Influence of Crystal Allomorph and Crystallinity on the Products and

Subscriber access provided by RMIT University Library

Article

The influence of crystal allomorph and crystallinity on the products and behavior of cellulose during fast pyrolysis Calvin Mukarakate, Ashutosh Mittal, Peter Nolan Ciesielski, Sridhar Budhi, Logan Thompson, Kristiina Iisa, Mark R Nimlos, and Bryon S Donohoe ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/acssuschemeng.6b00812 • Publication Date (Web): 19 Jul 2016 Downloaded from http://pubs.acs.org on July 22, 2016

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.

ACS Sustainable Chemistry & Engineering 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 37

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

ACS Sustainable Chemistry & Engineering

The influence of crystal allomorph and crystallinity on the products and behavior of cellulose during fast pyrolysis Calvin Mukarakate1*‡, Ashutosh Mital2‡, Peter N. Ciesielski2, Sridhar Budhi1, Logan Thompson1, Kristiina Iisa1, Mark R. Nimlos1, and Bryon S. Donohoe2* 1-National Bioenergy Center, and 2-Biosciences Center, National Renewable Energy Laboratory, 15013 Denver West Parkway, Golden, Colorado 80401, United States. Corresponding Author Correspondence should be addressed to [email protected], or [email protected] KEYWORDS cellulose, pyrolysis, crystallinity, allomorph, biomass conversion, biochar

ABSTRACT

Cellulose is the primary biopolymer responsible for maintaining the structural and mechanical integrity of cell walls, and during the fast pyrolysis of biomass may be restricting cell wall expansion and inhibiting phase transitions that would otherwise facilitate efficient escape of

ACS Paragon Plus Environment

1


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 37

pyrolysis products. Here, we test whether modifications in two physical properties of cellulose, its crystalline allomorph and degree of crystallinity, alter its performance during fast pyrolysis. We show that both crystal allomorph and relative crystallinity of cellulose impact the slate of primary products produced by fast pyrolysis. For both cellulose-I and cellulose-II, changes in crystallinity dramatically impact the fast pyrolysis product portfolio. In both cases, only the most highly crystalline samples produced vapors dominated by levoglucosan. Cellulose-III, on the other hand, produces largely the same slate of products regardless of its relative crystallinity, and produced as much or more levoglucosan at all crystallinity levels compared to cellulose-I or II. In addition to changes in products, the different cellulose allomorphs affected the viscoelastic properties of cellulose during rapid heating. Real-time hot-stage pyrolysis was used to visualize the transition of the solid material through a molten phase, and particle shrinkage. SEM analysis of the chars revealed additional differences in viscoelastic properties and molten phase behavior impacted by cellulose crystallinity and allomorph. Regardless of relative crystallinity, the cellulose-III samples displayed the most obvious evidence of having transitioned through a molten phase.

Introduction Loss of potentially useful carbon to light gases and char remains an important technical barrier to the economic and sustainable deployment of fast pyrolysis for the conversion of lignocellulosic biomass to transportation fuels.1 Since the biomass feedstock itself remains the dominant operational cost for biorefineries, maximizing the conversion of biomass to fuels and chemicals is vital to the feasibility of these processes.2 Typically, fast pyrolysis produces a solid byproduct, bio-char, that contains 15-25% of the carbon of the initial feedstock.3 While this


2

Page 3 of 37

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


stranded chemical energy can be used to provide process heat, it also represents an opportunity to drive a greater percentage of the initial biomass feedstock into more valuable products.4 The current state of fast pyrolysis technology provides few control points to adjust in an effort to increase the yield of bio-oil from biomass via fast pyrolysis.5 6 Heating rate, residence time, and particle size remain the main determinants that affect the ratio of high-value condensable organics to lower value products such as water, char, and light gases.7 Plant cell walls feature a fiber-reinforced, composite matrix design. The cellulose microfibrils provide the fiber reinforcement and the mechanical resistance to tensile and sheer forces. The hemicellulose and lignin components comprise the composite matrix. Recent work visualizing the phase transitions that solid biomass undergoes during pyrolysis8,9 and rheology studies on biomass biopolymers and whole biomass10 have suggested a unique role for cellulose in inhibiting the progression of biomass conversion to bio-oil during fast pyrolysis.11 Time-resolved analysis of product evolution during fast pyrolysis has also suggested that the major biopolymers in biomass volatilize roughly in the order of lignin (often with a broad volitilazation profile), hemicellulose, and finally cellulose.12 Furthermore, in an excellent study on the impact of ball milling cellulose to reduce its crystallinity Wang et al. suggest that modified, amorphous cellulose can more easily transition through a liquid phase during dehydration reactions to levoglucosan, while highly crystalline cellulose remains a solid.13,14 The same group also reported higher yields of anhydro-sugars from the pyrolysis of high crystallinity cellulose compared to ball-milled material.13 These independently observed phenomena are all relevant when considering the role of cellulose in the transition of biomass through the molten phase during pyrolysis. Cellulose’s role as the cell wall polymer chiefly responsible for maintaining structural and mechanical integrity of the material, may be restricting expansion of the cell wall


3


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 37

and inhibiting phase transitions that would otherwise facilitate efficient volatilization and escape of fast pyrolysis products from the biomass particle. The native form of cellulose in plants is a mixture of cellulose-Iβ and cellulose-Iα.15,16 Thermochemical treatments have been developed to selectively modify cellulose properties such, as changing the crystalline allomorph of the cellulose microfibrils from the native cellulose-I to either cellulose-II or cellulose-III (Figure 1), or changing the crystallinity. 17,18 Either of these modes of changing the physical properties of the cellulose could generate biomass with dramatically altered conversion performance. However, it remains unclear what, if any, physical properties of cellulose would be most desirable to alter for enhanced conversion by fast pyrolysis. Cellulose-I is the native form of cellulose in plant cell walls and exhibits 2-

Figure 1. Molecular models of cellulose allomorphs. Cellulose-I is the native form of cellulose in plant cell walls, including cotton linters, and exhibits 2-dimensional intralayer hydrogen-bonding networks. Cellulose-II can be generated by ionic liquid treatment and re-arranges cellulose chains into antiparallel sheets exhibiting a 3-dimensional network of intralayer and interlayer hydrogen bonding. Cellulose-III can be generated by treatment with anhydrous ammonia and, like cellulose-II, forms intralayer and interlayer hydrogen bonds, but the chains remain parallel like in cellulose-I. Black represents carbon atoms, red represents oxygen, white represents hydrogen atoms, and blue dashed lines represent hydrogen bonds.


4

Page 5 of 37

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


dimensional intralayer hydrogen-bonding networks.19 Cellulose-II can be generated by ionic liquid treatment and re-arranges cellulose into antiparallel sheets exhibiting a 3-dimensional network of both intralayer and interlayer hydrogen bonding. Cellulose-III can be generated by treatment with anhydrous ammonia and, like cellulose-II, forms intralayer and interlayer hydrogen bonds, but the chains retain a parallel orientation like cellulose-I. In addition to thermochemical treatments of biomass to modify cellulose properties, plants may be engineered to synthesize cellulose with altered crystallinity. In the model plant system, Arabidopsis thaliana, plants exist that have been modified to produce less crystalline cellulose indicating that this property of cellulose is genetically controllable.20 Cellulose physical properties impact both decrystallization and depolymerization,21 and thereby may be affecting the reaction mechanism pathway during pyrolysis.22 The mechanisms and kinetics of cellulose pyrolysis have been well explored and described in the literature.23-28 Here, we test whether modifications in two physical properties of cellulose, crystallinity and crystalline allomorph, change its behavior in fast pyrolysis. Specifically, we determine the slate of pyrolysis products by molecular beam mass spectroscopy (MBMS), identify and quantify the products with GCMS/FID, and investigate the behavior of different cellulose allomorphs at elevated temperatures relevant to fast pyrolysis. The resultant chars are evaluated for evidence of variable phase-transition behavior. This work will ultimately inform novel concepts for biomass feedstock pretreatments and engineering that improve the yields and quality of bio-oil by fast pyrolysis by manipulating the molecular arrangement of cellulose. Results and Discussion X-ray diffraction of the cotton linter allomorphs confirms their crystalline form and crystallinity index


5


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 37

X-ray diffractograms from the different cellulose allomorph samples with varying degrees of crystallinity are presented in Figure 2. Evidence for the change from cellulose-I to cellulose-II is the appearance of a doublet in the XRD for the 10Ī and 002 peaks (at 2θ values of about 20 and 22) as shown in Figure 2B. Figure 2C shows the X-ray diffractograms for cellulose-III samples prepared at different degrees of crystallinity. The change from cellulose-I to cellulose-III can be observed by following the position of the 002 peak shifting from a 2θ value of 23 to 21 as cellulose-I is completely converted to cellulose-III. Chemical composition analysis of all the allomorphs of cellulose-I presented in the supplementary data (Table S1) was performed using NREL LAPs29 and showed 99-100% glucan content for all the samples suggesting that no degradation of cellulose occurred during the chemical treatments.

Figure 2. XRD diffractograms of the different cellulose allomorphs, prepared from cotton linters, at varying degrees of crystallinity. The crystallinity index of the samples are CI-H=62, CI-M=44, CI-L=34, CII-H=57, CIIM=49, CII-L=40, CIII-H=60, CIII-M=49, CIII-L=32. Cellulose-II (B) shows a characteristic doublet in the 10Ī and 002 peaks at 2θ values of about 20 and 22. Cellulose III (C) is characterized by the shift in position of the 002 peak from a 2θ value of 23 to 21.

Dark-field microscopy shows the range in physical form among the cellulose allomorph samples To investigate the general morphology of the cotton linter materials and determine if the thermochemical treatments used to convert cellulose-I to cellulose-II or III at varying crystallinity had any additional impact on the size or structure of the particles, we examined the samples by dark field optical microscopy (Figure 3). An initial observation we made while


6

Page 7 of 37

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


handling the samples was that the different allomorph samples behaved differently even during preparation of the microscope slides: the cellulose-I material, in particular the medium and low crystallinity samples, were challenging to disperse and tended to clump rather that separating into fine particles. This was consistent whether the cellulose samples were kept dry or dispersed in water. Representative fields of view showing the dispersion patterns and general morphology of the individual cotton linter particles are shown in Figure 3.

Figure 3. Dark field micrographs of cellulose allomorph particles showing microscale morphology and patterns of dispersion. The images suggest that the cellulose-I material comprise particles with a relatively high aspect ratio and a tendency to clump. Aggregates are enhanced in the medium and low crystallinity cellulose-I samples. The thermochemical treatments that induce cellulose transformations into the cellulose-II and cellulose-III allomorphs have also resulted in particles with lower aspect ratios and less tendency to form clumps.

In addition to differences in clumping which may relate to variable surface properties, there were slight differences in the size and morphology of the linter particles. The cellulose-I control sample particles (high crystallinity index, CI-H) appeared typical of cotton linters. They ranged in size from 30-90 µm and displayed particle aspect ratios of ~2-3. The phosphoric acid swollen


7


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 37

cellulose (PASC) cellulose-I samples (CI-M, CI-L) appeared different with few discrete particles visible separate from the clumped material. The PASC samples had the appearance of linter cell walls having been at least partially delaminated and formed into layered sheets of less organized cellulose compared to the distinct particles of the other seven samples. All of the cellulose-II and cellulose-III samples appeared similar in size and aspect ratio with most particles ranging from 30-60 µm in size with aspect ratios ~2. Some of the CII and CIII particles had a tendency to clump in the low crystallinity samples, but none to the extent of the low crystallinity cellulose-I material. These analyses suggest that there are mesoscale properties of the cellulose-I samples at medium and low crystallinity that could impact their behavior in fast pyrolysis by imposing transport limitations, but no reason to suspect the minor differences in particle-scale properties would explain any differences among the rest of the allomorph samples. Molecular beam mass spectroscopy reveals differences in the profile of pyrolysis products released from the cellulose allomorphs The impact of variations in crystal structure on the distribution of products from fast pyrolysis was investigated by molecular beam mass spectroscopy (MBMS). The pyrolysis of pure cellulose at 500°C and analysis of products using the MBMS have been reported in previous studies12,30-33, and has been shown to produce levoglucosan m/z 162, 1,4:3,6-dianhydro-α-Dglucopyranose m/z 144 (this also results from an ionization fragment of levoglucosan), furfuryl alcohol m/z 98, and hydroxyacetaldehyde m/z 60. The peaks at m/z 57 and 73 are typical fragment ions from levoglucosan. Levoglucosan can also result in fragment ions at m/z 60 and 98.33 The spectra in Figure 4 were recorded from the pyrolysis of cellulose-I samples with


8

Page 9 of 37

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


Figure 4. py-MBMS spectra of CI-cellulose samples with varying crystallinity.

varying degrees of crystallinity, and it shows that they all produce similar peaks. However, the distributions of peak intensities are different for the three CI samples. The spectrum of sample CI-H is dominated by fragment ions of levoglucosan (m/z 57, 60, 73, 98 and 144), these ions were observed during pyrolysis of levoglucosan using chemical ionization.34 The most intense peak is at m/z 60, which could be due to hydroxyacetaldehyde, but the absence of the expected fragment ions of hydroxyacetaldehyde at m/z 31 and 32 rules out this assignment.33 Alternatively, the m/z 60 species is a levoglucosan fragment ion, which would be consistent with the rest of the spectra indicating that the pyrolysis of sample CI-H produces primarily levoglucosan. The spectrum of sample CI-M is dominated by m/z 126, which is due to 5HMF. The fragment ions for 5HMF are at m/z 69, 97 and 110.33 This means that pyrolysis of CI-M produces 5HMF and levoglucosan as the major products. Additional evidence for this can be deduced by comparing the intensities of the levoglucosan fragment ion at m/z 60, which represents about half in the CI-M spectrum. The spectrum for CI-L is dominated by m/z 43,


9


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 10 of 37

which is usually observed in the presence of carbonyl compounds such as acetol. The species at m/z 60 is partly hydroxyacetaldehyde (evidenced by presence of m/z 31 and 32 fragment ions)33 and also levoglucosan. This indicates that pyrolysis of CI-L produces low molecular weight products and low amounts of levoglucosan as shown by the weak m/z 144 peak. It is clear from this analysis that the MBMS produces complex spectra comprising a mixture of molecular peaks and fragment ions, so to simplify the data and understand the product distributions produced by all the cellulose samples with different crystal structure we employed multivariate analysis. To analyze the data with multivariate analysis, mass spectra of pyrolysis products were collected in four replicates for each of the cellulose samples. These data were analyzed using multivariate curve resolution optimized by alternating least squares (MCR-ALS) to elucidate which chemical compounds were contributing the most variance to the MBMS results. The MCR-ALS analysis has been successfully applied to simplify and find trends in MBMS data.12,30-32,35 The top 100 out of 500 masses with the largest variances were chosen for MCR-ALS analysis, producing a data set with dimensions of 36 samples by 100 masses. We optimized the MCR-ALS analysis to produce three pure components (PCs). Increasing the PCs to more than three did not improve the residual error. The reconstructed spectra of the three principal components are presented in Figure 5.


10

Page 11 of 37

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


Figure 5. Reconstructed spectra for each pure component (PC13) from MCR-ALS analysis of fast pyrolysis of cellulose allomorphs reveal differences in the primary products. PC1 (green) contains peaks typical of sugar dehydration reactions such as levoglucosan (m/z 162). The unique peak in PC2 (blue) is a carbonyl fragment (m/z 43). PC3 (green) is dominated by 5hydroxy-methyl furfural (m/z 126).

The first component, PC 1 (shown in red) contains compounds typical of pyrolysis products from levoglucosan and is similar to the most intense peaks in the spectrum recorded for CI-H. The peaks observed in this PC are all consistent with products from levoglucosan m/z 162, 144, 114, 98, 73, 60, 57 and 44. Levoglucosan is normally considered to be the primary product of cellulose pyrolysis. The second principal component, PC2 (shown in blue) contains a strong contribution from a compound at m/z 43, which is attributed to a carbonyl fragment (acetyl, CH3-CO-) and is similar to the most intense peaks for CI-L. The third principal component, PC3 (shown in green) is largely dominated by 5-hydroxymethyl furfural observed at m/z 126 and is similar to the most intense peaks for CI-M. The scores of these three PCs as a function of each of the different cellulose samples are displayed in Figure 6, and provide clear evidence that variations in the crystal structure affect the distribution of products formed during pyrolysis. Across all three cellulose allomorphs, the slate of products produced by the samples with the highest crystallinity is dominated by levoglucosan.


11


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 12 of 37

Figure 6. MCR-ALS analyses of the py-MBMS data for the allomorphs suggests significant difference in product slate both across variable crystallinity within each allomorph as well as among the cellulose allomorphs. Within cellulose-I, levoclucosan is the dominant component at all three levels of crystallinity, but its relative proportion decreases with decreasing crystallinity. Within cellulose-II, levoglucosan dominates the high crystallinity allomorph, while carbonyl dominates the medium and low crystallinity. Within cellulose-III, levoglucosan dominates all three samples at similar proportions regardless of crystallinity.

This effect is especially prominent in cellulose-I and III, where the contribution of levoglucosan is more than twice the sum of the other two components. In the case of cellulose-I, these results suggest that crystallinity plays an important role in determining the products, as different degrees of crystallinity produce mass spectra with carbonyl component scores. These trends are also in good agreement with the most intense peak in the pyrolysis gas chromatography mass spectroscopy (py-GCMS) data shown in Figure 7.


12

Page 13 of 37

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


However, this trend with crystallinity does not extend to the other two crystal allomorphs. Cellulose II, at both medium and low crystallinity gave products rich in carbonyl fragments. Meanwhile cellulose-III consistently produced levoglucosan as a dominant product, regardless of the relative crystallinity of the material. Py-GCMS and FID provides positive identification and semi-quantification of the primary pyrolysis products The py-GCMS/FID system complements MBMS analysis by providing positive identification of products and quantifying their abundance. Pyrolysis of all nine cellulose samples produced furfural, 1,2cyclopentanedione, 5-hydroxymethyl furfural, anhydroxylopyranose and levoglucosan as major products (>1% area yields).13,36 Table 1 lists compounds Figure 7. Py-GCMS chromatograms recorded from pyrolysis of the CI, CII and CIII samples with varying crystallinity. The major peaks are identified and reported in Table 1.


13


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 14 of 37

observed from pyrolysis of the cellulose samples including their relative amounts as a function of total area, only products with yields greater than one percent of the total area are listed in this table. It is clear that the yields of these products were different among the samples. Figure 7 shows example chromatograms recorded from pyrolysis of the CI, CII and CIII samples.

Table 1. Py-GCMS peak identification.

As can be seen from Figure 7A, CI-H produced the greatest amount of levoglucosan, followed by CI-M and the CI-L produced the least levoglucosan. The CI-M produced the greatest amount


14

Page 15 of 37

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


of 5HMF followed by CI-H and then CI-L. On the other hand, CI-L produced the greatest amounts of low molecular weight compounds. Figure 7B shows that CII-H produces a greater amount of levoglucosan than CII-M and CII-L. Figure 7C shows that all CIII samples produces roughly the same amounts of levoglucosan. These results are in good agreement with the pyMBMS work. The areas derived from the py-GCFID were divided by the mass of each sample to give semi-quantitative yields of the pyrolysis products and are plotted in Figure 8. Panel 8A shows that high crystalline samples for CI and CII allomorphs produced more levoglucosan compared to their low crystalline counterparts. The CI samples show a linear decrease in levoglucosan yield with crystallinity. However, all CIII samples produced high yields of levoglucosan. These results are in good agreement with the py-MBMS results discussed above. We also pyrolyzed an amorphous (crystallinity index=0) sample prepared by phosphoric acid treatment and it produced a similar levoglucosan yield as the CIII samples. This seems counter intuitive and goes against the trend established by the other cellulose-I samples, but could be due to the completely amorphous sample becoming an again more homogenous cellulose form after pretreatment. Wang et al.,13 also reported that highly crystalline cellulose gave a slightly higher yield of levoglucosan than the amorphous cellulose during fast pyrolysis at 500ºC. CI-H gave the highest yield of levoglucosan because it produced only eight products with yields greater than one percent of the total area (Figure 8F). Levoglucosan accounted for 73 % of the total area. On the other hand, CI-M and CI-L produced 22 and 25 products respectively. The


15


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 37

Figure 8. Analysis of the py-GCMS chromatograms reporting the area per mass of sample pyrolyzed. (A) levoglucosan; (B) other anhydro sugars; (C) furans; (D) cyclopentenones; (E) light oxygenates; (F) number of products >1% in abundance. CI-H produces the greatest amount of levoglucosan, followed by the CIII samples and then the CII-H material. CI-H also produces the fewest additional species besides levoglucosan, followed by CIII-M, CIII-L, CII-H, and CIII-H in that order. All three CIII cellulose samples produce similarly high amounts of levoglucosan

yield of levoglucosan produced from CI-M was 40 % of the total area. This was followed by 5HMF with yield of 7 % of the total area. Figure 8C shows yields of furanic compounds, and as expected CI-M produced the highest amount of 5HMF, followed by the CIII samples and the


16

Page 17 of 37

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


amorphous sample. All nine samples produced roughly the same amount of furfural, with CI-L, CII-M and CII-L producing slightly lower amounts. These three samples however, produced more 2(5H)-furanone compared to approximately the same amounts produced for the CI-H, CIIH, CIIIs and amorphous samples. The yields of 1,2-cyclopentanedione products also follow the same trend as the 2(5H)-furanone, with the CI-L, CII-M and CII-L producing more of this product compared to the other samples (Figure 8D). This could imply different thermal decomposition mechanisms between highly crystalline, more homogenous samples compared to the less crystalline heterogeneous ones. The yield of levoglucosan from CI-L was 21 % of the total area. This was followed by acetol (2-Propanone, 1-hydroxy-) with 7 % yield of the total area. The CI-L produced the least amount of levoglucosan, followed by CII-M and CII-L. This is because these samples produced lower molecular weight products (Figure 8E), for acetol (2-Propanone, 1-hydroxy-) and pyruvic acid. CI-L produced the highest amount of acetol and it was around eight times the amount produced by CI-H. Acetol and pyruvic acid are responsible for the m/z 43 fragment ion observed in the pyMBMS spectra (Figure 4 and 5). It is interesting to note that pyrolysis of all nine cellulose samples produced mainly the 1,6anhydro-β-D-glucopyranose isomer of levoglucosan, while the 1,6-anhydro-β-D-glucofuranose isomer was produced in small amounts and its yields from the CI-L, CII-M and CII-L samples was zero. Even though the pyranose isomer was the major product, pyrolysis of these samples produced significant amounts of furanic compounds (5HMF, furfyryl, etc.) as opposed to pyranic compounds. This could mean that the thermal decomposition of cellulose proceeds via an intermediate, which forms furanic compounds.36 The CI-L, CII-M and CII-L samples without


17


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 37

significant amounts of the furanose isomer did not form the 2,5-Dimethyl-4-hydroxy-3(2H)furanone compound. Thermogravimetric analysis reveals a connection between phase change and pyrolysis products To further explore how the crystalline form of cellulose could be directly impacting the mechanism of pyrolysis and char formation we performed thermogravimetric analysis (TGA). The mass loss and mass derivative data are presented in Figure 9. There is an interesting trend in the data when considered in light of the MCR-ALS and py-GCMS analysis. The samples whose pyrolysis products were strongly dominated by levoglucosan also resulted in the highest mass loss (lowest char yield). This includes cellulose-I at the highest crystallinity, cellulose-II at the highest crystallinity, and cellulose-III at all three crystallinity indexes. Also, this same group

Figure 9. Thermo gravimetric analysis plots of mass loss (top) and mass derivative (middle). CI-H, CII-H, and CIII at all three crystallinity levels exhibited the highest mass loss and the highest onset of mass loss temperature.


18

Page 19 of 37

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


shared the characteristic of having the highest onset of mass loss temperature, all over 300°C, and they exhibited excellent reproducibility in the TGA measurements. These results suggest that cellulose producing the most levoglucosan also gives the lowest char yields and devolatilizes at consistent, higher temperatures. Real-time hot stage microscopy tracks progression through the molten phase during pyrolysis We performed in-situ, hot stage microscopy to determine if the differences in cellulose crystal structure affect the molten phase behavior during pyrolysis in a way that could be visualized. We have used this apparatus in previous work to gain insight into the expansion and contraction of poplar biomass particles during pyrolysis.2 Like our previous results with poplar sections, there was evidence for a volumetric expansion during the pyrolysis of these pure cellulose samples (Figure 10). The percent area covered by 2D transmitted light micrographs of the cellulose samples was measured by image analysis and the cellulose-II and cellulose-III samples in particular exceed the starting percent area coverage before they begin contracting and shrinking. This slight increase in area above the starting value that can be seen in the graphs may be from an overall expansion of the particles or from a slumping of the particles onto the glass slide as they melt. Either explanation is indicative of a phase change. A contraction period followed this initial stage and continued as individual particles shrunk in size before their size eventually stabilized around 450°C.


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

Page 20 of 37

Figure 10. Hot stage microscopy analysis. Bright field micrograph frames of real-time hot stage pyrolysis of cellulose allomorphs and estimation of the peak molten phase transition temperature. Image analysis of the change in area fraction occupied by the cellulose particles during heating. A gallery of individual frames captured during the real-time microscopy of hot stage pyrolysis experiments is presented in Figure 10. The top row of images represents the starting condition of the cellulose allomorph materials at the starting temperature (25°C). The bottom row of images is of the cellulose chars at 450°C. The middle row contains images selected by visual inspection of the real-time video files that represent the peak rate of particle shrinkage. The image analysis data (Figure 10, graphs at bottom) track the change in the image area fraction occupied by the cellulose particles during heating from room temperature to 500ºC. The differences in the point of inflection, where the cellulose samples begin to shrink, and the rate of volume change


20

Page 21 of 37

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


supports the idea that both cellulose crystallinity (for CI, and CII) and allomorph can impact the behavior of cellulose pyrolysis. Scanning electron microscopy (SEM) analysis of char structure further reveals evidence of molten phase behavior Scanning electron microscopy (SEM) was used to characterize the microscale morphology and surface texture of the cellulose chars following pyrolysis. Lower magnification images shown in Figure 11A reveal the general size and shape of the char particles. Variations in crystallinity appeared to have little effect on the morphology of the chars, however significant differences were observed between chars produced from different crystal allomorphs. Char produced by the cellulose-I samples still exhibit structural similarities to cotton linter cell walls and appears more fibrous than char produced from the other allomorphs. Cellulose-II char appeared as smaller fragments relative to the cellulose-I char. While this char less resembled the original

Figure 11. Scanning electron microscopy of cellulose chars. (A) Low-magnification images showing the microscale morphology of the chars. (B) High-magnification images showing the surface texture of the chars. The magnification is constant within each image set, designated by the scale bars top left images of each set.


21


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

Page 22 of 37

cotton linter cells walls than did the cellulose-I char, it still maintained a generally fibrous morphology. Cellulose-III char displayed a structure distinct from the other chars, and retained even less likeness to cotton linter cell walls than did the cellulose-II char. This material appeared as interconnected filaments that had become fused to one another in some regions, in contrast to the other samples that displayed less interconnectivity and more distinct boundaries between particles. Structural differences are even more apparent in the higher magnification micrographs shown in Figure 11B. The micro-scale surface texture of char produced by both cellulose-I and II allomorphs, particularly from the high crystallinity samples, still displays features that appear to be dictated by bundles of cellulose microfibrils and cell wall structure, while the surface of the cellulose-III char was much smoother. Such smoothing of the surface texture likely arose from the molten state adopted by the cellulose during pyrolysis that facilitated a reduction in surface texture. Similar morphological rearrangements are common in other systems that involve biphasic interfaces where the liquid surface re-structures to minimize interfacial surface area and lower the thermodynamic free energy of the system. These observations are indicative of the differences of the viscoelastic properties of the cellulose allomorphs at high temperatures, and suggest the molten phase of the cellulose-III allomorph encountered during these experiments is less viscous than the molten phase experienced by the other allomorphs over the same temperature range. Furthermore, in the context of the product slate measured from cellulose-III by MBMS, this unique morphology hints at an interesting relationship between the apparent viscosity of the molten phase and a route to the production of levoglucosan. Specifically, all of the cellulose-III chars appeared to exhibit a relatively low viscosity in the molten phase regardless of their initial crystallinity, and the mass spectra observed for all the cellulose-III


22

Page 23 of 37

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


samples were dominated by levoglucosan. On the other hand, the pyrolysis of the high crystallinity cellulose-I samples was also dominated by levoglucosan while the char morphology showed the least evidence of having transitioned through a molten phase. This relationship between cellulose crystal forms and their high temperature properties suggests the potential for an additional control point that could be used to tune the slate of products from fast pyrolysis. This work suggests that biomass pretreatments known to alter the crystal structure of cellulose may be able to direct the slate of fast pyrolysis products to a more uniform, more desirable set of intermediates. Additionally, efforts to gain molecular control over the synthesis of plant cellulose by modifying the cellulose synthase complex could indeed help overcome remaining barriers for economic and sustainable deployment of fast pyrolysis of biomass. The potential exists to eventually modify energy crops to produce cellulose with properties better tailored for conversion. Conclusions We have shown that the crystal allomorph and relative crystallinity of cellulose can impact the slate of primary products produced by fast pyrolysis. For both cellulose-I and cellulose-II, changes in crystallinity dramatically alter the product yields from fast pyrolysis. For cellulose-II, only the most highly crystalline samples produced vapors dominated by levoglucosan. For cellulose-I, the dominance of levoglucosan as the major product decreased with decreasing crystallinity. Cellulose-III, on the other hand, produced largely the same slate of products, dominated by levoglucosan, regardless of relative crystallinity. In addition to changes in products, different cellulose allomorphs affected phase change and viscoelastic properties of cellulose during rapid heating. TGA and real-time hot-stage pyrolysis analysis indicated that the same samples that produced the most levoglucosan, CI-H, CII-H, and CIII at all three


23


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 24 of 37

crystallinity levels also had the highest onset of mass loss temperatures and the lowest char yields. SEM analysis of the chars revealed additional differences in viscoelastic properties and molten phase behavior. Regardless of relative crystallinity, the cellulose-III samples displayed the most obvious evidence of having transitioned through a molten phase. In contrast, the highcrystallinity cellulose-I appeared to volatilize without substantial phase transition as evidenced by char morphology. Materials and Methods Generation of cellulose allomorphs from cotton linters The cellulose allomorphs were prepared from cotton linters (CAS Number 9004-34-6; catalogue number 22183, Sigma-Aldrich) that contain cellulose of the Iβ allomorph with a starting crystallinity index (CI) of 62. Cellulose-I at varying degrees of crystallinity were prepared by using a method described by Hall et al., 2010.37 Briefly, 1 g of dry cellulose powder was added to 30 mL of ice-cold concentrated phosphoric acid to generate phosphoric acid swollen cellulose (PASC). The slurry was allowed to react at 0°C for 40 min with occasional stirring. After 40 min, 20 ml of ice-cold acetone was added to the slurry followed by stirring and filtration on a sintered glass crucible. The filtered sample was further washed three times each with 20 ml of ice-cold acetone and DI water. The resulting cellulose was freeze-dried. The cellulose-I samples at reduced crystallinity index (cellulose I medium crystallinity CI-44 (CI-M), and cellulose I low crystallinity CI-34 (CI-L)) were prepared by varying the phosphoric acid concentration. The designations of high, medium and low crystallinity are relative to the sample set used in this study.


24

Page 25 of 37

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


Cellulose-II samples at different degrees of crystallinity were prepared by aqueous NaOH treatment of cellulose-I (Fluka cotton linters, catalog number 22183) carried out according to the method described by Atalla et al.38 For obtaining the least crystalline cellulose-II sample, a sample of the cotton linters was soaked and stirred in an aqueous NaOH solution (23 wt %) at a concentration of 5 g of cotton linters per 100 g of NaOH solution at 25°C for 50 min under a nitrogen atmosphere. This mixture is designated as “mix A”. The NaOH solution was removed from this mix A by filtration and then the filter cake was extensively washed with distilled water until the wash pH ≈ 7. This sample is designated CII-L. For obtaining a medium crystallinity cellulose-II sample, a portion of mix A was diluted with deionized water to 17.5 mass percent NaOH and heated at 70°C for 50 min. This solution was then further heated at 70°C under a nitrogen cover at successive NaOH concentrations of 15, 12, 10, 8, 4, and 1 mass percent NaOH. The lower concentrations of NaOH were obtained by dilution with warm deionized water. 50 min of heating was used for each NaOH concentration. The mix obtained after the 50-min heating at 1 mass percent NaOH is designated as “mix B.” After the final washing with the 1 mass percent NaOH, the cellulose was extensively washed with deionized water at 70 °C to obtain a pH of ≈ 7. The filtered sample is designated CII-M. In order to obtain a highly crystalline cellulose-II sample, a portion of mix B was dewatered by filtration and immersed in glycerol at 100°C in a stainless steel reaction vessel (Parr Instrument Co., Model 4520, Moline, IL). After heating the sample at 145°C for 7 days, the sample was cooled to ≈ 100°C and washed with boiling deionized water. This sample is designated CII-H. All samples were freeze-dried after thorough washing. Cellulose-III samples were prepared by the method described earlier.39 Briefly, a sample of cellulose-I (Fluka cotton linters, catalog no. 22183) with a dry equivalent weight of ≈ 4 g was


25


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 26 of 37

placed in a stainless steel reaction vessel (Parr Instrument Company, Model 4714, Moline, IL). The reaction vessel was clamped shut, weighed, and then chilled in a dry ice acetone bath at 75°C to facilitate transfer of liquid ammonia into the reactor at atmospheric pressure. Anhydrous liquid ammonia was added slowly to the reaction vessel in the ratio ≈ 2.5 g ammonia per gram of cellulose using a stainless steel transfer tube from a liquid ammonia cylinder. After adding ammonia, the vessel was immediately weighed and then cooled in the dry ice acetone bath at 75°C for 15 min. The vessel was then immersed in a water bath maintained at ≈ 25°C until the temperature of the vessel was -33°C, after which it was removed from the water bath and allowed to vent into a hood until all the ammonia had evaporated. The cellulose sample clumped together and retained the shape of the vessel. This clumping may be because the ammonia evaporates slowly at ambient pressure. The sample was gently broken up to a powder using a mortar and pestle. The sample obtained is designated CIII-L. In order to obtain a cellulose-III sample of medium crystallinity, cotton linters were treated at -75°C for 15 min as described above, then the vessel was immersed in a water bath maintained at 25°C for 5 min. After completion of this reaction period and prior to its removal from the water bath, the treatment was terminated by immediately depressurizing the vessel. The ammonia treated cellulose was removed from the vessel and left in the hood overnight until all ammonia had evaporated. This sample is designated CIII-M. In order to obtain a cellulose-III sample of high crystallinity, cotton linters were treated at -75°C for 15 min as described above and then the vessel was immersed in a water bath maintained at 25°C for 30 min. The vessel was then placed in a preheated fluidized sand bath (Techne Inc., Burlington, NJ) maintained at 130°C for 1 h. After this final heat treatment and prior to its removal from the sand bath, the reaction vessel was depressurized. After releasing the ammonia, the vessel was cooled in a water bath maintained at 25°C. The


26

Page 27 of 37

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


cellulose sample was removed from the vessel and left in the hood overnight until the ammonia had evaporated. This sample is designated CIII-H. The samples were analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES) (supplementary Table S2). The samples were analyzed for degree of polymerization previously.40 X-ray diffraction measurements The crystallinity indexes (CI) of cellulose samples were measured by X-ray diffraction (XRD) by using a Rigaku (Tokyo, Japan) Ultima IV diffractometer with CuKα radiation having a wavelength λ (Kα1) = 0.15406 nm generated at 40 kV and 44 mA. The diffraction intensities of dried samples placed on a quartz substrate were measured in the range of 8 to 42° 2θ using a step size of 0.02° at a rate of 2° min-1. The crystallinity indexes (CI) of the cellulose samples were measured according to the amorphous subtraction method described by Park and colleagues.41 Briefly, a diffractogram of a 36-hour ball-milled cotton linter cellulose sample mentioned above was subtracted from the other cellulose samples to remove the influence of the amorphous component in the diffractograms. The ratio of the integrated area of each subtracted diffractogram to the area of the original was then calculated and multiplied by 100 to give the CI value of the sample. Darkfield microscopy and image analysis Darkfield images were captured on a Nikon Eclipse E800 microscope with an attached RTKE color SPOT camera and SPOT software (Diagnostic Instruments, Inc., Sterling Heights, MI). Full-frame resolution (1600 × 1200) with a 20X 0.75 NA Plan Apo objective lens is 0.367 µm/pixel. Particle size and aspect ratios were measured by image processing and analysis on binary images using FIJI42 (http://fiji.sc/Fiji) as described previously.43


27


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 28 of 37

Autosampler pyrolysis-molecular beam mass spectroscopy (py-MBMS) 4.0 mg of each cellulose sample was loaded into four separate reactor cups. Pyrolysis was performed using a single shot micro-furnace pyrolyzer equipped with an autosampler. The reactor cups were dropped one at a time into a micro pyrolyzer, providing rapid heating. The micro pyrolyzer consists of a quartz pyrolysis tube surrounded by a tubular furnace to provide uniform heating throughout. The furnace was calibrated to read the centerline temperature of the pyrolysis tube. Continuous 3.0 SLM flow of helium gas was maintained through the pyrolysis tube throughout the experiments providing the inert atmosphere for pyrolysis while sweeping and diluting the emerging products out of the reactor as they were produced. The micropyrolyzer assembly was directly connected to a molecular beam mass spectrometer (MBMS) for on-line measurement of the vapors. In the MBMS system the gas sample undergoes rapid, adiabatic expansion through a 250 µm orifice into a vacuum chamber held at ~100 mtorr.30,32 This expansion cools the gas and effectively arrests the chemistry occurring in the reactor and improves the sensitivity of the instrument. Further, cooling the gas helps minimize fragmentation during ionization and improves mass spectral signals. The cooled gas is skimmed into a molecular beam as it is drawn into another vacuum stage. This molecular beam is then ionized with an electron impact ionization source (22.5 eV) of a quadrapole mass spectrometer, yielding positive ions. The scan rate is set at about 1.0 ms/amu. The MBMS is set to measure over an m/z range of 10-500. Multivariate analysis of the MBMS spectra Multivariate analysis was used to identify groups of mass spectral peaks that were correlated in the product vapors. We used the multivariate curve resolution optimized by alternating least


28

Page 29 of 37

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


squares (MCR-ALS) routine available in the statistical analysis software package The Unscrambler (Camo Software AS, version 9.7).30,32,35 Multivariate curve resolution (MCR) resolves the principal component analysis results into mathematically constructed components which have derived sub-spectra that are used to partition the original variance of the data set into estimates of the concentrations of the components. This allows determination of elution profiles of the components in an unresolved mixture of two or more constituents, assuming the data has enough degrees of freedom to identify the separate sources of variance. The Unscrambler MCR algorithm is based on pure variable selection from principle component analysis (PCA) loadings to find the initial estimation of spectral profiles, and then alternating least squares (ALS) to optimize resolved spectral and concentration profiles. We included the constraints for producing non-negative concentration profiles and non-negative mass spectra. Autosampler pyrolysis-gas chromatography mass spectroscopy (py-GCMS) Py-GCMS analysis was performed to measure the product distribution of different allomorphs of cellulose and complement the results obtained from py-MBMS.30 A tandem micropyrolyzer (Rx-3050TR, Frontier Laboratories, Japan) equipped with an autosampler (AS-1020E) and microjet cryo-trap (MJT-1030Ex) coupled to gas chromatograph (7890B, Agilent Technologies, USA) interfaced with MS (5977A, Agilent Technologies, USA) and a flame ionization detector (FID) was used. 500 µg of cellulose were charged into deactivated stainless steel cups and loaded into the autosampler. This mass of cellulose has been shown in previous studies to be well below the sample mass that becomes mass transfer limited.44 The cups were plunged into the pyrolysis zone maintained at 500°C. The pyrolysis vapors were captured using a liquid nitrogen trap (80°C, held inside the GC oven) and released into the inlet of the gas chromatograph. The GC was equipped with a dean switch coupled to a two-way splitter, which directed products into


29


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 30 of 37

three columns. This system allowed non-condensable vapors and gases (boiling points below 80°C) to go through the first column (GS-GASPRO, Frontier Laboratories, Japan) and then to a TCD. The condensable vapors underwent separation through a two-way splitter sending products into two capillary columns (Ultra Alloy-5, Frontier Laboratories, Japan) with a 5 % diphenyl and 95 % dimethylpolysiloxane stationary phase. One column directed products into the FID and the other column directed products into the MS. The oven was programmed to hold at 40°C for 4.5 min followed by heating to 300°C at the ramp rate of 20°C min-1. The separated pyrolysis vapors were identified using the NIST GCMS library (http://chemdata.nist.gov). The semiquantification of products was performed from areas as detected by FID. Thermogravimetric analysis The mass loss during pyrolysis was evaluated by heating the samples in nitrogen in a TA Instruments Q500 Analyzer (TA Instruments, New Castle, DE). The samples were heated to 250°C at 20°C/min and further to 450°C at 10°C/min, and the mass losses were recorded. Real-time hot-stage microscopy In-situ microscopy of cellulose samples at elevated temperatures was performed with an Instec HCS621G heating stage equipped with a SCT200 temperature controller and WinTemp control software (Instec, Inc., Boulder, CO), manufacturer-modified to provide temperatures in the range of -190°C to approximately 700°C. The stage includes a 3 mm light aperture for image acquisition and gas-tight seals to prevent penetration of ambient gases. Cellulose samples were sandwiched between two 19 mm diameter glass coverslips and secured on the silver heat block with a thin-bored piece of aluminum and screws connected directly into the heating block. The reaction chamber was purged with N2 before heating and maintained at a constant flow during


30

Page 31 of 37

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


heating. For all hot-stage experiments, the heating block heating rate was approximately 200°C/min. Instec-supplied, NREL-modified SPOT software camera control macros were used to capture image sequences with the RT KE color SPOT camera and record time and temperature from the hot-stage temperature controller via WinTemp software. Images were captured at a rate of approximately 1 frame every three seconds. Image sequences were exported to FIJI42 (http://fiji.sc/Fiji) for segmentation and image analysis to measure the area occupied by the cellulose particles. Scanning electron microscopy Cotton linter cellulose allomorph and pyrolysis char samples were mounted on aluminum stubs with conductive carbon adhesive and sputter coated with 10 nm of gold prior to imaging. Imaging was performed with a FEI Quanta 400 FEG instrument under high vacuum conditions using an accelerating voltage of 15 keV. For full instructions, please see the journal’s Instructions for Authors. Depending on the journal, the manuscript may include sections such as an introduction, experimental details (sections titled Experimental Methods, Experimental Section, or Materials and Methods), theoretical basis (sections titled Theoretical Basis or Theoretical Calculations), results, discussion, and conclusions. ASSOCIATED CONTENT Supporting Information. Table S1. Compositional analysis. Chemical composition analysis of all the allomorphs of cellulose-I suggests that no degradation of cellulose occurred during their preparation.


31


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

Sample

Arabinan Galactan

Glucan

Xylan

Ash

Page 32 of 37

Fluka CL Cel I

0.0%

0.0%

99.0 ± 0.3%

0.8%

0.1%

Fluka CL Cel II 145ºC

0.0%

0.0%

99.8 ± 0.8%

0.0%

0.2%

Fluka CL Cel III 130ºC

0.0%

0.0%

99.1 ± 01.8%

0.7%

0.1%

Table S2. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) analysis reported as PPM (µg/g) of the cellulose samples. If the element was below the instrument detection limit, it is indicated by BDL. ICP-AES analysis suggests that the ash content of the samples is insufficient to have a major influence on their pyrolysis behavior.45

Sample

CI-H

CI-M

CI-L

CII-H

CII-M

CII-L

CIII-H

CIIIM

Dilution

98.0

251.0

413.0

148.0

185.0

165.0

176.0

130.0

171.0

862.0

Al

7.7 ± 0.08

18.6 ± 0.11

16.1 ± 0.11

12.6 ± 0.14

10.5 ± 0.20

6.0 ± 0.06

5.0 ± 0.15

8.9 ± 0.06

5.6 ± 0.06

25.8 ± 0.76

Ca

21.5 ± 0.11

56.7 ± 0.66

69.6 ± 0.20

85.2 ± 1.24

159.5 ± 1.61

28.1 ± 0.29

17.0 ± 0.06

16.1 ± 0.03

14.3 ± 0.09

167.9 ± 1.70

Fe

2.4 ± 0.10

64.4 ± 0.57

10.3 ± 0.94

32.4 ± 0.36

24.4 ± 0.33

25.5 ± 0.20

27.2 ± 0.32

9.1 ± 0.11

13.5 ± 0.04

321.1 ± 3.32

K

BDL

BDL

BDL

BDL

BDL

BDL

BDL

BDL

BDL

BDL

Mg

5.9 ± 0.09

16.0 ± 0.20

27.4 ± 0.02

20.8 ± 0.31

30.6 ± 0.46

6.6 ± 0.03

5.3 ± 0.04

4.9 ± 0.02

5.1 ± 0.04

36.3 ± 0.24

Na

28.1 ± 0.12

50.6 ± 1.80

109.3 ± 1.30

251.3 ± 2.85

510.3 ± 3.05

33.7 ± 0.33

26.2 ± 0.82

35.7 ± 0.37

34.8 ± 0.43

87.4 ± 7.93

P

BDL

BDL

BDL

BDL

BDL

BDL

BDL

BDL

BDL

BDL

S

32.9 ± 0.94

BDL

23.5 ± 1.63

29.2 ± 1.45

22.8 ± 2.63

30.2 ± 1.83

37.8 ± 0.68

34.7 ± 1.71

86.6 ± 2.66

BDL

CIII-L

Amorph

AUTHOR INFORMATION Corresponding Author


32

Page 33 of 37

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


Correspondence should be addressed to [email protected], or [email protected] Author Contributions ‡These authors contributed equally to this work. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. ACKNOWLEDGMENT The cellulose allomorph generation, py-MBMS, py-GCMS, and TGA work presented here was supported by the U.S. Department of Energy, Bioenergy Technologies Office (DOE-BETO) under contract number DE-AC36-08GO28308 with the National Renewable Energy Laboratory (NREL). The imaging and image analysis of cellulose particles and chars was supported as part of the Center for Direct Catalytic Conversion of Biomass to Biofuels (C3Bio), an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Award Number DE-SC0000997. The authors would also like to thank Steve Deutch for preforming ICP analysis and David Robichaud, Mark Jarvis, Gregg Beckham and Mike Crowley for helpful discussions. REFERENCES (1) Mettler, M. S.; Vlachos, D. G.; Dauenhauer, P. J. Top ten fundamental challenges of biomass pyrolysis for biofuels. Energ Environ Sci 2012, 5 (7), 7797. (2) Kazi, F. K.; Fortman, J. A.; Anex, R. P.; Hsu, D. D.; Aden, A.; Dutta, A.; Kothandaraman, G. Techno-economic comparison of process technologies for biochemical ethanol production from corn stover. Fuel 2010, 89, S20. (3) Bridgwater, A. V. Review of fast pyrolysis of biomass and product upgrading. Biomass Bioenerg 2012, 38, 68. (4) Mettler, M. S.; Vlachos, D. G.; Dauenhauer, P. J. Top ten fundamental challenges of biomass pyrolysis for biofuels. Energy & Environmental Science 2012, 5 (7), 7797.


33


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

Page 34 of 37

(5) Kersten, S.; Garcia-Perez, M. Recent developments in fast pyrolysis of ligno-cellulosic materials. Curr Opin Biotech 2013, 24 (3), 414. (6) Czernik, S.; Bridgwater, A. V. Overview of applications of biomass fast pyrolysis oil. Energ Fuel 2004, 18 (2), 590. (7) Bridgwater, A. V. Review of fast pyrolysis of biomass and product upgrading. Biomass and Bioenergy 2012, 38 (C), 68. (8) Haas, T. J.; Nimlos, M. R.; Donohoe, B. S. Real-Time and Post-reaction Microscopic Structural Analysis of Biomass Undergoing Pyrolysis. Energy & Fuels 2009, 23 (7), 3810. (9) Mettler, M. S.; Paulsen, A. D.; Vlachos, D. G.; Dauenhauer, P. J. The chain length effect in pyrolysis: bridging the gap between glucose and cellulose. Green Chem 2012, 14 (5), 1284. (10) Dufour, A.; Castro-Díaz, M.; Marchal, P.; Brosse, N.; Olcese, R.; Bouroukba, M.; Snape, C. In Situ Analysis of Biomass Pyrolysis by High Temperature Rheology in Relations with1H NMR. Energy & Fuels 2012, 26 (10), 6432. (11) 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. Energ Environ Sci 2011, 4 (10), 4306. (12) Jarvis, M. W.; Haas, T. J.; Donohoe, B. S.; Daily, J. W.; Gaston, K. R.; Frederick, W. J.; Nimlos, M. R. Elucidation of Biomass Pyrolysis Products Using a Laminar Entrained Flow Reactor and Char Particle Imaging. Energy & Fuels 2011, 25, 324. (13) Wang, Z. H.; McDonald, A. G.; Westerhof, R. J. M.; Kersten, S. R. A.; Cuba-Torres, C. M.; Ha, S.; Pecha, B.; Garcia-Perez, M. Effect of cellulose crystallinity on the formation of a liquid intermediate and on product distribution during pyrolysis. Journal of Analytical and Applied Pyrolysis 2013, 100, 56. (14) Wang, Z.; Pecha, B.; Westerhof, R. J. M.; Kersten, S. R. A.; Li, C. Z.; McDonald, A. G.; Garcia-Perez, M. Effect of Cellulose Crystallinity on Solid/Liquid Phase Reactions Responsible for the Formation of Carbonaceous Residues during Pyrolysis. Ind Eng Chem Res 2014, 53 (8), 2940. (15) Nishiyama, Y.; Langan, P.; Chanzy, H. Crystal structure and hydrogen-bonding system in cellulose 1 beta from synchrotron X-ray and neutron fiber diffraction. J Am Chem Soc 2002, 124 (31), 9074. (16) Nishiyama, Y.; Sugiyama, J.; Chanzy, H.; Langan, P. Crystal structure and hydrogen bonding system in cellulose I(alpha) from synchrotron X-ray and neutron fiber diffraction. J Am Chem Soc 2003, 125 (47), 14300. (17) Atalla, R. H.; Ellis, J. D.; Schroeder, L. R. Some Effects of Elevated-Temperatures on the Structure of Cellulose and Its Transformation. J Wood Chem Technol 1984, 4 (4), 465. (18) Mittal, A.; Katahira, R.; Himmel, M. E.; Johnson, D. K. Effects of alkaline or liquidammonia treatment on crystalline cellulose: changes in crystalline structure and effects on enzymatic digestibility. Biotechnol Biofuels 2011, 4, 41. (19) Atalla, R. H.; Vanderhart, D. L. Native cellulose: a composite of two distinct crystalline forms. Science 1984, 223 (4633), 283. (20) Harris, D. M.; Corbin, K.; Wang, T.; Gutierrez, R.; Bertolo, A. L.; Petti, C.; Smilgies, D. M.; Estevez, J. M.; Bonetta, D.; Urbanowicz, B. R.et al. Cellulose microfibril crystallinity is reduced by mutating C-terminal transmembrane region residues CESA1A903V and CESA3T942I of cellulose synthase. Proc Natl Acad Sci U S A 2012, 109 (11), 4098.


34

Page 35 of 37

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


(21) Beckham, G. T.; Matthews, J. F.; Peters, B.; Bomble, Y. J.; Himmel, M. E.; Crowley, M. F. Molecular-level origins of biomass recalcitrance: decrystallization free energies for four common cellulose polymorphs. J Phys Chem B 2011, 115 (14), 4118. (22) Seshadri, V.; Westmoreland, P. R. Concerted Reactions and Mechanism of Glucose Pyrolysis and Implications for Cellulose Kinetics. J Phys Chem A 2012, 116 (49), 11997. (23) Antal, M. J. In Advances in Solar Energy; Boer, K. W.;Duffle, J. A., Eds.; American Solar Energy Society: New York, 1983. (24) Antal, M. J.; Varhegyi, G. Cellulose Pyrolysis Kinetics - the Current State Knowledge. Ind Eng Chem Res 1995, 34 (3), 703. (25) Ponder, G. R.; Richards, G. N. A Review of Some Recent Studies on Mechanisms of Pyrolysis of Polysaccharides. Biomass Bioenerg 1994, 7 (1-6), 1. (26) Calonaci, M.; Grana, R.; Hemings, E. B.; Bozzano, G.; Dente, M.; Ranzi, E. Comprehensive Kinetic Modeling Study of Bio-oil Formation from Fast Pyrolysis of Biomass. Energ Fuel 2010, 24, 5727. (27) Dupont, C.; Chen, L.; Cances, J.; Commandre, J. M.; Cuoci, A.; Pierucci, S.; Ranzi, E. Biomass pyrolysis: Kinetic modelling and experimental validation under high temperature and flash heating rate conditions. J Anal Appl Pyrol 2009, 85 (1-2), 260. (28) Ranzi, E.; Cuoci, A.; Faravelli, T.; Frassoldati, A.; Migliavacca, G.; Pierucci, S.; Sommariva, S. Chemical Kinetics of Biomass Pyrolysis. Energ Fuel 2008, 22 (6), 4292. (29) Sluiter, J. B.; Ruiz, R. O.; Scarlata, C. J.; Sluiter, A. D.; Templeton, D. W. Compositional analysis of lignocellulosic feedstocks. 1. Review and description of methods. J Agric Food Chem 2010, 58 (16), 9043. (30) Mukarakate, C.; Zhang, X. D.; Stanton, A. R.; Robichaud, D. J.; Ciesielski, P. N.; Malhotra, K.; Donohoe, B. S.; Gjersing, E.; Evans, R. J.; Heroux, D. S.et al. Real-time monitoring of the deactivation of HZSM-5 during upgrading of pine pyrolysis vapors. Green Chem 2014, 16 (3), 1444. (31) Evans, R. J.; Milne, T. A. Molecular Characterization of the Pyrolysis of Biomass .2. Applications. Energy & Fuels 1987, 1 (4), 311. (32) Evans, R. J.; Milne, T. A. Molecular Characterization of the Pyrolysis of Biomass .1. Fundamentals. Energy & Fuels 1987, 1 (2), 123. (33) Shin, E. J.; Nimlos, M. R.; Evans, R. J. Kinetic analysis of the gas-phase pyrolysis of carbohydrates. Fuel 2001, 80 (12), 1697. (34) Evans, R. J.; Milne, T. A.; Soltys, M. N.; Schulten, H. R. Mass-Spectrometric Behavior of Levoglucosan under Different Ionization Conditions and Implications for Studies of Cellulose Pyrolysis. Journal of Analytical and Applied Pyrolysis 1984, 6 (3), 273. (35) Mukarakate, C.; Watson, M. J.; ten Dam, J.; Baucherel, X.; Budhi, S.; Yung, M. M.; Ben, H.; Iisa, K.; Baldwin, R. M.; Nimlos, M. R. Upgrading biomass pyrolysis vapors over [small beta]-zeolites: role of silica-to-alumina ratio. Green Chem 2014, 16 (12), 4891. (36) Patwardhan, P. R.; Satrio, J. A.; Brown, R. C.; Shanks, B. H. Product distribution from fast pyrolysis of glucose-based carbohydrates. Journal of Analytical and Applied Pyrolysis 2009, 86 (2), 323. (37) Hall, M.; Bansal, P.; Lee, J. H.; Realff, M. J.; Bommarius, A. S. Cellulose crystallinity--a key predictor of the enzymatic hydrolysis rate. FEBS J 2010, 277 (6), 1571. (38) Atalla, R. H.; Nagel, S. C. Annealing and Increased Order in Cellulose Ii. Journal of Polymer Science Part C-Polymer Letters 1974, 12 (10), 565.


35


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

Page 36 of 37

(39) Mittal, A.; Katahira, R.; Himmel, M. E.; Johnson, D. K. Effects of alkaline or liquidammonia treatment on crystalline cellulose: changes in crystalline structure and effects on enzymatic digestibility. Biotechnol Biofuels 2011, 4. (40) Goldberg, R. N.; Schliesser, J.; Mittal, A.; Decker, S. R.; Santos, A. F. L. O. M.; Freitas, V. L. S.; Urbas, A.; Lang, B. E.; Heiss, C.; da Silva, M. D. M. C. R.et al. A thermodynamic investigation of the cellulose allomorphs: Cellulose(am), cellulose II(cr), cellulose II(cr), and cellulose III(cr). The Journal Of Chemical Thermodynamics 2015, 81 (C), 184. (41) Park, S.; Baker, J. O.; Himmel, M. E.; Parilla, P. A.; Johnson, D. K. Cellulose crystallinity index: measurement techniques and their impact on interpreting cellulase performance. Biotechnol Biofuels 2010, 3 (1), 10. (42) Schindelin, J.; Arganda-Carreras, I.; Frise, E.; Kaynig, V.; Longair, M.; Pietzsch, T.; Preibisch, S.; Rueden, C.; Saalfeld, S.; Schmid, B.et al. Fiji: an open-source platform for biological-image analysis. Nat Methods 2012, 9 (7), 676. (43) Ciesielski, P. N.; Wang, W.; Chen, X.; Vinzant, T. B.; Tucker, M. P.; Decker, S. R.; Himmel, M. E.; Johnson, D. K.; Donohoe, B. S. Effect of mechanical disruption on the effectiveness of three reactors used for dilute acid pretreatment of corn stover Part 2: morphological and structural substrate analysis. Biotechnol Biofuels 2014, 7, 47. (44) Zhang, J.; Nolte, M. W.; Shanks, B. H. Investigation of Primary Reactions and Secondary Effects from the Pyrolysis of Different Celluloses. Acs Sustain Chem Eng 2014, 2 (12), 2820. (45) Patwardhan, P. R.; Satrio, J. A.; Brown, R. C.; Shanks, B. H. Influence of inorganic salts on the primary pyrolysis products of cellulose. Bioresource Technol 2010, 101 (12), 4646.


36

Page 37 of 37

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


FOR TABLE OF CONTENTS USE ONLY The influence of crystal allomorph and crystallinity on the products and behavior of cellulose during fast pyrolysis Calvin Mukarakate, Ashutosh Mital, Peter N. Ciesielski, Sridhar Budhi, Logan Thompson, Kristiina Iisa, Mark R. Nimlos, and Bryon S. Donohoe

Cellulose is responsible for maintaining the structural and mechanical integrity of plant cell walls. Both the crystal allomorph and relative crystallinity of cellulose impacts the slate of primary products produced by fast pyrolysis. This work suggests that biomass with altered cellulose structure could help overcome barriers for deployment of fast pyrolysis of biomass.


37

Influence of Crystal Allomorph and Crystallinity on the Products and

Recommend Documents