Multi-technique characterization of biochar formation

Engineering Conferences International ECI Digital Archives Biochar: Production, Characterization and Applications Proceedings 8-20-2017 Multi-technique characterization of biochar formation Anthony Dufour CNRS Nancy, France Yann Le Brech CNRS Nancy, France Luc Delmotte CNRS Mulhouse, France Jésus Raya CNRS Strasbourg, France Nicolas Brosse University of Lorraine, France See next page for additional authors Follow this and additional works at: http://dc.engconfintl.org/biochar Part of the Engineering Commons Recommended Citation Anthony Dufour, Yann Le Brech, Luc Delmotte, Jésus Raya, Nicolas Brosse, Guillain Mauviel, and Roger Gadiou, "Multi-technique characterization of biochar formation" in "Biochar: Production, Characterization and Applications", Franco Berruti, Western University, London, Ontario, Canada Raffaella Ocone, Heriot-Watt University, Edinburgh, UK Ondrej Masek, University of Edinburgh, Edinburgh, UK Eds, ECI Symposium Series, (2017). http://dc.engconfintl.org/biochar/50 This Abstract and Presentation is brought to you for free and open access by the Proceedings at ECI Digital Archives. It has been accepted for inclusion in Biochar: Production, Characterization and Applications by an authorized administrator of ECI Digital Archives. For more information, please contact franco@bepress.com.

Authors Anthony Dufour, Yann Le Brech, Luc Delmotte, Jésus Raya, Nicolas Brosse, Guillain Mauviel, and Roger Gadiou This abstract and presentation is available at ECI Digital Archives: http://dc.engconfintl.org/biochar/50

MULTI-TECHNIQUE CHARACTERIZATION OF BIOCHAR FORMATION A. Dufour a*, Y. Le Brech a, M. Castro-Diaz b, L. Delmotte c, J. Raya d, R. Gadiou c, S. Leclerc a, N. Brosse e, P. Marchal a, G. Mauviel a, C. Snape b a CNRS, Univ. Lorraine, Nancy, France. b Univ. Nottingham, UK. c CNRS, Univ. Hte. Alsace, Mulhouse, Fr. d CNRS, Univ. Strasbourg, Fr. e Univ. Lorraine, Nancy, Fr. * anthony.dufour@univ-lorraine.fr Alba, ECI Biochar, 22/08/2017

The physical and chemical mechanisms of char formation are complex Molecular network of polymers and minerals Intermediate material Hundreds of volatiles Char

Prof. Manuel Garcia-Perez has pointed out an important question during his Key Note of this conference: «We need to adress the question why biochar conserves the macro-structure of biomass while many components soften?» This talk deals with this question.

Biochar is formed through an intermediate soft material. Wood char particles, slow pyrolysis, from mm to μm: bubbles formation at μm length scale (Chem. Eng. Res. Des., 89 (10), 2136, 2011) Biomass components (notably lignin and hemicelluloses) soften. Klason, 1901; Göring, 1963; Sharma, 2014

Important to understand how the soft material impacts char formation. Molecular network of polymers and minerals Intermediate soft material Hundreds of volatiles Char In-situ analysis of interest to understand the intermediate, labile and soft material In-situ 1 H NMR and rheology have been extensively used for understanding metaplast during coal pyrolysis (Sato, 1979, Lynch, 1988, Castro-Diaz, 2005) but not yet for biomass.

Basics of in-situ rheology Pyrolysis is conducted inside the rheometer. Applied normal force Transducer Measured stress and phase angle between stress and strain ARES G2 rheometer N 2, 5K/min 2 parallel plates Biomass pellet (~2mm) Displacement on vertical axis Applied strain amplitude at an angular oscillation Motor Determination of viscous and elastic moduli based on phase angle Swelling and shrinking of particles based on the displacement between plates (results not shown)

In-situ rheology during cellulose pyrolysis: it stays mainly hard & elastic (under slow cond.) 1.2E+06 Elastic modulus (G') Increase in viscous behaviours 100 80 G' & G'' (Pa) 8.0E+05 4.0E+05 60 40 Mass loss (%) Max. 20 Viscous mod. (G'') mass 0.0E+00 loss rate 0 150 200 250 300 350 400 5K/min Temperature ( C) Energy & Fuels, 26 (10), 6432, 2012.

Lignin presents very different behaviours than cellulose. G' & G'' (Pa) 1.0E+09 1.0E+07 1.0E+05 Glass transition Volatiles Elastic modulus (G') Softening Formation of a 100 carbon foam by volatiles pushing the visco-elastic material. 80 Solidification of char Viscous mod.(g'') 1.0E+03 50 150 250 350 60 40 20 0 Mass loss (%) Temperature ( C)

Comparison between cellulose, xylan, lignin and miscanthus for elastic modulus evolution 1.0E+09 Elastic modulus (G', Pa) 1.0E+07 1.0E+05 Miscanthus Cellulose Xylan Lignin 1.0E+03 150 200 250 300 350 400 Temperature ( C) Lignin softening is not seen in the native network of biomass while G is 3 orders of magnitude lower & then higher than cellulose.

Simplified scheme of visco-elastic properties of fractionated polymers and native biomass Energy & Fuels, 26 (10), 6432, 2012.

This finding explains why char globally keeps the same macro-structure of biomass cells (at slow pyrolysis) although forming an intermediate soft material. Cellulose maintains the overall structure of macro-pores under slow conditions.

Basics of in-situ 1 H NMR Pyrolysis is conducted inside a specific NMR probe (heated up to 500 C). In-situ 1 H NMR analyses protons transverse relaxation ( 2 ) signal as a function of temperature = Proton magnetic resonance thermal analysis (PMRTA) (Lynch, 1988, Sakurov, 1993, 2004, Castro-Diaz, 2005) Protons in rigid structures = strong magnetic coupling = short relaxation time Protons in mobile structures like liquids = weaker magnetic interactions = long relaxation time

1 H NMR spectra were deconvoluted into Gaussian (solid-like) and Lorentzian (liquid-like) distribution functions Fraction of mobile protons (% H L ) (or of fluid phase ) calculated as: % H L = Lorentzian area Lorentzian + Gaussian areas Intensity (arbitrary unit) 4.0E+08 3.0E+08 2.0E+08 1.0E+08 Empirical Calculated (Gaus. + Lor. comp.) Gaus. comp. Lor. comp. % H L = 8.8% Intensity (arbitrary unit) 5.0E+08 4.0E+08 3.0E+08 2.0E+08 1.0E+08 % H L = 43.0% 0.0E+00-4.0E+04-2.0E+04 0.0E+00 2.0E+04 4.0E+04 Frequency (Hz) Miscanthus at 200 C 0.0E+00-4.0E+04-2.0E+04 0.0E+00 2.0E+04 4.0E+04 Frequency (Hz) Miscanthus at 300 C Fluid phase increases from 200 to 300 C ChemSusChem, 5, 1258, 2012

Comparison between 1 H NMR (% of mobile protons) and rheology (elastic modulus) 1.0E+10 400 C Lignin Elastic modulus (G', Pa) 1.0E+08 1.0E+06 1.0E+04 325 C Resolidification 100 C 200 C 275 C Xylan Softening 200 C During resolidification, molecular mobility is still present in a solid-like material (inside cages). 1.0E+02-10 10 30 50 70 90 110 % Fluid H Energy & Fuels, 26 (10), 6432, 2012.

Comparison between 1 H NMR (% of mobile protons) and rheology (elastic modulus) Elastic modulus (G', Pa) 1.6E+06 1.1E+06 6.0E+05 Miscanthus 375 C 250 C 150 C 400 C Mobility in an elastic material 1.0E+05-10 10 30 50 % Fluid H 350 C 350 C 325 C 300 C Cellulose Mobility of cellulose is developed in an elastic solid-like material ( cavities Mamleev, JAAP, 2009). Miscanthus very complex Other complementary methods need to be used.

In-situ analysis were completed by ex-situ 13 C NMR analysis of char for characterization of the chemical moieties in char. Same biomass pyrolyzed in an analytical fixed bed reactor Good control of mass transfers, similar mass loss than in TGA (mg) = chemical regime

Cross Polarisation/Magic Angle Spinning (CP/MAS) method is fast but not quantitative because of different magnetisation transfer rates from 1 H to 13 C depending on 13 C nuclei environments. Direct polarisation (DP) is quantitative but very long (~ 10 days for one spectrum). A specific CP/MAS method has been found to give similar spectra than DP/MAS but 50 times faster. + 2D 1 H- 13 C solid state NMR at 750MHz Anal. Chem., 87 (2), 843, 2015 Carbon, 108, 165, 2016

Quantitative evolution of main chemical moieties in Miscanthus chars (C molar balance) Formation of aromatic carbons mainly from holocellulose (for this biomass) Carbon, 108, 165, 2016

An exemple on the combination of various analytical methods for lignin char formation TG Mass loss DSC Glass transition, heats of reactions 13 C and 31 P NMR Evolution of chemical moieties In-situ rheology Swelling/shrinking, softening/solidification In-situ 1 H NMR Mobility GPC-UV LDI FT ICR MS Molecular weight distribution of major mobile species Analysis of some mobile species ACS Sust. Chem. Eng. 5 (8), 6940, 2017

An exemple on the combination of various analytical methods on lignin char Soft material Solid fused char Lignin All mobile species by in-situ 1 H NMR Main mobile species by THF extract. then GPC Smaller mobile species by LDIMS Crosslinked structure Heating

Physical-chemical mechanism of lignin char formation proposed based on these methods Swelling Softening Solidification Tg 135 C Increase of condensed aromatic C Decrease of protonated & oxygenated aromatic C Decrease of side chain carbons Start of Fluidity at 140 C 100% Fluidity Significant mass loss Increase in C and decrease in O number 0 50 100 150 200 250 300 350 400 Temperature ( C) ACS Sust. Chem. Eng. 5 (8), 6940, 2017

To conclude, various techniques are needed to understand char formation. We want to share our devices You are welcome in Nancy! Our coffee break The Stanislas square

Supplementary slides

Biomass pyrolysis produces an intermediate viscoelastic material evidenced by microscopic analysis Xylan and lignin form a viscoelastic material at slow and fast heating rates (Fisher et al., 2002) Cellulose before pyrolysis After pyrolysis Cellulose produces a viscous material at high heating rates (Boutin et al., 1998)

Structural evolution of biomass material undergoing pyrolysis analysed by microscopic analysis. Arrows show pyrolysis product coating the internal cell wall of a wood fiber (Haas et al., En&Fuels, 2009)

The mechanisms of the visco-elastic material formation from biomass are not yet understood This intermediate visco-elastic material is very important for: mass transfer mechanisms (Jarvis, 2011, Dufour, 2011) chemical mechanisms (Lédé, 2002, Castro-Diaz, 2005) and the selectivity of thermo-chemical processes In-situ 1 H NMR and rheology have been extensively used for understanding metaplast formation in coal (Sato, 1979, Lynch, 1988, Castro-Diaz, 2005) but not yet for biomass

Experiments was conducted on polymers carefully extracted from biomass (miscanthus) Miscanthus EtOH/Tol (1/2) 24h EtOH 24h Extractivefree Misc. HCl [5M]/ HF [5M] De-ashed Miscanthus Miscanthus Ethanol organosolv Lignin Holocelluloses NaClO 2 / AcOH 70 C, 2x 2h HCl [2,5M] 100 C, 4h Elemental and mineral analysis of all samples given in Dufour et al., ChemSusChem, 2012 Cellulose

It is important to extensively analyse: cellulose composition and structure: cristallinity index (by XRD), degree of polymerization (DP by GPC), ash content (Dufour et al. ChemSusChem, in press) lignin structural compositions by 13 C, 31 P NMR, FT-IR spectroscopies and GPC, ashes, etc. (El Hage, 2010) Lignin extracted by the organosolv process is close to the native one (El Hage, 2010)

Mass loss of samples (5K/min, ~1mg of sample, TG-DSC 111, Setaram France) 100 Lignin Mass fraction (%) (mass / initial mass) 80 60 40 20 Miscanthus Xylan Cellulose 0 150 250 350 450 Temperature ( C) Temperature of maximum mass loss rate: xylan 275 C, cellulose 340 C, lignin - 380 C.

Basics of mechanical spectroscopy Normal force (200g) 2 parallel plates Transducer Measured stress ( ) and phase angle ( ) between stress and strain 0 sin ( t+ ) 0 (max. amplitude) time Biomass pellet (~2mm) Motor Applied strain amplitude ( =0.1%) at an angular oscillation ( =1Hz) 0 sin ( t) 0 (max. amplitude) time G = cos ( ) 0 / 0 G = sin ( ) 0 / 0 tan( ) = G /G G (Pa) = elastic modulus proportional to the mechanical elastic energy stored and reversibly recovered G (Pa) = viscous modulus proportional to the mechanical energy irreversibly lose through viscous dissipation tan δ > 1, mainly viscous; tan δ < 1, mainly elastic

Rheological signature of xylan 1.0E+08 100 Elastic mod. (G') 1.0E+07 80 G' & G'' (Pa) 1.0E+06 Softening 60 40 Mass loss(%) 1.0E+05 Viscous modulus (G'') Fast resolidification 20 1.0E+04 150 250 350 0 Temperature ( C)

Rheological signature of miscanthus 100 G' & G'' (Pa) 1.6E+06 1.1E+06 Elastic mod. (G') Xylan 5.5E+05 Cellulose 20 Viscous mod.(g'') 5.0E+04 0 150 200 250 300 350 400 Temperature ( C) 80 60 40 Mass loss (%) Lignin softening is not seen in the native network consistent with 1 H NMR

Comparison between cellulose, xylan, lignin and miscanthus for tan( ) evolution 2 tan( ) 1 Lignin Xylan Cellulose Mainly VISCOUS Mainly ELASTIC Miscanthus 0 150 200 250 300 350 400 Temperature ( C)

Simplified scheme of visco-elastic properties of fractionated polymers and native biomass

Contents Background: importance of pyrolysis and intermediate liquid phase Polymers extractions and composition Basics and new insights from in-situ high temperature 1 H NMR Basics and new insights from in-situ high temperature rheometry Comparison of 1 H NMR and rheometry results

In-situ 1 H NMR analyses protons transverse relaxation signal as a function of temperature = Proton magnetic resonance thermal analysis (PMRTA) (Lynch, 1988, Sakurov, 1993, 2004, Castro-Diaz, 2005) Protons in rigid structures = strong magnetic coupling = short relaxation time Protons in mobile structures like liquids = weaker magnetic interactions = long relaxation time

Mobile protons are very important for pyrolysis mechanisms Mobile protons stabilise the free radicals formed when bridges break and can control the composition of the pyrolysis products as tars (Solomon, 1993) Interactions between cellulose and lignin (Hosoya, 2009) could act through the mobile H transfer from H donors intermediate products to H acceptors

Very different fluid phase development is analysed between cellulose, lignin, xylan and miscanthus 100 Lignin 80 % Fluid H 60 40 Miscanthus Xylan 20 Cellulose 0 150 200 250 300 350 400 Dufour et al., ChemSusChem, 2012 50mg, 5K/min Temperature ( C)

In synthetic blend, fluid phase is developed at a lower temperature (150 C) than in miscanthus (200 C) and to a higher max. fluid H fraction. 100 80 Synthetic blend %H L max De-ashed miscanthus % Fluid H 60 40 Miscanthus 20 0 150 200 250 300 350 400 Dufour et al., ChemSusChem, 2012 Temperature ( C) Same finding for synthetic holocelluloses and true holocelluloses (not shown).

Simplified scheme of fractionated polymers mobility upon pyrolysis Dufour et al., ChemSusChem, 2012

Mobility of the native network is shifted and reduced due to links between lignin/xylan and xylan/cellulose Cellulose maintains the rigidity of the native network

Extraction of lignin, holocellulose and cellulose from Miscanthus Aromatic clusters mainly formed from holocellulose (stable C yield in aromatics from lignin). No apparent interaction on the formation of aromatic C yields between native network and separated polymers.

2D HETCOR 1H-13C NMR of intermediate chars

Fixed bed, 1 H NMR and DSC under same mass transfer conditions (fixed bed, same carrier gas velocity)

1 H NMR in-situ method 1H NMR transverse relaxation signal was stimulated by a solid echo pulse sequence (90 -T-90 ). The rate of decrease in intensity of 1H NMR signal depends on the strength of the magnetic coupling. 1H in solid/rigid > strong static coupling > short life 1H NMR signal (Gaussian) 1H in mobile/liquid-like > weaker static magnetic interactions > longer life signal (Lorentzian) The Fourier-transformed time domain decay of the solid echo pulse deconvoluted in Gaussian and Lorentzian.

Principe de l analyse RMN H Molécules placées sous un haut champ magnétique continu (300Mhz=7 Tesla) Quelques spins (ppm) s orientent le long du champ magnétique Sans champ, les spins sont dans des positions plus aléatoires On applique une onde radio (impulsion de ~μs à 300MHz si 7 Tesla de champ) Les protons passent à un niveau de plus haute d énergie Une bobine (perpendiculaire au champ) mesure le retour à l équilibre des protons. Relaxation : dissipation de l énergie (envoyée par l onde radio) dans les autres atomes environnants les protons.

Principe de l analyse RMN H Mesure de la relaxation par la bobine Field Induction Decay Transformée de Fourier Un solide absorbe plus vite l énergie qu un liquide Les noyaux des protons dans un solide auront donc un temps de relaxation plus court que les protons dans un liquide. Le pic sera donc plus éfilé pour les liquides.

Effect of minerals (K) on mobility, DSC, TGA, etc. ChemSusChem, 2016

A new model has been developped based on the calculation of characteristic times Chem. Eng. Res. Des. 2011. -51-

TEM of various carbon structures in char App. Catal. 2015