IEA Bioenergy Task 34: Direct Thermochemical Lique Faction

PyNe 41, IEA Bioenergy Task 34Page 4 of 24Comparison of pyrolysis in screw-type reactors Continued continuedcondensation strategies to allow fora feasible data basis.Table 2: Process conditions of the experimentsScrew-type reactors have beenquestioned to be representative forfast pyrolysis (Elliott et al., 2017).This hypothesis can be investigatedwith the present study because fastpyrolysis experiments have beenconducted as part of the IEABioenergy Task 34 Round Robinand also because beech wood hasbeen used additionally. The mostmeaningful basis to compare bio-oilyields is on the basis of the organicliquid yield, which excludes thewater content of liquid product(s)(see Figure 1). The organic liquidyield obtained in this study frombeech wood fast pyrolysis iscomparable to the range of 50-55 %reported elsewhere for beech wood(Beaumont & Schwob, 1984;Greenhalf et al., 2013).Unfortunately, results in literatureare often only reported for the bio-oilyield, i.e. including water. Such datafor beech wood fast pyrolysis reportsbio-oil yields of around 61-70 %(Beaumont & Schwob, 1984; DiBlasi & Branca, 2001; Jendoubi etal., 2011; Wang et al., 2005). TheseDur.PythonSTYXFeed rate4-5 h 5-6 kg h-13-4 h 3 kg h-1Rxtr.temp.500 C500 CHeating rate 10.000 K min-1 100 K min-1results are in line with bio-oil yieldsobtained from beech wood fastpyrolysis in the present study, too(69.9 0.9 %).Vapourres. time 4 s 20 sCondens.temp(s)90 C; 20 C80 C; 15 Cfor subsequent gasification (Dahmenet al., 2012; Nicoleit et al., 2016).Finally, the quality of the producedfast pyrolysis bio-oil from thefeedstocks supplied by the IEABioenergy Task 34 Round Robindoes not show a significant deviationfrom bio-oils produced by fluidizedbed or ablative reactors (Elliott et al.,2017). Values can be easily checkedbecause there has been only oneparticipating laboratory that applieda twin-screw technology. The onlyexception to the quality parametersis the high solids content. Thecomparably poor solid separationprior to condensation is due to thecurrent project requirements inwhich the condensate(s) are mixedwith the solids to produce a bioslurryObviously, the reactor type used toachieve fast pyrolysis conditions isof minor importance as long asheating rate and hot vapourresidence time requirements aremet. It is concluded that the screwtype reactor used in this study issuitable to achieve all necessary fastpyrolysis conditions and inconsequence to produce a ‘typical’fast pyrolysis bio-oil. All data indicatethat it is a suitable representative forfast pyrolysis to allow for acomparison with intermediatepyrolysis.The organic liquid yields obtainedfrom fast pyrolysis are significantlyhigher than from intermediatepyrolysis, as expected. It is(Continued on page 5)Yield, dry feedstock mass basis (%)6051.250494035.63021.125.623.120Organic liquid (FP)18.4Organic liquid (IP)Water of reaction (FP)16.7Water of reaction (IP)100Beech wood Poplar woodBlendWheat strawFigure 1: Organic liquid and water of reaction yields (FP: fast pyrolysis, IP: intermediatepyrolysis). Error bars indicate the difference between two experimental runs.

Page 5 of 24PyNe 41, IEA Bioenergy Task 34Comparison of pyrolysis in screw-type reactors Continued continuedinteresting to observe that the yieldsfrom fast pyrolysis are more heavilyaffected by the ash content of thefeedstock. While more than twice asmuch organic liquid is produced incase of beech wood only 50 % moreis produced for the ash-rich wheatstraw. Consequently, the advantageof high organic liquid yields has tobe evaluated with specialconsideration of the type offeedstock used. The high amount ofreaction water produced duringintermediate pyrolysis does notaffect bio-oil quality negatively.Water content in the intermediatepyrolysis bio-oil of around 10 % hasbeen observed with the currentcondensation and product recoverydesign.The yield of several marker specieshas been evaluated for the obtainedbio-oils. The results support theexpected ongoing pyrolytic reactionsin case of intermediate pyrolysis.Acetic acid yields are increased andso is the yield of phenol on theexpense of guaiacol. It can even beobserved that benzene andnaphthalene are produced withyields one order of magnitude higherthan for fast pyrolysis.Surely, this study is a first step toallow for a more detailed comparisonof these two pyrolysis technologies.It provides the required quantitativedata to show differences in productdistribution. The next step is todirectly compare bio-oil quality andrelate it to the differences in bio-oilyield. Such data is a mandatoryprerequisite to allow for a rationaldecision for suitable pyrolysistechnology dependent on feedstockand product use.ReferencesBeaumont, O., Schwob, Y. 1984.Influence of Physical and ChemicalParameters on Wood Pyrolysis. Ind EngChem Process Des Dev, 23, 637-641.Dahmen, N., Henrich, E., Dinjus, E.,Weirich, F. 2012. The bioliq bioslurrygasification process for the production ofbiosynfuels, organic chemicals, andenergy. Energ Sust Soc, 2(3).Di Blasi, C., Branca, C. 2001. Kinetics ofPrimary Product Formation from WoodPyrolysis. Ind Eng Chem Res, 40(23),5547-5556.Elliott, D.C., Meier, D., Oasmaa, A., vande Beld, B., Bridgwater, A.V., Marklund,M. 2017. Results of the InternationalEnergy Agency Round Robin on FastPyrolysis Bio-oil Production. Energ Fuel,31(5), 5111-5119.Funke, A., Richter, D., Niebel, A.,Dahmen, N., Sauer, J. 2016. FastPyrolysis of Biomass Residues in aTwin-screw Mixing Reactor. J Vis Exp,115, e54395.Funke, A., Tomasi Morgano, M., Leibold,H., Dahmen, N. 2017. ExperimentalComparison of Fast and IntermediatePyrolysis. J Anal Appl Pyr, 124, 504-514.Greenhalf, C.E., Nowakowski, D.J.,Harms, A.B., Titiloye, J.O., Bridgwater,A.V. 2013. A comparative study of straw,perennial grasses and hardwoods interms of fast pyrolysis products. Fuel,108, 216-230.Jendoubi, N., Broust, F., Commandre,J.M., Mauviel, G., Sardin, M., Lede, J.2011. Inorganics distribution in bio oilsand char produced by biomass fastpyrolysis: The key role of aerosols. JAnal Appl Pyr, 92(1), 59-67.Neumann, J., Binder, S., Apfelbacher,A., Gasson, J.R., Ramírez García, P.,Hornung, A. 2015. Production andcharacterization of a new qualitypyrolysis oil, char and syngas fromdigestate - Introducing the thermocatalytic reforming process. Journal ofAnalytical and Applied Pyrolysis, 113,137-142.Nicoleit, T., Dahmen, N., Sauer, J. 2016.Production and Storage of GasifiableSlurries Based on Flash-PyrolyzedStraw. Energy Technology, 4(1), 221229.Tomasi Morgano, M., Leibold, H.,Richter, F., Seifert, H. 2015. Screwpyrolysis with integrated sequential hotgas filtration. Journal of Analytical andApplied Pyrolysis, 113, 216-224.Wang, X., Kersten, S., Prins, W., vanSwaaij, W.P.M. 2005. Biomass Pyrolysisin a Fluidized Bed Reactor. Part 2:Experimental Validation of ModelResults. Ind Eng Chem Res, 44, 87868795.Contact:Dr.-Ing. Axel FunkeFast Pyrolysis GroupHermann-von-Helmholtz-Platz 1Building 72776344 Eggenstein-Leopoldshafen,GermanyPhone: 49 721 608-22391Fax: 49 721 608-22244Email: axel.funke@kit.eduWeb: http://www.kit.edu/

PyNe 41, IEA Bioenergy Task 34Page 6 of 24Increasing production and isolation of phenols via pyrolysis oflignocellulosic biomassCharles MullenUSDA-ARSMuch of the research in biomasspyrolysis over the last decade hasfocused on modifications to thepyrolysis process to produce stable,partially deoxygenated bio-oils. Thishas been done, usually, with theultimate goal of reducing the oxygencontent, after upgrading, to near zerofor use in advanced hydrocarbon biofuels. However, there is now arealization that to be commerciallyviable, biorefineries will need toproduce high value chemical ormaterials co-products in addition tofuels, just as petroleum refineries do.This is where preservation of some ofthe inherent oxygen content, that wehave long sought to remove, may be anadvantage. Nearly limitless usefuloxygenated chemicals, from highvolume commodities to premiumpharmaceutical precursors areproduced through oxidation ofpetroleum hydrocarbons, sooxygenated biomass derivedreplacements may be advantageous.USDA-ARS has begun researchingways to increase the yield of certaintargeted oxygenated hydrocarbonsduring pyrolysis and also isolate themfrom the complex mixture. Oneexamples is phenolics. The phenolicaromatic C-O bond is among the mostdifficult of all the C-O bonds found inbio-oils to break, and phenols are foundFigure 1. Comparison of yields from selected compounds from switchgrass (SWG)and Cellulose (Cell) over HZSM-5 (23) and KZSM-5 (23) at 5/1 catalyst/biomass and500 ºC. Alkyl phenols phenol, o-,m-,p-cresols, 2,4-dimethyl phenol and 4-ethylphenol. Methoxy phenols guaiacol, 4-methylguaiacol, syringol. BTEX benzene,toluene, ethyl benzene, xylenes. Error bars represent one standard deviation.among the remaining oxygenates inboth the CFP and TGRP processes.Furthermore, phenol is more valuablethan its completely deoxygenatedanalog benzene. For these reasonsphenols are among the first oxygenatedchemicals we are targeting forincreased production during pyrolysis.Two types of phenols are producedfrom pyrolysis of lignocellulosicbiomass, alkyl phenols which can bederived from either the cellulose orlignin polymer, and methoxylatedphenols which are generated directlyfrom lignin depolymerization. Recentdevelopments in our laboratory haveincreased the production of bothclasses of phenols from pyrolysis ofbiomass. In many studies of HZSM-5zeolite catalyst deactivation, alkylphenols are often found as an“intermediate” product, not produced byfresh catalysts, but by partiallydeactivated catalysts at levels wellabove non-catalytic pyrolysis. For thisreason, we sought to understand theproperties of these partially deactivatedzeolites, and mimic them to maximizeproduction of alkyl phenols. Catalystdeactivation can occur by deposition ofcarbon deposits on catalysts or bypoisoning by alkali metals; both ofthese have the effect of decreasing theBrønsted acid site density. Wetherefore produced a simple series ofcatalysts, reducing the acid site densityby exchanging the acid sites withpotassium (K).As shown in Figure 1, when the KZSM5 catalyst was used for the CFP ofswitchgrass, there was a 74% decreasein the production of aromatichydrocarbons compounds and 3-foldincrease in the production of theselected alkyl- and methoxy phenolscompared to the parent HZSM-5. Inthe case of methoxy phenols, the yieldwas lower than that achieved noncatalytically, which suggests they werenot formed catalytically but rather their(Continued on page 5)

Page 7 of 24PyNe 41, IEA Bioenergy Task 34Increasing production and isolation of phenols continuedcatalytic conversion to other productswas limited by the use of a less activecatalyst. However, the alkyl phenolyield was substantially higher ( 4-foldincrease) than that produced noncatalytically, meaning the alkyl phenolswere the product of a catalytic reaction,not merely a consequence ofdiminished catalytic conversion.Similarly, there was an increase in theproduction of furans, particularly 2-,whose production increased to a yieldof 4.6 mg/g, whereas only tracedecrease in the average molecularweight of the bio-oil. Lignin is also themost abundant natural source ofaromatic compounds in the biosphere,which makes it an attractive renewablefeedstock for the production of biofuels,commodity chemicals, and other valueadded products. Unfortunately, limitedsuccess has been achieved indevelopment of effectivedepolymerization strategies to unlockthis potential.Table 1. Effect of adding 1,4-butandiol to lignin during microwavepyrolysis on bio-oil yield, molecular weight, dispersity index and phenolicselectivity.Eqs. of 1,4-Butanediol oil Mw1187.9637.6443.4332.0134.9Bio-oil Mn350.5261.8221.9181.3102.7Bio-oil PDI3.392.442.001.831.31Methoxy-phenols selectivity(wt%)59315562Alkyl-phenols selectivity (wt%)9591694538Bio-oil Yieldamounts were observed for HZSM-5(23) and non-catalytic pyrolysis.In our work, lab scale microwavepyrolysis of lignin has been performedand the liquid products obtained areWhen cellulose was used as thecomposed of smaller polymericstarting material rather thancomponents and moderate yields ofswitchgrass, the same trend ismonomeric phenols. However, uponobserved, and the production of alkylthe addition of 1,4-butanediol,phenols from cellulose (not just lignin)repolymerization reactions that limit theis verified. Selected alkyl phenols were yield of monomeric and other reducedproduced in a yield of about 9 mg/gmolecular weight products are inhibited.from cellulose over KZSM-5 (23),As shown in Table 1, a 85-90%compared with 3.8 mg/g over HZSM-5 reduction in the average molecular(23) and only trace amounts nonweight of the liquid products wascatalytically. Similarly, the yield of 2observed concurrent to an overallmethyl furan increased to over 5 mg/gincrease in liquid yield. At theover KZSM-5 (23), compared with 1optimized ratio of 2:1 lignin to 1,4mg/g using HZSM-5 (23). More details butanediol (w/w), the yield of selectedof this work can be found in our recent monomeric phenols increased threeACS Sustainable Chemistry andfold to 3.4 wt% (based on feedstock),Engineering publication [1].while the yield of mono-aromatichydrocarbons decreased byMeanwhile, in another development inapproximately 90%. The addition of theour laboratory we have also found that diol co-reactant also led to a significantmicrowave co-pyrolysis of lignin withshift in selectivity towards thehigh boiling diols can increase its netproduction of methoxy-phenolsdepolymerization, resulting in an(guaiacols, syringols) over nonincrease in production ofmethoxylated alkyl-phenols (phenol,methoxyphenols along with an overallcresols, etc.). The results obtainedmay help lead to the development ofnovel lignin co-processing methods.We believe that the addition of the diolmay quench radical or acid initiatedreploymerization reactions leading tothe observed higher yield and lowermolecular weight, and also thatpreservation of the methoxy groupsmay also decrease the occurrence ofrepolyemrization reactions duringpyrolysis. More details on thisdevelopment can be found in our recentjournal publication [2].We are currently building off thesedevelopments to further increase theproduction of phenolics from cellulose,lignin and biomass. This includesdevelopment of catalysts moreselective for phenols, transfer ofoperations to a continuous larger scaleprocess and development of methodsfor the selective isolation of phenolsfrom bio-oil.References[1] Mullen, C.A., Tarves, P.C., Boateng,A.A. Role of Potassium Exchange inCatalytic Pyrolysis of Biomass overZSM-5: Formation of Alkyl phenols andFurans. ACS Sustainable Chemistryand Engineering, 5, 2154-2162. 2017.[2] Tarves, P. C., Mullen, C.A., Strahan,G. D., Boateng, A.A. Depolymerizationof Lignin via Co-pyrolysis with 1,4Butanediol in a Microwave Reactor.ACS Sustainable Chemistry andEngineering, 5, 988-994. 2017ContactCharles A. Mullen, PhDResearch ChemistSustainable Biofuels and CoproductsEastern Regional Research Center,ARS, USDA600 E. Mermaid Lane, Wyndmoor PA19038 USAcharles.mullen@ars.usda.gov 1-215-836-6916

PyNe 41, IEA Bioenergy Task 34Page 8 of 24Applications of Phenolic Oil Derived from Fast PyrolysisLow-temperature, low-pressurehydrogenationMarjorie R. RoverThe superficial similarities betweenpetroleum and bio-oil haveencouraged efforts to upgrade it in thesame way as petroleum. Any facilecomparison is overshadowed by thefact that petroleum contains non-polarhydrocarbons that are relatively stable,requiring elevated temperatures andpressures to encourage chemicaltransformations, whereas bio-oilconsists of oxygenated organiccompounds whose high degree offunctionality makes them chemicallyreactive even at low temperatures andpressures. Due to the reactive natureof the phenolic oil, low-temperature,low-pressure hydrogenation (LTLP-H)was utilized to produce a stable, lowviscosity product at high yields [4].Low-temperature, low-pressurehydrogenation, which was performedat 21 C and 1 bar pressure, resultedin high carbon yields: 98.2% to 98.7%for phenolic oils from corn stover andred oak, respectively. The viscositiesof phenolic oil samples (measured at60 C) dropped dramatically afterhydrogenation: cornstover fraction 1decreased by 81% and fraction 2decreased by 47%. Red oak fraction 2Bioeconomy InstituteIowa State UniversityFast pyrolysis of lignin results in amixture of phenolic monomers andoligomers sometimes referred to aspyrolyticlignin although phenolic oil isIndividuala moreHighlights:accurate characterization.Iowa State University (ISU) hasdevelopeda fractionatingInsideStory2 recoverysystem that collects phenolic oil has3part of a InsideheavyStoryends fraction[1, 2]. Asimple waterextractionisabletoInside Story4separate the water-insoluble phenolicInside solubleStory sugars5oil from waterandanhydrosugars [3].Last Story6Phenolic oil presents 40-45% of thecarbon content of the products ofpyrolysis. Most of the aromatic contentof the lignin is preserved. Its oxygencontent is approximately 20-24%,much lower than that of thecarbohydrate-derived products ofpyrolysis, giving it an attractive heatingvalue for production of fuels.However, it is extremely reactive instorage or during thermal processing,which can complicate its refining,although in some cases this reactivitycan be exploited. This paper exploressome options for upgrading phenolicoil into various products.Fig. 1 - Gas chromatography simulated distillation (SimDist ASTM D2887) ofhydrotreated phenolic oil utilizing sulfided CoMo and Ru-Pd catalysts.experimental runs.heavy ends showed a more dramaticdrop in viscosity, decreasing by 99%.This viscosity thinning afterhydrogenation is in sharp contrast tothe viscosity thickening observed byresearchers who used highertemperatures and pressures tohydroprocess bio-oil. Viscosityreductions are usually associated withmore severe hydrocracking, whichreduces the molecular weight of heavyorganic compounds. Cracking ofphenolics clearly did not occur duringLTLP-H. Further analytical testingsuggested that viscosity thinning wasthe result of the self-solvation power ofproduced alcohol duringhydrogenation.(Continued on page 9)

Page 9 of 24PyNe 41, IEA Bioenergy Task 34Applications of Phenolic Oil Derived from Fast Pyrolysis continuedCatalytic hydroprocessing ofphenolic oilsHydroprocessing can deoxygenateand crack phenolic oil into fuel-rangehydrocarbon molecules asdemonstrated in a collaborationbetween ISU and Pacific NorthwestNational Laboratory (PNNL) [5]. Redoak and corn stover were pyrolyzedand the phenolic oil recovered usingwater extraction of the heavy ends atISU. The phenolic oils from these twofeedstocks were shipped to PNNLwhere they were catalyticallyhydrotreated to remove heteroatoms.The carbon yields were as high as90% carbon yield to liquidhydrocarbons from the phenolic oil.The liquid product from hydrotreatedred oak phenolic oil was analyzed bygas chromatography simulateddistillation. Figure 1 indicates asignificant portion of the hydrotreatedproducts fell within gasoline rangemolecules (42-52%), whereas, 43%fell within diesel range molecules [3].Lignocol was well below that of mostcoals.Coal substitute from curingphenolic oilFractionated phenolic oil as binderin asphalt and pavementThe reactivity of phenolic oil can beexploited to produce an attractive lowash, low sulfur coal substitute calledLignocol. Phenolic oil heated to 105220 C for 2 hours polymerizes to avitreous solid resembling coal withyields of 87.4-94.5 % depe

bioenergy into the market, with hard questions from the audience generating interesting debate. Topics included both global and Demonstrating sustainable bioenergy and Scaling it up Fig. 1: Biomass production and fuel storage in Sweden destined for transport diesel fuel. December 2017 IEA Bioenergy faction nfo: m /