UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Characterization of lignin molar mass and molecular conformation by multi-angle light scattering Ji, Lun 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2019_september_ji_lun.pdf [ 6.62MB ]
Metadata
JSON: 24-1.0379859.json
JSON-LD: 24-1.0379859-ld.json
RDF/XML (Pretty): 24-1.0379859-rdf.xml
RDF/JSON: 24-1.0379859-rdf.json
Turtle: 24-1.0379859-turtle.txt
N-Triples: 24-1.0379859-rdf-ntriples.txt
Original Record: 24-1.0379859-source.json
Full Text
24-1.0379859-fulltext.txt
Citation
24-1.0379859.ris

Full Text

  CHARACTERIZATION OF LIGNIN MOLAR MASS AND MOLECULAR CONFORMATION BY MULTI-ANGLE LIGHT SCATTERING by   Lun Ji   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in   THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (FORESTRY)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  July 2019  © Lun Ji, 2019    ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:  Characterization of Lignin Molar Mass and Molecular Conformation by Multi-Angle Light Scattering  submitted by Lun Ji in partial fulfillment of the requirements for the degree of Master of Applied Science  in Forestry  Examining Committee: Scott Renneckar  Supervisor  Feng Jiang Supervisory Committee Member  Shawn Mansfield  Supervisory Committee Member Vikramaditya G. Yadav Additional Examiner   Additional Supervisory Committee Members:  Supervisory Committee Member  Supervisory Committee Member      iii  Abstract There are many obstacles that hinder the understanding and hence the utilization of softwood kraft lignin (SKL). The lack of reliable measurements for lignin molecular weight and the corresponding molecular conformation hampers proper elucidation of structure-property relationships. Conventional gel permeation chromatography (GPC) is unable to robustly measure the molecular weight because of a lack of calibration standards with a similar structure to lignin. Further, the potential for interactions between lignin and the column gel packing delays separation, changing the mechanism from a strict hydrodynamic radius interpretation. In the present work, the determination of absolute molar mass of technical lignin was conducted utilizing gel permeation chromatography (GPC) combined with multi-angle light scattering (MALS). In order to clarify the light scattering profile, six SKL fractions, homogeneous in both structure and size were obtained by a combination of ultrafiltration and organic solvent fractionation. Further information on the molecular structure was studied utilizing a differential viscometer combined with quantitative 1D and 2D nuclear magnetic resonance spectroscopy (NMR) methods for chemical and structural analysis of functional groups and interunit linkages, respectively. Separated lignin fractions were used to enhance the clarity of light scattering profiles by narrowing the molecular weight distribution of lignin fractions so the larger polymers would not dominate the scattering. For solvent fractionated materials, the acetone soluble fraction had a lower molecular weight than the acetone insoluble fraction. In addition, hydrodynamic behaviour was acquired based on viscosity and molecular mass of fractionated samples. Acetone soluble lignin was found to possess a more compact structure relative to the acetone insoluble fraction, due to a significantly lower “α” value in the Mark-Houwink-Sakurada (MHS) plot. This compact  iv  geometry was supported by the structural analysis from NMR showing the acetone soluble part contained fewer native linkages and aliphatic side chains, which suggests the samples were considerably degraded. The relative degree of compactness (branching degree) was quantified by comparing hydrodynamic behaviour of SKL fractions with a “linear” lignin reference, as represented by enzymatic milled acidolysis lignin (EMAL), and it was found that lower molecular mass samples contained more branches than higher molecular mass fractions.    v  Lay summary Lignin is a by-product of the pulp and paper industry and is currently under intensive research as an alternative material to petroleum-based materials. However, the complex structure of lignin greatly obstructs the utilization of lignin, and thus there is a need to better understand lignin structure-property relationships. In this research, light scattering was implemented to characterize the molar mass and molecular conformation of fractionated softwood kraft lignin. The study highlighted the importance of fractionation in adding clarity to light scattering characterization. Absolute molecular weight and structural information measured by NMR and differential viscometer analysis on various lignin fractions was obtained and compared.   vi  Preface  In chapter 6 and chapter 7, Lun did all the experiments and data analysis. Volunteer students Hanna Wang and Ricky Hua helped with sample preparations. Dr. Renneckar, Dr. Cho, Amanda Johnson, Liyang Liu provided me with valuable advice in experiment design and data analysis.  In chapter 8, Liyang Liu advised me with 31P NMR and 13C NMR experimental protocols and analysis. HSQC was conducted by Mark Okon from the Mclntosh research group in the Department of Biochemistry. I prepared and analyzed the results myself.   Last, Prof. Scott Renneckar, Dr. Muzaffer Karaaslan, Dr. Shawn Mansfield, and Dr. Feng Jiang contributed to editing and advising of data presentation of this thesis. Dr. Cho and Liyang Liu provided valuable comments to improve the manuscript.  vii  Table of contents Abstract .......................................................................................................................................... iii Lay summary ...................................................................................................................................v Preface............................................................................................................................................ vi Table of contents ........................................................................................................................... vii List of tables .................................................................................................................................. xii List of figures ............................................................................................................................... xiv List of abbreviations ................................................................................................................... xvii Acknowledgements .................................................................................................................... xviii Dedication .................................................................................................................................... xix  Introduction ................................................................................................................... 1 1.1 Lignin ......................................................................................................................... 1 1.1.1 Lignin structure ............................................................................................... 1 1.1.2 Delignification from biomass ......................................................................... 4 1.1.2.1 Kraft lignin ............................................................................................... 4 1.1.2.2 Lignosulfonates (LS) ................................................................................ 5 1.1.2.3 Alkali (soda) lignin ................................................................................... 6 1.1.2.4 Organosolv lignin ..................................................................................... 6 1.1.3 Chemical structure of technical lignin ............................................................ 6 1.2 Utilization of lignin ................................................................................................... 8  Literature reviews.......................................................................................................... 9 2.1 Lignin fractionation ................................................................................................... 9 2.1.1 Selective precipitation ..................................................................................... 9  viii  2.1.2 Membrane filtration ........................................................................................ 9 2.1.3 Solvent fractionation ..................................................................................... 10 2.1.4 Summary for fractionation of lignin ............................................................. 11 2.2 Lignin molar mass characterization ......................................................................... 16 2.2.1 Molecular weight and molecular weight characterization ............................ 16 2.2.2 Molecular weight characterization of lignin ................................................. 16 2.2.2.1 Gel permeation chromatography – conventional calibration analysis ... 17 2.2.2.1.1 Solvent system ................................................................................. 20 2.2.2.1.2 Column gel ...................................................................................... 22 2.2.2.1.3 Lignin derivatives ............................................................................ 22 2.2.2.1.4 Other factors .................................................................................... 23 2.2.2.1.5 Comments on GPC conventional calibration analysis .................... 24 2.2.2.2 GPC – universal calibration ................................................................... 24 2.2.2.3 Vapor pressure osmometry (VPO) ......................................................... 25 2.2.3 Absolute molecular weight characterization ................................................. 25 2.2.3.1 Mass Spectrometry ................................................................................. 26 2.2.3.2 Ultracentrifuge and sedimentation equilibrium or velocity ................... 27 2.2.3.3 Membrane osmometer ............................................................................ 28 2.2.3.4 NMR end group titration ........................................................................ 29 2.3 Light scattering and light scattering in lignin characterization ............................... 29 2.3.1 Introduction to light scattering theory........................................................... 30 2.3.2 Light scattering in lignin characterization .................................................... 31 2.3.3 Issues in the light scattering of lignin ........................................................... 35  ix  2.3.3.1 Aggregation ............................................................................................ 35 2.3.3.2 Fluorescence and absorbance ................................................................. 36 2.3.3.3 The small size of lignin .......................................................................... 37 2.3.3.4 Heterogeneity of lignin ........................................................................... 38 2.4 Lignin molecular conformation characterization .................................................... 38 2.4.1 Conformation and topology characterization methods ................................. 39 2.4.1.1 Viscometry and pulsed gradient spin echo (PGSE) – NMR .................. 40 2.4.1.2 Dynamic light scattering (DLS) ............................................................. 41 2.4.1.3 Small angle neutron scattering and small angel x-ray scattering ........... 42 2.4.1.4 NMR chemical structure analysis ........................................................... 42 2.4.2 Lignin conformation analysis ....................................................................... 43 2.4.3 Branching degree .......................................................................................... 49 2.4.3.1 Viscosity method .................................................................................... 49  Hypothesis and objective ............................................................................................ 51  Materials and experimental methods .......................................................................... 55 4.1 Materials .................................................................................................................. 55 4.2 Sample preparation and flow chart .......................................................................... 55 4.2.1 Softwood kraft lignin fractionation ............................................................... 55 4.2.2 Acetone fractionation .................................................................................... 55 4.2.3 TFF fractionation (ultrafiltration) ................................................................. 56 4.3 Enzymatic milled acidolysis lignin preparation ...................................................... 58 4.3.1 Extractive-free wood fiber preparation ......................................................... 59 4.3.2 Ball milling and enzymatic hydrolysis ......................................................... 60  x  4.3.3 Lignin extraction and mild acid hydrolysis .................................................. 60 4.4 Determination of molecular weight of lignins GPC and MALS ............................. 61 4.4.1 Lignin acetylation ......................................................................................... 61 4.4.2 Refractive index increment measurement ..................................................... 61 4.4.3 GPC measurements with multi-detectors ...................................................... 61 4.5 Chemical analysis .................................................................................................... 62 4.5.1 Quantitative 31P NMR analysis ..................................................................... 62 4.5.2 Quantitative 13C NMR analysis .................................................................... 64 4.5.3 2D NMR (HSQC) analysis ........................................................................... 66 4.6 Elemental analysis ................................................................................................... 67  Result – lignin molecular weight characterization ...................................................... 69 5.1 Overall fractionation yield of solvent and membrane separated lignin ................... 69 5.2 MW from GPC ........................................................................................................ 70 5.2.1 Dn/dc values of fractionated lignin ............................................................... 70 5.2.2 Absolute molecular weight ........................................................................... 72 5.2.2.1 Fluorescence of lignin ............................................................................ 78 5.2.2.2 Comparing light scattering data with conventional analysis .................. 80 5.2.2.3 Lignin degree of polymerization ............................................................ 82  Result of lignin molecular conformation characterization .......................................... 84 6.1 GPC behaviour of lignin .......................................................................................... 84 6.2 Mark–Houwink–Sakurada plots .............................................................................. 84  Result of lignin chemical analysis............................................................................... 88 7.1 Kraft lignin NMR .................................................................................................... 88  xi  7.1.1 Aromatic unit as a reference and internal standard ....................................... 89 7.1.2 Interunit linkages .......................................................................................... 90 7.1.3 Functional groups.......................................................................................... 95 7.1.4 Degree of branching ...................................................................................... 99 7.1.5 Structure analyses of SKL fractions ........................................................... 103 7.2 Elementary analysis ............................................................................................... 106 7.3 Enzymatic milled wood lignin NMR ..................................................................... 107  Quantification of branching ...................................................................................... 109  Conclusion................................................................................................................. 116 Bibliography ................................................................................................................................118 Appendix ......................................................................................................................................138 Appendix A Light scattering profile ................................................................................ 138 Appendix B Acetylation correction ................................................................................. 140 Appendix C EMAL light scattering ................................................................................. 141 Appendix D Molecular density ........................................................................................ 142 Appendix E PS and lignin intrinsic viscosity .................................................................. 143 Appendix F Molecular weight by AF4 and Light Scattering .......................................... 144 F.1 Introduction ........................................................................................................................................ 144 F.2 Method .................................................................................................................................................. 144 F.3 Result and discussion .................................................................................................................. 145 Appendix G NMR result .................................................................................................. 147 G.1 NMR spectral ................................................................................................................................... 147 G.2 MW and Number of Native linkages............................................................................... 161  xii  List of tables  Table 1.1 The ratio of major linkages in lignin from softwood and hardwood6 ............................. 3 Table 2.1 Different membrane separation processes .................................................................... 10 Table 2.2 Review of lignin fractionation techniques .................................................................... 12 Table 2.3 Different types of measurement methods (A=absolute, R=relative, E=equivalent)47 .. 16 Table 2.4 Paper review of conventional calibration method in lignin characterization ............... 18 Table 2.5 MW characterization by different polar solvent system ............................................... 21 Table 2.6 MW from VPO 53 .......................................................................................................... 25 Table 2.7 Lignin characterization from MS .................................................................................. 26 Table 2.8 Lignin analysis by ultracentrifugation and sedimentation equilibrium ........................ 28 Table 2.9 MW from NMR titration and GPC method .................................................................. 29 Table 2.10 Paper review on light scattering on lignin .................................................................. 32 Table 2.11 Molecular architecture and MHS plot parameter α for dilute solution ....................... 41 Table 2.12 Reviews on lignin hydrodynamics .............................................................................. 45 Table 4.1 Quantification of various lignin moieties by 31P NMR ................................................ 63 Table 4.2 Quantification of various lignin moieties by 13C NMR ................................................ 65 Table 4.3 Quantification of various lignin moieties by 2D-HSQC NMR .................................... 67 Table 5.1 The dn/dc value of lignin fractions after different incubation periods ......................... 71 Table 5.2 Molecular weight averages for lignin fractions measured by MALS ........................... 73 Table 5.3 Molecular weight with and without fluorescence filters .............................................. 79 Table 5.4 Molecular weight averages for lignin fractions measured by conventional analysis ... 81 Table 5.5 Degree of lignin polymerization ................................................................................... 83 Table 6.1 α value and related polymer conformation in dilute solution ....................................... 85  xiii  Table 7.1 Information on interunit bonds abundance of SKL fractions as obtained (units per 100 Ar) ................................................................................................................................................. 90 Table 7.2 Information on functional groups abundance of SKL fractions ................................... 98 Table 7.3 Information on branching of SKL fractions ............................................................... 101 Table 7.4 Reviews on lignin C5 substitution from NMR results ................................................ 102 Table 7.5 Information and structural characterization of kraft lignin fractions .......................... 104 Table 7.6 Elemental composition of lignin fractions .................................................................. 106 Table 7.7 Information and structural characterization of EMAL fractions as obtained by HSQC, 13C NMR, and 31P NMR ............................................................................................................. 108 Table 8.1 Molecular weight averages for EMAL fractions measured by MALS and conventional analysis ........................................................................................................................................ 110  xiv  List of figures Figure 1.1 Three canonical lignin monomers. [a] p-coumaryl alcohol (H-lignin), [b] guaiacyl-lignin (G-lignin), [c] syringyl-lignin (S-lignin) .............................................................................. 2 Figure 1.2 Major linkages in lignin from softwood and hardwood  ............................................... 3 Figure 1.3 Reaction mechanism during kraft pulping process ....................................................... 5 Figure 2.1 Calibration curve of GPC using polystyrene as standards  ......................................... 20 Figure 2.2 Zimm plot .................................................................................................................... 31 Figure 2.3 Examples of conjugated structure ............................................................................... 37 Figure 2.4 Repeat units with different functionality, represented by T = terminal units; L = linear units; D = dendritic units  ............................................................................................................. 43 Figure 3.1 Scheme of the thesis. [a] first topic relates to issues of GPC; [b] the second topic relates to the branched structure of lignin ..................................................................................... 52 Figure 4.1 Laboratory arrangement of a TFF system ................................................................... 57 Figure 4.2 Lignin fractionation process and lignin fractions obtained ......................................... 57 Figure 4.3 Enzymatic milled acidolysis lignin (EMAL) isolation process flow chart ................. 59 Figure 4.4 Quantitative 31P NMR spectrum of [a] SKL and [b] hardwood EMAL ...................... 64 Figure 4.5 Quantitative 13C NMR spectrum of SKL, 1: β-O-4 total +β-β + β-5; 2: γ -O-Alk, secondary-OH; 3: primary-OH ..................................................................................................... 66 Figure 5.1 Scheme for SKL fractionation ..................................................................................... 70 Figure 5.2 RI profile vs elution volume of lignin fractions [a] AIKL, and [b] ASKL ................. 72 Figure 5.3 Molecular weight distribution vs elution volume of lignin fractions measured by light scattering: [a] AIKL>10kDa & ASKL>10kDa; [b] AIKL 3-10kDa & ASKL 3-10kDa; [c] AIKL <3kDa & ASKL <3kDa ................................................................................................................ 75  xv  Figure 5.4 [a] Number average molecular weight distribution and [b] hydrodynamic radius for lignin fractions. Red line was median value and blue box shows the upper and lower quartiles. The whiskers indicated the highest and lowest value, and outline values were excluded ............ 76 Figure 5.5 Peak deconvolution of RI profile of AIKL>10kDa ..................................................... 77 Figure 5.6 Zimm plot of detectors with and without filter bands of [a] AIKL 3-10kDa and [b] ASKL 3-10kDa ............................................................................................................................. 79 Figure 5.7 Molecular weight with and without fluorescence filters ............................................. 80 Figure 5.8 Molecular weight of lignin as estimated by light scattering and conventional analysis....................................................................................................................................................... 81 Figure 6.1 MHS plot of [a] AIKL fractions (AIKL>10kDa, AIKL3-10kDa, AIKL<3kDa), [b] ASKL fractions (ASKL>10kDa, ASKL3-10kDa, ASKL<3kDa), [c] AIKL prior to TFF fractionation and [d] ASKL prior to TFF fractionation ................................................................ 86 Figure 7.1 13C NMR spectrum of AIKL>10 kDa fraction; 1: Alk-O-Ar, α -O-Alk, 2: γ -O-Alk, OHsecondary, 3: OHprimary .................................................................................................................. 92 Figure 7.2 2D HSQC Spectrum of AIKL fraction – [a] aliphatic region; [b] aromatic region .... 93 Figure 7.3 Main structures in kraft lignin identified 2D NMR ..................................................... 94 Figure 7.5 Elimination aliphatic sides from kraft pulping ............................................................ 97 Figure 7.6 31P NMR spectrum of AIKL>10 kDa fraction ............................................................ 99 Figure 7.7 Branching structure on [a] C1 and [b] α position....................................................... 101 Figure 7.8 Hypothesized lignin structure of different lignin fractions ....................................... 105 Figure 8.1 MHS plot of AI-EMAL and AS-EMAL ................................................................... 111  xvi  Figure 8.2 [a] Trendlines in MHS plot of AIKL (3 fractions), AKL (3 fractions), AI-EMAL, and AS-EMAL, [b] branching ratio at different MW for AIKL(a) and ASKL, assuming “e” equals to 0.5, 0.75, 1, 1.5............................................................................................................................ 112 Figure 8.3 Branch points per 100 lignin units of [a] AIKL fractions and [b] ASKL fractions .. 115    xvii  List of abbreviations AIKL = Acetone insoluble softwood kraft lignin ASKL = Acetone soluble softwood kraft lignin  Bn = Branching units per molecule dn/dc = Refractive incremental index  DC = Degree of condensation  DP = Degree of polymerization  EMAL = Enzymatic milled acidolysis lignin gM = Branching ratio GPC = Gel permeation chromatography  HSQC = Heteronuclear single quantum coherence LS = Light scattering  MW = Molecular weight  NMR = Nuclear magnetic resonance  PDI = Polydisperse index  PS = polystyrene  Rg = Gyration radius  Rh = Hydrodynamic radius  RI = Refractive index  SKL = Softwood kraft lignin  SEC = Size exclusion chromatography  TFF = Tangential flow filtration  THF = Tetrahydrofuran   xviii  Acknowledgements I really appreciate the help of my supervisor Dr. Renneckar, who led me to the world of science and guided me in the field of lignin research. He kindly supported every decision I made and allowed me to learn from my experiences. Dr. Mansfield and Dr. Jiang are my committee members and raised lots of insightful questions and suggestions and also helped with insightful edits. Mijung Cho and Liyang Liu are two senior colleagues who really helped me a lot. I thank them for teaching me all the lab work techniques, correcting my mistakes, and sharing experiences with me. I appreciate everyone who worked with me, Muzaffer, Amanda, Saurabh, Ricky, Hanna, Ruilin, Jiufang, Kim, and others in the different groups within wood science that helped me.  Thank you everyone, because we shared our thoughts and sometimes criticized each other. There was happiness and arguments. But what’s important is, we did it for the same pursuit of making personal progress and contributing to our field of research. I enjoy the days being with you, working with you, chatting with you, gossiping with you, travelling with you. That’s one of the most precious experiences that I shall never forget.   xix  Dedication  This study is dedicated to my parents who provide their moral, spiritual, emotional, and financial support. In addition, best wishes to all my friends, we shall support each other in this continuous life journey.   1   Introduction  1.1 Lignin  1.1.1 Lignin structure  Lignin is an important structural component of terrestrial plants. The content of lignin varies among different plant species and accounts for 27-33%, 18-25% and 17-24% in softwood, hardwood, and grass species, respectively1. Lignin’s main role within the cell wall is to enhance the mechanical properties2. This aromatic polymer also aids in the resistance to microbial and fungal degradation, while assisting in the transportation of water and other nutrients due to its hydrophobicity relative to the polysaccharides3.  Lignin is polymerized from three different phenolic alcohol monomers: p-coumaryl alcohol (H-lignin), coniferyl alcohol (guaiacyl-lignin), and sinapyl alcohol (syringyl-lignin) (Figure 1.1). The content of each type within the cell wall is controlled genetically, thus composition depends upon the plant species. Softwood lignin is derived from over 95% coniferyl alcohol with the remainder p-coumaryl alcohol; hardwood lignin is derived from both coniferyl and sinapyl alcohol with a minor amount of p-coumaryl alcohol. Grass species contain a polymer from all three monomers2, and some vegetable sources contain high amounts of ferulic acid4.     2     (a) (b) (c) Figure 1.1 Three canonical lignin monomers. [a] p-coumaryl alcohol (H-lignin), [b] guaiacyl-lignin (G-lignin), [c] syringyl-lignin (S-lignin)  These monolignols are transported from the cytoplasm to the cell wall structure where they undergo enzyme-mediated dehydrogenative polymerization and form amorphous polymers5, known as lignins. The polymerization process happens via coupling of resonance stabilized radicals on the aromatic units. The most common linkages resulting from this process in “native lignin”, represented by milled wood lignin, include β-O-4, β-5, β-1, β-β, 5-5, and 4-O-5 (Figure 1.2), The corresponding composition of linkages in softwood and hardwood species is listed in the Table 1.16. The diversity of the monomers and the multiple chemical bonding sites results in the complex nature of lignin. In addition, due to chemical modification occurring during lignin extraction, the characterized lignin structure greatly depends upon the extraction process as lignin readily can repolymerize via radical induced coupling. These factors complicate the clear understanding of lignin structure.  Noticeably, hardwood lignin contains syringyl lignin (S-lignin), indicating one less active site on phenyl rings due to the existence of a second methoxy group, as compared with guaiacyl lignin (G-lignin). Thus, hardwood lignin contains more β-O-4 bonds as the primary backbone  3  linkage, and as such, has a lower possibility of branching with a lower degree of condensed structures from carbon-carbon bonds (C-C) at the open 5 position.  Table 1.1 The ratio of major linkages in lignin from softwood and hardwood6 Bonds Softwood % Hardwood % β-O-4 35-60 50-70 β-5 11-12 4-9 β-β 2-3 3-4 5-5 10 5 4-O-5 <4 7      β-O-4 β-β    β-5 / α-O-4 4-O-5 5-5 Figure 1.2 Major linkages in lignin from softwood and hardwood7   4  1.1.2 Delignification from biomass In the pulp and paper industry, delignification refers to the process where lignin is depolymerized and modified, typically into small fragments, so that it becomes soluble in pulping liqour3. The degraded soluble product from this process is called technical lignin to differentiate the altered material from its native state. Depending on the pulping process, there are four major types of industrial technical lignin: kraft lignin, lignosulfonates, alkali (soda) lignin, and organosolv lignin.   1.1.2.1 Kraft lignin  Kraft pulping is currently the most popular pulping process in North America (NA) with a closed loop recycling of its caustic pulping additives3,8. It accounts for approximately 80% of the total lignin produced in NA3. In the process, wood chips are mixed with sodium hydroxide and sodium sulfide and heatedto 150-170°C for approximately 2 hours. 90-95% of the total lignin within the wood chips is degraded and dissolves into the alkali solution9. Along with acidic carbohydrate breakdown products and inorganics, this solution of lignin is known as black liquor. In the process, the degradation of lignin mainly occurs at the cleavage of α and β ether bonds from the propyl side chain resulting in small soluble fragments. Other C-C linkages, like β-5, β- β, are more stable and therefore only minimal degradation of these linkages occurs. The depolymerization process also facilitates repolymerization of free phenolic radicals10. Therefore, there is a high percentage of condensed structures in technical lignin. In addition, a small amount of elemental sulfur is incorporated in lignin’s structure due to the presence of hydrosulfide anions that typically react with benzylic carbons. The detailed reaction mechanism is shown Figure 1.3.   5  After delignification, kraft lignin is precipitated from black liquor via acidification with CO2. Recovered lignin is further washed by sulfuric acid to maximize quantity and purity. In a new process, lignin is resuspended in acidic water; this is a commercialized process, known as LignoBoost. This resuspension is in contrast to Indulin AT lignin that is washed on a filter press11. Another developing process is the LignoForce system, where black liquor is oxidized by oxygen before CO2 precipitation. LignoForce lignin yields lower ash content and higher lignin purity without residual hydrogen sulfide12,13.    Figure 1.3 Reaction mechanism during kraft pulping process  1.1.2.2 Lignosulfonates (LS)  Sulfite pulping was the most popular process in North America for a number of years and it yields high purity cellulose fibers14. An aqueous solution of sulfite or bisulfite salt is heated with wood chips at temperatures between 140°C and 160°C and pH between 1.5 to 2. During this delignification process the lignin becomes modified with sulfate esters. The resulting  6  lignosulfonates are negatively charged in aqueous media, and thus, become water-soluble as anionic polyelectrolytes.    1.1.2.3 Alkali (soda) lignin  Soda processing is widely implemented in non-woody biomass, including wheat straw and bagasse. Sodium hydroxide is used to degrade lignin at 140-170°C in a sealed environment. The process is similar to the kraft process where major cleavage occurs at α and β ether bonds. The recovered lignin has high purity but low molecular weight15. Commercial lignins are sold under the brand name Protobind™ and serve as a commercial source of sulfur-free lignin.  1.1.2.4 Organosolv lignin Lignin can undergo acidolysis in organic solvents resulting in cleavage of the native -O-4 linkages. This process involves the utilization of an aqueous-organic solution, such as ethanol, acetone, or methanol, along with an acid catalyst such as sulfuric acid3. The isolated lignin typically has low carbohydrate contamination and has no sulfur, making it a relatively pure technical lignin. This material has relatively low molecular weight with a lower polydispersity, outperforming other lignins in a variety of applications.    1.1.3 Chemical structure of technical lignin  Though technical lignin differs from native lignin in chemical structure, the basic repeating unit remains similar, as the methoxy group on benzene is not substantially degraded16,17. Native S/G/H ratios affect the structure in technical lignin, as S lignin prevents recondensation reactions in the pulping process. Currently, 13C NMR and thioacidolyisis are two commonly used methods  7  to characterize S/G content. In 13C NMR spectra, S (108-110 ppm), G (113-108 ppm) and H (160-155 ppm) content are separated into three distinct peaks. In thioacidolysis, lignin is degraded into monomers through cleavage of the -O-4 bonds and is analyzed by GC-MS18.  Technical lignin varies from native lignin in chemical linkages, molecular weight, and functional groups. As mentioned, many ether bonds are degraded during delignification processes, and newly formed radicals undergo repolymerization into rather complex structures2. Many techniques can give partial information on technical lignin structure, including thioacidolysis, ESI-MS, and NMR. 13C-1H two-dimensional hetero-single quantum coherence (2D HSQC) NMR is a developing technique with insight in native lignin linkages, including β-O-4′, β-5’ and β-β’. This method can assess lignin within the solubilized cell wall, as well as native-like, lignin isolated under mild conditions such as ball milling19. Both of these cases offer a unique comparison to technical lignin which undergoes significant disruption.  Functional groups are another important factor affecting lignin properties. Hydroxyl groups are one of the dominant functional groups, and are classified into aliphatic and phenolic hydroxyl groups. 1H and 13C NMR provides quantification of hydroxyl groups, but acetylated samples have to be used to avoid significant overlap with ether groups within the spectra20. 31P NMR excels in enabling hydroxyl groups characterization providing differentiation of the hydroxyls along with the carboxylic acid groups of the derivatized sample. A phosphine reagent, such as 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxphospholane (TMDP), reacts with OH groups on the lignin forming phosphorylated derivatives. NMR spectra contain the phosphorous shift that is affected by the nucleus and adjacent chemical bonds. Thus, different OH groups are clearly separated into distinct regions.     8  1.2 Utilization of lignin  Currently, there are potential 50-100 million tons of technical lignin produced as a by-products of the pulp and paper industry every year3,21. However, more than 98% of lignin is burned, utilized as energy in the chemical recovery boilers22. Further, there is a considerable need for a renewable source for industrial feedstocks as a supplement or substitute for petroleum-based chemicals10,23. Therefore, research into value-added lignin-based products has gained considerable momentum in the past few years, and many discoveries and innovative lignin products have been developed relative to mature applications for lignosulfonates including dispersants, emulsion stabilizers, concrete additives, surfactants, and binders4. Prototypes of technical lignin products include lignin-based carbon fibers, lignin-derived polymers (polyurethanes, polyesters, and epoxy resin), and fine chemicals (vanillin) 3. However, most of the lignin products have very limited performance or cannot meet the specifications of petroleum analogs24,10. There are many obstacles that hamper the utilization of lignin. Among them, the heterogeneity of lignin and a lack of reliable characterization methods for molecular mass of lignin, as reviewed below, making it difficult to establish a reliable lignin molecular structure25. Hence, it is important to better characterize the absolute molar mass distribution of lignin and the related molar mass dependent molecular conformation.    9   Literature reviews  2.1 Lignin fractionation Due to the complexity of technical lignin, fractionation methods have been proposed for narrowing the variance of lignin properties. The overall objective from these processes is to obtain specific molecular weight fractions with uniform chemical structure. There are three different types of fractionation methods, selective precipitation, organic solvent dissolution, and membrane fractionation.   2.1.1 Selective precipitation  Selective precipitation isolates lignin fractions at different pH values and temperatures. For example, kraft lignin is fractionated by gradually acidifying black liquor. This method is the most straightforward as there are no additional solvents required to separate the lignin. However, the disadvantage of this method is the poor repeatability, as pH value and subsequent protonation is not easy to control relative to lignin structure26.   2.1.2 Membrane filtration  Due to the development of new membrane technology, ultrafiltration has been widely introduced into lignin fractionation27,28,29. Depending upon the pore size, three separation processes are distinguished as microfiltration, ultrafiltration, and nanofiltration, corresponding to membrane cutoffs (0.1-10um, 1-10nm, and less than 1nm respectively) (Table 2.1)30.      10  Table 2.1 Different membrane separation processes Process Membrane pore size Membrane material Microfiltration 0.1-10 µm Ceramics, metal oxides, polymers (cellulose nitrate or acetate, PVDF, PTFE, polysulfone, polyamides) Ultrafiltration 1-10 nm Ceramics, polysulfone, PP, nylon 6, PTFE, PVC, acrylic copolymers Nanofiltration <1 nm Cellulosic acetate and aromatic polyamide  Depending on filtration technique, it can be classified as dead-end filtration and tangential flow filtration (TFF). In dead-end filtration, fluid flows towards a membrane under applied pressure and particles accumulate on the filtration membrane27. In a TFF system, fluid flows tangentially along the surface of the membrane under an applied pressure so that the retained components do not build up on the membrane28. Therefore, TFF has great potential in scaling up the separation process as the method limits membrane fouling. Due to excellence in separation, TFF has been employed in the lignin fractionation system, especially for ultrafiltration system31.   2.1.3 Solvent fractionation  Solvent fractionation separates lignin depending upon the relative solubility and the MW distribution of the resultant lignin fractions is significantly narrowed32. Different solvent systems were proposed in the last few decades with seminal papers by Morck et al (1988)33,34,35. Sequential extraction systems use a different solvent to extract lignin step by step. For example, Gioia et al (2018) sequentially extracted kraft lignin wih ethyl acetate (EtOAc), ethanol (EtOH), methanol (MeOH), and acetone18. The sequence of organic solvent was designed so that lignin fractions with the lowest molecular weight were separated first. In another case, Saito et al. (2014) extracted lignin with 5 repeated methanol extractions, generating six lignin fractions36. Lignin fractionation with a single solvent simplified separation. However, the drawback was that  11  the methanol insoluble part accounted for 50% of the total mass and had increased polydispersity index (PDI). Another procedure sequentially precipitated dissolved lignin in a binary solvent system37. In this method lignin was first fully dissolved into a solvent. The polarity of the solvent environment was gradually altered by adding nonsolvent, so that lignin of polar differences was separated. Some of the commonly used binary solvent systems include methanol/methylene chloride, ether/acetone, diethyl ether/acetone, or acetone/hexane38,39,40.   2.1.4 Summary for fractionation of lignin Much research had been conducted on lignin fractionation. Through fractionation, purified lignin with decreased MW distribution outcompeted unfractionated technical lignin in many applications22. Generally, if lignin had better solubility in select organic solvents or smaller size, it also had a lower Tg, lower MWD, less aliphatic OH, higher phenolic OH, and increased sulfur content. Fractionation systems and resultant chemical structure of lignin fractionation are summarized and compared in Table 2.2.          12  Table 2.2 Review of lignin fractionation techniques Selective precipitation Lignin PH range Characterization chemical structure related Ref Black Liquor  Empty fruit bunch pH: 4.8, 4.0, 3.3, 2.0 1.5 Fractions MW PDI Raw (pH 2.0) 2010 1.60 4.8 2160 1.41 4.0 1990 1.40 3.0 1980 1.42 2.0 2070 1.43 1.5 2050 1.42  S:V:H= 18-21:10-12:1 (S lignin: Vanillin: H lignin) pH↑⇒ yield & purity↑ 199941 Black liquor 7.5% soda pulping  Raw straw  pH: 12.64, 11.08, 10.41, 9.16, 6.50, 5.40, 4.55, 2.57, 0.72 Fractions MW PDI Raw (pH 2.0) 3135 1.66 9.16 2160 1.41 6.50 1990 1.40 5.40 2120 1.86 2.57 2432 1.86 0.72 3501 1.84  pH↑⇒ yield ↓& purity↑ pH↑⇒impurities↑ 200926      Membrane fractionation HKL, SKL Black liquor (31% solids) TFF (4, 6, 20, 100kDa) Polyethersulphone membrane   200842 Black Liquor SKL Ceramics membranes  (5, 10, 15kDa) Fractions  MW PDI Raw   5654 3.01 >15kDa 6300 3.10 15kDa 3544 1.87 MW↑⇒ organic matter Inorganic matter, lignin content in OM ↓ No difference in H:G:S  201027  13  10kDa 2022 2.14 5kDa 1806 1.92  >15 contaminated  Black Liquor  SKL TFF (1Kda)  Ceramic membranes pH, T°C Salt (Na2SO4)  Acetylated Non-acetyl Alkali Fractions MW PDI MW PDI MW PDI Raw pH 9 3525 4.2 2071 2.9 2005 7.5 <1 pH 9 1096 2.2 899 1.9 1316 7.1 <1 pH 6.5 1074 2.1 864 1.9 1234 7.6 <1 pH 4 973 2.1 806 1.9 1227 6.3  Precipitation pH↓ ⇒ MW↓,  MW↓⇒S% ↑ MW distribution (PDI) ↓  201328 Black liquor Ceramic and polymeric membranes  (200Da-1kDa)  Lignin content, total dissolved solids  201329 SKL pH>13  solid content -17.6%  ultrafiltration (1, 5, 10kDa)   Ceramic membranes made of TiO2 and ZrO2 Fractions MW PDI Raw 20200 4.1 >10 33500 3.5 >5 28200 3.5 5-10 4900 2.2 1-5 4700 2.3 0-5 4100 2.4 0-1 2700 2.1  MW ↑⇒aliphatic OH ↑, phenolic OH ↓, Tg (70-170°C)↑, S%↑, molecular size↑ Low MW ⇒ soften↑ + high-temperature crosslinking 201443 HKL  Black liquor  TFF tubular ceramic membranes (5, 15 and 50 kDa) Fractions MW PDI Raw 10323 1.19 50kDa 12029 1.23 15kDa 9576 1.15 5kDa 9641 1.13 <5kDa 8879 1.11  MW ↑⇒ S/G ↓, condensed structure↓, phenolic monomers↓, total OH↑, 201844      Solvent fractionation  14  Softwood kraft lignin MeOH SEC +  DMF & DMF/LiCl Fractions MW PDI LS MW dndc Raw 7190 3.05 96100 0.1867 MeOH ISo 14900 2.28 188000 0.1837 MeOH So 3000 1.90 86600 0.1636 In DMF/LiC  MW distribution (PDI) ↓ MW↑⇒ aliphatic OH ↑, phenolic OH ↓ MW↑⇒ Tg Td ↑ 201436 SKL KL1: pine (bleached) KL2: pine  KL3:spruce  Acetone / n-hexane  Acetone  Fractions MW PDI L2 15000 10 F1 58000 19 F2 4000 1.8 F3 2100 1.3 F4 3100 1.1 MW ↓⇒ phenolic OH↑ , aliphatic ↑ , S% slightly changes   201431 wheat straw organosolv lignin Acetone / n-hexane  Acetone Fractions MW PDI Raw 5000 5.9 Ace ISo 23600 21 Hex 1st 7700 4.3 Hex 2nd 3900 3.5 Hex So 3500 5.7  MW distribution (PDI) ↓ MW ↓ ⇒ β-O-4 ↓ , phenolic OH↑ , aliphatic no specific result  Acetone soluble part, O/C ratio ↓, O content ↓; MW↓⇒ G lignin → 201637 SKL lignoboost ethyl acetate (EtOAc), Ethanol (EtOH), methanol (MeOH), Acetone Fractions  MW  PDI  KL 6800 6.1 EtOAc 1000 1.5 EtOH 2100 1.9 MeOH 3200 2.0 Acetone 5400 2.0 Residue 17700 2.8  MW↑⇒ Condensation↑, aliphatic OH ↑, phenolic OH ↓, Carboxylic acid↑, S%↑, ↑  201845 Kraft lignin Alcohol,  Ketone  Fractions MW-So PDI MW-ISo PDI MeOH 2040 3.52 14120 3.66 EtOH 1400 3.04 10710 3.35 Solubility: alcohol < acetate < ketone  201632  15  Ester   EtOAc + EtOH + MeOH + Acetone  SEC DMSO/LiCl 1-propOH 790 2.82 7800 5.20 i-propOH 760 2.62 6720 5.84 t-buOH 780 3.00 5670 8.72 Acetone 2450 3.95 15760 3.87 MEK 1300 4.06 10390 5.56 EtOAc 710 2.37 7800 4.11  Comparing THF and DMSO system: intermediate MW range are similar but diverge in some fractions         16  2.2 Lignin molar mass characterization  2.2.1 Molecular weight and molecular weight characterization Molecular weight is an important property of a polymer. Macromolecules with the same chemical structure can differ in properties due to different molecular weight (MW). For example, glass transition temperature (Tg) and mechanical strength increase with molecular weight and levels off after a critical point46.   2.2.2 Molecular weight characterization of lignin  There are many different techniques that have been applied to lignin molecular weight characterization, as summarized in Table 2.3. Absolute molar mass refers to a direct measure of the polymer molecular weight without any assumptions of chemical or physical properties of the sample. Relative molar mass measures the molecular weight relative to standards, which have the same chemical structure as the analyte. Equivalent molar mass also directly computes the molar mass when knowledge of the analyte structure is known47.   Table 2.3 Different types of measurement methods (A=absolute, R=relative, E=equivalent)47 Method Type Molecular Weight Range, g/mol Mean Value measured Membrane Osmometry A 104-106 Mn Gel Permeation Chromatography R 102-107 Mn, MW Vapor Phase Osmometry R <105 Mn End-group Analysis E <105 Mn Mass Spectrometry  E <105 Mn, MW Static Light Scattering A 102-108 MW Sedimentation Equilibrium A <106 MW Solution Viscosity R 102-108 M   17  2.2.2.1 Gel permeation chromatography – conventional calibration analysis Gel permeation chromatography (GPC), also referred to as size exclusion chromatography (SEC) provides molecular weight distribution information about a polymer48. It separates macromolecules by molecular size, where larger molecules elute earlier than smaller sized molecules after passing through a column of gel beads. This mechanism is known as steric exclusion. A GPC column is filled with a polymer gel of different sized pores and volumes. During elution, small molecules spend more time diffusing into pores and back, and therefore they elute later49. Afterward, the coupled downstream detectors, usually ultraviolet (UV) or refractive index (RI), detects the elution and concentration of the separated polymers and records the elution time. In physics, elution volume is mathematically modeled as  𝑉𝑒 = 𝑉0 + 𝐾𝑑 ∗ 𝑉𝑖 where, V0 is the total volume of solvent outside the pores, Kd is distribution coefficient, and Vi is the total volume of solvent inside the pores. Kd ranges from 0~1, indicating the volume of pores that is available for polymer diffusion50.  Prior to the experiment, a calibration curve must be obtained from a series of standards of different MW, and the relationship between molecular weight and elution time is plotted (Figure 2.1). By referring to the calibration curve, the MW at a certain elution volume is measured.  As the most popular method for MW determination, GPC has been applied and intensively researched for many lignin types. Discussions on parameters such as column gel types, mobile phase systems, lignin types, and lignin derivatives are list in Table 2.4.   18  Table 2.4 Paper review of conventional calibration method in lignin characterization Lignin sample Gel type Solvent Standards Reference Organic solvent  MWL, DHP Microgel & styragel THF(ac) & DMF PS + RI 200351 KL, OSL Ultralstyragel  THF (ac) PS + RI/Visco 199352 AL, OSL PL Gel  Cosmosil HPLC Si-gel 5SL + silica based column THF (ac) DMF/0.2M LiCl PS + UV/RI  200453 KL, LS, OSL  SDVB  THF + QAM PS & biphenyl + UV 199454 LS, KL, AL PSS gram (polyacrylate methacrylate) PSS PFG-PRO(silica) DMSO/water/0.05M LiBr DmAc/0.11M LiCl  Pullulan + RI  200555 200656 LS. KL, OSL, Soda  Rigid cross-linked silica particles  DMSO/ LiBr (0.5% w/v) PSS + UV/RI GPC + APC 201725 OSL PLgel  THF (ac) (acb) PS + RI 201648 EMAL MW Styragel HR5E and HR 1 THF (acb) PS + UV 200757 LS Polyacrylate methacrylate  PFG-PRO DMSO/water/0.05M LiBr DMAc/0.11M LiCl Pullulan/glucose/cellobiose Polyethylene glycol; Polyethylene oxide; 200656 MWL  DHPs Microgel  µStyragel 1000Å column DMF THF  PS +LS/RI/UV 200351  19  THF-LiBr SKL HR5E, HR1, HR0.5 THF DAMc/LiCl DMF/LiCl PS + UV/RI 199958 KL PSS Gram (Polyester copolymer network) DMSO THF (acb) PS + UV/RI 201632 SDVB: styrene–divinylbenzene; QAM: quaternary amine methyltrioctylammonium chloride; SDS: sodium dodecyl sulfate      Aqueous system  LS Jordi Glucose-DVB Water/DMSO/ Na2HPO4-4H2O/SDS  200259 KL  TSK-gel Ultrastyragel 0.1M NaOH THF (ac)  Pullulan + UV PS + UV  201328  TSK Bio-Gel P-60 NaOH NaOH/ Na2CO3 or KCl PSS + UV/RI 198560 LS, Soda  Ultrahydrogel +  Ultrastyragel NaOH/ Na2CO3 or NaCl Pulluan + YV  200061 KL, soda  Superdex 75 gel (Pharmacia) 0.1N NaOH Protein standards + lignin monomer/oligomer + UV 201262     20   Figure 2.1 Calibration curve of GPC using polystyrene as standards48  2.2.2.1.1 Solvent system  Based on the eluent, GPC systems are classified as either organic and aqueous solvent systems. In an aqueous system, sodium hydroxide (NaOH) was mostly adopted for lignin analysis. Historically, it was utilized due to the ease of buffer preparation and applied in the characterization of lignosulfonate and soda lignins63. However, many publications reported an unrepeatable, pH-dependant elution profile, which resulted from two reasons. First, lignosulfonates behave like a polyelectrolyte and its conformation expands at higher pH and is further impacted by salt concentration64, resulting in an increase in molecular size and an overestimation of MW. Second, LS can aggregate in NaOH when the pH is not high enough. As such, an empirical critical alkali concentration of 0.1M was suggested, above which LS was stable and aggregation suppressed64,65. Sodium and lithium salt, like Na2CO3, LiBr, or LiCl were also introduced in the system for increased ionic strength, while not increasing pH. However, high alkalinity and salt content impairs columns and detectors in the long term.   21  In an organic solvent system, both non-polar (or low-poplar) (THF) and polarized solvent (DMAc, DMSO, and DMF) have been intensively utilized and compared for lignin analysis. THF is currently the most frequently used buffer because of its wide compatibility with a wide range of columns, excellent elution capacity (THF accelerates the process of analyte absorption with column gel), low UV absorption (maximum at 212nm)66, and low polarity (low light scattering and good RI response ensures a clear background and lower noise level). However, the disadvantage is that lignin does not directly dissolve in THF. Therefore, acetylation was necessary to enhance the solubility of lignin in THF by transforming the hydroxyl groups into acetyl groups. Acetylation extends the analysis time to several days, with the need for chemical modification, and modifies the lignin structure. Moreover, though lignin particles visually disappear in THF, few micro-aggregations may still exist in THF. This process was not universal for all lignins, as lignosulphonates were not soluble on a THF system25.  In order to solve the solubility problem, researchers have tried to increase the polarity of the solvent system either by introducing LiBr or LiCl into THF51, or utilizing polar organic solvent like DMF/DMSO/DMAc. However, except for avoiding tedious acetylation step, none of these solvents outperforms THF (Table 2.5). More experimental data is required to better understand these phenomena58,51,32.  Table 2.5 MW characterization by different polar solvent system Ref  MW characterized in a different solvent system   199958 MW: DMF/LiCl >DMAc/LiCl > THF  200351 DMF resulted in quite different from THF and THF-LiCl.  THF and THF-LiCL results were similar  200656 Results in DMAc/LiCl is similar to DMSO/H2O/LiBr for lignosulfonate 200767 Review paper: 0.5M NaOH similar to DMAc and DMSO  201632 DMSO and THF, MW correlate well in the intermediate MW range but diverges at both ends.   22   2.2.2.1.2 Column gel  Currently, Agilent PLgel™ and Waters Styragel™ are the two most common column brands in lignin analysis48. In aqueous systems, TSK-GEL or ultrahydrogel series columns made from hydroxylated PMMA are utilized. In organic solvent, porous polystyrene/divinylbenzene (PS/DVB) particles are the most widely adopted material due to its wide pore sizes68 and high hydrophobicity69. Rigid cross-linked silica particle-based columns have been considered when resistance to high pressure is necessary in ultra-high pressure chromatography systems.  A series of round robin tests were conducted to test the reproducibility of lignin MW measurements on different GPC columns and solvent systems. NREL (1991) claimed quite large inter-laboratory variations for non-wood lignin on both THF and DMF systems70. Another test performed in 2007 reported low reproducibility of lignin with different columns and solvent systems67. Lignosulfonates may be the only lignin type that showed good repeatability56. The reason for this might be related to the adsorption phenomena arising from interactions between lignin and the column gel71.   2.2.2.1.3 Lignin derivatives  As lignin originally does not dissolve in THF, which is the most popular buffer for GPC system, different derivatization methods have been developed to ensure the solubility of lignin. Major preparation methods include acetylation, acetobromination, and silylation. In the process of acetylation, pyridine and acetic anhydride were used as reactants, and all hydroxyl groups asre transformed into acetyl groups. The polar associative interactions amongst hydroxyl groups are suppressed, such that modified lignin is compatible with THF67.   23  Another technique, acetobromination used acetyl bromide in glacial acetic acid for the reaction, and it yielded the same chemical structure and elution profiles as that from the acetylation process72,57,73. The major advantage of the technique was reducing the reaction time from a few days to approximately 1 hour. Silylation was also used on samples containing high level of carboxyl groups, like oxidized lignin, which interacts strongly with non-polar columns, and resulted in a long-tail shape profile74.   2.2.2.1.4 Other factors  There are also other factors that might affect the molecular weight results, but not systematically verified in publications. This includes operating temperature, calibration standards, GPC sample preparation, and peak selection during analysis. The temperature should impact the GPC as the separation process is governed by entropy50. 25°C, 35°C and 60°C are the most commonly reported GPC operating temperatures. One paper indicated the particle size of sulfated lignin increased with temperature from 293K to 333K in an aqueous solvent, as studied by dynamic light scattering75. In terms of sample preparation, minimal sample should be injected for sufficient signal response, whereas excess sample diminishes the GPC separation quality. Empirically, an injection concentration of 2.5-10 mg/mL was proper for a polymer with an MW <400,000 g/mol as suggested by Agilent49. Polystyrene (PS), dextran, and polystyrene sulfonate (PSS) are the commonly used calibration standards, which is largely defined by the solvent system.      24  2.2.2.1.5 Comments on GPC conventional calibration analysis In summary of GPC conventional calibration analysis, the major advantages of this method included wide accessibility, ease of use, and relatively good intra-laboratory repeatability. Nevertheless, it also has two major limitations. First, it measures the relative molecular weight, which can largely deviate from the true value due to the structural differences between branched lignin and linear polystyrene (a common THF standard). Second, it lacks universal applicability to all lignin types and has very limited inter-laboratory reproducibility48 67, especially for lignin with higher PDI and MW76,73,77. Lignin interaction with GPC column gel during elution causes a delay and an underestimation of MW. Moreover, lignin structure varies with species and fractions, leading to an even more unpredictable result67. Therefore, there is a need for reliable characterization of absolute molecular weight which is independent of the separation method and the conformation of the molecule, and provides more accurate information on lignin structure78,79.   2.2.2.2 GPC – universal calibration  Universal calibration measures the relative molecular weight dependent on the viscosity properties. A standard calibration curve is generated from a set of known standards and the relationship between intrinsic viscosity ([𝜂]) and molecular weight (MW) is used for better insight into molecular weight. As shown in the equation below, there is a theoretical relation between [𝜂] and MW, and the constants, 𝐾 and 𝛼. This equation is known as Mark-Houwink-Sakurada equation.  [𝜂] = 𝐾 ∗ 𝑀𝛼  25  Therefore, by measuring the intrinsic viscosity of an analyte, the molecular weight can be calculated based on the 𝐾, 𝛼 value of standards. As conventional and universal calibration often use the same calibration standards, the molecular weight from the two methods are usually similar56 80.  2.2.2.3 Vapor pressure osmometry (VPO) VPO depends strongly on the presence of low molecular weight fractions in the lignin53. The calibration curve is first developed indicating the relationship between the vapor pressure and MW of standards. Then, the vapor pressure of lignin at the same concentration is measured, and MW is derived from the calibration curve. In 2004, a paper reviewed this method, but did not further develop the protocol due to its low repeatability53 (Table 2.6).   Table 2.6 MW from VPO 53 Lignin fractions  MW Straw lignin  1420 Hemp lignin  640 Hardwood lignin  670 Flax lignin 600  2.2.3 Absolute molecular weight characterization  Absolute molecular weight is a direct measure of the physical properties of a molecule without the need for standards. There are five major absolute molecular weight (or equivalent molecular weight) characterization methods in lignin analysis, namely light scattering, mass spectrometry, ultracentrifugation and sedimentation equilibrium, membrane osmometer, and NMR end group titration.    26  2.2.3.1 Mass Spectrometry  In mass spectrometry characterization, a sample is first ionized into charged fragments. These ions are further accelerated in an electric and magnetic field, and fractionated according to the mass-to-charge ratio. Downstream detectors measure the charge and mass of the resultant ions. There are different types of mass spectrometers depending on the ionization and mass selection method, for instance, electrospray ionization mass spectrometry (ESI-MS) and matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF-MS)81.  In the case of lignin, the advantage of MS is the versatility and ease of use. It gives both structural information and molecular weight of a lignin fragment with a single run82. The disadvantage mainly associates with the ionization process, when significant lignin degradation and re-polymerization occurs. It greatly diminishes the reliability and the repeatability of measurements. The molecular weight measured is usually <1000 Da, less than other characterization methods (Table 2.7).  Table 2.7 Lignin characterization from MS Lignin type Techniques Molecular weight Chemical structure Ref MWL KL ESI-MS Softwood MWL 2500 Softwood OSL 1900 Hardwood MWL 2400 HKL 800   199982 Grass lignin MALDI-TOF MW: 100-800Da Qualitatively: 𝛽-5, 𝛽-O-4 200383 MWL ESI-MS MW:200-7000Da,      ave: 2500Da 𝛽-O-4; 𝛽- 𝛽;  𝛽-5 structure 201284   27  2.2.3.2 Ultracentrifuge and sedimentation equilibrium or velocity  In ultracentrifuge experiments, a sample is first centrifuged at a very high speed, as high acceleration as 1 ∗ 106 G. Meanwhile, UV or refractive index monitors the real-time concentration. The final steady state, when diffusion balances sedimentation, is analyzed by utilizing the Mason-Weaver equation below: 𝑀 = (2𝑅𝑇(1 − 𝑉𝜌)𝜔2) (𝑑𝑙𝑛𝑐𝑑𝑟2) where, M is absolute MW, c is a concentration, r is the radial distance from the center of rotation in cm, T is time, V is the partial specific volume, 𝜌 is the density of the solution, 𝜔 is the angular velocity of the rotor, and R is the gas constant85. This method offers absolute MW and was used to calibrate the GPC system several decades ago. However, the major limitation is the long analysis time, as it takes up to several days for running and reaching equilibrium. Today, it is nearly a supplementary method in the area of synthetic polymer analysis47. Due to the lack of recent examinations, the results are debatable and hard to verify. As common difficulties with lignin, like aggregation, were not reported in this method, ultracentrifugation might show promise, if the process can be streamlined.          28  Table 2.8 Lignin analysis by ultracentrifugation and sedimentation equilibrium Lignin type Solvent system Molecular weight Ref LS 0.1N NaCl MW: 70,000 Rh: 0.7-7nm 196886 MWL, OSL, AL, Steam exploded lignin THF Lignin Rotor speed MW SEL 15000 9100  30000 5400 OSL 15000 8000  30000 3100  400000 2300 AL 15000 9000  30000 4900 MWL 15000 10600  20000 10200  30000 5200  40000 2400  199085 KL 0.1M NaOH MW: 41800 198479  2.2.3.3 Membrane osmometer Membrane osmometer analysis indirectly measures the number average molecular weight of a polymer. A semipermeable membrane separates one chamber containing pure solvent and the other containing the polymer solution. The osmotic pressure across the membrane is measured and used to calculate the number average molecular weight87. The problem for lignin analysis is that lignin is so heterogeneous that many small molecules can pass through the membrane, and therefore impact the overall MW estimation.    29  2.2.3.4 NMR end group titration NMR end group analysis quantifies the end groups of a linear polymer, which equals the number of polymer chains. By referring to the structure of the repeat unit and total analyzed amount, the degree of polymerization and the number average molecular weight can be derived88. It is a new method in the area of lignin analysis, but has received a significant attention. However, the major debate concerns a clear definition of the end group in lignins. In the literature88, milled wood lignin end groups are considered as phenolics, and number of lignin chains = phenolic OH – (𝛽-𝛽′ + 𝛽-1 + 5-5’OH). However, the analysis is inaccurate with branching occurring in the lignin. In addition, this method is not applicable to technical lignin, where a clear repeating backbone structure does not exist. When applied to native-like lignin, measured MW was lower than the GPC method (Table 2.9). The authors argued that the difference resulted from extensive aggregation in GPC separation; hence the results were more similar with VPO and MS analysis where aggregation does not impact the measurement. Despite all the limitations, this method avoided the aggregation effect in GPC analysis. Table 2.9 MW from NMR titration and GPC method Lignin GPC (Da) NMR (Da) S-MWL 14200 1800 S-EMAL 7300 1400 Hard MWL 9600 2600  2.3 Light scattering and light scattering in lignin characterization Light scattering is a technique that provides measurement of the absolute weight average molecular weight, radius of gyration, and the second viral coefficient (A2) of a polymer in solution. More information, like polydispersity index (PDI) and number average molecular weight (Mn), can be determined in a single run by combination with a separation method like  30  GPC. Other techniques, like viscometer and refractive index, can also be coupled with light scattering so that hydrodynamic radius and refractive index increment (dn/dc) can be measured.   2.3.1 Introduction to light scattering theory  Light is an oscillating wave of electric and magnetic fields. From a physical standpoint, light interacts with matter in five ways: transmission, absorption, reflection, fluorescence, and scattering89. In light scattering, the matter emits light in all directions. To be specific, when the radiation interacts with a particle, light causes its subatomic particles to become polarized, and the oscillating charges radiate light in all direction at the same wavelength as the incident light.  With a given incident light, the scattered intensity of the radiated light depends on the polarizability of the matter, molar mass, concentration, and the scattering angle90,91. The polarizability and angular variation are related to the index of refraction and size of the particle, respectively92. Therefore, by measuring the 1) intensity of light scattering at different angles, 2) the concentration of polymer, and 3) index of refraction, the weight average molar mass and subsequent distribution can be determined. The mathematic relation is expressed follows: 𝐼(𝜃)𝑠𝑐𝑎𝑡𝑡𝑒𝑟𝑒𝑑 ∝ 𝑀 ∗ 𝐶 ∗ (𝑑𝑛𝑑𝑐)2∗ 𝑃(𝜃) 𝐾 ∗𝑐𝑅(𝜃)=1𝑀 ∗ 𝑃(𝜃) 𝑃(𝜃) ∝ 1 −  𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡 ∗  𝑠𝑖𝑛 (𝜃2)2∗ (𝑟𝑔)2 𝑅(𝜃) = 𝐼(𝜃)𝑠𝑐𝑎𝑡𝑡𝑒𝑟𝑒𝑑/𝐼(𝜃)𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 where 𝐼(𝜃)𝑠𝑐𝑎𝑡𝑡𝑒𝑟𝑒𝑑 is the intensity of light scattered, M is the absolute molecular weight, C is the real time concentration, 𝑃(𝜃) is the angular dependence as a function of angel (𝜃), K is a constant as a function of dn/dc, and 𝑟𝑔 is the gyration radius.  In light scattering analysis, scattered light intensity is plotted versus angle to better solve the equations and analyze goodness of fit, and this plot is known as Zimm plot (Figure 2.2).   31   Figure 2.2 Zimm plot 2.3.2 Light scattering in lignin characterization  Light scattering has been applied in lignin absolute molecular weight characterization for many years93,94,95. The major advantage of using this method is avoiding potential errors resulting from calibration. When light scattering is coupled with differential viscometer and refractive index detection, more information regarding the molecular conformation and hydrodynamics can be derived. Based on these methods for lignin, the MW from light scattering analysis are generally in the range from 8K – 100K (Table 2.10). It is generally an order higher than GPC conventional calibration results. The abnormally high MW from light scattering is debatable, and lacks validation from other experimental methods. Many plausible explanations are raised and three of them have received wide acceptance. One explanation regards the structural difference between lignin and the calibration standards as lignin is more compact, causing the GPC methods to underestimate the size. The others relate to aggregation and fluorescence that cause light scattering values of lignin to be larger95. Major research topics related to light scattering include absolute MW, lignin aggregation behaviour, fluorescence, and different solvent system were summarized in Table 2.10. 32   Table 2.10 Paper review on light scattering on lignin Lignin Solvent system Characterization result Conclusion and notes Ref Lignosulfonate  Fractionated by H2O/EtOH NaCl buffer  Fractions MW, 436nm MW, 546nm 1 8600 12300 2 15200 15700 3 21600 19200 4 26400 24400 5 37900 40000 6 48800 52600 7 138000 122000     MW↑⇒ fluorescence↑ Linear relationship between MW & diffusion coefficient ⇒ lignin: non-rigid chains 195593 Kraft lignin Low angle LS  DMF – 318.2K DMF – 350.7K 0.5N NaOH 303.2K  Lignin MW [𝜂] DMF Solvent a F1 22017 7.0-7.2 DMF 0.11 F2 28620 7.5-7.6 DMF 0.13 F3 36078 7.5-7.7 NaOH 0.23 F4 48442 7.8-7.9   F5 59528 8.1-8.2        Lignin spherical particles, and slightly solvated with solvent  199563 Lignosulfonate   Fractionated by H2O/EtOH SEC + MALLS Aqueous phosphate (Na2HPO4-H2O) +  DMSO and SDS, pH 10.5 633nm, 60°C,  Fractions  MW PDI dn/dc Raw  64000 8.8 0.195 F-70 4600 1.5 0.186 F-60 8000 1.3 0.196 F-55 15000 1.5 0.200 F-50 34000 1.9 0.204 F-45 68000 2.3 0.203 F-40 398000 3.5 0.205      Shows upswing curve at the end;  A2: 7 × 10−3, high MW↓⇒S% ↑0.39 – 0.64  No Rg  200259  33  MWL, G-DHPs   THF fractionation SEC +MALLS DMF THF/THF LiBr (acetylation) 632.8nm, Fluorescence filter  MW: 105 – 108; Rg: 100-150nm Lignin MW-THF PDI MW-LiBr PDI MWL-R 21700 4.3 24700 3.2 MWL-S 14700 3.4 19000 3.0 MWL-IS 54000 4.8 37300 2.4  Shows upswing profile;  Bimodal, which doesn’t disappear with time MW↓ ⇒ Rg↑ slightly 200396 SKL,  NaOH at different pH,  DMSO, THF-LiBr,  No acetylation  Different injection amount; storage time; Solvent  system,  MW: 104 – 106 Lignin aggregated at pH below 8.5 Very broad and complex peaks  200694 SKL (NL and NREL, LL( unlyophilized) MALLS + 9 solvent   + AFM  Lignin MW Rh Rg Dn/dc NL/water 5.4 × 105 33 46.7 0.19 NL/pH 12 5.0 × 105 19.5 19.8  NL/pH 14 6.9 × 104 30 15.6 0.20 NREL/water 4.8 × 105 25.3 24.8  NREL/pH 12 1.4 × 104 8.4 16.6  NREL/pH 14 2.7 × 104 33.5 36.6  LL/water 6.2 × 105 12.5 20.9  LL/pH 23 1.9 × 104 19.6 --- 0.25 Water to NaOH ⇒ dn/dc↑, MW, rg ↓; AFM detects rh  Lyophilization support aggregation  200697 EMAL softwood & hardwood THF & 0.01N NaOH; 633nm, 25°C; Acetobromination  1nm & 10nm filter  Lignin MW A dn/dc EMAL-S-THF 1.5 × 105 3.0 × 106 0.28 EMAL-S-NaOH 5.4 × 105 1.5 × 106 0.18 EMAL-H-THF 1.1 × 104 1.1 × 104 0.17 Days↑⇒ dn/dc ↑, MW↓ 1nm filter band, 60% overestimation; 200898  34  sodium lignosulfonate (NaLS) fractions  GPC – 0.1N NaNO3 LS - NaNO3 / NaCl DLS  Lignin  NaNO3 MW GPC MW LS Dn/dc Rg Rh F1 5K 14K 1.77 58.86 53.58 F2 10K 50K 1.77 55.86 51.14 F3 19.2K 263K 1.79 51.02 48.47 F4 27.5k 634K 1.82 44.07 44.3 NaCl      F1  8.7K 1.8  3.5 F2  22.5K 1.8  3.4 F3  82.5K 1.8  4.3 F4  157.6K 1.8  5.1       Slow mode in DLS eliminates aggregation  NaCl ⇒ A2 ↓ , not good solvent  201399 Technical lignin SEC + MALLS DMSO + LiBr  658/785 nm  Lignin MW (kDa) PDI dn/dc Indulin AT 14 2.4 0.15 Lignoboost 16 3.8 0.16 Soda 19 5.3 0.15 OSL HW 10 3.1 0.14 MWL 27 4.9 0.12 DHP 26 5.5 0.11 Biolignin 187 62.3 0.12      MW↑⇒ dn/dc ↑; 785nm has lower fluorescence;  201895  35   2.3.3 Issues in the light scattering of lignin  Light scattering is very powerful method in polymer chemistry and biological material analysis94,95, however, it is underutilized in the field of lignin analysis. Light scattering results in different results from GPC calibration, and the repeatability is also low. Four major issues hinder the reliability of lignin absolute molecular weight determination from light scattering, namely aggregation, fluorescence, the small size of lignin, and heterogeneity of lignin where large particles dominate scattering results.   2.3.3.1 Aggregation  The first issue relates to the formation of aggregates, which are clusters of lignin molecules assembled together by intermolecular forces79,80. In a light scattering experiment, such a cluster is characterized as a single particle with a molecular weight representing the sum of a number of lignin molecules. Hydrogen bonding is considered the major intermolecular force driving the process simply because of the existence of hydroxyl groups on lignin molecules. However, exhaustive acetylation of lignin does not avoid high molecular weight complexes, when hydroxyl groups are masked. Alternatively, Sarkanen raised suggested that π – π stacking due to homo-lumo interactions between aromatics governs the associative process79.  The existence of aggregations is supported by four factors: the abnormally large MW up to 106Da as characterized in light scattering experiments51,100,59, the bimodal profiles found in various chromatographs51,57, the instability of lignin in solution (lignin property changes with environment, like incubation time, concentration, temperature, and additives),75,101,102,101,98,51,103  36  and the effect of lignin derivatization (lignin modification has great significance on lignin structure)104,73.  Different methods were developed to limit lignin aggregations and improve solubility by lignin derivatizations (acetylation, acetobromination, and silylation), different buffer systems (NaOH, NaOH + LiCl, DMF, THF, DMAc), the addition of various additives (quaternary ammonium pairs54,105, LiCl/LiBr salts), sonication106, purification107, and controlling the injected lignin concentration56. Nevertheless, none of the methods were fully effective for the removal of aggregates and creates issues for a reproducible solution amongst different laboratories.    2.3.3.2 Fluorescence and absorbance  When excited by light or electromagnetic radiation, the excited electrons at a higher energy orbital is paired with the second electron at the ground state orbital. As the excited electron returns to the ground state, the energy emits in the form of a photon at lower energy of the incoming light108. Fluorescence usually occurs for substances with conjugated structures, for example, quinine (Figure 2.3 [a]) 109,110. Natural lignin also contains a limited conjugated structure (Figure 2.3 [b]), and more conjugated structures form as a result of condensation in the kraft pulping process (Figure 2.3 [c]). Fluorescence scattering intensity has been shown to be further enhanced by aggregations induce emission (AIE) in a solvent100, 111.   37     [a] Quinine [b] G lignin [c] Condensed lignin monomer through 5-5 bond Figure 2.3 Examples of conjugated structure  Fluorescence leads to an overestimation of lignin MW in light scattering. Fluorescence is quite well detected by light scattering experiments, as detectors with filter band have a lower signal response59,100. In addition, Raman spectroscopy using 785nm also detects lignin fluorescence112. However, there is no published data on lignin fluorescence intensity excited in the infrared region.  There are three major ways to decrease the fluorescence of lignin. The first method uses higher wavelength incident light as the excitation; fluorescence was the strongest from 350-450nm, and decays with increasing of wavelength113. Therefore, the wavelength of incident light used in lignin light scattering is usually 633nm or higher. The second is to add fluorescence filters. As the wavelength of fluorescence is always higher than the incident light, the filters can be installed on detectors to remove light at a higher wavelength. The last method is adding fluorescence quenchers such as iodide and bromide. It is commonly used for protein characterization, but currently not employed in lignin research109.    2.3.3.3 The small size of lignin  In a light scattering experiments, there is theoretically no angular dependence for a polymer smaller than 10nm, which is empirically equivalent to 100kDa for a linear polymer. The  38  threshold MW shall even be larger for a branched material, like lignin. Therefore, lignin theoretically has no angular dependence. As mentioned above, light scattering data is plotted in a Zimm plot, and linearly extrapolated to determine the molecular weight. However, usually lignin has poor linear fitting in Zimm plots, so inherent variability is observed in this system.   2.3.3.4 Heterogeneity of lignin  The last issue involves the heterogeneity of lignin and irregular structure. First, lignin usually has a complex light scattering profile due to its wide MW distribution. As light scattering is more sensitive to high MW particles, the RI signal only relates to the concentration of the analyst, and there should only be a small overlapping region between the light scattering and RI peaks. As a result, part of the lignin cannot be accurately characterized due to an insufficient signal from either light scattering or RI. The second reason relates to the incapability of GPC separation arising from lignin-column interactions. If lignin molecules are not separated effectively, high MW lignin can overshadow low MW lignin. Last, different lignin fractions have different dn/dc value that should be predetermined prior to light scattering experiments, and these values are dependent upon both lignin structure and aggregation56.   2.4 Lignin molecular conformation characterization In polymer physics, a polymer is defined mainly by architecture, degree of polymerization (DP), and chemical composition, all of which influence the molecular conformation. Molecular conformation is defined as the spatial structure of a polymer or a description of the relative location of its repeat units; it is also referred to as “molecular shape”114. Molecular conformation depends upon three factors: chain flexibility/stiffness, intramolecular interactions between  39  repeating units, and the interactions with the surrounding solvent87. Examples of molecular conformation include a sphere, linear coil, or rod shape. Moreover, polymer conformation changes with the surrounding environment. For example, a polymer in a good solvent tends to be a more expanded shape, while it will collapse in a poor solvent. A theta solvent, at a given temperature, has equal interactions among the solvent and the secondary interactions along the chains. Further, polymer topology refers to the molecular architecture or configuration of a polymer, such as linear, branched, crosslink, ring, star-branched, or ladder. It has a major impact on the properties of a polymer, and the topology is formed during synthesis. Conformation and topology are two inter-related concepts and one can deduce the other depending on certain assumptions.   2.4.1 Conformation and topology characterization methods  The topology of a molecule can be understood from either the chemical structure or conformational properties. Hydrodynamics is used to study the molar mass, conformation, and interaction properties of the polymer in solution conditions by observing the relative motions of the polymer to the solvent115. The diffusion coefficient, also known as diffusivity, describes the diffusion mobility of polymers in the solvent. The higher the diffusion, the faster the solute and solvent diffuse into each other96. It is further classified into two concepts: rotational diffusion associated with the overall orientation of particles or molecules, and translational diffusion (Dt) measuring the particle position in space. The Dt coefficient is proportional to temperature and inversely proportional to particle size, sphericity friction (fs), and inter-particle interactions. Hydrodynamic radius is “the radius of a hypothetical hard sphere that diffuses at the same rate as the particle under examination116.” In other words, it is the radius considering Brownian motion.  40  The parameter is affected by molecular size, solvent-solute interactions – excluded volume effects, and solvent viscosity. Another related parameter is the intrinsic viscosity ([η]), which is a measure of a solute’s contribution to the viscosity (m2/s) of a solution and calculated as below. In real measurements, it equals to specific viscosity for a dilute solution.  Relative viscosity: 𝜂𝑟 =  𝜂𝜂0 Specific viscosity: 𝜂𝑠𝑝 =  𝜂−𝜂0𝜂0= 𝜂𝑟 − 1 Reduced viscosity: 𝜂𝑟𝑒𝑑 =  𝜂𝑠𝑝𝑐=𝜂𝑟−1𝑐 Inherent viscosity: 𝜂𝑖 =  𝐼𝑛(𝜂𝑟)𝑐 Intrinsic Viscosity: [𝜂] = lim𝑐→0𝜂𝑠𝑝𝑐   As lignin is too small to be directly observed, most structures of lignin molecules are deduced based on the measurement and analysis of the above physical parameters. The following context introduces the techniques that have been applied in lignin conformational analysis and the corresponding results.   2.4.1.1 Viscometry and pulsed gradient spin echo (PGSE) – NMR Viscometry and PGSE-NMR are two techniques relying that help to understand the relationship amongst the diffusion coefficient (D), molar mass (MW), and intrinsic viscosity117. An empirical equation is developed to relate hydrodynamics to polymer architecture as follows:  [𝜂] = 𝐾𝜂 ∗ 𝑀𝛼   𝐷 = 𝐾𝑑 ∗ 𝑀−𝑏 where, Kη and Kd are constants, and α and b value are affected by the polymer-solute system. For a polymer as a random coil in a solvent, α = 0.5 indicates the theta condition, and higher values indicate better solvents, as the polymer chain will interact with the solvent. The  41  relationship between the α value and polymer architecture for a dilute solution is shown in Table 2.11.   Table 2.11 Molecular architecture and MHS plot parameter α for dilute solution 𝜶 Polymer architecture 0-0.5 Solid spheres 0.5-0.8 Linear random coils 0.8-2 Stiff chain, rigid rod   In GPC software analysis programs, such as Astra™, Rh can also be derived from viscosity assuming lignin is an Einstein sphere, which is a non-interacting rigid sphere. Rh is calculated from the following equations, where ϕ is the volume fraction and γ = 2.5 for spheres and larger for nonspherical particles, Mw is the absolute molecular weight and Vh is the hydrodynamic volume of the molecules. [𝜂] =𝛾𝑁𝐴𝑉ℎ𝑀𝑤 𝜂 = 𝜂0 (1 +  𝛾𝜙) 𝑟ℎ = (3𝑉ℎ4𝜋)13  2.4.1.2 Dynamic light scattering (DLS)  DLS detectors measure the fluctuation of light scattering signals, based on the translational diffusion coefficient (Dt) of the solute; this parameter is converted to the hydrodynamic radius by referring to Stokes-Einstein equation: 𝐷𝑡 =𝐾𝐵𝑇6𝜋𝜂𝑅ℎ  42  where, KB is Boltzmann’s constant, T is temperature (Kelvin), η is the viscosity of the solvent. Furthermore, by measuring Rg from light scattering or X-ray scattering, the ratio of Rh and Rg reveals the shape of a polymer; Rg/Rh is generally greater than 1 for extended architecture and decreases for compact structures. For instance, a ratio is 0.77 for solid sphere, 1.4 for a three-star polymer, and 1.65 for linear flexible polymers97,116.   2.4.1.3 Small angle neutron scattering (SANS) and small angel x-ray scattering (SAXS) X-rays and neutrons can be used in scattering experiments at small angles to investigate the architecture of polymers like lignin 118,119. X-ray and neutrons have rather small wavelengths that can penetrate molecules and reveal more detailed information on its conformation and topology while avoiding fluorescence issues. The characterization of size ranges from 1nm to 1000nm.   2.4.1.4 NMR chemical structure analysis Apart from detecting functional groups, a molecular architecture (branches) can also be determined from NMR by measuring the linkages and functional groups of the repeat units. For example, a repeat unit with a functionality of two polymerizes into a linear polymer, and a mixture of monomers of two and three functional groups would form a polymer with a certain degree of branching. Quantitative calculation of the number of branching points is given below120. 𝐷𝐵 =𝑇+𝐿𝑇+𝐿+𝐷   For higher molar mass;  𝐷𝐵 =2𝐷2𝐷+𝐿   For oligomers;  43  where, DB is the degree of branching, T is terminal units, one functionality reacted, L is linear units, two functionalities reacted, and D is dendritic units, three functionalities reacted (Figure 2.4).   Figure 2.4 Repeat units with different functionality, represented by T = terminal units; L = linear units; D = dendritic units120  2.4.2 Lignin conformation analysis  The understanding of lignin structure has a long history (Table 2.12), and there is still a lack of a comprehensive picture of its structure and conformation , although most of the techniques described above have been utilized. There are three major phases throughout the analysis on lignin structures. At the very beginning, most researchers described lignin as a compact sphere particle such as a rigid sphere, swollen sphere, compact sphere, as well as impermeable coils or microgel fragment particles. Three major pieces of evidence support this conclusion. First, lignin in the solid state was shown as particles when analyzed with SEM121. Second, studies on the hydrodynamics of lignin generally gave a small viscosity and low Mark-Howink parameter “α”. An α value from a viscometer or b value from PGSE-NMR of 0.1-0.3 is an indicator of a hard sphere122,123,52. Third, lignin polymerized from three basic monomers, among which, G and H lignin has a functionality of three or more. From the view of polymerization, it should naturally  44  yield a crosslink polymer. Researchers also pointed out that the spherical particle differs in hydrodynamic behaviour with the surrounding environment65. Because of this observation, some proposed lignin as a swollen particle or a microgel surrounded by loose coiling chains. During the second phase of investigation, some researchers argued lignin may be an irregular shape, such as a disk-like shape, a crosslinked sheet, or a rigid rod. First, a lignosulfonate film characterized by electron microscope showed a thickness of 1-3nm independent of molecular mass, which is much lower than the size of the other two dimensions124. Next, from the view of hydrodynamics, researchers acquired a shape factor of 4.9 for kraft lignin using pulsed field gradient NMR (PFGNMR), which indicated an oblate structure with an axial ratio of 18125,126. Milled wood lignin and dioxane lignin showed the same feature that one dimension was much smaller than the other two. Another few papers investigated lignin structure by SAXS and SANS122, 119, 127. They observed clear evidence of lignin aggregated in the shape of a cylinder in both DMSO and NaOD. The radius was less than 1nm, and the length ranged from 2 to 10nm, or higher. The third phase is still under development, with two major aspects. First, the technical lignin structure differs from its isolated fractions. Crestini et al. (2017) compared the NMR spectrum of acetone fractionated lignin and found acetone soluble lignin has more side chains and shorter backbones than acetone insoluble part indicating a more compact structure of ASKL104. Second, MWL might be a linear oligomer as suggested from NMR analysis88. Currently, more researchers have analyzed lignin by NMR to avoid the debatable aggregations in wet chemistry. Because of this, new discoveries might be made if we can combine the classical techniques and new NMR analysis to characterized well-fractionated lignin88,21,128. 45   Table 2.12 Reviews on lignin hydrodynamics Sample Technique Findings Ref Hydrodynamics  Alkali lignin LS + Viscometer Sedimentation equilibrium NaHCO3-NaOH  Fractionated lignin MW: 50K – 48000K; [𝜂] = 𝐾 ∗ 𝑀𝑤0.32 ⇒ hard sphere; 𝑆𝑤 = 𝐾 ∗ 𝑀𝑤0.52 ⇒ random coil; MW↑⇒ Branching ratio, g (0.92-0.59) ↓; 1960129 Lignosulfonate Viscometer  LS 0.01M-0.1M NaCl  MW: ~37K – 1300K; Rg: 50-100nm [𝜂] = 𝐾 ∗ 𝑀𝑤0.33;  𝐷 = 𝐾 ∗ 𝑀𝑤−0.33 [𝜂] & Rg↑⇒ IE↓ Lignosulfonate: free charges on the surface; “Microgel particles surrounded by a layer of loosely coiling chains” 196065 Lignosulfonate EM Flat disk: 10nm diameter, 2nm thickness  MW: 143K  Thickness: 𝑇 = 4 ∗𝑀𝑤𝜋𝐷2𝑁𝜌 1979124 SKL; Organosolv;  Preparative SEC;  SEC 0.5M NaOH; ultracentrifugation;  0.1M NaOH;  Expansion factor: 2.5-3.7; “Expanded random coil conformation without effects of long chain branching”; MW: 1680 ~ 140,000 from ultracentrifugation;  198280 Lignosulfonate  Electric field  UV-viscometer  “Changes conformation from a compact sphere to a non-free unwinding coil under electric field.”  1991130  46  Kraft lignin PGSE-NMR  Unacetylated -NaOD 0.1M Acetylated -CHCl3 1.0M Unacetylated: Rh: 2.05-2.28 nm,  𝐷𝑠~𝑀−𝑎  𝑎 = 0.39  Acetylated:  Rh: 0.5-1.31 nm,  𝐷𝑠~𝑀−𝑏  𝑏 = 0.3 ⇒ branched macromolecule; [C] ∝ 1/Ds ⇒ Compact and flat conformation; Shape factor (axial ratio): Perrin Equation: oblate structure + axial ratio ~ 18 Assuming at ‘0’ [C] limit & Flory- Fox polymer; 1991125 Kraft Lignin DLS turbidity, PGSE-NMR OH-, ionic↓ or T°C↑ ⇒ Lignin particle → colloidal → coagulation; Fractal dimension (dt): 𝑅ℎ~𝑡−𝑑𝑡 Stability ratio↑⇒ Fractal dimension↑2-2.4; Aggregations: 100 nm - 1-2µm; 2002131 lignosulfonate Viscometer   MW: 1.3K – 38.7K [𝜂] = 𝐾 ∗ 𝑀𝑤0.15;      [𝜂] = 7.4 − 13.3𝑚𝑙𝑔 2003117 Kraft lignin  PFG-NMR; SEC  0.1 M NaOH/NaOD MW: 1910 - 41400 1.0 to 2.2 nm  Lignosulfonate  Literature data  Viscosity and MW  𝐿~𝑀𝑤 .70±0.05; 𝑟~𝐿.70±0.05; (L: backbone length, r: radius) A randomly branched polyelectrolyte; Dependence of polyelectrolyte expansion on MW ⇒ not microgel; 200864  Biology point of view  A cross-linked network polymer  2004132  47      SAXS & SANS SKL; S organosolv; steam explosion; soda; MWL SAXS USAXS Dry state &0 .1 NaOH Dried lignin: Specific area: 0.5 to 60m2/g: MWL >> technical lignin Surface fractal dimension (Ds): 2-3 Lignin in 0.1N NaOH: Rg: 1-3 nm thick, while the length (5-9 nm) in solution;  2004118  Organosolv lignin  Lignin: “rigid and complex, ranging from nanogels to hyperbranched macromolecules, not linear oligomers or physical assemblies of oligomers122”  SKL;  Organosolv HW & SW  Wheat Lignin  SANS (in DMSO): lignin associate to form rigid rod-like/cylindrical  (2-10nm length, 0.5-1nm radius) Lignin branching degree affects association behaviour SKL HW: Lignin [C] ↑⇒ cylinder size↓ & Rg↓(500-100nm)[C]> overlap [C] Wheat Lignin [C] ↑⇒ Rg↑ 2015119 SKL, HKL, corncob soda lignin SANS SANX Relation between Aggregation & functional groups and substructures  Aliphatic↑ & G content (𝜋 − 𝜋) & chain entanglement↑ ⇒ aggregation↑; In DMSO: R: 0.6-1.1nm, L: 5-9nm; NaOD: R: 0.4-0.6, L:6.2-10.6nm133 2017133 Softwood L SANS/SANX Lignin surface dimension is invariant over the aggregates length scale from 0.1-100nm127. 2011127      48  NMR analysis     Soft & hard  MWL  NMR (QQ-HSQC, 31P, DFRC)  End group: polymeric units:  C9 formula, the amounts of phenolic groups, (𝛽-𝛽), (𝛽 -1), and (5- 5’) ⇒ DP DP from NMR data ⇒ linear oligomer (not affected by aggregation) Number of intermonomer linkages ≈ Number of units; ⇒ low branching degree 201188 Soft & hard  MWL NMR  Softwood: 90% G, hardwood: S≈G%; + linkage information + computational analysis⇒ hardwood lignin more linear structure.  201521 MWL  Fractionation  NMR  MW↑⇒p-hydroxyphenyl↑; linked to linear arabinoxylan MW↓⇒G lignin & guaiacyl units↑+ tricin unit; “arabinoxylan chains esterified by ferulic acid and cross-linked via ferulic acid dimerization” 2016128 SKL Acetone fractionated 1D & 2D NMR, SEC & MS  MW: 6000; PDI: 6.2 Acetone Insoluble: branched polymeric material;  Acetone soluble: more branched, less polymeric.  2017134  49  2.4.3 Branching degree  Quantitative data about branching topology is essential to understanding the structure of lignin, where branching ratio is defined as the size ratio between the branched polymer and its linear counterpart135. The fundamental assumption is a polymer of the same DP decreases in size with an increase in the degree of branching (gM).  𝑔 = (𝑅𝐵𝑟𝑎𝑛𝑐ℎ𝑅𝐿𝑖𝑛𝑒𝑎𝑟)2 𝑀 Different, but equivalent, methods were introduced to deal with a polymer possessing an unknown radius, including the use of viscosity measurement techniques. This is the case for lignin because it is too small to accurately measure the radius with light scattering.   2.4.3.1 Viscosity method  By comparing the intrinsic viscosity of branched polymers with its linear counterpart, at the same MW, the branch ratio (gM) can be derived135. The equation is expressed as below, where [η] is intrinsic viscosity, “e” is the drainage parameter, M represents the same molecular weight. 𝑔𝑀𝑒 = 𝑔′ = ([𝜂]𝑏𝑟𝑎𝑛𝑐ℎ𝑒𝑑[𝜂]𝑙𝑖𝑛𝑒𝑎𝑟)𝑀 Branching degree (gM) is a very abstract concept, but it can be related to branching unit per molecule and branch frequency (𝜆) which describes number of branch points per 100 units. Assuming lignin only contains bifunctional and trifunctional units, the number of branching units per molecule (Bn) and branch frequency (𝜆) is given in the following equations:  𝑔𝑀 = [(1 +𝐵𝑛7)12+4𝐵𝑛9𝜋]−12  50  𝜆 =𝑈𝑛𝑖𝑡𝑀𝑤 × 𝐵𝑛 × 100𝑀𝑤 In all, the viscosity method requires measurement of the absolute MW and intrinsic viscosity of the analyte and a linear reference standard. Assuming lignin only contains trifunctional units as the branching points, the number of branch points can be quantified from the view of hydrodynamics.     51   Hypothesis and objective  As a by-product of the pulp and paper industry, lignin is an abundant and underutilized raw material. There are many obstacles that hinder the use of lignin, especially the complexity of its structure and a lack of reliable characterization methods. As such, the overall objective of my thesis was to better characterize the absolute molar mass of lignin and molecular conformation by light scattering. Two major topics are discussed in my work. The first topic relates to utilizing the GPC conventional calibration method (Figure 3.1[a]), which is currently the most popular method in lignin molecular weight characterization. The major advantages of the method are ease of operation with well-developed theory and working procedures and good intra-laboratory repeatability67. Nevertheless, the intrinsic disadvantages of this method are poor inter-laboratory reproducibility and incapability of characterizing absolute molar mass. This issue is because lignin molecules are structurally different from calibration standards, and moreover, technical lignin structure varies with types and fractions so that no standards can represent the entire lignin distribution or different lignin types. In addition, due to its inherent branched structure, lignin may interact with column packing materials resulting in a delay of elution and an underestimation of molecular weight.     52   Figure 3.1 Scheme of the thesis. [a] first topic relates to issues of GPC; [b] the second topic relates to the branched structure of lignin Therefore, an absolute molar mass characterization technique is necessary because it provides information that is abs independent of separation method. In our experimental setup, light scattering was selected as it employs absolute Mw characterization of a polymer95. However, apart from the high initial cost, light scattering has four major obstacles in lignin characterization: aggregation, fluorescence, small lignin size, and heterogeneity of lignin size and structure. For minimizing the aggregation effect in THF buffer, lignin was acetylated prior to analysis and incubated for 48 hours prior to analysis. Peak deconvolution was conducted to analyze the effect of aggregations. Further, 785nm incident light was selected and 10nm width  53  filter bands were installed to remove most of the fluorescence. In terms of the large variance due to small molecular size, at least six repeats were performed for each sample and the average value used in this work. In regard to the heterogeneous nature of lignin, I hypothesized that fractionation could improve the of light scattering profile and reveal the structural difference amongst different fractions. Without fractionation, low MW lignin will be overshadowed in light scattering, as lignin has a wide Mw distribution and light scattering is more sensitive to high Mw particles. In addition, the dn/dc value, which depends on polymer molecular weight and varies with fractions, was predetermined for each lignin fractions to provide a better calculation of the MW.  The combination of acetone and tangential flow fractionation (TFF) was used for softwood kraft lignin (SKL) fractionation because solvent fractionates lignin based on its solubility or polarizability, while TFF fractionates based on the size of molecules. Thus, lignin fractions that are homogeneous in both chemical structure and size were obtained. In order to help validate the light scattering result and explore the potential of light scattering, intrinsic viscosity was determined by differential viscometer. Based on viscosity properties, hydrodynamic behaviour and molecular conformation were evaluated by the adoption of Mark-Houwink-Sakurada equation. The result of molecular conformation analysis was further compared with chemical analysis as determined by NMR. Such a comparative evaluation of results from three techniques should provide more information on lignin structure and lend credence to the light scattering results.  Another topic related to the lignin is understanding its branched nature (Figure 3.1 [b]) and potential interaction with the separation efficiency in columns. An upswing curve was found in the lignin GPC and light scattering profile of Mw against elution volume (Figure 5.3). Such a  54  “hook” shape might result from two reasons. One explanation is the interaction between GPC column and branched lignin resulting in a delay in elution of high Mw particles. The other possibility is the stronger fluorescence emission from smaller molecular weight lignin, which elutes at the end and leads to an overestimation of the molar mass. Therefore, I attempted to understand this “hook” profile and proposed that that we can quantify the degree of branching by comparing light scattering profiles of technical lignin with a linear reference standard. This method was introduced by Zimm, known as the branch ratio equation78. The key to this method is to find a “linear” reference sample with a similar unit structure as lignin and have a reliable measurement of the absolute molar mass and intrinsic viscosity. In my experimental design, hardwood milled wood lignin from eucalyptus was used due to its high syringyl content (70%), lower condensation degree (approaching zero), and similar chemical composition. Last, I hypothesize that AF4 might minimize the interaction of lignin and the “system”, and outcompete the GPC system in lignin characterization. AF4 has been used to successfully characterize several branched polymers136, and I tried to conduct similar work with to lignin. This chapter is included in the Appendix E.   55   Materials and experimental methods 4.1 Materials  Industrial softwood kraft lignin (SKL, Indulin-AT) was provided from WestRock Company, SC, USA. Lignin was dried under vacuum at 50 °C and stored in a vacuum desiccator prior to all experiments. All reagents were obtained from Fisher Scientific, VWR Corporation, and Sigma-Aldrich, and used as received. Tetrahydrofuran (THF) employed for GPC was analytical grade, and deuterated chloroform (CDCl3) was obtained from Cambridge Isotope Laboratories.  4.2 Sample preparation and flow chart 4.2.1 Softwood kraft lignin fractionation  All lignin preparations indicated below were duplicated. Prior to fractionation, lignin was first washed with acidic water to remove impurities. 100g of SKL was vigorously stirred for 8 hours in 1 litre of acidic water (0.01M hydrochloric acid, pH =2). The lignin solid was then separated via filtration under vacuum using Whatman filter paper. The washing process was repeated twice. Lignin solids were then washed with distilled water, air-dried overnight, and vacuum dried at 50°C.  4.2.2 Acetone fractionation  Acid washed and dried lignin (100g) was dispersed into 500mL acetone incrementally and stirred for 6 hours at room temperature. The acetone insoluble fraction (AIKL) was isolated via vacuum filtration. The acetone soluble fraction (ASKL) was air dried to remove acetone, and the solid fraction was further dried in a vacuum oven at 50°C.    56  4.2.3 TFF fractionation (ultrafiltration)  Acid washed, acetone-soluble (ASKL) and acetone-insoluble (AIKL) lignins were further fractionated by tangential flow fractionation system (TFF). The cross-flow ultrafiltration was carried out on a Spectrum system equipped with a polymeric membrane column, a peristaltic pump (KrosFlo TFF system, KMPi), three feed permeate and retentate pressure sensors, a back-pressure valve transmembrane pressure (TMP) controller, a process reservoir, and a collection tank. The arrangement of the TFF system is shown in Figure 4.1. Two membrane columns (molecular weight cut-offs of 3kDa and 10kDa) were made from modified polyethersulfone (mPES), with a diameter of 3.5 cm, surface area of 540 cm2, and maximum operating pressure of 30psi. All parts of the TFF system were purchased from Spectrum Lab (Repligen, Inc.).          57   Figure 4.1 Laboratory arrangement of a TFF system  Figure 4.2 Lignin fractionation process and lignin fractions obtained Acid washed and dried lignin was dissolved in 0.1N sodium hydroxide (NaOH) at a concentration of 10g/L and stabilized for 48 hours. Prior to the fractionation, the column was cleaned with 0.1N NaOH and rinsed with distilled water. An integration test was conducted to ensure no cracks or damages to the column by circulating distilled water for 20 minutes prior to the experiment. Lignin liquor was first screened through a 0.65 m column filter to remove the insoluble particles and ash. Then, the batch mode was used where lignin liquor was passed  58  through the membrane (buffer permeate) until a volume reduction factor of five was achieved, corresponding to a volume reduction of 80%. This indicated that over 90% of the sample was successfully fractionated according to technical notes137. The process was performed at room temperature, and the transmembrane pressure (TMP) was increased stepwise, from 2 to 5 psi. The resultant fractions, including <3kDa, 3-10kDa, and >10kDa, were precipitated by slowly dropping the pH to 2 with HCl. The precipitated lignin was then washed by centrifugation twice with acidic water and then distilled water, respectively. The fractionated samples were freeze-dried and stored in a desiccator. The overall process is shown in Figure 4.2. In total, nine different lignin fractions were obtained – raw lignin (SKL), ASKL, AIKL, ASKL >10kDa, AIKL >10kDa, ASKL 3-10kDa, AIKL 3-10kDa, ASKL <3kDa, and AIKL <3kDa.  4.3 Enzymatic milled acidolysis lignin preparation  Enzymatic milled acidolysis lignin (EMAL) with high syringyl (S) content was used as representative to native lignin. Eucalyptus wood chips are used for extractions. The process flow chart is shown in Error! Reference source not found.   59   Figure 4.3 Enzymatic milled acidolysis lignin (EMAL) isolation process flow chart  4.3.1 Extractive-free wood fiber preparation  The wood chips were first ground to pass a 20 mesh on a Wiley mill. Extractives were removed by Soxhlet extraction using a mixture of toluene/ethanol (2:1) for 8 hours. The solid residue was then washed with ethanol to remove the toluene and Soxhlet extracted with 95% ethanol until the alcohol became colorless. The remaining solid residue was transferred into a 1000 mL Erlenmeyer flask and mixed with 500 mL of distilled water. The flask was heated for 1 hour in a boiling hot water bath. After cooling to room temperature, the mixture was filtrated and washed with distilled water. Extractive-free wood fibers were air dried overnight and further dried in a vacuum oven at 50°C.   60  4.3.2 Ball milling and enzymatic hydrolysis  1 gram of extractive-free wood fiber was subjected to rotary ball milling (Retsch PM200) at 600rpm for 6 hours at 10 min intervals with 5 min interval breaks. Rotation speed was set to 600 rpm. Ball milled wood powder was then suspended in citrate buffer at pH 4.8 at a 2% consistency. 50 mg of Ctec3 enzyme, which is a mixture of biomass degradation enzymes was added to the 1 gram of aliquot wood powder, and another 2% (w/v) Tween20 was added as a detergent and emulsifier. The mixture was incubated in a shaker at 50°C for 48 hrs, operating at 150 rpm. The enzymatically hydrolyzed residue, known as crude lignin was filtered out, washed with acidic water, then washed with distilled water, and freeze-dried.   4.3.3 Lignin extraction and mild acid hydrolysis 5 grams of crude lignin was added into a 100 mL mixture of dioxane and acid water at pH=2 (96:4 v/v). Lignin was extracted by reflux for 2 hours under a flow of nitrogen. The supernatant, containing lignin in dioxane, was separated by centrifugation and neutralized with sodium bicarbonate (NaHCO3). It was then added dropwise into acidic water (pH=2). The precipitated lignin was separated by centrifugation and washed twice with acidic water and distilled water. The resultant enzymatic milled acidolysis lignin (EMAL) was dried and stored in a vacuum oven at 50°C. Dried EMAL was acetone-fractionated into soluble and insoluble fractions following the same procedure as described for SKL.    61  4.4 Determination of molecular weight of lignins by gel permeation chromatography (GPC) and multi-angle light scattering (MALS) 4.4.1 Lignin acetylation  Prior to GPC analysis, 100 mg of lignin sample was first acetylated using 3 mL acetic anhydride and 3 mL anhydrous pyridine. The reaction was carried out for 48 hours at room temperature under a flow of nitrogen and constant stirring. Then, the acetylated lignin was precipitated with hydrochloric acid (pH = 2) and isolated through vacuum filtration using 0.45 m PTFE filter paper. The resulting acetylated lignin solid was dried in vacuum oven at 50°C.   4.4.2 Refractive index increment measurement  Specific refractive index increment (dn/dc) for different lignin and lignin fractions were determined using batch mode on WYATT 477-TREX Optilab T-rex.  The wavelength of incident light was 785nm, and the flow cell temperature was set to 35°C. THF was used as the eluting solvent. Kraft lignin solutions were prepared at various concentrations (0.1mg/mL, 0.5mg/mL, 1mg/mL, 1.5mg/mL and 2mg/mL), and filtered through 0.45μm PTFE filters. Then, the lignin samples were injected via a syringe pump at a rate of 0.2 mL/min until a plateau RI profile was achieved. dn/dc values were tested for freshly prepared samples and stabilized samples after 48 hours. Each lignin sample was run in duplicate.   4.4.3 GPC measurements with multi-detectors  The GPC system consisted of an Agilent 1260 ISO pump, guard column, three Styragel columns (HR 4, HR 3, and HR 1, Waters, Milford, MA), combined with a UV detector (1260 VWDVL UV, Agilent), a MALLS detector (DAWN HELEOS-II, WYATT 800-H2HC), a  62  differential viscometer (WYATT 323-V2), and a refractive index detector (Optilab T-rex, WYATT 477-TREX).  The three columns cover the entire molecular range of lignin (300-500K) ensuring complete separation. The wavelength of incident light was 785nm. Fluorescence filters of 10nm bandwidth were installed on the even numbered light scattering detectors to remove most of the fluorescence. Polystyrene standard (MW=30K) was used for the calibration of light scattering equipment and normalization of the 90° detector. In addition, the columns were calibrated with sets of polystyrene standards (MW 1300, 2500, 5780, 9000, 17500, 30000, 200000) received from Pressure Chemical Company.  The acetylated lignins were dissolved in THF at a concentration of 2mg/mL. Lignin solutions were then stabilized for 48 hours at room temperature and filtered with 0.45μm PTFE syringe filter. 100μL of lignin solution was injected and analyzed. The system temperature was maintained at 35°C and THF (HPLC grade, Fisher Scientific) was used as the eluent. All the data was collected and processed by ASTRA 6 software. Six measurements were collected for each lignin sample.  4.5 Chemical analysis  4.5.1 Quantitative 31P NMR analysis 20mg of lignin sample was dissolved in a 400 μL mixture of pyridine and CDCl3 (1.6/1, v/v). Immediately thereafter, 40μL of 5.6 mg/mL relaxation reagent (chromium (III) acetylacetonate), 100μL of 9mg/ml internal standard, N-hydroxy-5-norbornene-2,3-dicarboximide (e-HNDI), and 50μL of 2-chloro- 4,4,5,5-tetramethyl-1,2,3-dioxaphospholane (TMDP) was added as the internal standard. Fully dissolved lignin solution was transferred into a  63  5mm NMR tube. Quantitative 31P NMR was performed with an inverse gated decoupling pulse. Processing parameters were as follows: number of scans 800, relaxation delay 5s, acquisition time 1.4 s, pulse length 6 μs, and 90° pulse width. The system was calibrated by the residue water content in TMDP as phosphitylated derivative at 132.2 ppm. Functional groups assignments are shown in Table 4.1138,139,140.   Table 4.1 Quantification of various lignin moieties by 31P NMR Functional Groups Chemical Shift (ppm) Aliphatic hydroxyl group (AlOH) 150-145 Aromatic hydroxyl group (ArOH) 144-136 Carboxylic acid (COOH) 136-133 C5 substituted 144.5-140 Guaiacyl 140.2-139        64    Figure 4.4 Quantitative 31P NMR spectrum of [a] SKL and [b] hardwood EMAL  4.5.2 Quantitative 13C NMR analysis 150mg of lignin sample was dissolved in 450μL of DMSO-d6. 60 μL of 50mg/mL relaxation agent (chromium (III) acetylacetonate) in DMSO-d6 and 15mg of trioxane was added as the internal standard. The solvent was then transferred into a 5mm NMR tube. The NMR processing parameters were as follows: number of scans 20000, relaxation delay 2 s, acquisition time 1.4 s, pulse length 8.15 μs, and 90° pulse. DMSO (39.5 ppm) was used for signal calibration. The functional groups assignments are listed in Table 4.220,141. Note that due to significant overlap in the aromatic region, it was difficult to quantify linkages on benzene (a) (b) PPM  65  groups, such as C5-substitution and degree of condensation. The aromatic region (100-160 ppm) was set to a value of 600, and all the integrals were reported per 100 aromatic units.       Table 4.2 Quantification of various lignin moieties by 13C NMR Functional Groups Chemical Shift (ppm) and calculations Quinones 175-177 + 180-182 Aliphatic COOR 175.0-168.0 Conjugated COOR 168.0-166.0 H lignin 163.0-156.0 Aromatic H 125.0- 98.0 G lignin (hardwood) / (softwood) 114.0-110.0 / 114.0-98.0 S lignin 110.0-98.0 OMe 58.0-54.0 β-O-4 + β-β + β-5 (1) 90.0-82.5 Alkyl-O-alkyl (Oxygenated aliphatic) 90.0-58.0 –  aliphatic OH Alk-O-Ar, α-O-Alk  90.0-78.0 γ -O-Alk, secondary OH (2) 78.0-65.0  OH primary (3) 65.0-58.0 Degree of condensation (DC) 200 + G lignin – Aromatic H   66   Figure 4.5 Quantitative 13C NMR spectrum of SKL, 1: β-O-4 total +β-β + β-5; 2: γ -O-Alk, secondary-OH; 3: primary-OH  4.5.3 2D NMR (HSQC) analysis Heteronuclear single quantum coherence (HSQC) NMR was performed as follows: 40mg of lignin sample was dissolved in 600μL of DMSO-d6 solution and transferred into a 5mm NMR tube. The 13C-1H spectrum was acquired using Bruker “hsqcetgpsp.3_m4” pulse program with the following parameters: matrices of 4 data points for the 1H and 256 data points for 13C were collected with an interscan delay (D1) of 750 ms, 4 number of scans and a spectral width from 12.67 to -3.30 ppm for 1H and 210 to -30 ppm for 13C. Spectra were referenced to the DMSO-d6 signal (2.50/39.5 ppm). Topspin 3.5 software was adopted to process the spectrum including Fourier transform, baseline correction, and calibration. A semiquantitative analysis of the HSQC spectra were performed based on the previous paper142. Specifically, part of aromatic compounds was integrated and defined as the internal standard.   PPM  67  Table 4.3 Quantification of various lignin moieties by 2D-HSQC NMR Functional group  13C/1H The integrals expressed per number of correlations (∫ ) β-O-4 (A) 71.9/4.9 ∫ 𝐴𝛼 β-5 (B) 87.7/5.5 ∫ 𝐵𝛼 β-β (C) 85.5/4.6 ∫ 𝐶𝛼 Guaiacyl lignin (G) 110.2/6.9 ∫ 𝐺2 Syringyl lignin (S)  104.2/6.7 ∫ 𝑆2,6 /2 p-hydroxyphenyl lignin (H)  128.2/7.2 ∫ 𝐻2,6 /2 p-benzoate (Pb) 131.6/7.7 ∫ 𝑃𝐵2,6 /2 p-coumarate (PCA) 130.1/7.5 ∫ 𝑃𝐶𝐴2,6 /2 Dibenzodioxocin 81.5/4.75  Stilbene (St) 126.6/6.9 ∫ 𝑆𝑡𝛼, 𝛽 /2 5-hydroxymethylfurfural 122/7.5 ∫ 5ℎ 2 Quinones  122.3/5.7   4.6 Elemental analysis  Elemental analysis was conducted on a Carlo Erba Elemental Analyzer EA 1108. The contents of sulfur, carbon, oxygen, nitrogen, and hydrogen were measured for all lignin fractions.  The unit molar mass was calculated together with the 13C NMR analysis (Section 7.1 and 7.2). The basic assumption was that the SKL consists of guaiaycl (G) units only (C9 + methoxy group). The weight percent of methoxy groups (-OCH3) was first calculated from quantitative 13C NMR, and then the weight percent of carbon (C), hydrogen (H) and oxygen (O) within the methoxy group was calculated using the following equations143. [𝐶𝑂𝐶𝐻3] =[𝑂𝐶𝐻3] ∗ 12.0131.035 Equation 4.1  68  [𝑂𝑂𝐶𝐻3] =[𝑂𝐶𝐻3] ∗ 16.0031.035 Equation 4.2 [𝐻𝑂𝐶𝐻3] =[𝑂𝐶𝐻3] ∗ 1.008 ∗ 331.035 Equation 4.3 The weight ratio of carbon (C) in the C9 structure was calculated by subtracting the mass of C in the methoxy groups from the entire weight percent derived from elemental analysis.  [𝐶9] = [𝐶] − [𝐶𝑂𝐶𝐻3] Equation 4.4 Based on this, the index of methoxy, H, O, and S were calculated using the following equations:  [𝑛𝑂𝐶𝐻3] =[𝑂𝐶𝐻3] ∗ 108.0931.035 ∗ [𝐶9] Equation 4.5 [𝑛𝐻] =([𝐻] − [𝐻𝑂𝐶𝐻3]) ∗ 108.091.008 ∗ [𝐶9] Equation 4.6 [𝑛𝑂] =([𝑂] − [𝑂𝑂𝐶𝐻3]) ∗ 108.0919.0 ∗ [𝐶9] Equation 4.7 [𝑛𝑆] =[𝑆] ∗ 108.0932.065 ∗ [𝐶9] Equation 4.8    69   Result – lignin molecular weight characterization  5.1 Overall fractionation yield of solvent and membrane separated lignin The yield of each lignin fraction after vacuum drying is shown in Figure 5.1. After the initial acetone dissolution, the yield of the insoluble and soluble fraction were 60.4% and 39.6%, respectively. For the membrane fractionation, the >10kDa fractions accounted for a significant portion of the mass (60–70%), and the yield decreased with membrane cut-offs, where only approximately 15% was obtained for the low molecular weight sample (<3kDa). It should be noted that anhydrous acetone was used and all the lignin samples were vacuum dried for 48 hours prior to fractionation, because moisture content significantly increased the yield of acetone soluble fraction as observed in initial trials. The yield of acetone fractionation was similar to those obtained in other studies focusing on acetone fractionation, where AIKL of softwood kraft lignin ranged from 50 to 60%31,37,144,145. Similarly, in most previous investigations examining membrane fractionation, high MW materials also accounted for the major part of lignin. For example, Huang et al. (2017) reported a yield of 50.4% for >5kDa fraction146, and Shao et al. (2014) reported a yield of 54.3% for >10kDa fraction147. Compared with the binary fractionation, where acetone soluble lignin is precipitated by gradually adding another solvent, high MW fraction was also the main constituent in terms of percentage by weight31,144. .   70   Figure 5.1 Scheme for SKL fractionation 5.2 MW from GPC 5.2.1 Dn/dc values of fractionated lignin  The dn/dc values of kraft lignin before and after 48 hours of equilibrium are given in Table 5.1, where a significant decrease is apparent across all lignin fractions. Empirical evidence indicated that a dn/dc value can change over time when a polymer dissolves. For example, Suto et al. (1986) observed dissolution of ethyl cellulose accompanied by a slight decrease in the dn/dc value within the first thirty minutes148. This phenomenon was explained as an association effect, because a polymer takes time to dissolve and reach equilibrium in solution. Undissolved polymers might form aggregates characterized by a high dn/dc value, which is proportional to the MW. Therefore, the disassociation process will break-apart aggregates leading to a decrease in both MW and dn/dc values. However, Contreras et al. (2008) observed an opposite trend, where the dn/dc of milled wood lignin in THF and NaOH increased with incubation time. It is likely that the divergent findings stems from the differences in samples used, because MWL could undergo complex disassociation processes and alter its interactions, that in turn would  71  change the dn/dc value98. It is also noteworthy that the dn/dc values of lignin fractions after 48 hours equilibrium decreased with MW, which is to be expected as dn/dc is proportional to MW.  Although the MW of ASKL was lower than that of AIKL vide infra, the ASKL fractions had a higher dn/dc value. This result indicates a structural difference between AIKL and ASKL because dn/dc is proportional to the polarizability of a polymer. In studies in this field, most dn/dc values pertaining to SKL ranged from 0.14 to 0.16 in organic solvent such as THF, DMF and DMSO94,98,149,95, although the equilibrium time was not usually reported in the study. These values were consistent with the freshly prepared solvent (Table 5.1). However, in the present study, dn/dc values after 48 hours were used because all samples were incubated for 48 hours prior to analysis for consistency.  Table 5.1 The dn/dc value of lignin fractions after different incubation periods  dn/dc freshly prepared (mL/g) dn/dc after 48 hours (mL/g)  Raw SKL 0.1512 0.0908 AIKL 0.1554 0.0823 ASKL 0.1429 0.1003 AIKL>10kDa 0.1474 0.1301 ASKL>10kDa 0.1719 0.1495 AIKL3-10kDa 0.1445 0.1246 ASKL3-10kDa 0.1611 0.1229 AIKL<3kDa 0.1441 0.1210 ASKL<3kDa 0.1597 0.1279 Each sample was conducted in duplicate    72  5.2.2 Absolute molecular weight  In light scattering experiments, small molecules theoretically have no angular dependence, so the Zimm plot inevitably has a low fitting parameter with large variance. In order to maximize the accuracy and reliability of the findings, each measurement in the present study was estimated with six replicates, and the average value was employed. Moreover, lignin fluorescence was minimized by using an incident laser light of 785 nm wavelength and 10 nm fluorescence filter bands on the detectors. In addition, lignin absorbance and laser intensity fluctuation were auto-corrected by dividing light scattering signals by the forward monitor as recently highlighted by Zinovyev et al (2018).95   Figure 5.2 RI profile vs elution volume of lignin fractions [a] AIKL, and [b] ASKL  73   Figure 5.2 shows the refractive index (RI) elution profile of all lignin fractions, and light scattering profiles can be found in Appendix A. As can be seen, fractionation led to the separation of lignin samples, creating a right shift in the RI profiles, indicating a decrease in molecular weight, as smaller molecules take longer time to elute. The MW of AIKL and ASKL obtained via light scattering measurements was also plotted as a function of elution volume (Figure 5.3). For ASKL and AIKL the molar masses were similar for the different fractions, except for the AIKL >10kDa sample, that shifted the peaks to a lower elution value. The calculated MW and PDI are summarized in Table 5.2, indicating that lignin Mw ranged from 4 kDa to 30 kDa for ASKL fractions and 11 kDa to 115 kDa for AIKL fractions.  Compared with unfractionated SKL (PDI 2.10), the MW distribution significantly narrowed after fractionation (1.2−1.6). More importantly, the wide RI and light scattering profiles of the initial unfractionated lignin were replaced by narrow and more defined peaks, and the calculated MW distribution was significantly improved (Figure 5.3).  Table 5.2 Molecular weight averages for lignin fractions measured by MALS  Mn (kDa) Mw (kDa) PDI dn/dc Rh (nm)* RAW SKL 35.72 ± 11.2 74.47 ± 32.43 2.10 0.091 3.53 AIKL 59.97 ± 22.93 82.47 ± 45.25 1.37 0.082 3.60 ASKL 13.66 ± 8.02 18.83 ± 7.08 1.39 0.100 2.17 AIKL>10kDa 73.03 ± 14.67 114.93 ± 14.44 1.61 0.130 4.95 ASKL>10kDa 26.23 ± 9.91 30.75 ± 13.96 1.17 0.150 3.24 AIKL3-10kDa 8.95 ± 4.17  12.49 ± 6.50 1.39 0.125 1.90 ASKL3-10kDa 7.13 ± 4.15 9.12 ± 4.24 1.30 0.123 1.70 AIKL<3kDa 8.83 ± 3.93 11.46 ± 6.24 1.29 0.121 1.77 ASKL<3kDa 3.42 ± 2.06 4.27 ± 2.11 1.25 0.128 1.22  74  *: analyzed from viscosity properties. Each sample represents the average of 6 replicates.    75    Figure 5.3 Molecular weight distribution vs elution volume of lignin fractions measured by light scattering: [a] AIKL>10kDa & ASKL>10kDa; [b] AIKL 3-10kDa & ASKL 3-10kDa; [c] AIKL <3kDa & ASKL <3kDa  Based on these results, it appears that the AIKL had a higher MW than ASKL, because smaller polymers, like ASKL usually dissolve more readily. In terms of tangential flow filtration (TFF), Figure 5.4 [a] illustrated the MW distribution of lignin fractions and its relationship between membrane cutoffs. The TFF system did indeed fractionate the lignin from high to low MW, although measured lignin MW distribution did not perfectly match the TFF membrane cutoffs. ANOVA analysis resulted in p-values less than 0.05 among different fractions except for the AIKL 3-10kDa and AIKL <3kDa fraction, illustrating significant difference in the MW  76  distribution between most fractions. Furthermore, the similar radius between ASKL3–10 kDa (1.77nm) and AIKL<3 kDa (1.7nm) fractions illustrated the challenge of separating these complex polymers (Figure 5.3 [b]). The observation that lignin MW exceeded the corresponding membrane cutoffs MW may be because the TFF membrane is classified by MW cutoffs (MWCO) in ultrafiltration, although it filters based on hydrodynamic radius. Empirically, particles with a MW 3–6 times larger than MWCO can still squeeze through the membrane and is also dependent upon molecular shape150, 150. Moreover, after correcting the overestimation due to acetylation, the MW of lignin align better with the membrane cutoffs (Appendix D). Nevertheless, it has been shown that TFF is more effective at fractionating high MW than low MW fractions, which is important to remove these fractions for light scattering results.    Figure 5.4 [a] Number average molecular weight distribution and [b] hydrodynamic radius for lignin fractions. Red line was median value and blue box shows the upper and lower quartiles. The whiskers indicated the highest and lowest value, and outline values were excluded   Aggregation was observed in AIKL>10 kDa fraction, indicating an overestimation of the results. AIKL>10 kDa samples resulted in an abnormally high MW (300 kDa) with a very early (a) (b)  77  elution peak. Second, RI detected a bimodal profile, which may be related to the high MW aggregates54,151 or indicative of two distinct samples, as was reported previously with MWL samples15257,96. Finally, a front tail was found, where part of the lignin eluted even earlier than unfractionated AIKL. To assess the extent to which MW was affected by aggregates, the RI profile was deconvoluted into three distinct peaks (Figure 5.5). The first peak accounted for 34.7% of the total mass and MW ranges from 80 to 400 kDa. The second peak accounted for 33.6% and MW ranges from 40 kDa to 100 kDa. Based on this analysis, it must be acknowledged that AIKL >10kDa was likely significantly affected by aggregations.   Figure 5.5 Peak deconvolution of RI profile of AIKL>10kDa  In addition, a MW variability of around 50% was noticed. Such a high deviation is inevitable for lignin due to its small molecular size, for which no angular dependence exists in Zimm plots. Appendix A illustrated the Zimm plot of lignin solutions, where poor fitting was found for the raw lignin and was slightly improved for other samples. Therefore, each sample was repeated a number of times to decrease the variance and improve the reliability of the measurement.  78   5.2.2.1 Fluorescence of lignin  Lignin fluorescence increases the signal detected and leads to an overestimation of MW results. To estimate the effect of fluorescence, measurements from light scattering with and without filter bands were compared, and the ratio of MW is given in Table 5.3 Molecular weight with and without fluorescence filtersand Figure 5.7. The lignin samples that were analyzed without filters clearly showed lower MW than the samples analyzed with filters-- the overestimation was 6–85% dependent upon sample type. These results clearly show the impact of fluorescence was more significant on lower MW lignin fractions. The overestimation gradually rose from 21% to 85% and 15% to 52% for AIKL and ASKL fractions, respectively. This may be because light scattering intensity decreases with MW, while fluorescence is only dependant on fluorophore content. In other words, analytes scatter less light for lower MW fractions, while fluorescence remains the same because the fluorophore content did not change. Another reason may be related to the lignin structure, as supported by NMR analysis (Section 7.1). Low MW lignin contains condensed and potentially conjugated structures serving as fluorophore compounds, such as quinone and p-coumarate (PCA). It is also worth noticing the fluorescence effect between AIKL and ASKL fractions was inconsistent. The major reason can be found in the close analysis of the Zimm plot. Since fluorescence has no angular dependence, when it dominates the signal response, the fitting line of detector with and without fluorescence filters may differ significantly in both the slope and intersect at the same point on y axis, as shown in the example of ASKL 3-10kDa fractions (Figure 5.6). This is the challenge in analyzing the impact of fluorescence, especially when comparing AIKL and ASKL, as their structures differ significantly, leading to major differences in fitting slope.  79    Figure 5.6 Zimm plot of detectors with and without filter bands of [a] AIKL 3-10kDa and [b] ASKL 3-10kDa  Table 5.3 Molecular weight with and without fluorescence filters  Mw (kDa) (W|O filters) Mw (kDa)  (W Filters) MW (W|O filters) / MW (W Filters) AIKL 87.36 ± 40.01 82.47 ± 45.25 1.06 ASKL 23.35 ± 6.92 18.83 ± 7.08 1.23 AIKL >10kDa 138.78 ±17.46 114.93 ± 14.44 1.21 ASKL >10kDa 35.47 ± 12.00 30.75 ± 13.96 1.15 AIKL 3-10kDa 21.60 ± 16.04 12.49 ± 6.50 1.73 ASKL 3-10kDa 13.60 ± 8.96 9.12 ± 4.24 1.71 AIKL <3kDa 21.21 ± 7.18 11.46 ± 6.24 1.85 ASKL <3kDa 8.04 ± 4.26 4.27 ± 2.11 1.88   80   Figure 5.7 Molecular weight with and without fluorescence filters 5.2.2.2 Comparing light scattering data with conventional analysis Conventional analysis, based on calibration standards, is the most commonly used technique in lignin MW characterization. A comparison of light scattering and conventional analyses revealed that the light scattering results are in general 5–10 times larger (Table 5.3). There are many plausible reasons for the difference between the two techniques. The interaction between lignin and the column could delay the elution of samples, which would result in an underestimation compared to conventional analysis. The structural difference between a compact lignin and linear standards could also lead to an underestimation when using conventional analysis. The remaining fluorescence effect may result in an overestimation of the light scattering results. Finally, as light scattering is more sensitive to high MW particles, low MW lignin fractions that scatter insufficient light may be overshadowed when mixed with high MW lignin, resulting in an overestimation by light scattering. The ratio of the two techniques varies a great deal. For example, a MW measurement by light scattering for methanol fractionated lignin was  81  reported to be 100 times larger than the results obtained by conventional analysis36. In another light scattering experiment in DMSO, an overestimation of 2–5 times was reported95. Table 5.4 Molecular weight averages for lignin fractions measured by conventional analysis  Mn (kDa) Mw (kDa) PDI Ratio of LS / Convention Raw SKL 1.70 ± 0.14 6.01 ± 1.11 3.52 12.41 AIKL 2.01 ± 0.46  6.17 ± 0.20 3.10 13.36 ASKL 0.62 ± 0.07 1.58 ± 0.19 2.57 11.92 AIKL>10kDa 6.11 ± 1.51 16.42 ± 1.53 2.70 7.00 ASKL>10kDa 1.19 ± 0.15 3.41 ± 0.38 2.85 9.02 AIKL3-10kDa 1.45 ± 0.80 1.61 ± 0.76 1.11 7.76 ASKL3-10kDa 0.91 ± 0.05 1.34 ± 0.06 1.47 6.81 AIKL<3kDa 0.82 ± 0.11 1.33 ± 0.16 1.56 8.61 ASKL<3kDa 0.40 ± 0.08 0.76 ± 0.13 1.63 5.62   Figure 5.8 Molecular weight of lignin as estimated by light scattering and conventional analysis   82  5.2.2.3 Lignin degree of polymerization It is not straight forward to establish the degree of polymerization for lignin. One significant reason is that the unit formula is unavailable, as the C9 structure is greatly modified during the pulping process. Therefore, various assumptions and approximations have been developed for the estimation of unit molar mass. Based on these methods, SKL consists of G lignin (C9 + methoxy group), the unit formula and unit molar mass were calculated to be 175-190 (Table 5.5). The number average degree of polymerization (DP) was further calculated based from the Mn value. The unit molar mass was 170–190 dependent upon the fraction, which is similar to other studies on lignin unit molar mass144,45. The 3–10 kDa and <3 kDa lignin fractions were close to an oligomer or low molecular weight polymer, with a DP of 20–50, while AIKL>10 kDa is quite a large molecule, with a DP over 300. The calculated DP was much higher than most published data whereby a DP of 10–20 was reported45,153; this result was mainly because of the large MW from light scattering compared to conventional analysis as mentioned above, and wide MW distribution used for its calculation. Compared with literature results, such a wide distribution (20-300) reflects the fact that lignin properties change with fractions. For example, lignin thermal stability and glass transition temperature (Tg) significantly increase with higher MW fractions55,144 because the thermal properties of a polymer change with DP and level off after a critical point.  However, AIKL<3 kDa had a similar DP to that of AIKL 3–10 kDa. This result may be due to the possible difficulty of ultrafiltration when particle sizes are similar and the difficulty of light scattering to reveal highly accurate data for small diameter materials.  83  Such a finding highlights the inability of light scattering to accurately distinguish two similar fractions as a large variance is inevitable for a small diameter lignin sample.  Table 5.5 Degree of lignin polymerization  Unit Formula Unit Molar Mass DP AIKL C9H8.36O2.60(OCH3)0.87S0.09 188.91 317 ASKL C9H8.11O1.85(OCH3)0.92S 0.06 177.34 79 AIKL >10kDa C9H8.33O2.88(OCH3)0.77S 0.06 189.53 385 ASKL >10kDa C9H7.73O2.91(OCH3)0.82S 0.05 190.62 138 AIKL 3-10kDa C9H8.45O2.68(OCH3)0.75S 0.07 185.95 48 ASKL 3-10kDa C9H7.96O2.34(OCH3)0.82S 0.05 181.90 39 AIKL <3kDa C9H8.19O2.47(OCH3)0.74S 0.07 182.05 48 ASKL <3kDa C9H9.08O2.18(OCH3)0.67S 0.05 175.22 19 Each sample was analyzed in duplicates.   84   Result of lignin molecular conformation characterization  6.1 GPC behaviour of lignin  In the GPC chromatograms there is a clear upswing in the curve when plotting lignin MW as a function of elution time (Figure 6.1), which was previously reported.59,95,94 As discussed earlier, there are two possible reasons for this observation. First, the possible interactions between branched polymers and the column gel packing matrix, whereby molecules with a higher degree of branching would elute at later times resulting in an overestimation of the MW. Second, fluorescence typically has a greater impact on lower MW lignin, which would elute at the end of the run, and it would again result in an overestimation of MW. It is very difficult to quantify how much of these two factors individually affect the results. Nevertheless, the result may qualitatively suggest that lignin is a branched polymer, which opens up the possibility of delving further into the lignin structure by light scattering and other methods such as viscosity measurements.   6.2 Mark–Houwink–Sakurada plots The conformation and structure of lignin were further investigated by hydrodynamic behaviour, which was characterized by viscous measurements and further related to the molecular conformation by the Mark–Houwink–Sakurada (MHS) equation. In the experimental setup, intrinsic viscosity was equated with specific viscosity, as they were approximately the same in a diluted solution90. As shown in Figure 7.1 [a] and [b], the intrinsic viscosity of different fractions of ASKL and AIKL were combined and plotted against MW (known as a MHS plot). This plot is a bridge between viscosity, MW, and molecular conformation because a polymer in solution should follow the rule of [𝜂] = 𝐾 ∗ 𝑀𝛼, where [𝜂] is intrinsic viscosity, M  85  is molecular weight, K is a constant, and the α value is conformation coefficient relating to polymer conformation in a dilute solution. The smaller the α value, the more compact the polymer is in solution154.  Table 6.1 α value and related polymer conformation in dilute solution 𝜶 value Polymer conformation 0-0.5 rigid spheres 0.5-0.8 linear random coils 0.8-2 stiff chain, rigid rod  The Mark–Houwink coefficient α was 0.20 for ASKL and 0.34 for AIKL. Both 𝛼 values suggested that lignin is a hard sphere, whereby the α value of ASKL was significantly lower than ASKL, indicating that ASKL has a more compact structure154. However, when comparing the MHS plot of unfractionated AIKL and ASKL (Figure 6.1 [c] and [d]), there was minimal difference between the α values (AIKL=0.2509, ASKL=0.2308). Most literature reported a similar hydrodynamic behaviour ([𝜂] = 𝐾 ∗ 𝑀0.1−0.3)122,123,52 for various types of unfractionated lignin in different solutions. For instance, in a publication where softwood MWL was fractionated by THF, α value of MWL-insoluble and MWL-soluble was shown to be 0.28 and 0.27 respectively96.   86      Figure 6.1 MHS plot of [a] AIKL fractions (AIKL>10kDa, AIKL3-10kDa, AIKL<3kDa), [b] ASKL fractions (ASKL>10kDa, ASKL3-10kDa, ASKL<3kDa), [c] AIKL prior to TFF fractionation and [d] ASKL prior to TFF fractionation  87   As conjecture, an apparent similar α value for AIKL and ASKL before TFF fractionation indicates that the low MW lignin fraction was not well detected by both light scattering and differential viscometer due to insufficient sensitivity. Apart from the large variance in light scattering results, the intrinsic viscosity of lignin is very low. For instance, the viscosity of a lignin molecule is only approximately a quarter of that of linear polystyrene standards, at the MW of 104 Da, which is very hard to be accurately measured155 (Appendix E). Therefore, the background noise of unpurified lignin was very high and consequently resulted in a R2 value of 0.5. The trend most likely was dominated by the larger molecular weight components that impact viscosity. Furthermore, the low viscosity at equivalent MW of polystyrene standards suggested a random coil structure was not an appropriate model for lignin. In addition to the use of the MHS plot, an effective hydrodynamic radius of lignin fractions was calculated from intrinsic viscosity based on the assumption that dissolved lignin is an Einstein sphere (hard sphere), as given in Error! Reference source not found.. Lignin had a radius of 1.7–3.5 nm, and ASKL was generally smaller in size than AIKL fractions. On one hand, this was because ASKL fractions had lower MW and DP than their corresponding AIKL fractions. On the other hand, ASKL may have a more compact structure than AIKL. As shown in Figure 6.1, in the 4 kDa to 10 kDa region, where the MW of ASKl and AIKL overlapped, AIKL generally had higher intrinsic viscosity at the same MW, and thus larger size as the hydrodynamic radius is proportional to viscosity90. The molecular density was also compared in the Appendix A.3.    88   Result of lignin chemical analysis  7.1 Kraft lignin NMR    Based on the light scattering results, different lignin fractions exhibited different hydrodynamic behaviours and molecular conformation in solvent, which suggested significant structural differences among the lignin fractions. In this section, the chemical structure of lignin was analyzed based on a series of lignin NMR spectra and the results were compared with light scattering and hydrodynamic analyses. The 13C NMR of acetylated samples, 31P of phosphitylated derivatives, and 13C-1H 2-dimensional heteronucleaer single quantum coherence NMR of unmodified lignin were acquired to provide careful characterization of lignin functional and structural elements.  The phenomenon that lignin structure varies with fractions has been reported many times in the literature45,31. Despite the variance in fractionation methods, some properties are always intertwined. For example, a lignin fraction of comparatively lower MW will also have better solubility in organic solvents, more phenolic OH, less aliphatic OH, and a lower content of ether bonds. Some researchers have argued that this is because native lignin is comprised of different fractions with varying chemistry and can be separated according to the solubility in various solvents. Others propose that MW significantly impacts the lignin structure because the different fractions undergo different degrees of degradation during the pulping operations. Additional information can provide insight to the differences by analyzing MW and chemical structure separately, which are two distinct properties of a polymer. In this chapter, by combining acetone fractionation and TFF fractionation, AIKL and ASKL fractions with a similar MW were compared so that the effect of MW was minimized when analyzing the impacts of acetone fractionation. Likewise, the effect of MW was analyzed within  89  AIKL and ASKL fractions so that the impact of solubility was ruled out. By combining light scattering and NMR analyses, a detailed elucidation of lignin chemical structure, MW, and the relationship between lignin fractions was generated for these kraft lignin samples.  7.1.1 Aromatic unit as a reference and internal standard The HSQC NMR is a type of 2D NMR that is used for the semi-quantitative determination of interunit linkages in lignin. The system detects C–H bonds under different environments and provides a response signal with the intensity proportional to the number of bonds. All the signals were normalized to the number of C2–H bonds on the aromatic ring, as this hydrogen is least affected during pulping, and provides a representative number of G lignin units. Guaiacyl lignin content for the softwood kraft lignin was largely the only source of aromatic rings in lignin and thus rendered as a constant value. As such, C2–H was taken as an internal standard, and all linkages were reported on the basis of G lignin or aromatic ring equivalents156. In a similar semi-quantiative fashion, for 13C NMR, 100–160 ppm were assigned to unsaturated structures and mostly aromatic rings in native lignin; however, 1,3,5-trioxane was added as an internal standard for quantitative evaluation134. As kraft lignin was significantly restructured in the pulping process, unsaturated components, such as stilbenes, aryl enol ethers and cinnamyl groups, may exist and were assigned to the 100–160 ppm region in the 13C NMR profile. To obtain a pure aromatic region, the number of unsaturated components was determined from HSQC and subtracted from the 100–160 ppm region134. When analyzing 31P NMR data, N-hydroxy-5-norbornene-2,3-dicarboximide (e-HNDI) was used as an internal standard, and the data were reported as mmol/g.   90  7.1.2 Interunit linkages Detected interunit linkages in kraft lignin consisted of β-aryl ether (β-O-4), phenyl coumaran (β-5), pinoresinol (β-β), diphenyl ethane (β-1), stilbene, and linkages such as 4-O-5 and α-O-4 (Table 7.1). Most ether bonds are native linkages, but they are severely degraded during the pulping process. Structures such as 5-5 and stilbene may form from repolymerization or degradation reactions during kraft pulping. Table 7.1 Information on interunit bonds abundance of SKL fractions as obtained (units per 100 Ar)  Raw AIKL ASKL AIKL >10kDa ASKL >10kDa AIKL 3-10kDa ASKL 3-10kDa AIKL <3kDa ASKL <3kDa β-O-4′a 7.74 14.01 4.23 17.48 4.37 6.9 3.89 6.62 1.67 β-5′a 2.25 4.38 1.52 6.94 2.91 2.38 1.73 1.86 1.16 β–β′a 3.10 4.72 2.94 7.41 4.94 5.38 4.35 4.28 4.16 Stilbene 2.21 0.85 1.02 0.1 0.83 2.53 2.68 1.27 6.31 β-O-4′ + β-β + β-5b 11.63 17.87 5.78 14.07 6.71 11.31 4.71 8.1 4.24 Alk-O-Ar, α-O-Alk b 14.45 25.57 6.65 17.17 9.11 13.10 5.11 10.44 5.81 a: Obtained from HSQC; b: Obtained from 13C NMR;   HSQC is a powerful 2D NMR method that shows distinct peaks in the aliphatic and aromatic regions of lignin, and provides more information on interunit linkages than 1D NMR. However, a major disadvantage of the technique is the limit of detecting condensed aromatic regions where protons are no longer directly attached to the ring, such that structures on the phenyl ring cannot be directly quantified. From the aliphatic region of the HSQC NMR spectrum (Figure 7.2) and the quantified results (Table 7.1), the content of all native linkages (β-O-4′, β-β, β-5) in ASKL was significantly lower than that of AIKL across all MW fractions. AIKL  91  fractions contained 50 to 160% more total native linkages (β-O-4′ + β-β + β-5) than its corresponding ASKL fractions. In addition, other linkages that formed during the pulping process, such as stilbenes, accounted for a higher content in the lower MW fraction and ASKL fractions (Figure 7.2). Taken together, these results suggested structural differences among the two solvent fractionated lignins. Furthermore, in the 13C NMR spectra (Figure 7.1), 58–98 ppm was assigned to oxygenated aliphatic groups, including both oxygenated aliphatic moieties involved in inter-unit linkages and hydroxyl groups. However, the reliability of the quantitative result was significantly reduced by the observed overlap among functional groups. Thus, the quantification process involved certain assumptions and calculations considering the low abundance of interunit bonds in technical lignin which leads to a high noise to signal ratio in this region. Moreover, carbohydrate contaminants were also identified in raw SKL, AIKL, and AIKL>10kDa fractions, and complicated the quantification. Despite the relatively low reliability, analysis of interunit linkages yielded a similar result to HSQC. Across all MW fractions, AIKL fractions contained 1.5–1.6 times more native linkages, approximately twice as many linkages at α and β positions and 60–480% more ether bonds than ASKL. Comparing with EMAL values (Table 7.7) and those reported in the literature for softwood EMAL157,20, more than 80% the native ether bonds were degraded after kraft pulping. Hence, the kraft pulping process, significantly modifies the lignin. However, this study provides additional insight to structural differences between AIKL and ASKL fractions suggesting either the process or the location/type of lignin is more sensitive to degradation schemes. Assuming that native ether bonds functioned as backbone linkages in lignin, ASKL should be a more disrupted and degraded structure than AIKL due to the presence of fewer interunit linkages. Further, lignin may undergo a repolymerization process during delignification; if ASKL is more degraded, then  92  it may contain increased levels of stilbenes and other condensed compounds, such as quinones (Q), p-coumarate (PCA), and 5-hydroxymethylfurfural (F) derived from degraded carbohydrates. These compounds are typically formed in repolymerization as β-O-4 and β-5 units broken and converted into aryl enol ethers and stilbene-like structures due to oxidization158. Another possible explanation for the structural difference between AIKL and ASKL was in regard to the composition of native lignin, as represented by enzymatic milled wood acidolysis lignin (EMAL) (Table 7.7). There was an approximately 15% acetone soluble fraction in EMAL, which contained fewer native linkages. Thus, ASKL might be the product of degradation of acetone soluble fraction in native lignin. However, the limitation was that the EMAL characterized was extracted from eucalyptus, a hardwood lignin, which was structurally different from softwood. However, this conjecture, provides an important line of scientific inquiry of where this difference may arise (i.e. lignin structure in the middle lamellae and secondary wall may differ).   Figure 7.1 13C NMR spectrum of AIKL>10 kDa fraction; 1: Alk-O-Ar, α -O-Alk, 2: γ -O-Alk, OHsecondary, 3: OHprimary PPM  93    Figure 7.2 2D HSQC Spectrum of AIKL fraction – [a] aliphatic region; [b] aromatic region   (b) (a)  94    Phenyl coumaran (β-5′) (A) Arylglycerol-β-aryl ethers (β-O-4′) (C)  Pinoresinols (β–β′) (B)    Stilbene (St) Dibenzodioxocin p-coumarate (PCA)    Carbohydrate (X) 5-hydroxymethylfurfural (F) Quinones (Q) Figure 7.3 Main structures and degradation products in kraft lignin identified by 2D NMR  In terms of structural dependence on MW, Error! Reference source not found. shows the relationship between total native linkages per 100 aromatics against calculated degree of  95  polymerization (DP) for various lignin fractions. The content of native linkages increased as DP increased and gradually leveled off at high MW. As found with HSQC analysis, compared with >10 kDa fractions, the total content of native linkages in <3 kDa fractions decreased by 60% in AIKl and by 40% in ASKL. This change can be explained by assuming kraft pulping to be a random chain cleavage process. Chain cleavage of the larger polymers at the beginning of the process resulted in a greater decrease in DP, while the cleavage of small polymers towards the end of the process had less effect on DP. Therefore, the remaining bonding is inversely proportional to overall DP (𝑙𝑖𝑛𝑘𝑎𝑔𝑒𝑠 = 𝑎 + 𝑏/𝐷𝑃).   7.1.3 Functional groups  Hydroxyl groups and carboxylic acids were determined using 31P NMR analysis (Figure 7.5). Carboxylic acid groups were present in softwood kraft lignin, and there was a significant increase relative to the EMAL. The carboxylic acid content was significant and accounted for 0.5-0.9 mmol/g; the content was higher in ASKL fractions and overall was increased for the lower MW fractions. Further analysis of the solvent fractionated lignin, compared with hardwood EMAL and literature values159,88, indicated that aliphatic OH content was significantly lowered and phenolic OH groups increased considerably (Table 7.2). This change may be related to the significant degradation of aliphatic side chains and cleavage on aryl-ether bonds as discussed above. Aliphatic OH content was higher in AIKL than ASKL across all MW fractions, while phenolic OH content was higher in the ASKL fractions. The difference between AIKL and ASKL became smaller as MW was reduced for the samples. Furthermore, the total amount of OH functional groups for ASKL and AIKL was very close (1-12% difference). Lower aliphatic  96  OH and higher phenolic OH in the ASKL samples suggested a significant cleavage of ether bonds and formation of phenolic OH along with  carbon loss37.  In terms of the relationship of the hydroxyl content and MW, it appeared that as MW decreased, aliphatic OH decreased, while phenolic OH were generally stabilized, and the total OH content slightly decreased. This phenomena was not commonly reported in solvent fractionated samples, but was shown in a study on lignin ultrafiltration where both phenolic OH and aliphatic OH decreased with MW44. On the one hand, it may be simply because AIKL>10kDa fraction contained carbohydrate impurities as indicated in the 2D-NMR spectra. As well, this finding might indicate the degradation of aliphatic chains was more closely related to the reduction of MW. As shown in Figure 7.4, delignification involves elimination of aliphatic side chains from quinone methide intermediates due to oxidization, and aliphatic OH will be lost in the form of formaldehyde.         97       Figure 7.4 Elimination aliphatic sides from kraft pulping  In addition, the ratio between aliphatic OH and aromatic OH was used to represent the degree of degradation in the pulping process. AIKL was almost double the value of ASKL for all MW fractions indicating samples with smaller values were modified to a greater extent. This ratio was consistent at 0.35 for ASKL, and decreased from 0.96 to 0.55 from AIKL >10kDa to AIKL 3-10 & <3kDa fractions157. (c) (a) (b)  98  Based on the analysis of the hydroxyl groups it can be concluded that ASKL was a more degraded structure than AIKL. Further, the chemical structure of ASKL seemed to be more similar across different MW fractions, while the aliphatic chains in the AIKL fractions might be more disrupted at lower MW. Generally, these data were in line with most conclusions in the literature that the amount of aliphatic OH was higher in insoluble fractions and higher MW fractions, while phenolic OH content contradicted to what was previously reported160, 35. From this analysis it appeared the minimal differences in the structure of the ASKL fraction revealed that this sample was more uniform across different MW, whereas the structure of AIKL was increasingly disrupted at the lower MW range.  Table 7.2 Information on functional groups abundance of SKL fractions  Raw AIKL ASKL AIKL  >10kDa ASKL  >10kDa AIKL  3-10kDa ASKL  3-10kDa AIKL  <3kDa ASKL  <3kDa Hydroxyl groups (mmol/g) Aliphatic OHc 2.32 3.15 1.72 3.61 1.85 2.16 1.48 1.95 1.53 Phenolic OHc 3.53 3.76 4.89 3.77 4.83 3.92 4.55 3.49 4.28 Ali / Phenolic OHc 0.66 0.84 0.35 0.96 0.38 0.55 0.33 0.56 0.36 Total OHc 6.11 7.16 6.93 7.70 7.08 6.38 6.35 5.68 6.16 COOHc 0.45 0.46 0.46 0.59 0.87 0.99 1.27 0.83 0.96 c: Obtained from 31P NMR. Each sample analysis was conducted in duplicate   99   Figure 7.5 31P NMR spectrum of AIKL>10 kDa fraction  7.1.4 Degree of branching  In the literature there have been minimal attempts to determine the degree of branching in lignin, and seldom are there any quantitative data. The major reason concerns the difficulties in defining a clear backbone and side chains of lignin. Further, the branching structure may form during the pulping process153. During delignification, free radicals form and transfer on the phenyl ring, and create potential new side chains at the C5, C6 position. Crestini et al. (2018) proposed that sulfur can be transformed into a biradical during pulping process and result in oxidative coupling reactions on phenols. Phenoxy radicals might form at positions ortho and para on the ring during delignification, resulting in biphenyl and biphenyl ether structures. These reactions would lead to branching that is not found in typical native structures. Based on the assumption that native ether linkages are backbone structures and any substitutions at the C5 forms side chains, branching degree quantification equals to the measurements of the number of C5 substitutions, although this neglects C1 substitutions. The most direct method for C5 quantification of lignin’s condensed structure is based on 31P NMR analysis of C5 substituted PPM  100  free phenolics, and semi-quantification of 5 substituted linkages using HSQC. As shown in Table 8.2, C5 substitution was 5-20% higher in ASKL than AKIL across all MW fractions. Meanwhile, the degree of substitution decreased with lower MW samples. With HSQC NMR dibenzodioxocin, β-5 and 4-O-5 linkage were directly measured to reveal potential branch points. β-5 is a native linkage, while 4-O-5 was not detected during the analysis, perhaps due to the presence of a negligible amount. The dibenzodioxocin content between AIKL and ASKL was inconsistent. Yue et al. (2016) claimed the detection of 4-O-5 bonds142 and determined an amount of a negligible amount of 1% 4-O-5 units per 100 aromatics in SKL.  The second quantification method for branching depends on the assumption that C5 substitution relates to the amount of Ar-H142 found on the aromatic units of lignin. As each aromatic ring in softwood kraft lignin contains a carbon chain on C1, a methoxyl group on C3, and a hydroxyl group or an ether bond on C4, there are ideally three hydrogens on each aromatic ring. The presence of C5 substitution, like β-5 and 5-5, reduces the amount of C5-H and overall Ar-H. As shown in the 13C NMR, the Ar-H amount ranged from 225-240 per 100 aromatics for all fractions, indicating the existence of a side chain or branch point for 60-75% of the aromatics. The Ar-H content was similar between ASKL and AIKL with a difference of less than 1% for each MW fraction. This data suggests that both fractions have equivalent amount of branching. There was an approximate 6% increase in Ar-H amount from >10kDa to <3kDa fraction. Similarly, Balakshin et al. (2016) proposed another method, where “branching” was described by a term, “degree of condensation” (DC)161,141. This method corrects for the inaccuracy of G content, which is generally <100 in technical lignin. They developed the equation that DC = 200 + G – Ar-H. Based on this method, the degree of condensation was 35-55% with no significant difference between ASKL and AIKL, while it slightly decreased from 50 to 40 from >10kDa to  101  <3kDa fractions. In summary of the different C5 measurements (Table 7.3), 31P detected higher C5 substitution in ASKL, while 13C yielded a similar result between AIKL and ASKL, and HSQC detected more β-5 in AIKL. In light of the relationship between branching degree and MW, there was clearly a decrease in branching with a decrease in MW, and a difference for lignin fragments with free phenolics.    Figure 7.6 Branching structure on [a] C1 and [b] α position   Table 7.3 Information on branching of SKL fractions  Raw AIKL ASKL AIKL >10kDa ASKL >10kDa AIKL 3-10kDa ASKL 3-10kDa AIKL <3kDa ASKL <3kDa DCb 42.59 48.64 36.87 53.92 58.37 37.50 40.34 40.74 37.68 C5 substituted phenolc 1.66 1.79 1.87 1.53 1.84 1.29 1.36 1.24 1.37 Aromatic C–Hb 242.88 234.77 251.88 230.66 225.53 242.14 241.25 241.07 243.93 dibenzodioxocina 0.62 1.01 1.17 0.3 0.64 2.43 1.45 1.73 1.87 Methoxyb 74.55 81.46 74.26 78.98 74.20 75.97 76.26 75.15 82.70 a: Obtained from HSQC (unit: per 100 aromatics); b: Obtained from 13C NMR (unit: per 100 aromatics);  c: Obtained from 31P NMR (unit: mmol/g);   A similar inconsistency in substitution was also reported previously 36,37, 45,44,134, as summarized in Error! Reference source not found.. This phenomenon can be explained by the f(a) (b)  102  ollowing two reasons. On the one hand, all techniques have their limitations. 31P NMR provides incomplete information, as it only detects C5 substitution on free phenolic units, while quantification of branching by 13C NMR involves many assumptions, which potentially magnifies errors. On the other hand, the basic assumption that C5 substitution represents total branching is incomplete. For example, other positions might also form side chains such as C1, or at benzyl carbons, and at locations where methoxy groups were lost during pulping (Figure 7.6).  Table 7.4 Reviews on lignin C5 substitution from NMR results Fractions C5-subsititution ArH β-5 Ref MeOH ISo 1.84 2.2 - 36 MeOH So 1.83 2.1 - AIOL 4.1  7 37 ASOL-250 (hexane) 6.6  5.7 ASOL-750 (hexane) 7.2  7.0 ASOL-800 (hexane) 11.5  8.9 LFEtOAc 28  1 45 LFEtOH 40  2 LFMeOH 45  3 LFAcetone 50  4 LR50kDa   185  44 LR15kDa  192  LR5kDa  195  AIKL 1.8 249 2.6 134 AIKL300 (hexane) 2.5 223 1.8 AIKL500 (hexane) 2.4 214 1.8 AIKL900 (hexane) 2.5 205 0.9 AIOL: acetone insoluble organosolv lignin;  ASOL: acetone soluble organosolv lignin   103   7.1.5 Structure analyses of SKL fractions  Based on the NMR analysis, we tried to elucidate the differences in structure among lignin fractions. First, there was a clear difference between the AIKL and ASKL fractions. AIKL had a higher content of native linkages, indicating a lower degree of degradation and potentially a more extended structure. AIKL also contained more aliphatic hydroxyl groups and less phenolic hydroxyl groups, suggesting less degradation of aliphatic side chains and reduced breakage of ether bonds, relative to ASKL. The higher content of stilbene and other conjugated compounds was a sign of significant modification in ASKL fractions. This result was consistent with what was observed with light scattering and viscosity measurements. However, due to the inconsistent measurement of C5 substitutions, there was no solid evidence indicating a difference in the number of branching points between the AIKL and ASKL fractions. Regarding the relationship between MW and chemical structure, higher MW fractions showed a higher content of native linkages, lower aromatic substitution, and fewer aliphatic hydroxyl groups. Nevertheless, a significant difference in chemical structure between high and low MW fractions was not found. This result indicated that ASKL fractions were more uniform than AIKL fractions, as the ratio of aliphatic to phenolic were similar with different MW fractions within the ASKL fraction; the ratio varied significantly amongst the AIKL fractions. The data supports that the aliphatic chains in AIKL might be degraded more in the lower MW fractions. Based on the bonding patterns and bond frequency derived from NMR data, a visual representation of one of many possible structures was given Figure 7.7 for the lignin fractions.     104      Table 7.5 Information and structural characterization of kraft lignin fractions  Raw AIKL ASKL AIKL  >10kDa ASKL  >10kDa AIKL  3-10kDa ASKL  3-10kDa AIKL  <3kDa ASKL  <3kDa Quinonesa - - 0.62 - 0.37 0.92 0.91 0.65 4.51 p-coumarate (PCA) a - - 0.07 - - - - - 0.54 5-hydroxymethyl- furfurala 1.05 0.53 0.55 0.51 0.99 1.27 1.67 1.49 2.15 Gb 85.47 83.41 88.75 84.58 83.90 79.64 81.59 81.81 81.61 Hb 5.57 2.35 1.43 6.48 2.17 3.09 3.75 4.95 - Aliphatic COORb 4.17 1.96 5.82 6.17 10.92 12.53 13.38 4.70 4.82 conjugated COORb - - 0.88 - 1.69 3.34 2.56 0.55 0.14 a: Obtained from HSQC (unit: per 100 aromatics); b: Obtained from 13C NMR (unit: per 100 aromatics); c: Obtained from 31P NMR (unit: mmol/g);         105    AIKL>10 kDa ASKL>10 kDa   AIKL<3 kDa ASKL<3 kDa Figure 7.7 Hypothesized lignin structures of different lignin fractions based on the frequency of interunit linkages.      106  7.2 Elementary analysis  The elemental content of the lignin fractions were determined by elemental analysis (Table 7.6). Together with methoxyl content, the unit formula of lignin was derived for each fraction (Table 5.5).  Table 7.6 Elemental composition of lignin fractions Lignin C H S O C:H:O (mol:mol) Raw 62.74 5.82 1.11 30.34 2.8 : 3.1 : 1 AIKL 60.22 5.58 1.47 32.13 2.5 : 2.8 : 1 ASKL 65.60 6.00 0.98 27.43 3.2 : 3.5 : 1 AIKL >10kDa 59.56 5.41 0.99 34.05 2.3 : 2.5 : 1 ASKL >10kDa 59.52 5.15 0.82 34.52 2.3 : 2.4 : 1 AIKL 3-10kDa 60.71 5.55 1.12 32.63 2.5 : 2.7 : 1 ASKL 3-10kDa 62.84 5.56 0.92 30.69 2.7 : 2.9 : 1 AIKL <3kDa 62.11 5.54 1.15 31.21 2.7 : 2.8 : 1 ASKL <3kDa 64.28 6.14 0.84 28.75 3.0 : 3.4 : 1 Each sample analysis was conducted in duplicate.  There were differences regarding the elemental composition between the AIKL and ASKL fractions. The sulfur content was slightly higher in AIKL, and increased slightly as MW decreased. Most lignin ultrafiltration experiments reported a significantly higher sulfur content in lower MW fractions43,55,28, while a similar value was found in solvent fractionation144,121. As the lignin material was acid washed and solvent fractionated, most inorganic sulfur compounds were removed, and the remaining sulfur was covalently bonded to lignin molecules. Therefore, the relationship between sulfur content and lignin fractions may relate to the pulping mechanism. This mechanism in the pulping process has the nucleophile bisulfide (HS-) attacking the intermediate quinone methide, and leads to the cleavage of the ether bond at the β-position. The  107  introduced bisulfide would prevent further condensation and favour the solubility of lignin molecules in base162,163. Therefore, a high sulfur content may be an indicator of more severe degradation, however this was not observed as AIKL had more native linkages than ASKL. The oxygen content was slightly lower for ASKL, while the carbon content was higher than the AIKL. This data supports the proposed loss of aliphatic hydroxyl groups in ASKL.    7.3 Enzymatic milled wood lignin NMR Enzymatic milled wood lignin was prepared as a “linear” reference, although it is not reported to be linear except in one study153. Hardwood (Eucalyptus) was selected because of the high S content and potential high linearity. EMAL lignin was fractionated into acetone soluble (AS-EMAL) and insoluble fractions (AI-EMAL) to ensure the solubility, with the latter containing more lignin-carbohydrate content. As shown in Table 8.4, both fractions contained around 70% S lignin and 30% G content. In 31P and 13C NMR analysis, the C5 substituted phenol and degree of condensation (DC) were both close to zero indicating the guaiacyl rings were not substituted, suggesting the structure had a rather low degree of branching. However, the sum of native linkages (β-O-4 + β-5 + β-β) were only 69 and 56 per 100 aromatics for AI-EMAL and AS-EMAL. This suggests a low degree of polymerization153, or the fact that some other linkages were not characterized by the method. Despite this unknown result, such a low degree of condensation indicated that it was a potential linear reference, or at least less branched, and useful to compare the degree of lignin structure alternation after kraft pulping.     108  Table 7.7 Information and structural characterization of EMAL fractions as obtained by HSQC, 13C NMR, and 31P NMR  AI-EMAL  Eucalyptus AS-EMAL   Eucalyptus Interunit Bonds Abundance (per 100 Ar)  β-O-4′a 51 36 β-5′a 4 5 β–β′a 14 15 β-O-4′ + β-β + β-5b 55.54 46.13 Alk-O-Ar, α-O-Alkb 285.54 190.45 DCb 1.00 0.00 Functional Groups Abundance (per 100 Ar) G ligninb 29 30 S ligninb 63 62 Methoxy (OMe)b 153.34 142.55 Aromatic C–Hb 212.70 214.19 Functional Groups Amount (mmol/g) Aliphatic OHc 4.496 4.513 C5 substituted phenolc 0.01 0.09 G OHc 0.836 1.194 S OHc 0.535 0.707 Total OHc 1.371 1.901 COOHc 0.101 0.136 a obtained from HSQC, b obtained from13C, c obtained from 31P.  109   Quantification of branching As reviewed earlier, the extent of branching of a polymer can be quantified by comparing the viscosity of a sample with a linear standard because a branched polymer is smaller than the equivalent linear structure at the same MW. Based on this, I proposed an alternative method for the quantification of kraft lignin linearity by comparing kraft lignin with a “more-linear” lignin model. The selection of a proper linear standard is the key to obtaining an accurate measurement of a branching ratio, which is simpler to do with synthetic polymers such as polystyrene. The standard should be of a similar repeat unit as kraft lignin and of high linearity. Therefore, enzymatic milled acidolysis lignin (EMAL) extracted from Eucalyptus was selected. Hardwood EMAL is comprised of both S and G subunits, with the S lignin component having reduced branching spots; the EMAL process prevents condensation at the 5 position, leading the lignin with mainly branches arising during cell wall biosynthesis. The linearity from this lignin was relatively high, as supported by NMR results, which showed an S content of 60–70 per 100 aromatics and a degree of condensation close to zero (Table 7.5). However, in the light scattering measurements, EMAL showed strong aggregations in GPC due to large molecular size and contamination of lignin-carbohydrate complexes (LCCs). To obtain a relatively clear light scattering profile and for better comparability to SKL fractions, EMAL was fractionated by acetone into acetone soluble EMAL (ASEMAL) and acetone insoluble EMAL (AIEMAL). The MW results are listed in Table 8.1 with the MHS plot shown in Figure 8.1. Note the NMR result is given in the next section. The MHS plot for EMAL lignin has a similar lower intrinsic viscosity, but the AIEMAL has a much higher upper limit. For this lignin the α value was 0.404, greater than the ASEMAL of 0.33. As shown in the light scattering profile of AIEMAL, there is a clear bimodal profile, which was a sign of aggregation, and the calculated MW exceeded 200  110  kDa based on light scattering data. ASEMAL had a clear elution profile and an MW of 11.8 kDa, which was equivalent to a DP of approximately 45. In addition, the effect of fluorescence on EMAL was less than 5%, suggesting that much of the fluorescent chromaphores were induced by kraft pulping. Before proceeding to analyze branch quantification, it is necessary to comment on EMAL standards that will be used as reference. First, the standards are partially branched because ASEMAL and AIEMAL had an α value of 0.33 and 0.40 in its MHS plot, respectively, as shown in Figure 8.1. Second, AIEMAL may contain aggregations that decreased the precision for hydrodynamic behaviour measurement, while ASEMAL had a very low DP; the response of this latter sample had hydrodynamic behaviour in between a random coil polymer and an oligomer. As a result of not having a perfectly linear model, the branching quantification was an analysis of how much native lignin conformation was modified during the pulping process, albeit one sample was a hardwood and the other sample was a softwood. Considering the significant aggregation effect in AIEMAL fraction, ASKL was selected as the linear reference.  Table 8.1 Molecular weight averages for EMAL fractions measured by MALS and conventional analysis  Light scattering analysis  Conventional analysis  Mn (kDa) MW (kDa) PDI Rh (nm)*  Mn (kDa) MW (kDa) PDI Ratio of LS /Convention ASEMAL-Eu 8.92 11.77 1.32 2.00  1.69 3.42 2.03 3.44 AIEMAL-Eu 176.19 214.63 1.21 6.80  1.98 8.52 4.42 25.18 *: analyzed from viscosity properties   111   Figure 8.1 MHS plot of AI-EMAL and AS-EMAL As shown in Figure 8.3 [a], the branching ratio (gM) was calculated by comparing the intrinsic viscosity of SKL fractions with ASEMAL standards at the same MW, mathematically expressed as Equation 8.1. Factor e is known as drainage parameter, which equals to 0.5 for a “non-draining polymer”, 1.0 for free draining polymer and 1.5 for Flory- Fox polymer164. Depending on branching structure, Zimm suggested value of 0.5 for star polymer under theta condition165, while Berry demonstrated that a value of 1.5 best represented a comb-type polymer166, and a value of 0.75 was recommended by Bohdanecky for general use167. In Figure 8.3 [b] & [c], the branching ratio of AIKL and ASKL at different drainage factor values is given as a function of MW.  𝑔𝑀𝑒 = 𝑔′ = ([𝜂]𝑏𝑟𝑎𝑛𝑐ℎ𝑒𝑑[𝜂]𝑙𝑖𝑛𝑒𝑎𝑟)𝑀 Equation 8.1  112     Figure 8.2 [a] Trendlines in MHS plot of AIKL (3 fractions), AKL (3 fractions), AI-EMAL, and AS-EMAL, [b] branching ratio at different MW for AIKL(a) and ASKL, assuming “e” equals to 0.5, 0.75, 1, 1.5.  113  The branching ratio is an abstract concept, but it can be converted into two more specific terms—branching units per molecule (Bn) and the number of branch points per 100 units (𝜆) according to the Equation 8.2 and Equation 8.390. In calculating the number of branch points, two assumption are made. The first assumption considers the analyzed lignin slice as a monodisperse polymer. The second assumption is that lignin only contains trifunctional groups as branch points. 𝑔𝑀 = [(1 +𝐵𝑛7)12+4𝐵𝑛9𝜋]−12 Equation 8.2 𝜆 =𝑈𝑛𝑖𝑡𝑀𝑤 × 𝐵𝑛 × 100𝑀𝑤 Equation 8.3 As shown in Figure 8.3 [a] & [b], branch points per 100 units at various e values were plotted against MW assuming an average unit MW of 190 Da. As shown in Figure 8.3 [c], at an e value of 0.75, in the range between 3 kDa and 40 kDa, where lignin MW was mostly distributed, branch frequency (𝜆) of ASKL was higher at higher MW, and the distribution was narrower than AIKL fractions. In addition, the branch frequency obtained was 5 to 30 for both AIKL and ASKL fractions, which is lower than the value of the degree of condensation around 40 per 100 aromatics obtained from the 13C NMR (Table 7.3). The trend that branch frequency decreased with an increase in MW also differs from 13C and 31P NMR results. As the linear reference was not perfect, the branch points calculated should be slightly lower than the true value. The increasing branch frequency with MW suggests that high MW lignin was less modified as compared with EMAL. Besides, the MW of linear standards mainly was distributed between 3kDa to 40kDa, and thus, the outside region was extrapolated and may differ from real behaviour. As a result, the absolute calculation on low and high MW region fall outside the  114  calibration region. In all, as compared with the EMAL control, high MW lignins were less modified than lower MW lignins and up to one fourth of the lignin subunits served as branching points. AIKL had a similar degree of branching when compared to the ASKL sample.  Compared with the NMR analysis, the major advatage of this method is avoiding the debatable defination on the branch structure of lignin via specific substitution. Because of this, inconsistent data was obtained from NMR when using different techniques. For example, Crestini et al. (2017) reported an increase in branch points at lower MW for fractionated lignin134, while Costa et al. (2018) observed an opposite trend on lignin fractions by ultrafiltration44. The major limitation of this method is the difficulties in finding a perfect linear reference method with wide enough MW distribution to cover the range of the samples.         115     Figure 8.3 Branch points per 100 lignin units of [a] AIKL fractions and [b] ASKL fractions  at different drainage factors “e” of 0.5, 0.75, 1, 1.5;  [c] Comparison of branch points between AIKL and ASKL at “e” of 0.75.   Reference range Reference range Reference range  116   Conclusion  Recently, studies have attempted to elucidate the structure and MW of softwood kraft lignin. However, there are still questions that surround the measurement of lignin MW. The commonly used GPC system with polystyrene calibration lacks the ability for effective technical lignin fractionation, as branched lignin may interact with column gel and result in a delay in elution relative to linear standards. As such, any measurement would be an oversimplification without effective fractionation. For this reason, pre-fractionation by acetone and membrane filtration was introduced for lignin analysis. From this perspective, experimental purification methods were used to remove some variation for the different lignin fractions from both solubility and size perspectives, to better understand the structure of lignin via light scattering analysis. Furthermore, NMR was used to indirectly support findings from light scattering, as there are no ideal lignin standards to validate light scattering data. The use of fractionated lignin enhanced the clarity of light scattering chromatograms providing better resolved data. Based on this data, the lignin fractions spanned a wide MW distribution from 4kDa to 110kDa. For solvent fractionated samples, AIKL had a higher molecular weight than its corresponding ASKL fractions. The absolute molecular weight, as characterized by light scattering, was 5-10 times larger than that of conventional GPC analysis. Fluorescence, when comparing scattered light with and without band pass filters, led to an overestimation of light scattering results by 20-80%, dependent upon the fraction analyzed. The influence of fluorescence on MW was less severe for ASKL than that on AIKL, and the impact of fluorescence decreased with increasing MW. Nevertheless, some limitations of light scattering of lignin was based on the inherent size of lignin, which limited accurate determination of Rg  117  from its Zimm plot. As well, certain samples (AIKL>10kDa) had evidence of aggregation that would shift the MW values to larger numbers that were note reasonable.  In addition to light scattering, further research into technical lignin structure was conducted with viscosity and NMR analysis. The combination of light scattering and the use of a differential viscometer revealed the hydrodynamic behaviour of lignin based on MHS plots. An α value of 0.20 was found for ASKL and 0.32 for AIKL, which indicated a slightly more extended structure for AIKL. Such finding aligned with the NMR results that show more native linkages (β-O-4, β-5, β-β) and aliphatic side chains in AIKL. Additionally, it was found that most linkages and functional groups decreased at a decreasing MW. Less native linkages indicated a more degraded molecule, and the changing rate of bond frequency was typical in cleavage reactions of networks.  Finally, the branching degree was investigated as a new method to quantify lignin branching by comparing the hydrodynamic behaviour of lignin with a “linear” reference, a hardwood enzymatic mild acidolysis lignin (EMAL). Analysis of branching degree revealed a similar condensation degree to the NMR results, and the branching level of ASKL decreased with molecular weight. However, the drawback of this method was the imperfect linear reference, and the result should be interpreted as the change in lignin branching during delignification. As a whole, this research characterized the absolute molar mass of technical lignin by light scattering and demonstrated a new method for semi-quantitive measurement of lignin branching. The structure of various “homogeneous” lignin fractions by a series of techniques was obtained limiting the variability within light scattering Zimm plots.    118  Bibliography (1)  Cui, C.; Sadeghifar, H.; Sen, S.; Argyropoulos, D. S. Toward Thermoplastic Lignin Polymers; Part II: Thermal & Polymer Characteristics of Kraft Lignin & Derivatives. BioResources 2013, 8 (1), 864–886. (2)  Petridis, L.; Schulz, R.; Smith, J. C. Simulation Analysis of the Temperature Dependence of Lignin Structure and Dynamics. J. Am. Chem. Soc. 2011, 133 (50), 20277–20287. (3)  Upton, B. M.; Kasko, A. M. Strategies for the Conversion of Lignin to High-Value Polymeric Materials: Review and Perspective. Chem. Rev. 2015, 116, 2277. (4)  Zhao, Z.; Moghadasian, M. H. Chemistry, Natural Sources, Dietary Intake and Pharmacokinetic Properties of Ferulic Acid: A Review. Food Chem. 2008, 109 (4), 691–702. (5)  Berlin, A.; Balakshin, M. Industrial Lignins: Analysis, Properties, and Applications; Elsevier, 2014. (6)  Henriksson, G. S == IáÖåáå. 121–146. (7)  Datta, R.; Kelkar, A.; Baraniya, D.; Molaei, A.; Moulick, A.; Meena, R. S.; Formanek, P. Enzymatic Degradation of Lignin in Soil: A Review. Sustain. 2017, 9 (7). (8)  Schorr, D.; Diouf, P. N.; Stevanovic, T. Evaluation of Industrial Lignins for Biocomposites Production. Ind. Crops Prod. 2014, 52, 65–73. (9)  Kalia, S.; Kaith, B. S.; Kaur, I. Pretreatments of Natural Fibers and Their Application as Reinforcing Material in Polymer Composites-a Review. Polym. Eng. Sci. 2009. (10)  Kai, D.; Tan, M. J.; Chee, P. L.; Chua, Y. K.; Yap, Y. L.; Loh, X. J.  119  Towards Lignin-Based Functional Materials in a Sustainable World. Green Chem. 2016, 18 (5), 1175–1200. (11)  Silva, E. A. B. da; Zabkova, M.; Araújo, J. D.; Cateto, C. A.; Barreiro, M. F.; Belgacem, M. N.; Rodrigues, A. E. An Integrated Process to Produce Vanillin and Lignin-Based Polyurethanes from Kraft Lignin. Chem. Eng. Res. Des. 2009. (12)  Kouisni, L.; Gagné, A.; Maki, K.; Holt-Hindle, P.; Paleologou, M. LignoForce System for the Recovery of Lignin from Black Liquor: Feedstock Options, Odor Profile, and Product Characterization. ACS Sustain. Chem. Eng. 2016, 4 (10), 5152–5159. (13)  Kouisni, L.; Holt-Hindle, P.; Maki, K.; Paleologou, M. The LignoForce System???: A New Process for the Production of High-Quality Lignin from Black Liquor. Pulp Pap. Canada 2014, 115 (1), 18–22. (14)  Mansouri, N. E. El; Salvadó, J. Structural Characterization of Technical Lignins for the Production of Adhesives: Application to Lignosulfonate, Kraft, Soda-Anthraquinone, Organosolv and Ethanol Process Lignins. Ind. Crops Prod. 2006, 24 (1), 8–16. (15)  Sahoo, S.; Seydibeyoĝlu, M. Ö.; Mohanty, A. K.; Misra, M. Characterization of Industrial Lignins for Their Utilization in Future Value Added Applications. Biomass and Bioenergy 2011, 35 (10), 4230–4237. (16)  Salanti, A.; Zoia, L.; Orlandi, M.; Zanini, F.; Elegir, G. Structural Characterization and Antioxidant Activity Evaluation of Lignins from Rice Husk. J. Agric. Food Chem. 2010, 58 (18), 10049–10055. (17)  Moxley, G.; Gaspar, A. R.; Higgins, D.; Xu, H. Structural Changes of Corn  120  Stover Lignin during Acid Pretreatment. J. Ind. Microbiol. Biotechnol. 2012, 39 (9), 1289–1299. (18)  Van De Pas, D. J.; Nanayakkara, B.; Suckling, I. D.; Torr, K. M. Comparison of Hydrogenolysis with Thioacidolysis for Lignin Structural Analysis. Holzforschung 2014, 68 (2), 151–155. (19)  Mansfield, S. D.; Kim, H.; Lu, F.; Ralph, J. Whole Plant Cell Wall Characterization Using Solution-State 2D NMR. Nat. Protoc. 2012, 7 (9), 1579–1589. (20)  Xia, Z.; Akim, L. G.; Argyropoulos, D. S. Quantitative 13C NMR Analysis of Lignins with Internal Standards. J. Agric. Food Chem. 2001, 49 (8), 3573–3578. (21)  Liu, W.-J.; Jiang, H.; Yu, H.-Q. Thermochemical Conversion of Lignin to Functional Materials: A Review and Future Directions. Green Chem. 2015, 4888–4907. (22)  Gellerstedt, G. Softwood Kraft Lignin: Raw Material for the Future. Ind. Crops Prod. 2015, 77, 845–854. (23)  Editor, P. C. K. L. Quality Living Through Chemurgy and Green Chemistry; 2016. (24)  Azadi, P.; Inderwildi, O. R.; Farnood, R.; King, D. A. Liquid Fuels, Hydrogen and Chemicals from Lignin: A Critical Review. Renew. Sustain. Energy Rev. 2013. (25)  Sulaeva, I.; Zinovyev, G.; Plankeele, J.-M.; Sumerskii, I.; Rosenau, T.; Potthast, A. Fast Track to Molar-Mass Distributions of Technical Lignins. ChemSusChem 2017. (26)  García, A.; Toledano, A.; Serrano, L.; Egüés, I.; González, M.; Marín, F.; Labidi, J. Characterization of Lignins Obtained by Selective Precipitation. Sep. Purif.  121  Technol. 2009, 68 (2), 193–198. (27)  Toledano, A.; García, A.; Mondragon, I.; Labidi, J. Lignin Separation and Fractionation by Ultrafiltration. Sep. Purif. Technol. 2010. (28)  Helander, M.; Theliander, H.; Lawoko, M.; Henriksson, G.; Zhang, L.; Lindström, M. E. Fractionation of Technical Lignin: Molecular Mass and PH Effects. BioResources 2013. (29)  Arkell, A.; Olsson, J.; Wallberg, O. Process Performance in Lignin Separation from Softwood Black Liquor by Membrane Filtration. Chem. Eng. Res. Des. 2014, 92 (9), 1792–1800. (30)  Toledano, A.; García, A.; Mondragon, I.; Labidi, J. Lignin Separation and Fractionation by Ultrafiltration. Sep. Purif. Technol. 2010, 71 (1), 38–43. (31)  Cui, C.; Sun, R.; Argyropoulos, D. S. Fractional Precipitation of Softwood Kraft Lignin: Isolation of Narrow Fractions Common to a Variety of Lignins. ACS Sustain. Chem. Eng. 2014, 2 (4), 959–968. (32)  Duval, A.; Vilaplana, F.; Crestini, C.; Lawoko, M. Solvent Screening for the Fractionation of Industrial Kraft Lignin. Holzforschung 2016, 70 (1), 11–20. (33)  MORCK, R.; YOSHIDA, H.; KRINGSTAD, K. P.; HATAKEYAMA, H. Fractionation of Kraft Lignin by Successive Extraction with Organic-Solvents .1. Functional-Groups, C13-NMR-Spectra and Molecular-Weight Distributions. Holzforschung 1986. (34)  Mörck, R.; Reimann, A.; Kringstad, K. P. Fractionation of Kraft Lignin by Successive Extraction with Organic Solvents: III. Fractionation of Kraft Lignin from Birch. Holzforschung 1988.  122  (35)  Yoshida, B. H.; Kringstad, K. P.; Pulp, S.; Hatakeyama, H. Fractionation of Kraft Lignin by Successive Extraction with Organic Solvents. 1987, 41 (3), 171–176. (36)  Saito, T.; Perkins, J. H.; Vautard, F.; Meyer, H. M.; Messman, J. M.; Tolnai, B.; Naskar, A. K. Methanol Fractionation of Softwood Kraft Lignin: Impact on the Lignin Properties. ChemSusChem 2014, 7 (1), 221–228. (37)  Lange, H.; Schiffels, P.; Sette, M.; Sevastyanova, O.; Crestini, C. Fractional Precipitation of Wheat Straw Organosolv Lignin: Macroscopic Properties and Structural Insights. ACS Sustain. Chem. Eng. 2016, 4 (10), 5136–5151. (38)  Teng, N. Y.; Dallmeyer, I.; Kadla, J. F. Effect of Softwood Kraft Lignin Fractionation on the Dispersion of Multiwalled Carbon Nanotubes. Ind. Eng. Chem. Res. 2013, 52 (19), 6311–6317. (39)  Keyoumu, A.; Sjödahl, R.; Henriksson, G.; Ek, M.; Gellerstedt, G.; Lindström, M. E. Continuous Nano- and Ultra-Filtration of Kraft Pulping Black Liquor with Ceramic Filters: A Method for Lowering the Load on the Recovery Boiler While Generating Valuable Side-Products. Ind. Crops Prod. 2004, 20 (2), 143–150. (40)  van de Pas, D.; Hickson, A.; Donaldson, L.; Lloyd-Jones, G.; Tamminen, T.; Fernyhough, A.; Mattinen, M. L. Characterization of Fractionated Lignins Polymerized by Fungal Laccases. BioResources 2011, 6 (2), 1105–1121. (41)  Sun, R.; Tomkinson, J.; Bolton, J. Effects of Precipitation PH on the Physico-Chemical Properties of the Lignins Isolated from the Black Liquor of Oil Palm Empty Fruit Bunch Fibre Pulping. Polym. Degrad. Stab. 1999, 63 (2), 195–200. (42)  Jönsson, A. S.; Nordin, A. K.; Wallberg, O. Concentration and Purification of Lignin in Hardwood Kraft Pulping Liquor by Ultrafiltration and Nanofiltration. Chem.  123  Eng. Res. Des. 2008, 86 (11), 1271–1280. (43)  Sevastyanova, O.; Helander, M.; Chowdhury, S.; Lange, H.; Wedin, H.; Zhang, L.; Ek, M.; Kadla, J. F.; Crestini, C.; Lindstr, M. E. Tailoring the Molecular and Thermo – Mechanical Properties of Kraft Lignin by Ultrafiltration. 2014, 40799 (May), 1–11. (44)  Costa, C. A. E.; Pinto, P. C. R.; Rodrigues, A. E. Lignin Fractionation from E. Globulus Kraft Liquor by Ultrafiltration in a Three Stage Membrane Sequence. Sep. Purif. Technol. 2018, 192 (September 2017), 140–151. (45)  Gioia, C.; Lo Re, G.; Lawoko, M.; Berglund, L. Tunable Thermosetting Epoxies Based on Fractionated and Well-Characterized Lignins. J. Am. Chem. Soc. 2018, 140 (11), 4054–4061. (46)  Balani, K. Physical and Mechanical Properties of Polymers. Biosurfaces A Mater. Sci. Eng. Perspect. 2015, 329–344. (47)  Oberlerchner, J. T.; Rosenau, T.; Potthast, A. Overview of Methods for the Direct Molar Mass Determination of Cellulose. Molecules 2015, 20 (6), 10313–10341. (48)  Lange, H.; Rulli, F.; Crestini, C. Gel Permeation Chromatography in Determining Molecular Weights of Lignins: Critical Aspects Revisited for Improved Utility in the Development of Novel Materials. ACS Sustain. Chem. Eng. 2016, 4 (10), 5167–5180. (49)  Agilent Technologies. An Introduction to Gel Permeation Chromatography and Size Exclusion Chromatography. Primer 2014. (50)  Podzimek, S. Size Exclusion Chromatography. Light Scatt. Size Exclusion Chromatogr. Asymmetric Flow F. Flow Fractionation 2011, 99–206.  124  (51)  Cathala, B.; Saake, B.; Faix, O.; Monties, B. Association Behaviour of Lignins and Lignin Model Compounds Studied by Multidetector Size-Exclusion Chromatography. J. Chromatogr. A 2003. (52)  Glasser, W. G.; Davé, V.; Frazier, C. E. Molecular Weight Distribution of (Semi-) Commercial Lignin Derivatives. J. Wood Chem. Technol. 1993, 134, 545–559. (53)  Gosselink, R. J. A.; Abächerli, A.; Semke, H.; Malherbe, R.; Käuper, P.; Nadif, A.; Van Dam, J. E. G. Analytical Protocols for Characterisation of Sulphur-Free Lignin. Ind. Crops Prod. 2004, 19 (3), 271–281. (54)  Majcherczyk, A.; Hüttermann, A. Size-Exclusion Chromatography of Lignin as Ion-Pair Complex. J. Chromatogr. A 1997, 764 (2), 183–191. (55)  Ringena, O.; Saake, B.; Lehnen, R. Isolation and Fractionation of Lignosulfonates by Amine Extraction and Ultrafiltration: A Comparative Study. Holzforschung 2005, 59 (4), 405–412. (56)  Ringena, O.; Lebioda, S.; Lehnen, R.; Saake, B. Size-Exclusion Chromatography of Technical Lignins in Dimethyl Sulfoxide/Water and Dimethylacetamide. J. Chromatogr. A 2006, 1102 (1–2), 154–163. (57)  Guerra, A.; Gaspar, A. R.; Contreras, S.; Lucia, L. A.; Crestini, C.; Argyropoulos, D. S. On the Propensity of Lignin to Associate: A Size Exclusion Chromatography Study with Lignin Derivatives Isolated from Different Plant Species. Phytochemistry 2007. (58)  Sjöholm, E.; Gustafsson, K.; Colmsjö, A. Size Exclusion Chromatography of Lignins Using Lithium Chloride/N,N-Dimethylacetamide As Mobile Phase. Ii. Dissolved and Residual Pine Kraft Lignins. J. Liq. Chromatogr. Relat. Technol. 1999, 22  125  (18), 2837–2854. (59)  Fredheim, G. E.; Braaten, S. M.; Christensen, B. E. Molecular Weight Determination of Lignosulfonates by Size-Exclusion Chromatography and Multi-Angle Laser Light Scattering. J. Chromatogr. A 2002, 942 (1), 191–199. (60)  Pellinen, J.; Salkinoja-Salonen, M. Aqueous Size Exclusion Chomatography of Industrial Lignins. J. Chromatogr. A 1985, 322 (C), 129–138. (61)  Chen, F.; Li, J. Aqueous Gel Permeation Chromatographic Methods for Technical Lignins. J. Wood Chem. Technol. 2000, 20 (3), 265–276. (62)  Pakkanen, H.; Alén, R. Molecular Mass Distribution of Lignin from the Alkaline Pulping of Hardwood, Softwood, and Wheat Straw. J. Wood Chem. Technol. 2012, 32 (4), 279–293. (63)  Dong, D.; Fricke, A. L. Intrinsic Viscosity and the Molecular Weight of Kraft Lignin. 1995, 36 (10), 2075–2078. (64)  Myrvold, B. O. A New Model for the Structure of Lignosulphonates. Part 1. Behaviour in Dilute Solutions. Ind. Crops Prod. 2008, 27 (2), 214–219. (65)  Rezanowich, A.; Goring, D. A. . Polyelectrolyte Expansion of a Lignin Sulfonate Microgel. J. Colloid Sci. 1960, 15 (5), 452–471. (66)  Specification for Tetrahydrofuran , Unstabilized , For UV , HPLC Grade ( HP842 ). No. 0, 842. (67)  Baumberger, S.; Abaecherli, A.; Fasching, M.; Gellerstedt, G.; Gosselink, R.; Hortling, B.; Li, J.; Saake, B.; De Jong, E. Molar Mass Determination of Lignins by Size-Exclusion Chromatography: Towards Standardisation of the Method. Holzforschung 2007, 61 (4), 459–468.  126  (68)  Williams, A.; Hagel, L. Column Handbook for Size Exclusion Chromatography. Column Handb. Size Exclusion Chromatogr.  1999, 27–74. (69)  Pbt, P. E. T.; Ht, S. Waters Styragel GPC Column Selection Chart. 2000, 2000. (70)  Chum, H. L.; Johnson, D. K.; Agblevor, F. A.; Robert, J.; Hames, B. R.; Milne, T. A.; Overend, R. P. STATUS OF THE LEA VOLUNTARY STANDARDS ACTIVITY- ROUND ROBINS ON WHOLE WOOD AND LIGNINS. 1701–1716. (71)  Ritter, A.; Schmid, M.; Affolter, S. Determination of Molecular Weights by Size Exclusion Chromatography (SEC) - Results of Round Robin Tests. Polym. Test. 2010, 29 (8), 945–952. (72)  Axelsson, L.; Franzén, M.; Ostwald, M.; Berndes, G.; Lakshmi, G.; Ravindranath, N. H. Characterization and Analysis of the Molecular Weight of Lignin for Biorefi Ning Studies. Biofuels, Bioprod. Biorefining 2012, 6 (3), 246–256. (73)  Asikkala, J.; Tamminen, T.; Argyropoulos, D. S. Accurate and Reproducible Determination of Lignin Molar Mass by Acetobromination. J. Agric. Food Chem. 2012, 60 (36), 8968–8973. (74)  Podzimek, S. Light Scattering. Light Scatt. Size Exclusion Chromatogr. Asymmetric Flow F. Flow Fractionation 2011, 37–98. (75)  Dem’yantseva, E. Y.; Lysogorskaya, N. P.; Klyubin, V. V.; Zaitseva, S. V. A Dynamic Light Scattering Study of the Temperature Dependence of the Size-Distribution Pattern and Aggregation Stability of Sulfate Lignin and Wood Resin in Aqueous Alkali Solution. Russ. J. Appl. Chem. 2002. (76)  Milne, T. A.; Chum, H. L.; Agblevor, F.; Johnson, D. K. Standardized  127  Analytical Methods. Biomass and Bioenergy 1992, 2 (1–6), 341–366. (77)  Sluiter,  a.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D.; Crocker, D. NREL/TP-510-42618 Analytical Procedure - Determination of Structural Carbohydrates and Lignin in Biomass. Lab. Anal. Proced. 2012, No. April 2008, 17. (78)  Podzimek, S. White Paper Wyatt Technology Corporation Branching Revealed : Characterizing Molecular Structure in Synthetic and Natural Polymers by Multi-Angle Light Scattering. (79)  Sarkanen, S.; Teller, D. C.; Stevens, C. R.; Mccarthy, J. L. Lignin. 20. Associative Interactions Between Kraft Lignin Components. Macromolecules 1984, 17 (12), 2588–2597. (80)  Sarkanen, S.; Sarkanen, S.; Teller, D. C.; Abramowski, E.; McCarthy, J. L. Lignin. 19. Kraft Lignin Component Conformation and Associated Complex Configuration in Aqueous Alkaline Solution. Macromolecules 1982, 15 (4), 1098–1104. (81)  Zabet-Moghaddam, M.; Heinzle, E.; Tholey, A. Qualitative and Quantitative Analysis of Low Molecular Weight Compounds by Ultraviolet Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry Using Ionic Liquid Matrices. Rapid Commun. Mass Spectrom. 2004, 18 (2), 141–148. (82)  Evtuguin, D. V.; Domingues, P.; Amado, F. L.; Neto, C. P.; Correia, A. J. F. Electrospray Ionization Mass Spectrometry as a Tool for Lignins Molecular Weight and Structural Characterisation. Holzforschung 1999, 53, 525–528. (83)  Evtuguin, D. V.; Amado, F. M. L. Application of Electrospray Ionization Mass Spectrometry to the Elucidation of the Primary Structure of Lignin. Macromol. Biosci. 2003, 3 (7), 339–343.  128  (84)  Richel, A.; Vanderghem, C.; Simon, M.; Wathelet, B.; Paquot, M. Evaluation of Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry for Second-Generation Lignin Analysis. Anal. Chem. Insights 2012, 7 (1), 79–89. (85)  Himmel, M. E.; Tatsumoto, K.; Grohmann, K. Molecular Weight Distribution of Aspen Lignins From Con Ventional Gel Permeation Chromatography, Universal Calibration and Sedimentation Equilibrium. J. Chromatogr. 1990, 498, 93–104. (86)  Gupta, P. R.; McCarthy, J. L. Lignin. XIV. Gel Chromatography and the Distribution in Molecular Size of Lignin Sulfonates at Several Electrolyte Concentrations. Macromolecules 1968, 1 (3), 236–244. (87)  Podzimek, S. Chapter 1 - Polymers. Light Scatt. Size Exclusion Chromatogr. Asymmetric Flow F. Flow Fractionation 2011, 1–36. (88)  Crestini, C.; Melone, F.; Sette, M.; Saladino, R. Milled Wood Lignin: A Linear Oligomer. Biomacromolecules 2011, 12 (11), 3928–3935. (89)  Chapter 5, Light and Matter. Electromagnetic Wave. (90)  Corporation, W. T. ASTRA 6 User ’ s Guide. 1. (91)  Tn, T. N. ASTRA Quick Guide. 1 (805), 1–21. (92)  01 - General MALS Theory 2014. (93)  Moacanin, J.; Felicetta, V. F.; Haller, W.; Mccarthy, J. L. Lignin. VI. Molecular Weights of Lignin Sulfonates by Light Scattering. J. Am. Chem. Soc. 1955, 77 (13), 3470–3475. (94)  Gidh, A. V.; Decker, S. R.; Vinzant, T. B.; Himmel, M. E.; Williford, C. Determination of Lignin by Size Exclusion Chromatography Using Multi Angle Laser Light Scattering. J. Chromatogr. A 2006, 1114 (1), 102–110.  129  (95)  Zinovyev, G.; Sulaeva, I.; Podzimek, S.; Rössner, D.; Kilpelainen, I.; Sumerskii, I.; Rosenau, T.; Potthast, A. Getting Closer to Absolute Molar Masses of Technical Lignins. ChemSusChem 2018. (96)  Cathala, B.; Saake, B.; Faix, O.; Monties, B. Association Behaviour of Lignins and Lignin Model Compounds Studied by Multidetector Size-Exclusion Chromatography. J. Chromatogr. A 2003, 1020 (2), 229–239. (97)  Gidh, A. V.; Decker, S. R.; See, C. H.; Himmel, M. E.; Williford, C. W. Characterization of Lignin Using Multi-Angle Laser Light Scattering and Atomic Force Microscopy. Anal. Chim. Acta 2006, 555 (2), 250–258. (98)  Contreras, S.; Gaspar, A. R.; Guerra, A.; Lucia, L. A.; Argyropoulos, D. S. Propensity of Lignin to Associate: Light Scattering Photometry Study with Native Lignins. Biomacromolecules 2008, 9 (12), 3362–3369. (99)  Qian, Y.; Deng, Y.; Guo, Y.; Yi, C.; Qiu, X. Determination of Absolute Molecular Weight of Sodium Lignosulfonates (NaLS) by Laser Light Scattering (LLS). Holzforschung 2013, 67 (3), 265–271. (100)  Contreras, S.; Gaspar, A. R.; Guerra, A.; Lucia, L. A.; Argyropoulos, D. S. Propensity of Lignin to Associate: Light Scattering Photometry Study with Native Lignins. Biomacromolecules 2008, 9 (12), 3362–3369. (101)  Deng, Y.; Feng, X.; Zhou, M.; Qian, Y.; Yu, H.; Qiu, X. Investigation of Aggregation and Assembly of Alkali Lignin Using Iodine as a Probe. Biomacromolecules 2011, 12 (4), 1116–1125. (102)  Lindstr??mn, T. The Colloidal Behaviour of Kraft Lignin - Part I.: Association and Gelation of Kraft Lignin in Aqueous Solutions. Colloid Polym. Sci.  130  Kolloid Zeitschrift Zeitschrift f??r Polym. 1979. (103)  Gan, L.; Zhou, M.; Yang, D.; Qiu, X. Aggregation and Adsorption Behaviors of Carboxymethylated Lignin (CML) in Aqueous Solution. Holzforschung 2012. (104)  Li, H.; Deng, Y. H.; Qiu, X. Q. Influence of Acetylation on the Microstructure of Alkali Lignin in Tetrahydrofuran. Wuli Huaxue Xuebao/ Acta Phys. - Chim. Sin. 2015, 31 (1), 128–136. (105)  Tolbert, A.; Akinosho, H.; Khunsupat, R.; Naskar, A. K.; Ragauskas, A. J. Characterization and Analysis of the Molecular Weight of Lignin for Biorefining Studies. Biofuels, Bioproducts and Biorefining. 2014. (106)  Gilca, I. A.; Popa, V. I.; Crestini, C. Obtaining Lignin Nanoparticles by Sonication. Ultrason. Sonochem. 2015, 23, 369–375. (107)  Guerra, A.; Filpponen, I.; Lucia, L. A.; Argyropoulos, D. S. Comparative Evaluation  of Three Lignin Isollation Protocols for Various Wood Species. J. Agric. Food Chem. 2006, 54, 9696–9705. (108)  Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd Edition, Joseph R. Lakowicz, Editor; 2006. (109)  Coletta, V. C.; Rezende, C. A.; Conceição, F. R. Da; Polikarpov, I.; Guimarães, F. E. G. Mapping the Lignin Distribution in Pretreated Sugarcane Bagasse by Confocal and Fluorescence Lifetime Imaging Microscopy. Biotechnol. Biofuels 2013, 6 (1), 43. (110)  Donaldson, L. A.; Radotic, K. Fluorescence Lifetime Imaging of Lignin Autofluorescence in Normal and Compression Wood. J. Microsc. 2013, 251 (2), 178–187.  131  (111)  Xue, Y.; Qiu, X.; Wu, Y.; Qian, Y.; Zhou, M.; Deng, Y.; Li, Y. Aggregation-Induced Emission: The Origin of Lignin Fluorescence. Polym. Chem. 2016, 7. (112)  Lupoi, J. S.; Singh, S.; Parthasarathi, R.; Simmons, B. A.; Henry, R. J. Recent Innovations in Analytical Methods for the Qualitative and Quantitative Assessment of Lignin. Renew. Sustain. Energy Rev. 2015, 49, 871–906. (113)  Donaldson, L.; Radoti??, K.; Kalauzi, A.; Djikanovi??, D.; Jeremi??, M. Quantification of Compression Wood Severity in Tracheids of Pinus Radiata D. Don Using Confocal Fluorescence Imaging and Spectral Deconvolution. J. Struct. Biol. 2010, 169 (1), 106–115. (114)  Fleming, R. J. An Introduction to Polymer Physics. Physicist. 2003. (115) . ._Polymer Physics Michael Rubinstein, Ralph H Colby-Standand.Pdf. (116)  Cotts, P. M. Characterization of Polymer Chain Architecture: Size Exclusion       Chromatography with Intensity and Dynamic Light Scattering and Viscometric       Detectors. Mult. Detect. Size Exclusion Chromatogr. Ser. (117)  Karmanov, A. P.; Monakov, Y. B. Lignin. Structural Organisation and Fractal Properties. Russ. Chem. Rev. 2003, 72 (8), 715–734. (118)  Vainio, U.; Maximova, N.; Hortling, B.; Laine, J.; Stenius, P.; Simola, L. K.; Gravitis, J.; Serimaa, R. Morphology of Dry Lignins and Size and Shape of Dissolved Kraft Lignin Particles by X-Ray Scattering. Langmuir 2004, 20 (22), 9736–9744. (119)  Ratnaweera, D. R.; Saha, D.; Pingali, S. V.; Labb??, N.; Naskar, A. K.; Dadmun, M. The Impact of Lignin Source on Its Self-Assembly in Solution. RSC Adv. 2015, 5 (82), 67258–67266.  132  (120)  Takayama, S. Characterization of Branched Polymers. Kobunshi 1989, 38 (2), 154–157. (121)  Hermiati, E.; Risanto, L.; Lubis, M. A. R.; Laksana, R. P. B.; Dewi, A. R. Chemical Characterization of Lignin from Kraft Pulping Black Liquor of Acacia Mangium. AIP Conf. Proc. 2017, 1803. (122)  Harton, S. E.; Pingali, S. V.; Nunnery, G. A.; Baker, D. A.; Walker, S. H.; Muddiman, D. C.; Koga, T.; Rials, T. G.; Urban, V. S.; Langan, P. Evidence for Complex Molecular Architectures for Solvent-Extracted Lignins. ACS Macro Lett. 2012, 1 (5), 568–573. (123)  大谷強. No Title日本における介護サ ビス 市場の 特徴 経済学的分析の 前提の 検討. (124)  Goring, D. A. I.; Vuong, R.; Gancet, C.; Chanzy, H. The Flatness of Lignosulfonate Macromolecules as Demonstrated by Electron Microscopy. J. Appl. Polym. Sci. 1979, 24 (4), 931–936. (125)  Garver, T. M.; Callaghan, P. T. Hydrodynamics of Kraft Lignins. Macromolecules 1991, 24 (2), 420–430. (126)  Can You Get Shape Information From Dynamic Light Scattering ? 1–3. (127)  Petridis, L.; Pingali, S. V.; Urban, V.; Heller, W. T.; O’Neill, H. M.; Foston, M.; Ragauskas, A.; Smith, J. C. Self-Similar Multiscale Structure of Lignin Revealed by Neutron Scattering and Molecular Dynamics Simulation. Phys. Rev. E - Stat. Nonlinear, Soft Matter Phys. 2011, 83 (6), 4–7. (128)  Zikeli, F.; Ters, T.; Fackler, K.; Srebotnik, E.; Li, J. Fractionation of Wheat Straw Dioxane Lignin Reveals Molar Mass Dependent Structural Differences. Ind. Crops  133  Prod. 2016, 91, 186–193. (129)  Gupta, P.; Goring, P. Physicochemical Studies of Alkali Lignins. Can. J. Chem. 1960, 38, 270–279. (130)  Kontturi, A. K.; Kontturi, K.; Niinikoski, P. Transport of Lignosulphonates under an External Electric Field. J. Chem. Soc. Faraday Trans. 1991, 87 (11), 1779–1783. (131)  Norgren, M.; Edlund, H.; Wågberg, L. Aggregation of Lignin Derivatives under Alkaline Conditions. Kinetics and Aggregate Structure. Langmuir 2002, 18 (7), 2859–2865. (132)  Ralph, J.; Lundquist, K.; Brunow, G.; Lu, F.; Kim, H.; Schatz, P. F.; Marita, J. M.; Hatfield, R. D.; Ralph, S. A.; Christensen, J. H.; et al. Lignins: Natural Polymers from Oxidative Coupling of 4-Hydroxyphenyl- Propanoids. Phytochem. Rev. 2004, 3 (1–2), 29–60. (133)  Zhao, W.; Xiao, L. P.; Song, G.; Sun, R. C.; He, L.; Singh, S.; Simmons, B. A.; Cheng, G. From Lignin Subunits to Aggregates: Insights into Lignin Solubilization. Green Chem. 2017, 19 (14), 3272–3281. (134)  Crestini, C.; Lange, H.; Sette, M.; Argyropoulos, D. S. On the Structure of Softwood Kraft Lignin. Green Chem. 2017, 19, 4104–4121. (135)  Yu, L. ‐P; Rollings, J. E. Low‐angle Laser Light Scattering–aqueous Size Exclusion Chromatography of Polysaccharides: Molecular Weight Distribution and Polymer Branching Determination. J. Appl. Polym. Sci. 1987, 33 (6), 1909–1921. (136)  Ruijter, K. De; Roessner, D.; Podzimek, S.; Scherrers, R.; Johann, C. Polymer Characterization by AF4-MALS. 5. (137)  Millipore. Protein Concentration and Diafiltration by Tangential Flow  134  Filtration. Tech. Br. 2003, 1–24. (138)  Sadeghifar, H.; Dickerson, J. P.; Argyropoulos, D. S. Quantitative 31P NMR Analysis of Solid Wood Offers an Insight into the Acetylation of Its Components. Carbohydr. Polym. 2014. (139)  Pu, Y.; Cao, S.; Ragauskas, A. J. Application of Quantitative31P NMR in Biomass Lignin and Biofuel Precursors Characterization. Energy Environ. Sci. 2011, 4 (9), 3154–3166. (140)  Liu, L.-Y.; Cho, M.; Sathitsuksanoh, N.; Chowdhury, S.; Renneckar, S. Uniform Chemical Functionality of Technical Lignin Using Ethylene Carbonate for Hydroxyethylation and Subsequent Greener Esterification. ACS Sustain. Chem. Eng. 2018, 6, acssuschemeng.8b02649. (141)  Balakshin, M. Y.; Capanema, E. A.; Santos, R. B.; Chang, H. M.; Jameel, H. Structural Analysis of Hardwood Native Lignins by Quantitative 13C NMR Spectroscopy. Holzforschung 2016, 70 (2), 95–108. (142)  Yue, F.; Lu, F.; Ralph, S.; Ralph, J. Identification of 4-O-5-Units in Softwood Lignins via Definitive Lignin Models and NMR. Biomacromolecules 2016, 17 (6), 1909–1920. (143)  Adler, E. Wood Science Anci Technology. 1977, 8. (144)  Jiang, X.; Savithri, D.; Du, X.; Pawar, S.; Jameel, H.; Chang, H. M.; Zhou, X. Fractionation and Characterization of Kraft Lignin by Sequential Precipitation with Various Organic Solvents. ACS Sustain. Chem. Eng. 2017, 5 (1), 835–842. (145)  Sadeghifar, H.; Wells, T.; Le, R. K.; Sadeghifar, F.; Yuan, J. S.; Jonas Ragauskas, A. Fractionation of Organosolv Lignin Using Acetone:Water and Properties of  135  the Obtained Fractions. ACS Sustain. Chem. Eng. 2017, 5 (1), 580–587. (146)  Huang, C.; He, J.; Narron, R.; Wang, Y.; Yong, Q. Characterization of Kraft Lignin Fractions Obtained by Sequential Ultrafiltration and Their Potential Application as a Biobased Component in Blends with Polyethylene. ACS Sustain. Chem. Eng. 2017, 5 (12), 11770–11779. (147)  Shao, L.; Zhang, X.; Chen, F.; Xu, F. Fast Pyrolysis of Kraft Lignins Fractionated by Ultrafiltration. J. Anal. Appl. Pyrolysis 2017, 128 (October), 27–34. (148)  Suto, S.; Ise, H.; Karasawa, M. Refractometry of Lyotropic Mesophase Systems: Ethyl Cellulose/M‐cresol and Ethyl Cellulose/Acetic Acid. J. Polym. Sci. Part B Polym. Phys. 1986, 24 (7), 1515–1524. (149)  Siochi, E. J.; Ward, T. C.; Haney, M. A.; Mahn, B. The Absolute Molecular Weight Distribution of Hydroxypropylated Lignins. Macromolecules 1990, 23 (5), 1420–1429. (150)  L Schwartz, K. S. Introduction to Tangential Flow Filtration for Laboratory and Process Development Applications. Pall Sci. Tech. Rep. 2002. (151)  Chum, H. L.; Johnson, D. K.; Tucker, M. P.; Himmel, M. E. Some Aspects of Lignin Characterization by High Performance Size Exclusion Chromatography Using Styrene Divinylbenzene Copolymer Gels. Holzforschung 1987, 41 (2), 97–108. (152)  Guerra, A.; Gaspar, A. R.; Contreras, S.; Lucia, L. A.; Crestini, C.; Argyropoulos, D. S. On the Propensity of Lignin to Associate: A Size Exclusion Chromatography Study with Lignin Derivatives Isolated from Different Plant Species. Phytochemistry 2007, 68 (20), 2570–2583. (153)  Crestini, C.; Melone, F.; Sette, M.; Saladino, R. Milled Wood Lignin: A  136  Linear Oligomer. Biomacromolecules 2011. (154)  Saunders, G.; Maccreath, B. A Guide to Multi-Detector Gel Permeation Chromatography Contents. Agil. GPC Backgr. 2012. (155)  Scattering, L. White Paper Wyatt Technology Corporation Branching Revealed : Characterizing Molecular Structure in Synthetic and Natural Polymers by Multi-Angle Light Scattering. (156)  Sette, M.; Wechselberger, R.; Crestini, C. Elucidation of Lignin Structure by Quantitative 2D NMR. Chem. - A Eur. J. 2011, 17 (34), 9529–9535. (157)  Constant, S.; Wienk, H. L. J.; Frissen, A. E.; Peinder, P. De; Boelens, R.; Van Es, D. S.; Grisel, R. J. H.; Weckhuysen, B. M.; Huijgen, W. J. J.; Gosselink, R. J. A.; et al. New Insights into the Structure and Composition of Technical Lignins: A Comparative Characterisation Study. Green Chem. 2016, 18 (9), 2651–2665. (158)  Gellerstedt, G.; Majtnerova, A.; Zhang, L. Towards a New Concept of Lignin Condensation in Kraft Pulping. Initial Results. Comptes Rendus - Biol. 2004, 327 (9–10), 817–826. (159)  Chakar, F. S.; Ragauskas, A. J. Review of Current and Future Softwood Kraft Lignin Process Chemistry. In Industrial Crops and Products; 2004. (160)  Yuan, T.; He, J.; Xu, F.; Sun, R. Fractionation and Physico-Chemical Analysis of Degraded Lignins from the Black Liquor of Eucalyptus Pellita KP-AQ Pulping. Polym. Degrad. Stab. 2009, 94 (7), 1142–1150. (161)  Balakshin, M. Y.; Capanema, E. A.; Santos, R. B.; Chang, H.; Jameel, H. Structural Analysis of Hardwood Native Lignins by Quantitative 13C NMR Spectroscopy. Holzforschung 2016, 70 (2).  137  (162)  Lignin: Historical, Biological, and Materials Perspectives; 1999. (163)  Casey, J. P. Pulp and Paper, Chemistry and Chemical Technology, Volume I. Bioresour. Technol. 1980. (164)  Zimm, B. H. Commentary: Reflections on “Dynamics of Branched Polymer Molecules in Dilute Solution,” by Bruno H. Zimm and Ralph W. Kilb,J. Polym. Sci., XXXVII, 19 (1959). J. Polym. Sci. Part B Polym. Phys. 2005. (165)  Berry, G. C.; Hobbs, L. M.; Long, V. C. Solution and Bulk Properties of Branched Polyvinyl Acetates Part III-Intrinsic Viscosity and Light Scattering Measurements. Polymer (Guildf). 1964. (166)  Bohdanecky, M. A Semiempirical Formulation of the Effect of Random Branching on Intrinsic Viscosity. Macromolecules 1977, 10 (5), 971–975. (167)  Bahary, W. S.; Hogan, M. P. Assessment of Branching in Polymers by Size Exclusion Chromatography with Light Scattering, Viscosity, and Refractive Index Detection. Int. J. Polym. Anal. Charact. 1996, 2 (2), 121–133.   138  Appendix  Appendix A  Light scattering profile  The normalized light scattering profile is given in Figure A.1, which shifts forward in elution volume, as similar to the RI profiles. This response is typical for polydisperse polymers as light scattering is proportional to the size and concentration (𝑀𝑤 ∗ 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛), whereby RI only is dependent on concentration.   Figure A.1 Light scattering profile of [a] AIKL and [b] ASKL  Figure A.2 shows the Zimm plot of lignin fractions, where the Rayleigh ratio (𝐾∗𝐶𝑅(𝜃)) was plotted as a function of angle (sin2 (𝜃2)). Seven detectors installed with fluorescence filter bands were used for measurements, and a few outlier values were unselected for the plot based on best  139  practices. The intercept of the fitting line on the y-axis was the value of 1/Mw, and the slope of the fitting line related to the radius of gyration for the molecule. It should be noted that each slice of the light scattering profile has a Zimm plot, and by calculating the Mw across the entire peak range, the overall Mw and PDI were obtained for the samples. As can be seen from the Zimm plots, most of the lignin fractions had data that did not clearly fit a trend, despite the removal of outlier values. Lignin has a small size relative to other polymers, limiting its angular dependence detection, this constraint also is the reason why light scattering results have a quite large variance.       140      Figure A.2 Zimm plot of lignin fractions. [a] AIKL, [b] ASKL, [c] AIKL>10kDa, [d] ASKL>10kDa, [e] AIKL 3-10kDa, [f] ASKL 3-10kDa, [g] AIKL <3kDa, [h] ASKL<3kDa  Appendix B  Acetylation correction  Because of the presence of acetyl groups, the absolute MW was overestimated by light scattering, and therefore, the real MW should be adjusted. The adjusted MW is given in Table B.1, and the procedure is given as follows; the amount of hydroxyl groups was first converted to the amount per 100 lignin units based on the hydroxyl content measured by 31P NMR (mmol/g), as shown in the equation below:  𝑂𝐻(𝑝𝑒𝑟 𝑢𝑛𝑖𝑡𝑠) = 𝑂𝐻(𝑚𝑚𝑜𝑙/𝑔) ∗𝑈𝑛𝑖𝑡 𝑀𝑊1000  141  the subunit MW was adjusted by adding the difference in molecular weight between a hydroxyl group (17 g/mol) and an ester group (59 g/mol):   𝐴𝑑𝑗𝑢𝑠𝑡𝑒𝑑 𝑢𝑛𝑖𝑡 𝑀𝑊 = 𝑢𝑛𝑖𝑡 𝑀𝑊 + 𝑂𝐻(𝑝𝑒𝑟 𝑢𝑛𝑖𝑡𝑠) ∗ 42 The real number average molecular weight was adjusted following the equation:  𝐴𝑑𝑗𝑢𝑠𝑡𝑒𝑑 𝑀𝑛 = 𝑀𝑛 ∗𝑢𝑛𝑖𝑡 𝑀𝑊𝐴𝑑𝑗𝑢𝑠𝑡𝑒𝑑 𝑢𝑛𝑖𝑡 𝑀𝑊  Table B.1 Adjusted molecular weight of lignin fractions  OH (per units) Unit MW Adjusted Unit MW Adjusted Mn (kDa) Adjusted DP AIKL 1.35 188.91 245.72 46.10 244 ASKL 1.23 177.34 228.96 10.58 60 AIKL >10kDa 1.46 189.53 250.82 55.18 291 ASKL >10kDa 1.35 190.62 247.30 20.22 106 AIKL 3-10kDa 1.19 185.95 235.78 7.06 38 ASKL 3-10kDa 1.16 181.90 230.41 5.63 31 AIKL <3kDa 1.03 182.05 225.48 7.13 39 ASKL <3kDa 1.08 175.22 220.55 2.71 15 Each sample was analyzed in duplicate.  Appendix C  EMAL light scattering  The light scattering profile of acetone insoluble enzymatic milled acidolysis lignin (AIEMAL) and acetone soluble enzymatic milled acidolysis lignin (ASEMAL) are given in Figure C.1. AIEMAL had a significant bimodal light scattering profile where the first peak barely had any RI response. It may be a sign of significant aggregations due to the existence of carbohydrate with a very small concentration of material. ASEMAL had a better profile in spite of the ‘hook’ curve at the end of elution. This response may arise from the existence of lignin  142  column interactions from branched materials or fluorescence from small chromophores although EMAL has less fluorescence than technical lignin.      Figure C.1 Molecular weight distribution, light scattering and RI profile of [left] AIEMWL and [right] ASEMW  Appendix D  Molecular density  The molecular density is calculated as equation below, where 𝜌 is the density, Rh is the hydrodynamic radius and Nv is Avogadro constant, which equals to 6.022×1023. 𝜌 =𝑀𝑛34 𝜋 𝑅ℎ3 ∗ 𝑁𝑣 Table D.1 shows the molecular density of all the lignin fractions. All AIKL fractions had a lower molecular density indicating a more extended structure than ASKL. In addition, as Mw decreased, molecular density increased and size decreased, which is to be expected as smaller molecules tend to be more compact with additional branching.       143    Table D.1 Hydrodynamic radius calculated from viscostar and molecular mass density  Mn (kDa) Rh (nm) Density (g/cm3) Raw SKL 35.72 3.53 0.32 AIKL 59.97 3.60 0.51 ASKL 13.66 2.17 0.53 AIKL >10kDa 73.03 4.95 0.23 ASKL >10kDa 26.23 3.24 0.30 AIKL 3-10kDa 8.83 1.90 0.51 ASKL 3-10kDa 7.13 1.70 0.58 AIKL <3kDa 8.95 1.77 0.65 ASKL <3kDa 3.42 1.22 0.75  Appendix E  PS and lignin intrinsic viscosity  Figure E.1 compares the intrinsic viscosity of AIKL, ASKL, and polystyrene. As can be found, at the same intrinsic viscosity, the Mw of lignin is much higher than the PS standards. It explains the quite large difference between relative Mw and absolute Mw for the lignin samples when PS is used as a calibration standard.    144   Figure E.1 Mark Houwink plot of AIKL, ASKl and polystyrene standards  Appendix F  Molecular weight by AF4 and Light Scattering F.1 Introduction  AF4 is a technology that fractionates samples based on the hydrodynamic radius, the same as GPC. The difference in techniques is that smaller species elute first, and larger fractions elute later. Very large particles are retained in the channel until the end of elution. AF4 is reported to perform better than GPC in branched polymer characterization as it minimizes the interactions between the analyte and system55, 109. Therefore, this method was adopted towards lignin analysis in order to provide an alternative to the possible interactions of lignin with the column gel.  F.2 Method   The AF4-MALS experiments were carried out with a short channel Eclipse AF4 for organic solvent coupled with MALS and RI detectors. Regenerated cellulose (5kDa) was used as the separation membrane. The length of the channel was 30cm and the spacer height was 350μm with THF as the eluent.  The detector flow was set to a constant rate of 0.7ml/min through the  145  entire procedure. A relaxation step was performed from 0-7 min under a focus flow of 2ml/min, during which the sample was injected and focused. Next, the system was switched to an elution mode, where a cross flow gradient of 3 to 0.1ml/min for 20 minutes was applied. The materials were maintained close to the membrane under cross flow, while separated based on their size and diffusion rate. After the elution step, the channel was flushed for another 8 min to remove all of the retained materials. Figure F.1 shows the entire processing procedures.   Figure F.1 AF4 processing flow rate F.3 Result and discussion  As the membrane cutoff was 5kDa, only AIKL/ASKL >10kDa fractions were tested to avoid potential inaccuracy due to penetration across the membrane. As shown in Figure F.2, Mw rose with elution volume as expected, indicating a successful fractionation with the AF4 system. The Mw of AIKL>10kDa and ASKL>10kDa was 67.15kD and 37.15kDa respectively. Both fractions had a similar Mw distribution from 104 to 3×105, whereby AIKL had a much larger PDI.  Compared with the GPC system in chapter 4, the Mw of AIKL was around half of the value, whereby that of ASKL was 20% higher. This result is in contrast to theory as Mw  146  measured by light scattering should be the same regardless of the fractionation techniques implemented if the fractionation method does not cause artifacts. The significant lower value of AIKL and slightly higher ASKL value may be explained by two reasons. First, in AF4, large particle aggregates may be retained on the membrane and flushed out at the end of the method. In terms of the 20% difference for ASKL results, it may result from ineffective fractionation leading to a mixture of high Mw and low Mw particles, in which case, low Mw lignin might be overshadowed. Table F.1 Lignin Molecular weight from AF4 and Light Scattering  Mn Mw PDI AIKL>10kDa 34.25 67.15 1.98 ASKL>10kDa 30.80 37.15 1.22    Figure F.2 Light Scattering profile and molecular weight of AIKL>10kDa (a) and ASKL>10kDa (b)   147   Figure F.3 RI profile of Polystyrene 30kDa in AF4 system   Appendix G  NMR result   G.1 NMR spectral HSQC of the aliphatic region and aromatic regions of SKL fractions were plotted in Figure G.1 and Figure G.2. 31P spectra of the lignin samples were plotted in Figure G.3 and 13C NMR were plotted in Figure G.4. Figure G.5, Figure G.6, and Figure G.7 show the HSQC, 31P NMR, and 13C NMR spectra of EMAL fractions, respectively. The summarized NMR data is given in Table G.1 Table G.1 Information and structural characterization of kraft lignin fractions as obtained by HSQC, 13C NMR, 31P NMR  Raw AIKL ASKL AIKL >10kDa ASKL >10kDa AIKL 3-10kDa ASKL 3-10kDa AIKL <3kDa ASKL <3kDa Interunit Bonds Abundance (per 100 Ar) Arylglycerol-β-aryl ethers (β-O-4′)b 7.74 14.01 4.23 17.48 4.37 6.9 3.89 6.62 1.67 Phenyl coumaran (β-5′) b 2.25 4.38 1.52 6.94 2.91 2.38 1.73 1.86 1.16 Pinoresinols (β–β′)b 3.10 4.72 2.94 7.41 4.94 5.38 4.35 4.28 4.16 Stilbeneb 2.21 0.85 1.02 0.1 0.83 2.53 2.68 1.27 6.31  β-O-4′ + β-β  + β-5a 11.63 17.87 5.78 14.07 6.71 11.31 4.71 8.1 4.24  148  Alkyl-O-alkyl a 37.75 64.23 17.49 28.77 10.26 30.18 5.20 26.76 16.18 Alk-O-Ar, α-O-Alk a 14.45 25.57 6.65 17.17 9.11 13.10 5.11 10.44 5.81 DC a 42.59 48.64 36.87 53.92 58.37 37.500 40.340 40.740 37.680  Functional Groups Abundance (per 100 Ar) G a 85.47 83.41 88.75 84.58 83.90 79.64 81.59 81.81 81.61 H a 5.57 2.35 1.43 6.48 2.17 3.09 3.75 4.95 - Methoxy (OMe) a 74.55 81.46 74.26 78.98 74.20 75.97 76.26 75.15 82.70 Aromatic C–H a 242.88 234.77 251.88 230.66 225.53 242.14 241.25 241.07 243.93 Aliphatic COOR a 4.17 1.96 5.82 6.17 10.92 12.53 13.38 4.70 4.82 conjugated COORa - - 0.88 - 1.69 3.34 2.56 0.55 0.14 Primary OH a 13.39 25.08 8.54 13.04 6.5 12.01 2.84 11.33 1.27  Functional Groups Amount (mmol/g) Aliphatic OH c 2.32 3.15 1.72 3.61 1.85 2.16 1.48 1.95 1.53 C5 substituted phenol c 1.66 1.79 1.87 1.53 1.84 1.29 1.36 1.24 1.37 Phenols c 3.53 3.76 4.89 3.77 4.83 3.92 4.55 3.49 4.28 p-hydroxy phenyl c 0.25 0.26 0.32 0.31 0.40 0.29 0.32 0.25 0.35 Ali / Phenol c 0.66 0.84 0.35 0.96 0.38 0.55 0.33 0.56 0.36 Total OH c 6.11 7.16 6.93 7.70 7.08 6.38 6.35 5.68 6.16 COOH c 0.45 0.46 0.46 0.59 0.87 0.99 1.27 0.83 0.96  Other Components Abundance (per 100 Ar) Sugar (X2) 1.57 2.72 - 4.64 - - 0.47 - - Sugar (X3) 0.67 1.51 - 3.29 - - 0.16 - - Sugar (X4) 0.46 1.13 - 2.93 - - - - - Quinones b - - 0.62 - 0.37 0.92 0.91 0.65 4.51 p-coumarate (PCA)b - - 0.07 - - - - - 0.54 5-hydroxymethyl- furfuralb 1.05 0.53 0.55 0.51 0.99 1.27 1.67 1.49 2.15 a from HSQC, b from 13C,  c from 31P.       149     (a) (b) (c)  150     (d) (e) (f)  151    Figure G.1 2D HSQC Spectrum of lignin fractions – aliphatic region; [a] AIKL, [b] ASKL, [c] AIKL>10kDa, [d] ASKL>10kDa, [e] AIKL 3-10kDa, [f] ASKL 3-10kDa, [g] AIKL <3kDa, [h] ASKL<3kDa       (g) (h)  152   7   (a) (b) (c)  153     (d) (e) (f)  154    Figure G.2 2D HSQC Spectrum of lignin fractions – aromatic region; [a] AIKL, [b] ASKL, [c] AIKL>10kDa, [d] ASKL>10kDa, [e] AIKL 3-10kDa, [f] ASKL 3-10kDa, [g] AIKL <3kDa, [h] ASKL<3kDa  (g) (h)  155      (a) (b) (c) (d)  156      (e) (f) (g) (h)  157   Figure G.3 31P NMR Spectrum of lignin fractions [a] Raw lignin SKL [b] AIKL, [c] ASKL, [d] AIKL>10kDa, [e] ASKL>10kDa, [f] AIKL 3-10kDa, [g] ASKL 3-10kDa, [h] AIKL <3kDa, [i] ASKL<3kDa     (i) PPM PPM PPM  158      PPM PPM PPM PPM  159    Figure G.4 13C NMR of lignin fractions [a] Raw lignin SKL [b] AIKL, [c] ASKL, [d] AIKL>10kDa, [e] ASKL>10kDa, [f] AIKL 3-10kDa, [g] ASKL 3-10kDa, [h] AIKL <3kDa, [i] ASKL<3kDa      PPM PPM  160    Figure G.5 2D NMR of EMWL; [a] AI-EMWL aliphatic region, [b] AS-EMWL aliphatic region, [c] AI-EMWL aromatic region, [d] AS-EMWL aromatic region     PPM  161   Figure G.6 13C NMR of EMWL; [a] AI-EMWL, [b] AS-EMWL    Figure G.7 31P NMR of EMWL; [a] AI-EMWL, [b] AS-EMWL  G.2 MW and Number of Native linkages  The relationship between MW and number of native linkages of AIKL and ASKL is given in Figure G.8Error! Reference source not found.. However, the line of the best fit was insufficient to investigate the relationship between MW and lignin structure and the line should PPM PPM PPM  162  be used to guide the reader. Overall, lower MW lignin samples were more degraded during delignification.   Figure G.8 Relationship between total native linkages per 100 aromatics and degree of polymerization for [a] AIKL and [b] ASKL   (b) (a) 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0379859/manifest

Comment

Related Items