Pyrolytic Reactions of Lignin within Naturally Occurring Plant

A hierarchy of structures was described, with elementary “lignocellulose” macromolecular structures making up wood cells, which combine to make up...
0 downloads 0 Views 430KB Size
Download PDF
Subscriber access provided by ONDOKUZ MAYIS UNIVERSITESI

Article

Pyrolytic reactions of lignin within naturally occurring plant matrices: Challenges in Biomass Pyrolysis Modeling due to Synergistic Effects Anthe George, Trevor J. Morgan, and Rafael Kandiyoti Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/ef501459c • Publication Date (Web): 02 Sep 2014 Downloaded from http://pubs.acs.org on November 10, 2014

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

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

Page 1 of 19

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

Energy & Fuels

Pyrolytic reactions of lignin within naturally occurring plant matrices: Challenges in Biomass Pyrolysis Modeling due to Synergistic Effects# Anthe George*ǂ, Trevor J. Morgan**, and Rafael Kandiyoti*** *

Combustion Research Facility, Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, USA ** Hawaii Natural Energy Institute, University of Hawaii at Manoa, 1680 East West Road, POST 109, Honolulu, Hawaii 96822, USA *** Chemical Engineering Department, Imperial College London, London SW7 2AZ, UK ǂ

Corresponding author (e-mail address: [email protected])

Abstract Systematically larger char yields were observed from the pyrolysis of chemically isolated lignins, compared to expected yields from the pyrolysis of lignins embedded in plant material. Naturally occurring lignins are known to be intermeshed with other plant constituents within the composite matrices of lignocellulosic biomass. An attempt was made to simulate their behaviour by pyrolyzing pellets prepared from mixtures of lignin and cellulose powders. However, the results gave char yield trends that did not conform to trends observed when pyrolyzing plant derived biomass. These findings are interpreted in terms of entirely different reaction pathways operating when lignins are pyrolyzed within naturally occurring biomass, compared to “pure” lignins or composite particles made from mixtures of fine powders. It appears char yield trends from the pyrolysis of lignocellulosic biomass are closely linked to the detailed morphology, as well as the chemical makeup, of the highly oxygenated plant derived material within which the lignin components of plants are embedded. The observed sensitivity of reaction pathways to plant specific structural (morphological) features poses added challenges in formulating realistic ab-initio mathematical models for predicting the pyrolysis chemistry of lignocellulosic biomass. #

Some of the data in this manuscript was presented in a recent review article.1 However, the contents of this manuscript have not hitherto been presented as a distinct and coherent paper examining synergistic effects governing the pyrolysis of biomass.

1 ACS Paragon Plus Environment

Energy & Fuels

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

Page 2 of 19

Introduction Most forestry and plant derived agricultural wastes contain high proportions of cellulose (~35 to 50 %, d.b.), hemicelluloses (~20 to 35 %, d.b.) and lignins (~10 to 25 %, d.b.).2,3,4 Some North European oak woods reportedly contain up to 35 % lignins.5 Lignocellulosic biomass normally also contains minor amounts of extractables, such as resins, oils, fats and waxes.3 Apart from notable exceptions such as wheat straw and rice husks, mineral matter contents in most plant derived material is of the order of 1 to 2 %.6,7 When the linear polymer cellulose is pyrolyzed, covalent bonds holding together the “chain” of single rings are cleaved releasing the single-ring compound laevoglucosan (1,6-anhydro-β, D-glucopyranose) as the most abundant primary product. Laevoglucosan is thermally sensitive and may degrade readily if exposed to temperatures above 250-300 °C, or contacted with heated surfaces for any length of time. In some of their pioneering work, Shafizadeh and co-workers pyrolyzed various biomass materials by pushing small metallic crucibles filled with sample into a preheated tubular furnace.8 When they pyrolyzed pure cellulose, they observed laevoglucosan yields approaching 50 %. This was a surprisingly high yield, given the opportunities for secondary thermal degradation and recombination reactions within that particular reactor configuration. When the thermal degradation of volatile products is, at least, partially suppressed higher yields of laevoglucosan may be obtained. Volatiles condensed and recovered during cellulose pyrolysis experiments in a purpose designed fluidized-bed reactor operating between 400 – 500 °C converted ~ 85 % of the original cellulose to laevoglucosan by mass.9 However, when Shafizadeh and co-workers pyrolyzed wood powders that contained between 35-50 % cellulose under similar conditions, they found less than 3 % of the sample mass in the form of laevoglucosan.10 Shafidazeh and co-workers also found higher yields of levoglucosan when the wood powders had undergone a mild acid washing compared to untreated wood powder.11 The effect was attributed to naturally occurring inorganics present in the wood samples, particularly to potassium and magnesium species. The low amounts of laevoglucosan in the tars/oils from the pyrolysis of naturally occurring lignocellulosic biomass has since been widely reported upon, despite cellulose being present in proportions as high as 50 % in the samples. To cite but one example, Fraga and co-workers pyrolyzed a sample of sugar-cane bagasse and found virtually no laevoglucosan in the tar/oil.12 A wire-mesh reactor was used for these experiments using small particle sizes (106152 µm dia.), under conditions where extra-particle secondary reactions of volatiles are minimized and with rapid quenching of the products of pyrolysis.12,13 During similar experiments with samples of silver birch wood, product tars/oils presented only trace concentrations of laevoglucosan. Taken together with results from the work by Shafizadeh and co-workers cited above, these data provided credible early evidence that component parts of lignocellulosic biomass do not pyrolyze independently. Meanwhile, the high yields of tars/oils produced during the rapid pyrolysis of lignocellulosic biomass open a wide field of possibilities for making transportation fuels from renewable feedstocks. In this paper, we will not discuss the extensive literature on the enduring challenges facing the conversion of these “liquids” to transport fuels.1,14,15,16,17,18 It nevertheless seems clear that a more detailed and realistic understanding of the pyrolytic 2 ACS Paragon Plus Environment

Page 3 of 19

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

Energy & Fuels

process would be helpful in optimizing the quality of “liquids” produced by biomass pyrolysis for eventual upgrading. Recently, excellent progress has been reported in the ab-initio modelling of the decomposition pathways of pure cellulose, based on the stated premise that “…understanding cellulose pyrolysis chemistry is…crucial for developing efficient biofuel production technologies.”19 Further progress was deemed to be possible through ab-initio mathematical models, simulating chemical reactions during the pyrolysis of naturally occurring biomass, based on the assumption that this could be done “…analogously from the problems of cellulose…”20 However, unlike other components of lignocellulosic biomass, the molecular structures of celluloses are well-defined, which helps in modelling. Furthermore, the cited work of Shafizadeh and co-workers10 on its own (going back to the 1960ies) should have been sufficient to show that the composite nature of plant derived biomass fundamentally alters pyrolytic reaction pathways, from those predominating during the pyrolysis of pure cellulose. In this context, the extent to which synergistic effects dominate the course of pyrolytic reactions would have a direct bearing on whether and how particular chemical reaction pathways can be identified and fitted into ab-initio mathematical models. This paper aims to develop a clearer view of the nature of synergistic interactions observed during the pyrolysis of lignins embedded within plant derived biomass, and the effect of these synergistic interactions on overall product distributions. To this end, char yields associated with the pyrolysis of isolated (“pure”) lignin samples have been compared with estimated char yields expected from the lignin component of composite, plant derived samples. In order to provide a wider field of vision, results from the literature have been collated alongside “in house” data. These findings have been compared with data from the pyrolysis of synthesized pellets made by compressing mixtures of fine lignin and cellulose powders, of variable composition. It is important to note that the approach outlined above not ideal due to the unavoidable differences in the chemical composition of native lignin embedded within plant material and that of chemically isolated lignins that will undoubtedly influence their respective pyrolysis behaviors. Nonetheless, it will be shown that it is possible to draw qualitative trends from the data that provides valuable insights into the synergistic effects that occur during pyrolysis. Finally, the level of chemical and morphological detail necessary for tracking the nature and magnitude of experimentally observed synergistic effects has been considered in relation to the information required for developing realistic ab-initio models of biomass pyrolysis. As an aside, we may recall that there are numerous commercially available mathematical models that claim the capability of predicting the behaviour of some solid fuels during pyrolysis. In what follows, we will try to examine the limits of what can be deduced about the pyrolytic reactions of biomass, using results from purpose designed experiments. It seems useful to point out that such commercial models usually require data from experiments described in the present paper and/or from similar research. Mathematical models that serve to fit – and eventually estimate – pyrolysis data, often through the use of adjustable parameters, will not be included in the present discussion.

Experimental 3 ACS Paragon Plus Environment

Energy & Fuels

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

Page 4 of 19

Samples: The cellulose sample used in these experiments was a pure, binder free microgranular powder for thin layer chromatography (Whatman, UK; Part No. 4061-050). Isolated lignin characteristics vary with extraction method and original plant material. The isolated lignin sample chosen for this study was a sample of softwood derived Kraft lignin by Holmen AB (Sweden), used as received. To prepare samples for the wire-mesh pyrolysis reactor (see below), pellets were made from mixtures of the sample powders, ~ 1 cm in diameter and a couple of millimetres thick. These pellets were then crushed and the 106-152 µm size fraction sieved for use in the pyrolysis experiments. Four composite pellets were prepared by mixing the fine lignin and cellulose powders in proportions that ranged from 100 % lignin to 100 % cellulose in “25 %-steps”. Wire-mesh pyrolysis reactor: An atmospheric pressure wire-mesh pyrolysis reactor was used for this study, as described elsewhere.13 Briefly, a stainless steel wire-mesh sample holder was held taut between two electrodes. The mesh also acts as a resistance heater and samples were heated at variable heating rates to the pre-set experimental temperature. The reactor was configured to stream a flow of sweep gas (helium at 0.1 m/s) through the sample holder, to minimize interactions between evolving volatiles and solid surfaces (heated mesh and pyrolyzing particles), in an attempt to reduce extra-particle secondary reactions of the volatiles. The experimental procedure and tar/oil trap have been described in the original publication.13 The particle size range selected for these experiments was 106−152 µm. Experiments described below have been carried out at two heating rates, 1 and 1,000 °C s-1, to a peak temperature ranging from 400 to 900 °C, with 30 s hold time at the peak temperature. Total volatiles were determined to a repeatability of ± 1 % and tar yields to ± 1.5-2 %.

Results & Discussion Pyrolysis of a Kraft lignin sample: Table 1 presents the total volatile, tar/oil and char results from pyrolysis of Kraft lignin (Holmen AB, Sweden) under atmospheric pressure. The experiments were carried out at two heating rates (slow 1 °C s-1, and fast 1,000 °C s-1) in the atmospheric pressure wire-mesh reactor described above.

Table 1. Total volatile, tar/oil and char yields from the atmospheric pressure pyrolysis of Kraft lignin, determined using slow (1 °C s-1) and fast (1,000 °C s-1) heating rates in the wiremesh reactor, with 30 s holding at peak temperature. Helium flowing through the sampleholder was 0.1 m s-1 in all experiments.21 Total volatiles % daf basis* Temperature °C

1°C s-1 400 44.7 600 57.2 900 58.7 * daf: dry ash free

1,000°C s-1 50.6 65.5 68.3

Tar/oil yield % daf basis* Heating rate 1°C s-1 1,000°C s-1 37.6 40.0 42.0 44.9 43.2 45.0

Char residue % daf basis* 1°C s-1 55.3 42.8 41.3

1,000°C s-1 49.4 34.5 31.7

4 ACS Paragon Plus Environment

Page 5 of 19

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

Energy & Fuels

As expected, the yields of total volatiles increased and the char residues diminished with increasing peak experimental temperature. As commonly observed,12,13 tar/oil yields tended to increase with temperature to a peak value and then to hold steady with further rising experimental temperature. In experimental systems where the reactor configuration allows secondary cracking and/or recombination reactions of evolved volatiles, tar/oil yields would be expected to decrease with increasing temperature, after rising to a peak value.22,23 Another clearly discernible trend in Table 1 is the (albeit limited) increase in tar/oil as well as total volatile yields when a faster heating rate was used. At fast heating rates (1,000 °C s-1 or faster), pyrolyzing solid wood or lignin particles have also been observed to display transient plastic phase behaviour, akin to the plastic behaviour observed when heating some coals.21 For biomass materials, the effect has been more recently referred to by Mettler et.al.20 In effect, increases in total volatile and tar/oil yields (e.g. Table 1) as well as softening (plastic) behaviour at high heating rates are common to a variety of coal24,25 and biomass20,21 derived samples. Industrial “coking” processes entirely rely on the plastic behaviour of many middle-to-high coals upon heating;26 such coals display a spread of abilities for coking (plastic) behaviour. Heating rates greater than 500 °C s-1 have been demonstrated to enhance the softening (plastic) behaviour of some low rank coals that do not ordinarily soften, when heated slowly (say at 1 °C s-1).24,25,27,28 The transient plastic deformation of low-rank samples such as lignin21 (estimated oxygen content: 37-38%) and of intermediate maturity samples, such as lignite particles29 (oxygen content: 26.5 %) are commonly observed, when the pyrolysis takes place in high-heating rate environments. Boutin et al.30 reported evidence for a short lifetime liquid species that is formed during the pyrolysis of pure cellulose. In their study, radiant flash pyrolysis was performed using a 5 kW xenon lamp with flash durations of ~ 1 s, with sample temperatures estimated to reach about 500 °C. The short lived liquid phase was found to have a different composition than typical fast pyrolysis oils, containing far fewer species that could be identified by GC-MS or HPLC. Instead, the transient liquid phase from cellulose mainly contained material with low degrees of polymerization, produced from splitting cellulosic polymers. Virtually no char residue was produced. As discussed elsewhere,1 the observed transient fluid (plastic) phase that is produced during biomass pyrolysis is probably related to a transient local abundance of native hydrogen, as also seen for some pyrolyzing coals. It appears furthermore, that many pyrolysis studies on biomass and coal are undertaken with little consideration of the many shared aspects of experimental design or the similar problems encountered in the interpretation of experimental results. When it comes to fundamental research, it seems clear that doing away with the compartmentalization of coal and biomass pyrolysis as separate spheres may give rise to useful synergies. Comparison of isolated lignin data with results from lignocellulosic biomass: Tables 2 and 3 present total volatile, tar/oil and char yields from the atmospheric pressure pyrolysis of silver birch wood and sugar cane bagasse samples. These experiments were conducted in the wire-mesh reactor described above. As already signalled in the Experimental section, the particle size range selected for these experiments was 106−152 µm.

5 ACS Paragon Plus Environment

Energy & Fuels

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

Page 6 of 19

Table 2. Total volatile, tar/oil and char yields from the atmospheric pressure pyrolysis of silver birch wood, determined using slow (1 °C s-1) and fast (1,000 °C s-1) heating rates in the wire-mesh reactor, with 30 s holding at peak temperature. Helium flowing through the sample-holder at 0.1 m s-1 was used as ambient gas in all experiments.21,31 (Reproduced from Ref. 31 with permission; Copyright 1991 Elsevier.) Total volatiles % daf basis* Temperature °C 400 500 700

1°C s 77 89 93

-1

Tar/oil yield % daf basis* -1

1,000°C s 89 96 99

900 93 * daf: dry ash free

99

Heating rate -1 -1 1°C s 1,000°C s 43 56 49 58 54 57 52

57

Char residue % daf basis* 1°C s 23 11 7

-1

7

-1

1,000°C s 11 4 1 1

Table 3. Total volatile, tar/oil and char yields from the atmospheric pressure pyrolysis of sugar cane bagasse, determined using slow (1 °C s-1) and fast (1,000 °C s-1) heating rates in the wire-mesh reactor, with 30 s holding at peak temperature. Helium flowing through the sample-holder at 0.1 m s-1 was used as ambient gas in all experiments.21,31 (Reproduced from Ref. 31, with permission; Copyright 1991 Elsevier.) Total volatiles % daf basis* Temperature °C 400 500 600 700 900 * daf: dry ash free

-1

1°C s 74.3 86.1 89.1 87.5 88.8

Tar/oil yield % daf basis* -1

1,000°C s 88.3 93.7 96.1 96.9 96.9

Heating rate -1 -1 1°C s 1,000°C s 37.0 49.2 42.4 56.4 45.4 54.4 45.6 53.7 45.4 53.7

Char residue % daf basis* -1

1°C s 25.7 13.9 11.1 12.5 11.2

1,000°C s 11.7 6.3 3.9 3.1 3.4

-1

The data in Tables 2 and 3 showed trends that were similar to lignin pyrolysis data presented in Table 1. In all cases, both total volatile and tar/oil yields increased with faster heating. It may be noted, however, that, when using fast heating with rapid and effective volatile removal from the reaction zone, small particles of wood (106 – 152 µm) gave remarkably low char yields: about 1 % residue between 600 and 900 °C. The analogous char yield from sugar cane bagasse (Table 3) was only marginally higher. However, the Kraft lignin sample (Table 1) gave much greater (> 30 %) amounts of char under equivalent experimental conditions, compared to samples from silver birch or sugar cane bagasse. The sugar cane bagasse and silver birch samples contain approximately 21 % and 27 % lignin, respectively. If the lignin, cellulose and other components of biomass were to pyrolyze independently, higher char yields would be expected from the sugar cane bagasse and silver birch than is observed experimentally – even assuming zero char yields from all other components of the samples. Conversely, the data in Table 1 showed much higher char 6 ACS Paragon Plus Environment

Page 7 of 19

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

Energy & Fuels

yields for the isolated lignin samples than would have been predicted, based on the amounts of char produced during the pyrolysis of the lignocellulosic biomasses and their estimated lignin contents. In other words, the lignin in the plant derived biomass should have produced more char, had the pyrolysis of biomass components been taking place independently, or nearly so. On the basis of these experiments, we appear to face a “char deficit” in experiments involving the pyrolysis of naturally occurring biomass samples, relative to experiments with chemically isolated lignin. A survey of lignin pyrolysis experiments: We naturally expect the chemical compositions and structures of chemically isolated lignins to change from one plant species to another and to vary equally widely as a function of the isolation method. Furthermore, there is no evidence to suggest that the structures and compositions of lignins that are isolated via chemical methods are similar to those of the lignins within the parent biomass. On the contrary, evidence has been reported for distinct differences in the chemical structure and composition of isolated lignins compared to lignin native within plant material.32 In attempting to test the “char deficit” hypothesis outlined above, variations in lignin structures may (at least to some extent) be evened out by surveying results from as broad a range of pyrolysis experiments as possible, using isolated (“pure”) lignin samples prepared using a wide selection of methods and from a variety of starting materials (i.e. plants). By using this admittedly inexact approach, we will seek to sketch some essentially qualitative conclusions. Table 4 shows a set of results from many different types of pyrolysis experiments, carried out using isolated lignin samples prepared via different chemical isolation methods and from a number of distinct plant species. In this context, a more robust comparison would require data from experiments performed using reactor configurations, which can provide information more closely related to the fundamental pyrolytic behaviour of the materials. When experiments are designed for characterizing the fundamental thermochemical behaviour of solid fuels, it is necessary to decouple sample specific behaviour from aspects of the results relating to sample and reactor configuration.1,33 It then becomes necessary to evaluate the behaviour of sample particles in isolation, to avoid intra-particle contact between volatiles and solids, as well as to prevent the degradation of volatile products. For example, both the freeboard height of a fluidized-bed pyrolysis reactor and the temperature distribution within it would tend to affect extents of tar cracking in the freeboard and alter tar yields. In addition, it is necessary to grind samples as finely as is practicable, in order to minimize intraparticle reactions of tar precursors. Particle diameters greater than ~150 µm tend to increase char yields and to give misleading results. Meanwhile, it is problematic to work with particles smaller than 70−80 µm because of static electrical effects. As in all design problems it is necessary to arrive at compromise solutions. In this context, we note that thermogravimetric (TG) balances are not constructed in a way to minimise losses of tar due to contact between volatiles and pyrolyzing solids, or to reduce secondary reactions of the volatiles that are generated during pyrolysis.35 As explained elsewhere,1,33 data from purpose-built pyrolysis reactors such as the wire-mesh and fluidized-bed types (as detailed elsewhere34) tend to provide data that is closer to the fundamental behaviour of sample materials than most other reactor designs. Nonetheless, 7 ACS Paragon Plus Environment

Energy & Fuels

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

Page 8 of 19

results from numerous other experimental configurations are presented in Table 4, in order to widen the field of available lignin samples and pyrolysis experiments. It is apparent that the allure of TG balances for studying the pyrolytic behaviour of biomass is as irrefutable as the unavoidable drawbacks of this type of instrument for performing pyrolysis experiments, as explained in some detail elsewhere.35

Table 4. Selection of pyrolysis char yields, from lignins prepared with diverse methods and pyrolyzed in diverse types of apparatus. (Adapted from Ref. 1 with additional data from Ref. 44, reproduced with permission; Copyright 2014 and 2013, Am. Chem. Soc.)

8 ACS Paragon Plus Environment

Page 9 of 19

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

Energy & Fuels

Authors (year)

Lignin preparation method

Pyrolysis method

Iatridis & Gavalas36 (1979)

Kraft lignin from Douglas fir

Early version of “captive sample” technique (wire-mesh reactor)

Chan and Krieger37 (1981)

pinewood Kraft lignin

volume heating by 1.5×1.5 dielectric-loss cm micro-wave pellet heating

Nunn et al. (1985)

Milled wood lignin from sweet gum hardwood

“heated grid” (wire-mesh) reactor

Caballero et al.39 (1996)

eucalyptus wood Kraft lignin

Pyroprobe instrument

38

Alcell and kraft lignins

Wang et al.41 (2009)

MWLb from Manchurian ash and Mongolian pine

De Wild et al.42 (2009)

1.Alcell organosolv process from mixture of hardwoods. 2. “GRANIT” Precipitation after pulping non-woody plants

Trinh et.al.44 2013 de Wild et.al.44,45 de Wild et.al.44,45 de Wild et.al.44,45 de Wild et.al.44,45

200 mg

≤100 µm thick flakes

powder

Thermogravimetric 10 mg balance (TGA) powder

Ferdous et al.40 (2002)

Beis et al.43 (2010)

Particle Size

Heating Rate 200-400 (approx.) °C s-1 20°C s-1 (est.)

1,000 °C s-1 nominal 20,000 °C s-1 5-15 °C min-1 5-15 °C min-1

Fixed bed reactor

Thermogravimetric Not available balance (TGA)

1°C s-1

Char yield data (Temp °C, Char %)

400 500

~76-78 ~ 56

600 750

~ 43 ~ 35

650-750 (est.)

33 %

307

96.9

527

50

1077

14.5

450

44a

700 900 800 Alcell Kraft

35a 33a 35 43

800 Alcell Kraft

38-42a 45-50a

37 and 26 %

Alcell Thermogravimetric powder balance (TGA) 5 °C min-1

Pre-heated fluidized-bed

1-3 mm pellets

Indulin AT 10 mg Thermo“Lignoboost” 40 %) of solid mass; the latter would clearly be in the process of thermal decomposition and probably quite reactive. In looking for leads to explain synergistic effects between pyrolyzing biomass components, leading to our “char deficit”, the thermal sensitivity of laevoglucosan provides a useful clue. With a sublimation temperature of 115 °C and an onset of decomposition around 300 °C, the volatilization of laevoglucosan at typical cellulose pyrolysis temperatures would be as rapid as its subsequent thermal degradation. We have already seen that much of the lignin microstructure would likely remain intact at these relatively low temperatures – likely to be in its initial stages of dehydration and carbonization. As secondary reactions of laevoglucosan (and possibly also hemicellulosederived volatiles) within the confined space of intermeshed microstructures lead to thermal degradation, the resulting highly oxygenated environment is likely to enhance the incipient thermal breakdown reactions of lignin microstructures. Thus, one explanation of the “lignin char deficit” would go through this highly oxygenated pyrolysis environment leading to the more effective degradation and volatilization of plant derived lignin compared to the pyrolysis of chemically isolated lignin. (iii) Order of relative tar reactivities: Tar yields measured at the exit of fluidized-bed pyrolysis reactors rise to a maximum with increasing temperature and then decline as tarcracking reactions in the bed and the reactor freeboard begin to destroy more tar than is produced by increasing the sample temperature.22,23,54 The temperature of the tar yield maximum tends to rise with the increasing thermochemical stability of the tar vapours. In an already cited fluidized bed pyrolysis study54, the low-middle rank bituminous coal (Linby coal; UK) showed a tar yield maximum around 580–590°C. For a more highly oxygenated lignite sample, the analogous maximum was observed at around 530 °C (not shown), while for cellulose itself and for silver birch wood (the same sample as in Table 2), the tar-yield maxima occurred between 425 – 450 °C. The full set of data may be found in Ref. 54. While the shapes of the asymmetric bell-shaped curves for Linby coal tar yields are similar to those for silver birch wood, the data show far greater proportions of tar survival at higher temperatures by the Linby coal tars. Taken together, these data show the significant differences between the thermal stabilities of tars from different sources. They also show the temperature ranges where tars produced during the pyrolysis of different samples are thermally cracked. In considering the factors affecting outcomes during the pyrolysis of naturally occurring biomass samples, the temperatures at which the thermal breakdown of each distinct biomass component is initiated and the thermal reactions of released volatiles (from the distinct biomass components) must be considered as concurrent effects. Although the data sets presented are incomplete, the clearly observed reactivity of tars/oils evolving during cellulose pyrolysis tend to confirm the reasons presented to explain why little or no laevoglucosan is found in tars from the pyrolysis of naturally occurring biomass materials. As mentioned 14 ACS Paragon Plus Environment

Page 15 of 19

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

Energy & Fuels

above, the “char deficit” observed during the pyrolysis of lignins embedded in naturally occurring lignocellulosic biomass is thought to arise from interactions (of the pyrolyzing lignin) with reactive breakdown products from more highly oxygenated components of the lignocellulosic biomass.

Data from differential scanning calorimetry: DSC (differential scanning calorimetry) experiments conducted using pure cellulose and the samples cited in Tables 1-3, provide a measure of confirmation for the proposition that the chemically reactive environment near pyrolyzing lignin matrices substantially alter the progression of their pyrolytic reactions. In the DSC experiments, pure cellulose and “pure” lignin showed characteristic single peaks, showing the temperature range over which the samples pyrolyzed. However, the two composite samples tested (sugar cane bagasse and silver birch wood) gave broader single peaks nearer the zone where the (pure) cellulose decomposition peak had originally appeared. The peak found during the decomposition of (“pure”) lignin could not be observed at all.21 By themselves, these observations from DSC do not represent a conclusive proof of the argument. However, taken together with the “lignin char deficit,” the evidence strongly suggests that the pyrolysis chemistry of lignins is fundamentally altered when thermal breakdown takes place within pyrolyzing matrices of highly oxygenated and more reactive species. Other lines of investigating synergistic phenomena during biomass pyrolysis: Thus far, observations of synergistic effects outlined above have relied on two effects to characterize interactions between distinct biomass components during pyrolysis: the “lignin char deficit,” and, the levels of laevoglucosan survival (or destruction), depending on the structure, defining the chemical environment within which the cellulose was pyrolyzed. Analogous effects pertaining to extractables, mineral matter and hemicelluloses have not been studied, since the char residue from these components appears to be small (cf. Tables 1-3). Meanwhile, recent work has provided valuable additional avenues for investigating synergistic interactions during biomass thermal breakdown. Couhert and co-workers57 determined gas evolution in an entrained flow reactor, during the pyrolysis of a set of five lignocellulosic biomass samples. Similar experiments were performed using xylan (as hemicelluloses), cellulose and several specimens of chemically isolated lignins. This work involved calculating the total gas evolution from the pyrolysis of lignocellulosic biomass, through the weighted sums from the evolution of gas from isolated components of biomass and the contents of biomass components in the naturally occurring species. The authors concluded that “…such a simple approach was not successful.” The authors suggested that the discrepancies might have been caused by the effect of the mineral matter contents of the naturally occurring biomass samples; however, systematic trends appeared difficult to discern. In an earlier study investigating the co-pyrolysis of coal and biomass, observations had been made suggesting that mineral matter contained in biomass might have a catalytic effect on the overall process.58 A recent review by Vassilev and co-workers presented a survey of mineral matter and extractable material in lignocellulosic biomass.3,6,7 Whilst this approach serves to supplement our knowledge on the descriptive aspects of biomass composition, there appears

15 ACS Paragon Plus Environment

Energy & Fuels

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

Page 16 of 19

to be a need for purpose designed experiments to investigate the behaviour of all biomass components during thermal breakdown. Implications for modelling pyrolysis reactions: While the mathematical modeling of pyrolysis reactors themselves is beyond the scope of this study, the conclusions reached do not imply the existence of insurmountable barriers to the modeling of the operation of larger scale pyrolysis/gasification reactors. Our findings do, however, point to the necessity of adopting alternative lines of attack in approaching the problem. One possible line of attack would entail the development of databases for product distributions determined during the pyrolysis of selected biomass samples, using purpose designed bench-scale experiments for generating information to be used in simulating and eventually designing larger scale process equipment and reactors. It thus appears necessary to devise experiments that are (i) able to tease out the fundamental thermal breakdown patterns of solid fuel particles, and (ii) an accompanying set of experiments, able to track the sequence of reactions that the released volatile products would undergo, in the form of secondary extra-particle reactions. Once these two strands of information are established, the usual rules of reactor modelling may be put to use, for calculating product distributions and reactor volumes. Discrepancies from simple models may then be used to formulate increasingly more realistic reaction schemes and reactor models, without a need to formulate the “detailed pyrolysis chemistry” of lignocellulosic biomass.

SUMMARY & CONCLUSIONS Experimental results have been presented showing that significantly larger char yields were observed when chemically isolated lignin samples were pyrolyzed compared to the pyrolysis of lignins naturally embedded in plant derived material. The apparent “lignin char deficit” has been confirmed by surveying a wide array of lignin pyrolysis experiments, which drew on a variety of original plant materials and lignin isolation methods. These results reflect the extent of synergistic effects between distinct plant components during the pyrolysis of naturally occurring lignocellulosic biomass. Within this framework, the “deficit” is consistent with chemical reactions between plant lignin and the reactive and highly oxygenated molecular fragments, generated by the prior thermal breakdown of more labile biomass components, viz. cellulose and hemicelluloses. Char yields from the pyrolysis of pellets prepared using mixtures of cellulose and lignin powders gave trends that did not conform to those from the pyrolysis of plant derived biomass. In addition to chemical makeup, the results show that char yield trends from the pyrolysis of lignocellulosic biomass are closely linked to the morphology, i.e. the microscopic architecture, of distinct biopolymers that hold plant material together. The complex nature of plant structure (and its effect on pyrolysis product distributions) would require that we include a high level of morphological mapping into the ab-initio modeling of the detailed chemistry of biomass pyrolysis. This seems a complicated and unrealistic procedure. The attempt to study the nature of synergistic interactions recognized during the thermal breakdown of naturally occurring biomass has led to the examination of several, probably inter-connected phenomena: (i) The intermeshing of natural biopolymers composed of distinct biomass components, (ii) The different thermal-breakdown temperature ranges of 16 ACS Paragon Plus Environment

Page 17 of 19

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

Energy & Fuels

cellulose, lignins and hemicelluloses, and, (iii) The differences between the thermal sensitivities (i.e. reactivities) of tars evolving from distinct plant components. The outlook summarized in this paper relies mainly on evidence relating to the pyrolysis of cellulose and lignins. Analogous problems relating to the pyrolysis of extractables, mineral matter and hemicelluloses remain to be investigated using meaningful, purpose designed experiments. It is concluded that the observed sensitivity of reaction pathways to plant specific structural features poses added challenges in formulating ab-initio models for tracking the “detailed pyrolysis chemistry” of lignocellulosic biomass. The complexity of the pyrolysis chemistry of such materials does not appear to justify simple generalizations. The evidence presented suggests that our general understanding of the parameters involved falls well short of predicting trends in product distributions, let alone arriving at quantitative predictions of the detailed chemistry of the pyrolytic processes of naturally occurring biomass. Given the complexities of synergistic interactions during pyrolysis and their probable dependence on species specific plant morphologies, attempting to develop ab-initio mathematical models (mathematical simulations) of the detailed chemistry of biomass pyrolysis, in a manner that would follow “…analogously from the problems of cellulose…”20 does not appear to be realistic. However, it is still possible to model the behaviour of larger scale pyrolysis/gasification reactors by combining conventional reactor design concepts with experimental data from experiments that are (i) capable of identifying the fundamental patterns of thermal breakdown, and (ii) designed to complement such data by tracking the extra-particle secondary reactions of volatile products.

1

Morgan, T.J.; Kandiyoti, R., Chemical Reviews 2014, 114, 1547. Hua-Jiang Huang, H-J.; Ramaswamy, S.; Tschirner, U.W.; Ramarao, B.V. Separation and Purification Technology 2008, 62, 1. 3 Vassilev, S.V.; Baxter, D.; Andersen, L.K.; Vassileva, C.G.; Morgan, T.J. Fuel 2012, 94,1. 4 Vassilev, S. V.; Baxter, D.; Andersen, L. K.; Vassileva, C. G. Fuel 2010, 89, 913 5 Gunther, C., Mosandl, A. 1986. The Composition of Oak and an Overview of its Influence on Maturation. updated: August 16, 2001. accessed: 19 November 2013; http://homedistiller.org/oak.pdf 6 Vassilev, S. V.; Baxter, D.; Andersen, L. K.; Vassileva, C. G. Fuel 2013, 105, 40. 7 Vassilev, S. V.; Baxter, D.; Andersen, L. K.; Vassileva, C. G. Fuel 2013, 105, 19. 8 Shafizadeh, F.; Fu, Y. L. Carbohydrate Res. 1973,29, 113. 9 Stiles, H.N., PhD Thesis, University of London, 1986. 10 Shafizadeh, F. J. J. Anal. & Appl. Pyrolysis 1982, 3, 283. 11 Richards, G. N.; Shafizadeh, F. Austral. J Chem 1978, 8, 1825. 12 Fraga, A.R.; Gaines, A.F.; Kandiyoti, R. Fuel 1991, 70, 803. 13 Li, C-Z.; Bartle, K.D.; Kandiyoti, R. Fuel 1993, 72, 3. 14 Pindoria, R.V.; Megaritis, A.; Herod, A.A.; Kandiyoti, R. Fuel 1998, 77, 1715. 15 Furimsky, E. Appl. Catal. A 2000, 199, 147. 16 Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044. 17 Sanna, A.; Andrésen, J.M. Chem SusChem 2012, 5, 1944. 2

17 ACS Paragon Plus Environment

Energy & Fuels

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

Page 18 of 19

18

Huber, G.W., T.P. Vispute, and K. Routray, Production of hydrogen, liquid fuels, and chemicals from catalytic processing of bio-oils, 2010, University of Massachusetts, USA; 54pp. Patent application. 19 Agarwal, V., Dauenhauer, P.J.; Huber, G.W.; Auerbach, S.M. J. Am. Chem. Soc. 2012, 134, 14958. 20 Mettler, M.S.; Vlachos, D.G.; Dauenhauer, P.J. Energy Environ. Sci.2012, 5, 7797. 21 Fraga-Araujo, A.R. PhD Thesis, University of London, 1990. 22 Tyler, R. J. Fuel 1979, 58, 680. 23 Tyler, R. J. Fuel 1980, 59, 218. 24 Hamilton, L.H. Fuel 1980, 59, 112. 25 Hamilton, L.H. Fuel 1981, 60, 909. 26 Habermehl, D.; Orywal, F.; Beyer, H.D. Chemistry of Coal Utilization Second Supplementary Volume (ed. Elliott, M. M.), Wiley, New York, p. 317. 27 Gibbins-Matham, J.R.; Kandiyoti, R. Energy & Fuels 1988, 2, 505. 28 Fukuda, K.; Dugwell, D.R.; Herod, A.A.; Kandiyoti, R. Energy & Fuels 2004, 18,1140. 29 Solomon, P.R.; Serio, M.A.; Carangelo, R.M.; Markham, J.R. Fuel 1986, 65, 182. 30 Boutin, O.; Ferrer, M., Lede, J. J. Analytical & Applied Pyrolysis 1998, 47, 13. 31

Fraga, A.R.; Gaines, A.F.; Kandiyoti, R. Fuel 1991, 70, 803.

32

Cybulska, I.; Brudecki, G.; Rosentrater, K.; Julson, J. L.; Lei, H. Bioresource technology 2012 118, 30. 33 Kandiyoti, R.; Herod, A.A.; Bartle, K.D. Solid Fuels and Heavy Hydrocarbon Liquids: Thermal Characterization and Analysis; Elsevier Science Publishers: Amsterdam, Oxford, London, New York, 2006; see Chapter 3. 34 Gonenc, Z.S.; Gibbins, J.R.; Katheklakis, I.E.; Kandiyoti, R. Fuel 1990, 69, 383. 35 Kandiyoti, R. Fuel 2002, 81, 975. 36 Iatridis, B.; Gavalas, G.R. Ind. Eng. Chem. Prod. Res. Dev.1979, 18(2), 127. 37 Chan, R.W-C.; Krieger, B.B. J. Applied Polymer Science 1981, 26, 1533. 38 Nunn, T.R.; Howard, J.B.; Longwell, J.P.; Peters, W.A. Ind. Eng. Chem. Process Des. Dev. 1985, 24, 844. 39 Caballero, J.A.; Font, R. ; Marcilla, A. J. Anal. & Applied Pyrolysis 1996, 36, 159. 40 Ferdous, D.; Dalai, A. K.; Bej, S.K. ; Thring, R. W. Energy & Fuels 2002, 16, 1405. 41 Wang, S.; Wang, K.; Liu, Q.; Gu, Y.; Luo, Z.; Cen, K.; Fransson, T. Biotechnology Advances 2009, 27, 562. 42 de Wild, P.; Van der Laan, R.; Kloekhorst, A.; Heeres, E. Environmental Progress & Sustainable Energy 2009, 28, 461. 43 Beis, S.H.; Mukkamala, S.; Hill, N.; Joseph, J.; Baker, C.; Jensen, B.; Stemmler, E.A.; Wheeler, M.C.; Frederick, B.G.; van Heiningen, A.; Berg, A.G.; DeSisto, W.J. BioResources 2010, 5, 1408. 44 Trinh, T.N.; Jensen,P.A., Sárossy, Z.; Dam-Johansen, K.; Knudsen, N.O.; Sørensen, H.R.; Egsgaard, H. Energy & Fuels 2013, 27, 3802. 45 de Wild, P.J.; Huijgen, W.J.J.; Heeres, H.J. Journal of Analytical and Applied Pyrolysis 2012, 93, 95. 46 Nowakowski, D.J.; Bridgwater, A.V.; Elliott, D.C.; Meier, D.; de Wild, P. Journal of Analytical and Applied Pyrolysis 2010, 88, 53. 47 Guerra. A; Filpponen, I. J. Agric. Food Chem. 2006, 54, 5939. 48 Li, C-Z.; Bartle, K.D.; Kandiyoti, R. Fuel 1993, 72, 1459. 18 ACS Paragon Plus Environment

Page 19 of 19

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

Energy & Fuels

49

Kandiyoti, R.; Herod, A.A.; Bartle, K.D., op. cit.; pp. 38-40. Dyrkacz, G.R.; Horwitz, E.P. Fuel 1982, 61, 3. 51 Dyrkacz, G.R.; Bloomquist, C.A.A.; Ruscic, L. Fuel 1984, 63, 1367. 52 Hosoya, T.; Kawamoto, H.; Saka, S. Journal of Analytical and Applied Pyrolysis 2007, 80, 118. 53 Zaror, C.A.; Hutchings, I.S.; Pyle D.L.; Stiles, H.N.; Kandiyoti, R. Fuel 1985, 64, 990. 54 Stiles, H. N.; Kandiyoti, R. Fuel 1989, 68, 275. 55 Yang, H.; Yan, R.; Chen, H; Lee, D-H.; Zheng, C. Fuel 2007, 86, 1781. 56 Schwaiger, N.; Witek, V.; Feiner, R.; Pucher, H.; Zahel, K.; Pieber, A.; Pucher, P.; Ahn, E.; Chernev, B.; Schroettner, H.; Wilhelm, P.; Siebenhofer, M. Bioresource Technology 2012, 124, 90. 57 Couhert, C.; Commandre, J-M.; Salvador, S. Fuel 2009, 88, 408. 58 Collot, A-G.; Zhuo, Y.; Dugwell, D.R.; Kandiyoti, R.; Fuel 1999, 78, 667. 50

19 ACS Paragon Plus Environment