zeolite catalyst, such as HZSM-5, promotes deoxygenation reactions, turning the oxygenated volatiles into compatible hydrocarbon fuels, mainly aromatics, such as benzene, toluene, and xylene (abbreviated BTX). However, if the hydrocarbons fail to desorb, they oligomerize into coke and remain in the pores of the catalyst, blocking the active sites [11]. CFP can be performed either in a single unit (in-situ) or in a sequence of two units (ex-situ). In Catalytic Fast Hydropyrolysis (CHP), this same process is performed in the presence of pressurized hydrogen gas. CHP offers the advantage of a reactant gas that could reduce coking, increase yields, and shift the composition towards alkanes and


Page 2 of 40

Page 3 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

cycloalkanes, making a hydrocarbon mixture with a similar composition to jet fuel. With a sufficiently abundant, cheap feedstock, CFP and CHP could meet much of the transportation needs of our society. Arundo Donax, also known as Giant Reed, is a fast-growing herbaceous weed. It requires little input, and grows at a rate of 37.7 tons per hectare per year, which is 34.8% faster than even Miscanthus x giganteus [12]. Furthermore, there is a need to eliminate it from the habitats it is invading [13]. The fast growing rate of the feedstock could be what is necessary to economically supply fuel for our transportation sector in the near future, while simultaneously eliminating it from sensitive ecosystems. For these reasons, Arundo Donax was selected as feedstock for the present study. Non-catalytic hydropyrolysis studies have shown that hydrogen by itself has a small impact on the products of pyrolysis. Meesuk et al [14] found that more gases are produced, at the expense of bio-oil and char. They hypothesize that the char undergoes hydrogasification in this environment. In the presence of an HZSM-5 catalyst, it has been reported that hydrogen does not increase liquid hydrocarbon production in semi-batch systems such as Py-GC/MS [15,16], but in continuous hydropyrolysis reactors it may limit solids production by hydrogasification of coke and thus could improve yields of hydrocarbons [14]. To the best of our knowledge, the first instance of a catalyst being used to upgrade bio-oil in-situ (without hydrogen) to more useful hydrocarbon products occurred in 2000, when Olazar et al used HZSM-5 to convert sawdust to liquid oil in a conical spouted bed reactor [17]. The yield of liquid hydrocarbons was 8.20 wt.%. Huber et al demonstrated the


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

effect of temperature from 500 to 670°C and WHSV from 0.1 to 1.7 h-1 of the catalytic fast pyrolysis of sawdust and found that as WSHV increased, toluene and xylene yields decreased, while naphthalene yields increased. They also found a maximum of total yield of aromatics of 14.0 C% at 600°C and 0.1 h-1. Generally, toluene yields decreased with temperature, while benzene and naphthalene yields increased [11]. In-situ vs. ex-situ CFP was studied by Iisa et al and the results showed that in-situ produced slightly more organic liquid and less coke [18]. In another work, they also found that it is difficult to remove oxygen from CFP oil via hydrotreating without compromising the carbon efficiency of the process [19]. Lappas et al summarized many findings, demonstrating that zeolites showed promise for decreasing the oxygen content of pyrolysis oils [20]. Literature on CHP has only been available for the last few years. Much of it has been thoroughly summarized by Resende [21]. Most studies have been done via PyGC/MS [15,16,22–25]. Gamliel et al tested the effects of catalyst properties on the products of CHP and found that acidic catalyst supports produced more aromatics and metal impregnated acidic supports made more methane and other light alkanes [23]. Thangalazhy-Gopakumar et al studied various transition metals impregnated on HZSM-5 at varying hydrogen pressures for hydropyrolysis at 650°C (coil temperature) [26]. With HZSM-5 without a transition metal, they found that the pressure did not have any effects until 27.5 bar, at which point the xylene yield decreased and the toluene increased. The total yield of hydrocarbons did not change. They noted that this contradicted their expectations, as increasing pressure would increase the presence of hydrogen free radicals and therefore should suppress coking reactions, allowing more aromatics to form. They


Page 4 of 40

Page 5 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

also found that transition metals increased the yield of aromatics by increasing the rate of hydrogenation, but only at pressures higher than 21 bar. Another study from ThangalazhyGopakumar et al also showed little difference in aromatic yields and composition between catalytic fast pyrolysis with HZSM-5 and catalytic fast hydropyrolysis with HZSM-5 at 550°C (coil) and 5.5 bar [15]. It is unclear, however, if this similarity between CFP and CHP remains at higher pressures. Jan et al tested CHP in-situ with HZSM-5 and Pd-HZSM-5 and reported that their results indicate that hydrogen radicals at high temperatures promote demethoxylation, decarbonylation, and dehydration reactions. This effect is accelerated with a transition metal doped HZSM-5 catalyst, which suggests that the transition metal may increase the effects of hydrogen radicals [27]. Zhu et al tested the mechanisms of hydropyrolysis by flowing anisole and hydrogen over bifunctional catalysts and reported the conversion of anisole to hydrocarbons. They found the acidic base, HBeta, promotes transalkylation to produce mostly phenolics, the metal catalyst without an acidic base (Pt/SiO2) promoted demethylation, hydrodeoxygenation, and hydrogenation, yielding phenol, benzene, and cyclohexane. The bifunctional PtHBeta catalyst produced large amounts of BTX by transalkylation and hydrodeoxygenation. They suggest that in the presense of PtHBeta, the aromatic ring is partially hydrogenated near the Caromatic-OH bond by a hydrogen radical at the Pt site. This is followed by rapid dehydration at the acid site which reforms the aromatic bond to produce a deoxygenated benzene ring [28].

CHP of lignocellulosic

biomass in a fluidized bed reactor has been reported by two groups so far [29–31]. Dayton et al tested NiMo on an unidentified catalyst support in a fluidized bed reactor and found a


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

maximum of carbon efficiency of 42.0 C% at 425°C, 20 bar, and 40 vol.% H2. They define carbon efficiency as the carbon yield of C4-C6 vapors plus the carbon yield of organic liquid [29]. The organic liquid at this temperature mostly consisted of polyaromatic hydrocarbons, phenolics, and monoaromatic hydrocarbons. The oxygen content was 2.5 wt.%. Marker et al tested the IH2® system developed at the Gas Technology Institute (GTI) with a variety of feedstocks, but little information was given about the type of catalysts used and the tests varied multiple conditions at a time, making comparisons difficult [30]. Further testing showed that scaling from a bench scale reactor to a 50 kg/day pilot reactor did not reduce the yield or quality of gasoline and jet-fuel range hydrocarbons. The pilot scale reactor was tested for 800 hours without change to the yields [32]. They performed a techno-economic analysis (TEA) of the IH2® process and determined that liquid transportation fuels could feasibly be made at a cost of $1.64/gallon in 2007 USD [31]. The purpose of this manuscript is to compare catalytic fast pyrolysis (performed in an inert environment) and catalytic fast hydropyrolysis (performed in hydrogen). To the best of our knowledge, such a comparison, performed in continuous reactors, is non-existent in the current literature. For CFP, we varied the temperature, weight hourly space velocity (WHSV, the feed rate of biomass divided by the weight of catalyst), and pressure. Then, we performed CHP at 35 bar and 400°C, and report the effect of temperature.

2) Experimental Methods We prepared the biomass by grinding stalks of Arundo Donax into particles less than 1 mm in size, averaging 0.34 mm in diameter, with a hammer mill (Thomas Wiley,


Page 6 of 40

Page 7 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Philadelphia, PA). The HZSM-5 pellets were commercially available (Pingxiang NaikeChemical Industry Equipment Packing Co, Pingxiang, CN) with a Si/Al ratio of 38. The diameter of the pellets was 2-3 mm. We selected the spherical form of the catalyst because it was a convenient choice to achieve fluidization of the mixture catalyst + sand. The sand is necessary to achieve proper fluidization with reasonably uniform temperature across the bed. With this choice, we were able to fluidize the reactor and simultaneously elutriate the product char out of the bed. By proceeding this way, we separated the coked catalyst from the char. No other commercially available catalyst could achieve proper fluidization in this configuration, due to the pneumatic feeding system for biomass, which adds an extra layer of complexity to fluidization. Many other studies use cylindrical pelletized catalyst, but this was not possible in our study. Fluidization characteristics were determined by the Wen and Yu equation (Error! Reference source not found.), which relates the Archimedes number to the minimum fluidization Reynold’s number [33]. We solved Equation 1 for the Reynold’s number, and obtained the minimum fluidization velocity (Umf,i) using the appropriate gas, solid, and reactor geometry parameters for each solid in the reactor mixture. Next, we calculated the combined minimum fluidization velocity (Um,,f,,mixed) according to Error! Reference source not found. [34], in which xi is the mass fraction of each component in the solid mixture. We determined the maximum flow rate of gas (to avoid bed elutriation) experimentally. Our reactor design, as shown in Figure 1, involves the feed of biomass to the fluidized bed pneumatically, as opposed to having the feeder directly connected to the


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 40

reactor. This design choice intended to minimize heat transfer to the biomass particles prior to their entrance into the bed. However, it also led to 6 L/min of the gas fed to the side of the reactor above the plate, which do not contribute to fluidization, but contribute to undesirable elutriation. For this reason, it was necessary to use a fluid velocity three times higher than the theoretical minimum. This characteristic also made it difficult to establish proper conditions to fluidize the catalyst by itself. To overcome this problem, we mixed a spherical catalyst and sand in a 1:1 ratio, which achieved fluidization over our entire range of pressures, without elutriating the bed. 𝐴𝑟 = 1650𝑅𝑒𝑝,𝑚𝑓 + 24.5𝑅𝑒2𝑝,𝑚𝑓 Equation 1: Wen and Yu equation relating Archimedes number to minimum fluidization velocity. P denotes particle, mf denotes minimum fluidization.

1 𝑈𝑚𝑓, 𝑚𝑖𝑥𝑒𝑑

𝑛

=

𝑥𝑖

∑𝑈

𝑖=1

𝑚𝑓,𝑖

Equation 2: Rincon et al equation for minimum fluidization velocity of a mixed medium

For our experiments, the alumina sand is not catalytic. We performed preliminary experiments that showed that the products and selectivity obtained in the fluidized bed (without added catalyst) are the same as what is obtained in pyroprobe experiments, which do not involve sand. We performed BET analysis of the catalyst to determine the surface area and pore dimensions. First, we vacuum-dried the samples using a degasser at 150°C overnight. Then, we weighed the samples and placed them in a Micromeritics TriStar II Plus (Micromeritics, Norcross, GA) BET analyzer to run for 24-48 hours.


Page 9 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

We performed catalytic fast pyrolysis (CFP) of 97-135 g biomass in a fluidized bed reactor system as shown in Figure 1. A magnetic stirrer in the hopper aided the biomass onto the auger. The auger carried biomass from the hopper to a funnel where it was pneumatically forced in the reactor. The reactor bed contained 200 g of 2-3 mm spherical catalyst mixed with 200 g of 100-500 µm alumina sand (Kramer Industries, Piscataway, NJ) to assist in fluidization (biomass to catalyst ratio ranging from 0.5-0.7). We preheated the gas going into the reactor with a heating tape (HTS Amptek, Stafford, TX). An electric furnace (Carbolite Gero, Sheffield, UK) surrounded the reactor to provide heat to the reaction. The products were carried from the reactor to the cyclone via a transfer line, which separated the char. We maintained the transfer line and cyclone temperatures at 275350°C throughout the run with additional heat tapes. The stream then flowed through the condensation system consisting of an impinger, three double-pipe condensers (the last two filled with 6 mm glass beads (Fisher Scientific, Hampton, NH)), and three methanol-filled bubblers. A chiller (Thermo Fisher Scientific, Waltham, MA) provided water at 5°C to the double-pipe condensers. The vapor at the first condenser was usually at 10-20°C in the atmospheric pressure experiments, and 100-150°C in the higher-pressure experiments. The second and third condensers were both below 10°C for the atmospheric pressure experiments. At higher pressures, the second condenser was usually 20-40°C and the third one was around 10°C. We placed the methanol bubblers in a dry-ice bath to ensure maximum collection of products into the methanol. Finally, for atmospheric pressure CFP, the permanent gases went through a coalescing filter (Swagelok, Solon, OH) and were


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

vented from the system. We collected a portion of the vented gases in gas sampling bags (Restek, Bellefonte, PA) throughout the entire run. After the run was completed, we collected the gas bags and analyzed their contents with a GC-TCD/FID 2014 (Shimadzu, Kyoto, JP) using a Supelco 60/80 Carboxen 1000 packed column (Supelco Analytical, Bellefonte, PA). We calibrated for carbon monoxide, carbon dioxide, ethylene, and nitrogen. We collected and weighed liquids in the collectors. We shook their contents thoroughly and dissolved a sample into methanol for GC-MS/FID 2010 (Shimadzu, Kyoto, JP) analysis using a SHRXI-5MS capillary column (30 m × 0.25 mm I.D. × 0.25 μm film thickness) capillary column . The GC inlet temperature was 300 °C, and the inlet split ratio was 30: 1. The oven temperature was programmed from 40 °C to 300 °C with a ramp of 10 °C/min. The initial and final temperatures were held for 4 and 5 min, respectively. We individually calibrated 48 compounds and assumed some of those to have the same response factors as compounds with a similar chemical structure (i.e., indene and indane, dimethylfuran and vinylfuran, etc.). Furthermore, we analyzed samples of the collected liquids in a Mettler Toledo V20 Karl Fisher Titrator (Mettler Toledo, Columbus, OH) for the yield of water. We accounted for liquids that adhered to the interior walls of the tubing of the condenser system by weighing all components before and after the reaction. We then dissolved the adhered liquids in methanol and analyzed them by GCMS/FID. We ran the contents of the bubblers through the GC-MS/FID without further dilution. To calculate the yields, we summed the products of the bubblers, multiplying the obtained concentrations by the amount of methanol in the bubblers, as measured at


Page 10 of 40

Page 11 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

collection. We collected the char from the cyclone separator, transfer line, and reactor. We separated the coked catalyst from the rest of the bed by sieving in order to differentiate coke and char, since the coked catalyst has a much larger diameter (2-3 mm) than the sand and char particles. We burned samples of the char from each collection point at 600°C for 6 hours in air to determine the amount of organic char and the amount of ash or sand. We analyzed samples of each component in a CHN 2400 elemental analyzer (Perkin-Elmer, Waltham, MA) to determine the elemental content of the char. Samples of coked catalyst underwent the same procedures to determine the coke yield and elemental composition. We performed some changes to the reactor system and analysis to make it suitable for high-pressure experiments. For ease of fluidization, we decreased the reactor inner diameter from 3.8 cm to 2.5 cm to increase the gas velocity, and switched from inert nitrogen to inert helium to reduce the density of the gas. Rather than venting the gas at the end of the system, we used a Maximator gas booster (Maxpro Technologies, Fairview, PA) to recycle the gas back to the reactor. This may have affected the gas composition, as the gaseous products went through the reactor multiple times. However, given the rate of production of gases internally, and the high flow rate/high gas pressure of the fluidizing agent (nitrogen or hydrogen), the reaction products (carbon monoxide) become extremely diluted, and the low resulting partial pressures suggest that the effect of recycling on the products collected is likely to be small. Finally, to prevent the methanol bubblers from releasing methanol vapor into the recycle stream and through the reactor system, we replaced the methanol with 3 mm glass beads. We took gas samples at the end of the run instead of throughout the run. We also


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

placed the furnace before the reactor to increase heat transfer to the gas feed since preheating the gas with heating tapes was insufficient at high pressure. High temperature heating tapes (HTS Amptek, Stafford, TX) wrapped around the reactor maintained the reaction temperature through the experiment. For high-pressure experiments, we analyzed the gas samples in a GC/TCD-FID with a HP Plot capillary column (Agilent Technologies, Santa Clara, CA) to account for C3+ vapors, whereas in the low-pressure CFP experiments, we assumed these to be negligible. We did this because we anticipated that CHP would produce C3+ vapors, whereas CFP would not. This was verified by analysis of the C3+ vapors from high pressure CFP. We performed the CFP experiment at 550°C, atmospheric pressure, with HZSM-5, and a WHSV of 1.1 in triplicates to determine the standard deviations. WHSV is defined in Equation 3. We assumed the standard deviations to be independent of the variables tested, and determined statistical significance in the data by two sample t-tests for all reported yields. We performed all analytical testing of products at least two times. Any differences in data that are determined to be insignificant are treated as equal. All data are reported on a carbon yield basis except where noted. 𝑊𝐻𝑆𝑉 (ℎ ―1) =

𝑓𝑒𝑒𝑑 𝑟𝑎𝑡𝑒 𝑜𝑓 𝑏𝑖𝑜𝑚𝑎𝑠𝑠 𝑖𝑛 𝑘𝑔/ℎ 𝑚𝑎𝑠𝑠 𝑜𝑓 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡 𝑖𝑛 𝑘𝑔

Equation 3: Definition of weight hourly space velocity

3) Results and Discussion The characterization of the Arundo Donax used in the present work can be found elsewhere [35]. The same source also presents the results of non-catalytic fast pyrolysis at 12 ACS Paragon Plus Environment

Page 12 of 40

Page 13 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

various temperatures. The main liquid products of non-catalytic fast pyrolysis are oxygenated compounds, mainly acetic acid in the case of Arundo Donax. In this case, the organic liquid yields (bio-oil minus water) are highest at the lowest tested temperature, 400°C (40.5 wt. %). This temperature of highest yields is lower than what is usually observed for most biomass feedstocks, mainly because of the high content of potassium (1.06 wt.%) [35]. Liquid hydrocarbons were not produced in non-catalytic fast pyrolysis of Arundo Donax.

3.1) Catalytic fast pyrolysis (CFP) 3.1.1) Effect of Temperature on Yields of CFP We performed catalytic fast pyrolysis of A. donax in nitrogen gas with HZSM-5, and evaluated the effect of temperature, weight hourly space velocity (WHSV), and pressure. We used 550°C, 1.0-1.2 h-1 WHSV, 1 bar, and HZSM-5 catalyst as the base case experiment, which we performed in triplicate. The yields of products from CFP at varying temperatures are shown in Figure 2. For the base case, the carbon balance was 99.9 C% and the total mass balance was 98.0 wt.%. The carbon yields are as follows: 23.3 C% char, 12.1 C% coke, 15.0 C% known liquid, 21.9 C% unknown liquid (totaling 36.9 C% liquid), and 27.7 C% gas. Known liquids were determined by GC/MS-FID analysis and total liquid carbon was quantified by CHN analysis. We quantified unknown liquid carbon by difference. We believe the unknown liquid to be oligomeric tars from the first impinger and the condenser walls composed of molecules that are too large to be analyzed via GC/MS-FID. The water yield


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

was 22.3 wt.%. The 15.0 C% known liquids are composed of 11.6 C% hydrocarbons, 1.6 C% phenolics, and 1.8 C% other oxygenates such as acetic acid, levoglucosan, and furan. The hydrocarbons produced were 6.4 C% toluene, 1.8 C% benzene, 1.6 C% C8 aromatics (xylenes, ethylbenzene, styrene), 0.3 C% other monoaromatic hydrocarbons (MAHs, C9 and C10 aromatics), and 1.2 C% diaromatic hydrocarbons (DAHs). The gases produced were 12.2 C% carbon monoxide, 9.3 C% carbon dioxide, 3.7 C% ethylene, and 2.5 C% methane. As the temperature increased from 400C to 625C, the total yield of gases increased from 17.8 to 37.1 C% due to increases in the yields of each component of the gas. Carbon monoxide is the most prevalent gas, followed by carbon dioxide. These make up 8.9 C% and 7.3 C% of the yields at 400°C and 18.3 C% and 10.1 C% of the yields at 625°C, respectively. A significant increase occurs in the yield of methane, which increases from 0.7 C% to 4.8 C% when the temperature increases from 400°C to 625°C. This indicates that increasing the temperature increases the prevalence of cracking reactions. We note that, without the presence of HZSM-5, A. donax produced very little ethylene at all (0.0 - 0.4 C%), and the yield of carbon monoxide (3.0 - 6.0 C%) usually matched that of carbon dioxide (3.9 - 6.1 C%), whereas with a catalyst, ethylene is produced in significant quantities (1.0 – 3.9 C%) and carbon monoxide is consistently more prevalent than carbon dioxide [35]. Therefore, we can conclude that ethylene and carbon monoxide are products of the catalytic reactions. The results suggest that carbon monoxide is a result of catalytic decarbonylation at the acid site (whereas carboxylation at the catalyst appears to take place


Page 14 of 40

Page 15 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

to a smaller extent), while the ethylene is possibly a product of the cracking of unsaturated hydrocarbons. The char and coke yields decrease with increasing temperature, from 31.5 C% and 14.0 C% at 400°C, to 22.6 C% and 10.7 C% at 625°C, respectively, possibly due to a reduction in the extent of recombination reactions resulting from an increase in the extent of competing cracking reactions. The yield of known liquids decreases with temperature until it plateaus at about 15 C%, when the temperature reaches 550°C. This aligns with our previous research, which showed that A. donax produces the most volatiles from non-catalytic fast pyrolysis at 400°C [35]. However, because increasing the temperature also increases the activity of the catalyst, the hydrocarbon yield increases with temperature until it plateaus at 550°C at a yield of 11.6 C%, when the increase in catalytic activity producing liquid hydrocarbons begins to be undermined by the increase in the extent of cracking reactions converting the volatiles into permanent gases. Benzene and di-aromatic hydrocarbon (DAH) yields increase with temperature to a maximum of 3.2 C% and 1.1 C% at 625°C. Toluene reaches a maximum at 550°C with a yield of 6.4 C% and C8 aromatics at 475°C with a yield of 2.3 C%. As the temperature increases, it is possible that the methyl groups in C8 species crack to form toluene and methane, and as it continues to increase, the toluene cracks to form benzene and additional methane. The high temperature may also crack polyaromatic coke and char to form more diaromatic hydrocarbons. The lowest temperature tested, 400°C, yielded less aromatic hydrocarbons (4.8 C%) and more phenolics (4.3 C%) and other


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

oxygenated liquid products (5.3 C%) than higher temperatures, suggesting a lower extent of catalytic deoxygenation under these conditions. Carlson et al tested the effects of temperature as well on CFP of pine sawdust and found a maximum of liquid hydrocarbons of 11.0 C% at 600°C (at a WHSV of 0.2 h-1) [11]. This aligns well with our findings, which is interesting given the presence of alkali metals in A. donax. We have previously shown that inorganics in A. donax reduce organic liquid yields by promoting secondary reactions [35]. It appears that the inorganics affect mostly non-catalytic fast pyrolysis, indicating that perhaps the alkali metals primarily act on the oxygenated volatiles during secondary cracking and condensation reactions. Since the catalyst consumes such oxygenated volatiles, preventing those from participating in the secondary reactions, the alkali salts do not seem to have the same effect on catalytic fast pyrolysis. Interestingly, Wang et al showed that alkali salts do have an effect on catalytic fast pyrolysis when the catalyst is in a secondary ex-situ reactor [36]. This may happen because once the alkali salts break down the pyrolysis vapors into gas and solids during the pyrolysis stage (secondary reactions), the catalyst in the secondary unit cannot upgrade those, thereby reducing total yields. For this reason, in order to avoid loss of hydrocarbon yields, in-situ catalytic fast pyrolysis is clearly preferable over ex-situ for biomass that contains high amounts of alkali salts. 3.1.2) Effect of WHSV on Yields of CFP The yields of products from CFP as function of WHSV at 550C and 1 bar in HZSM-5 are given in Figure 3. We varied the WHSV by adjusting the flow rate of biomass


Page 16 of 40

Page 17 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

into the reactor. Overall, the effect of WHSV on the yields in the range 0.5 – 1.9 h-1 is small. This suggests that the range of WHSV used allows for plenty of available catalytic sites for the volatile species produced. The situation changes when the WHSV increases from 1.9 to 3.3 h-1. At these conditions, yields of aromatic hydrocarbons drastically decreased from 11.6 to 6.9 C%. It appears that there are sufficient acid sites for conversion of the biomass until the WHSV exceeds 1.9 h-1. Above this WHSV, the concentration of volatile oxygenates may be too high for the available acid sites, preventing a significant portion of the volatiles from being converted into aromatic hydrocarbons. However, the yield of oxygenated volatiles does not increase, suggesting that these volatiles instead undergo secondary reactions to form other species, such as tar, which would be included in the “unknown liquids” component. Carlson et al found that as WHSV increases in CFP of pine sawdust, MAH yields decrease or stay the same from 0.1 to 1.7 h-1, and DAH yields increase [11]. The decrease in MAHs is consistent with our work, but the increase in DAHs is not. The yields of DAHs from A. donax CFP are lower than Carlson reports for pine at nearly all conditions. Differences in feedstock, temperature, and reactor set-up may explain the discrepancies. 3.1.3) Effect of Pressure on Yields of CFP We performed high pressure CFP (in He gas) of Arundo Donax at 400°C and 35 bar and the results are shown in Table 1, along with the results from atmospheric pressure CFP (in N2 gas) at 400°C for comparison. The residence times for both low and high pressure experiments is about 1 second. For the high pressure run, we recycled 100 % of


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the outlet gas. This means any gaseous products are fed back to the reactor, possibly participating in the reactions. However, the total amount biomass fed during the course of one experiment, and consequently the amount of gaseous products, is likely too small to significantly affect the results. For instance, the high pressure run produces about 16 L of gas, half of which is inert CO2. The volume of helium in the system at 35 bar is approximately 300 liters, indicating that reactive gases constitute a maximum of about 2.5 % of the atmosphere at the end of the experiment, with smaller percentages during the course of the run. The increase in pressure from 1 to 35 bar led to an increase in char yields from 31.5 C% to 38.1 C%. Known liquid yields decreased from 14.3 C% to 8.3 C%, especially the phenolics and other oxygenates, which decreased from 4.1 and 5.5 C% to 1.5 and 2.7 C%, respectively. It is possible that the higher frequency of collisions at higher pressure leads to a higher extent of recombination reactions, especially recombination of oxygenated volatiles, which converts some of the phenolic molecules and other oxygenates into char. This result is consistent with literature on high pressure pyrolysis in general, and seems to be independent of the catalyst. Venkatakrishnan et al reported that, for fast pyrolysis in a continuous reactor without a catalyst, higher pressure leads to an increase in solid yields, a decrease in liquid yields, and constant gas yields. [37]. The addition of hydrogen gas (without a catalyst) to replace the inert helium did not have a significant impact on the effect of pressure. The effects of pressure on the products of catalytic fast pyrolysis are less significant. C8 aromatic yields, other MAHs (C9+) and DAHs increased slightly, while toluene yields decreased from 1.8 C% to 0.8 C% and benzene yields


Page 18 of 40

Page 19 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

decreased slightly. The higher pressure appears to contribute to recombination and alkylation of the aromatics, favoring larger branched aromatic structures over the smaller C6 and C7 aromatics. The total yield of liquid hydrocarbons did not significantly change. Gamliel et al reported that pressure changes from 0 to 31 bar for CFP in a PyGC/MS system led to an increase in solid yields, decrease in gas yields, and constant liquid yields [24],which is consistent with the present work. In their work (Py-GC/MS system), the higher pressure possibly affected the heat transfer through the quartz tube, reducing the extent of cracking reactions of the volatiles and leading to recombination reactions.

3.2) Catalytic fast hydropyrolysis (CHP) 3.2.1) Effect of Hydrogen A comparison of CFP at high pressure with a similar CHP experiment in Figure 4 (both performed at 400°C, 35 bar, 1.1 h-1, and in the presence of HZSM-5) reveals that the hydrogen gas leads to an increase in the production of hydrocarbons over helium, from 4.3 C% to 6.2 C%, especially for C8 aromatics and DAHs, which nearly doubled in yield. Furthermore, some liquid alkanes (C4-C10) were produced, though only 0.3 C%. The yields of C2-C3 olefin gases increased from 0.5 to 3.2 C% in CHP relative to CFP, possibly due to the hydrogen reacting with volatiles that would have otherwise recombined to form coke, while the yield of carbon dioxide decreased from 7.9 to 5.8 C%. It is also possible that a portion of the coke undergoes hydrogasification, producing additional methane. The increase in olefins is possibly a consequence of the overall increase in hydrocarbons, as light olefins could originate from hydrocarbon secondary cracking. As


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

expected, the coke yield decreased (from 13.7 C% to 9.7 C%), and the yield of phenolics increased from 1.5 C% to 2.6 C%. The hydrogen gas may be decreasing the extent of the coking reactions, consequently leaving more volatiles available to react on the catalyst, increasing the yield of aromatic hydrocarbons, phenolics (which could be an intermediate catalytic deoxygenation product), and olefinic gases. Similar to the CFP case, the deoxygenation of the pyrolysis vapors does not appear to take place primarily via decarboxylation, but instead via dehydration, which produces additional water. This leaves more carbon atoms available for production of hydrocarbons. Another possible effect of the excess hydrogen in this system is to shift the equilibrium of the water-gas shift reaction in the reverse direction, consuming carbon dioxide and producing carbon monoxide and water. The decreased rate of coking could be a result of the presence of hydrogen, which interact with reactive volatiles, reducing the extent of coking reactions. Because of this, it is likely that the catalyst is not deactivated as quickly during the experiment and therefore it yields more zeolite adsorption products, such as BTX and ethylene, as shown in Figure 4. However, hydrogenation appears to be mostly limited to carbon-oxygen bonds, as the aromatic and olefin hydrocarbons saturation to become alkanes only takes place to a small extent (the total alkanes yield is 0.6 C%). Alkanes are an important component of hydrocarbon fuels such as gasoline and jet-fuel. We will discuss the possibility of alkane formation in the next section (temperature effects). Based on the results in Figure 4, we hypothesize that hydrogen at these conditions form olefinic gas and water. These reactions compete with the formation of coke.


Page 20 of 40

Page 21 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Overall, the experimental yields obtained from CFP at these conditions correspond to roughly 6.0 gallons of hydrocarbons per ton of Arundo Donax. For CHP, the yield of hydrocarbons increase to 8.9 gallons/ton. We note that this comparison was performed at 400°C and 35 bar for the two processes. CFP at 550°C and 1 bar led to 16.8 gallons/ton. We are confident that there is still significant room for improvement in hydrocarbon yields pending on the development of new catalysts for the CHP process. 3.2.2) Effect of Temperature on Yields of CHP Previous work with an equilibrium calculator at Colorado State University estimates that, when the temperature of pure hydrogen gas (without a catalyst) increases from 350°C to 400°C [38] the molar concentration of hydrogen radicals increases from 4.02 x 10-17 to 9.44 x 10-16 – a 24-fold increase. In the presence of a proper catalysts, we assumed that the increase in hydrogen radical presence correlates strongly with the reactivity of the hydrogen gas. The result of this increase can be highlighted by comparing the yields of CHP with HZSM-5 at 350°C and 400°C in Figure 5. The gas yield increases from 11.6 C% to 18.9 C%. Most notably, we see increases in carbon monoxide (5.1 C% to 8.6 C%), methane (0.2 C% to 0.9 C%), and olefins (0.8 C% to 3.2 C%), which are the gaseous products of catalytic deoxygenation. We also see an increase in total hydrocarbons from 3.8 C% and 6.2 C%, driven by increases in all aromatic hydrocarbons. Catalytic activity increases with temperature, as shown by the decrease in other oxygenates from 4.4 C% to 3.1 C% and increase in intermediate deoxygenation products such as phenolics from 1.5 C% to 2.6 C%. This suggests that hydrogen may assist in deoxygenation reactions, producing more hydrocarbons and less oxygenates, especially at higher radical 21 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 40

concentrations. Other reactions that appear to have significantly higher rates at 400C are decarbonylation and cracking of aromatics, evidenced by the significant increase in the yield of gaseous olefins. The small yield of liquid alkanes (0.2-0.3 C%) at both temperatures suggest that hydrogenation of aromatics only takes place to a small extent. This is in agreement with findings from Jan et al, [16] which showed that the equilibrium yields of hydrogenation products are very small at temperatures above 300C (hydrogenation is an exothermic process). Therefore, it appears that the large concentration of hydrogen radicals at 400C mostly affects deoxygenation reactions such as carbonylation and dehydration. Hydrogenation of aromatics only takes place to a small extent because of thermodynamic limitations.

3.2.3) Analysis of Catalyst Post-run

% 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑎𝑟𝑒𝑎 𝑙𝑜𝑠𝑠 𝑆𝑢𝑟𝑓𝑎𝑐𝑒 𝑎𝑟𝑒𝑎 𝑙𝑜𝑠𝑠 𝑟𝑎𝑡𝑖𝑜 = = 𝑏𝑖𝑜𝑚𝑎𝑠𝑠 𝑡𝑜 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡 𝑟𝑎𝑡𝑖𝑜

𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑎𝑟𝑒𝑎 𝑎𝑓𝑡𝑒𝑟 𝑒𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡 ) 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑎𝑟𝑒𝑎 𝑏𝑒𝑓𝑜𝑟𝑒 𝑒𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡 𝑡𝑜𝑡𝑎𝑙 𝑏𝑖𝑜𝑚𝑎𝑠𝑠 𝑓𝑒𝑑 𝑖𝑛 𝑒𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡 𝑡𝑜𝑡𝑎𝑙 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡

1―(

Equation 4: Definition of surface area loss ratio in units of % loss per gram biomass per gram catalyst

We performed BET analysis of the catalyst before and after some of the experiments. The results are shown in Table 2. The processes in both atmospheres (helium and hydrogen) reduce the micropore volume and surface area. However, the extent of reduction is larger in helium compared to hydrogen. CFP performed with helium using 200 g HZSM-5 and 118.2 g biomass resulted in a micropore volume reduction of 17.6 μL/g and


Page 23 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

a surface area reduction of 72 m2/g. CHP using 200 g HZSM-5 and 123.5 g biomass resulted in a micropore volume reduction of 11.2 μL/g and a surface area reduction of 52 m2/g. The coke formed/catalyst ratio was larger for CFP (0.040) than for CHP (0.035). We defined the surface area loss ratio in Equation 4 as the percent of surface area lost when one gram of biomass is pyrolyzed through one gram of catalyst. While high pressure CFP with helium gas led to a surface area loss ratio of 45.7%, CHP only led to a surface area loss ratio of 31.6%. Therefore, CHP, which produces less coke than CFP (5.63 C% as opposed to 6.77 C% in this case), is expected to allow for a longer catalyst life due to decreased deactivation rates of the catalyst by coking.

3.3) Overall Reaction Pathways The information generated in the present work allows us to expand the current knowledge on the lumped pathways for the hydropyrolysis process. Fast pyrolysis, catalytic fast pyrolysis, and catalytic fast hydropyrolysis all involve hundreds of chemical reactions. As a consequence, the goal of the present work is not to investigate the detailed mechanisms involved in the all of these reactions. Such studies would require the use of model compounds, which is beyond the scope of our work. However, for systems that use real biomass feedstocks, it is possible to describe overall, lumped reaction pathways which involve species such as char and tar, describing the process in general terms, and possibly a few reactions in the gas phase (such as hydrogasification and water-gas shift). One of the goals of our paper is to identify the differences in overall pathways between CFP and CHP. Figure 6 shows us the proposed pyrolysis pathway, which combines previous knowledge


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

from the literature with new findings from the present work. In proposing this pathway, we are assuming that the reactivity of the hydrogen gas in the presence of HZSM-5 is largely driven by the concentration of hydrogen radicals within the gas. Originally, the lignocellulosic polymers (cellulose, hemicellulose, and lignin) break down in the presence of heat into oxygenated volatiles. A portion of the volatiles cracks to form permanent gas and water, a path that is favored at high temperatures and long residence times. In contrast, low temperatures and slow heating rates promote the recombination of volatiles into char [9]. Increased pressure simply increases the yield of char. Some tar (high molecular weight volatiles) also result, especially at long vapor residence times. In the presence of a zeolite catalyst, such as HZSM-5, a portion of the volatiles adsorb and react via deoxygenation reactions (mostly decarbonylation and dehydration, with decarboxylation taking place to a smaller extent), generating hydrocarbons, mainly aromatics, such as benzene, toluene, and xylene (abbreviated BTX), and releasing CO, CO2 and water as by-products. With increased temperature, the absorbed organics produce more gaseous olefins such as ethylene, and branched aromatics such as toluene crack to form methane and benzene. However, if the hydrocarbons fail to desorb, they may oligomerize into coke and remain in the pores of the catalyst, blocking the active sites [11]. When hydrogen is introduced, it may interact with adsorbed organic species at the acid sites, producing more gaseous olefins and desorbed aromatic hydrocarbons at the expense of coke. The presence of hydrogen can also promote hydrogasification of the coke 24 ACS Paragon Plus Environment

Page 24 of 40

Page 25 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

and the reverse-water gas shift reaction [14]. These additional routes are the main difference between CFP and CHP in HZSM-5. Using results from our previous paper on fast pyrolysis of A. donax (washed and unwashed) without a catalyst [2], we are in a position to compare the four processes (fast pyrolysis, fast pyrolysis of washed biomass, catalytic fast pyrolysis, and catalytic fast hydropyrolysis), determining which situation results in the highest carbon recovery in the liquid, and the highest amount of drop-in fuels. The results are given in Table 3. Note that in Table 3 is comparing the highest liquid yields from each process, which are not necessarily obtained at identical conditions. We assumed the carbon content of the liquid products from fast pyrolysis to be the same as the carbon content of the biomass, given that there is not a catalyst imposing deoxygenation in this process. The data show that the highest yield of carbon is recovered in non-catalytic fast pyrolysis, probably because the catalytic processes promote decarbonylation and decarbonylation reactions, causing loss of carbon from the liquid to the gas products. However, the quality of this fuel is low, due to the high oxygen content and lack of hydrocarbons. CFP produces less liquid carbon than fast pyrolysis, but a significant percentage of that liquid is BTX, which is a separate phase from the bio-oil produced and can be used as a high-value additive for transportation fuels. As we have previously discussed, CHP produces a higher liquid yield than CFP when the two processes are compared at the same conditions. However, among all the conditions studied, the highest liquid yield for CFP (11.6 wt %) was higher than the highest yield for CHP (6.2 wt %). 25 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

This is related to the pressure requirements and temperature limitations in the operation of the fluidized bed reactor. It is likely, however, that the use of higher temperature and catalysts designed for hydrogenation reactions (such as metal-doped zeolites) may lead to significant increases in yields. Evidence for this hypothesis will be provided in our future work.

Conclusion We performed catalytic fast pyrolysis of A. Donax with HZSM-5 at varying temperatures, weight hourly space velocities, and pressures with and without hydrogen gas. The highest yield of liquid hydrocarbons for CFP was 11.6 C%. CHP produces more liquid hydrocarbons than CFP at identical conditions. Overall, CHP with HZSM-5 at 400C and 35 bar can produce 8.9 gal/ton of liquid hydrocarbons, as opposed to 6.0 gal/ton from CFP. The introduction of hydrogen also led to a 17 % reduction of coke produced relative to the coke produced under a helium atmosphere.

Acknowledgements This work was funded by Tree Free Biomass Solutions and the Weyerhaueser Endowmentin Paper Science and Engineering.

References [1]

F.A. Agblevor, S. Besler, Inorganic compounds in biomass feedstocks. 1. Effect on


Page 26 of 40

Page 27 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

the quality of fast pyrolysis oils, Energy & Fuels. 10 (1996) 293–298. doi:10.1021/ef950202u. [2]

D.. S. Chandler, F.L.P. Resende, Effects of warm water washing on the fast pyrolysis of Arundo Donax, Biomass and Bioenergy. 113 (2018) 65–74. doi:10.1016/j.biombioe.2018.03.008.

[3]

G. Luo, D.S. Chandler, L.C.A. Anjos, R.J. Eng, P. Jia, F.L.P. Resende, Pyrolysis of whole wood chips and rods in a novel ablative reactor, Fuel. 194 (2017). doi:10.1016/j.fuel.2017.01.010.

[4]

H. Hwang, S. Oh, T.S. Cho, I.G. Choi, J.W. Choi, Fast pyrolysis of potassium impregnated poplar wood and characterization of its influence on the formation as well as properties of pyrolytic products, Bioresour. Technol. 150 (2013) 359–366. doi:10.1016/j.biortech.2013.09.132.

[5]

D.J. Nowakowski, A. V. Bridgwater, D.C. Elliott, D. Meier, P. de Wild, Lignin fast pyrolysis: Results from an international collaboration, J. Anal. Appl. Pyrolysis. 88 (2010) 53–72. doi:10.1016/j.jaap.2010.02.009.

[6]

M.B. Pecha, J.I. Montoya, F. Chejne, M. Garcia-Perez, Effect of a Vacuum on the Fast Pyrolysis of Cellulose: Nature of Secondary Reactions in a Liquid Intermediate, Ind. Eng. Chem. Res. 56 (2017) 4288–4301. doi:10.1021/acs.iecr.7b00476.

[7]

P.R. Patwardhan, R.C. Brown, B.H. Shanks, Understanding the product distribution from biomass fast pyrolysis, Iowa State University, 2009. doi:10.1016/j.jaap.2009.08.007.

[8]

C. Dou, D.S. Chandler, F.L.P. Resende, R. Bura, Fast Pyrolysis of Short Rotation Coppice Poplar: An Investigation in Thermochemical Conversion of a Realistic Feedstock for the Biorefinery, ACS Sustain. Chem. Eng. (2017). doi:10.1021/acssuschemeng.7b01000.

[9]

A. V Bridgwater, Fast Pyrolysis of Biomass for Energy and Fuels, (n.d.).

[10]

R. Mahadevan, R. Shakya, S. Neupane, S. Adhikari, Physical and chemical properties and accelerated aging test of bio-oil produced from in situ catalytic pyrolysis in a bench-scale fluidized-bed reactor, Energy and Fuels. 29 (2015) 841–


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

848. doi:10.1021/ef502353m. [11]

T.R. Carlson, Y.-T. Cheng, J. Jae, G.W. Huber, Production of green aromatics and olefins by catalytic fast pyrolysis of wood sawdust, Energy Environ. Sci. 4 (2011) 145–161. doi:10.1039/C0EE00341G.

[12]

C. Williams, T. Biswas, A. Downie, Yield , salt tolerance and energy from Arundo donax : a potential biochar and biofuel crop, in: 1st Asia Pacific Biochar Conf., Queensland, AU, 2009.

[13]

J. Giessow, Jason; Casanova, Jason; Leclerc, Rene; MacArthur, Robert;Fleming, Genie;Giessow, Arundo Distribution and Impact Report, (2011).

[14]

S. Meesuk, J.P. Cao, K. Sato, Y. Ogawa, T. Takarada, Fast pyrolysis of rice husk in a fluidized bed: Effects of the gas atmosphere and catalyst on Bio-oil with a relatively low content of oxygen, Energy and Fuels. 25 (2011) 4113–4121. doi:10.1021/ef200867q.

[15]

S. Thangalazhy-Gopakumar, S. Adhikari, R.B. Gupta, M. Tu, S. Taylor, Production of hydrocarbon fuels from biomass using catalytic pyrolysis under helium and hydrogen environments, Bioresour. Technol. 102 (2011) 6742–6749. doi:10.1016/j.biortech.2011.03.104.

[16]

O. Jan, R. Marchand, L.C.A. Anjos, G.V.S. Seufitelli, E. Nikolla, F.L.P. Resende, Hydropyrolysis of lignin using Pd/HZSM-5, Energy and Fuels. 29 (2015) 1793– 1800. doi:10.1021/ef502779s.

[17]

M. Olazar, R. Aguado, J. Bilbao, A. Barona, Pyrolysis of sawdust in a conical spouted-bed reactor with a HZSM-5 catalyst, AIChE J. 46 (2000) 1025–1033. doi:10.1002/aic.690460514.

[18]

K. Iisa, R.J. French, K.A. Orton, M.M. Yung, D.K. Johnson, J. Dam, M.J. Watson, M.R. Nimlos, In Situ and ex Situ Catalytic Pyrolysis of Pine in a Bench-Scale Fluidized Bed Reactor System, (2016). doi:10.1021/acs.energyfuels.5b02165.

[19]

K. Iisa, R.J. French, K.A. Orton, A. Dutta, J.A. Schaidle, Production of low-oxygen bio-oil via ex situ catalytic fast pyrolysis and hydrotreating, Fuel. 207 (2017) 413– 422. doi:10.1016/j.fuel.2017.06.098.

[20]

A.A. Lappas, K.G. Kalogiannis, E.F. Iliopoulou, K.S. Triantafyllidis, S.D.


Page 28 of 40

Page 29 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Stefanidis, Catalytic Pyrolysis of Biomass for Transportation Fuels, (2016) 45–56. doi:10.1002/wene.16. [21]

F.L.P. Resende, Recent advances on fast hydropyrolysis of biomass, Catal. Today. 269 (2016) 148–155. doi:10.1016/j.cattod.2016.01.004.

[22]

F. Melligan, M.H.B. Hayes, W. Kwapinski, J.J. Leahy, Hydro-pyrolysis of biomass and online catalytic vapor upgrading with Ni-ZSM-5 and Ni-MCM-41, Energy and Fuels. 26 (2012) 6080–6090. doi:10.1021/ef301244h.

[23]

D.P. Gamliel, L. Wilcox, J.A. Valla, The effects of catalyst properties on the conversion of biomass via catalytic fast hydropyrolysis, Energy and Fuels. 31 (2017) 679–687. doi:10.1021/acs.energyfuels.6b02781.

[24]

D.P. Gamliel, G.M. Bollas, J.A. Valla, Bifunctional Ni-ZSM-5 Catalysts for the Pyrolysis and Hydropyrolysis of Biomass, Energy Technol. 5 (2017) 172–182. doi:10.1002/ente.201600136.

[25]

D.P. Gamliel, G.M. Bollas, J.A. Valla, Two-stage catalytic fast hydropyrolysis of biomass for the production of drop-in biofuel, Fuel. 216 (2018) 160–170. doi:10.1016/j.fuel.2017.12.017.

[26]

S. Thangalazhy-Gopakumar, S. Adhikari, R.B. Gupta, Catalytic pyrolysis of biomass over H+ZSM-5 under hydrogen pressure, Energy and Fuels. 26 (2012) 5300–5306. doi:10.1021/ef3008213.

[27]

O.D. Jan, R. Gustafson, No Title, (2014).

[28]

X. Zhu, L.L. Lobban, R.G. Mallinson, D.E. Resasco, Bifunctional transalkylation and hydrodeoxygenation of anisole over a Pt / HBeta catalyst, 281 (2011) 21–29. doi:10.1016/j.jcat.2011.03.030.

[29]

D.C. Dayton, J. Hlebak, J.R. Carpenter, K. Wang, O.D. Mante, J.E. Peters, Biomass Hydropyrolysis in a Fluidized Bed Reactor, Energy and Fuels. 30 (2016) 4879– 4887. doi:10.1021/acs.energyfuels.6b00373.

[30]

T.L. Marker, L.G. Felix, M.B. Linck, M.J. Roberts, Integrated hydropyrolysis and hydroconversion (IH2) for the direct production of gasoline and diesel fuels or blending components from biomass, part 1: Proof of principle testing, Environ. Prog. Sustain. Energy. 31 (2012) 191–199. doi:10.1002/ep.10629.


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[31]

E.C.D. Tan, T.L. Marker, M.J. Roberts, Direct Production of Gasoline and Diesel Fuels from Biomass via Integrated Hydropyrolysis and Hydroconversion Process— A Techno-economic Analysis, Environ. Prog. Sustain. Energy. 33 (2014) 609–617. doi:10.1002/ep. 11791.

[32]

T.L. Marker, L.G. Felix, M.B. Linck, M.J. Roberts, P. Ortiz-Toral, J. Wangerow, Integrated Hydropyrolysis and Hydroconversion (IH^2) for the Direct Production of Gasoline and Diesel Fuels or Blending Components from Biomass, Part 2: Continuous Testing, Environ. Prog. Sustain. Energy. 33 (2014) 762–768. doi:10.1002/ep.

[33]

R. Cocco, S.B.R. Karri, T. Knowlton, Introduction to fluidization, Chem. Eng. Prog. 110 (2014) 21–29.

[34]

J. Rincon, J. Guardiola, A. Romero, G. Ramos, Predicting the Minimum Fluidization Velocity of Multicomponent Systems, J. Chem. Eng. JAPAN. 27 (1994) 177–181.

[35]

D.S. Chandler, F.L.P. Resende, Effects of warm water washing on the fast pyrolysis of Arundo Donax, Biomass and Bioenergy. 113 (2018) 65–74. doi:10.1016/j.biombioe.2018.03.008.

[36]

K. Wang, J. Zhang, B.H. Shanks, R.C. Brown, The deleterious effect of inorganic salts on hydrocarbon yields from catalytic pyrolysis of lignocellulosic biomass and its mitigation, Appl. Energy. 148 (2015) 115–120. doi:10.1016/j.apenergy.2015.03.034.

[37]

V.K. Venkatakrishnan, J.C. Degenstein, A.D. Smeltz, W.N. Delgass, R. Agrawal, F.H. Ribeiro, High-pressure fast-pyrolysis, fast-hydropyrolysis and catalytic hydrodeoxygenation of cellulose: production of liquid fuel from biomass, Green Chem. 16 (2014) 792. doi:10.1039/c3gc41558a.

[38]

D. Dandy, Chemical Equilibrium Calculation, Color. State Univ. Bioanal. Microfluid. Progr. (2018). http://navier.engr.colostate.edu/code/code-4/index.html.


Page 30 of 40

Page 31 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure Captions Figure 1: Reactor schematic. 1) Biomass hopper. 2) Auger motor. 3) Feed funnel. 4) Feed line. 5) Pre-heater (high pressure only). 6) Fluidizing gas. 7) Fluidized bed reactor with catalyst and sand. 8) Transfer line. 9) Cyclone separator. 10) Impinger (cooled by water bath in high-pressure experiments). 11) Double-pipe condenser. 12) Glass bead-filled double-pipe condensers (x2). 13) Methanol bubblers (x3) (atmospheric pressure) or glass-bead-filled impingers (x3) (high pressure). 14) Coalescing filter. 15) Exhaust (atmospheric pressure) 16) gas booster (high pressure). TC: thermocouple. TD: transducer Figure 2: Yields of CFP as a function of temperature. Top left: total yields. Top right: Gas yields. Bottom: major liquid compounds. MAH: mono-aromatic hydrocarbon. DAH: di-aromatic hydrocarbon. The carbon balances are between 98.6-108.1 C%. All experiments performed at atmospheric pressure, a WHSV of 1.1 h-1, and in inert gas. Figure 3: Yields of CFP as a function of the weight hourly space velocity (in h-1). Top left: total yields. Top right: gas yields. Bottom: major liquid compound yields. The carbon balance is 94.3-102.4 C%. All experiments performed at atmospheric pressure, 550°C, and in inert gas. Table 1: Yields of products from CFP at two different pressures. Both experiments performed at 400°C, 1.1 h-1 WHSV, and in inert gas (N2 at low pressure, He at high pressure). Figure 4: Comparison of yields from high pressure CFP and CHPat 400°C, 35 bar, 1.1 h-1, in the presence of HZSM-5. Top left: total yields. Top right: yields of gaseous products. Bottom: yields of liquid products. Standard deviations for alkanes are not given because they were not produced in the base case. The carbon balances for these experiments are 94.5 and 98.0 C% for helium and hydrogen, respectively. Figure 5: Effects of temperature (in °C) on yields of CHPat 1.0-1.2 h-1 and 35 bar with HZSM-5. Top left: total yields. Top right: gaseous yields. Bottom: major liquid compounds. The carbon balances were 99.1 and 98.0 C% for 350°C and 400°C, respectively. Standard deviations for alkanes are not given because they were not produced in the base case. Table 2: Results of BET analysis of catalyst before and after high pressure catalytic fast pyrolysis and high pressure catalytic hydropyrolysis. Surface are loss ratio is defined in Equation 4. Conditions: 400°C, 35 bar, 1.1 h-1. Figure 6: Lumped mechanisms of CFP and CHP. The present work demonstrates the effect of pressure (P) and hydrogen radicals (H) whereas the effects of heat (Q), and zeolite acid sites (+) are reported in previous works. PAHs: polyaromatic hydrocarbons. HCs: hydrocarbons Table 3: Comparison of highest liquid yields and characteristics from fast pyrolysis, catalytic fast pyrolysis, and catalytic hydropyrolysis


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figures and Tables


Page 32 of 40

Page 33 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Page 34 of 40

Page 35 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1 bar Major Yields (C%) Char Coke Gas Known Liquids Unknown Liquids Water (wt%) Gas Yields (C%) CO CO2

35 bar

31.5 14.0 17.8 14.3 21.0 23.1

38.1 13.7 17.1 8.5 17.1 24.7

8.9 7.3

8.2 7.9

CH4

0.7

0.5

C2-C3 gases Liquid Yields (C%) Total HCs Benzene Toluene C8 aromatics Other MAHs DAHs Phenolics Other Carbon Balance

1.0

0.6

4.8 0.4 1.8 1.3 0.4 0.8 4.1 5.5 98.6

4.3 0.2 0.8 1.5 0.6 1.0 1.5 2.7 94.5


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Page 36 of 40

Page 37 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 40

Reaction gas

He

H2

Catalyst

HZSM-5

HZSM-5

Biomass fed (g)

118.2

123.5

Coke mass yield (wt.%)

6.77 wt.%

5.63 wt.%

Coke / g catalyst (g)

0.040

0.035

Micropore volume before (μL/g)

69.7 ± 2.4

69.7 ± 2.4

Micropore volume after (μL/g)

52.1

58.5

Pre-run surface area (m2/g)

267 ± 13.2

267 ± 13.2

Post-run Surface area (m2/g)

195

215

Surface area loss ratio

45.7%

31.6%


Page 39 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Total liquid

Page 40 of 40

Non catalytic,

Non catalytic,

Catalytic fast

Catalytic

unwashed,

washed,

pyrolysis,

hydropyrolysis,

400°C, 1 bar

475°C, 1 bar

550°C, 1 bar

400°C, 35 bar

40.5

50.2

21.1

23.8

0

0

11.6

6.2

40.5

50.2

36.8

30.0

Bio-oil

Bio-oil

Two phases:

Two phases:

BTX & bio-oil

BTX & bio oil

yield (w/o water) (wt%) Liquid hydrocarbon yield (C%) Carbon recovery in liquid (C%) Fuel type


Comparison between Catalytic Fast Pyrolysis and Catalytic Fast

Recommend Documents