Fast Pyrolysis of Cellulose by Infrared Heating

energies

Article Fast Pyrolysis of Cellulose by Infrared Heating

Takashi Nomura , Hinano Mizuno, Eiji Minami and Haruo Kawamoto *

Graduate School of Energy Science, Yoshida-Honmachi, Sakyo-Ku, Kyoto 606-8501, Japan; [email protected] (T.N.); [email protected] (H.M.); [email protected] (E.M.) * Correspondence: [email protected]

Abstract: The fast pyrolysis of cellulose produces levoglucosan (LG), but secondary pyrolysis reactions tend to reduce the yield. The present study assessed the fast pyrolysis of cellulose by infrared (IR) heating under nitrogen flow. Because the nitrogen was not efficiently heated, gaseous LG was immediately cooled, resulting in a maximum yield of 52.7% under optimized conditions. Slow nitrogen flow and a high IR power level provided a greater gas yield by raising the temperature of the cellulose, and the formation of CO could be used as an indicator of the gasification of LG. Glycolaldehyde (GA) was the major byproduct, and the GA yield remained relatively constant under all conditions. Accordingly, GA was not a secondary product from the LG but was likely produced from the reducing ends of cellulose and other intermediate carbohydrates. The pyrolysis of cellulose proceeded within a narrow region of carbonized material that absorbed IR radiation more efficiently. The bulk of each cellulose sample could be decomposed in spite of this heterogeneous process by maintaining fast pyrolysis conditions for a sufficient length of time. This technique is a superior approach to LG production compared with other fast pyrolysis methods based on heat conduction. Keywords: cellulose; fast pyrolysis; infrared heating; levoglucosan; glycolaldehyde

Citation: Nomura, T.; Mizuno, H.; Minami, E.; Kawamoto, H. Fast Pyrolysis of Cellulose by Infrared Heating. Energies 2021, 14, 1842. 1. Introduction https://doi.org/10.3390/en14071842 Owing to the depletion of fossil fuel resources and environmental concerns, there is currently significant interest in the use of renewable resources. Biomass has received Academic Editors: Fernando Rubiera much attention in this regard as the only renewable carbon source that can be converted González and Covadonga Pevida into useful chemicals and fuels. Cellulosic biomass accounts for the majority of biomass García resources on Earth, but the polymeric nature of these materials limits their widespread use. Accordingly, the development of efficient cellulose conversion technologies is very impor- Received: 16 March 2021 tant. Pyrolysis conducted under oxygen-free conditions is a potentially viable approach to Accepted: 23 March 2021 this goal because this technique is able to rapidly degrade stable cellulosic biomass [1,2]. Published: 26 March 2021 Fast pyrolysis is a process characterized by rapid heating to yield a mixture of liquid products, known as bio-oil, instead of char (that is, solid products) from lignocellulosic Publisher’s Note: MDPI stays neutral biomass [3–7]. The bio-oil produced can be utilized as a source of liquid fuels (gasoline and with regard to jurisdictional claims in diesel) and biochemicals. Levoglucosan (LG, 1,6-anhydro-β-D-glucopyranose) is one of the published maps and institutional affil- iations. major components of bio-oil obtained from cellulose [8–10] and can be used as a source of various chemicals. As an example, the hydrolysis of LG gives glucose, which can be converted to ethanol, lactic acid and other products via fermentation. It has been reported that the temperature of cellulose during fast pyrolysis is in the range of 400–450 ◦C, which is much higher than the value of approximately 350 ◦C Copyright: © 2021 by the authors. associated with slow pyrolysis [11–14]. This overheating can promote the recovery of LG, Licensee MDPI, Basel, Switzerland. which has a boiling point of 385 ◦C[14], by rapid evaporation. However, the efficient This article is an open access article production of LG from the fast pyrolysis of cellulose requires information regarding the distributed under the terms and conditions of the Creative Commons susceptibility of this compound to thermal degradation. Prior work has shown that LG ◦ Attribution (CC BY) license (https:// undergoes thermal polymerization in the vicinity of 250 C, which is much lower than the creativecommons.org/licenses/by/ temperature required for the formation of LG during cellulose pyrolysis [15–18]. These 4.0/). seemingly contradictory temperature values can be explained by the difference in LG

Energies 2021, 14, 1842. https://doi.org/10.3390/en14071842 https://www.mdpi.com/journal/energies Energies 2021, 14, 1842 2 of 13

reactivity in the molten and gas phases. In the gas phase, LG is stable up to 500–600 ◦C and fragments into gaseous products at higher temperatures [16]. The greater reactivity of molten LG can be explained by hydrogen bonding between LG molecules, which act as acidic and basic catalysts [19,20]. The literature therefore indicates that the capture of gaseous LG produced by the fast pyrolysis of cellulose is very important to obtain a high yield of this compound. To maintain such a high cellulose degradation temperature (400–450 ◦C), various types of furnaces have been developed and are classified into two types depending on the heating principle [3,6,11,12,21–28]; the entrained downflow, ablative reactor and fluid bed types are based on conductive heating, where biomass is heated by contact with a hot environment or hot metal surface [3,6,21–25]. The other type is based on radiation heating [11,12,26–28], and biomass is efficiently heated by adsorbing radiation such as infrared (IR) [11,12], condensed solar [26] and lasers [27,28]. This feature of selective heating favors the production of LG from cellulose. The precise control of the pyrolysis temperature is vital during the production of LG because gaseous LG is quickly converted into other products such as CO and H2 above 600 ◦C[16]. For this reason, radiation heating, which can heat materials selectively, may be superior to conductive heating. Shoji et al. [14] reported that a furnace temperature in excess of 600 ◦C was necessary to obtain a cellulose degradation temperature above 400 ◦C so as to maintain fast pyrolysis conditions with a small amount of cellulose in a preheated furnace. Under such conditions, the secondary degradation of LG is inevitable even in the gas phase. In order to suppress the secondary decomposition of LG in the gas phase, IR heating was selected for the pyrolysis of cellulose. When using radiation heating, the LG produced is efficiently cooled using a stream of nitrogen that has not been effectively heated due to insufficient IR absorption of nitrogen. The fast pyrolysis of cellulose using radiation heating has been previously reported [11,12,27–32], while most of the studies focus on the utilization of solar energy for pyrolysis [29,30] and the formation of melting substances during irradiation [11,12,27,28]. As for the production of LG, Suzuki et al. [21] reported the formation of LG and anhydro- oligocellosaccharides by irradiating a CO2 laser on cellulose, although the yields from cellulose were quite low. By using similar CO2 laser irradiation under nitrogen flow or vacuum conditions, Kwon et al. reported that the yield of LG reached a maximum at around 25% from Whatman CF11 cellulose powder. However, the efficient production of LG from cellulose by radiation heating has not been achieved, and the thermal degradation mechanisms under the irradiation conditions have not been fully clarified. In the present study, the production of LG from cellulose by IR heating under a nitrogen flow was studied. After the investigation of the effects of the experimental parameters on the product yield, the thermal degradation mechanism at the surface where IR is absorbed was evaluated.

2. Materials and Methods 2.1. Cellulose Samples Whatman No. 42 cotton filter paper (Whatman PLC, Maidstone, UK, pore size: 2.5 µm) and microcrystalline cellulose powder (Avicel PH-101, Asahi Kasei Corp., Tokyo, Japan) were used in the pyrolysis trials. These materials were employed as received without further purification. The Whatman filter paper was cut into 1.0 × 4.3 cm pieces weighing 30 mg (dry) before use.

2.2. Pyrolysis and Product Analysis Figure1 shows a diagram of the experimental set-up, in which an IR image furnace (RHL-E45N, ADVANCE RIKO, Kanagawa, Japan) was used to heat the cellulose. A quartz tube (internal diameter: 30 mm, length: 495 mm) was set inside a cylindrical furnace, and the IR radiation was focused on the center of the furnace. During trials with the Whatman cellulose sheets, each sample was placed on a stainless steel mesh situated at the center of Energies 2021, 14, x FOR PEER REVIEW 3 of 14

Energies 2021, 14, x FOR PEER REVIEW 3 of 14

Energies 2021, 14, 1842 the IR radiation was focused on the center of the furnace. During trials with the3 Whatman of 13 cellulose sheets, each sample was placed on a stainless steel mesh situated at the center of thethe reactor. IR radiation In the was case focused of the on Avicel the center cellulose of the trials,furnace. a quartzDuring boattrials (length:with the Whatman35 mm, width: 10thecellulose mm, reactor. height: sheets, In the6 each mm, case sample ofAS the ONE Avicelwas placedCorporation, cellulose on a trials,stainless Osaka, a quartz steel Japan) mesh boat containing (length:situated 35at the mm,the center width:sample of was the reactor. In the case of the Avicel cellulose trials, a quartz boat (length: 35 mm, width: placed10 mm, on height: the mesh 6 mm, and AS then ONE inserted Corporation, into the Osaka, reactor. Japan) A sampling containing bag the made sample of wasTedlar® (5 10 mm, height: 6 mm, AS ONE Corporation, Osaka, Japan) containing the sample was® L)placed containing on the methanol mesh and (30 then mL) inserted was attached into the reactor.to the outlet A sampling of the bagreactor made tube of Tedlarto collect the (5placed L) containing on the mesh methanol and then (30 inserted mL) was into attached the reactor. to the A outlet sampling of the bag reactor made tube of Tedlar to collect® (5 volatile products, and the air inside the reactor system was replaced with nitrogen before theL) containing volatile products, methanol and (30 themL) air was inside attached the reactorto the outlet system of wasthe reactor replaced tube with to collect nitrogen the eachbeforevolatile trial each products,by purging trial by and purging with the aira nitrogen withinside a the nitrogen flow reac tor(5 flow L/min)system (5 L/min) wasfor 5 replaced min. for 5 The min. with nitrogen The nitrogen nitrogen in before the in sample bagtheeach was sample trial released by bag purging was from released with the a vent nitrogen from before the flow vent the (5 before pyroL/min) thelysis for pyrolysis trial,5 min. and The trial, the nitrogen and process the in process the time sample timewas suffi- cientlywasbag sufficientlywas short released to accommodate short from to the accommodate vent the before whole the the pyrovolume wholelysis volume oftrial, injected and of injected the nitrogen. process nitrogen. time After was After the suffi- thenitrogen flownitrogenciently was short adjusted flow to was accommodate adjustedto the designated to the the designated whole value volume value(from of (from 1injected to 110 to L/min), 10nitrogen. L/min), the After thecellulose cellulose the nitrogen was was heated usingheatedflow an was usingIR adjusted output an IR energy outputto the designated energyin the inrange thevalue range of (fr0.5om of to 0.51 4.0 to to 10kW. 4.0 L/min), kW. the cellulose was heated using an IR output energy in the range of 0.5 to 4.0 kW. Sample bag Quartz glass tube Sample bag Quartz glass tube Cellulose Cellulose N2 N2

StainlessStainless steel steel mesh mesh

Infrared furnace MethanolMethanol Infrared furnace

FigureFigureFigure 1. The 1.1. The fast fast pyrolysis pyrolysispyrolysis system systemsystem incorporating incorporatingincorporating anan an infraredin infraredfrared imageimage image furnacefurnace furnace usedused used inin thisthis in study.thisstudy. study.

DuringDuring thethe pyrolysis,pyrolysis, aa colorlesscolorless mistmist escapedescaped fromfrom thethe reactorreactor andand waswas capturedcaptured inin During the pyrolysis, a colorless mist escaped from the reactor and was captured in thethe samplingsampling bag,bag, where where it it collected collected on on the the inner inner walls walls after after standing standing for for 30 30 min min (Figure (Figure2). theA2). sampling 5.0 A mL5.0 mL quantity bag,quantity where of neon of neonit gas collected wasgas was added on added tothe the innerto sample the wallssample bag afteras bag an standingas internal an internal standard, for 30standard, min after (Figure 2).whichafter A 5.0 which the mL contents thequantity contents were of analyzed wereneon analyzed gas by was micro by added micro gas chromatography togas the chromatography sample(GC, bag CP-4900,as (GC, an CP-4900,internal Varian Var- Inc.,standard, afterPaloian whichInc., Alto, Palo CA,the Alto, USA).contents CA, The wereUSA). chromatographic analyzed The chromato by conditions micrographic gas conditions included chromatography a 10included m MS5 (GC,a A 10 column, m CP-4900, MS5 Ar A Var- ◦ ianascolumn, Inc., the carrierPalo Ar asAlto, gas, the a carrierCA, column USA). gas, temperature a Thecolumn chromato temperature of 100graphicC, a of column 100conditions °C, pressure a column included of pressure 170 kPa a 10 andof m170 aMS5 A column,thermalkPa and Ar conductivitya thermalas the carrier conductivity detector gas, (TCD).a detectorcolumn Using (TCD).temperature these Using conditions, theseof 100 conditions, the°C, compounds a column the compounds pressure analyzed of 170 kPa(andanalyzed and their a thermal(and retention their conductivity times)retention were times) Nedetector (25.7were s), Ne(TCD). H (25.72 (27.0 s),Using s), H O2 (27.02 these(35.8 s), s), conditions,O2 CH (35.84 (60.9 s), CH the s) and4 (60.9compounds CO s) (79.8 s) (Figure S1A). A second channel on the instrument included a 10 m PoraPLOT Q analyzedand CO (and (79.8 theirs) (Figure retention S1A). times) A second were channel Ne (25.7 on s), the H 2instrument (27.0 s), O included2 (35.8 s), a CH 10 4 m(60.9 s) column with He as the carrier gas, a column temperature of 80 ◦C, a column pressure of andPoraPLOT CO (79.8 Q columns) (Figure with S1A). He as theA secondcarrier gas, channel a column on temperaturethe instrument of 80 °C,included a column a 10 m 190 kPa and a TCD. The associated analytes (and retention times) were CO2 (19.3 s), C2H4 PoraPLOTpressure ofQ 190column kPa and with a TCD.He as The the associatedcarrier gas, analytes a column (and temperature retention times) of 80 were °C, COa column2 (22.6(19.3 s),s), CC22H64 (22.6(25.3 s), s) andC2H C6 (25.33H6 (53.5 s) and s) C (Figure3H6 (53.5 S1B). s) (Figure S1B). pressure of 190 kPa and a TCD. The associated analytes (and retention times) were CO2 (19.3 s), C2H4 (22.6 s), C2H6 (25.3 s) and C3H6 (53.5 s) (Figure S1B).

FigureFigure 2.2. PhotographicPhotographic imagesimages ofof samplesample bagsbags immediatelyimmediately afterafter pyrolysispyrolysis andand 3030 minmin later.later.

AfterAfter eacheach trial,trial, thethe condensatescondensates inin thethe samplesample bagbag andand onon thethe wallswalls ofof thethe reactorreactor Figurewerewere 2. extractedextracted Photographic withwith methanol methanolimages of (200 (200sample mL),mL), bags afterafter imme whichwhichdiately thethe methanol methanolafter pyrolysis waswas evaporatedevaporated and 30 min under underlater. vacuum.vacuum. AA BrukerBruker AC-400 AC-400 (400 (400 MHz) MHz) spectrometer spectrometer was was subsequently subsequently used used toacquire to acquire a 1H a nuclearAfter magnetic each trial, resonance the condensates (NMR) spectrum in the of thesample resulting bag material and on in the dimethylsulfoxide walls of the reactor were extracted with methanol (200 mL), after which the methanol was evaporated under vacuum. A Bruker AC-400 (400 MHz) spectrometer was subsequently used to acquire a

Energies 2021, 14, x FOR PEER REVIEW 4 of 14

1H nuclear magnetic resonance (NMR) spectrum of the resulting material in dimethylsulfoxide (DMSO)-d6 (0.7 mL), containing maleic acid as an internal standard and hydrox- Energies 2021, 14, 1842 4 of 13 ylamine hydrochloride (NH2OH•HCl, 10 mg) for the in situ derivatization of aldehydes and ketones in the product mixture to the corresponding oximes. The yields of the products, including LG and glycolaldehyde (GA), were determined by comparing the peak areas(DMSO)- of thed6 (0.7characteristic mL), containing signals maleic of these acid ascompounds an internal with standard that and of the hydroxylamine internal standard hydrochloride (NH2OH•HCl, 10 mg) for the in situ derivatization of aldehydes and (Figure S2). The methanol, DMSO-d6, maleic acid and hydroxylamine hydrochloride were ketones in the product mixture to the corresponding oximes. The yields of the products, purchasedincluding LGfrom and Nacalai glycolaldehyde Tesque, (GA),Inc., Kyot wereo, determined Japan and by used comparing without the purification. peak areas of the characteristicThe solid residue signals remaining of these compounds on the stainless with that steel of the mesh internal was standard weighed (Figure and S2).the result isThe referred methanol, to as DMSO- the chard6 ,mass maleic herein. acid and The hydroxylamine mass of the solid hydrochloride carbonized were materials purchased adhering tofrom the Nacalaiinner surface Tesque, of Inc., the Kyoto, sample Japan boat and was used obtained without from purification. the mass loss after heating the boat inThe air solid at 600 residue °C for remaining 2 h. on the stainless steel mesh was weighed and the result is referredThe totemperature as the char mass in the herein. interior The of mass the of Avicel the solid cellulose carbonized in the materials sample adhering boat was to moni- toredthe inner using surface a fine of thermocouple the sample boat (0.25 was mm obtained in diameter, from the type mass K, loss Shinnetsu after heating Co., the Ltd., boat Ibaraki, in air at 600 ◦C for 2 h. Japan) connected to a thermologger (AM-8000, Anritsu Corporation, Kanagawa, Japan). The temperature in the interior of the Avicel cellulose in the sample boat was mon- Theitored tip of using the athermocouple fine thermocouple was embedded (0.25 mm inin diameter,the center type of the K, sample Shinnetsu in the Co., depth Ltd., direction.Ibaraki, Japan) connected to a thermologger (AM-8000, Anritsu Corporation, Kanagawa, Japan).Most The of the tip ofexperiments the thermocouple were repeated was embedded three times in the and center the results of the sample are shown in the as mean valuesdepth along direction. with standard deviations. Most of the experiments were repeated three times and the results are shown as mean 3.values Results along and with Discussion standard deviations. 3.1.3. ResultsEffects of and the Discussion Experimental Parameters on the Product Yield 3.1.The Effects effects of the of Experimental the experimental Parameters parameters on the Product on the Yield yields of LG and other products were evaluatedThe effectsusing ofthe the Whatman experimental cellulose parameters to evaluate on the yields the capacity of LG and of other this productspyrolysis were system to generateevaluated LG. using The the effects Whatman of modifying cellulose to the evaluate position the of capacity the cellulose of this pyrolysis in the reactor system (at to Points A,generate B and C LG. in Figure The effects 3), the of modifyingflow rate of the nitrogen position and of the the cellulose IR power in theon reactorthe product (at Points yield were evaluated.A, B and C Figure in Figure 4 summarizes3), the flow ratethe ofyields nitrogen of LG, and gas the and IR powerchar obtained on the product under various yield con- were evaluated. Figure4 summarizes the yields of LG, gas and char obtained under various ditions. The yield is not 100% in total due to the presence of unquantified products such as conditions. The yield is not 100% in total due to the presence of unquantified products water.such asIn water.general, In general,a large amount a large amount of water of wateris produced is produced during during the pyrolysis the pyrolysis of cellulose. of However,cellulose. the However, amount the of amount water ofwas water not wasquantified not quantified in the present in the present paper. paper.

140 mm

10 mm 10 mm N₂ ABC

+7.5 C̊ ( control ) +2.8 C̊ ( control ) Infrared furnace +27.6 C̊ ( with mesh) +2.5 C̊ ( with mesh)

Figure 3. Positions (A–C) of Whatman cellulose samples in the reactor and changes in temperature at the upstream Figure 3. Positions (A–C) of Whatman cellulose samples in the reactor and changes in temperature at the upstream and and downstream positions of the furnace during blank heating trials with and without a stainless steel mesh (4.0 kW/ downstream positions of the furnace during blank heating trials with and without a stainless steel mesh (4.0 kW/5 s irra- 5 s irradiation/10 L/min nitrogen ﬂow). The values shown here are the increases relative to room temperature. diation/10 L/min nitrogen flow). The values shown here are the increases relative to room temperature. The position of the cellulose was expected to affect the LG yield because the residence time of the volatile products in the IR irradiation area varied depending on the position (C > B > A). We anticipated that the extent to which secondary pyrolysis reactions of the gaseous LG proceeded would be increased with increases in the residence time. However, the LG yield, which ranged from 32% to 42%, was not greatly changed when modifying the sample position (Figure4) during the trial employing 4.0 kW radiation, 5 s irradiation and a 10 L/min nitrogen ﬂow. These results indicate that the degree of secondary degradation of the gaseous LG in response to IR irradiation was very limited, possibly because of the minimal absorption of IR radiation by the gaseous LG and the nitrogen carrier gas. In this respect, IR heating is superior to other conductive heating methods for the production of LG via the fast pyrolysis of cellulose.

EnergiesEnergies 2021, 202114, x, 14FOR, 1842 PEER REVIEW 5 of 13 5 of 14

50 50 100 4 kW, 10 L/min, 5 s LG 4 kW, 5 s Char 5L/min, 5 s 40 LG 40 80

30 30 60 LG 40 20 Gas 20 Gas 10 10 20 Gas Char Char

Yield Yield (wt%, cellulose base) 0 0 0 ABC 024681012 01234 Sample position N₂ flow (L/min) Power (kW)

FigureFigure 4. Effects 4. Effects of experimental of experimental parame parametersters on the on theyields yields of levoglucosan of levoglucosan (LG), (LG), gas gas and and char char from from the the pyrolysis pyrolysis of of What- man cellulose.Whatman cellulose. Unexpectedly, the gas yield was high (10.1 ± 5.6%) when the cellulose was set near the outletThe of position the furnace of the (Position cellulose A). was The secondaryexpected degradationto affect theof LG gaseous yield because LG to permanent the residence ◦ timegases of the such volatile as COand products H2 is significant in the IR aboveirradiation 600 C area in conjunction varied depending with short on residence the position (C times> B > ofA). 1–2 We s similar anticipated to those that in the the present extent study to which [18], andsecondary such high pyrolysis temperatures reactions could of the gaseoushave occurred LG proceeded at Position would A. This be increased is also supported with increases by the gas in composition,the residence as time. discussed However, thelater. LG yield, To evaluate which this ranged possibility, from the32% temperatures to 42%, was upstream not greatly and changed downstream when of themodifying IR thefurnace sample were position measured (Figure in blank 4) during tests with the trial and withoutemploying the stainless 4.0 kW radiation, steel mesh (Figure5 s irradiation3, and4.0 a kW,10 L/min 5 s irradiation, nitrogen 10 flow. L/min These nitrogen results flow). indicate The temperature that the degree was found of secondary to increase degra- by 4.7 and 25.1 ◦C without and with the mesh, respectively, when the nitrogen carrier dation of the gaseous LG in response to IR irradiation was very limited, possibly because gas passed through the IR irradiation zone. These increases were insufficient to promote of secondarythe minimal degradation absorption of of the IR LG radiation but could by possibly the gaseous haveraised LG and the the temperature nitrogen carrier of the gas. In cellulosethis respect, to promote IR heating the secondary is superior degradation to other conductive of LG at Position heating A. methods for the production of TheLG via nitrogen the fast flow pyrolysis rate was of expected cellulose. to affect the reaction yield by carrying volatile productsUnexpectedly, to a cooler the region gas yield of the was apparatus high (10.1 such that± 5.6%) they when did not the condense cellulose on was the hotset near thereactor outlet wall of the where furnace secondary (Position degradation A). The may secondary occur. The degradation flow rate could of also gaseous conceivably LG to per- manentaffect thegases cellulose such as degradation CO and H temperature2 is significant during above the pyrolysis600 °C in because conjunction the nitrogen with short residencewas not times efficiently of 1–2 heated s similar by the to IR those radiation. in the Fourpresent nitrogen study flow [18], rates and (1.0, such 3.0, high 5.0 andtempera- 10.0 L/min) were assessed at a constant IR power of 4.0 kW and constant irradiation time tures could have occurred at Position A. This is also supported by the gas composition, as of 5 s. The data show that the LG yield was increased with decreases in the flow rate from discussed10 to 3.0 later. L/min, To which evaluate can bethis explained possibility by, thethe greater temperatures heating efficiencyupstream at and lower downstream flow of ratesthe IR of furnace the cool were nitrogen. measured The gas in yields blank significantly tests with increasedand without when the the stainless flow rate steel was mesh (Figurefurther 3, decreased 4.0 kW, 5 to s 3.0irradiation, and 1.0 L/min, 10 L/min while nitrogen the LG yield flow). was The decreased temperature at 1.0 L/min was asfound a to increaseresult ofby secondary 4.7 and 25.1 degradation. °C without and with the mesh, respectively, when the nitrogen carrier gasThe passed IR power through directly the IR affected irradiation the heating zone. These efficiency increases of the cellulose.were insufficient Consequently, to promote secondarythe LG yield degradation decreased of significantly the LG but whencould the possibly IR power have was raised reduced the fromtemperature 1.0 to 0.5 of kW, the cel- lulosewhile to thepromote char yield the secondary was greatly de increasedgradation as of a resultLG at ofPosition the large A. amount of unreacted cellulose. Experiments with a longer irradiation time of 10 s were also conducted at 0.5 and The nitrogen flow rate was expected to affect the reaction yield by carrying volatile 1.0 kW (Table1). Prolonging the irradiation time increased the LG yield while lowering the productschar yield, to a and cooler the maximum region of LG the yield apparatus of 52.7% such was obtained that they after did 10 not s of condense irradiation on at anthe hot reactorIR power wall of where 1.0 kW. secondary The three-time degradation average may under occur. the same The conditions flow rate wascould47.5 also± 3.8% conceiva-. blyThese affect yields the cellulose were higher degradation than those temperatur reported ine during other studies the pyrolysis using radiation because heatingthe nitrogen was(10–30%) not efficiently and fluidized heated bed by (up the to 40%)IR radiation. [29–32]. For Four example, nitrogen the LG flow yields rates reported (1.0, 3.0, in the 5.0 and 10.0literature L/min) usingwere aassessed CO2 laser at werea constant less than IR 25%power [31 ].of The 4.0 greaterkW and sweep constant efficiency irradiation of the time of 5nitrogen s. The data flow systemshow that in the the present LG yield study was may increased increase thewith LG decreases recovery atin thethe expense flow rate of from 10 theto 3.0 occurrence L/min, ofwhich the secondary can be explained degradation by of the LG. greater heating efficiency at lower flow rates of the cool nitrogen. The gas yields significantly increased when the flow rate was further decreased to 3.0 and 1.0 L/min, while the LG yield was decreased at 1.0 L/min as a result of secondary degradation. The IR power directly affected the heating efficiency of the cellulose. Consequently, the LG yield decreased significantly when the IR power was reduced from 1.0 to 0.5 kW, while the char yield was greatly increased as a result of the large amount of unreacted cellulose. Experiments with a longer irradiation time of 10 s were also conducted at 0.5 and 1.0 kW (Table 1). Prolonging the irradiation time increased the LG yield while lower-

Energies 2021, 14, 1842 6 of 13

Table 1. Yields of LG, gas and char from the pyrolysis of Whatman cellulose while varying the IR power and irradiation time with a 5 L/min nitrogen ﬂow.

Yield (wt%, Cellulose Base) Infrared Irradiation LG Gas Char Power (kW) Time (s) 5 4.9 ± 5.6 (48.1 ± 1.8) * 0.9 ± 0.7 89.8 ± 7.9 0.5 10 35.0 ± 2.7 (51.3 ± 1.3) 4.0 ± 3.3 31.7 ± 5.6 5 41.3 ± 4.6 (51.9 ± 4.1) 1.6 ± 0.6 20.4 ± 8.1 1.0 10 47.5 ± 3.8 (48.3 ± 3.7) 4.3 ± 0.1 1.7 ± 0.8 * Figures in parentheses are estimated yields based on the amounts of cellulose that were degraded over 0–5 s and 5–10 s. The latter amount was calculated from the difference in mass between the char portions obtained at 5 s and at 10 s.

As shown by the numbers in parentheses, the yields of LG during the irradiation periods of 0–5 and 5–10 s were determined by assuming that the difference in the mass values of the char at 5 s and at 10 s equaled the amount of cellulose degraded during the irradiation time from 5 to 10 s. It should be noted that the estimated LG yields during the first and last 5 s intervals were not greatly different. This result illustrates a very important aspect of IR heating that is discussed in more detail below. These data establish that the IR power, nitrogen flow rate and sample position all affected the actual cellulose degradation temperature. When this temperature was greater than 600 ◦C, the LG yield decreased because of the conversion of the LG to permanent gases. The IR irradiation time was also found to have an important effect in terms of obtaining complete cellulose pyrolysis. The major byproducts obtained from the fast pyrolysis of cellulose under IR irradiation in this work were gaseous but the specific compounds changed depending on the pyrolysis conditions. In Figure5, the molar percentages of H 2, CO and CO2 are plotted against the gas yields from the various trials. At a gas yield less than 4%, H2 and CO2 were the main components, while at higher gas yields, CO was produced, and the molar ratios of CO to H2 became 1.3–2.5 along with lesser amounts of CO2. These results are reasonably explained with gas production from the secondary degradation of gaseous LG, based on prior reports that CO and H2 are selectively produced from LG in the gas phase above 600 ◦C[18]. Therefore, the formation of CO appears to be a useful indicator of the appearance of cellulose degradation temperatures greater than 600 ◦C. It should be Energies 2021, 14, x FOR PEER REVIEW noted that it was challenging to measure the actual sample temperature throughout the IR 7 of 14 irradiation process because of the nonuniform progression of the thermal degradation of the cellulose, as discussed further on.

80 80 100 H₂ CO CO₂ 80 60 60 60 40 40 40 20 20 20 Yield base) gas Yield (mol%, 0 0 0 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 Gas yield (wt%, cellulose base)

Figure 5. Proportions of H2, CO and CO2 in the gaseous products as functions of the overall gas yield from the fast pyrolysis Figure 5. Proportions of H2, CO and CO2 in the gaseous products as functions of the overall gas yield from the fast pyrolysis of Whatman cellulose. of Whatman cellulose.

GA was the other major byproduct, and Figure 6 summarizes the yields of GA obtained under various fast pyrolysis conditions compared with the overall gas yields. It is apparent that there were no correlations between the yields of both products and that the GA yield remained relatively constant as the various parameters were changed while the gas yield varied greatly. Accordingly, the formation of GA cannot be explained by the secondary degradation of LG.

4 kW, 10 L/min, 5 s 4 kW, 5 s 5 L/min, 10 s 20 Gas 15

5 GA

Yield (wt%, cellulose base) 0 ABC1351 Power (kW) Set point N₂ flow (L/min)

Figure 6. Yields of glycolaldehyde (GA) and overall gas yields (wt % based on amount of cellulose) from the pyrolysis of Whatman cellulose under various conditions.

In the pyrolysis of cellulose, GA has been reported to form in a higher temperature range than LG, and a negative correlation has been reported between the yields of these products [33]. Therefore, there was controversy over the GA formation pathway, although the literature [33–36] shows a direct formation pathway from cellulose that competes with LG formation. The observations in the present study clearly show the direct pathway of GA formation. GA is believed to form during the primary cellulose pyrolysis stage, likely via retro-Aldol-type fragmentation reactions of the reducing ends of cellulose and carbo- hydrate intermediates. Such reactions are favorable because stable six-membered cyclic transition states are involved (Figure 7). A similar mechanism has been proposed for the pyrolysis of cellobiose based on prior work using pyrolysis-GC/MS spectrometry in conjunction with 13C-labeled cellobiose [37].

Energies 2021, 14, x FOR PEER REVIEW 7 of 14

80 80 100 H₂ CO CO₂ 80 60 60 60 40 40 40 20 20 20 Yield base) gas Yield (mol%, 0 0 0 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 Gas yield (wt%, cellulose base)

FigureEnergies 20215. Proportions, 14, 1842 of H2, CO and CO2 in the gaseous products as functions of the overall gas yield from the fast pyrolysis7 of 13 of Whatman cellulose.

GA was the other major byproduct, and Figure 6 summarizes the yields of GA obtained GAunder was various the other fast major pyrolysis byproduct, conditions and Figure compared6 summarizes with the the yields overall of GA gas obtained yields. It is under various fast pyrolysis conditions compared with the overall gas yields. It is apparent apparent that there were no correlations between the yields of both products and that the that there were no correlations between the yields of both products and that the GA yield GAremained yield remained relatively relatively constant as constant the various as th parameterse various wereparameters changed were while changed the gas yield while the gasvaried yield greatly. varied Accordingly,greatly. Accordingly, the formation the of formation GA cannot of be GA explained cannot by be the explained secondary by the secondarydegradation degradation of LG. of LG.

4 kW, 10 L/min, 5 s 4 kW, 5 s 5 L/min, 10 s 20 Gas 15

5 GA

Yield (wt%, cellulose base) 0 ABC1351 Power (kW) Set point N₂ flow (L/min)

FigureFigure 6. Yields 6. Yields of glycolaldehyde of glycolaldehyde (GA) (GA) and and overall overall gas yields (wt(wt % % based based on on amount amount of cellulose)of cellulose) from from the pyrolysis the pyrolysis of of WhatmanWhatman cellulose cellulose under under various various conditions. conditions. In the pyrolysis of cellulose, GA has been reported to form in a higher temperature rangeIn the than pyrolysis LG, and aof negative cellulose, correlation GA has has been been reported reported to between form in the a higher yields oftemperature these rangeproducts than [33LG,]. Therefore,and a negative there was correlation controversy has over been the reported GA formation between pathway, the yields although of these productsthe literature [33]. Therefore, [33–36] shows there a direct was formationcontrovers pathwayy over the from GA cellulose formation that competespathway, with although theLG literature formation. [33–36] The observationsshows a direct in formatio the presentn pathway study clearly from show cellulose the direct that competes pathway with LGof formation. GA formation. The GAobservations is believed in to the form presen duringt study the primary clearly cellulose show the pyrolysis direct pathway stage, of GAlikely formation. via retro-Aldol-type GA is believed fragmentation to form during reactions the primary of the reducing cellulose ends pyrolysis of cellulose stage, and likely viacarbohydrate retro-Aldol-type intermediates. fragmentation Such reactions reactions are of favorable the reducing because ends stable of cellulose six-membered and carbo- cyclic transition states are involved (Figure7). A similar mechanism has been proposed for hydrate intermediates. Such reactions are favorable because stable six-membered cyclic Energies 2021, 14, x FOR PEER REVIEWthe pyrolysis of cellobiose based on prior work using pyrolysis-GC/MS spectrometry8 of in 14 transitionconjunction states with are13 C-labeledinvolved cellobiose(Figure 7). [37 A]. similar mechanism has been proposed for the pyrolysis of cellobiose based on prior work using pyrolysis-GC/MS spectrometry in conjunction with 13C-labeled cellobiose [37].

OH +

Reducing end HO

FigureFigure 7. 7.Mechanism Mechanism for for the the formation formation of of GA GA from from the the reducing reducing ends ends of of cellulose cellulose and and other other interme- inter- diatemediate carbohydrates carbohydrates during during the fastthe fast pyrolysis pyrolysis of cellulose. of cellulose. 3.2. The Mechanism for Cellulose Pyrolysis by Infrared Irradiation 3.2. The Mechanism for Cellulose Pyrolysis by Infrared Irradiation Figure8 presents photographic images of the char residues recovered from Whatman Figure 8 presents photographic images of the char residues recovered from Whatman cellulose after applying IR irradiation at either 0.5 or 1.0 kW. A colorless unreacted part cellulose after applying IR irradiation at either 0.5 or 1.0 kW. A colorless unreacted part clearly remained intact on the specimen after the trial at 0.5 kW, suggesting that the thermal clearly remained intact on the specimen after the trial at 0.5 kW, suggesting that the thermal degradation of cellulose occurred nonuniformly under such conditions. The pyrolysis in degradation of cellulose occurred nonuniformly under such conditions. The pyrolysis in the upstream region of the nitrogen flow (the right side of the apparatus) was particularly slow, presumably due to the effect of the cool nitrogen. This tendency was also observed at the higher irradiation power of 1.0 kW after 5 s of irradiation, although the unreacted part was completely pyrolyzed by prolonging the irradiation time to 10 s. High selectivity for LG formation was also maintained under these conditions, as shown in Table 1.

Figure 8. Photographic images of cellulose chars obtained under various fast pyrolysis conditions.

Figure 9 provides an enlarged microscopic view of the char obtained at 1.0 kW after 5 s and demonstrates the formation of a very narrow (less than 0.5 mm) film-like carbonized zone. Cellulose generates film-like char when fast pyrolysis conditions are achieved because it rapidly converts to molten intermediates prior to the evaporation of LG and other products. On the basis of these images, it appears that the thermal degradation of the cellulose occurred within a very small area adjacent to the narrow carbonized zone, which was heated quickly owing to the efficient absorption of IR radiation (Figure 10). This process spread over the cellulose sheet as the irradiation time span increased. Gas formation would also occur in this small area in conjunction with the evaporation of LG as temperatures greater than 600 °C were achieved. Boutin et al. [11] reported the formation of short-lived liquid species while irradiating the surface of a cellulose pellet with a concentrated xenon lamp. However, they did not analyze LG and did not report the progression of the pyrolysis zone to the stage where cellulose was completely degraded.

Energies 2021, 14, x FOR PEER REVIEW 8 of 14

OH +

Reducing end HO

Figure 7. Mechanism for the formation of GA from the reducing ends of cellulose and other intermediate carbohydrates during the fast pyrolysis of cellulose.

3.2. The Mechanism for Cellulose Pyrolysis by Infrared Irradiation Figure 8 presents photographic images of the char residues recovered from Whatman Energies 2021, 14, 1842 cellulose after applying IR irradiation at either 0.5 or 1.0 kW. A colorless unreacted8 ofpart 13 clearly remained intact on the specimen after the trial at 0.5 kW, suggesting that the thermal degradation of cellulose occurred nonuniformly under such conditions. The pyrolysis in the theupstream upstream region region of the of nitrogen the nitrogen flow ﬂow (the (theright right side sideof the of apparatus) the apparatus) was particularly was particularly slow, slow,presumably presumably due to due the to effect the effect of the of cool the coolnitrogen. nitrogen. This Thistendency tendency was wasalso also observed observed at the at thehigher higher irradiation irradiation power power of 1. of0 kW 1.0 kWafter after 5 s of 5 sirradiation, of irradiation, although although the unreacted the unreacted part partwas wascompletely completely pyrolyzed pyrolyzed by prolonging by prolonging the irradi the irradiationation time time to 10 to s. 10 High s. High selectivity selectivity for forLG LGformation formation was was also also maintained maintained under under these these conditions, conditions, as shown as shown in Table in Table 1. 1.

FigureFigure 8.8. PhotographicPhotographic imagesimages ofof cellulosecellulose charschars obtainedobtained underunder variousvarious fastfast pyrolysispyrolysis conditions.conditions.

FigureFigure9 9 provides provides an an enlarged enlarged microscopic microscopic view view of theof the char char obtained obtained at 1.0 at kW1.0 kW after after5 s and5 s and demonstrates demonstrates the the formation formation of a of very a very narrow narrow (less (less than than 0.5 mm)0.5 mm) film-like film-like carbonized carbon- zone.ized zone. Cellulose Cellulose generates generates film-like film-like char char when when fast fast pyrolysis pyrolysis conditions conditions are are achievedachieved becausebecause itit rapidlyrapidly convertsconverts toto moltenmolten intermediatesintermediates priorprior toto thethe evaporationevaporation ofof LGLG andand otherother products.products. On the basisbasis ofof thesethese images,images, itit appearsappears thatthat thethe thermalthermal degradationdegradation ofof thethe cellulosecellulose occurredoccurred withinwithin aa veryvery smallsmall areaarea adjacentadjacent toto thethe narrownarrow carbonizedcarbonized zone,zone, whichwhich waswas heatedheated quicklyquickly owingowing to thethe efficientefficient absorptionabsorption of IR radiationradiation (Figure 1010).). ThisThis processprocess spreadspread overover thethe cellulosecellulose sheetsheet asas thethe irradiationirradiation timetime spanspan increased.increased. GasGas formationformation would also also occur occur in in this this small small area area in in conjunction conjunction with with the the evaporation evaporation of LG of LGas temperatures as temperatures greater greater than than 600 600°C were◦C were achieved. achieved. Boutin Boutin et al. et [11] al. [reported11] reported the thefor- Energies 2021, 14, x FOR PEER REVIEWformationmation of short-lived of short-lived liquid liquid species species while while irradiating irradiating the surface the surface of a ofcellulose a cellulose pellet pellet9 withof 14

Energies 2021, 14, x FOR PEER REVIEWwitha concentrated a concentrated xenon xenon lamp. lamp. However, However, they they did did not not analyze analyze LG LG and and did did not not report 9 thetheof 14 progressionprogression ofof thethe pyrolysispyrolysis zonezone toto thethe stagestage wherewhere cellulosecellulose waswas completelycompletely degraded. degraded.

FigureFigure 9. 9. MicrographicMicrographic image image of of the the dark dark carbonized carbonized area area of of the the cellulose cellulose char char obtained obtained at at 1.0 1.0 kW kW after 5 s. afterFigure 5 s. 9. Micrographic image of the dark carbonized area of the cellulose char obtained at 1.0 kW after 5 s.

：Carbonized layer ：Pyrolysis area ：Pyrolysis direction ：Carbonized layer ：Pyrolysis area ：Pyrolysis direction Figure 10. A diagram showing the process occurring during the fast pyrolysis of a Whatman cellu- loseFigureFigure sheet 10. 10. underA A diagram diagram IR irradiation. showing showing the the process process occurring occurring during during the the fast fast pyrolysis pyrolysis of a of Whatman a Whatman cellulose cellu- sheetlose sheet under under IR irradiation. IR irradiation. The above data obtained using the Whatman cellulose sheets provide useful insights into theThe pyrolysis above data in the obtained planar usingdirection. the Whatman To complement cellulose this sheets information, provide usefulpyrolysis insights be- haviorinto the in thepyrolysis thickness in the direction planar was direction. also ex Toamined complement using the this Avicel information, cellulose pyrolysis powder. Inbe- thesehavior trials, in the a 300 thickness mg portion direction of this was powder also ex wasamined transferred using the to Avicel a quartz cellulose boat and powder. irradi- In atedthese at trials,2.5 kW a for300 30 mg s under portion a 5of L/min this powder nitrogen was flow. transferred Similar experiments to a quartz usingboat and only irradi- 100 mgated samples at 2.5 kWwere for also 30 sconducted under a 5 toL/min understand nitrogen the flow. effect Similar of sample experiments loading. using Under only these 100 conditions,mg samples LG were was also obtained conducted in 44.3% to understand (with 100 mg the samples) effect of sampleand 48.5% loading. (300 mg)Under yields, these whichconditions, were comparable LG was obtained to those in 44.3%obtained (with from 100 the mg Whatman samples) celluland 48.5%ose sheets. (300 mg)The yields,film- likewhich chars were seen comparable in the images to inthose Figure obtained 11 indicate from thatthe Whatmanfast pyrolysis cellul wasose achieved sheets. The during film- theselike charsexperiments. seen in theIt should images also in Figurebe noted 11 thatindicate a high that LG fast yield pyrolysis was observed was achieved when usingduring eitherthese 100 experiments. or 300 mg Itsamples, should indicatingalso be noted that that the a sample high LG amount yield was did observednot affect whenthe ability using toeither perform 100 fastor 300 pyrolysis. mg samples, indicating that the sample amount did not affect the ability to perform fast pyrolysis.

Figure 11. Photographic images of quartz boats after applying IR radiation to (A) 300 mg and (B) 100Figure mg Avicel 11. Photographic cellulose samples images (1 of kW, quartz 5 L/min boat snitrogen after applying flow, 30 IR s). radiation to (A) 300 mg and (B) 100 mg Avicel cellulose samples (1 kW, 5 L/min nitrogen flow, 30 s). The temperature in the middle of the cellulose sample in the boat was monitored using Thea fine temperature thermocouple in thein trialsmiddle with of athe power cellulose level sampleof 1.0 kW, in the a 20 boat s irradiation was monitored time andusing a 5 aL/min fine thermocouple nitrogen flow. in The trials irradiation with a power time oflevel 20 sof was 1.0 insufficientkW, a 20 s irradiationto completely time pyrolyzeand a 5 theL/min cellulose, nitrogen meaning flow. Thethat irradiationthe sample timetemperature of 20 s wasduring insufficient cellulose todegradation completely couldpyrolyze be assessed. the cellulose, During meaning these trials,that the the sample bottom temperature of the sample during boat cellulose was covered degradation with aluminumcould be assessed.foil so that During the sample these only trials, received the bottom radiation of the from sample the top. boat Figure was covered12 provides with photographicaluminum foil images so that of the the sample boat after only the received trial, alongradiation with from a plot the of top. the Figure internal 12 providessample temperaturephotographic over images time. of It theis evidentboat after that the the tria surfacel, along of with the asample plot of was the darkenedinternal sample to a temperature over time. It is evident that the surface of the sample was darkened to a

Energies 2021, 14, x FOR PEER REVIEW 9 of 14

Figure 9. Micrographic image of the dark carbonized area of the cellulose char obtained at 1.0 kW after 5 s.

：Carbonized layer ：Pyrolysis area ：Pyrolysis direction

Figure 10. A diagram showing the process occurring during the fast pyrolysis of a Whatman cellulose sheet under IR irradiation.

Energies 2021, 14, 1842 9 of 13 The above data obtained using the Whatman cellulose sheets provide useful insights into the pyrolysis in the planar direction. To complement this information, pyrolysis behavior in the thickness directionThe above datawas obtainedalso examined using the using Whatman the celluloseAvicel cellulose sheets provide powder. useful insightsIn these trials, a 300 mginto portion the pyrolysis of this in powder the planar was direction. transferred To complement to a quartz this boat information, and irradi- pyrolysis ated at 2.5 kW for 30behavior s under in thea 5 thicknessL/min nitrog directionen wasflow. also Similar examined experiments using the Avicel using cellulose only powder. 100 In mg samples were alsothese conducted trials, a 300 mgto understand portion of this the powder effect was of transferred sample loading. to a quartz Under boat and these irradiated at 2.5 kW for 30 s under a 5 L/min nitrogen ﬂow. Similar experiments using only 100 mg conditions, LG wassamples obtained were in also44.3% conducted (with 100 to understand mg samples) the effect and of48.5% sample (300 loading. mg) yields, Under these which were comparableconditions, to those LG was obtained obtained fr inom 44.3% the (withWhatman 100 mg cellul samples)ose and sheets. 48.5% The (300 film- mg) yields, like chars seen in thewhich images were in comparable Figure 11 to thoseindicate obtained that fromfast thepyrolysis Whatman was cellulose achieved sheets. during The ﬁlm-like these experiments. Itchars should seen inalso the be images noted in Figure that a11 high indicate LG that yield fast was pyrolysis observed was achieved when duringusing these experiments. It should also be noted that a high LG yield was observed when using either either 100 or 300 mg samples, indicating that the sample amount did not affect the ability 100 or 300 mg samples, indicating that the sample amount did not affect the ability to to perform fast pyrolysis.perform fast pyrolysis.

Figure 11. Photographic images of quartz boats after applying IR radiation to (A) 300 mg and (B) 100 mg Avicel cellulose Figure 11. Photographic images of quartz boats after applying IR radiation to (A) 300 mg and (B) samples (1 kW, 5 L/min nitrogen flow, 30 s). 100 mg Avicel cellulose samples (1 kW, 5 L/min nitrogen flow, 30 s). The temperature in the middle of the cellulose sample in the boat was monitored The temperatureusing in athe fine middle thermocouple of the in cellulose trials with asample power levelin the of 1.0boat kW, was a 20 monitored s irradiation time using a fine thermocoupleand a 5 L/min in trials nitrogen with flow. a power The irradiation level of time1.0 kW, of 20 a s was20 s insufficient irradiation to completelytime pyrolyze the cellulose, meaning that the sample temperature during cellulose degradation and a 5 L/min nitrogencould flow. be assessed. The irradiation During these time trials, of the 20 bottom s was of insufficient the sample boat to completely was covered with pyrolyze the cellulose,aluminum meaning foil sothat that the the sample sample onlytemperature received radiation during from cellulose the top. degradation Figure 12 provides could be assessed. Duringphotographic these images trials, of the the boatbottom after of the the trial, sample along with boat a plotwas of covered the internal with sample aluminum foil so thattemperature the sample over only time. received It is evident radiation that the from surface the of top. the Figure sample 12 was provides darkened to a photographic imagesgreater of the extent boat than after the interior,the tria basedl, along on the with left halfa plot of the of samplethe internal wherethe sample surface has been removed. The temperature plot demonstrates a relatively slow rate of temperature temperature over time.rise inside It is theevident cellulose that layer. the The surface final temperature of the sample of 250 ◦ Cwas at 20darkened s was lower to than a the minimum value of 350 ◦C required for the efficient thermal degradation of cellulose. The brief pause in the temperature increase observed in the range of 100–150 ◦C is attributed to the volatilization of water. When the IR irradiation time was extended to 30 s, only a small

amount of film-like char remained in the boat, as can be seen in Figure 11, and LG was obtained in a 50.0% yield. These results indicate that the thermal degradation of cellulose proceeded non-uniformly from the surface that received the IR radiation, as illustrated in Figure 13. Figure 14 illustrates the heat and mass transfer processes and related events in the thin surface layer in which fast pyrolysis occurred during the IR irradiation of the Avicel cellulose. Cellulose tends to reflect IR, so the absorption of IR radiation by the cellulose was not sufficient to raise the sample temperature rapidly to above 400 ◦C. However, once solid carbonized materials were produced at the surface, this region underwent fast heating as it became more efficient at absorbing the radiation. Energies 2021, 14, x FOR PEER REVIEW 10 of 14 Energies 2021, 14, x FOR PEER REVIEW 10 of 14

greater extent than the interior, based on the left half of the sample where the surface has greater extent than the interior, based on the left half of the sample where the surface has been removed. The temperature plot demonstrates a relatively slow rate of temperature been removed. The temperature plot demonstrates a relatively slow rate of temperature rise inside the cellulose layer. The final temperature of 250 °C at 20 s was lower than the rise inside the cellulose layer. The final temperature of 250 °C at 20 s was lower than the minimum value of 350 °C required for the efficient thermal degradation of cellulose. The minimum value of 350 °C required for the efficient thermal degradation of cellulose. The brief pause in the temperature increase observed in the range of 100–150 °C is attributed brief pause in the temperature increase observed in the range of 100–150 °C is attributed to the volatilization of water. When the IR irradiation time was extended to 30 s, only a to the volatilization of water. When the IR irradiation time was extended to 30 s, only a small amount of film-like char remained in the boat, as can be seen in Figure 11, and LG small amount of film-like char remained in the boat, as can be seen in Figure 11, and LG was obtained in a 50.0% yield. These results indicate that the thermal degradation of cel- was obtained in a 50.0% yield. These results indicate that the thermal degradation of cel- Energies 2021, 14, 1842lulose proceeded non-uniformly from the surface that received the IR radiation, as illus- 10 of 13 lulose proceeded non-uniformly from the surface that received the IR radiation, as illustrated in Figure 13. trated in Figure 13.

Figure 12. (A) FigurePhotographic 12. (A) images Photographic of a quartz images sample of a quartzboat after sample IR irradiation boat after for IR irradiation20 s (less than for 20the s 30 (less s required than the for 30 s required for Figure 12. (A) Photographic images of a quartz sample boat after IR irradiation for 20 s (less than the 30 s required for complete pyrolysiscomplete of Avicel pyrolysis cellulose) of Avicel and cellulose) (B) the temperature and (B) the temperature inside the cellulose inside the over cellulose time over(1 kW, time 5 L/min (1 kW, 5nitrogen L/min nitrogen ﬂow). complete pyrolysis of Avicel cellulose) and (B) the temperature inside the cellulose over time (1 kW, 5 L/min nitrogen flow). flow).

： Carbonized layer ：Pyrolysis area ：Pyrolysis direction ： Carbonized layer ：Pyrolysis area ：Pyrolysis direction

EnergiesFigure 2021 13., 14A, xFigure diagramFOR PEER 13. showing AREVIEW diagram the showing progression the ofprogression the fast pyrolysis of the fast of Avicel pyrolysis cellulose of Avicel powder cellulose in a quartz powder boat in ina response11 of 14 Figure 13. A diagram showing the progression of the fast pyrolysis of Avicel cellulose powder in a to IR irradiation.quartz boat in response to IR irradiation. quartz boat in response to IR irradiation. Figure 14 illustrates the heat and mass transfer processes and related events in the Figure 14 illustrates the heat and mass transfer processes and related events in the thin surface layer in which fast pyrolysis occurred during the IR irradiation of the Avicel thin surface layer in which fast pyrolysis occurred during the IR irradiation of the Avicel cellulose. Cellulose tends to reflect IR, so the absorption of IR radiation by the cellulose cellulose. Cellulose tends to reflect IR, so the absorption of IR radiation by the cellulose was not sufficient to raise the sample temperature rapidly to above 400 °C. However, once was not sufficient to raise the sample temperature rapidly to above 400 °C. However, once solid carbonized materials were produced at the surface, this region underwent fast heat- solid carbonized materials were produced at the surface, this region underwent fast heating as it became more efficient at absorbing the radiation. ing as it became more efficient at absorbing the radiation.

Figure 14. Physicochemical processes occurring in the pyrolysis zone, in which the thermal degra- Figure 14. Physicochemical processes occurring in the pyrolysis zone, in which the thermal degrada- dation of cellulose proceeds in a very narrow area near the carbonized layer. tion of cellulose proceeds in a very narrow area near the carbonized layer.

TheThe resulting resulting heat heat energy energy was was transferred transferred to to the the adjacent adjacent cellulose cellulose to to promote promote rapid rapid thermalthermal degradation. degradation. TheThe temperature temperature of of the the degrading degrading cellulose cellulose was was rapidly rapidly raised raised to to aboveabove 400 400◦ C°C and and liquid liquid intermediates intermediates were were generated. generated. The The evaporation evaporation of of LG LG and and other other productsproducts is is an an endothermic endothermic process, process, while while the the formation formation of charof char is exothermicis exothermic [38 –[3840–].40]. As As a resulta result of theof the greater greater contribution contribution of the of formerthe former endothermic endothermic process, process, the fast the pyrolysis fast pyrolysis was overallwas overall endothermic endothermic and so and required so required continued continued heat input heat input from thefrom carbonized the carbonized layer. layer. The heatThe transferheat transfer to the interiorto the interior of the celluloseof the cellulose was relatively was relatively slow, based slow, on based the low on thermal the low thermal conductivity of cellulose (0.04 W/mK) [41,42], and so only a narrow region experienced high temperatures.

4. Conclusions The pyrolysis of cellulose using IR irradiation was studied, resulting in the following conclusions. 1. The sample position in the IR furnace, the IR power level and the nitrogen flow rate all affected the thermal degradation of cellulose by modifying the cellulose degradation temperature. 2. Under the optimum conditions, LG was obtained in a 52.7 wt % yield from Whatman cellulose sheets (infrared power: 1.0 kW, nitrogen flow: 5 L/min, irradiation time: 10 s). 3. The gas yield increased when the cellulose was overheated to above 600 °C as a result of the secondary degradation of LG. This effect could be monitored by tracking the formation of CO, although in situ temperature measurements were difficult due to the nonuniform progression of the cellulose thermal degradation. 4. GA was the other major product of cellulose degradation, and the yield of this compound was not correlated with the gas output, suggesting that it was not a secondary LG degradation product. GA was evidently produced during the primary cellulose pyrolysis stage via the retro-Aldol fragmentation of the reducing ends of cellulose and other intermediate carbohydrates. 5. The thermal degradation of cellulose occurred in a nonuniform manner in response to IR irradiation, with the formation of a narrow carbonization layer. This layer was rapidly heated by efficiently absorbing IR and, in turn, heated the adjacent cellulose. This process then propagated throughout the cellulose to maintain a high LG output rate. 6. The use of IR heating during the production of LG from cellulose offers several advantages compared with other fast pyrolysis methods based on heat conduction. The latter methods require the cellulose to be ground and heated quickly to maintain a

Energies 2021, 14, 1842 11 of 13

conductivity of cellulose (0.04 W/mK) [41,42], and so only a narrow region experienced high temperatures.

4. Conclusions The pyrolysis of cellulose using IR irradiation was studied, resulting in the following conclusions. 1. The sample position in the IR furnace, the IR power level and the nitrogen flow rate all affected the thermal degradation of cellulose by modifying the cellulose degradation temperature. 2. Under the optimum conditions, LG was obtained in a 52.7 wt % yield from Whatman cellulose sheets (infrared power: 1.0 kW, nitrogen flow: 5 L/min, irradiation time: 10 s). 3. The gas yield increased when the cellulose was overheated to above 600 ◦C as a result of the secondary degradation of LG. This effect could be monitored by tracking the formation of CO, although in situ temperature measurements were difficult due to the nonuniform progression of the cellulose thermal degradation. 4. GA was the other major product of cellulose degradation, and the yield of this compound was not correlated with the gas output, suggesting that it was not a secondary LG degradation product. GA was evidently produced during the primary cellulose pyrolysis stage via the retro-Aldol fragmentation of the reducing ends of cellulose and other intermediate carbohydrates. 5. The thermal degradation of cellulose occurred in a nonuniform manner in response to IR irradiation, with the formation of a narrow carbonization layer. This layer was rapidly heated by efficiently absorbing IR and, in turn, heated the adjacent cellulose. This process then propagated throughout the cellulose to maintain a high LG output rate. 6. The use of IR heating during the production of LG from cellulose offers several advantages compared with other fast pyrolysis methods based on heat conduction. The latter methods require the cellulose to be ground and heated quickly to maintain a high sample temperature, while the IR heating methods allow the use of any cellulose, regardless of size. Infrared power can also be easily controlled by changing the electric power. 7. These results give insights into the production of biochemicals and biofuels via LG and pyrolysis-based saccharification.

Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/en14071842/s1, Figure S1: GC chromatogram of gaseous products from Whatman cellulose after pyrolysis by infrared heating (4kW, nitrogen ﬂow: 1 L/min, 5 s). (A) column: MS5A, carrier gas: Ar, column temperature: 100 ◦C. (B) column: PoraPLOT Q, carrier gas: He, column temperature: 80 ◦C., Figure S2: 1H-NMR spectrum of soluble products in methanol from Whatman cellulose after pyrolysis by infrared heating (4kW, nitrogen ﬂow: 1 L/min, 5 s) Author Contributions: Conceptualization, T.N. and H.K.; methodology, T.N. and H.K.; formal analysis, T.N. and H.M.; investigation, T.N. and H.M.; resources, H.K.; writing—original draft preparation, T.N.; writing—review and editing, E.M. and H.K.; visualization, T.N. and E.M.; supervision, H.K.; project administration, T.N. and H.K.; funding acquisition, T.N. and H.K. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the JST Mirai Program, Japan (Grant no. JPMJMI20E3) and by the Japan Society for the Promotion of Science (JSPS) KAKENHI program (Grant nos. JP19H03019 and JP20J12367). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Data is contained within the article or Supplementary Material. Energies 2021, 14, 1842 12 of 13

Acknowledgments: The authors would like to express their sincere gratitude for the support of the Shin-Etsu Chemical Co., Ltd. We thank Michael D. Judge for editing a draft of this manuscript. Conﬂicts of Interest: The authors declare no conﬂict of interest.

Abbreviations LG levoglucosan GA glycolaldehyde

References 1. Jahirul, M.I.; Rasul, M.G.; Chowdhury, A.A.; Ashwath, N. Biofuels production through biomass pyrolysis-A technological review. Energies 2012, 5, 4952–5001. [CrossRef] 2. Uddin, M.N.; Techato, K.; Taweekun, J.; Rahman, M.M.; Rasul, M.G.; Mahlia, T.M.I.; Ashrafur, S.M. An overview of recent developments in biomass pyrolysis technologies. Energies 2018, 11, 3115. [CrossRef] 3. Bridgwater, A.V. Review of fast pyrolysis of biomass and product upgrading. Biomass Bioenergy 2012, 38, 68–94. [CrossRef] 4. Kersten, S.; Garcia-Perez, M. Recent developments in fast pyrolysis of ligno-cellulosic materials. Curr. Opin. Biotechnol. 2013, 24, 414–420. [CrossRef] 5. Sharifzadeh, M.; Sadeqzadeh, M.; Guo, M.; Borhani, T.N.; Murthy Konda, N.V.S.N.; Garcia, M.C.; Wang, L.; Hallett, J.; Shah, N. The multi-scale challenges of biomass fast pyrolysis and bio-oil upgrading: Review of the state of art and future research directions. Prog. Energy Combust. Sci. 2019, 71, 1–80. [CrossRef] 6. Venderbosch, R.; Prins, W. Fast pyrolysis technology development. Biofuels Bioprod. Biorefin. 2010, 4, 178–208. [CrossRef] 7. Perkins, G.; Bhaskar, T.; Konarova, M. Process development status of fast pyrolysis technologies for the manufacture of renewable transport fuels from biomass. Renew Sustain. Energy Rev. 2018, 90, 292–315. [CrossRef] 8. Shafizadeh, F.; Furneaux, R.H.; Cochran, T.G.; Scholl, J.P.; Sakai, Y. Production of levoglucosan and glucose from pyrolysis of cellulosic materials. J. Appl. Polym. Sci. 1979, 23, 3525–3539. [CrossRef] 9. Kwon, G.J.; Kim, D.-Y.; Kimura, S.; Kuga, S. Rapid-cooling, continuous-feed pyrolyzer for biomass processing: Preparation of levoglucosan from cellulose and starch. J. Anal. Appl. Pyrolysis 2007, 80, 1–5. [CrossRef] 10. Itabaiana Junior, I.; Avelar Do Nascimento, M.; De Souza, R.O.M.A.; Dufour, A.; Wojcieszak, R. Levoglucosan: A promising platform molecule? Green Chem. 2020, 22, 5859–5880. [CrossRef] 11. Boutin, O.; Ferrer, M.; Lédé, J. Flash pyrolysis of cellulose pellets submitted to a concentrated radiation: Experiments and modelling. Chem. Eng. Sci. 2002, 57, 15–25. [CrossRef] 12. Boutin, O.; Ferrer, M.; Lédé, J. Radiant flash pyrolysis of cellulose—Evidence for the formation of short life time intermediate liquid species. J. Anal. Appl. Pyrolysis 1998, 47, 13–31. [CrossRef] 13. Lédé, J.; Blanchard, F.; Boutin, O. Radiant flash pyrolysis of cellulose pellets: Products and mechanisms involved in transient and steady state conditions. Fuel 2002, 81, 1269–1279. [CrossRef] 14. Shoji, T.; Kawamoto, H.; Saka, S. Boiling point of levoglucosan and devolatilization temperatures in cellulose pyrolysis measured at different heating area temperatures. J. Anal. Appl. Pyrolysis 2014, 109, 185–195. [CrossRef] 15. Pictet, A. Sur la transformation de la lévoglucosane en dextrine. Helv. Chim. Acta 1918, 1, 226–230. [CrossRef] 16. Fukutome, A.; Kawamoto, H.; Saka, S. Processes forming gas, tar, and coke in cellulose gasification from gas-phase reactions of levoglucosan as intermediate. ChemSusChem 2015, 8, 2240–2249. [CrossRef] 17. Hosoya, T.; Kawamoto, H.; Saka, S. Different pyrolytic pathways of levoglucosan in vapor- and liquid/solid-phases. J. Anal. Appl. Pyrolysis 2008, 83, 64–70. [CrossRef] 18. Kawamoto, H.; Murayama, M.; Saka, S. Pyrolysis behavior of levoglucosan as an intermediate in cellulose pyrolysis: Polymeriza- tion into polysaccharide as a key reaction to carbonized product formation. J. Wood Sci. 2003, 49, 469–473. [CrossRef] 19. Kawamoto, H.; Hosoya, T.; Ueno, Y.; Shoji, T.; Saka, S. Thermal stabilization and decomposition of simple glycosides in the presence of aromatic substances in closed ampoules: The role of OH π hydrogen bonding. J. Anal. Appl. Pyrolysis 2014, 109, 41–46. [CrossRef] 20. Kawamoto, H.; Ueno, Y.; Saka, S. Thermal reactivities of non-reducing sugars in polyether—Role of intermolecular hydrogen bonding in pyrolysis. J. Anal. Appl. Pyrolysis 2013, 103, 287–292. [CrossRef] 21. Suzuki, J.; Azuma, J.; Koshijima, T.; Okamura, K.; Okamoto, H. Characterization of mono- and oligosaccharides produed by CO2 laser irradiation on cellulose. Chem. Lett. 1983, 12, 481–484. [CrossRef] 22. Dauenhauer, P.J.; Colby, J.L.; Balonek, C.M.; Suszynski, W.J.; Schmidt, L.D. Reactive boiling of cellulose for integrated catalysis through an intermediate liquid. Green Chem. 2009, 11, 1555–1561. [CrossRef] 23. Teixeira, A.R.; Mooney, K.G.; Kruger, J.S.; Williams, C.L.; Suszynski, W.J.; Schmidt, L.D.; Schmidt, D.P.; Dauenhauer, P.J. Aerosol generation by reactive boiling ejection of molten cellulose. Energy Environ. Sci. 2011, 4, 4306–4321. [CrossRef] 24. Lédé, J.; Li, H.Z.; Villermaux, J.; Martin, H. Fusion-like behaviour of wood pyrolysis. J. Anal. Appl. Pyrolysis 1987, 10, 291–308. [CrossRef] 25. Luo, G.; Chandler, D.S.; Anjos, L.C.A.; Eng, R.J.; Jia, P.; Resende, F.L.P. Pyrolysis of whole wood chips and rods in a novel ablative reactor. Fuel 2017, 194, 229–238. [CrossRef] Energies 2021, 14, 1842 13 of 13

26. Weldekidan, H.; Strezov, V.; Town, G. Review of solar energy for biofuel extraction. Renew Sustain. Energy Rev. 2018, 88, 184–192. [CrossRef] 27. Nordin, S.B.; Nyren, J.O.; Back, E.L. An indication of molten cellulose produced in a laser beam. Text. Res. J. 1974, 44, 152–154. [CrossRef] 28. Schroeter, J.; Felix, F. Melting cellulose. Cellulose 2005, 12, 159–165. [CrossRef] 29. Zeng, K.; Gauthier, D.; Soria, J.; Mazza, G.; Flamant, G. Solar pyrolysis of carbonaceous feedstocks: A review. Sol. Energy 2017, 156, 73–92. [CrossRef] 30. Hopkins, M.W.; DeJenga, C.; Antal, M.J. The flash pyrolysis of cellulosic materials using concentrated visible light. Sol. Energy 1984, 32, 547–551. [CrossRef] 31. Kwon, G.J.; Kuga, S.; Hori, K.; Yatagai, M.; Ando, K.; Hattori, N. Saccharification of cellulose by dry pyrolysis. J. Wood Sci. 2006, 52, 461–465. [CrossRef] 32. Westerhof, R.J.M.; Oudenhoven, S.R.G.; Marathe, P.S.; Engelen, M.; Garcia-Perez, M.; Wang, Z.; Kersten, S.R.A. The interplay between chemistry and heat/mass transfer during the fast pyrolysis of cellulose. React. Chem. Eng. 2016, 1, 555–566. [CrossRef] 33. Piskorz, J.; Radlein, D.S.A.G.; Scott, D.S.; Czernik, S. Pretreatment of wood and cellulose for production of sugars by fast pyrolysis. J. Anal. Appl. Pyrolysis 1989, 16, 127–142. [CrossRef] 34. Piskorz, J.; Radlein, D.; Scott, D.S. On the mechanism of the rapid pyrolysis of cellulose. J. Anal. Appl. Pyrolysis 1986, 9, 121–137. [CrossRef] 35. Richards, G.N. Glycolaldehyde from pyrolysis of cellulose. J. Anal. Appl. Pyrolysis 1987, 10, 251–255. [CrossRef] 36. Banyasz, J.L.; Li, S.; Lyons-Hart, J.L.; Shafer, K.H. Cellulose pyrolysis: The kinetics of hydroxyacetaldehyde evolution. J. Anal. Appl. Pyrolysis 2001, 57, 223–248. [CrossRef] 37. Degenstein, J.C.; Murria, P.; Easton, M.; Sheng, H.; Hurt, M.; Dow, A.R.; Gao, J.; Nash, J.J.; Agrawal, R.; Delgass, W.N.; et al. Fast pyrolysis of 13C-labeled cellobioses: Gaining insights into the mechanisms of fast pyrolysis of carbohydrates. J. Org. Chem. 2015, 80, 1909–1914. [CrossRef][PubMed] 38. Milosavljevic, I.; Oja, V.; Suuberg, E.M. Thermal effects in cellulose pyrolysis: Relationship to char formation processes. Ind. Eng. Chem. Res. 1996, 35, 653–662. [CrossRef] 39. Mok, W.S.L.; Antal, M.J. Effects of pressure on biomass pyrolysis. II. Heats of reaction of cellulose pyrolysis. Thermochim. Acta 1983, 68, 165–186. [CrossRef] 40. Cho, J.; Davis, J.M.; Huber, G.W. The intrinsic kinetics and heats of reactions for cellulose pyrolysis and char formation. ChemSusChem 2010, 3, 1162–1165. [CrossRef] 41. Di Blasi, C. Numerical simulation of cellulose pyrolysis. Biomass Bioenergy 1994, 7, 87–98. [CrossRef] 42. Majumdar, A.; Mukhopadhyay, S.; Yadav, R. Thermal properties of knitted fabrics made from cotton and regenerated bamboo cellulosic fibres. Int. J. Therm. Sci. 2010, 49, 2042–2048. [CrossRef]