technique on the graphitization of a biomaterial, lignin, where methods were employed to improve the interactions between the lignin matrix and the graphene oxide additives. The results from the above-described work in the phenolic resin and biomaterial prompted the need for an in-depth understanding of the templating mechanisms and their contributing factor to graphitization. In this regard, the thesis then scrutinizes the predominant force and operative mechanism driving graphitization-physical templating versus chemical templating. Finally, the thesis assesses the influence of the properties of modified hard carbons in energy storage applications and provides strategies for performance improvement. Collectively, the important contribution of this thesis is centered on the development of 2D nanomaterial templating in inducing graphitization of hard carbons (which requires no purification process for the resultant graphitic carbon), understanding the templating mechanisms interplay in modifying and tailoring crystalline properties in hard carbons and lastly, highlighting the electrochemical performance of the modified hard carbons in energy storage applications.

v

Table of Contents .

Abbreviation Meaning 2D Two-dimensional Å Angstroms AGP Autogenic Pressure APP Atmospheric Pressure ATR Attenuated Total Reflectance BET Brunauer-Emmett-Teller BJH Barrett-Joyner-Halenda BSU Basic Structural Unit ℃ Celsius CB Carbon Black C-C Carbon-Carbon CE Coulombic Efficiency CNS Carbon Nanosheets CNTs Carbon Nanotubes CT Coal Tar CTP Coal Tar Pitch Cu Copper d(002) Interplanar Spacing D-band Disorder band DRIFTs Diffused Reflectance Infrared Fourier Transform FCE First Cycle Efficiency FTIR Fourier Transform Infrared Spectroscopy FWHM Full Width at Half Maximum G Graphene G-band Graphitic band GO Graphene Oxide GPL Graphene Nanoplatelet hBN Hexagonal Boron Nitride HCL Hydrochloric Acid viii HRTEM High Resolution Transmission Electron Microscopy HTT High Treatment Temperature La Crystallite Size Lc Crystallite Height Li Lithium LDCMs Lignin Derived Carbon Materials LiB Lithium-ion Battery mAh Milliamp hour MD Molecular Dynamics Mg Magnesium MPa Mega Pascal MWCNTs Multi-wall Carbon Nanotubes N Novolac NG Non-graphitizing carbon nm Nanometers NMP 1-methyl-2-pyrrolidinone O Oxygen OH Hydroxyl Group PAH Polyaromatic Hydrocarbon PAN Polyacrylonitrile PLM Polarized Light Microscopy PR Phenolic Resin PVC Polyvinyl Chloride PVDF Polyvinylidene Fluoride ReaxFF Reactive Force Field RGO Reduced graphene oxide RSFs Relative Sensitivity Factor SA Surface Area SAED Selected Area Electron Diffraction SEI Solid Electrolyte Interface SEM Scanning Electron Microscopy ix TEM Transmission Electron Microscopy TGA Thermogravimetric Analysis wt.% Weight Percent XPS X-ray Photoelectron Spectroscopy XRD X-ray Diffraction x Figure 2.1. Thermogravimetric Analysis (TGA) of (a) carbonized systems under atmospheric conditions (b) carbonized systems under autogenic conditions; derivative %weight curves of (c) samples carbonized under atmospheric conditions (d) samples carbonized under autogenic conditions; derivative %weight curves of (c) samples carbonized under atmospheric conditions (d) samples carbonized under autogenic conditions................. 18 Figure 2.2 X-ray diffraction (XRD) patterns of the three-model systems showing a progression of crystallinity with temperature (a) APP_resole (b) AGP_resole (c) APP_resole+PVC (d) AGP_resole+PVC (e) APP_PVC and (f) AGP_PVC. ................................................. 21 Figure 2.3 Lattice parameters extracted from deconvoluted XRD spectra across the temperature series. d(002) plot of (a) APP_resole (b)AGP_resole (c)APP_PVC (d) AGP_PVC; (e) Lc plot of AGP systems; (f) d(002) plot of AGP_resole+PVC. ........................................ 23 Figure 2.4 Raman Spectroscopy analysis of (a) APP carbonized systems graphitized at 1500℃ (b) AGP carbonized systems graphitized at 2500℃ (c) APP carbonized systems graphitized at 2500℃ (d) AGP carbonized systems graphitized at 2500℃; Raman analysis of (e) crystallite length (La) and (f) ID/IG ratio comparison of APP and AGP samples at 2500℃. 26 Figure 2.5 Morphology study of model systems graphitized at 2500℃ using Scanning Electron Microscopy (SEM). (a) APP_resole (b) AGP_resole (c) APP_ resole+PVC (d) AGP_resole+PVC (e)APP_PVC and (f) AGP_PVC. .................................................. 29 Figure 2.6 TEM and SAED images of APP_PVC (a,b) and AGP_PVC (c,d)...................... 31 Figure 2.7 TEM and SAED images of APP_resole (a,b) and AGP_resole (c,d) .................. 32 Figure 2.8 TEM and SAED images of APP_resole+PVC (a,b) and AGP_resole+PVC (c,d) 34 Figure 3.4 Transmission Electron Microscope (TEM) of RGO2-N (a, b, c), GO1-N (d, e, f), and GO2-N (g, h, i). ....................................................................................................... 49 xi Figure 3.5 Selected area electron diffraction (SAED) pattern of (a) Novolac (N) (b) GPL-N, (c) RGO1-N, (d) RGO2-N, (e) GO1-N, (f) and GO2-N. ................................................. 51 Figure 3.6Lattice parameters extracted from deconvoluted XRD spectra of novolac (N), GPL-N (5.3at. % O), RGO1-N (14.4at. % O), RGO2-N (15.4at. % O), GO1-N (30.8at. % O) and GO2-N (35.3at. % O) vs. %oxygen content of the additives (a) d (002) versus %oxygen content (b) La versus %oxygen content (c) Lc versus %oxygen content. ..................... 54 Figure 3.7(a) Basic novolac polymerization mechanism79 (b) FTIR showing reduction of -OH and -C-O groups due to addition of additives. ............................................................ 57 Figure 4.5 Transmission electron microscopy (TEM) and selected area diffraction patterns of (a, b) lignin (c, d) lignin-rGO and (e, f) lignin-GO composites. ....................................... 77 Figure 5.1 SEM images of (a) G1, (b) G2 and (c) hBN. .................................................... 86 Figure 5.2 XPS spectra of (a) G1, (b) G2 and (c) hBN. ..................................................... 88 Figure 5.3 TEM images of (a) G1, (b) G2 and (c) hBN. .................................................... 90 Figure 5.4 TGA weight loss curve of (a) G-N and hBN-N and (b) carbonized G-N and hBN-N; XRD spectra of the graphitized samples at the varied weight percent of additives (c) G-N and (d) hBN-N; Raman spectra of the graphitized samples at the varied weight percent of additives (e) G-N and (f) hBN-N. ............................................................................. 93 Figure 5.5 XRD lattice parameters at various weight percentages, 0.5 wt%, 2.5 wt% and 5 wt%: interplanar d(002) spacing of (a) G-N and (b) hBN-N; Crystallite size (La) of (c) G-N and (d) hBN-N; Crystallite height (Lc) of (e) G-N and (f) hBN. ........................................ 95 Figure 5.6 SEM images of (a) 2.5 wt% G-N, (b) 5 wt% G-N, (c) 2.5 wt% hBN-N and (d) 5 wt% hBN-N. ................................................................................................................... 97 xii Figure 5.7 TEM images and SAED patterns of (a, b) 2.5 wt% G-N, (c, d) 5 wt% G-N, (e, f) 2.5 wt% hBN-N and (g, h) 5 wt% hBN-N. ...................................................................... charge capacity of at different C-rates and (b) Coulombic efficiency (CE) at 0.1C in 20 cycles; (c) Initial voltage charge/discharge profile of novolac at 0.1C (also representative of GPL-N and RGO-N initial cycles). xiii List of Tables

Table 2.1 Raman parameters for 2500C APP and AGP samples ........................................ Table 3.1 Summary of additive oxygen content measured by XPS..................................... Table 4.1 Chemical composition of GO and RGO additives measured by XPS ................... Table 4.2 Raman analysis of lignin, lignin-GO and lignin-RGO ........................................ Table 5.1 Table of elemental composition of hBN, G1, and G2 ......................................... Table 5.2 XRD parameters comparison of physical and chemical templates. All additives are at 2.5 wt.%. ................................................................................................................ Table 6.1 Summary of capacity and first cycle efficiency of hard carbon anodes derived from biomass and polymers. ............................................................................................. Table 6.2 First cycle efficiency and specific capacity at 0.1C for novolac samples. ............. Table 6.3 Capacity with first-cycle efficiency .................................................................. xiv

First, I would like to thank my advisor, Dr. Randy Vander Wal, for his support and guidance; I very much appreciate all the opportunities and funding provided for me to succeed and make it to the end of my Ph.D. journey.

I must also thank my committee members for their insightful suggestions, support, and advice in past years. Special thanks to Dr. Ramakrishnan Rajagopalan for hands-on tutelage and guidance on electrochemistry and battery work. Gratitude to all collaborators and sponsors of my research.

I would like to thank my lab mates, Akshay Gharpure, James Heim, Mpila Nkiawete for making a conducive work environment and all your help in lab set-ups. The journey would not have been as fun without you guys around. Gratitude also extends to the post doc, Dr. Bindu and junior lab mates.

I must also not forget to thank the wonderful staff of the Energy Institute (Brad and Ronnie)

and the materials characterization lab who always lent a helping hand in accomplishing tasks and created an open atmosphere for conducting research.

To all the wonderful friends I met here at Penn State, I am grateful for your encouragement and support throughout this journey. Similarly, I must also thank Dr. Anthony Olorunnisola for your mentorship, wisdom and guidance in critical times.

My mother, Azuka Florence Ike, and my siblings (Chisom, Chioma, Uche and George), deserve endless thanks and the deepest gratitude for their love, support and patience.

xv Funding Acknowledgement This material is based upon work supported by the National Science Foundation under Grant No. 2306042 and the American Chemical Society Petroleum Research Fund under ACS PRF Grant 59973-DN4. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the American Chemical Society.

As the world moves toward a greener energy future, the demand for graphite is increasing.

Graphite is a key material in lithium-ion batteries (LIBs) contributing more than 40% by weight in each battery 1,2 . There are many other applications of graphite such as lubricants, fuel cells, aluminum electrodes, brake linings, electronics, solar panels, electric trains etc. [3][4][5] . These applications are realized due to the remarkable properties of graphite including good electrical conductivity, softness, and high crystallinity 6 . The demand for graphite has earned it a spot on the U.S. critical mineral list as it is projected that there will be a short supply in the future 7 .

Currently, the major sources of graphite are from mining and petroleum products (cokes and pitches), which are unsustainable and have negative environmental consequences 6 . This has created the urgent need to find sustainable precursors and methods for making synthetic graphite or graphitic-like materials.

There are two main heat treatment stages by which carbonaceous precursors are transformed into graphite: an initial heat treatment referred to as carbonization and a second higher temperature process often referred to as "graphitization", though with the recognition that neither graphite nor graphitic carbons may result, depending upon the precursor. Carbonization is the process of increasing the carbon content by heating under inert atmosphere to decompose organic matter and remove heteroatoms and volatiles 8,9 . During carbonization, carbon precursors are heated to between 500 °C and 1000°C and may go through a mesophase stage (ultimately yielding a so-called "soft carbon") or char in place (ultimately yielding a so-called "hard carbon") 6 . After carbonization, the sample is then passed through a graphitization heat treatment above ~2,500°C indicate a highly ordered material (graphite). The degree of graphitization (if any at all) depends on the carbon precursor in use.

Franklin was first to propose models for graphitizing and non-graphitizing carbons and identified their differences using the schematic shown in Figure 1.1 10 . Franklin's work stated that non-graphitizing carbons cannot be converted into graphite even after high temperature treatment (HTT) because of the presence of crosslinks, internal hydrogen, and sample viscosity which prevents initial stacking (or rearrangement) during carbonization 10,11 . This then results in randomly oriented crystallites in the resulting carbon structure as seen in Figure 1.1a. Graphitizing carbons on the other hand will form layered parallel and well-ordered crystallites after HTT as seen in Figure 1.1b.

'polymeric carbons' from work done on a phenol-hexamine resin (Figure 1.1c). The model of non-graphitizing carbon was described as graphite ribbons that twist and bend in a random way 6 .

The graphite layers are depicted as branches that allow formation of three covalently connected regions aligned in different directions to form a 'strong confluence'. 'Weak confluences' are described as the non-covalent interactions between ribbon layers similar to those that hold graphite layers together 6 .

Oberlin objected to the model proposed by Jenkins stating questionable interpretation of electron micrographs. This was because only a part of the structure was being imaged (the resolved {002} fringes which are visible when parallel to the electron beams). Therefore, the structure appearance may be misleading and could be more cage-like rather than the ribbon-like structure being proposed by Jenkins 12 . Oberlin's 12,13 work identified graphitic carbons as being built from basic structural units (BSU) and classified stages of graphitization with the BSU developing in each stage (Figure 1.1d).

The BSU (see Figure 1.1d) in stage 1 develops distorted columns that pile up parallelly and independently in stage 2. Oberlin stated the columns in stage 2 are close to one another due to release of heteroatoms that occur around 800℃. Above 1500℃ (stage 3), the BSU disappears as columns do not remain independent and aromatic layers gets hooked edge to edge to form distorted wrinkled layers. Also, in this stage 3, 1500℃ -1900℃, turbostratic structure begins to disappear and three-dimensional order increases, Lc increases. Stage 4 begins above 2100℃ as inplane defects are removed leaving a well ordered and graphitic structure. As insightful and groundbreaking as Oberlin's work was, it is worth mentioning that this model can only be applied to graphitizable carbons and does not work for non-graphitizable carbons. Her study attributed the non-graphitizability of carbons from oxygen-rich precursors to very small volumes of preferred orientation of the BSUs that occur early in the carbonization process 13 .

Marsh also detailed structural changes that occur in the graphitization of mesophase (Figure 1.1e). It can be seen in the diagram that as temperature increases, the structure of the graphene layer becomes less defective 14 . Marsh attributed this structural change to single atom movement within the graphene layer. At temperatures between 1400℃ and 1700℃, wrinkles and bends can be seen in the graphene layer (which is like Oberlin's model). However, Marsh states that there is still little presence of in-plane defects even after 2000℃ as this are never completely annealed out 14 . This is in opposition to Oberlin model which states in-plane defects are completely removed after 2100℃.

Marsh and Rodriguez-Reinoso opposed using a graphitic or turbostratic crystallites model for non-graphitizable carbon. This is because non-graphitizable carbons (hard carbons) are unable to migrate and form graphene-like sheet layers at temperatures below 900℃ 15 . Marsh et al., 16 described non-graphitizable microporous carbon (hard carbon) as being defective micro-graphene layers primarily composed of six-membered ring systems with stress and strain accommodated by other sized ring systems. Using the study done by Marsh and Rodriguez-Reinoso, the stages of formation of hard carbon when heated to HTT was purposed as shown in 16 In this illustration (Figure 1.2) derived from Marsh et al., it is shown that the graphenic sheet are crumpled, wrinkled, and disordered even after HTT. This is because the final product of hard carbons retains its precursor's structure due to crosslinking, high oxygenation, and inability to go through a fluid stage during carbonization. Therefore, the graphenic sheet of hard carbon cannot be unfolded or flattened to increase stacking of the lamellae 16,18 . It is also worth noting that the process of graphitization is polluting and very energy intensive. For example, there are releases of volatiles and hazardous gases during the carbonization stage 16,19 . These are greenhouse gases and are the major causes of global warming and environmental pollution 20 .

There has been much research interest to find suitable methods by which to graphitize hard carbons. The motivation lies in their abundance, low cost, high carbon content, sustainability, and unique electronic and structural properties 3,20 . For decades, researchers have explored processes to graphitize hard carbons such as phenolic and furan resins, and even biomaterials like lignin and sugar. Phenolic and furan resins are thermoset resins well known for their unique properties such as thermal stability and mechanical strength. These properties make them useful for various industry and commercial applications 21,22 . For example, phenolic resins (PR) reinforced with fillers have been molded into heat resistant objects such as electrical connectors, appliance handles and used as coatings 23 . There are two types of PR: novolac and resole. Resoles are prepared under basic conditions with an excess molar ratio of formaldehyde to phenol while novolac are prepared under acidic conditions with a less than an equimolar ratio of formaldehyde to phenol and require a curing agent (hexamethylenetetramine) to crosslink the novolac 24 .

Other studies have shown that adding catalysts to phenolic resins can accelerate their graphitization. Ren et al., 25 used a Ni-Zn-B alloy to modify phenolic resin. It was found that the catalyst improved graphitization by approximately 30% at temperatures lower than 1400℃, however, as temperatures increased beyond 1400℃, the catalytic ability of the Ni-Zn-B alloy decreases due to its severe sintering. Rastegar et al., 26 used Ni(NO3)2 as a catalyst to modify phenolic resin and study micro morphology and graphite crystallite during carbonization. It was found that when the pyrolysis temperature gradually increases, the micro-morphology of phenolic resin carbon gradually changes from porous carbon nanotubes to carbon cells and onion peel carbon. It should be noted that the residual catalysts from this process are usually detrimental to most applications of the graphite 3,22 . It is therefore crucial to find a more desirable and less harmful additive or catalyst to promote graphitization of hard carbons.

In recent years, there has been research interest in using nanomaterials such as graphene and carbon nanotubes (CNTs) as additives (or catalysts) for the graphitization of hard carbons. It has been proposed that the aromatic ring base of the hard carbon precursor will have a 𝜋 -𝜋 interaction with the sp 2 backbone of the nanomaterials causing a rearrangement or realignment of the disordered hard carbon to graphite during HTT 3 . This process is known as templating. Chen et al. 27 used multi-walled carbon nanotubes to modify furan resin of carbon/carbon composite matrix. Analysis by Raman and PLM showed induced graphitization of the furan resin the areas around the CNTs after HTT. This was attributed to the π-π conjugated bonds between sp 2 hybrid multi-walled carbon nanotubes and resin carbon surface. Yi Shoujan et al., 27 studied the effect of graphene oxide (GO) on the graphitization of furan resin. It was found that the graphitization degree of furan resin carbon increased with increasing the graphene oxide content up to 29 wt.% and heat treatment temperature up to 2400 ℃. However, Yi Shoujan et al., 27 stated the mechanism by with GO accelerated the graphitization of the furan resin was unknown. Liang et al., 28 used carbon black and CNTs as additives to improve graphitization degree of phenolic resins and found that the CNTs caused a high degree of graphitization in the phenolic resin matrix compared to the carbon black.

The templating method using nanomaterials has been applied to improving the graphitization of biomaterial like sugar. Madhu et al., 29 used GO and reduced graphene oxide (RGO) to catalyze the graphitization of sugar. It was shown that GO additive improved the graphitic quality of the sugar. This was attributed to radical sites of the GO reactively bonding to the decomposing sugar matrix during HTT. This process was called 'reactive templating.' The same effect was not observed for the reduced graphene oxide (RGO) due to the weak interaction between the extended aromatic π network of the RGO and a non-aromatic sp 3 precursor sugar.

These results show that templating can prove to be an effective method for the graphitization of hard carbons such as thermoset resins and renewable biomaterials. Biobased and renewable precursors such as lignin are being researched heavily as alternatives to petroleum precursors and mined graphite for use as energy storage and composite materials [30][31][32] . Exploring the effect of templating in the graphitization of biomaterials will be an interesting and insightful study for their potential applications. Finally, other external factors such as pressure and temperature can influence the graphitization of carbon precursors. Understanding how these factors play a role in the final graphitic quality of a carbon precursor is a key step to determining the right parameters and method suitable for graphitization.

The purpose of this work is to develop and provide in-depth understanding of the templating technique for graphitizing hard carbon (non-graphitizing) precursors. In addition, this work aims to explore and contribute knowledge on the impact of heat-treatment parameters (e.g., pressure, temperature) in the graphitization of carbon precursors. Finally, the application of modified hard carbons as anodes in lithium-ion batteries was investigated. Organized in seven chapters, the core of the dissertation centers on chapter two through six. In chapter two, a study on the effect of carbonization method on final quality of graphite is presented. The importance of this chapter lies in discerning whether pressure influences or improves graphitization of carbon precursors. Three model systems were used in this study -a non-graphitizing precursor (phenolic resin, resole to be specific), a graphitizing precursor (polyvinyl chloride, PVC) and a 50:50 blend of the two. Samples were carbonized at atmospheric pressure or under autogenic pressure conditions and then subjected to a series of higher HTTs while characterizing the graphitization degree at each temperature. Chapter three presents a study on the effect of two graphene oxide (GO) derived additives in the templating induced graphitization of novolac. The GO derived additives, selected based on their oxygen content, were added to the novolac matrix, and passed through several heat-treatment processes. The working hypothesis is based on the additives acting as templates that allow the novolac matrix to align during graphitization (physical templating) as well as form reactive radicals that provide further bonding to the decomposing matrix during HTT (chemical templating). Chapter four investigates graphitization of bio-derived material (alkali lignin) by templating method using GO additives. The templating concept from the previous chapter is applied in the graphitization of alkali lignin. The lignin-GO samples were passed through an initial carbonization heat treatment and subjected to a series of HTT to determine graphitization degree at each temperature. This chapter demonstrates templating as a technique for converting bio-derived material to graphitic carbon. Chapter 5 distinguishes the templating mechanismschemical vs. physical templating -and their individual contributing effect on the graphitization of the novolac matrix. The use of 2D additives in the form of graphene and hexagonal boron nitride (hBN) with zero to minimal oxygen functionalities provided the means to understand how the extensive sp² network of these materials promote matrix ordering through templating. All heattreatment and characterizations conditions were kept constant for effective comparisons with results in chapter 3. Chapter 6 then tests modified hard carbons as anodes in lithium-ion batteries. This chapter analyzes the influence of the hard carbon properties on electrochemical performance and provides strategies to improve this performance. Lastly, chapter 7 presents the dissertation conclusions and future work recommendations.

The content of this chapter is adapted from a paper published in Carbon Trends and referenced as:

Ike, S., & Vander Wal, R. (2024). Effect of carbonization methods on graphitization of soft and hard carbons. Carbon Trends, 16, 100382.

Synthetic graphite dominates over 70% of the battery anode market, yet its energyintensive processing results in substantial carbon emissions. Researchers are exploring novel synthetic routes to reduce this environmental impact. This study investigates the influence of carbonization methods on the graphitization behavior of soft and hard carbons. Using a threemodel system phenolic resole (hard carbon), polyvinyl chloride (PVC) (soft carbon), and a 50:50 blend of resole and PVC. Carbonization was conducted under autogenic pressure (AGP) and atmospheric pressure (APP) at 500℃ for 5 hours, followed by high-temperature treatment at varying temperatures. By using various techniques, including X-ray diffraction, Raman spectroscopy, transmission electron microscopy, and selected area diffraction, non-graphitizing precursors exhibited improved properties under autogenic pressurized carbonization, such as larger crystallite size, sharp crystalline peaks, lower ID/IG ratio, and narrow G-full width half-maximum an indication of graphitic like material by lowering amorphous phase at high temperature. For graphitizing or soft carbon precursors, the method of carbonization did not impact the graphitization level. The most significant finding was the enhanced graphitic nature observed in hard carbon under AGP conditions, without the need for any catalyst. These materials hold promise for battery applications due to their desirable characteristics.

Carbonization is the process of concentrating and purifying carbon by denaturing organic matter with heat in the presence of little to no oxygen 33 . Carbonization can be classified into two regimes: atmospheric and pressurized. The atmospheric method involves heating the sample in an open boat or vessel placed in a furnace. The sample is typically heated at a slow heating rate (0.1℃/m to 2℃/min) to temperatures between 500℃ -1000℃, and a long residence time ranging from 0.5 hours to 24 hours or more 33,34 . During carbonization, various reactions such as deoxygenation, dehydrogenation, and dealkylation occur as the carbon precursor decomposes resulting in the evolution of hydrocarbons (short aliphatic chains) and volatile gases such as CO and CO2 35,36 . The release of these hydrocarbons and volatile gases leads to a lower carbon yield content of the carbonized product 37,38 . Carbonization performed under pressure is an attempt to increase the carbon yield in comparison to atmospheric carbonization [39][40][41] . Under pressure, the evolution of hydrocarbons and volatiles is suppressed which may lead to an increase in the carbonization yield 37,42 .

Pressurized carbonization is typically carried out in different ways: (1) carbonization under pressure built up by the decomposition gases of the precursor (autogenic) which stands in contrast to (2) carbonization under constant pressure, and (3) carbonization under hydrothermal conditions 37 . Carbonization under autogenic pressure is done by heating the carbon precursor from ambient to carbonization temperature with gradual pressure increase from decomposition gas built up in a closed vessel. Under autogenic conditions, pressure cannot be kept constant; it is strongly dependent on the temperature and the amount of sample used 37 . Conversely, pressure can be kept constant by using an autoclave during carbonization heat treatment.

Hydrothermal carbonization can be done at temperatures either above 400°C, below 250°C, or within these limits. The pressure is built up by water vapor and depends on the amount of water and volume of autoclave used 37,43 . A consequence of carbonizing under pressure is that it may change the carbonization process (gas solubility, mesophase viscosity, carbonaceous intermediates, carbon yield, etc.) which can in turn lead to the resulting carbon being different in structure, property, and even particle morphology from that obtained without pressure 37,44 . This makes pressurized carbonization a possible method for controlling the structure and texture of resultant carbons. Michio Inagaki et al., 37 stated that the principal purposes for carbonization under pressure include modification of carbonization behavior, improving the carbon yield, densifying the resultant carbon, changing graphitization behavior, or obtaining specific particle morphology of the resultant carbons.

There have been many studies on the effect of carbonization methods on the carbon content of precursors. Hosomura et al., 45 studied the effect of pressure carbonization on carbon-carbon (C-C) composites and conventional materials such as phenolic resins and pitches. They found that pressure carbonization does not affect the carbon yield of the hard carbons (phenolic resin and furfuryl alcohol); the carbon yield obtained was 55% irrespective of pressure. However, there was a significant increase in the carbon yield for pitches (of low molecular weight) carbonized under pressure. Texture changes were also observed wherein the pressure carbonization increased the proportion of fine mosaic relative to coarse texture as seen by polarized light microscopy (PLM).

Another study by Ayache et al., [12] also observed a change in the texture of the resultant carbon after the carbonization of polyethylene and anthracene at 650°C under a pressure of 30 MPa.

Characterizations from scanning electron microscopy (SEM), transmission electron microscopy (TEM), and optical microscopy showed that both materials had radial texture phases after pressure carbonization. Inagki et al., 46 carbonized coal tar pitches at 650°C under a pressure of 30 MPa in closed and open gold tubes. The carbon yield of the original pitch was 96% and 51% under the closed and open systems respectively. The low carbon yield in the open system was attributed to hydrocarbons and volatiles lost during carbonization whereas in the closed system, the decomposition gases react with the mesophase, increasing the high carbon content to greater than 90%.

Despite these findings, however, there are comparatively few studies of carbonization pressure effects on the graphitization behavior and final (graphitic quality) of the final carbon product. In one such study, Kamiya et al., 47 graphitized three polyvinyl alcohol (PVC) samples under normal pressure after preheating (at 1470℃, 1590℃, and 1620℃) under pressure of 5 kbar.

These authors found that the treatment of soft carbon under high pressure had no accelerating effect on the carbon graphitization in the subsequent heat treatment under normal pressure.

This study, however, does not specifically examine hard carbons which would be a useful contribution to understanding if pressure carbonization can influence graphitization behavior.

Okamoto et al., studied the effect of pressure on carbonization and subsequent graphitization on phenolic resins 48 . Phenolic resins were carbonized at 650°C under atmospheric or 100MPa pressures and iron oxide was used as a catalyst. Subsequent heat treatments were conducted at 1200°C, 1500°C, and 1900°C. The study measured carbon yield, density, and crystallinity as well as optical observations. Results indicated that pressurized carbonization of phenolic resin advanced graphite crystallization accelerated by iron oxide powder. This study's focus was on the catalyst effect and the sole effect of pressure during carbonization was not discussed. This therefore highlights the need for a rigorous fundamental study on the effect of pressure on nongraphitizing carbons. Moreover, understanding the graphitization of hard carbons is on the uprise and has become a hot topic in the scientific community [49][50][51] and industry as the precursor to synthetic graphite due to its abundance and environmental friendliness.

In this work, we selected two precursors non-graphitizing (hard carbon, resole), graphitization (soft carbon, polyvinyl alcohol (PVC)), and their blend (50:50 by weight) to study the effect of carbonization methods on graphitization. During carbonization, we subjected the samples to atmospheric and autogenic pressure carbonization and subsequent high-temperature heat treatment. Detailed analysis from X-ray diffraction (XRD), Raman spectroscopy, transmission electron microscope (TEM), and selected area diffraction (SAED) showed a trend of improved graphitic quality in the form of larger crystallite size (La), sharp crystalline peak, lower ID/IG ratio, narrow G-full width half-maximum for the non-graphitizing precursors because of pressurized carbonization. For the graphitizing or soft carbon precursor pressure carbonization does not show any change in graphitization level. The improved graphitic nature in hard carbon under pressure without the addition of any catalyst is an important finding as they may be used for battery applications.

Polyvinyl chloride (PVC -molecular weight of 62,000) was used as received (Sigma Aldrich) without any further modification. The resin was a resole from Supelco (Plenko 14946 resole). Material preparations included the resole, PVC, and a blend (50/50 resole+PVC) of each.

Samples were placed in an oven at 100℃ for 24 hours to remove moisture.

The three samples (PVC, resole, and resole + PVC) were carbonized by two methods, atmospheric and autogenic pressure. Carbonization under atmospheric pressure (APP) was done in an open boat in a Thermolyne 2110 tube furnace while carbonization under autogenic pressure (AGP) was done in a customized pressurized reactor. The sample was wrapped in brass foil and inserted into the reactor. The reactor was then pressurized with nitrogen to purge all oxygen out and checked for any leaks. The pressure in the reactor typically ranges between 500 -1000 psi during the carbonization process. After carbonization, the reactor is allowed to cool down, and opened, and the carbonized sample is collected and ready for graphitization. Carbonization by both methods was done at 500°C for 5 hours.

The conventional and pressure carbonized samples (six samples in total) were graphitized at 1000°C, 1500°C, and 2500°C. The 1000°C graphitization was done in a Thermolyne 2110 tube furnace at 15°C/minute. The 1500°C was done in a GSL-1700X-UL furnace at 10°C /min temperature ramp up to 900°C and then 5°C/min to 1500°C. Graphitization was done at 2500°C in a Centorr Vacuum Industries series 45 graphitization furnace, heating at 15°C/min to 1000°C, 10°C/min to 2000°C and finally 5°C/min to 2500°C. All graphitization runs were performed under an inert atmosphere.

Thermogravimetric analysis (TGA) was done in a Q600 (TA Instruments, USA). Precarbonized samples were subjected to a temperature ramp test (30 -900 °C at 10 °C/min). The conventional and pressure carbonized samples were subjected to the same temperature ramp test, however, with an isothermal hold at 900 °C for 120 minutes to study the carbon content of each sample.

The X-ray diffraction patterns were collected using a Malvern PANalytical Empyrean diffractometer equipped with Cu source (λ ≅ 1.54A°), para-focusing optics and PIXcel 3D

detector. The spectrum was scanned in the 2θ range from 10° to 90°. The background subtraction, peak fitting and quantification were done using MDI JADE® software.

Raman spectra was collected using a Horiba LabRAM HR Evolution equipped with a 300 groove/mm grating and a 532nm laser. The spectra were acquired in DuoScan™ mode which increases the statistical significance of the data by rastering over a wider area. At least 5 measurements were collected for each sample to ensure the analysis was representative.

SEM images were taken with field-emission SEM: Apreo. Samples were prepared by placing a few milligrams on a carbon taped pin stub holder. To obtain FESEM images, an acceleration voltage of 7kV and a working distance between 11mm to 7mm was maintained.

Transmission electron microscope samples were prepared by sonicating a few milligrams (mg) of graphitized material in ethanol and then a droplet of the suspension was placed on a copper (Cu) supported lacey carbon grid and allowed to dry. The samples were imaged using a FEI TalosTM F200X scanning/transmission electron microscope equipped with an FEG source providing 0.12 nm resolution. The instrument was operated at 200 kV and the samples were imaged at various magnifications in the ranges. Selected area electron diffraction (SAED) patterns were taken concurrently with TEM imaging.

We first calculated the carbon yields for all samples (Figures 2.1a, b) for both atmospheric carbonization (APP) and autogenic pressure carbonization (AGP). APP_resole has a carbon yield of 72%, slightly lower but comparable to the carbon yield of 79% measured for AGP_resole.

APP_PVC has a carbon yield of 90% which is higher than the carbon yield measured for AGP_PVC (71%). The higher degree of decomposition in AGP_PVC may be due to trapped volatiles in the matrix that were unable to escape from the closed reactor during carbonization.

The same trend is observed in the mixture where the carbon content for APP_resole+PVC (88%)

is higher than AGP_resole+PVC (66%). Generally, there is a low degree of decomposition observed across most samples because a lot of volatiles have been lost during carbonization. The TGA derivative weights of the carbonized samples by both methods are plotted in Figures 2.1c and 2.1d. For samples carbonized under APP conditions (Figure 2.1c), APP_resole shows two decomposition peaks at approximately 515°C and 700°C. Similarly, APP_PVC has two decomposition peaks, one at 250°C and a broad peak from 595°C to 780°C, and the blend (APP_resole+PVC) has two very pronounced decomposition peaks that occur at the same temperature as resole. The decomposition peak at 250℃ in PVC is not seen in the blend. On the other hand, resole carbonized under autogenic condition (AGP_resole) shows three decomposition peaks at 75℃, 115℃, and 668℃ (Figure 2.1d). AGP_PVC has multiple decomposition peaks above 600°C while maintaining the broad peak from 595°C to 780°C observed in the derivative curve of APP_PVC. The blend, AGP_resole+PVC shows four decomposition peaks at 75°C, 300°C, 495°C and 562°C. For both carbonization methods, the rate with PVC and resole+PVC losses is much higher than pure resole resin. This is expected to lead to high porosity in PVC and resole+PVC as alluded to in previous work 52 . progression of crystallinity with temperature. Samples were heat treated at different temperatures: 1000℃, 1500℃, and 2500℃. The heat treatment at 500℃ is the carbonization temperature used for all samples. APP_resole (Figure 2.2a) has very broad peaks at 500℃, 1000℃, and 1500℃; this

There is no marked difference in the XRD spectra of the graphitization of resole carbonized under APP and AGP conditions. To summarize, resole is a non-graphitizable, hard carbon meaning it does not go through a mesophase fluid stage which allows for rearrangement and stacking of graphene sheets, rather it simply chars in place during carbonization. This is due to the formation of highly cross-linked and curved structures facilitated by the presence of oxygen functional groups. Even though there is an observed improved intensity/growth of the d(002) peak of APP and AGP resole graphitized at 2500℃, it is however broad and does not compare to the very sharp and intense d(002) peak of pure graphite [18-23]. The improved intensity of the d(002) peak of resole can be attributed to small crystallites that formed at 2500℃ but their formation does not signify long-range order. of the mixture even at an HTT of 2500℃. Although the d(002) peak intensity increases as

temperature increases, it is still very broad at 2500℃, which is not representative of graphite [18-23]. A similar trend is illustrated in the XRD spectra of the AGP_resole+PVC mixture (Figure

Even though the mixture exhibited the same characteristic trend in carbon yield as PVC (soft carbon), from this result, it appears that the hard carbon, resole, dominates the structure after heat treatment and impedes the graphitization of the mixture. system graphitized at different temperatures. d(002) is the typical primary measure for graphitization. As d(002) decreases, graphitic quality increases. In Figure 2.3a. the d(002) trend

for APP_resole is shown. As resole is heated and treated to higher temperatures, d(002) decreases, again signaling an increase in graphitic quality. Despite the monotonic decrease in d(002) with increasing HTT, the d(002) value at 2500℃ is 0.346nm, higher than that of pure graphite at 0.335nm. This is consistent with the non-graphitizing characteristic of resole. The same trend can be seen in AGP_resole graphitized across the temperature series in Figure 2.3b. At 2500℃, the d(002) for AGP_resole is 0.346nm. These results show that there is no significant difference between the graphitization of resole carbonized under atmospheric or autogenic conditions. value decreases as temperature increases, reaching a value of 0.337 nm at 2500℃, close to 0.335 nm for pure graphite [54][55][56][57][58][59] . PVC is a graphitizable carbon, and it passes through the mesophase stage during carbonization which allows for rearrangement and graphite sheet stacking at HTT.

The same trend can be seen in AGP_PVC in Figure 2.3d where at 2500℃, the d(002) is 0.339nm.

These results show that there is no significant difference between the graphitization of APP_PVC and AGP_PVC. The Lc values across the temperature series for systems carbonized under autogenic conditions are plotted in Figure 2.3e. Lc is the crystallite height and as Lc increases, graphitic quality increases. For all systems, there is an increase in Lc value at 2500℃, which signifies an increase in graphitic quality. The Lc value for the resole and the mixture (resole+PVC) is 3nm. Thus, the non-graphitizing precursor dominates the graphitic quality and properties of the mixture after HTT. Meanwhile, the Lc value for PVC (14nm) is significantly higher than that of resole and the mixture (resole+PVC). Raman spectroscopy is another technique used to characterize the graphitic quality of An equal intensity of the D and G bands can be observed in each sample, a hallmark of nongraphitic carbon. The D-band arises from a ring-breathing mode originating from disordered edges as a result of defects caused by the presence of sp 3 carbons dangling from the edges, holes in the lattice, etc. 13,33,60 . This suggests that 1500℃ is a temperature too low to remove defects, enable sheet stacking, and thus improve graphitic quality. The same trend can be seen in the Raman spectra for the AGP carbonized systems graphitized at 1500℃ in APP_resole and AGP_ resole after HTT at 2500℃ in Figure 2.4c and 2.4d, respectively. This is because resole is a non-graphitizable carbon, it does not go through a fluid mesophase stage during carbonization which allows for initial stacking and rearrangement of graphene sheets, it simply forms char. The spectra of APP_resole+PVC and AGP_resole+PVC in Figure 2.4c and 2.4d, respectively, also show the presence of a prominent D-band. This signifies that the mixture is nongraphitizable. Although PVC is a soft carbon, these results indicate that when mixed with resole (a hard carbon), the resultant mixture takes the structure and graphitization behavior of the hard carbon. The La values calculated from Raman were plotted (Figure 2.4e). Resole and resole+PVC have low La values unlike PVC (graphitizing/soft carbon). There is an improvement A more in-depth Raman analysis reveals distinctions between the APP and AGP samples, as illustrated in Table 2.1. The initial column presents the intensity ratio of the D to G band (ID/IG), a metric reflecting defect and disorder levels in samples. Decreasing ID/IG values correspond to heightened graphitic carbon content. APP_PVC and AGP_PVC exhibit comparable low ID/IG values of 0.11 and 0.15, respectively, indicative of low disorder and high graphitic quality. Conversely, the resole and resole+PVC mixture exhibit higher ID/IG ratios, suggesting increased defect levels and diminished graphitic quality. Notably, AGP_resole and AGP_resole+PVC display lower ID/IG ratios than their APP counterparts, indicating fewer defects in the AGP samples. The subsequent columns detail the G-peak position, and the G-full width half-maximum (G-FWHM) provide clear evidence of the disparity in graphitization or enhanced graphitic content between AGP and APP non-graphitic samples (resole and resole+PVC). The smaller, narrower G-FWHM signifies higher graphitic quality. AGP_resole and AGP_resole+PVC exhibit lower G-FWHM values than their APP counterparts, while APP_PVC and AGP_PVC show comparable G-FWHM values.

Further analysis of Raman parameters, including the ratio of the 2D to G band (I2D/IG), underscores improved graphitic layers in AGP samples, as they exhibit higher I2D/IG values than their APP counterparts. This trend is also observed in the full-width half-maximum (FWHM) ratio of the 2D to G band in the last column, with higher values for AGP samples compared to their APP counterparts (except for PVC, which shows comparable values). These findings suggest lower defects and enhance graphitic content in AGP samples, particularly in the non-graphitic samples, AGP_resole and AGP_resole+PVC. It appears that carbonization under pressure can influence the formation of larger crystallites and improve the overall structure.

Table 2.1 Raman parameters for 2500C APP and AGP samples Scanning electron microscopy (SEM) was used to study the morphology of heat-treated samples at 2500℃. Figures 2.5a and 2.5b show SEM micrographs of APP_resole and AGP_resole, respectively. There is a high density of small aggregates of carbons with no observation of orderliness in APP_resole. AGP_resole shows flakes and aggregates appearing in wavy-like morphology. This morphology is due to the non-graphitizing nature of phenolic resin which forms disordered structures even after HTT at 2500°C. Figures 2.5c and 2.5d show SEM micrographs of the APP_resole+PVC and AGP_resole+PVC, respectively. A lack of a layered structure is observed in both images. Samples appear as aggregated carbons and large flakes with no orientation or layers. This evidence supports results from XRD and Raman which show the blend to exhibit hard carbon properties, unable to graphitize even after HTT at 2500°C. Conversely, the SEM images of APP_PVC (Figures 2.5e) and AGP_PVC (Figures 2.5f) show layered structures of large graphite flakes indicative of high degree of orderliness. PVC is a graphitizing carbon that forms graphitic layers after HTT at 2500°C. AGP_PVC also has some sections of roughness arising from aggregates (non-homogeneity). This evidence supports results from Raman where a low D-band intensity is seen in the spectra. Samples I D /I G G-Peak position (cm-1) G FWHM I 2D /I G 2DFWHM/GFWHM APP_PVC 0.11 1585 22 0.36 2.47 AGP_PVC 0.15 1586 24 0.42 2.41 APP_Resole 1.44 1592 58 0.64 1.25 AGP_Resole 0.81 1593 48 0.83 1.49 APP_resole+PVC 1.55 1589 59 0.87 1.26 AGP_resole+PVC 1.1 1590 44 0.98 1.55 Transmission electron microscopy (TEM) was used to investigate the nanostructure distinctions between the graphitized AGP and APP samples. Corresponding selected area electron diffraction (SAED) patterns were collected to provide insights into their crystallinity, emphasizing subtleties that may arise from the differing carbonization methods. The nanostructure of APP_PVC (depicted in Figure 2.6a) reveals sheet stacking without wrinkles or curvature, indicative of a graphitizing carbon. The accompanying SAED pattern (Figure 2.6b) forms a polycrystalline ring, representing periodic crystallites along grain boundaries. This nanostructure and the supporting SAED pattern align with XRD and Raman results, confirming that PVC manifests as a graphitic carbon at 2500°C. Similarly, AGP_PVC (Figure 2.6c) exhibits sheet stacking and the presence of lattice fringes, signifying long-range order in the material. The corresponding SAED pattern (Figure 2.6d) is also polycrystalline with a sharp and defined crystallite orientation like APP_PVC. In contrast to the graphitic and well-ordered nanostructure observed in PVC, APP_resole (depicted in Figure 2.7a) unveils a non-graphitic, fullerenic curved nanostructure. The corresponding selected area electron diffraction (SAED) pattern (Figure 2.7b) appears amorphous, characterized by halo rings, thus confirming the non-graphitic nature of resole -an The characteristic of a non-graphitizing nanostructure persists in the resole+PVC mixture.

In APP_resole+PVC (Figure 2.8a), a multitude of curved lamellae is evident, and the corresponding selected area electron diffraction (SAED) pattern (Figure 2.8b) showcases amorphous halo rings. Similarly, the AGP mixture exhibits a comparable nanostructure to its APP counterpart (Figure 2.8c), albeit with instances of short-stack lamellae highlighted in the red box. This is further elucidated by the SAED pattern (Figure 2.8d), which reveals a combination of amorphous and polycrystalline patterns-consistent with the trends observed in the materials.

These observations underscore the formation of larger crystallites in the AGP carbonized samples which was also confirmed by Raman analysis. Pressure carbonization improves the formation of larger crystallites in the non-graphitizing precursors and can therefore influence their various industry applications such as for energy storage where materials with larger crystallite size (La)

have been shown to lead to improved capacity. 2.5. Conclusion This work showed improved graphitization/graphitic quality in non-graphitizing precursors because of pressurized carbonization. XRD crystalline peak at 26.2° is much sharper in autogenic pressure carbonization for both resole and resole+PVC signifying more graphitic material of the This work is important because of the growing need to identify procedures for the graphitization of hard carbon precursors to synthetic graphite with the potential for energy storage applications.

The current work provides an interesting finding that shows metallic catalysts can be eliminated in graphitizing hard carbon when it is done under autogenic pressure. This may be a welcoming progress in the quest to find alternative graphite precursors for synthetic graphite manufacturing.

A follow-up study to this work by deliberately applying external pressure during carbonization may be of interest to determine the improved graphitic quality of carbons.

The content of this chapter is adapted from a paper published in Carbon Trends and referenced as:

Ike, S. N., & Vander Wal, R. (2024). Templating-induced graphitization of novolac using graphene oxide additives. Carbon Trends, 16, 100388.

Increasing graphite demand for energy storage applications creates the need to make graphite using precursors and processes that are affordable and friendly to the environment. Nongraphitizing precursors such as biomass or polymers are known for their low cost and sustainability; therefore, graphitizing them will be an accomplishment. In this work, a process of converting a non-graphitizing precursor, phenolic resin novolac (N), into a graphitic carbon is presented. This was achieved by the addition of five additives categorized as graphene oxide (GO)

and its derivatives with varied oxygen concentrations. The hypothesis is that the additives act as templates that promote matrix aromatic alignment to their basal planes during carbonization (physical templating) in addition to forming radical sites that bond to the decomposing matrix (chemical templating). Results showed that the addition of reduced graphene oxide (RGO)

additives of approximately 15.4 at.(%) oxygen content to the novolac matrix (RGO-N) show the best graphitic quality. In contrast, the addition of GO additive of twice or more oxygen content ≥ 30.8 at.(%) to the novolac matrix (GO-N) led to poor graphitic quality. This suggests that there is an optimum amount of oxygen content in GO additives needed to induce graphitization of the novolac matrix.

Graphitization is an energy intensive process. The Acheson process is still the preferred method for graphitization of carbon precursors. This process operates at temperatures above 2800°C and requires a long processing time leading to the high cost of synthetic graphite 65 . The process utilizes precursors such as petroleum cokes and tar pitches which are graphitizable carbon precursors [66][67][68][69][70][71] . However, these precursors are unsustainable and come from polluting sources.

With high demand for graphite in traditional applications such as electrodes in aluminum refining and electric arc furnaces steel industries and now with the forecasted exponential rise of electric vehicles 1,2,72 and associated lithium ion batteries, there is an urgent need to find cost-effective carbon precursors and alternative approaches for graphitization. For example, carbon precursors such as phenolic and furan resins and even biomaterials like lignin and sugar are of major interest because of the abundance, low cost, high carbon content, less pollution, sustainability of these precursors, and unique electronic and structural properties 3,18 . However, converting these nongraphitizable precursors into a highly ordered graphitic structure presents a challenge.

Accompanying this challenge is interest in reducing the energy and cost associated with graphitization by using lower temperatures.

Attempts to understand controlling factors on graphitization has stimulated different research works and many factors have been attributed to the cause of graphitization of nongraphitizable carbon matrices. In a study by Lanticse-Diaz et al., anisotropic structures were formed in the area around the carbon nanotubes (CNT) in a furan resin/CNT carbon-carbon (C-C) composite after heat treatment at 2800℃ 73 . It was suggested that interaction between the furan resin matrix and CNT (diameter of 100-300 nm) at the interface prevents the matrix from shrinking during graphitization resulting in a stress gradient beyond the interface leading to stress graphitization. Upon analysis, stress-induced graphitization assumes the following: First, wetting of the carbon nanotube by the matrix and, second, a perpendicular (stress-induced) force that propagates to some degree to cause matrix restructuring into a graphitic form. However, it is unclear how and why the matrix retains sufficient integrity to propagate such a stress during severe pyrolysis and decomposition. Originally the concept emerged from observations with continuous fiber composites wherein the extended fibers were considered to resist matrix contraction [Reznik et al]. However, for MWCNT with variable curvature, twists, and bends, it is unclear how such stress develops over straight-lengths of a few 10's of nanometers. The explanation seems to relate the internal change of a matrix, namely formation of graphitic structure, to an extrinsic physical constraint such as the boundary of carbon fibers or nanocarbon additives. Tzeng et al., 74 studied a CNT reinforced phenolic resin C-C composite and observed the formation of graphitic rods at 2000℃. Also, in a study done by Nam et al., acid treated CNT were dispersed into a polyaniline (PANI) matrix and heat treated at temperatures above 1500℃. They observed crystalline structures

in the composites heated to 1700℃. It is important to note that the temperatures at which graphitic structures were observed in these studies is far below the temperature reported for stress graphitization as proposed by Lanctice-Diaz et al 73 .

In related work, Saha et al., 75 explored the templating ability of nanomaterials, CNT and graphene, in the graphitization of polyacrylonitrile derived fibers (PAN) using reactive force field (ReaxFF) and molecular dynamics (MD). Depending on temperature, two mechanisms were proposed for graphitization: physisorption and chemisorption. The former, referred to here as "physical templating", occurred as temperature approached 2200K (~1900℃) where templating is governed by 𝜋 -𝜋 interactions between the 𝜋-conjugated system of the nanofillers and 𝜋electrons of the C≡N groups and/or all the carbon rings of the PAN fibers. The latter chemisorption, referred to here as "chemical templating", occurred as temperature approaches 3000K (~2700℃)

where functional groups such as C≡N on the carbon matrix become 'leaving groups' and form radical sites for covalent bonding with the nanofiller. Gao et al., 32 also used atomistic ReaxFF and large-scale MD simulations to explain the mechanism by which graphene could modify the microstructure in graphene reinforced carbon fibers. This study found that there are dangling bonds on the graphene edges, which form bonds with the polymer matrix, serving as "catalytic seeds"

for the formation of larger graphitic structure. Ma et al., 76 reported an increase in graphitization in a carbon/phenolic resin C-C composite reinforced with graphene oxide (GO). Their MD simulations showed that even small concentrations of GO served as a nucleating agent for the formation of graphitic structure. These authors found that the oxygen-containing functional groups on the GO play a significant role in enhancing the interfacial reactions between the GO and phenolic resin matrix. To date there is yet little experimental study on this theory to understand the role of oxygen functional groups in graphene additives and its influence in graphitization behavior of a non-graphitizable carbon matrix. Different studies have been conducted and have ascribed varied causes (stress, strain, physical and chemical templating) to the observed graphitization of hard carbon matrices/precursors. Presently, mechanistic understanding is lacking by which these factors induce graphitization, let alone ranking. Another drawback is that there is no experimental work to provide a basis for (or validation of) modeling studies for the oxygen content of additives upon templating of non-graphitizing carbon precursors. Moreover, the contributions and balance between oxygen content and sp 2 framework upon graphitization remain untested. These knowledge gaps motivated the current study.

In this work, the effect of templating using graphene oxide (GO) additives to graphitize novolac (a non-graphitizing precursor) was demonstrated. These additives were selected based on their varied oxygen content. The hypothesis is that the oxygen functional groups of the GO additives would act as leaving groups and form radical sites for bonding and alignment with matrix radicals during high temperature treatment (HTT). This process is referred to as chemical templating. Furthermore, the aromatic rings derived from the novolac matrix interact with the sp 2 network of the two-dimensional nanomaterials (as identified by the ReaxFF based study 75 , thereby aiding in formation of layered graphite material, a process referred to as physical templating.

Novolac was prepared with phenol to formaldehyde molar ratio less than one using concentrated hydrochloric acid (HCl) as a catalyst. The resulting solution was stirred with a magnetic stirrer and heated to approximately 70°C as the phenol dissolved. Once the phenol completely dissolved, 5ml of concentrated HCl was added in drops using a pipette. After a few minutes, polymerization reaction was visible as the solution began to change from a clear to a milky color followed by a spontaneous and dramatic bubble rise in the beaker. A whitish-pink precipitate then formed.

Five graphene oxide-derived additives were selected as templating agents: two graphene oxides (GO), two reduced graphene oxides (RGO) and a graphene nanoplatelet (GPL). GPL (5.3 at.(%) O), RGO (14.4 at.(%) O) and GO (30.8 at.(%) O) materials were gotten from the vendor, Cheap Tubes. RGO (15.4 at.(%) O) was obtained from the graphene supermarket and GO (35.3 at.(%) O) was purchased from Abalonyx. Table 3.1 provides a breakdown of the additives' oxygen content as measured by XPS. All additives were used as received from the vendors. A predetermined weight percentage of each was first dispersed into methanol and sonicated for six minutes. Novolac (N) with each additive was then left to mix and stir overnight resulting in a thick viscous black liquid or semi-solid with clay-like consistency.

Table 3.1 Summary of additive oxygen content measured by XPS

Carbonization was done in a customized pressurized tubing reactor. The sample was wrapped in brass foil and inserted into the reactor. The reactor was then pressurized with nitrogen to purge out oxygen and checked for any leaks. Carbonization was done at 500°C for 5 hours. The pressure in the reactor typically ranges between 500 -1500psi during the carbonization process.

After carbonization is completed, the reactor is allowed to cool down, opened, and the carbonized sample is collected and ready for graphitization.

Carbonized samples were weighed, placed in graphite crucibles, and put into a Centorr Vacuum Industries series 45 graphitization furnace. Graphitization was performed at 2500°C for 1 hour under the argon atmosphere.

The X-ray diffraction patterns were collected using Malvern PANalytical Empyrean diffractometer equipped with Cu source (λ ≅ 1.54A°), para-focusing optics and PIXcel 3D detector. The spectrum was scanned in the 2θ range of 10° to 90°. The background subtraction, peak fitting and quantification were done using MDI JADE® software.

Transmission electron microscope samples were prepared by sonicating a few milligrams (mg) of graphitized material in ethanol and then a droplet of the suspension was placed on a copper (Cu) supported lacey carbon grid and allowed to dry. The samples were imaged using a FEI TalosTM F200X scanning/transmission electron microscope equipped with FEG source providing 0.12 nm resolution. The instrument was operated at 200 kV and the samples were imaged at various magnifications in the ranges. Selected area electron diffraction (SAED) was taken concurrently with TEM imaging.

Raman spectra was collected using Horiba LabRAM HR Evolution equipped with 300 groove/mm grating and a 532nm laser. The spectra were acquired in DuoScan™ mode which increases the statistical significance of the data by rastering over a wider area. At least 5 measurements were collected for each sample to ensure that the analysis was representative.

XPS experiments were performed using a Physical Electronics VersaProbe II instrument equipped with a monochromatic Al Kα X-ray source (hν = 1,486. nm (95% of the signal originated from this depth or shallower). Quantification was done using instrumental relative sensitivity factors (RSFs) that account for the X-ray cross section and inelastic mean free path of the electrons.

Graphene oxide additives were first sonicated in methanol for dispersion 77,78 ; then this sonicated solution was added before the polymerization of the novolac phenolic resin (Figure 3.1a). These additives were selected based on their oxygen content (Supporting Information microscope (TEM) (Supporting Information Figure S1 and S2). To determine the percentage weight (wt.%) of the additive that gives the best graphitic quality, varied additive weight percentages were tested in the novolac matrix (N): 1, 1.75, 2.5, 3.5, and 5 wt.(%). X-ray diffraction (XRD) analysis was applied to determine the level of graphitization in graphitic materials. The best additive percent was gauged based on measuring the d(002) spacing wherein 2.5 wt.(%) was optimal. Figures 3.1b and c show the d(002) values for RGO1-N and GPL-N . There is a decrease in d(002) spacing with an increase in additive wt.(%) up to 2.5wt.(%) while an increase in the d(002) spacing was observed as the additive wt.(%) increased beyond 2.5 wt.(%). Therefore, the lowest d(002) spacing occurs at 2.5wt.(%) and this was the optimal weight percent of additive to induce maximum graphitization in the novolac matrix. A similar trend was observed when crystallite height (Lc) and crystallite length (La) were measured; the maximum crystallite size occurs with 2.5wt.(%) of GO additive (see supporting information, Figure S3). Figure 3.2a and 3.2b shows the XRD spectra of 2500℃ heat treated novolac, GPL-N, RGO-N and GO-N samples of varied additive oxygen content. Pure novolac has a very broad peak d(002) peak (at ~26̊ degrees) consistent with a non-graphitizing carbon. Phenolic resins have been shown not to graphitize at HTTs even above 2500℃. This is because they do not pass through a Further evidence of graphitization can be seen using transmission electron microscopy (TEM) (Figures 3.3 and 3.4). Figure 3.3 shows the TEM images of pure novolac, GPL-N and RGO1-N. Pure novolac (Figures 3.3a-c), depicts curved and ribbon-like nanostructure, short range order, and indicative of a non-graphitizing matrix. However, the TEM images of GPL-N and RGO1-N, Figures 3.3d-f and 3.3g-i respectively, illustrate improved nanostructure with sheet-like stacks and lattice fringes signifying long range order and indicative of a graphitizing material. This Selected area electron diffraction (SAED) was used to further study the graphitization level RAMAN and TEM analyses. In addition, the SAED pattern and TEM images of both RGO-N materials demonstrate RGO as the best templating additive for novolac to obtain a high-quality graphitic material. It is therefore evident that GO and its derived additives are acting as templates to direct the matrix of pure novolac matrix from a non-graphitizing to a graphitizing one; however, there is a limit of oxygen content in GO additives to effect graphitization of the novolac matrix. To further confirm the effect of oxygen groups as stated in the hypothesis, a comparison was made between the oxygen content of the graphene oxide additives and the graphitic quality of their corresponding heat-treated materials using XRD parameters d(002), Lc and La. The d-spacing value d(002) is a primary measure of graphitic quality. O content, a negative correlation occurs and the level of graphitic quality decreases.

Likewise, the presence of oxygen functional groups on graphene oxide (GO) has been shown to improve the graphitization of phenolic resins elsewhere [24-26]. In a related study, a very small concentration (0.1 wt.(%) of GO) in phenolic carbon-carbon composites served as a nucleating agent leading to the formation of graphitic structure. That study found that the oxygencontaining functional groups on the GO played a significant role in enhancing the interfacial reactions between the GO and phenolic resin matrix and improved graphitization of the composite (24). The improved GO-novolac matrix interactions occur through the esterification reaction as well as the 𝜋 -𝜋 stacking overlap between the GO and phenolic units in the matrix [25]. In molecular dynamics simulations, Papkov et al., [26] identified a templating effect of GO upon the graphitic structure of continuous carbon nanofibers made from polyacrylonitrile (PAN). Fibers containing GO (1.4 wt.(%)) showed significantly improved graphitic order compared to the pristine fibers which was attributed to a templating effect of the GO that caused formation of preferentially oriented graphitic crystallites.

Operationally, in the present study, with more oxygen groups, there are more radical sites formed for bonding with the matrix and better alignment of the matrix derived cyclic and aromatic structures aided by the 𝜋-electron network and pseudo-two-dimensional morphology of the GO additives. This is observed for up to 15.4 at.(%) O content in GO additive. Thereafter, larger amounts of oxygen groups lead to a disrupted sp 2 framework and cannot effectively act as a template to direct and align the structure of matrix from a non-graphitizing to a graphitizing one.

In this work, GO-derived additives (RGO and graphene nanoplatelets (GPL)) are shown to induce graphitization of novolac (a non-graphitizing precursor) by templating. Additionally, it was found that the templating mechanism also depends on the matrix or non-graphitizing precursor.

For instance, novolac appears to respond well (graphitize) when RGO is added.

In order to identify the chemical interactions between the GO materials and novolac, an alternative configuration was implemented. Thin films of novolac (N), RGO-N, and GPL-N were made by doctor blading and rapidly inserted into a furnace pre-heated to 300℃. FT-IR analysis was conducted on the films to gain insight on reactions that occur between the GO additives and the matrix. Based on the spectroscopic analysis for this early stage of carbonization, a possible mechanism for the templating induced graphitization of novolac by graphene oxide derivatives is proposed. For chemical reactive templating to occur, oxygen functional groups of the GO additives leave during carbonization and form radical sites for bonding with the decomposing matrix. Upon exposure to 300℃, novolac polymerization reactions occur, 79 forming intermediate structures such as hydroxymethyl phenol as seen in Figure 3.7a. The methylol group (C-OH) of hydroxymethyl phenol may undergo condensation reactions to form CH2 linkages. Similarly, the methylol group may also react with the oxygen functional groups of GO derivatives. From XPS analysis, the oxygen functional groups detected in the GO additives were -O-C=O, C=O, and C-O. The -O-C=O group is known to populate the edge sites, which will be the more reactive site for bonding of the matrix.

In the FT-IR spectra (Figure 3.7b), the -OH group (at ~3500cm -1 ) decreases for RGO-N and GPL-N compared to pure novolac. The expectation here is that the methylol group from novolac react with the -O-C=O group of the GO derived additives leading to a loss of -OH group in the form of water resulting in an ester coupling. The bands at 1730cm -1 and 1250cm -1 (Figure 3.7c) are attributed to both stretching of C=O and C-O of the ester groups that reveals formation of methyl ester via covalent bonding. This evidence supports the bonding that takes place between the novolac matrix and GO additive which is necessary for reactive templating upon HTT.

Meanwhile, the two-dimensional morphology of the GO derived additives could also play a role in graphitization through 𝜋 -𝜋 interactions with the novolac matrix as identified by ReaxFF modeling. Experimentally this has been referred to as 'confinement' or structure-directing graphitization 80 . Huang et al. used graphene with no oxygen functional groups to induce graphitization of polyacrylonitrile (PAN) resulting in highly crystalline graphite films. This would suggest that the sp 2 network of the GO additives could contribute to aromatic ring alignment in the adjacent matrix. The combined contributions of chemical and physical templating were manifested in the improved graphitic quality of RGO-N and GPL-N compared to pure novolac.

In this work, a technique to convert non-graphitizable precursors into graphitizable ones by the addition of graphene oxide and its derivatives was demonstrated. It was shown that the addition of additives with oxygen content ≥15.4 at.(%) to the novolac matrix can lead to highly graphitic material after high temperature treatment (2500°C). XRD analysis showed the decrease of d(002) and increase of crystallite size (La) and crystallite height (Lc) for GPL-N and RGO-N materials (≥ 15.4 at.(%) O). In addition, RGO-N materials of similar additive oxygen content (15.4 at.% and 14.4 at.%) gave parallel results and showed the best graphitic quality after HTT as measured by XRD. Results of improved graphitization caused by the addition of the GO derived additive were supported by Raman and TEM characterizations. The working hypothesis is that these GO derived additives act as chemical and physical templates to induce the graphitization of the novolac matrix. Chemical templating occurs as oxygen groups of the GO derived additives form reactive radicals that provide bonding to the decomposing novolac matrix. Also, aromatic rings formed during carbonization of novolac align with the sp 2 network of the two-dimensional graphene oxides, thereby aiding in the formation of layered graphite material, a process called physical templating. Lastly, the results show there is an optimum amount of oxygen content in GO additives needed to induce graphitization of the novolac matrix as XRD results show the addition GO additive of higher oxygen content (≥ 30.8% O) to the novolac matrix (GO-N) led to poor graphitic quality after HTT. This is attributed to large amounts of oxygen groups which create a disrupted sp 2 framework and cannot effectively act as a template to direct and align the structure of matrix from a non-graphitizing to a graphitizing one. This work is important because of the current need for graphitic materials, especially for energy storage applications. Additionally, the templating technique could be a more environmental and cost-effective method to turn nongraphitizing precursors into graphitizable ones.

Lignin is the second most abundant natural polymer (derived from biomass) after cellulose 81 . It accounts for up to 30 wt.% of wood 82,83 . It has been reported that over 130 million tons of waste lignin are produced in the paper making industry annually 84 . However, its use has been limited; most of it is burned to produce process heat or discharged into river directly thus causing serious environmental pollution 85 . Lignin possess many remarkable properties such as high carbon content, good thermal stability, biocompatibility, degradability, etc. 86,87 which makes it a great potential precursor for several industrial and commercial applications. These properties combined with lignin's low cost, abundance, and renewable origin have created the motivation and research interest in finding functional applications and converting lignin to high-value products 88 . Ligninderived carbon materials (LDCMs) are being considered for batteries 88,89 , supercapacitors, carbon fibers 86,90,91 , catalysis, and environmental applications. Developing these applications from lignin would create a way to integrate renewable and sustainable materials for future energy demands and address environmental issues 92,93 .

However, it has been a challenge to graphitize LDCMs into high value products or for use as a precursor for carbon composite materials. This is because lignin is a non-graphitizable carbon that does not go through a liquid crystalline state (mesophase stage) during carbonization due to its oxygen-rich backbone which facilitates crosslinking and curvature in its structure. This makes LDCMs amorphous hard carbons that are difficult to graphitize after high temperature treatment (HTT) which is in sharp contrast to the petroleum-based precursors (petroleum coke and pitch) that form highly graphitic materials after HTT. to improve the graphitic quality of kraft lignin derived from eucalyptus by heat treatment of samples up to 2800℃ (3073K) 94 . The samples used were initially washed with sulfuric acid and thermally treated to obtain different chars/ash content. After HTT of the medium ash char (0.17%) lignin sample up to 2800℃, characterizations from XRD and Raman showed progressive structural order as HTT increased which suggested an improved graphitic quality of the lignin sample. This improvement was attributed to the presence of inorganic matter impurities present in the lignin based on its ash content.

Catalytic graphitization of kraft lignin to graphene-based structures using four different transitional metals (Ni, Cu, Fe, and Mo) has also been reported 95 . The main product after heat treatment at 1000℃ were multi-layer graphene-encapsulated metal nanoparticles along with some graphene sheets/flakes. Raman spectra showed the presence of D and G bands which suggested a low degree of graphitization and it was found that the particle sizes and graphene shell layers were affected by the promoted metals.

Liu et al., 82 used alkali lignin as a carbon precursor to prepare carbon nanosheets (CNSs) by freeze-casting of a lignin aqueous dispersion and then direct carbonization at 900℃. The CNSs were prepared from 5, 10, and 20 mg/mL lignin aqueous dispersions. The graphitic quality was measured using Raman analysis wherein the lignin derived nanofibers showed prominent D and G bands, indicating a low graphitic quality of the fibers. It was also found that the graphitic degree of CNSs decreased with increasing concentration of lignin precursor dispersion because thinner lignin sheets facilitate easier removal of volatile products during carbonization.

While the use of catalysts and presence of impurities has been studied to accelerate or induce graphitization of LDCMs (with varied success), there is little work exploring the graphitization of lignin by templating using two-dimensional materials such as graphene oxide (GO). There is however research to suggest that templating by GO has the potential to aid graphitization of biomaterials. The use of GO as a templating agent has proven to be successful in the graphitization of biomaterials such as sugar. It was shown that GO additives improved the graphitic quality of the sugar due to radical sites of the GO reactively bonding to the decomposing sugar matrix during HTT; this process was called chemical templating 29 . Another proposed factor that could be contributing to improved graphitization is the interaction between the emerging aromatic domains within the matrix and the sp2 framework of the GO leading to alignment and orientation (physical templating). These results suggest that templating graphitization of lignin using GO is worth exploring. The same oxygen functional groups on GO that were key to sugar graphitization potentially could similarly graphitize lignin by reactive templating.

Furthermore, studies using reactive force field (ReaxFF) molecular dynamics (MD) and simulations have supported mechanisms of chemical reactive and physical templating. One such study by Rajabpour et al., 96 ReaxFF molecular dynamics simulations showed the addition of graphene to polyacrylonitrile (PAN) matrix improved alignment of 6-member carbon rings possibly due to with π-π interactions. Similarly, Papkov et al., 97 ran simulations using carbon nanotubes (CNTs) and graphene as templates in PAN matrix and found improved alignment between the surface of the templates and the newly formed carbon during the carbonization process. Results from the simulations showed active sites for bonding were not found on the surface of the templates, but instead, on the edge sites where C-H bonds break and form radical sites for bonding with the PAN matrix. With these findings, it is possible that oxygen groups at the edge sites of GO templates could play a similar role in the bonding and alignment of lignin matrix.

In this work, we report the effect of two-dimensional materials, graphene oxide (GO) and its derivative, reduced graphene oxide (RGO), as templating aids for the graphitization of lignin (alkali derived). The hypothesis is that these additives act as templates to graphitize lignin using 𝜋 -𝜋 interactions (physical templating) and by radical bonding between the oxygen functional groups present on these GO materials and lignin matrix during high temperature treatment (chemical templating). It was found that lignin samples with either GO or RGO showed improved d(002), La, Lc values compared to pure lignin when all samples were treated at 2500°C. Raman spectra further confirmed the improved graphitization in lignin-GO and lignin-RGO by measuring less disorder in these materials compared to pure lignin. Furthermore, transmission electron microscopy images and selected area electron diffraction patterns revealed ordered nanostructures and defined polycrystalline patterns in the lignin-GO and lignin-RGO samples. This work presents a method to synthesize graphitic-like material using carbon-based templates, meaning there is no need for further purification of the final material as in the case of transition metal catalysts.

Graphene oxide (GO) and reduced graphene oxide (RGO) were used as template additives in the lignin matrix. Table 4.1 provides a breakdown of the elemental composition as measured by XPS. Lignin, alkali was obtained from Sigma Aldrich (8068-05-1). The nanomaterials additives, RGO and GO were obtained from Cheap Tubes. Lignin-RGO and lignin-GO were made using 2.5 wt.% of each additive. The mixtures were sonicated in methanol for 15 mins and left in the oven at 200°C for 5 hours.

The three samples (lignin, lignin-GO, and lignin-RGO) were carbonized at 500°C for 5

hours in an inert atmosphere using a tubing bomb reactor.

The carbonized samples were graphitized at 1000°C, 1500°C and 2500°C. The 1000°C graphitization was done in a Thermolyne 2110 tube furnace at 15°C/minute. The 1500°C was done in a GSL-1700X-UL furnace at a 10°C/min temperature ramp up to 900°C and then 5°C/min to 1500°C. Graphitization was done at 2500°C in a Centorr Vacuum Industries series 45 graphitization furnace, heating at 15°C/min to 1000°C, 10°C/min to 2000°C and finally 5°C/min to 2500°C. All graphitization runs were performed under an inert atmosphere.

FT-IR was done in a Bruker vertex 80 spectrometer. IR spectra was collected using attenuated total reflected (ATR) method for the pure lignin and stabilized samples at 200°C.

Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was used to collect spectra for samples carbonized at 500°C.

Thermogravimetric analysis (TGA) was done in a Q600 (TA Instruments, USA). Samples (lignin, lignin-GO and lignin-RGO) were subjected to a temperature ramp test (30 -900 °C at 10 °C/min).

The X-ray diffraction patterns were collected using Malvern PANalytical Empyrean diffractometer equipped with Cu source (λ ≅ 1.54A°), para-focusing optics and PIXcel 3D

detector. The spectrum was scanned in the 2θ range of 10° to 90°. The background subtraction, peak fitting and quantification were done using MDI JADE® software.

Transmission electron microscope samples were prepared by sonicating a few milligrams (mg) of graphitized material in ethanol and then dropped on a copper (Cu) supported lacey carbon grid and allowed to dry. The samples were imaged using a FEI TalosTM F200X scanning/transmission electron microscope equipped with FEG source providing 0.12 nm resolution. The instrument was operated at 200 kV and the samples were imaged at various magnifications in the ranges. Selected area electron diffraction (SAED) patterns were taken concurrently with TEM imaging.

Raman spectra was collected using a Horiba LabRAM HR Evolution instrument equipped with 300 groove/mm grating and a 532nm laser. The spectra were acquired in DuoScan™ mode which increases the statistical significance of the data by rastering over a wider area. At least 5 measurements were collected for each sample to make the analysis more representative.

SEM images were taken with a field-emission SEM, an Apreo 5. Samples were prepared by placing a few milligrams on a carbon taped pin stub holder. To obtain FESEM images, an acceleration voltage of 7kV and a working distance between 11mm to 7mm was maintained.

XPS experiments were performed using a Physical Electronics VersaProbe II instrument equipped with a monochromatic Al Kα X-ray source (hν = 1,486.7 eV) and a concentric hemispherical analyzer. Charge neutralization was performed using both low energy electrons (<5 eV) and argon ions. The binding energy axis was calibrated using sputter cleaned Cu (Cu 2p3/2 = 932.62 eV, Cu 3p 3/2 = 75.1 eV) and Au foils (Au 4f 7/2 = 83.96 eV). Peaks were charged with reference to C-C (sp 2 ) band in the carbon 1s spectra at 284.5 eV. Measurements were made at a takeoff angle of 45° with respect to the sample surface plane. This resulted in a typical sampling depth of 3-6 nm (95% of the signal originated from this depth or shallower). Quantification was done using instrumental relative sensitivity factors (RSFs) that account for the X-ray cross section and inelastic mean free path of the electrons.

The thermal degradation behavior of lignin, lignin-RGO, and lignin-GO samples was evaluated using thermogravimetric analysis (TGA) (Figure 4.1a). The temperature ramp (up) was performed from room temperature to 900°C followed by an isothermal hold for two hours at 900°C. At 100°C, the thermogram of the pure lignin sample shows a weight loss of more than 5% compared to GO-and RGO-lignin samples, each with a loss of less than 1%. This is the temperature at which absorbed water is lost. At 400°C, the weight loss of lignin was 32%, compared to lignin-GO (29%), and lignin-RGO (28%) mixtures. The trend of more weight loss in lignin continues up to the final temperature 900℃ where lignin loses more than half its weight (57%) while for both lignin-GO and lignin-RGO, the corresponding weight losses were less than 50% (Figure 4.1b).

Lignin is usually stabilized to increase carbon yield in carbon fibers production 31,86,88 ; TGA analysis has shown higher carbon content in the lignin-GO and -RGO samples. It appears graphene oxide and reduced graphene oxide are stabilizing/crosslinking with lignin leading to the high carbon content observed. Physical evidence of crosslinking is also seen in (Figure 4.1c, d and e).

At 200°C, the lignin-GO and lignin-RGO samples formed solid structures at 200°C unlike pure lignin which remained in powder form.

To confirm the occurrence of crosslinking reactions, FT-IR was used to measure the changes in functional groups of lignin, lignin-GO and lignin-RGO materials. between GO additives and lignin causing loss of water leading to reduction in -OH group. The peaks at 1591cm -1 and 1450cm -1 appear to increase in lignin-GO and -RGO. These peaks are attributed to phenyl ring vibrations and a -CH stretch deformation, respectively 98,99 . Likewise, the peaks at 1338 cm -1 and 1266 cm -1 are attributed to the C-O-C vibration increase, consistent with etheric bridge formation. These changes suggest an interaction between lignin and GO additives at 200℃ causing a loss of -OH, possibly in the form of water and formation of new bonds such C-O-C at ~1200cm -1 . This could explain the solidification of the lignin-GO and lignin-RGO samples.

At 500℃, a complete reduction in the -OH group at 3300cm -1 and an increase at 1458cm -1 for lignin-GO additives is observed (especially for lignin-RGO). Everything else remains essentially the same. The 1458cm -1 peak can be attributed to improvement of the alkyl stretch, -CH2 symmetric scissoring. This indicates improved bonding between the lignin and GO additives. X-ray Diffraction (XRD) is a primary technique used to determine the crystallinity of materials. Figure 4.2a shows the XRD spectra of lignin at the different HTTs. There is no sign of graphitic-like materials (low crystallinity/amorphous quality) observed for lignin at 1000°C and 1500°C indicated by the low intensity and very broad d(002) peak as well as undefined peaks at XRD parameters are typically used to measure graphitization: interlayer spacing (d(002)), crystallite size (La), and crystallite height (Lc). The d( 002) is a primary measure of graphitization and a decrease in d(002) implies a higher degree of graphitization (improved graphitic quality).

The d(002) peak was deconvoluted into two phases, the amorphous and crystalline phase, due to its broadness and asymmetry. Also, an increase in La and Lc values signifies an increase in graphitic quality. In Likewise, the Lc values for lignin-GO and lignin-RGO are 9nm and 8nm, respectively, compared to that for pure lignin (5nm). This indicates a greater degree of graphitization (or improved graphitic quality) for lignin-RGO and lignin-GO. The La and Lc trend at 2500℃ is clearly depicted in the line graph in Figure 4.3d. There was a 50% increase in La for both lignin-GO and lignin-RGO compared to pure lignin. Also, there is an 80% increase in Lc for lignin-GO and a 60% increase in the Lc for lignin-RGO compared to pure lignin. The increase in La and Lc is attributed to the templating effect of the graphene oxides on lignin. Raman spectroscopy is a second important technique for characterizing the graphitic quality of carbon materials. The G peak at 1580cm -1 is the characteristic peak of sp 2 hybrid structure, which represents the symmetry and crystallinity of graphene materials; the D peak at 1350 cm -1 is a defect peak, representing the defects and disorder of graphite layers 100101 . Figure Conversely, pure lignin exhibits a higher ID/IG ratio (0.38), suggesting increased defect levels and diminished graphitic quality. The subsequent columns detail the G-peak position, and the G-full width half-maximum (G-FWHM) provide clear evidence of the disparity in crystalline content between pure lignin and GO additive samples. The smaller, narrower G-FWHM signifies higher graphitic quality. Lignin-GO and lignin-RGO exhibit lower G-FWHM values than pure lignin.

Further analysis of Raman parameters, including the ratio of the 2D to G band (I2D/IG), underscores improved graphitic layers in lignin-GO and lignin-RGO samples, as they exhibit higher I2D/IG values. This trend is also observed in the full-width half-maximum (FWHM) ratio of the 2D to G band in the last column, with higher values for lignin-GO and lignin-RGO compared to pure lignin.

These findings suggest lower defects and enhanced crystalline content in lignin-GO and lignin-RGO sample. It appears that addition of GO derived additives into the lignin matrix can influence the formation of larger crystallites and improve the overall structure. d), and lignin-RGO (e, f) carbons. The TEM images of lignin (Figure 4.5a) show a disordered nanostructure indicative of a non-graphitizing matrix. Curved structures yet embedded also support disorder in the nanostructure. The SAED patterns of lignin (Figure 4.5b) show both an amorphous halo ring pattern and the presence of nano-crystallites (which arose from the HTT at 2500℃) but do not imply long-range order in the pure lignin sample. However, the TEM images of lignin-rGO (Figure 4.5c) and lignin-GO (Figure 4.5e) show an improved nanostructure with sheet stacking and the presence of lattice fringes indicative of a graphitizing matrix. This suggests

The hypothesis for this work is that graphene oxide additives can induce graphitization of the lignin matrix by the process of templating during heat treatment. Characterization by XRD showed improvement in the graphitic quality (measured by lattice parameters, crystallite height, and size) in lignin-GO and lignin-RGO. There was an overall 50% increase in crystallite size, La, and between a 60 to 80% increase in crystallite height, Lc, for lignin-RGO and lignin-GO compared to pure lignin. This improvement in graphitic quality was supported by Raman analysis which revealed lower disorder in these lignin-GO and -RGO materials compared to pure lignin.

Furthermore, TEM images and SAED patterns showed orderliness in the nanostructure with presence of lattice fringes, sheet stacking and sharp and defined polycrystalline patterns. All this evidence points to an increase in the graphitic quality facilitated by the addition of GO and RGO 2D templates. Graphene oxides are well known for their two-dimensional sp 2 framework which can serve as 'confinement' for directing the structure of housing matrix. We propose this as the operative feature by which the GO additive directs the nano-to micro-structure of the lignin matrix.

With the extended sp 2 framework, the GO and RGO act as a 'template' to guide graphitic development by process of π-π interactions. This appears as the major driving force of graphitization. Considering the presence of oxygen groups on the graphene layer creates a mixed sp 2 -sp 3 state, however, there is still a large portion (between 40 to 60%) which are unfunctionalized and provides the sp 2 framework for templating the lignin matrix into graphitic materials. Moreover, during heat treatment, the de-functionalization of the GO and RGO additives extends the sp 2 framework and further guides the templating of lignin. In addition, possible chemical bonding between the GO additives and lignin may extend the two planar frameworks.

Related ReaxFF molecular dynamics simulations also support templating mechanisms of the GO derived additives 26 . Mechanisms depicted by Gharpure et al 26 are similar to observations and trends seen in the above experimental results of lignin-GO and lignin-RGO. It can therefore be surmised that GO-additives act as templating agents resulting in the change in structure and properties of lignin. The use of the 2D carbon-based additives in the form of graphene oxide as a 'template graphitization agent' is a novel process in transforming the structure of bio precursors such as lignin.

In this work, the templating effect of graphene oxide (GO) and reduced graphene oxide (RGO) additives on the graphitization of lignin was demonstrated., XRD, Raman, and TEM characterizations were done to characterize the templating effect of the GO additives on the lignin matrix. Results showed some improved graphitic quality in the lignin-GO and lignin-RGO samples compared to pure lignin at 2500℃. It was also found that lignin-GO and lignin-RGO exhibited equal or comparable degree of graphitization. Results also reveal operative feature by which the GO additive directs the nano-to micro-structure of the lignin matrix: first, the extended sp 2 framework, the GO and RGO act as a 'template' to guide graphitic development by process of ππ interactions. This appears as the major driving force of graphitization. In addition, during heat treatment, the de-functionalization of the GO and RGO additives extends the sp 2 framework and further guides the templating of lignin. The use of GO-additives acts as templating agents resulting in the change in structure and properties of lignin to form graphitic materials provides the advantage of no further purification of the final material as in the case of transition metal catalysts.

The content of this chapter is adapted from a paper under final review in Carbon Trends and referenced as:

Ike, S. N., & Vander Wal, R. Distinguishing Physical vs. Chemical Templating Mechanisms for Inducing Graphitization in Novolac Matrix. Chemical Templating Mechanisms for Inducing Graphitization in Novolac Matrix.

Chapter 3 investigated the templating ability of graphene oxide-derived additives to induce graphitization of the novolac matrix. The findings led to two working hypotheses: physical and chemical templating. The additives act as templates that promote matrix aromatic alignment to their basal planes during carbonization (referred to here as physical templating) in addition to forming radical edge sites that bond to the decomposing matrix (referred to here as chemical templating). However, results mainly underscored the role of functional groups the GO additiveschemical templating. The aim of this current work seeks to differentiate the contributions of the operative mechanisms and shed light on their effect on graphitization. To study this, 2D materials with minimal oxygen functionalization, graphene, and hexagonal boron nitride (hBN), were used as templates to induce graphitization. First, the optimum weight percent of the physical additives in the novolac matrix was determined and the graphitic quality was measured by characterizations from X-ray Diffraction and Raman spectroscopy. Results revealed that hBN did not induce the graphitization of novolac matrix and was attributed to the absence of a sp² framework in hBN, unable to provide the crucial π-π interactions with the aromatic rings of the matrix. In contrast, the results gotten from two graphene additives mirrored one another and showed improved graphitization of the novolac matrix-reduced d(002) and improved Lc values, however, La values were low compared to that obtained from 'chemical templates' used in chapter 3. From these results, it was deduced that both mechanisms are operative and have a working interplay in inducing graphitization of the novolac matrix. While physical templating offers control over longrange order in the form of crystallite height, chemical templating contributes to carbon reorganization and lateral growth extent.

Two-dimensional materials, represented by graphene and its derivatives, possess a distinctive templating capability, serving as fundamental building blocks for synthesizing carbon materials across diverse applications. Notably, the interlayer spaces within these materials function as either confinement spaces or reactive sites during templating processes. Additionally, graphene derivatives, like graphene oxides, exhibit additional versatility due to functional groups, influencing the guidance and direction of new material formation [1,2]. Numerous studies have investigated the templating ability of 2D material, graphene, in influencing carbon growth and structure. In one such study, Guoping Yang et al [3] synthesized a phenolic resin-based porous carbon composites with graphene and potassium hydroxide activation. After carbonization heat treatment, the study measured crystallinity by XRD and Raman and cited improved crystallinity aided by the presence of graphene in the carbon-composites as compared to pure phenolic resin.

In another study, Shi et al. utilized graphene as a template for MoS2 growth, highlighting the significance of weak van der Waals forces in facilitating MoS2 layer growth on the graphene surface [4]. Similarly, Huang et al. demonstrated that graphene, devoid of oxygen functional groups, induces graphitization of polyacrylonitrile (PAN), emphasizing the sp 2 network's contribution to aromatic ring alignment [5]. Gharpure et al also reported a small amount of graphene additive aided mesophase formation in pitches and acted as a templating agent during heat-treatment to catalyze the emergence of graphitic quality, with a majority of the improvements seen below 2000 °C [7]. Moreover, the size of graphene filler also plays a role in the degree of graphitization as found in the work done by Singh et al. [8] who studied the effect of lateral dimension of graphene as a filler in a graphene-anthracene C-C composite. Other studies also reveal that graphene's physical lattice orientation and van der Waals forces govern atomic-scale chemical reactions [6]. This templating process, known as physical templating, relies on the unique characteristics of 2D materials. Notably, graphene layered structure, devoid of permanent polarity or functional groups makes it an ideal template for nanomaterial synthesis under van der Waals interactions.

Earlier chapter (chapter 3) underscored the role of functional groups in graphene derivatives, such as graphene oxide and reduced graphene oxide, as reactive sites for bonding and directing precursor novolac matrix formation-a process termed chemical templating. Chapter 3 elucidated that graphene oxide additives with varying oxygen content induce graphitization of nongraphitizing phenolic resin, namely novolac. The working hypothesis suggests that oxygen groups depart during heat treatment, forming radical sites for bonding with the decomposing novolac matrix. While chemical templating influences graphitization, the chapter claimed the significance of physical templating persists due to the sp2 framework of 2D materials and their interactions with the novolac matrix.

This chapter aims to scrutinize the predominant force and operative mechanism driving graphitization-physical templating versus chemical templating. The investigation centers on the role of physical templating, employing 2D materials such as graphene and hexagonal boron nitride (hBN) with minimal oxygen content to induce novolac matrix graphitization. To ensure a meaningful comparison, consistent weight percentages of additives and heat treatment steps were maintained, aligning with previous chapter.

Novolac resin was synthesized using a phenol-to-formaldehyde molar ratio below 1, catalyzed by concentrated hydrochloric acid (HCl). The mixture was stirred magnetically and heated to approximately 70°C until the phenol dissolved completely. Subsequently, 5 ml of concentrated HCl was added dropwise using a pipette. Within minutes, the polymerization reaction became evident. The initially clear solution transformed into a milky suspension, followed by a rapid and dramatic bubble formation in the beaker. Finally, a whitish-pink precipitate solidified.

The additives, graphene 1 (G1), graphene 2 (G2), and hexagonal boron nitride (h-BN),

were chosen as physical templating agents at various weight percentages: 0.5%, 2.5%, and 5%.

G1 was sourced from MTI Corporation and G2 additive from XG Sciences while h-BN was obtained from Sigma Aldrich. All additives were used as received from the vendors. To achieve homogeneous dispersions, a predetermined weight percentage of each additive was first dispersed in methanol and sonicated for six minutes. This ultrasonic treatment effectively exfoliated the layered structures of the additives, promoting their interaction with the Novolac resin.

Subsequently, the Novolac (N) with each additive was mixed and stirred overnight. This extended mixing period ensured the thorough incorporation of the dispersed additives, resulting in thick, viscous black liquids or semi-solids with a clay-like consistency.

Carbonization was performed in a custom-designed pressurized tubular reactor. The sample was carefully wrapped in brass foil and placed within the reactor. To ensure complete oxygen removal, the reactor was purged with nitrogen and checked for leaks. Carbonization was then conducted at 500°C for 5 hours under a pressure range of 500-1500 psi. Following the process, the reactor was allowed to cool down gradually before opening. Finally, the carbonized sample was collected for subsequent graphitization.

Carbonized samples were weighed, placed in graphite crucibles, and put into a Centorr Vacuum Industries series 45 graphitization furnace. Graphitization was performed at 2500°C for 1 hour under the argon atmosphere.

XPS analysis was conducted using a Physical Electronics Versa Probe II equipped with a monochromatic Al Kα X-ray source (1486.7 eV) and a concentric hemispherical analyzer.

Charge neutralization employed both low-energy electrons (<5 eV) and argon ion sputtering.

Calibration of the binding energy scale was achieved using sputter-cleaned copper (Cu 2p3/2 = 932.62 eV, Cu 3p3/2 = 75.1 eV) and gold (Au 4f7/2 = 83.96 eV) foils. Peak referencing utilized the C-C (sp2) component of the carbon 1s spectrum at 284.5 eV. Measurements were performed at a 45° takeoff angle relative to the sample surface, resulting in a typical sampling depth of 3-6 nm (95% signal contribution from this depth). Quantification relied on instrument-specific relative sensitivity factors (RSFs) accounting for X-ray cross-sections and electron inelastic mean free paths.

The X-ray diffraction patterns were collected using Malvern PANalytical Empyrean diffractometer equipped with Cu source (λ ≅ 1.54A°), para-focusing optics and PIXcel 3D

detector. The spectrum was scanned in the 2θ range of 10° to 90°. The background subtraction, peak fitting and quantification were done using MDI JADE® software.

Transmission electron microscope samples were prepared by sonicating a few milligrams (mg) of graphitized material in ethanol and then dropped on a copper (Cu) supported lacey carbon grid and allowed to dry. The samples were imaged using a FEI TalosTM F200X scanning/transmission electron microscope equipped with FEG source providing 0.12 nm resolution. The instrument was operated at 200 kV and the samples were imaged at various magnifications in the ranges. Selected area electron diffraction (SAED) patterns were taken concurrently with TEM imaging.

Raman spectra were collected using a Horiba LabRAM HR Evolution instrument equipped with 300 groove/mm grating and a 532nm laser. The spectra were acquired in DuoScan™ mode which increases the statistical significance of the data by rastering over a wider area. At least 5 measurements were collected for each sample to make the analysis more representative.

SEM images were taken with a field-emission SEM, an Apreo 5. Samples were prepared by placing a few milligrams on a carbon-taped pin stub holder. To obtain FESEM images, an acceleration voltage of 7kV and a working distance between 11mm to 7mm were maintained.

First, characterization of the physical template additives was conducted to understand individual properties such as morphology, elemental composition, and nanostructure.

Scanning electron microscopy (SEM) revealed distinct differences in the surface morphology of graphene (G1 and G2) and hexagonal boron nitride (HBN) templates. agglomerated into larger clusters. This morphology significantly differs from the flaky texture observed for the graphenes. X-ray spectroscopy (XPS) analysis (Table 5.1) confirmed the presence of boron and nitrogen as expected in hBN, but also revealed traces of organic molecules, potassium, oxygen, and even a minor silicate component. G1 and G2 have equivalent elemental composition. Notably, low amounts of oxygen ~1.1% and sulfur content of ~ 0.1% were detected in the G1 sample, and ~1.4% oxygen, and 0.2% sulfur for G2. The C1s spectra (Figure 5.2a-c) presented a narrow peak for G1 (Figure 5.2a) centered at 284.5 eV with a full-width half maximum (FWHM) of 0.80, indicating primarily sp² hybridized carbon and minimal defects for G1. The G2 spectra (Figure 5.2b) also reveals a narrow C1s FWHM of 0.66 revealing a more structured carbon system. hBN a broader FWHM of 2.01 (Figure 5.2c) indicating more defects which might be because of functionalities by the trace elements revealed in Table 5.1. The differences in the nanostructures of graphenes (G1 and G2) and hBN were further elucidated in the TEM images. Graphene (G1) in Figure 5.3a shows the sheets that appear transparent in lower magnification images indicative of a few layers (2-4 layers). This sheet stacking is also seen in the higher magnification image. G2 (Figure 5.3b) also revealed wrinkled transparent sheets which can be estimated to be a few layers ~2-4 layers. All similar characteristics were observed in G1. Unlike the graphene sheets, hBN reveals a curved (fullerene) ribbon-like nanostructure (Figure 5.3c). How these additive structures influence their effect in inducing the graphitic templating of the novolac matrix is explored in the next section. As explained in the methodology section, these physical template additives were added into the novolac matrix, carbonized, and graphitized. The two graphene additives (G1 and G2) gave very similar results; therefore, the discussion on these additives will be referred to as graphene (G) onwards. Figure 5.4a shows thermal stability for pristine graphene and hBN. Graphene undergoes a swift, single decomposition step at 600°C, while hBN loses adsorbed water at 100°C

and remains stable for an extended temperature range. Both additives are introduced into the novolac matrix and carbonized at 500°C (Figure 5.4b). Carbonized G-N retains its characteristic decomposition at 600°C, while hBN-N exhibits a distinct 400°C decomposition peak reflecting the surrounding novolac's breakdown. Carbonization boosts thermal stability, evident in the higher onset temperatures for weight loss across all samples compared to their pristine counterparts. The stable carbon structure formed during carbonization is the driving force behind this improvement.

To probe the effect of the physical template additives in inducing graphitization of the novolac matrix, their weight concentration was varied at 0.5%, 2.5%, and 5% and passed through carbonization and graphitization heat treatment. XRD analysis (Figure 5.4c and 4d) reveals the structural transformations at varying additive weight percentages. In Figure 5.4c at 0.5 wt% loading, a broad d(002) peak at ~26° indicates the presence of non-graphitizing carbon within the matrix. However, increasing the graphene content to 2.5 wt% brings a remarkable shift. The d(002) peak narrows significantly, accompanied by the emergence of the (004) and (006) peaks of the (001) series. This points towards a more ordered stacking along the c-axis a sign of enhanced graphitization. Further concentration beyond 2.5 wt% (5 wt% G-N) leads to the broadening of the d(002) peak. This suggests an optimal window for graphene's influence is at 2.5%, where it effectively promotes the matrix's transformation towards a more graphitic structure. On the other hand, regardless of the loading (0.5 wt% to 5 wt%), the XRD spectra for hBN in Figure 5.4d consistently displays a broad d(002) peak, characteristic of non-graphitizing carbon. This implies that hBN, unlike graphene, does not significantly impact the graphitization process within the novolac matrix. To gain deeper insights into the degree of graphitization, lattice parameters, d(002), crystallite height (Lc), and crystallite size (La) extracted from deconvoluted XRD curves will be discussed later. Raman spectroscopy provides a complementary analysis to XRD [10,11]. details about their graphitic nature. In Figure 5.4e, the intensity of the D-band at 1350 cm⁻¹ provides a signature of defects and disorder in graphitic layers. For 0.5 wt% G-N, a prominent Dband signifies significant lattice defects. However, the intensity drastically reduces in the 2.5 wt% G-N sample, aligning perfectly with the XRD observations of enhanced graphitization.

Interestingly, the D-band reemerges in the 5 wt% G-N sample, suggesting an oversaturation effect where excessive graphene hinders the ordering process. This reinforces the notion of an optimal 2.5 wt% loading for promoting graphitization as observed from XRD. A similar trend can be seen in the hBN-N samples (Figure 5.4f). The persistent D-band across all weight percentages confirms the lack of significant graphitization under hBN's influence. This further corroborates the XRD findings. The ratio of D-band intensity to G-band intensity (ID/IG) offers a quantitative measure of the defect concentration and graphitization degree in each sample. This combined analysis using XRD and Raman spectroscopy shows how graphene and hBN differ in influencing the graphitization of the novolac matrix at high temperatures. While graphene exhibits an optimal window for promoting order and graphitization at 2.5 wt% loading, hBN appears to hold minimal influence over this process. XRD spectra of the graphitized samples at the varied weight percent of additives (c) G-N and (d) hBN-N; Raman spectra of the graphitized samples at the varied weight percent of additives (e) G-N and (f) structures. This confirms the decline in graphitic quality at higher graphene concentrations. Figure 5.6c shows 2.5 wt% hBN-N as rough, fused aggregates with no evidence of sheet stacking, indicative of its non-graphitic nature. However, a closer look at the 5 wt% hBN-N sample (Figure 5.6d) reveals the minimum emergence of some stacked sheet-like features amidst the roughness and smaller disordered flakes. While this suggests a slight increase in order compared to lower hBN loadings, it falls short of exhibiting the well-defined long-range order or high graphitic quality observed in the 2.5 wt% G-N sample. Thus, SEM analysis reinforces the overall obervation:

graphene promotes graphitization at an optimal 2.5 wt% loading, while hBN has minimal influence on the process. This multi-technique approach provides a comprehensive understanding of the distinct roles these additives play in shaping the graphitized structures of the novolac matrix. Transmission electron microscope (TEM) images and corresponding SAED pattern (Figure 5.7) illustrate the nanostructure and crystallite pattern of the materials. 2.5 wt% G1-N (Figure 5.7a) shows a well-ordered nanostructure characterized by long-range order and presence of lattice fringes. The corresponding sharp polycrystalline SAED pattern (Figure 5.7b) supports the ordered crystallites. However, the TEM image of 5 wt% G1-N (Figure 5.7c) does not show the same level of ordered nanostructure as the 2.5 wt% G1-N. There is a mixture of curved and ordered sheets, and the corresponding SAED pattern (Figure 5.7d) shows both polycrystalline and amorphous properties. This evidence is supported by XRD and Raman which showed a reduced graphitic quality in comparison to 2.5 wt% G-N. In contrast, the TEM and SAED pattern of 2.5 wt% hBN-N (Figure 5.7e and 5.7f) and 5 wt% hBN-N (Figure 5.7g and 5.7h) reveal amorphous properties represented by curved fullerenic nanostructure and halo rings SAED pattern which is consistent with results from XRD and Raman. This therefore adds proof that hBN was not an effective template to induce the graphitization of the novolac matrix.

Unlocking high-performance carbon materials demands precise control over their graphitic structures. This study investigated the interplay between physical and chemical templating mechanisms governing the graphitization of a novolac matrix, utilizing two-dimensional (2D) nanomaterials as templates. The work explored the potential of graphene, with its extensive sp² network, to promote matrix ordering through physical templating, where aromatic rings in the novolac interact with the 2D framework, facilitating stacking and alignment. Our findings show that graphene addition led to a lower d(002) spacing (indicating increased interlayer order) and higher crystallite height (Lc) compared to the pristine matrix, novolac [9]. However, results from hexagonal boron nitride showed that despite possessing a similar layered structure, it exerts minimal impact on graphitization, as evidenced by the higher d(002) and lower Lc values in hBN-N. This could be attributed to the absence of C-C linkages in hBN's sp² framework, hampering the crucial π-π interactions with the aromatic rings of the matrix. To further elucidate the relative significance of each mechanism, we compared physically templated G-N and hBN-N with chemically templated graphene oxide derivatives (rGO-N and GPL-N) [9] (see Table 5.2). The presence of oxygen functional groups on rGO and GPL introduces leaving groups that can promote radical formation during high-temperature treatment (HTT). These radicals interact and align with matrix radicals, favoring a more ordered graphitic structure. From the table, the d(002) and Lc values for rGO-N and GPL-N are statistically comparable to those of G-N suggesting enhanced graphitization. The reverse trend is observed for La values, rGO-N and GPL-N have higher La values compared to G-N. From these observations and results, it can be deduced that templates which possess a good amount of oxygen functional groups, seemingly located on the edge sites that form radical sites during heat treatment allow for growth along the AB axis and therefore, improve La. On the other hand, the sp 2 hybridization of the basal plane allows for growth along the C-axis, improving d(002) and Lc. This is also supported in the results where oxygen content does not contribute reactive (i.e. chemical) templating along the c-axis for reason that the tested rGO and GPL have varied total oxygen content, and thus similarly different basal and edge site oxygen content [9]. Yet, for the varied basal [O]-atom content, Lc appears to be independent, leading to the above conclusion.

In other words, while chemical templating may contribute to initial carbon reorganization or lateral carbon growth, a well-developed, long-range ordered structure in terms of crystallite height is promoted by physical templating. From the result, it can be deduced that both mechanisms are operative in inducing graphitization of the novolac matrix.

In conclusion, our investigation into the graphitization of a novolac matrix using graphene and hexagonal boron nitride (hBN) as templates has revealed distinctive influences on the structural and thermal properties of the resulting composites. Graphene, at an optimal loading of 2.5 wt%, acts as a template catalyst for enhanced graphitization, promoting long-range order and larger crystallite sizes. In contrast, hBN shows minimal impact on the novolac matrix's graphitization process. The study compared physical templating, represented by graphene, and chemical templating, illustrated by graphene oxide derivatives. Our findings suggest while physical templating with graphene appears to offer greater control over long-range order in the form of crystallite height, chemical templating through graphene oxide derivatives can contribute to carbon reorganization by carbon lateral growth. Understanding the precise roles of each mechanism and their potential synergy holds significant promise for tailoring the synthesis of advanced carbon materials with desired properties for diverse applications. Studies have shown crystallite size strongly influences the electrical resistivity and thermal conductivity as well as modulus of the graphite material [12-16]; this research presents a technique for tailoring these specific crystallite properties to meet such needs. Future research will delve deeper into the interplay of template functionalities, optimize processing conditions, and explore complementary characterization techniques to fully elucidate the mechanisms governing the precise control over graphitization processes. The insights gained from this study pave the way for tailoring the synthesis of advanced carbon materials with desired properties for diverse applications, ranging from thermal management to structural reinforcements.

The growing need for greener and cleaner energy has put a lot of focus on the development of better and high-performance materials for renewable energy storage applications. Lithium-ion batteries (LiBs) have been at the forefront as an efficient energy storage system and are widely used in portable electronics and commercial electric vehicles (1). In general, LiBs consist of anode, cathode, electrolyte, and separator. The anode material is a key regulator of the LIBs as it determines cell performance such as cycle life, rate capability and energy density (2). Therefore, research into developing better and higher quality anodes materials is important and necessary for high performance LiBs. Carbonaceous materials in the form of natural or synthetic graphite have dominated as anodes in LiBs. This is because of their excellent ability to reversibly intercalate lithium-ions and operate at low voltage of 0.1V in reference to Li+/Li (2). Graphite also has a relatively good energy density of 372 mAh/g(3). However, performance issues at low temperatures due to resistance within the electrode, the charge-transfer process and reduced rate of lithium diffusion are major drawbacks of graphite anodes (4, 5). Other carbonaceous materials such as hard carbons have shown promising characteristics in terms of cycling ability, fast charging performance and better capacity at lower temperature(6, 7) that make them practical and viable candidates as an alternative anode material in LiBs. Moreover, the morphology of hard carbons with short-range domains, pores, defective and edge sites provide distinctive characteristics for the transport and storage of Li-ions (8).

The mechanism of lithium-ion storage and transfer differs in graphite than for hard carbons (Figure 6.1). For Graphite, Li-ions carried by the electrolyte insert into the interlayer space to form graphite intercalation compound (GIC) (9-11). After staging process (a phenomenon where GICs are sorted into graphite layers), the GIC becomes LiC6, and intercalation is complete (Figure 6.1a)(9-11). This is accompanied by a change in the stacking order of the graphite from ABAB to establish a solid relationship among the selection of precursors, preparation parameters, structural and morphological properties, and electrochemical performances of the obtained hard carbons

In this chapter, we tested the electrochemical properties of hard carbons, novolac phenolic resin and bio precursor lignin, which were modified using two different techniques. Lignin was modified using an emulsion polymerization technique with poly furfuryl alcohol (PFA). Emulsion polymerization is a versatile and scalable technique used to synthesize polymer microspheres with particle size ranging from a few nanometers to microns (21-23). The spherical shape is formed with the assistance of a structure directing agent (i.e., template), a surfactant, which aggregates to a micelle and are crosslinked by the addition of a strong acid. The size and shape of the polymeric particle is also influenced by the polymerization kinetics (24). This technique is done to control the morphology of the carbon while preserving the internal properties such as porosity, pore size distribution and surface area (24). Secondly, from chapter 3, novolac, a phenolic resin was modified to a graphitic carbon using the templating technique with 2D materials such as graphene nanoplatelets (GPL) and reduced graphene oxide (RGO). These 2D materials in the form of graphene oxide additives act as templates (structure directing agents) that promote matrix aromatic alignment to their basal planes during carbonization (physical templating) in addition to forming radical sites that bond to the decomposing matrix (chemical templating) to guide the structure from a non-graphitic to a graphic carbon. To understand how these precursors and preparation parameters influence the electrochemical properties, a Galvanostatic test at different C-rates was done for 20 cycles and electrochemical parameters such as the first cycle reversibility, columbic efficiency, and final capacity (mAh/g) were compared to determine their ultimate performance as anodes in LiBs.

The materials for this experiment, lignin (L), poly furfuryl alcohol (PFA) and Pluronic F-127, were purchased from Sigma-Aldrich and were used as received. To make the carbon spheres, pluronic F-127 was dissolved in a mixture of ethanol and water. This solution was allowed to stir for several minutes to allow surfactant dissolve. Concentrated hydrochloric acid (HCL) was added dropwise as a polymerization initiator. This was followed by the addition of lignin. The mixture was allowed to stir for about 30 minutes and PFA was added. The solution was allowed to rest and stir for 24hrs. After 24hrs, 5M sulfuric acid was added to solution and the solution was heated to 90℃ until most of solvent evaporated and a thick viscous residue remained. This was followed by vigorous washing of the residue by method of centrifuging using distilled water. The collected solid was allowed to dry in a 90℃ oven for 24hrs. The dried sample was carbonized at 800℃ under inert atmosphere. With the intended goal being to make a more sustainable anode material from a bio precursor, the percentage of lignin in the mixture was increased to understand if performance will improve with higher biomass as active material. The weight percent of lignin in the synthesis process varied at 30 wt.%, 50 wt.%, and 70 wt.% to make a 30L/70P, 50L/50P and 70L/30P sample, respectively.

Novolac, RGO-N and GPL-N active materials were used as prepared in chapter 3. All characterizations of these samples such as x-ray diffraction, Raman spectroscopy and transmission electron microscopy can also be found in chapter 3.

BET Nitrogen gas adsorption measurement combined with Barrett-Joyner-Halenda (BJH) method was used to determine pore size distribution and cumulative pore volumes of the samples.

Samples were characterized by N 2 adsorption using Micromeritics instrument at the temperature of liquid nitrogen. Prior to analysis, the samples were outgassed in a vacuum at 200 °C for 3-5 h.

BET surface areas were determined from the data points in the relative pressure range between 0.05 and 0.2. black Super P (5 wt.%). These materials were mixed and dissolved in 5mL 1-methyl-2pyrrolidinone (NMP) and allowed to stir for 24hrs. The electrode slurry was cast on a copper foil and dried at room temperature for 24hrs followed by heating under vacuum at 90℃ for another 24hrs.The electrodes were punched with a 16mm punch; the cells were assembled in an Argon filled glovebox using LiPF6 as the electrolyte, a separator, small lithium foil, spacer, and wafer spring. The cells were allowed to rest for 48hrs before electrochemical testing. The cells were cycled Galvano statistically at various C-rates, 0.1 C, 0.2C, 0.5C, 1C, 2C, and 5C. Data was collected for 20 cycles between the cut-off voltage of 0.01V and 3V, respectively. All cycling tests were conducted on Arbin Instruments battery testing system.

Step-by-step process of electrode preparation and coin-cell assembly.

The electrochemical performance of graphite was measured as a baseline for testing and analysis. Figure 6.3a presents the capacity and CE for 10 charge/discharge cycles for graphite.

The theoretical capacity of graphite is 372 mAh/g (1, 7, 20). From the plot, the capacity at the first few cycles is lower due to the irreversible capacity losses from the formation of the solid electrolyte interface (SEI) layer. The SEI layer is a thin film (approximately 10nm thick) that is formed on the surface of the electrode after the first few charge/discharge cycles which only allows the passage of Li+ ion to and from the electrode while providing an electronic barrier to avoid further decomposition of the solvent. It consists of components from electrolyte such as lithium carbonate (Li2CO3) and Lithium Fluoride (LiF) (25, 26). The formation of SEI layer is very important and necessary for reversible and stable operation of the lithium-ion cell. The coulombic efficiency (CE) defined as the discharge to charge ratio also tells the same story as values greater than 100% are measured in these cycles. From the 4 th cycle, capacity begins to increase and stabilizes to 342 mAh/g with the CE value maintained at 100%. Figure 6.3b shows the charge/discharge curve for graphite. As noted in the introduction, lithium-ion is intercalated into graphite layers, and this is reflected in the low voltage plateau region of the voltage profile. This differs from the charge/discharge curve of the hard carbon where there is a sloping down instead of a plateau as a result of adsorption mechanism of Li-ion uptake into defects or edge sites (16). 6.4.2. Electrochemical Performance of Novolac Resin Modified by Templating Technique Figure 6.4a demonstrates the capacity as a function of the various C-rates for novolac,

RGO-N and GPL-N. The plot shows that capacity decreases with an increase in C-rate. The initial capacity at 0.1C for novolac was measured at 230 mAh/g which then decreased to 172 mAh/g at 0.2C and finally to 120 mAh/g at 5C. For RGO-N, the capacity was measured at 159 mAh/g at 0.1C, 117 mAh/g at 0.2C and further decreased to 90 mAh/g at 5C while GPL-N had a capacity of 179 mAh/g at 0.1C and decreased to 75 mAh/g at 5C. At higher C-rates, a higher current is given to the cell and the cell is required to charge/discharge at a faster rate of time. If the transfer of ions is not fast or efficient enough for power dissipation, capacity decreases as is the case with the trend

.5 depicts the capacity trend with number of cycles at a fixed charge/discharge current density 37mAh/g corresponding to 0.1C rate for 30L/70P, 50L/50P, and 70L/30P. As seen in the capacity plot for 30L/70P, the first cycle capacity begins at 270 mAh/g but drops to 105 mAh/g at the 2 nd cycle sample (Figure 6.5a). In subsequent cycles, capacity stabilizes, and a final capacity of 78 mAh/g was measured. Overall, about 70% of the initial capacity was lost. In the same trend, a higher initial capacity, 500 mAh/g, is measured for the 50L/50P sample as shown in Figure 6.5b. This is an interesting find as it shows that a higher lignin content results in a higher initial capacity. However, due to large irreversibility, a huge capacity is lost at the 2 nd cycle (195 mAh/g). No major loss in capacity is observed after the formation of the SEI layer in subsequent cycles and a final capacity of 155 mAh/g is measured which is about twice the final capacity of the 30L/70P sample. Lastly, the 70L/30P sample (Figure 6.5c) measured a high initial capacity of 374 mAh/g and a final capacity of 137 mAh/g. The first cycle efficiency at 0.1C for 30L/70P, 50L/50P and 70L/30P were measured at 37%, 40%, and 39%, respectively, Table 6.3. The large irreversibility phenomenon can be explained based on the properties of hard carbons. Hard carbons, specifically biomaterials are known for their large surface area which provides an advantage for a high initial capacity of the anode; however, a disadvantage to this large surface area is also substantial loss of that capacity due to large consumption and trapping of Li-ions during the formation of the SEI layer in the first few cycles leading to a low capacity as seen in the lignin samples. Also worth noting, coulombic efficiency for all samples stabilized to 100% after the first few cycles. 50L/50P and (c) 70L/30P.

Factors such as surface area (SA) and porosity influence electrochemical performance of the hard carbon material as anodes and were analyzed to understand the behaviors of the lignin and novolac samples. The nitrogen adsorption/desorption isotherm of 30L/70P, 50L/50P and 70L/30P are displayed in Figure 6.7. The shape of the isotherm for the samples reflects a type IV samples is detailed on the table. The result of the pore size distribution gives good insight to the exhibited electrochemical performance of each sample.

Studies have shown rate capacity is proportional to micropores and rate retention is proportional to mesopores (30, 31). The increased micropore volume of the 50L/50P and 70L/30P

samples facilitates a high-rate capacity especially at lower C-rates (i.e., 155mAh/g and 135 mAh/g, respectively at 0.1C). This is the opposite case for 30L/70P with much lesser microspores underscoring its low-rate capacity (i.e., 78 mAh/g at 0.1C). Furthermore, the presence of bimodal distribution of mesopores in the 50L/50P and 70L/30P samples could provide a good reservoir which facilitates shorter diffusion path for rapid ion transfer and creates better accessibility (mesopore transfer porosity) (32). This is a key characteristic that aids the rate retention performance even at higher C-rates. Even though the 30L/70P sample had equivalent mesopore volume as the 50L/50P sample, its lack or lessened micropores volume affects its capacity and does not provide the optimum pore size distribution conditions for high performance. On the other hand, the 50L/50P presents the optimum pore size distribution for good capacity and rate retention performance. The SA for all samples is also displayed on Table 6.4. As expected, all materials have a very high SA. 30L/70P sample has a SA of 345 m 2 /g while 50L/50P and 70L/30P samples have SA of 671m 2 /g and 673 m 2 /g, respectively . The measured SA confirms the reason for the large irreversibility and low FCE. Due to the samples' large SA, there are lots of exposed edge sites for the irreversible trapping of Li-ion and the formation of thick SEI layer. Table 6.4. BET surface area and BJH pore size distribution analysis for lignin samples.

Phenolic resin, novolac, demonstrated different electrochemical characteristics than bio precursor, lignin. For instance, the novolac samples had higher FCE at 0.1C (Table 6.2). This is due to fewer edge functionalities as a result of heat treatment (novolac samples were heat treated at 2500℃, chapter 3). However, the clogged pores by additives (RGO and GPL) inhibited sites for Li-ion storage and diffusion resulting in lower reversible capacity than pure novolac. The BET measurement of the novolac produced a very low surface area due to non-equilibration of the nitrogen gas at the given related pressure and time duration of experiment. But, phenolic resins are known for their microporosity (33, 34) and would exhibit characteristics of a type I. This isotherm is characteristic for microporous materials showing micropore filling but no multilayer Sample BET Surface (m 2 /g) Total pore volume (cm³/g) Mesopore volume (cm³/g) Micropore volume (cm³/g) meso to micro pore ratio 30L/70P 345 0.179 0.0381 0.141 0.27 50L/50P 671 0.289 0.0366 0.252 0.15 70L/30P 673 0.325 0.0629 0.262 0.24

adsorption (27, 28). The microporosity of the novolac samples underscore higher capacity at 0.1C, although the lack of mesopore/macro-pore structure would affect the rate retention performance at higher C-rate.

To test this, the capacity rate retention was calculated for novolac and 50L/50P sample (Figure 6.8). The percent capacity retention is used as a predictor of long-term life and performance in lithium-ion cells. From the plot, the capacity retention of novolac at 5C was 52%, but recorded higher at 68% for the 50L/50P sample. This shows that 50L/50P retained higher capacity at a higher C-rate compared to novolac. While Novolac has the advantage of more microporosity which is accessible at low C-rates, therefore, achieving higher capacity at 0.1C, the drawback to this is that as rate increases, lithium-ion diffusion in micropores is significantly reduced. Therefore, only 52% of the original capacity is retained at higher c-rate. The opposite is true for lignin. There is better accessibility because of the presence of micropores and transport bimodal mesopores. The transport mesopore aids lithium-ion diffusion at high c-rate causing it to retain higher percentage of the initial capacity (68%). The slow electrolyte transport in a microporous system is the reason why micropores in carbon materials are unable to yield good rate performance, especially at high current density, but the existence of mesopores is favorable for high-rate performance (35). Better accessibility (from larger surface area) could also mean there is large irreversibility due to the formation of thicker SEI layer. The thick SEI layer also influences the lower specific capacity of the 50L/50P sample compared to novolac. A solution to minimize SEI thickness and improve performance is to add electrolyte additives to make thinner the SEI layer. Such additives include vinylidene carbonate and fluoroethylene carbonate (FEC). In the next section we discuss in-depth solutions and strategies to optimize electrochemical performance of hard carbons based on results gotten in this work.

For the hard carbons tested in this work, it was demonstrated that factors such as texture/porosity and even synthesis conditions have significant effect on their electrochemical performance. These are challenges that affect the capacity, rate capability retention and even first cycle efficiency of the battery cell. Therefore, it is necessary to optimize the conditions and properties of the hard carbon anodes for better performance. The following are considerations for optimization and fabrication strategies to improve performance and meet satisfactory standards for practical applications.

1. Engineering porosity -The mechanism of lithium-ion storage in hard carbons is centered around adsorption of Li-ions in pore and defective sites, therefore the porosity, defects and even interlayer spacing are key to enhancing electrochemical performance (20, 36).

Engineering porosity by tailoring the porous network architecture with hierarchical pore structure (micro, meso and macro pores) creates the optimum conditions for efficient Liion storage leading to high specific capacity and rate capability (35).

2. Optimizing/Regulating Surface Area -A large surface area has an adverse impact by causing larger SEI formation and irreversible consumption of Li-ions during this process.

This also has a negative impact in the FCE due to exposed edge sites. Decreasing surface area is an effective method to reduce irreversible capacity loss and improve FCE (37). Some studies have suggested a low surface area of less than 40 m 2 /g, or even less than 10 m 2 /g is preferable to suppress side reactions at charging and discharging and improve performance (27).

3. Improving Li-on Transport by Element Doping -Inefficient transport of lithium-ions within the cell inhibit electrochemical performance and life cycle. Modifications to enhance charge-carrier-transport kinetics is necessary for better Li-ion transport (38, 39).

Substituting carbon atoms of the anode with heteroatoms such as nitrogen, boron, phosphorous, etc. that have electron acceptor or donor capability can change bond distances and incite disorder/defect in the carbon structure which can aid Li-ion transport through modulated band structure of the graphene layers (40, 41). As an example, studies have found that N-doped carbons can also give high-rate capabilities, along with long-term cycling stability due to improvement in the electrical conductivity and wetting behavior of the electrode/electrolyte contact (42). Doping in the form of single or multi-element is a welcome approach to improving properties and electrochemical behavior of the anode materials (43, 44). Implementing these strategies can improve electrochemical properties of hard carbons and lead to high performance for practical applications.

This dissertation aimed at developing techniques and expanding the knowledge on inducing the graphitization of hard carbons (non-graphitizing precursors), contributing valuable insights and data to the influence of processing parameters, 2D-template additives, and precursor degradation chemistry. The challenges faced in graphitizing hard carbons including the use of catalysts that caused impurities that are detrimental to the final product, and the benefits of using hard carbons as alternative precursors for energy storage applications, underscore the importance of understanding and applying techniques to graphitize hard carbons. This research is focused on the templating-induced graphitization of hard carbons using 2-dimensional materials and testing modified hard carbons in energy storage applications.

The scope of the investigation is expanded between chapters 2 through 6. In Chapter 2, the effect of carbonization pressure on the graphitization of hard and soft carbons was explored. The findings reveal that pressure generated during carbonization can lead to an enhanced crystalline nature/improved crystallinity in hard carbons. On the other hand, carbonization pressure did not impact the graphitization level of soft carbons. Future research should explore the impact of deliberately applying a range of fixed external pressure during carbonization. The results gotten from that test could create more knowledge on how pressure affects growth of crystallite size which is beneficial in the performance of hard carbons as anode material in energy storage applications.

In chapter 3, the study delved into the templating-induced graphitization of phenolic resin, novolac using graphene oxide additives. The results showed that the addition of GO additives with oxygen content up to 15.4 at.(%) induced graphitization and led to improved graphitic quality of the novolac matrix. In contrast, the addition of GO additive of twice or more oxygen content ≥ 30.8 at.(%) to the novolac matrix (GO-N) led to poor graphitic quality. This finding led to the working hypothesis that these GO derived additives act as chemical and physical templates to induce the graphitization of the novolac matrix. Chemical templating occurs as oxygen groups of the GO derived additives form reactive radicals that provide bonding to the decomposing novolac matrix. Also, aromatic rings formed during carbonization of novolac align with the sp 2 network of the two-dimensional graphene oxides, thereby aiding in the formation of layered graphite material, a process called physical templating. Future research can delve deeper into the interplay of template functionalities e.g., phenols vs. carboxyl vs. carbonyls groups or other non-oxygen functionalities such as sulfur or nitrogen groups, optimize processing conditions, and explore complementary characterization techniques to fully elucidate the mechanisms governing the precise control over graphitization processes. The insights gained from this study pave the way for tailoring the synthesis of advanced carbon materials with desired properties for diverse applications, ranging from thermal management to structural reinforcements.

The study then expanded into inducing the graphitization of a hard carbon bio precursor, lignin, using graphene oxide additives. The results revealed a 50% increase in the crystallite size (La) and between a 60 to 80% increase in crystallite height (Lc) caused by the addition of the GO additives to the lignin matrix. However, we find that oxygen content in the GO additives do not play a role in the degree of graphitization. Moreover, the method of processing (direct mixing, acid digestion and hydrothermal treatment) produced similar results which emphasizes the lack of degradation of the lignin matrix influence how the GO additives interact and induce graphitization.

Extension of this research should focus on exploring processing methods to accelerate the degradation bio precursors such as lignin into simpler components (lower molecular weight compounds) which would be beneficial for its interactions with templating agents during heat treatment. Other bio precursors such as cellulose, hemicellulose, corn starch, paper pulp can be used as carbon matrices for template graphitization.

To differentiate the physical and chemical templating mechanism effect, chapter 5 explored graphitization of novolac using 2D-tempaltes in the form of graphene and hexagonal boron nitride with minimal to zero oxygen groups. Results revealed that hBN did not induce graphitization of novolac and this was attributed to the absence of a sp² framework in hBN that was unable to provide the crucial π-π interactions with the aromatic rings of the matrix. In contrast, the graphene additive led to an improved graphitization of the novolac. Comparing the results with of graphene oxide induced templating in chapter 2, it was surmised that both mechanisms are operative; while physical templating offer control over long-range order in the form of crystallite height due to π-π interactions with the aromatic rings of the matrix, chemical templating contributes to carbon reorganization and lateral growth extent due to radical edge site bonding from leaving oxygen groups. Future research can explore comparative ReaxFF atomistic simulations to complement the observed experimental findings. The use of other additives such as carbon nanotubes (CNTs) can be explored individually or together with graphene/graphene oxide.

Lastly, the study examined the application of modified hard carbons, lignin (modified by emulsion polymerization) and phenolic resin (modified by templating), as anodes in lithium-ion batteries. Electrochemical measurements revealed novolac samples had higher first cycle efficiency and specific capacity at 0.1C rate compared to the lignin samples. However, the lignin samples had higher capacity retention across higher C-rates. Porosity and textural characterizations gave insights to the observed trends by identifying differences in pore size distribution (micro, meso and macro pores) which influenced electrochemical performance. Based on the results, the lignin sample (50L/50P) consisted of a bimodal distribution of mesopores which is favorable for lithium-ion diffusion at high c-rates while novolac exhibits microporosity which aids high-capacity performance at lower c-rates. Extension of this research should focus on strategies to improve hard carbon anode performance such as engineering hierarchical porosity, controlling particle size (by jet milling etc.), optimizing surface area and even using methods like elemental doping.