Processing 2D Nanomaterials: Study of Colloidal and Chemical Stability of Boron Nitride and Ti3c2tx Nanosheets

PROCESSING 2D NANOMATERIALS: STUDY OF COLLOIDAL AND CHEMICAL STABILITY OF BORON NITRIDE AND TI3C2TX NANOSHEETS

A Dissertation by TOUSEEF HABIB

Submitted to the Office of Graduate and Professional Studies of Texas A&M University in partial fulfillment of the requirements of the degree of DOCTOR OF PHILOSOPHY

Chair of Committee, Micah J. Green Committee Members, Jodie L. Lutkenhaus Miladin Radovic Mustafa Akbulut Head of Department, M. Nazmul Karim

May 2019

Major Subject: Chemical Engineering

Copyright 2019 Touseef Habib ABSTRACT Two-dimensional (2D) nanomaterials have been hotly investigated since the identification of graphene in 2004. With so much research focus it has been difficult to translate promising laboratory results to commercial success. The main reason for this is the processing of such nanomaterials. Processing at a nanoscale is challenging because the materials at that scale either aggregate to their parent structure or chemically degrade, rendering them useless. In this work, liquid processing advancement of two novel nanosheets are presented; boron nitride nanosheets

(BNNSs) and Ti3C2Tx MXenes. BNNSs are difficult to process at the nanoscale because they tend to aggregate. We have demonstrated the viability of co-solvents in processing BNNSs to obtain high yields while still maintaining a high quality BNNSs. Ti3C2Tx MXenes on the other hand are easier to process but they oxidize. The chemical degradation was examined various media and suggestions made to store Ti3C2Tx MXenes in solid media to decrease the rate of oxidation.

Additionally, we demonstrated a never-before-seen property of Ti3C2Tx MXenes: their propensity to rapidly heat under radio frequency fields. We explored how polymer/ Ti3C2Tx composite film architecture and composition affect both conductivity and RF responsiveness. This new property can be exploited for various RF-based applications where localized heating is desired.

ACKNOWLEDGEMENTS Majority of the credit for this PhD goes to my parents, they always pushed me to go further and never settle. With their unconditional love, support, and belief in me (even when I did not) inspired me to carry on with my studies. To my brothers, thank you for the initial push and for the constant support over the last five years; you both always kept my spirits up.

I am grateful to my research group members, past and present. Fahmida Irin, Dorsa Parviz, Yueyi Xu, and Brandon Sweeney. They welcomed me in the group with open arms. To Smit Shah, Xiofei Zhao, Nutan Patil, Muhammad Anas, and Wanmei Sun; thank you for your contributions, I could not have finished my work without your help and backing. I also want to thank Thomas Achee and Eliza Price, the wondergrads who labored for me in the laboratory and helped me get my experiments done. I also want to acknowledge Pritishma Lakhe, Martin Pospisil, Kai Morikawa, Joseph Gerringer, Saerom Yu, Morgan Plummer, and the rest of the Green group for their input and contributions.

To Dr. Miladin Radovic and Dr. Jodie Lutkenhaus, thank you. As committee members and collaborators, your ideas and feedback has helped me immensely. Thank you Dr. Mustafa Akbulut, I appreciated you input during the committee meetings. I also want to acknowledge Hyosung An, Evan Prehn, and the other members of the Radovic and Lutkenhaus research groups; it was exciting collaborating with you guys.

I am eternally grateful to my advisor Dr. Micah Green. Despite being burdened with the grueling schedule of academia, he always made time for his students. He was instrumental in my growth as a professional and as a researcher. Always leading by example, he set a high standard for what it means to be a scientist and a family man.

I want to thank Allah, The Creator of all. None of this would be possible without him. “O my Lord, increase me in knowledge” Quran, 20:114.

iii

CONTRIBUTORS AND FUNDING SOURCES This work was supervised by a committee consisting of Professor Micah Green (advisor), Jodie Lutkenhaus, and Mustafa Akbulut from the Department of Chemical Engineering and Professor Miladin Radovic from the Department of Material Science. Some data from Chapter 2 were provided by Professor Rajesh Khare from Texas Tech University’s Department of Chemical Engineering. All other work was completed by the student independently. My research was supported by Artie McFerrin Department of Chemical Engineering with their graduate student assistantship. Funding for this work was provided by the National Science Foundation under CAREER award CMMI-1253085 as well as a 2014 DuPont Young Faculty Award. Additional funding was provided by the U.S. National Science Foundation (Grant CMMI- 1760859) and TAMU Energy institute.

iv TABLE OF CONTENTS Page

ABSTRACT ...... ii

ACKNOWLEDGEMENTS ...... iii

CONTRIBUTORS AND FUNDING SOURCES ...... iv

TABLE OF CONTENTS ...... v

LIST OF FIGURES ...... vii

LIST OF TABLES ...... xi

CHAPTER 1 – INTRODUCTION ...... 1

Boron nitride nanosheets (BNNSs) ...... 1 Synthesis & Processing ...... 2 Applications ...... 5 Challenges ...... 9

Ti3C2Tx MXene nanosheets ...... 9 Synthesis & Processing ...... 10 Applications ...... 12 Challenges ...... 20

CHAPTER 2 – COSOLVENTS AS LIQUID SURFACTANTS FOR BORON NITRIDE NANOSHEET (BNNS) DISPERSIONS ...... 22

Summary ...... 22 Introduction ...... 22 Materials & Methods ...... 24 Results & Discussion ...... 28 Conclusion ...... 41

CHAPTER 3 – OXIDATION STABILITY OF TI3C2TX MXENE NANOSHEETS IN SOLVENTS AND COMPOSITE FILMS ...... 42

v Summary ...... 42 Introduction ...... 42 Materials & Methods ...... 44 Results & Discussion ...... 47 Conclusion ...... 61

CHAPTER 4 – TI3C2TX MXENE/POLYMER COMPOSITES HEAT IN RESPONSE TO RADIO FREQUENCY (RF) FIELDS ...... 62

Summary ...... 62 Introduction ...... 62 Materials & Methods ...... 63 Results & discussion ...... 66 Conclusion ...... 77

CHAPTER 5 – CONCLUSION ...... 78

Boron nitride nanosheets ...... 78

Ti3C2Tx MXene nanosheets ...... 79

NOMENCLATURE ...... 82

REFERENCES ...... 83

APPENDIX – TEMPLATE-FREE 3D TITANIUM CARBIDE (TI3C2TX) MXENE PARTICLES CRUMPLED BY CAPILLARY FORCES ...... 102

Summary ...... 102 Introduction ...... 102 Materials & Methods ...... 103 Results & Discussion ...... 107 Conclusion ...... 114

vi LIST OF FIGURES Page Figure 1-1; Ball milling hBN to BNNSs (reprinted from Nano Letters)17 ...... 2

Figure 1-2: Molecular dynamic simulation of PVP attaching on the surface of BNNS (Reprinted from Phys. Chem. Chem. Phys.)28 ...... 3

Figure 1-3: CVD growth of multi-layer BNNSs (Reprinted from Nature Comm.)42 ...... 4

Figure 1-4: Boron nitride barrier protecting copper surface from microbial corrosion (reprinted from ACS Nano)47 ...... 6

Figure 1-5: Photo of BNNS/PVA composite (reprinted from Nanoscale)52 ...... 8

Figure 1-6: Possible MAX phases (reprinted from Thin Solid Films)59 ...... 9

Figure 1-7: Potential MXene compositions (reprinted from Nature Reviews Material)71 ...... 11

Figure 1-8: Synthesis of Ti3C2Tx from a) Ti3AlC2 (parent MAX phase), to b) Ti3C2Tx clay, then c) to Ti3C2Tx nanosheets (reprinted from Technol.-Mat. Sci. Edit, J. Mater. Chem. A, Chem. Commun)77,78,79 ...... 12

Figure 1-9: TEM images of the CdS/Ti3C2Tx cauliflower-like structure. The scale bars are 200 nm (reprinted from Nature Communications)82 ...... 13

Figure 1-10: Water flux of water and salt solutions with different cation charge through the 86 Ti3C2Tx/PVDF membranes (reprinted from Phys. Chem. Letters) ...... 15

Figure 1-11: Proposed shielding mechanism; reflection, internal reflections, and dipole interactions (reprinted from Science)87 ...... 16

Figure 1-12: Comparison of Ti3C2Tx nanosheets and GO with a) E. coli, and b) B. subtillis (reprinted from ACS Nano)89 ...... 17

Figure 1-13: LBL assembly process of Ti3C2Tx and PDAC (reprinted from Science Advances)98 ...... 19

Figure 2-1: The relationship between absorbance (measured at 400 nm) and concentration allows us to calculate extinction coefficient ...... 26

Figure 2-2: a) TEM image of co-solvent dispersion. b) TEM image of solvent exchanged dispersion ...... 30

vii t-butanol yielded a “frozen” dispersion; the BNNSs were kinetically trapped after freezing. c) Redispersion in co-solvent was successful as the BNNSs flakes “dissolved” into solvent...... 32

Figure 2-4: a) TEM image of redispersed BNNSs in pure t-butanol. b) TEM image of redispersed BNNSs in co-solvent mixture. Both TEM images reflect the quality of the dispersions...... 33

Figure 2-5: FTIR-ATR of vacuum-filtered, solvent-exchanged dispersion, pure water, pure t- butanol, vacuum filtered co-solvent dispersion, and freeze dried BNNSs powder ...... 34

Figure 2-6: End configurations of one replica structures for (a) ethanol-BNNS-water system, (b) t-butanol-BNNS-water system, (c) 1-butanol-BNNS-water system, and (d) 1-hexanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity...... 35

Figure 2-7: End configurations of two different replica structures for ethanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity...... 36

Figure 2-8: End configurations of two different replica structures for t-butanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity...... 36

Figure 2-9: End configurations of two different replica structures for 1-butanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity...... 36

Figure 2-10: End configurations of two different replica structures for 1-hexanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity...... 37

Figure 2-11: Radial distribution functions between (a) carbon atoms of alcohols and BNNS atoms, (b) oxygen atoms of alcohols and BNNS atoms, and (c) oxygen atoms of water molecules and BNNS atoms. RDFs for each system are shown by the following colors: ethanol-BNNS-water system (dotted dark-green line), t-butanol-BNNS-water system (solid red line), 1-butanol-BNNS-water system (dash-dot dark yellow line), and 1-hexanol- BNNS-water system (dashed dark-blue line)...... 38

Figure 2-12: Cluster size probability distributions of (a) ethanol molecules (dark-green bars), (b) t-butanol molecules (red bars), (c) 1-butanol molecules (dark yellow bars), and (d) 1-hexanol molecules (dark-blue bars) in water...... 40

Figure 2-13: Simulated snapshots of (a) ethanol clusters in water, (b) t-butanol clusters in water, (c) 1-butanol clusters in water, and (d) 1-hexanol clusters in water. Water molecules are not shown for the sake of clarity...... 40

Figure 3-1: Overview of experimental procedure: After synthesis of Ti3C2Tx MXene nanosheets from parent MAX phases, samples are dispersed in various media ...... 47

viii

Figure 3-2: Conductivity vs time measurements with error bars (standard deviations) for a) Ti3C2Tx films in air, b) films made from aged Ti3C2Tx dispersed in ice, and c) Ti3C2Tx/PVA composite films. The decrease in conductivity is indicative of increasing oxidation...... 48

Figure 3-3: MXene clay aged in various organic solvents for 8 months. After 8 months, the MXenes were vacuum filtered into films then their conductivity was measured. Films made from aged MXenes in acetone and acetonitrile possessed measurable conductivity than other solvents...... 49

Figure 3-4: Decrease in conductivity with increasing TiO2 in MXene clay that were aged in various solvents. 1 D is one day, 1 M is 1 month, and 8 M is 8 months...... 50

th th Figure 3-5: Ti3C2Tx colloidal dispersion in water on the 0 day and the 14 day ...... 51

Figure 3-6: UV-vis spectra of Ti3C2Tx/water dispersion at 14 days. The sample has been diluted 1/10. The absorbance at 580 nm is 0.236 and the extinction coefficient at the same wavelength is 1167.2 ml mg-1 m-1. Using Beer-Lambert’s Law, the concentration was calculated to be 0.202 mg ml-1 ...... 52

Figure 3-7: XPS of a) vacuum filtered film drawn from fresh (Day 0) Ti3C2Tx colloidal dispersion; the TiO2 content in this sample was 11.06% and b) vacuum filtered film drawn th from the same Ti3C2Tx colloidal dispersion on the 14 day; the TiO2 content was 55.80% ...... 53

Figure 3-8: (top) Conductivity as a function of time for samples stored in humid and dry 4+ environments. (bottom) Ti (TiO2) content of RH 80% sample with time...... 54

Figure 3-9: a) Ti3C2Tx in acetone; Ti3C2Tx powder sedimented to the bottom b) Ti3C2Tx in acetone after vortex mixing which form a temporary colloidal solution; this was vacuum filtered to obtain vacuum filtered film c) Ti3C2Tx in acetonitrile; Ti3C2Tx powder sedimented to the bottom d) Ti3C2Tx in acetonitrile after vortex mixing which form a temporary colloidal solution; this was vacuum filtered to obtain vacuum filtered film ...... 56

Figure 3-10: Master curve for normalized conductivity variation of films made from MXenes dispersed in various media over time...... 58

Figure 3-11: UV absorbance of various layer pair (LP) LBL films: a) as-prepared, b) after 2 weeks, c) after 1 month (four weeks), and d) UV absorbance (at 770 nm) of all of LP LBL films at various times ...... 59

Figure 3-12: Change in conductivity of MXene film with increasing UV exposure time ...... 60

Figure 4-1: a) SEM image of delaminated Ti3C2Tx MXene flake, and b) cross-sectional SEM image of neat MXene film ...... 66

Figure 4-2: Photo of a) the tiny capacitor and the sample; the tiny capacitor generates the RF field, and b) the FLIR camera and the RF field generator ...... 67

Figure 4-3: a) schematic of the RF apparatus and the Ti3C2Tx MXene composite sample, b) same schematic but with the RF fringing field turned on which heats the sample (observed with FLIR camera), c) FLIR image of 50 wt.% composite, and d) plot of the heating rate vs frequency to determine the resonant frequency (highest heating rate) of each sample...... 68

Figure 4-4: Raw data of RF sweeps. The RF fields were turned on at 3W for 2 seconds then turned off (0W) for 13s for every frequency from 1-150 MHz ...... 69

Figure 4-5: Raw data of temperature vs time for a) Day 0 samples at 1W, and b) Day 0 samples at 3W ...... 70

Figure 4-6: Raw data of temperature vs time for a) Day 30 samples at 1W, and b) Day 30 samples at 3W ...... 71

Figure 4-7: conductivity at each composition for Day 0 (fresh), b) temperature vs time graph for 10 wt.% composite on Day 0 and at 1W, c) the rise in temperature for each Day 0 sample (at 1W and 3W) vs composition, d) the rise in temperature for each Day 0 sample (at 1W and 3W) vs conductivity ...... 73

Figure 4-8: a) conductivity of Day 0 and Day 30 samples, and b) temperature rise for Day 30 samples (at 1W and 3W) vs composition. Tmax was the max temperature reached during the span of 12 seconds...... 74

Figure 4-9: Ti3C2Tx’s titanium X-ray photoelectron spectroscopy (XPS) of a) Day 0 sample where the TiO2 content is roughly 4.5% and b) Day 30 sample where the TiO2 content is roughly 30% ...... 75

Figure 4-10: Thermal cycling on Day 0 at 3W of a) 5 wt.% sample, b) 10 wt.% sample, c) 50 wt.% sample. The first five cycles of d) 5 wt.% sample, e) 10 wt.% sample, f) 50 wt.% sample ...... 76

LIST OF TABLES Page Table 1-1: Summary of advantages and disadvantages of different BNNSs synthesis/processing techniques...... 5

Table 1-2: Mechanical test results of Ti3C2Tx/UHMWPE composites (reprinted from Materials and Design)100 ...... 20

Table 2-1: Concentration, ζ potential, and average lateral size of all dispersions ...... 29

Table 2-2: Concentration, ζ potential, and average lateral size of rehydrated dispersions ...... 33

Table 3-1: Conductivity (S m-1) of films made from aqueous MXene dispersions with varying dispersion age ...... 51

Table 3-2: Ti3C2Tx film conductivity vs UV exposure time data ...... 60

Table 4-1: Conductivity of the samples by composition on Day 0 and Day 30 ...... 72

Table 4-2: Comparison of RF heating of the fresh (Day 0) and aged (Day 30) set of samples at 1W and 3W of power ...... 75

Table 4-3: Conductivities before and after thermal cycling on Day 0 at 3W of 5 wt.%, 10 wt.%, 50 wt.% sample. The % drop in conductivity is not very significant ...... 76

CHAPTER 1 – INTRODUCTION Interest in 2D materials have exploded since the isolation of two-dimensional (2D) graphene flakes in 2004 by Novoselov and Geim.1 They isolated graphene with scotch tape, peeling graphene away from the parent graphite material. Upon its isolation, graphene generated immense buzz due to its material properties. It has been reported a single graphene sheet possesses an estimated specific surface area of 2630 m2 g-1, sheet resistance of 10−6 Ω cm, thermal conductivity (theoretical) as high as 5300 W m−1⋅K−1, melting point as high 6000 K (theoretical), measured tensile strength of 130 GPa, and Young’s Modulus of 1 TPa.2-6 As a result, graphene is a promising material for use in wide range of applications from membranes to catalyst supports. Even with all the great material properties of graphene, its use remains very limited due to processing difficulties in producing high quality graphene (single to a few layers) in a scalable and cost effective method. The discovery of graphene also led to exploration of other 2D materials like boron nitride, molybdenum disulfide, tungsten disulfide, phosphorene, and MXenes. Just like graphene, all these 2D materials face similar challenges in production. In this dissertation, I will cover processing aspects of two 2D nanomaterials: boron nitride and Ti3C2Tx MXenes.

Boron nitride nanosheets (BNNSs) A single layer of hexagonal boron nitride (hBN) consists of covalently bonded boron and nitrogen atoms (in a hexagonal ring structure similar to graphene); each sheet is held together by van der Waals forces originating from the partial charges of boron and nitrogen.7 Isolating single sheets (or even few layered sheets) of hBN is the goal.

Individual boron nitride sheets (BNNSs) at the 2D level possess tremendous properties (optical, mechanical, and thermal) that can be exploited to create devices and/or functional materials with a wide range of applications. The bandgap of boron nitride is roughly 5.9 eV, making it electrically insulating.8 It has been reported that a single layer boron nitride starts oxidizing at 700 ˚C but it can be stable up to 850 ˚C, making it a good candidate for coatings geared towards oxidation and corrosion resistance. 9-11 The thermal conductivity of a single layer BNNS is calculated to be 600 W m-1 K-1.12 The experimentally measured Young’s modulus of a few layered BNNS is 1.16 TPa, which rivals that of graphene while the theoretical Young’s modulus for a single layer BNNS is calculated to be in the range of 0.71-0.97 TPa.6,13 Unfortunately, obtaining few layer BNNSs in a scalable manner remains a challenge and when

1 BNNSs are isolated, they aggregate back (restack), losing their 2D properties (aggregated BNNSs have properties similar to their starting material).

Synthesis & Processing Pacilé et al. first reported isolation of few layered BNNSs in 2008 using the same mechanical cleavage technique (scotch tape method) used by Novoselov and Geim in their discovery of graphene.14 This method has very low yields and is not scalable. A more efficient and scalable mechanical exfoliation method was reported by Li et al., who used ball milling process to produce BNNSs. Ball milling shears apart the hBN layers creating few layered BNNSs. Even though promising, ball milling does create BNNSs with small flake sizes (~250 nm), limiting them for applications that require larger flakes.15 Although later studies improved the ball milling process by exfoliating BNNSs with larger flake sizes (high as 1.5 µm) and with exfoliation yield as high as 18% (Figure 1-1).16,17

Figure 1-1; Ball milling hBN to BNNSs (reprinted from Nano Letters)17

A more scalable method for BNNSs production is through liquid phase exfoliation. The starting material (hBN) is exfoliated in a liquid with the aid of mechanical energy, often using cavitation. Using water as the solvent is preferable because it is safe, abundant, and industrially friendly. However, BNNSs are hydrophobic and have a propensity to restack in water. It has been reported sonicating BNNSs in water causes OH functionalization on the edges of the nanosheet.8,18 Similarly, heat treatment of hBN in air yields hydroxylated boron nitride nanosheet.19 Treating hBN with hot stream of water vapor (850 ˚C) functionalizes the edges with OH groups making

2 them compatible with hydrogels.20 Other methods to make BNNSs soluble in water include: functionalization by Lewis base interaction, oxygen covalent functionalization, nitrene covalent functionalization, π–π stacking, and functionalization through phosphoric and sulfuric acid.21-26

Another method to disperse BNNSs in water is to use a surfactant. Bari et al. reported using polyvinylpyrrolidone (PVP), a polymer that physically adsorbs on the basal plane of the BNNS and prevents aggregation (Figure 1-2).27 We also demonstrated further processing BNNS (and other types of 2D nanosheets) into a crumpled 3D morphology by a spray drying method. The advantages of crumpled nanosheets are they are aggregation resistant, and are also easier to transport (easier to handle crumpled powder than liquid dispersion). Such 3D crumpled BNNS may potentially be used as encapsulation material and lubricant additives.28

Figure 1-2: Molecular dynamic simulation of PVP attaching on the surface of BNNS (reprinted from Phys. Chem. Chem. Phys.)28

BNNSs can be directly exfoliated in organic solvents as well. Prior studies have shown BNNSs can be dispersed in N-methyl-pyrrolidone (NMP), isopropanol (IPA), chloroform, dimethylformamide (DMF) and other organic solvents.29,30 Liu et al. altered the BNNS surface with polystyrene and poly (methyl methacrylate) to enhance the stability of BNNSs in various organic solvents. The polystyrene interacts with the BNNSs due to π–π stacking (a non-covalent

3 interaction due to π bonds), and the poly(methyl methacrylate) extends outward, forming a micelle around the sheet.31

Chemical vapor deposition (CVD) is a bottom up approach (unlike mechanical and liquid phase exfoliation which are top down approach); under specific conditions, high quality (single to few layer) BNNS can be grown on a substrate Figure 1-3). Successful CVD of BNNSs have been synthesized by gas phase precursors (BF3/NH3, BCl3/NH3, B2H6/N, and B10H14/NH3 mixture), 32-40 liquid precursors (B3N3H6, B3N3H3Cl3, and B3N3Cl6), and solid precursors (H3NBH3). To obtain stoichiometric layers of BNNSs, it is important to maintain the 1:1 stoichiometric ratio of

B/N; choosing B3N3Cl6 and H3NBH3 as the precursors makes it easier to maintain this ratio. Typical CVD substrates are Cu and Ni, but successful growth of BNNSs have been demonstrated on polycrystalline Pt and Fe.41,42 Song et al. used CVD to grow two to five layer of boron nitride film on various substrates. The grown film possessed bandgap of 5.5 eV and elastic modulus (measured using nanoindentation) was in the range of 200-500 N/m.43

Figure 1-3: CVD growth of multi-layer BNNSs (Reprinted from Nature Comm.)42

BNNSs can also be synthesized via solutions phase chemistry. Weng et al. reported chemical synthesis of BNNS by reacting g-C3N4 and B(OH)3 (thermal substitution) to synthesize water soluble hydroxylated BNNSs which displayed high biocompatibility with promising applications in drug delivery.44 Lei et al. reported a thermal process to synthesize porous BNNSs

4 with starting materials of boron trioxide, guanidine hydrochloride, and methanol. The synthesized porous BNNSs possesses pore volume of 1.09 cm3, BET surface area of 1.427 m2 g-1, and pore size distribution range of 20-100 nm.45

Table 1-1 lists out the advantages and disadvantages of different BNNSs synthesis procedures.

Table 1-1: Summary of advantages and disadvantages of different BNNSs synthesis/processing techniques

Process Advantages Disadvantages Mechanical Exfoliation  Can obtain few layered BNNSs  Surface defects  Can be done in atmospheric  Small flakes sizes conditions  Post processing aggregation  Easy to control processing steps concerns (milling time, ball size, temperature) Liquid phase exfoliation  Can be done in atmospheric  Use of harmful organic solvents conditions  Requires post processing steps  Easy to control processing steps to remove solvent or surfactant, (agitation time, agitation speed, Removal of solvent and/or choice of solvent/surfactant) surfactant leads to aggregation  The surfactant can be tailored for concerns. specific polymers (for composites)  Processing volume inversely  Scalable correlates with the quality of BNNSs obtained Chemical vapor deposition  Growth of high quality BNNSs  Done in harsh conditions (high (CVD) (single to few layers) temperatures and sometimes  No aggregation concerns non-atmospheric pressures)  Expensive  Difficult to scale up Solution phase chemistry  Easy to obtain precursors  Separation step is required to  BNNSs can be tailored with specific obtain BNNSs from unwanted functionalization and/or reaction products morphology for the target  Done in harsh conditions (high application temperatures)

Applications Due to their chemical stability and inertness, BNNSs are a promising material for corrosion and oxidation resistance. Sun et al. synthesized (liquid phase exfoliation) BNNSs/polyvinyl butyral (PVB) coatings for corrosion resistance. The researchers demonstrated through electrochemical testing that corrosion resistance performance increases with increasing BNNSs

5 content (the coating was applied on a copper surface). At 0.1 wt.% BNNSs, the reduction in corrosion rate was ~27 times than the pure PVB coating. At 1.0 wt.% BNNSs, the reduction in corrosion rate was roughly 67,000 times than the pure PVB coating.46 Chilkoor et al. demonstrated success in reducing microbial corrosion induced by a Desulfovibrio alaskensis G20, a sulfate reducing bacteria. A single layer BNNS grown on a copper surface by chemical vapor deposition displayed a corrosion inhibition efficiency of ~91%. Such coatings (CVD grown BNNS) are much thinner than commercially available polyaniline coatings used for similar purposes (Figure 1-4).47 Liu et al. coated nickel with CVD grown BNNS for oxidation resistance. The BNNS coating was able to protect the metal’s surface up to 1100 ˚C at oxygen rich environments. The authors claim the single to two layer BNNS coating is one of the thinnest material to survive such harsh conditions.47 Due to their chemical stability, BNNSs were also investigated as fillers for flame retardant coatings. Zulkurnain et al. demonstrated BNNSs incorporated coatings that burned at lower temperatures; a low of 138 ˚C compared to a high of 250 ˚C for coatings without BNNSs.48

Figure 1-4: Boron nitride barrier protecting copper surface from microbial corrosion (reprinted from ACS Nano)47

Due to BNNSs’ hydrophobicity, there is research interest in using BNNSs to prepare hydrophobic surfaces and utilizing BNNSs for water treatment applications. Pakdel et al. coated

(through chemical vapor deposition) Si/SiO2 substrate with 5 nm thick BNNSs. The contact angle of water on the coating was as high as 159˚, indicating a highly hydrophobic surface.49 Lei et al.

6 synthesized porous BNNSs using a thermal process for water treatment. BNNSs are hydrophobic but they adsorb and absorb various organic materials (oils, organic solvents, and dyes). The thermally synthesized porous BNNSs were demonstrated to absorb 33 times their own weight. Moreover, porous BNNSs are recyclable because of their chemical stability; the adsorbed materials can be burned off at high temperatures without degrading the BNNSs.45

BNNSs have been shown to be a good support substrate for other nanoparticles. Huang et al. obtained few layered BNNSs by liquid phase exfoliation in ethylene glycol. The exfoliated BNNSs were then used as a support for noble metal particles (Ag, Au, Pt, etc.). The hybrid BNNSs/metal catalyst demonstrated high catalytic activity for p-nitrophenol reduction.50 Duan et al. reported a carbon coated composite of Fe3O4 nanoparticles on a BNNSs support (as a substrate); this hybrid material was used as an anode for lithium ion batteries. The hybrid BNNSs/

Fe3O4/carbon anode displayed higher electrical performance than neat Fe3O4. The BNNSs support was theorized to act as a flexible support for the volume expanding Fe3O4 particles during the electrochemical process.51

One of the promising applications of nanomaterials is their use as fillers in polymer matrix to enhance the polymer’s material properties. Khan et al. developed composites with BNNSs and polyvinylalcohol (PVA) via solution casting (Figure 1-5); the composites displayed increased modulus and strength with only 0.1 vol.% BNNSs content. Beyond 0.1 vol.% though, the modulus and strength decreased.52 Wu et al. developed silane treated BNNS/styrene-butadiene rubber (SBR) composites using two-roll milling process. The silane interacts with both BNNSs and SBR, leading to better dispersion of BNNSs in SBR. A composite with 10.5 vol.% silane treated BNNSs was 253% more thermally conductive than neat SBR. It was also shown that BNNS/SBR composites without silane possessed the lowest thermal conductivity because of agglomeration of BNNS within SBR.53 Xie et al. melt processed BNNSs with poly (ethylene terephthalate) (PET) to develop composites with good oxygen barrier properties. A composite with 0.017 vol.% BNNSs displayed 42% decrease in oxygen permeability compared to neat PET.54 Su et al. functionalized BNNT/BNNS by 1-pyrenebutyric acid (PBA) and dispersed it in an epoxy matrix to develop composites with high thermal conductivity and high electrical resistance (properties desired for thermal interface materials). The BNNT/BNNS epoxy composites (2 wt.%) possessed thermal conductivity of 0.47 m−1 K−1 and an electrical resistivity of 9.3 × 1016 Ω cm, higher than the neat

7 epoxy.55 Yan et al. developed an epoxy composite with 1 vol.% BNNT/BNNS fillers which were non-covalently functionalized by pyrene carboxylic acid. The epoxy composite displayed an increase of 95% in thermal conductivity and an increase of 57% in Young’s Modulus while still maintaining similar resistivity to that of the neat epoxy.56 Hu et al. reported preparation of BNNS 3D hydrogel (hydrogels are 3D polymeric network with high water content). Due to their mechanical properties, they have diverse applications as a soft material. Nanomaterial are added to hydrogels to enhance the crosslinks of the polymer to strengthen the network, which in turn enhances mechanical properties of the hydrogel. Molten citric acid was used to treat hBN and produce water soluble BNNSs. The stability of BNNSs in water was due to generation of NH2 and OH groups on the basal plane of the nanosheets rather than linkage of citric acid molecules on the surface. The treated soluble BNNSs was used to synthesize a hydrogel (95 wt.% water) with polyacrylamide. The hydrogel displayed elongation and compression of >10,000% without breakage.57

Figure 1-5: Photo of BNNS/PVA composite (reprinted from Nanoscale)52

Challenges BNNSs have numerous promising applications, but processing remains a challenge. There is a tradeoff between obtaining high quality BNNSs and processing at large scales. Adding dispersants is a method to obtain high quality BNNSs while maintaining a high scale of production, but the amount of dispersant required limits the end application. Therefore, it is important to find a processing technique that can yield high quality BNNSs while maintaining a high scale of production without restricting the end application.

Ti3C2Tx MXene nanosheets MXenes are materials derived by etching its parent MAX phase. The MAX phase consists of early transitional metal (M), group 13/14 element (A), and carbide or nitride (X). The chemical 58 formula for MAX compounds are Mn+1AXn, where subscript n is a number from 1-3. MAX phase compounds are characterized as metallic-ceramics because of their high electrical and thermal conductivity (like metals) while still maintaining resistance to oxidation and thermal shock (like ceramics).58 Figure 1-6 shows the possible MAX phases.59

Figure 1-6: Possible MAX phases (reprinted from Thin Solid Films)59

Discovery of MAX phase compounds were first reported in the 1960s by Nowotny, but pioneering work on MAX phase occurred in the 1990s, led by Barsoum and El-Raghy when they 60 successfully synthesized and characterized Ti3SiC2; the first phase pure MAX compound. Since then, interest in MAX compounds have grown due to their material properties and their material diversity. Material properties in MAX compounds can be toggled by altering the M component.

The synthesis of phase pure MAX compounds are straightforward but energetically expensive to synthesize because it requires sintering at high temperatures (1300-1400 ˚C).61

Synthesis & Processing Once MAX compounds are synthesized, MXenes can be derived by etching out the A phase. The resulting product is Mn+1XnTx where the subscript n is a number from 1-3, T stands for terminal (or surface) groups (O, OH, Cl and F) and the subscript x denotes the number of terminal 62 groups. The first synthesized MXene was Ti3C2Tx etched from the parent Ti3AlC2 in 2011. Along 63-71 with Ti3C2Tx, 18 other MXenes have been synthesized (shown in Figure 1-7). Typical synthesis route involves chemical etching of MAX phase compounds in hydrofluoric acid (HF). The HF etches the A layer and the unstable bonds from the M (formerly M-A bonds) are stabilized by forming new bonds with components from the solution like oxygen, hydroxyl groups, fluorine, and chlorine (depending on the synthesis procedure). In essence, the M-A bond has been replaced by weaker M-X bond. The terminal or surface groups are key differentiators in MXenes compared to other nanomaterials because they allow dispersibility in water without the need for a dispersant 72 unlike most nanosheets. From the synthesis of TI3AlC2 MAX phases to the etching of Ti3C2Tx MXenes, some of the synthesis steps involved are hazardous with potential for dust explosions, runaway reactions, and toxic exposure. Lakhe et al. explored these safety concerns and made recommendations for safe handling and practices.73

Figure 1-7: Potential MXene compositions (reprinted from Nature Reviews Material)71

Ti3AlC2 etched with 50 vol.% HF solution yields Ti3C2Tx with hydroxyl groups as majority of the surface groups.62 Ghidiu et al. proposed a safer route for MXene production; instead of using

HF, an aqueous solution with HCl and LiF can be used to successfully etch Ti3AlC2 into Ti3C2Tx. LiF and HCl forms HF in an aqueous solution but at a much lower concentration.74 With this etching route, myriad of surface groups are obtained. The authors also tried etching with various 74 fluoride salts (instead of LiF) and with H2O2 (instead of HCl) with very limited success. The Li cations that forms from the disassociation of LiF intercalate the interlayer spacing of the Ti3C2Tx + + clay causing expansion of the structure. Researchers were able to ion exchange Li ions with Na , K+, Rb+, Mg2+, and Ca2+ and investigate the following changes in structure.75 Sun et al. successfully demonstrated an electrochemical etching method without the use of any fluoride ions, 76 although it was for a different MXene, Ti2AlC instead of Ti3AlC2.

Figure 1-8 shows the synthesis procedure from MAX phase to nanosheets. Once etched,

Ti3C2Tx possesses an accordion morphology), which is sometimes referred to as clay (Figure

1-8b). The accordion morphology (spacing in between the sheets) is due to the removal of the Al layer. The etched Ti3C2Tx sheets are held together by Van der Waals forces which can be easily overcome (to separate the nanosheets) with mechanical agitation. Ti3C2Tx MXenes are delaminated in water via bath sonication then centrifuge the dispersion to obtain stable Ti3C2Tx dispersion in the supernatant. Different research groups have different sonication and centrifugation time to obtain stable Ti3C2Tx nanosheet dispersion.

Ti3AlC2 MAX Ti3C2Tx MXene Ti3C2Tx nanosheets a b c

HF etching Delamination

Applications

Due to the high electrical conductivity of Ti3C2Tx MXenes, they are promising material for

Lithium ion batteries (LIB) anodes. Sun et al. reported using Ti3C2Tx MXene clay intercalated with dimethyl sulfoxide (DMSO) for use as an LIB anode. Through electrochemical + characterization, the researchers found the Li intercalated the Ti3C2Tx layers leading to charge storage. It was theorized Ti3C2Tx with majority fluorine surface groups instead of oxygen surface group would enhance the electrochemical performance because irreversible capacity is associated 80 with oxygen functional groups. Luo et al. prepared Ti3C2Tx composite for LIB anode with volumetric capacitance of 1375 mAh cm-3, more than twice that of highest reported graphite -3 81 electrode (550 mAh cm ). The composite was first prepared by intercalating Ti3C2Tx clay with

LiOH then immersing the intercalated Ti3C2Tx in dispersion of Sn (IV) with polyvinylpyrrolidone (PVP) as the dispersant. The PVP prevented agglomeration of the Sn (IV) and ensured uniform distribution of the Sn (IV) nanoparticles within Ti3C2Tx clay, which effectively acted as a matrix.

The superior performance of the composite can be attributed to the high conductivity of Ti3C2Tx

12 clay, accessibility to ample ion-exchange sites in Ti3C2Tx clay, and expansion/contraction that 81 occurs during the electrochemical reaction of the Ti3C2Tx clay accordion structure.

Ti3C2Tx MXenes was used as co-catalyst (instead of rare earth metals like Pt) along with cadmium sulfide (CdS) for water splitting for hydrogen production.82 A hydrothermal method was used to synthesize CdS/Ti3C2Tx composites which possessed cauliflower-like structure (Figure 1-9). The metallic properties (charge carriers) and terminal sites or surface functional groups made

Ti3C2Tx good candidate for a co-catalyst. During the hydrothermal treatment, the fluorine functional groups were replaced by oxygen and hydroxyl groups (contributed from aqueous solution). The oxygen and hydroxyl groups contributed to stronger reactions with water molecules. It was reported the hydrogen production was 14,342 µmol h-1g-1, with a quantum efficiency of 82 40.1% at 420 nm. In another publication, Ti3C2Tx composite with TiO2 and CuO was fabricated via cupric nitrate decomposition that demonstrated improved photocatalytic activity than TiO2 (the standard photocatalyst).83

Figure 1-9: TEM images of the CdS/Ti3C2Tx cauliflower-like structure. The scale bars are 200 nm (reprinted from Nature Communications)82

Xie et al. fabricated catalyst where Ti3C2Tx nanosheets were used as the catalyst support for platinum nanoparticles. The developed catalyst possessed half-wave potential of 0.847 V,

13 higher than 0.834 V of commercially available Pt catalyst supported on carbon black. Ti3C2Tx/Pt catalyst also displayed electrochemical surface area loss of 15.66% after 10,000 cycles compared 84 to 40.80% loss in Pt/C catalysts. The superior performance of the Ti3C2Tx supported Pt catalyst can be attributed to the interaction of the Pt and the terminal groups of the Ti3C2Tx nanosheets that prevent Ostwald ripening, dissolution of Pt particles, and prevent Pt agglomeration; these are common problems found with Pt catalyst supported by carbon material. The eventual goal is to utilize Ti3C2Tx/Pt catalyst for proton exchange membrane fuel cells that can generate a high energy output with low environmental hazardous emissions.84

Liu et al. discovered that magnesium hydride (MgH2) with 5 wt.% Ti3C2Tx clay composites started dehydrogenating at 185 ˚C, almost 100 ˚C lower than the dehydrogenation temperature of pure MgH2 (MgH2 by itself is an attractive material for hydrogen storage because of its low cost, reversibility, and high hydrogen capacity of 7.6 wt.%).85 Roughly 6.2 wt.% of hydrogen content was liberated within one min in the 5 wt.% Ti3C2Tx/MgH2 composite sample compared to 85 requiring 80 minutes for pure MgH2 to release the same amount of hydrogen. The hydrogen uptake of the composite was also vastly superior to the pure MgH2 sample; at 100 ˚C, the 5 wt.%

Ti3C2Tx/MgH2 composite sample absorbed 5.5 wt.% while the pristine sample absorbed less than 1% at the same conditions. The authors attributed the vastly superior performance of the composite 85 to the layered structure of Ti3C2Tx that enhance kinetics of hydrogen molecules.

Ren et al. fabricated Ti3C2Tx MXene membranes by vacuum filtering Ti3C2Tx MXene nanosheet dispersion onto a polyvinylidene difluoride (PVDF) filter. They then investigated the permeation of water molecules and various salt molecules based on ion charge. Differing flux of each compound indicates potential for sieving and separation. Due to the low interlayer spacing of the Ti3C2Tx membrane (~6 Å), the permeation of larger molecules is reduced while the permeation of molecules with radii > 6.4 Å are prevented.86 The charge of molecules also plays an important role; cations with single charge causes the expansion of the interlayers leading to quicker permeation rate and cations with multiple charge causes shrinkage of the interlayers leading to slower permeation rates. Water flux rates of some salts solutions are shown in Figure 1-10. The shrinkage and expansion of the interlayers by the cations are result from the interaction between the Ti3C2Tx terminal groups and the passing cations. It was also reported that the water flux through the membrane was 37.4 L bar-1 h-1 m-2 compared to 6.5 L bar-1 h-1 m-2 for graphene oxide membrane

86 of similar thickness. This significance of the study indicates the usefulness of Ti3C2Tx membranes as a sieve (by charge and size) and also for potential applications in water desalination. The potential in utilizing MXenes for separation is abundant because 1) possibility of treating the terminal groups for specific interactions to acquire desired permeation and 2) possibility of utilizing other types of MXenes to fabricate membrane (instead of Ti3C2Tx) that can potentially alter the separation mechanism by allowing for more diversity or uniformity of molecules to permeate through.

Figure 1-10: Water flux of water and salt solutions with different cation charge through the Ti3C2Tx/PVDF membranes (reprinted from Phys. Chem. Letters)86

Ti3C2Tx nanosheets possess excellent electromagnetic waves (EMW) shielding properties. Current material for EMW shielding are metals; the thicker the material the better it is in shielding. However, thicker material translates to higher costs and weight, which is of great concern to the aviation/space industry. EMW shielding is measured in decibels, 10-30 dB provides low level of shielding, 60-90 dB is considered a high level of shielding, and anything above 90 dB is considered exceptional shielding. Ti3C2Tx and other MXenes are great material for shielding EMW because 1) most of the EMW are reflected back due to the free electrons on the surface, 2) the EMW that

15 passes through attenuates between the MXene layers loosing energy, and 3) the terminating groups, especially F interacts with the EMW leading to polarization losses (Figure 1-11).87 The second and third reason mentioned gives Ti3C2Tx and other MXenes the advantage over other materials over in terms of EMW shielding. Additionally, Han et al. demonstrated the usefulness of TiO2 nanocrystals and amorphous carbon in absorption of electromagnetic waves; both compounds can be synthesized on the MXene surface by annealing at controlled temperatures.88

Figure 1-11: Proposed shielding mechanism; reflection, internal reflections, and dipole interactions (reprinted from Science)87

Rasool et al. investigated the viability of using Ti3C2Tx dispersions for anti-bacterial behavior. The investigators compared Ti3AlC2 (MAX phase), Ti3C2Tx clay, and Ti3C2Tx nanosheets in their ability to inhibit growth of two bacteria; Escherichia coli (E. coli) and Baccilus subtilis (B. subtilis). Ti3AlC2 suspension displayed growth inhibition of 14.39% and 18.34,

Ti3C2Tx clay dispersion displayed growth inhibition of 30.55% and 33.60% , and Ti3C2Tx

16 nanosheets dispersion displayed growth inhibition of 97.70% and 97.04% for E. coli and B. subtilis 89 respectively. From various imaging techniques, it was determined that Ti3C2Tx nanosheets performed well because the nanosheets physically damage the bacterial membrane causing cell leakage leading eventually to cell lysis. Another theorized mechanism suggested by the authors mention the activity of the Ti3C2Tx nanosheets terminal groups; the terminal groups interact with the bacterial cell membrane, preventing uptake of key nutrients and eventually leading to cell death. Increasing concentration of Ti3C2Tx nanosheet dispersion decreases the survival rate of bacteria; growth inhibition of 99% were achieved with Ti3C2Tx nanosheet dispersion concentration -1 89 of 200 µm mL . The anti-bacterial activity of Ti3C2Tx nanosheet dispersion is superior the performance of graphene oxide (GO) dispersion, which was considered the standard for 2D material used for anti-bacterial purposes (Figure 1-12).89 Moving forward, a thorough investigation into the health and environmental hazards of Ti3C2Tx nanosheets needs to be conducted to understand the biological effects of Ti3C2Tx beyond bacteria.

a b

Figure 1-12: Comparison of Ti3C2Tx nanosheets and GO with a) E. coli, and b) B. subtillis (reprinted from ACS Nano)89

Ti3C2Tx MXenes can also be utilized in developing biosensors. Zhang et al. found that

Ti3C2Tx MXenes can bind to exosomes. Exosomes are biomarkers useful to predicting, diagnosing, and monitoring cancer-linked health issues. The exosome (Cy3-CD63) can adsorb on the4 surface of MXene through hydrogen bonding and chelation interaction (with the Ti ions). The Ti3C2Tx

MXenes based sensor possesses wide linear range of 104 – 109 particles ml-1 and detection limit of 3 -1 90 1.4x10 particles ml . There have also been reports of Ti3C2Tx MXene biosensor with detection limit of 125 particles µL-1, 100 times lower than current methods.91 Kumar et al. functionalized

Ti3C2Tx MXene nanosheets with aminosilane and utilized them to covalently bind with carcinoembryonic antigen (CEA), a biomarker found in types of cancer responsible for half of the cancer deaths worldwide. The detection range for the functionalized Ti3C2Tx MXenes were 0.0001 – 2000 mg ml-1 with sensitivity of 37.9 µA ng-1 ml cm-2 per decade.92

Thin films made from Ti3C2Tx demonstrate excellent properties and are easy to fabricate. A thin film can be fabricated by simple drop casting, spin coating, and vacuum filtration of a

Ti3C2Tx dispersion; all simple and inexpensive processing steps. Macroscopic properties on a thin film can be measured, which in correlated to the 2D material it is fabricated from. Dillon et al. reported fabrication of Ti3C2Tx MXene thin films (by spin casting) that possessed in-plane conductivity of 650K S m-1.93 The reported conductivity value was the highest among thin films fabricated from solution based 2D materials. The in-plane conductivity of the thin film was in the same order of magnitude as a single Ti3C2Tx nanosheet, indicating the conductivity of single sheet can be replicated for a larger macroscopic area. This phenomenon was ascribed to coplanar alignment of the nanosheets during spin casting. Optical characterization of the thin films revealed free-electron plasma oscillations above 1130 nm, indicating potential for utilizing Ti3C2Tx in 93 plasmonic applications. Hantanasirisakul et al. used spray coating method to fabricate Ti3C2Tx MXene thin films that were not only transparent but also flexible (spray coated on polyester substrate) for the purposes of utilizing them for electrodes and sensors.94 The flexible thin film generated a piezo response upon deformation and this behavior was characterized. The researchers also investigated the effects of different intercalants from NaOH to DMSO to tune the optical properties prior to fabricating the film. Reversible changes in transmittance were observed based 94 on intercalated compound. Xu et al. were able to fabricate Ti3C2Tx electrode films using electrophoretic deposition (EPD). The EPD synthesized films possessed high capacitance (~140 F g-1) and showed good cycling performance with negligible loss in capacitance even after 10,000 runs. The performance was attributed to the layered structure of the film due to EPD process and the optimized pore size distribution of 4 nm that enabled ion diffusion.95 Romer et al. spin-coated

Ti3C2Tx dispersion onto glass substrate to produce Ti3C2Tx thin films with thickness of 13 nm. The films were treated with O and H plasma; in the oxidized state, the resistivity of the film was 5.6

µΩ m and in the reduced state, the resistivity was found to be 4.6 µΩ m. The change in resistivity 96 is reversible and indicates potential for use of Ti3C2Tx films as gas sensors. Additionally, the researchers also found changes to resistivity with changes to relative humidity. The film resistance under vacuum was 243 Ω and under 80% humidity, the resistivity increased roughly 26 times to 6340 Ω.96 Thin films for humidity sensing were developed by An et al. based on this principle.97

An et al. also developed Ti3C2Tx films on deformable substrates through layer-by-layer (LBL) process (Figure 1-13). The flexible Ti3C2Tx composites demonstrated recoverable electrical resistance response under a high bending radius of 2.5 mm and a high amount of tensile strain (40%).98

98 Figure 1-13: LBL assembly process of Ti3C2Tx and PDAC (reprinted from Science Advances)

Ti3C2Tx MXene have been used as fillers for polymers to create composites with enhanced material properties than the virgin polymer. Ti3C2Tx MXene-composites were first investigated by

Ling et al. Ti3C2Tx MXene nanosheets were mixed with each polyvinyl alcohol (PVA) and polydiallyldimethylammonium chloride (PDDA) to make a Ti3C2Tx/polymer composites. The 4 -1 Ti3C2Tx/PVA composite displayed conductivity of 2.2x10 S m , about a magnitude lower than 5 -1 the conductivity of pure Ti3C2Tx MXene nanosheet film (2.4x10 S m ), but still vastly superior than the non-conductive virgin PVA polymer.99 In another work, composites were prepared with ultrahigh molecular weight polyethylene (UHMWPE). The Ti3C2/UHMWPE had higher hardness, better mechanical properties, and creep resistance than pure UHMWPE. Better material properties in the composite can be attributed to higher amount of crystallization due to increased nucleation

19 sites provided by Ti3C2Tx allowing for higher degree of crystallization and larger interface contributing to further load transfer between the filler (Ti3C2Tx) and matrix (UHMWPE). Table 100 1-2 show the mechanical test results of Ti3C2Tx/UHMWPE composites. Shahzad et al. fabricated Ti3C2Tx composites with sodium alginate (SA) for electromagnetic shielding purposes.

The Ti3C2Tx/SA composite had electromagnetic shielding effectiveness of around 55 dB; lower than the 92 dB from pure Ti3C2Tx nanosheet film. However, the 55 dB is still considered good shielding with much lower Ti3C2Tx content and more importantly, the composite was an order of 87 magnitude thinner than the pure Ti3C2Tx nanosheet film. SA was chosen as the matrix because of interaction between oxygen functional groups of SA and the terminating groups of the Ti3C2Tx MXenes.

100 Table 1-2: Mechanical test results of Ti3C2Tx/UHMWPE composites (reprinted from Materials and Design)

Challenges

Ti3C2Tx MXenes possess excellent material properties, making them attractive for a wide range of applications from coatings to photocatalysts. They are also hydrophilic, making

20 processing in water simple unlike other nanosheets. However, they are prone to oxidation and chemically degrade over time. Therefore, it is important to assess oxidation during processing and develop mitigation strategies so they will have a long shelf life to be properly utilized for various applications.

CHAPTER 2 – COSOLVENTS AS LIQUID SURFACTANTS FOR BORON NITRIDE NANOSHEET (BNNS) DISPERSIONS* Summary Despite a range of promising applications, liquid-phase exfoliation of boron nitride nanosheets (BNNSs) is limited, both by low yield in common solvents as well as the disadvantages of using dissolved surfactants. One recently reported approach is the use of co-solvent systems to increase the as-obtained concentration of BNNS; the role of these solvents in aiding exfoliation and/or aiding colloidal stability of BNNSs is difficult to distinguish. In this paper, we have investigated the use of a t-butanol/water co-solvent to disperse BNNSs. We utilize solvent- exchange experiments to demonstrate that the t-butanol is in fact essential to colloidal stability; we then utilized molecular dynamics simulations to explore the mechanism of t-butanol/BNNS interactions. Taken together, the experimental and simulation results show that the key to the success of t-butanol (as compared to the other alcohols of higher or lower molecular weight) lies in its ability to act as a “liquid dispersant” which allows it to favorably interact with both water and BNNSs. Additionally, we show that the stable dispersions of BNNS in water/t-butanol systems may be freeze-dried to yield non-aggregated, redispersible BNNS powders, which would be useful in an array of industrial processes.

Introduction Boron nitride nanosheets (BNNSs) are 2D materials that can be exfoliated from hexagonal boron nitride (hBN). BNNSs have been dubbed “white graphene” because they possess qualities similar to graphene (thermal and mechanical strength), with two major exceptions; they are electrically insulating (band gap ~ 6 eV) and chemically inert.101 Due to their material properties, BNNSs can be utilized as anti-corrosion coatings, anode for lithium ion batteries, water-repellent coatings, water treatment adsorbents, catalyst support materials, and as polymeric composite fillers to enhance thermal, mechanical, and barrier properties.20,53-57,102-109

For scalable production of BNNSs, the most common route is exfoliation (via sonication or shear mixing) in liquid solvents, followed by a separation step (typically centrifugation) to

* Reprinted with permission from Touseef Habib, Dinesh Sundaravadivelu Devarajan, Fardin Khabaz, Dorsa Parviz, Thomas C. Achee, Rajesh Khare, and Micah J. Green. Langmuir, 2016, 32 (44), pp 11591–11599. Copyright © 2016 American Chemical Society

22 remove nanosheets that are not colloidally stable.110 Sonication utilizes acoustic energy that induces cavitation; the resulting flow fields exfoliate layers from the parent material. Long sonication times lead to a greater degree of exfoliation, but the average lateral size of nanosheets in the dispersion decreases. Post sonication, centrifugation is used to separate the exfoliated nanosheets from the unexfoliated material. These processes are involved/tedious and must be carefully tuned to obtain dispersions with nanosheets at high concentration, high yield, high lateral size, and low thicknesses.111

Unfortunately, BNNSs lack colloidal stability in most common solvents; the nanosheets tend to aggregate, which is problematic for both processability and the final material properties of BNNS-based films, coatings, and composites. Typical methods to prevent BNNS re-aggregation involve the use of dispersants to make BNNS colloidally stable.27,31 However, this also introduces a diluent component in the system that is counterproductive to the performance of BNNS-based films and coatings. Therefore, there is a need to process BNNSs without the need of dispersant additives in pure solvents. Recent studies suggest that co-solvent may allow for improved yield without the need of a dispersant.112-114 We explore this in further detail below.

Although BNNSs are not soluble in water, Lin et al. argued that sonication may induce edge functionalization (hydroxyl groups) such that exfoliated BNNSs sufficiently repel each other and form a stable colloid.110 However, the concentration and lateral size of nanosheets in aqueous BNNS dispersions are quite low. To increase BNNSs concentration, Marsh et al. dispersed hBN in co-solvents consisting of water and various polar organic solvents.115 From absorbance data, it was determined that the co-solvent system composed of t-butanol and water (60-40 wt.% respectively) resulted in the highest as-obtained concentration.115 The authors attribute the success of t-butanol to the steric effects caused by its size because of which BNNSs are kept separated. Although steric effects caused by the size of butanol play a role, an important factor is the alcohol size dependent change in the relative magnitudes of hydrophobicity and hydrophilicity, which make t-butanol an optimum solvent for efficient dispersion of BNNS.

In the current chapter, we explore this issue using both experiment and simulation. In the experiments, solvent exchange is used as a means to decouple solvent effects on (i) exfoliation and (ii) colloidal stability. Our experimental results suggest that beyond simple exfoliation, t-butanol plays a critical role in the stability of BNNSs dispersions. Molecular dynamics simulations confirm

23 its effectiveness as a liquid dispersant due to its amphiphilic nature. We show that alcohols of lower molecular weight do not effectively shield BNNS from water, while alcohols of higher molecular weight are immiscible in water. On the other hand, t-butanol provides both advantages (i.e. effective shielding of BNNS from water and also good miscibility with water), allowing it to preferentially migrate to the BNNS surface and provide colloidal stability.

Materials & Methods Dispersion Preparation

60 ml of a co-solvent mixture of t-butanol (308250-1L from Sigma Aldrich) and deionized (DI) water were prepared at a 60-40wt.% ratio respectively in a 100 ml beaker. Hexagonal boron nitride was added at a concentration of 2 mg ml-1. The top of the beaker was covered by a parafilm with a punctured hole in the center for sonicator tip insertion. The parafilm was required because it prevented t-butanol from evaporating out. The sonicator tip (Q-Sonica) was inserted halfway into the dispersion so that it was equidistant from both the bottom of the beaker and top of the dispersion; all dispersions were sonicated for 90 minutes. The sonicator power is set by the amplitude and the amplitude for all experiments were 30% of the maximum value. The output voltage is 1000V, output frequency is 20KHz, and the usual energy output with ½ inch sonicator horn was ~115K Joules after 90 minutes of sonication. After sonication, the dispersion was centrifuged for 4 hours at 3500 rpm. After the centrifugation, the supernatant was extracted and the concentration, ζ potential, and lateral size was measured. Pure water dispersions were prepared using the same sonication-centrifugation method described above in 50 ml of DI water and hBN concentration of 2 mg ml-1. A similar procedure was followed to prepare ethanol-water as well as 1-butanol-water BNNSs dispersion, but the BNNSs were not colloidally stable. BNNSs dispersion with co-solvent mixture of hexanol and water was also attempted, but was not pursued further due to immiscibility between the solvents

Solvent exchange technique

40 ml of co-solvent dispersion was poured into the dialysis bag and clamped tightly at both ends. The bag was placed in a 4000 ml beaker and the beaker was filled with water until the 1000 ml mark. Solvent exchange occurs very rapidly but we kept the dialysis bag submerged in the water bath for 6+ hours. After 6+ hours, the water bath was replaced with another 1000 ml of fresh

24 water. The dialysis bag was submerged for another 6+ hours. At the end of the process, the dispersion was centrifuged for four hours at 3500 rpm.

Dispersions with Polyvinylpyrrolidone (PVP10 from Sigma Aldrich) were prepared using the procedure above; the only difference being dispersants were added at 1mg ml-1 prior to solvent exchange. After adding 1mg ml-1 of dispersant, the dispersion was bath sonicated for 15 minutes for good mixing and then solvent exchanged using the procedure described above.

Absorbance

UV-vis spectroscopy (Shimadzu UV-vis 2550) was used to determine the concentration of dispersions. The extinction coefficient for co-solvent dispersion was calculated by plotting known concentrations against their absorbance at 400 nm wavelength. The concentrations were determined by vacuum filtration. 2 ml of co-solvent dispersion was vacuum filtered in 0.2 μm pore sized filter paper. The mass of the filter paper was measured twice, once before filtration (empty filter paper) and then with filtrate (filter paper was kept in an oven at 60°C for greater than six hours to evaporate out all the moisture so that the only mass contribution would be from the filtrate and filter paper). The mass difference before and after filtration was the mass of the boron nitride nanosheets. Therefore, concentration of the dispersion was calculated from mass and volume, and the absorbance was determined from UV-vis spectroscopy. The procedure was repeated again by diluting 2 ml of co-solvent dispersion to 6 ml (dilution by adding 4 ml of co-solvent) and its absorbance was measured again. The concentrations were plotted against absorbance (at 400 nm) to confirm the linear relationship, as shown in Figure 2-1. The extinction coefficient was calculated using Beer-Lambert Law (A = cLα; A is absorbance, c is concentration, L is the path length of the cuvette which is 0.01 m, and α is the extinction coefficient) and was found to be 125.5 ml m-1 mg-1.

Figure 2-1: The relationship between absorbance (measured at 400 nm) and concentration allows us to calculate extinction coefficient

Preparation of samples for FTIR-ATR

Films were prepared through vacuum filtration; the solvent exchanged dispersion was vacuum filtered onto a Teflon membrane and left to dry overnight. Solvent exchanged dispersion powder was then scrapped off from the film to obtain FTIR-ATR spectra. Similarly, the vacuum- filtered film from the co-solvent exchanged dispersion was prepared by vacuum filtering the co- solvent dispersion onto a Teflon filter paper, drying it overnight, and then washing it thrice with 20 ml of DI water. After extensive washing, the filter paper was dried again overnight; t-butanol is very volatile, so drying overnight would evaporate out excess t-butanol.

Rehydration

T-butanol is a solvent that can be freeze dried.116 Therefore, mixture of t-butanol and water was considered a good candidate for freeze drying. Freeze drying is a gentle processing technique that eliminates the solvent while preventing nanosheet aggregation as compared to other methods of drying.117 Freeze drying BNNS co-solvent dispersion yielded white powder product, pure BNNS. The as obtained BNNS powder was then re-dispersed at a concentration of 0.5mg ml-1 in two different solvents: pure t-butanol and the co-solvent mixture.

Simulation Method

All simulation systems contained a boron nitride nanosheet of size 48 Å × 42 Å that was solvated in water-alcohol co-solvent mixture. Our focus is on the interplay of the interactions between BNNS-alcohol, BNNS-water and alcohol-water molecule pairs. In the absence of chemical modifications, the edges of BNNS make negligible contributions to these interactions and thus the size of BNNS does not affect the results as long as it is large enough to allow adsorption of a large number of alcohol or water molecules, as is the case here. Each system consisted of a single BNNS and 15,625 water molecules. The number of alcohol molecules in the ethanol-BNNS-water, t-butanol-BNNS-water, 1-butanol-BNNS-water and 1-hexanol-BNNS- water systems were 512, 216, 216 and 216 respectively. Each system contained at least 50,000 atoms and the edge length of the cubic box was about 80 Å.

The simulation methodology was very similar to that employed in our recent work27. In particular, BNNS was modeled using the three-body Tersoff potential118 while the TIP3P model119 was employed for water. The interactions of alcohol molecules as well as the non-bonded interactions between BNNS-water and BNNS-alcohol entities were represented by the general AMBER force field (GAFF) parameters.120,121 The partial charges on the atoms of the alcohol molecules were obtained by the AM1-BCC method122,123, whereas those for BNNS were taken from the literature.124

The SHAKE algorithm125 was applied to constrain the bond lengths and the bond angles of water molecules. The van der Waals and electrostatic interactions were truncated at a cut-off distance of 12 Å; tail corrections and the particle-particle particle-mesh (PPPM) method were utilized to account for the long-range interactions.126 Simulations were carried out at constant temperature and pressure conditions of T = 300 K and P = 1 atm. Nose-Hoover thermostat and barostat127 were applied to maintain the temperature and pressure for this purpose. All of the MD simulations were carried out using the LAMMPS package128 with a time step of 1 fs (femtosecond).

The structure i.e. dispersion/aggregation of alcohol molecules in the co-solvent mixture was analyzed using clustering analysis. For the clustering analysis, the cluster size probability distribution of alcohol molecules (which quantifies the probability of alcohols forming clusters of different size in the water medium) in water was evaluated in the absence of BNNS. For this purpose, two alcohol molecules were considered to be part of the same cluster if the distance between the heavy atoms of the two molecules was less than the cluster cutoff distance. Following

27 our previous work, 129,130 the values of the cluster cutoff distance for the alcohol-BNNS-water systems were obtained from the location of the first minimum in the alcohol-alcohol radial distribution function as determined from the pure alcohol simulations. Thus, an alcohol molecule was considered to belong to a cluster if it resided within the cluster cutoff distance of any of the alcohol molecules in the cluster.131

Results & Discussion In this chapter, we investigate the role of t-butanol in stabilizing BNNSs in co-solvent systems using both colloidal experiments and molecular dynamics simulation. We hypothesize that the t-butanol effectively acts as a liquid dispersant, physically adsorbing on the basal plane surface of BNNS to prevent aggregation. The hydrocarbon chain of the t-butanol interacts with the BNNS preventing aggregation while the OH group of t-butanol interacts with water (due to hydrogen bonding) keeping the t-butanol-BNNS complex stable in water.

Experimental Results

To evaluate this hypothesis, we prepared BNNS dispersions in both water as well as t- butanol/water co-solvent mixture. We then manipulated the co-solvent dispersion through solvent exchange (to pure water) and by adding dispersants (PVP). For each dispersion type, we measured BNNS concentration (measured by UV-vis spectroscopy), ζ potential, and lateral size (measured by dynamic light scattering). These results are listed in Table 2-1. By comparing these metrics, we can determine the effects of solvent composition independent of the exfoliation effects. (In contrast, co-solvent mixtures of both ethanol/water as well as 1-butanol/water were prepared for BNNSs dispersion, but the BNNSs sedimented post centrifugation. Hexanol-water co-solvent mixtures were immiscible.)

The procedure for the initial exfoliation and dispersion follows: The parent hBN material was sonicated in the solvent to exfoliate hBN to BNNSs; this suspension was then centrifuged to obtain a two-phase system. The bottom portion consisted of unexfoliated parent material and aggregated BNNSs, while the supernatant contained suspended BNNSs. The ability of the dispersion to remain stable in the supernatant throughout centrifugation provides a check for colloidal stability.

Table 2-1: Concentration, ζ potential, and average lateral size of all dispersions

a b c d

Solvent- Solvent- Pure water Co-solvent exchanged exchanged dispersion dispersion dispersion + dispersion PVP Concentration (mg 0.002 0.213 0.003 0.055 ml-1)

ζ potential (mv) -39.1 ± 0.9 -11.6 ± 1.6 -20.9 ± 0.3 -22.6 ± 0.4

Average lateral size 200.4 ± 0.4 741.1 ± 24.8 371.5 ± 5.3 344.8 ± 5.8 (nm)

We first compare the as-obtained dispersions for both pure water and the co-solvent system (Table 2-1a-b). The as-obtained concentration was significantly higher for the co-solvent system than for the BNNSs dispersed in water; this is consistent with the results of Marsh et al. As discussed earlier, we hypothesize that this higher concentration is not only due to exfoliation effectiveness but also because the t-butanol acts as a liquid dispersant, interacting with the BNNS and water molecules preventing aggregation. (Figure 2-2a depicts representative TEM images of BNNSs in the co-solvent system; the nanosheets are a few layers thick, with lateral sizes in the 0.1-1 μm range.) The data (Table 2-1a-b) also show that the lateral size is higher in the co-solvent system and the ζ potential is lower in the co-solvent. Our hypothesis may shed light on why this is the case. In the water-only system, Lin et al. argued that hydroxyl edge functionalization occurs during sonication and these edge functional groups keep the BNNSs suspended. Presumably, these edge effects would have a larger contribution in the nanosheets with a small lateral size. Indeed, our data show that the water-only system contains chiefly small nanosheets. In contrast, the co- solvent system displays higher concentration and higher lateral size despite the lower ζ potential. This may occur because of a fundamentally different stabilization mechanism; rather than edge functionalization alone, the t-butanol presence on the basal plane sterically (not electrostatically) prevents aggregation. This mechanism is not size-dependent, allowing larger nanosheets to remain dispersed rather than aggregating. These larger nanosheets would have a lower contribution from edge functionalization, which accounts for the lower ζ potential.

Figure 2-2: a) TEM image of co-solvent dispersion. b) TEM image of solvent exchanged dispersion

The preferential adsorption of t-butanol on BNNS surfaces is a plausible mechanism that provides an explanation for the difference in concentration between pure water and the co-solvent mixture. In fact, prior studies have shown that polyvinylalcohol (PVA) in water can sterically stabilize BNNSs.52 PVA is a long chain polymer but the individual units of PVA have a similar structure to t-butanol. Given the demonstration of PVA as a dispersant for BNNS in water, it is plausible that t-butanol will similarly adsorb (akin to a dispersant), presumably with the –OH group oriented away from the BNNS surface while 1-butanol is not able to achieve this conformation on the BNNS surface. This may be one of the contributing factors for why 1-butanol was less effective than t-butanol as a co-solvent. This aspect is further discussed in the simulation section.

To experimentally validate our hypothesis, we implemented a solvent exchange technique to remove t-butanol from the co-solvent dispersion and replace it with water. These solvent- exchanged dispersions were centrifuged after the solvent exchange process to eliminate aggregated BNNSs that formed due to changes in solvent composition. Indeed, the concentration of dispersed BNNSs dropped to values those observed for the water-only system (Table 2-1c); this confirms that the solvent affects colloidal stability, not just exfoliation efficiency. Figure 2-2b shows TEM images of a solvent-exchanged dispersion, in contrast to the original co-solvent dispersion in Figure 2-2a. Both TEM images display thinly layered BNNSs, reflecting the quality of the dispersion. Dark spots in Figure 2-2 a could be contaminants that were introduced during sample preparation (the sample was dried with a household hair drier in laboratory). Nanosheets in Figure

2-2b are almost half the size of the ones in Figure 2-2a. This is consistent with our DLS data in Table 2-1a-b.

Interestingly, the BNNSs that survived the solvent exchange procedure possessed lateral sizes that were roughly twice that of the water-only dispersions (Table 2-1a, Table 2-1c). The larger lateral size may indicate that some larger sheets remain stable because of residual t-butanol. In a separate experiment, 1 ml/mg of PVP was added prior to solvent exchange in order to mitigate re-aggregation during the solvent exchange process; PVP is known to act as a dispersant for BNNSs in a range of solvents.27 The solvent exchange dispersion with PVP was able to arrest aggregation during solvent exchange procedure as evident by the higher concentration compared to the solvent exchanged dispersion (Table 2-1c-d).

Co-solvent dispersions were freeze dried to obtain BNNS powders. One of the key properties of nanosheet powders is their ability to be redispersed into solvents or polymer melts. We re-dispersed these powders in both pure t-butanol and co-solvent at concentrations of 0.5 mg ml-1. Rehydration was also attempted in pure water, but as expected, BNNSs did not redisperse. All three dispersions can be viewed in Figure 2-3.

Figure 2-3: a) Redispersion in pure water was not successful as big flakes could be seen floating around and eventually sedimented to the bottom. b) Redispersion in pure t-butanol was challenging because of its freezing point around room temperature. Redispersing in pure t-butanol yielded a “frozen” dispersion; the BNNSs were kinetically trapped after freezing. c) Redispersion in co-solvent was successful as the BNNSs flakes “dissolved” into solvent.

The dispersion quality (as measured by ζ potential and lateral size) of the redispersed samples are summarized in Table 2-2. Successful redispersion without sonication in the co-solvent mixture suggests that the t-butanol molecules were able to stabilize the as-exfoliated BNNSs. The redispersed co-solvent dispersion possesses similar ζ potential to the original co-solvent system. The average lateral size of the co-solvent redispersed system increased relative to the original co- solvent dispersion. This may be caused by selective loss of small nanosheets during freeze drying. TEM images of these redispersed samples are shown in Figure 2-4; this data suggest not only successful exfoliation but also non-aggregation during the entirety of the process, from initial dispersion to freeze drying and finally to rehydration. The inset in Figure 2-4a displays the low number of layers suggesting a lack of agglomeration in pure t-butanol. Similarly, the inset in Figure 2-4b also shows few layer nanosheets and a lack of agglomeration in the co-solvent mixture.

Table 2-2: Concentration, ζ potential, and average lateral size of rehydrated dispersions

Redispersed in co- Redispersed in t- Redispersed in

solvent butanol water Concentration (mg 0.5 0.5 0.5 ml-1) ζ potential (mv) -10.7 ± 0.4 N/A N/A Average lateral size 1167.7 ± 63.4 804 ± 57.8 N/A (nm)

Figure 2-4: a) TEM image of redispersed BNNSs in pure t-butanol. b) TEM image of redispersed BNNSs in co-solvent mixture. Both TEM images reflect the quality of the dispersions.

FTIR-ATR spectra (Figure 2-5) suggest residual t-butanol in the solvent-exchanged sample. To further confirm that t-butanol adsorbs on BNNSs, FTIR-ATR analysis was performed. FTIR-ATR spectra of pure water and t-butanol were obtained from their respective liquid phases. FTIR-ATR spectra from both vacuum-filtered solvent-exchanged dispersion and vacuum-filtered co-solvent dispersion were obtained from powders that were scraped off from their respective films. Even after vacuum filtering, washing, and drying, the solvent-exchanged sample still displays the characteristic alkane C-H peak (2850-3000 cm-1) and characteristic alcohol OH peak (3200-3500 cm-1), confirming the presence of t-butanol. The samples without any OH peak were the vacuum-filtered co-solvent BNNS dispersion and the freeze dried BNNSs powder.

Figure 2-5: FTIR-ATR of vacuum-filtered, solvent-exchanged dispersion, pure water, pure t-butanol, vacuum filtered co-solvent dispersion, and freeze dried BNNSs powder

Simulation Results†

The experimental results suggest that t-butanol not only aids in exfoliation but also acts as a liquid dispersant. We argue that the amphiphilic character of alcohols plays a central role in stabilizing BNNS in water, and this balance can be modulated by alcohol chain length. To gain deeper insight into this phenomenon, we performed molecular dynamics simulations of alcohol- BNNS-water systems using atomistically detailed models. The chain length dependence of the BNNS stabilization ability of alcohols was investigated by focusing on four systems: (1) ethanol- BNNS-water, (2) t-butanol-BNNS-water, (3) 1-butanol-BNNS-water, and (4) 1-hexanol-BNNS- water.

BNNSs by nature are hydrophobic because they cannot form hydrogen bonds. Alcohols, on the other hand, are amphiphilic in character since the hydroxyl group can form one or more hydrogen bonds with water thus imparting hydrophilic character to them, while the hydrocarbon part of the alcohols is hydrophobic. It is expected that the hydrocarbon chain of the alcohols will wrap around BNNS thus shielding it from water, whereas the end hydroxyl groups will form hydrogen bonds with water, in turn stabilizing the assembly in water. Alcohols thus can act as a

† Simulation work done by Dr. Khare’s group

34 liquid surfactant and avoid phase separation in the system. Since the hydrophobic character of alcohols increases with an increase in the chain length, this aspect is expected to change the ability of the alcohols to stabilize BNNS.

The focus of this study was on the molecular structure in the alcohol-BNNS-water systems. In order to improve statistics, three replicas were studied for each chemical system. Each of the alcohol-BNNS-water model structures was equilibrated for a period of 2.5 ns at a temperature of 300 K, this stage was followed by a production run of 8 ns duration. The end configurations for one of the replicas for each type of alcohol system are displayed in Figure 2-6 (the configurations for the other two replicas for each alcohol system are shown in Figure 2-7 to Figure 2-10. Inspection of Figure 2-6 indicates a chain length dependent behavior of the alcohol systems: ethanol molecules are dispersed throughout the system, most of the butanol (both t-butanol and 1- butanol) molecules accumulate near BNNS but some butanol molecules continue to be dispersed in the system, whereas all hexanol molecules seem to have clustered around BNNS.

Figure 2-6: End configurations of one replica structures for (a) ethanol-BNNS-water system, (b) t-butanol- BNNS-water system, (c) 1-butanol-BNNS-water system, and (d) 1-hexanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity.

Figure 2-7: End configurations of two different replica structures for ethanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity.

Figure 2-8: End configurations of two different replica structures for t-butanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity.

Figure 2-9: End configurations of two different replica structures for 1-butanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity.

Figure 2-10: End configurations of two different replica structures for 1-hexanol-BNNS-water system after 8 ns of production run. Water molecules are not shown for the sake of clarity.

The difference in the behavior of the two types of butanol molecules is not clear from the graphics in Figure 2-6; the origins of this difference are in the packing of these molecules on the BNNS surface and the miscibility of these alcohol molecules with water. To further elucidate the first effect, i.e., to quantify the alcohol molecule packing around BNNS, the radial distribution functions (RDF) of the alcohol (or water) molecules around BNNS were calculated. The RDF quantifies the probability of finding an atom of an alcohol (or water) molecule at a given separation distance from an atom of BNNS; the probability is normalized by the corresponding probability in the case of uniform distribution of molecules. As seen in Figure 2-11a, all of the alcohols have a higher probability of occurrence near BNNS than that for being in the bulk liquid, with ethanol displaying the smallest probability for being near BNNS. Among the longer alcohols, carbon atoms of butanol have a higher probability of being near BNNS than hexanol carbon atoms at shorter separations (r < 12 Å), while at longer separations (12 Å < r < 32 Å), hexanol carbons atoms have a higher probability for being near BNNS Figure 2-11b shows a subtle, yet important difference in the conformation of t-butanol and 1-butanol molecules near BNNS surface. Oxygen atoms of both types of butanol molecules have a higher probability of occurrence near BNNS than the oxygen atoms of ethanol and hexanol molecules, however, the height of the first peak is higher in t-butanol RDF, and also the peak position in t-butanol RDF is at a longer distance from the BNNS surface than in the 1-butanol RDF. These observations indicate that t-butanol not only packs efficiently near BNNS, it packs such that its hydroxyl group is further away from the BNNS, thus making it more accessible to the surrounding water molecules for the purpose of hydrogen bond

37 formation. The packing of water molecules around BNNS is quantified by the RDF shown in Figure 2-11c. Water depletion near BNNS (r < 30 Å) was observed for all four alcohol systems. At small separations from BNNS (r < 8 Å), water concentration is the lowest for t-butanol, while it is the highest (although overall magnitude is still very small) for hexanol. We attribute the results in Figure 2-11a - Figure 2-11c to the qualitative differences in the packing of alcohol molecules around BNNS (see Figure 2-6a-d): while some ethanol molecules are around BNNS, majority of the ethanol molecules are dispersed throughout the system; most of the butanol molecules pack around BNNS and uniformly cover it on all sides; whereas the hexanol molecules pack around BNNS but the packing is asymmetric with most hexanol molecules being on one side of BNNS with a small fraction being on the other side of BNNS.

Figure 2-11: Radial distribution functions between (a) carbon atoms of alcohols and BNNS atoms, (b) oxygen atoms of alcohols and BNNS atoms, and (c) oxygen atoms of water molecules and BNNS atoms. RDFs for each system are shown by the following colors: ethanol-BNNS-water system (dotted dark-green line), t- butanol-BNNS-water system (solid red line), 1-butanol-BNNS-water system (dash-dot dark yellow line), and 1-hexanol-BNNS-water system (dashed dark-blue line).

This packing behavior of alcohol molecules around BNNS is governed by the competition between the alcohol-BNNS and the alcohol-water interactions, i.e., by the alcohol-water miscibility. To further elucidate the water-alcohol interactions, starting from the end configurations of the MD simulation trajectories of the alcohol-BNNS-water systems, the BNNS sheets were removed from the systems, and the systems were subjected to MD simulations. The dispersion of alcohol molecules in the simulated systems was monitored using clustering analysis. These MD runs allowed the alcohol-water systems to attain equilibrium in the absence of the interactions with BNNS. The cluster size distributions in the ethanol and hexanol containing systems equilibrated on a much faster time scale than the cluster size distribution in the butanol containing systems. Thus, ethanol-water and hexanol-water systems were equilibrated for a period of 4 ns followed by a production run of 8 ns, whereas the two butanol-water systems were equilibrated for a period of 12 ns, followed by a production run of 20 ns. The resulting cluster size distributions are plotted in Figure 2-12 and representative snapshots of the systems are shown in Figure 2-13. As seen from Figure 2-12a and Figure 2-12d, both ethanol and hexanol exhibited a unimodal cluster size distribution, albeit at opposite ends of the size range. Thus, ethanol which is strongly hydrophilic, is well dispersed in the water phase with 95% of ethanol molecules occurring in clusters of size smaller than 20. On the other hand, hexanol which has a stronger hydrophobicity due to the presence of the longer hydrocarbon tail, mostly occurs as a large, single cluster with a few hexanol molecules (less than 1%) lying outside this cluster. Butanol, owing to the intermediate length of its hydrocarbon tail, strikes a balance between the hydrophilic and the hydrophobic interactions. Thus, both types of butanol molecules exhibit a bimodal cluster size distribution (see Figure 2-12b and Figure 2-12c), but with a systematic difference in their behavior that is indicative of the difference in their miscibility with water. In particular, approximately 26% of t- butanol molecules reside in small clusters of size one to twenty molecules, while only 9% of 1- butanol molecules exist in clusters of size less than 20 (i.e. much higher fraction of t-butanol molecules are dispersed in water than is the case for 1-butanol). For t-butanol, 74% molecules reside in a large cluster of size in the range of 140 to 180 molecules, whereas for 1-butanol, 91% of molecules reside in clusters of size greater than 180 molecules. Thus, the higher miscibility of t-butanol with water (compared to miscibility of 1-butanol) allows it to simultaneously have favorable interactions with both BNNSs and water, in turn, allowing it to act as a liquid dispersant for BNNS in water.

Conclusion From our findings, BNNS stability in co-solvent extends beyond surface tension and exfoliation, as the organic solvent plays the role of liquid dispersant to effectively stabilize nanosheets. This is supported by TEM images and FTIR-ATR spectra that confirm the presence of t-butanol molecules post solvent exchange step. Additionally, molecular dynamics simulations in which performance of t-butanol was compared with that of other alcohols of various chain lengths, also supports our assertion. Longer carbon chain length increases hydrophobicity leading to weaker interaction of alcohol molecules with water molecules. Short chain length translates to stronger alcohol interaction with the water molecules, leading to unstable dispersion. Butanol possesses the optimum carbon chain length to allow for both BNNS stabilization and miscibility with water. Among the two butanol molecules studied, t-butanol exhibits superior performance for dispersing BNNS due to two reasons: its packing behavior near BNNS which allows its hydroxyl group to be farther away from BNNS, thus making it easily accessible to water, and that its higher miscibility with water leads to favorable interactions with both BNNS and water. These factors allow t-butanol to act as a “liquid surfactant” for stabilizing BNNS in water. Our results show promise for industrial-scale production of BNNSs; the field of nanosheet exfoliation has often faced the difficult decision of using dispersants (surfactants, polymers) to increase yield, with the disadvantage of impurities in the final product. A liquid-phase dispersant system would avoid such disadvantages while still allowing for higher yields.

CHAPTER 3 – OXIDATION STABILITY OF TI3C2TX MXENE NANOSHEETS IN SOLVENTS AND COMPOSITE FILMS‡ Summary

Ti3C2Tx belongs to the fascinating family of MXenes, 2D materials with an attractive combination of functional properties suitable for applications such as batteries, supercapacitors, and strain sensors. However, fabrication of devices and functional coatings based on Ti3C2Tx remains challenging as they are prone to chemical degradation by oxidation into TiO2. In this paper, we examine oxidation of Ti3C2Tx in various media (air, liquid, and solid) via conductivity measurements to assess the shelf life of Ti3C2Tx MXenes. The oxidation of Ti3C2Tx was observed in all media used in this study, but it is fastest in liquid media and slowest in solid media (including polymer matrices). We also show that conventional indicators of MXene oxidation, such as changes in color and colloidal stability, are not always reliable. Finally, we demonstrate acceleration of oxidation under exposure to UV light.

Introduction MXenes are 2D materials consisting of M (early transitional metal), X (carbon or nitrogen) and T (terminal groups). MXenes can be represented with the general formula Mn+1XnTx, where n = 1,2,3 and x represents the number of terminal groups. They are commonly derived from the parent MAX (or Mn+1AXn) phases by selective etching of the A element (a group 13 or 14 element). Out of over 70 known MAX phases with different composition, only a few (~20) have actually been experimentally converted into MXenes.62,70 The most common type of MXene are titanium carbides (Ti3C2Tx), which are obtained from the parent titanium aluminum carbide (Ti3AlC2) MAX phase by removal of aluminum layer by etching in HF or mixture of HCl and fluoride salts; 62,76 however, there is a research interest in pursuing etching without use of fluoride ions. Ti3C2Tx possess high electrical conductivity, excellent electromagnetic shielding properties, and high in- plane stiffness.63,87,88,132,133

The terminal groups (such as -F, -OH, and =O) of Ti3C2Tx MXenes are polar, making them hydrophilic and suitable for processing in water without the need for a dispersant. Aqueous

‡ Reprinted with permission from Touseef Habib, Xiaofei Zhao, Smit A. Shah, Yexiao Chen, Wanmei Sun, Hyosung An, Jodie L. Lutkenhaus, Miladin Radovic & Micah J. Green. npj 2D Materials and Applications volume 3, Article number: 8 (2019)

Ti3C2Tx nanosheet dispersions can be assembled into films, polymer composites, or even into 3D crumpled morphologies.79,100,134 These MXene-based bulk materials have been utilized for hydrogen storage, antibacterial films, flexible electronics, water desalination, and absorption of heavy metals.86,89,99,135,136 Additionally, because of their high conductivity and ease of processing, there is substantial research interest in utilizing these materials for batteries, supercapacitors, and strain sensors.81,98,133,137-141

Despite the promising properties of MXenes, questions about their chemical stability linger. Although a number of reports do not mention this, Ti3C2Tx MXenes are known to oxidize over time. This hampers the utility of Ti3C2Tx (and other MXenes) in numerous applications. Only a few prior studies have explored this problem, but methods for preventing or delaying oxidation of MXenes are still not well understood. Ghassemi et al. demonstrated controlled oxidation of

Ti3C2Tx MXenes under air to obtain TiO2 by varying heating rate, temperature range, and exposure time and showed that slow heating rates cause formation of rutile TiO2 particles while quick heating rates result in formation of anatase TiO2 particles. The formation of TiO2 grains were observed through transmission electron microscopy.142 Halim et al. observed oxidation on the surfaces of air-aged free-standing disks made by cold-pressing multilayer various MXene nanosheets (Ti3C2Tx, Ti2CTx, Ti3CNTx, Nb2CTx and Nb4C3Tx). The authors observed an increase 143 of oxygen content on the surface with time. Maleski et al. dispersed Ti3C2Tx in a wide range of organic solvents (polar protic, polar aprotic, and nonpolar) and analyzed the colloidal dispersion quality using Hildebrand and Hansen solubility theory. The authors noted the color change in

Ti3C2Tx/water dispersion from black to white, while the Ti3C2Tx in organic solvent dispersions stayed black. (Ti3C2Tx MXenes are generally black and TiO2 particles are white.) They proposed that the lack of color change indicates slower oxidation rate in organic solvents when compared to 144 water, suggesting that the water molecules are reacting with Ti3C2Tx. Zhang et al. studied degradation of MXene flakes in aqueous solutions and reported the following: (i) degradation of MXenes flakes are size-dependent, (ii) decreasing colloidal stability is correlated with MXene oxidation, and (iii) they outlined an oxidation preventive storage method: storing MXene-water dispersions in hermetically sealed Argon-filled vials at 5˚C. The authors suggest that dissolved oxygen in the MXene dispersion plays a role in oxidation.145

Despite these efforts, our current understanding of comparative oxidation of MXenes in different media and conditions remains elusive. In this paper, we examine the oxidation of

Ti3C2Tx in air, liquid (water, acetone, acetonitrile), and solid (ice and polymer) media to understand how storage and/or dispersion media influences Ti3C2Tx oxidation processes. Under air, we observe a continual decrease in conductivity with longer exposure time. In aqueous dispersions, we observe a sharp decrease in electric conductivity while the dispersion retains its dark color and colloidal stability, even when the majority of it is oxidized. Ti3C2Tx dispersed in organic solvents (acetone and acetonitrile) had sharp drop in conductivity after being aged 14 days in each media. In air and solid media (frozen samples and polymer composites), the drop in conductivity was slower than that in liquid media.

Materials & Methods

Synthesis of Ti3AlC2 MAX phase

Commercial Ti (44 μm average particle size, 99.5% purity), Al (44 μm average particle size, 99.5% purity) and TiC powders (2 – 3 μm average particle size, 99.5% purity), (all from Alfa

Aesar, MA, USA), were used as starting raw materials to synthesize Ti3AlC2 MAX phase. To prepare homogeneous powder mixtures, Ti, Al and TiC powders were first weighed to achieve Ti:Al:C=3.0:1.2:1.8 ratio and mixed together using ball-milling with zirconia beads in a glass jar at the speed of 300 rpm for 24 hours. Then, the bulk high-purity Ti3AlC2 samples were sintered at temperature of 1510 oC for 15 mins with a loading of 50 MPa using Pulsed Electric Current System

(PECS). To fabricate high-purity Ti3AlC2 powder, the PECSed sample was first drill-milled and sieved to obtain powder with particle sizes below 44 µm.79

Synthesis of Ti3C2Tx MXene clay

Ti3C2Tx MXene clay was synthesized by etching Al from the Ti3AlC2 phase using technique described by Ghidiu et al.74 Concentrated hydrochloric acid (HCl, ACS reagent, 37% w/w Sigma-Aldrich) was diluted with DI water to obtain 30 mL of 6 M HCl solution. This solution was transferred to a polypropylene (PP) beaker and 1.98 gm of lithium fluoride (LiF, 98+% purity, Alfa Aesar) was added to it. This dispersion was stirred for 5 minutes using a

44 the cap to avoid buildup of hydrogen gas. The reaction mixture was stirred at 40 ºC for about 45 hours. The slurry product was centrifuged and washed with deionized (DI) water to remove the unreacted HF and water soluble salts. This washing process was repeated until pH of the filtrate reached a value of about 5. Reaction product is collected at the bottom of the polypropylene 79 centrifuge tubes and is extracted as Ti3C2Tx MXene clay.

Intercalation and delamination of Ti3C2Tx MXene clay

Ti3C2Tx MXene clay was intercalated with dimethyl sulfoxide (DMSO) and eventually bath sonicated to obtain an aqueous dispersion of delaminated Ti3C2Tx MXenes following procedure described in more detail by Mashtalir et al.132 DMSO (ReagentPlus, >99.5%, Sigma-

Aldrich) was added to Ti3C2Tx MXene to form a 60 mg/ml suspension followed by about 18 hours of stirring at room temperature. After intercalation, excess DMSO was removed by several cycles of washing with DI water and centrifugation at 5000 rpm for 4 hours. The intercalated Ti3C2Tx MXene clay’s suspension in deionized water was bath sonicated for 1 hour at room temperature followed by centrifugation at 3500 rpm for 1 hour to separate the heavier components.79

Preparation of Ti3C2Tx vacuum filtered film before freeze- drying

After centrifugation, the supernatant was collected and vacuum filtered on a PTFE membrane (0.45 µ pore size). The vacuum filtered film was then vacuum dried overnight. This was the reference for all vacuum filtered samples.

Freeze-drying Ti3C2Tx nanosheets

After centrifugation, the collected supernatant was stored in a freezer (< 0˚C) then freeze- dried (Labconco FreeZone) for three days to obtain Ti3C2Tx nanosheet powder.

Ti3C2Tx in organic solvent

The freeze dried Ti3C2Tx powder was added to 100 ml of acetone and acetonitrile at a -1 concentration of 1 mg ml . Ti3C2Tx powder in both organic solvents was mixed with the aid of a vortex mixer to form a temporary colloidal solution as all the Ti3C2Tx powder sedimented out after few minutes (for both acetone and acetonitrile). Therefore to obtain a vacuum filtered film, the solution was vacuum filtered right after vortex mixing. The vacuum filtered film was then vacuum dried (to eliminate any moisture) before the conductivity was measured.

Ti3C2Tx in water

-1 The Ti3C2Tx powder (concentration of 1 mg ml ) was added to 100 ml of water and shaken with the aid of a vortex mixer; the powder dispersed upon contact with water. 10 ml of sample was drawn out every week to prepare a vacuum filtered film. The film was then vacuum dried overnight (room temperature) before its electrical conductivity was measured.

Ti3C2Tx in ice

40 ml from the water dispersion was drawn out and separated into four 10 ml samples. The four samples were stored in the freezer to freeze the samples. To prepare a film to measure electrical conductivity, each sample was thawed by submerging the sample container in room temperature water. After thawing, the sample was vacuum filtered to obtain a film; the film was then vacuum dried overnight before its electrical conductivity was measured.

Ti3C2Tx in polymers

Ti3C2Tx powder and polyvinyl alcohol (PVA) (one sample with 50-50 wt.% and another with 10-90 wt.% respectively with total solid concentration of 1 mg ml-1) were bath sonicated for 15 minutes, then vacuum filtered to obtain a polymer composite film. These films were vacuum dried overnight (room temperature) before their electrical conductivity was measured.

UV oxidation

Ti3C2Tx dispersion is vacuum filtered for 30 minutes to obtain a buckypaper film, then it was vacuum oven dried for overnight to remove all moisture. Conductivity of the Ti3C2Tx film was measured before UV exposure and after exposure. All experiments were done under dark housing where the only source of light the sample was exposed to was emitted from the UV lamp (254 nm, 4 Watt, 0.16 Amp).

Characterization

Conductivity measurements were done using 4 Point Resistivity Probe powered by Keithley 2000, 6221, and two 6514. XPS measurements were conducted in Omnicron XPS. Zeta potential was measured using Malvern Zetasizer ZS90. UV-vis measurements were done using Shimadzu UV−vis 2550.

Results & Discussion

To prepare Ti3C2Tx MXene clay, we followed a previously reported procedure for etching 79,146 the A layer from the parent Ti3AlC2 (MAX phase) in a mixture of LiF and HCl for 45 hours.

After etching, the Ti3C2Tx (MXene) clay was washed with DI water. The Ti3C2Tx powder was intercalated with dimethyl sulfoxide (DMSO) and then solvent exchanged to water. The Ti3C2Tx MXene in water was bath sonicated and centrifuged and the supernatant of dispersed, delaminated

Ti3C2Tx nanosheets was collected. These dispersions were freeze dried to obtain Ti3C2Tx nanosheets, which were used to prepare samples by re-dispersing them in liquids (DI water, acetone, and acetonitrile) and in solid media (ice and polymer matrices). The experimental procedure is schematically depicted in Figure 3-1.

Figure 3-1: Overview of experimental procedure: After synthesis of Ti3C2Tx MXene nanosheets from parent MAX phases, samples are dispersed in various media

We used the electrical conductivity of vacuum filtered MXene films as an indicator of the degree of oxidation in air. Prior studies show that increased oxygen content in titanium oxide films leads to lower conductivity; similar trends also exist for graphene conductivity and oxidation.147- 149 After delamination, aqueous dispersions of Ti3C2Tx nanosheets were vacuum filtered to obtain a Ti3C2Tx nanosheet buckypaper. The buckypaper was vacuum-dried overnight to remove excess moisture and its electrical conductivity was measured to be 2.49x104 ± 1.16x103 S m-1; this was used as a starting conductivity value for studying the Ti3C2Tx the oxidation in air and water. The buckypaper was kept at room temperature for the duration of the experiment, and its electrical conductivity was measured with time. Over a period of two months, these Ti3C2Tx vacuum filtered films exposed to atmospheric air displayed a strong decrease in conductivity (Figure 3-2a). After 27 days of exposure, the conductivity was roughly 7% of the original value of 2.49x104 ± 1.16x103 S m-1, suggesting rapid oxidation. The conductivity on the 64th day was 4.90x102 S m-1, less than 2% from the original 0th day measurement. This suggests a strong decrease in conductivity between the 0th and 27th day, and a weaker rate of decrease thereafter. In other words, the sample had more reactive sites in the early stages of the experiment such that oxidation occurs much faster. As the number of reactive sites decreases with time, the oxidation rate becomes slower. The fluctuations on the later days (27th day an onwards) can be attributed to changes in air humidity since humidity 96 has been shown to affect the electrical conductivity of Ti3C2Tx MXenes.

For our study of oxidation in liquid media, three solvents (water, acetone, and acetonitrile) were used. We chose acetone and acetonitrile because non-delaminated MXenes were observed to

48 be the most resistant in these solvents (Figure 3-3 and Figure 3-4). The freeze-dried Ti3C2Tx nanosheet powder was re-dispersed into all three liquid solvents at a concentration of 1 mg ml-1 using a vortex mixer. The freeze-dried powder dispersed back in water forming a colloidal dispersion but it did not disperse in acetone and acetonitrile. All the samples were prepared on the same day.

Figure 3-4: Decrease in conductivity with increasing TiO2 in MXene clay that were aged in various solvents. 1 D is one day, 1 M is 1 month, and 8 M is 8 months.

We first discuss the oxidation of Ti3C2Tx in aqueous dispersions (Table 3-1). After storing MXenes in water for one week, a portion of the dispersion was vacuum filtered into a film and its electrical conductivity was measured; the conductivity of the MXene sample dropped from 2.49x104 ± 1.16x103 to 8.52x103 ± 1.21x103, or by more than 65% compared to the original sample. After 14 days, a portion of the remaining dispersion (supernatant) was again vacuum filtered into a film and its conductivity was measured. The conductivity of this MXene buckypaper was below 3 -1 the measurement threshold (< 10 S m ), suggesting significant oxidation of Ti3C2Tx. The results are in agreement with prior literature; Zhang et al. reported complete oxidation of their Ti3C2Tx MXene dispersion within the span of two weeks as well.145 Their claim was based on drop in UV absorbance while ours is based on a drop in conductivity. Although our water-Ti3C2Tx system was (initially) a stable colloidal dispersion and black in color (Figure 3-5), UV-vis measurements (Figure 3-6) revealed a massive drop in concentration to 0.2 mg ml-1 from the original 1 mg ml-1. The remaining stable colloidal particles had a ζ potential of -25.1 mV, indicative of strong electrostatic repulsion between the dispersed particles. This suggests that the Ti3C2Tx colloidal particles may be oxidized but not to the extent that they become colloidally unstable. X-ray photoelectron spectroscopy (XPS) analysis revealed a high TiO2 content of 55.8%, confirming oxidation (Figure 3-7). These data collectively suggest that Ti3C2Tx colloidal stability and black color are not always directly correlated with degree of oxidation.144,145

Table 3-1: Conductivity (S m-1) of films made from aqueous MXene dispersions with varying dispersion age

Day 0 Day 7 Day 14

4 3 -3 Water 2.5 ± 0.1 x10 8.5 ± 1.2 x10 < 10

Acetone 7.9 ± 3.7 x103 2.4 ± 0.9 x102 4.0 ± 0.6 x101

Acetonitrile 3.0 ± 0.2 x103 1.3 ± 0.1 x102 9.6 ± 0.5 x101

th th Figure 3-5: Ti3C2Tx colloidal dispersion in water on the 0 day and the 14 day

Figure 3-7: XPS of a) vacuum filtered film drawn from fresh (Day 0) Ti3C2Tx colloidal dispersion; the TiO2 content in this sample was 11.06% and b) vacuum filtered film drawn from the same Ti3C2Tx colloidal th dispersion on the 14 day; the TiO2 content was 55.80%

In addition to investigating chemical stability of Ti3C2Tx in aqueous dispersions, we also investigated effects of humidity on the chemical stability of Ti3C2Tx buckypaper. The Ti3C2Tx buckypaper samples were stored in three different relative humidity (RH) conditions; 0%, 50%, and 80%. The conductivity of these samples were measured and the results are shown in Figure

3-8a. The samples stored in a dry condition (RH 0%) maintained their conductivity well over the span of three weeks, while the samples stored in humid conditions experienced significant drops in conductivity. The wettest sample kept in RH 80% experienced the highest drop in conductivity; the increase in TiO2 content in this sample correlates well with the decrease in conductivity as shown in Figure 3-8b.

Figure 3-8: (top) Conductivity as a function of time for samples stored in humid and dry environments. 4+ (bottom) Ti (TiO2) content of RH 80% sample with time.

The freeze-dried Ti3C2Tx powder formed a temporary stable colloidal dispersion in acetone and acetonitrile (Figure 3-9) after vortex mixing even though the Ti3C2Tx powder did sediment out after a few minutes. Vacuum filtered films were prepared by vacuum filtering the temporary dispersions right after vortex mixing (results in Table 3-1). By day 14, Ti3C2Tx aged in all three of the solvents experienced a similar drop in conductivity. However, by day 21, samples prepared from aging in each acetone and acetonitrile possessed a higher conductivity than the sample prepared from water. This indicates Ti3C2Tx oxidation in acetone and acetonitrile occurs at a lower rate than in water. The continued oxidation in acetone and acetonitrile can be attributed to their hygroscopicity. It is probable they (acetone and acetonitrile) absorbed and retained water during the experiment. Ti3C2Tx MXenes also have a high affinity for water molecules, such that any water present in the atmosphere and/or the solvents would interact with the nanosheets and contribute to oxidation over time.75 Overall, our data suggests storing MXenes in liquid media is conducive to oxidation.

Oxidation in ice was assessed by first dispersing Ti3C2Tx in water then freezing it below 0 ˚C. The ice samples were thawed by keeping the vials in room temperature water for 20 minutes. Post thawing, the samples were vacuum filtered to obtain buckypaper; the buckypaper was vacuum dried overnight to remove excess moisture and then its electrical conductivity was measured. The electrical conductivity (as seen in Figure 3-2b) compared to the original was 80%, 33%, and 23% over 7, 14, and 21 days respectively. (In comparison, Zhang at al. reported a 43% drop in the MXene concentration from the fresh sample after 25 days in their system where the samples were

56 kept at 5˚C and pressurized under Argon; presumably, the drop in concentration is the result from the MXenes sedimenting out.) Overall, the frozen samples retained electrical conductivity (within the same order of magnitude) quite well compared to the samples in liquid medium indicating slower oxidation. The slower oxidation can be attributed to slower kinetics due to the solid media and lower temperature. Ti3C2Tx in a solid medium display slower oxidation rates. Even so, the oxidation process cannot be entirely prevented.

Ti3C2Tx has also shown potential as filler in polymers with enhanced electrical conductivity, mechanical performance, and electromagnetic interference shielding performance.87,88,99,100,150 However, no studies have explored any decrease in performance of these materials due to Ti3C2Tx nanosheet oxidation. This issue is critical because of the potential commercial applications of these Ti3C2Tx/polymer composites. If MXene/polymer composites degrade over a short time scale (1-3 weeks), then the long-term utility of such materials becomes compromised, especially for materials that rely on electrical conductivity.

To assess degradation of MXenes in polymer composites, we synthesized Ti3C2Tx/ polyvinyl alcohol (PVA) composites and measured the decrease in electrical conductivity with time (Figure 3-2c). PVA is a water soluble polymer with a repeat unit of [CH2CH(OH)]n. It is widely used commercially because of its hydrophilicity, biodegradability, and non-toxicity.151 The freeze-dried Ti3C2Tx nanosheet powder was used to prepare these vacuum-filtered polymer composites at the following ratios: 1) 50-50 wt. % Ti3C2Tx to PVA respectively and 2) 10-90 wt.

% Ti3C2Tx to PVA respectively; both samples were kept under atmospheric conditions for the duration of the experiment. The conductivity of 50-50 wt. % Ti3C2Tx/PVA sample was roughly 40% of its original value on the 30th day and 20% of the original value by the 57th day. The th conductivity of 10-90 wt. % Ti3C2Tx/PVA sample was roughly 7% of the original value on the 29 th day and 4% of the original value by the 50 day. Similar to the Ti3C2Tx film in air, there seems to be a strong decrease in conductivity within the first four weeks, followed be a more gradual decrease in conductivity thereafter (seen in detail in Figure 3-10). This indicates a slowdown in the oxidation of Ti3C2Tx, most likely due to decreasing number of reactive sites with time. Note that the two polymer composite samples (regardless of the amount of polymer) and the Ti3C2Tx film in air sample follow a similar trend. This suggests the mechanism of oxidation in polymer

57 composite is not affected by the amount of polymer and the hydrophilic polymer does not form an effective protective barrier to prevent oxidation.

Figure 3-10: Master curve for normalized conductivity variation of films made from MXenes dispersed in various media over time

These findings are consistent with observations made on our recently reported layer-by- layer (LbL) films.98 Over the course of four weeks, the absorbance of these MXene/PDAC LbL films decreased, suggesting oxidation (Figure 3-11).

We also examine oxidation under UV irradiation. From past studies, Ti3C2Tx dispersions demonstrated strong absorbance in the wavelength range of 250-300 nm.79 Therefore, we hypothesized that exposure of Ti3C2Tx to UV light in this wavelength range would accelerate oxidation. Based on our data (Figure 3-12 and Table 3-2), there is a strong downtrend in conductivity with longer UV time exposure, suggesting an increase in oxidation. It took only 24 hours of UV exposure under atmospheric conditions to cause conductivity loss of over >85%, while a similar sample under atmospheric conditions and stored in relative darkness experienced conductivity loss of over >85% in 27 days. This has serious implications for Ti3C2Tx MXenes composites made from UV cured polymers; UV curing may degrade Ti3C2Tx MXenes within the monomer.

Figure 3-12: Change in conductivity of MXene film with increasing UV exposure time

Table 3-2: Ti3C2Tx film conductivity vs UV exposure time data

UV exposure time (hours) Conductivity (S m-1)

2 2 0 5.94x10 ± 1.42x10 1 3.96x102 ± 9.89x101 2 3.51x102 ± 1.37x102 3 2.93 x102 ± 6.43x101 4 2.91 x102 ± 6.68x101 5 2.76 x102 ± 8.15x101 6 2.18 x102 ± 7.80x101 8 1.73 x102 ± 5.85x101 24 8.70 x101 ± 3.69x101

We argue that there are two reasons for accelerated Ti3C2Tx oxidation under UV light. The first is the transformation of UV light to heat, leading to increased oxidation kinetics. It has been demonstrated that Ti3C2Tx MXenes can absorb light and convert it to heat with almost 100% 152 efficiency. The second is the generation of radicals under UV light that may attack the Ti3C2Tx surface, causing further oxidation. We surmise that the UV light interacts with the TiO2 causing 153 O2∙ and OH∙ radical formation from the oxygen and moisture present in the environment. With longer UV exposure times, more radicals were generated, leading to more oxidation. Even so, more experiments need to be done to properly assess UV effects on Ti3C2Tx oxidation.

Conclusion

We assessed Ti3C2Tx oxidation behavior in air, liquid, and solid media; we observed oxidation is the slowest in solid media and the fastest in liquid media. In water, Ti3C2Tx oxidizes within two weeks and data suggests that the dispersion’s black color and colloidal stability are not reliable indicators of oxidation, contrary to prior reports. From a storage perspective, Ti3C2Tx can be preserved in ice (and to a lesser extent in some organic solvents) to decrease oxidation or freeze- dried to form a redispersible powder. Ti3C2Tx composites, regardless of the amount of polymer displays similar profile in terms of decrease in conductivity, suggesting the polymer does not act as a barrier to decrease oxidation. Additionally, we find that UV exposure of Ti3C2Tx films accelerates oxidation. We anticipate that this will be one of many studies needed within the MXene community to understand and prevent oxidation and allow for longer shelf-life and reliable functional properties.

CHAPTER 4 – TI3C2TX MXENE/POLYMER COMPOSITES HEAT IN RESPONSE TO RADIO FREQUENCY (RF) FIELDS Summary

We report a new property of Ti3C2Tx MXene nanosheets; they heat in response to low power radio frequency fields, and the heating response is highly dependent on the conductivity of the sample. Ti3C2Tx/polymer composites with different compositions were synthesized to obtain samples with different conductivity. Our study suggests that there is a range of composite conductivity (10 – 1000 S m-1), where the RF induced heating is highest. In addition, since MXenes are prone to oxidation which decreases conductivity, RF heating can be used to monitor

Ti3C2Tx oxidation. This finding that MXene-family materials heat unlocks a new material library for potential RF based applications from cancerous tumor ablation to welding or bonding materials together.

Introduction MXenes are a family of 2D materials discovered in 2011.62 They possess impressive functional properties which can be utilized for catalysts, batteries, and sensors, among 79,82,97,98,133 others. MXenes are obtained from the parent Mn+1AXn phase, where M is an early transitional metal, A is a group 13 or 14 element, X is either carbon or nitrogen, and n can be number 1, 2, or 3. Once the A layer is removed, MXene is obtained. Ti3C2Tx is the most studied

MXene (Tx are the terminal groups) and it is etched from the parent Ti3AlC2 MAX phase (M = Ti,

A = Al, X = C, and n = 2). One of the compelling properties of Ti3C2Tx MXene is its conductivity, with reported values upward of 2.4x105 S m-1, a number similar to multi-layered graphene.99 Recent discoveries of conductive nanomaterials rapidly heating in response to RF prompted our hypothesis that Ti3C2Tx MXenes nanosheets will respond to RF as well.

Radio frequencies (RF) lie between 3 kHz to 300 MHz on the electromagnetic spectrum and, as the name implies, are commonly used for radio communication. RF has also been demonstrated to be useful for ablation of cancerous tumors, food processing and treatment, welding and bonding materials, and more recently for direct heating of carbon nanotubes.154-164 Sweeney et al. reported the first-ever heating of polymer nanocomposites; it was demonstrated that multiwalled carbon nanotubes embedded in epoxy can be utilized for industrial bonding applications. Using RF fields, the epoxy was cured at a faster rate than the current commonly

62 practiced method of oven curing.165 More importantly, this can be cured in an oven-free environment which allows for targeted heating of thermosets.

Most of the prior reports about the relationship between Ti3C2Tx MXenes and electromagnetic waves are about the use of Ti3C2Tx MXenes for electromagnetic interference 87,88,166 (EMI) shielding; this is also related to Ti3C2Tx MXene’s excellent conductivity. The high amount of charge carriers on the MXene surface causes the EM waves to reflect. The waves that are absorbed instead of reflected are weakened by internal attenuation between the MXene layers.

Ti3C2Tx also outperformed other types of MXenes in electromagnetic shielding due to its higher 87 conductivity. However, to our knowledge, no one has ever attempted to examine if Ti3C2Tx MXene (in composites or as a pure film) heat from any kind of electromagnetic waves. We believe this is the first report to probe the low-power RF-heating behavior of Ti3C2Tx MXenes.

Here we evaluate the hypothesis that Ti3C2Tx MXene composites will heat in response to

RF. We demonstrate that a pure Ti3C2Tx MXene film will reflect RF waves because of its high conductivity, but a Ti3C2Tx MXene polymer composite with lower conductivity may absorb RF waves and heat. Using a forward looking infrared (FLIR) camera, we were able to observe the heating responses for Ti3C2Tx MXene/polymer composites with different compositions under low- power RF waves. The results indicate a trend where the heating with RF waves correlates with conductivity.

Materials & Methods

Synthesis of Ti3AlC2 MAX phase

Commercial Ti (44 μm average particle size, 99.5% purity), Al (44 μm average particle size, 99.5% purity) and TiC powders (2 – 3 μm average particle size, 99.5% purity), (all from Alfa

(PECS). To fabricate high-purity Ti3AlC2 powder, the PECSed sample was first drill-milled and sieved to obtain powder with particle sizes below 44 µm.79

Synthesis of Ti3C2Tx MXene clay

Polytetrafluoroethylene (PTFE) magnetic stirrer at room temperature. Ti3AlC2 MAX phase powder was slowly added to the HCl+LiF solution to prevent overheating as the reaction is exothermic. The PP beaker was capped to prevent evaporation of water and a hole was made in the cap to avoid buildup of hydrogen gas. The reaction mixture was stirred at 40 ºC for about 45 hours. The slurry product was centrifuged and washed with deionized (DI) water to remove the unreacted HF and water soluble salts. This washing process was repeated until pH of the filtrate reached a value of about 5. Reaction product is collected at the bottom of the polypropylene 79 centrifuge tubes and is extracted as Ti3C2Tx MXene clay.

Intercalation and delamination of Ti3C2Tx MXene clay

Ti3C2Tx/polymer composites

Ti3C2Tx powder and polyvinyl alcohol (89000-98000, 99+% hydrolyzed, Sigma Aldrich) were bath sonicated for 15 minutes, then vacuum filtered on a polysulfone membrane (with pore size of 0.2 µm) to obtain a polymer composite film. All the composites were made with total mass of 10 mg; this ensure similar areal density for all composites. The composite films were vacuum

64 dried overnight (room temperature) before their electrical conductivity was measured prior to RF experiments.

RF experiments

The RF power source was a signal generator (Rigol Inc., DSG815) and 500 W amplifier (Prana R&D, GN500D. The experimental setup is shown in Figure. We used a non-direct contact fringing field RF applicator. It comprised of two parallel copper strips with a 2 mm spacing on a Teflon slab. All composite films were placed on a 1mm thick glass slide to prevent any damage to RF applicator. We monitored the temperature profile using a Forward-Looking Infrared camera (FLIR systems Inc., A655sc). For optimal RF heating, an impedance match between the source and system is essential to ensure maximum power transfer. An auto tuner changes C (capacitance), L (inductance), R (resistance) elements of circuit at a selected operating frequency to get an impedance match. An alternative to autotuner is to operate at resonant temperature analysis. The samples were exposed to RF ON state (power = 3 W) for 3 s followed by off state (0.0001 W) for 12 s between frequency range of 1MHz to 150 MHz. The on and off type sweep was used to instantaneously heat the sample followed by gradual cooling at each frequency. The frequency sweep plots generated from this method was run through a MATLAB simulation code to get dT/dt verses frequency data. The MATLAB code calculated the heating rates as a function of frequency determined by selecting the points when the power is switched on and ~0.8 s into each cycle and calculating the slope between the points. We selected the resonant frequency for our heating experiments, where dT/dt response was maximum. RF heating response of the films was measured at resonant frequency for 3 W and 1 W RF power. Thermal cycling experiments were performed on films that showed RF response. RF power was switched on (3W) and off (0.0001 W) for 30 s respectively for 50 cycles at resonant frequency.

Characterization

Results & discussion

Synthesis of Ti3C2Tx/polymer composites started with Ti3AlC2, the parent MAX phase of

Ti3C2Tx. Ti3C2Tx was etched from Ti3AlC2 with LiF + HCl (forms hydrofluoric acid) then delaminated in water to obtain Ti3C2Tx nanosheets (Figure 4-1a) using a previously reported procedure.73,79 The delaminated nanosheets along with polyvinyl alcohol (PVA) were used to synthesize Ti3C2Tx/PVA composites via vacuum filtration at initial compositions of 1, 5, 10, 25, 50, 75, 100 wt.% MXene. Figure 4-1b shows a cross-sectional image of the 100 wt.% film. PVA is a commonly used commercial polymer with a repeat unit of [CH2CH(OH)]n. It is also 151 biodegradable and hydrophilic, making it easy to process. Ti3C2Tx/PVA composites have been used to demonstrate the macroscopic benefits of utilizing Ti3C2Tx MXene nanosheets as fillers. Numerous studies have highlighted the excellent electrical conductivity, thermal stability, and 99,167,168 mechanical strength of Ti3C2Tx/PVA composites. Similarly, we use Ti3C2Tx/PVA composites to demonstrate the viability of RF heating.

a b

10 µm 1 µm

Figure 4-1: a) SEM image of delaminated Ti3C2Tx MXene flake, and b) cross-sectional SEM image of neat MXene film

The RF apparatus is shown in Figure 4-2; the sample is placed on a capacitor that generates the RF field (Figure 4-3a-b). All the samples were probed via RF frequency sweep (Figure 4-3c)

66 at 3W to identify the resonant frequency with the greatest heating rate. The frequency sweep was programmed such that the sample was exposed to 3W RF fields for 2 seconds (power turned on) followed by 12 seconds of cooling time (power turned off) at each frequency from 1-150 MHz as shown in Figure 4-4. From the data in Figure 4-4, the heating rate as a function of frequency was calculated by finding the slope (change in temperature) in the 2s period when the power was turned on. Upon identifying the resonant frequency (98 – 100 MHz for all samples), the samples were exposed to RF fields at the resonant frequency for a 12 second period (Figure 4-5 and Figure 4-6) to probe their heating profiles. These samples were exposed to with RF (at 1W and 3W) the same day as they were synthesized; these results are labeled “Day 0” samples (fresh samples). A different set of samples from the same batch were stored under ambient conditions and were treated again with RF (at 1W and 3W) after 30 days to probe their heating profile; these results are labeled as “Day 30” samples (aged sample). The conductivity of these samples on Day 0 and Day 30 were measured by four-point probe and are listed in Table 4-1.

a b Copper strips MXene sample FLIR camera RF field Teflon generator board Capacitor setup

10 mm

Figure 4-2: Photo of a) the tiny capacitor and the sample; the tiny capacitor generates the RF field, and b) the FLIR camera and the RF field generator

96 ˚C a b c

21 ˚C

(˚C/s)

/ dT

Frequency (MHz)

Temperature (C) Temperature 35

20 0 500 1000 1500 2000 Time (s) 1wt% 5wt% 10wt% 25wt% 50wt% 75wt% 100wt%

Figure 4-4: Raw data of RF sweeps. The RF fields were turned on at 3W for 2 seconds then turned off (0W) for 13s for every frequency from 1-150 MHz

50 a 45 40 35 30

Temperature (C) Temperature 25 20 0 2 4 6 8 10 12 Time (s)

1wt% 5wt% 10wt% 25wt% 50wt% 75wt% 100wt%

b 180 160 140 120 100 80 60

Temperature (C) Temperature 40 20 0 0 2 4 6 8 10 12 Time (s) 1wt% 5wt% 10wt% 25wt% 50wt% 75wt% 100wt%

Figure 4-5: Raw data of temperature vs time for a) Day 0 samples at 1W, and b) Day 0 samples at 3W

45 a 40

Temperature (C) Temperature 25

20 0 2 4 6 8 10 12 Time (s) 1wt% 5wt% 10wt% 25wt% 50wt% 75wt% 100wt%

80 b

40 Temperature (C) Temperature 30

20 0 2 4 6 8 10 12 Time (s) 1wt% 5wt% 10wt% 25wt% 50wt% 75wt% 100wt%

Figure 4-6: Raw data of temperature vs time for a) Day 30 samples at 1W, and b) Day 30 samples at 3W

The conductivity of the Day 0 samples are shown in Figure 4-7a. It is clear from the data that the amount of MXene content affects the conductivity value of the composite. The conductivity of the 1wt.% composite was below the measurement threshold (< 10-3 S m-1). The

71 jump in the conductivity value of the 5 wt.% sample (2.41 ± 0.18x10-1 S m-1) from the 1 wt.% sample suggests lack of a percolation network for the 1 wt.% composite.

Table 4-1: Conductivity of the samples by composition on Day 0 and Day 30

MXene composition Conductivity (S/m) Conductivity (S/m) Day % drop from Day 0 to (wt.%) Day 0 30 30 5 4 100 1.26 ± 0.03x10 3.58 ± 0.24x10 - 71.46% 3 3 75 3.54 ± 0.33x10 1.05 ± 0.06x10 - 70.41% 2 2 50 4.32 ± 0.57x10 1.18 ± 0.07x10 - 72.56% 2 1 25 1.10 ± 0.13x10 2.82 ± 0.12x10 - 74.47% 0 0 10 5.01 ± 1.51x10 1.69 ± 0.01x10 - 66.24% -1 -3 5 2.41 ± 0.18x10 < 10^ N/A -3 -3 1 < 10 < 10 N/A -3 -3 1 < 10 < 10 N/A

The RF-induced heating of the samples are shown by the increase in the maximum temperature reached during the time vs. temperature experiments (Figure 4-5 and Figure 4-6 displays the temperature vs. time data for all samples). Figure 4-7b displays temperature vs time data for the 10 wt.% composite sample at Day 0 and 1W; the ambient temperature (Tamb) and the maximum temperature (Tmax) reached are marked. The data demonstrates that a steady temperature is eventually reached upon RF exposure. The temperature rise (Tmax - Tamb) from RF heating of the samples as a function of composition and conductivity are shown in Figure 4-7c and Figure 4-7d respectively. Even at different powers (1W and 3W), the shape of the plot in Figure 4-7c-d suggests that an optimal range of conductivity (10 – 1,000 S m-1) exists where the MXene composites will absorb RF waves and heat. Conversely, there is little RF heating for samples at the extreme sides of the composition range. Similarly, the highest composition sample of 100 wt.% (1.26 ± 0.03x105 S m-1) also did not heat under RF fields. This can be attributed to the reflection of RF fields. Shahzad et al. demonstrated that Ti3C2Tx MXenes surface possess high charge carriers that are responsible for reflecting electromagnetic waves.87 Similar results with carbon nanotubes (CNTs) embedded in polymers have been reported suggesting minimal microwave and RF response at both low and high CNT loadings. Likewise, the 5 wt.% CNT loaded sample was the most RF responsive and it possessed conductivity of around 102 S m-1. Similarly, the 25 wt.%

Ti3C2Tx MXenes loaded sample was also the most RF responsive and it possessed a similar conductivity, 1.10 ± 0.13x102.165,169

a b

)

1 -

Tmax

Tamb

Temperature (˚C) Temperature Conductivity Conductivity m (S

Composition (wt%) Time (s)

c d

for Day 0 Day for

amb

–

max

T T

Composition (wt%) Conductivity (S m-1)

We previously reported that Ti3C2Tx MXenes are prone to oxidation in various media; as 170 a result, their conductivity drops over time. To observe the RF response of oxidized Ti3C2Tx MXene composites, we studied samples stored in ambient conditions for 30 days (Day 30 samples). The conductivities of all the samples decreased over the period of 30 days and the data are shown in Figure 4-8a. The drop in conductivity from Day 0 to Day 30 is clear for all composites and the conductivity of the 5 wt.% composite sample by Day 30 was below the measurement threshold (< 10-3 S m-1). With the exception of the 5 wt.% samples, the conductivity

73 trend for the Day 30 set of samples is similar to the Day 0 set of samples, suggesting oxidation occurred at a similar rate for all composites Figure 4-8b shows Day 30 samples possesses a similar RF heating profile to that of Day 0 samples; again suggesting there is a range of conductivity (10 – 1,000 S m-1) where RF heating is optimal, even at different power levels of 1W and 3W. Table 4-2 shows the comparison between the Day 0 and Day 30 samples. For both power values, 75 wt.% sample experienced a significant jump in heating even though the conductivity decreased. However, the decreased conductivity (1.05 ± 0.06x103 S m-1) is closer to the 10 – 1000 S m-1 range. At the lower conductivity on Day 30 for the 75 wt.% composite, there is more RF wave absorption leading to increased heating. There is a major drop in heating in the 25 wt.% composite by Day 30 at both power. The conductivity of this sample (2.82 ± 0.12x101 S m-1) on Day 30 dropped below the optimal range from Day 0. To understand the amount of oxidation that occurred within 30 day,

X-ray photoelectron spectroscopy (XPS) was done on the 100 wt.% sample; the TiO2 content increased by 25% as shown in Figure 4-9. We previously reported that conductivity can be used as a rapid metric to assess MXene oxidation in lieu of other time intensive.170 Similarly, the RF response can be used as another rapid, functional method to assess the degree of oxidation.

a b

)

for Day 30Day for

amb

–

max

Conductivity Conductivity (S m T

Composition (wt%) Composition (wt%)

Figure 4-8: a) conductivity of Day 0 and Day 30 samples, and b) temperature rise for Day 30 samples (at 1W and 3W) vs composition. Tmax was the max temperature reached during the span of 12 seconds.

Table 4-2: Comparison of RF heating of the fresh (Day 0) and aged (Day 30) set of samples at 1W and 3W of power

Power = 1W T - T @ Day 30 max amb Change from Day 0 to T - T @ Day 0 (˚C) MXene Comp (%) max amb (˚C) 30 100 0.14 0.41 66.3% 75 3.63 13.32 72.7% 50 8.60 9.32 7.7% 25 19.19 12.50 -53.5% 10 6.03 5.58 -8.0% Power = 3W 100 0.13 0.80 83.5% 75 19.00 34.60 45.0% 50 37.88 33.62 -12.6% 25 108.75 39.41 -175.9% 10 25.61 22.38 -14.4%

470 465 460 455 450 Titanium XPS TiO2 (3/2) TiO2 (1/2) b

470 465 460 455 450 Titanium XPS TiO2 (3/2) TiO2 (1/2)

Figure 4-9: Ti3C2Tx’s titanium X-ray photoelectron spectroscopy (XPS) of a) Day 0 sample where the TiO2 content is roughly 4.5% and b) Day 30 sample where the TiO2 content is roughly 30%

We also evaluated whether oxidation directly affects the cyclability of RF heating; any drop in conductivity will affect RF heating. We performed repeated experiments where the sample was exposed to RF and then cooled; the data confirms the cyclability of RF heating. 5, 10, 50 wt.% composites were used for thermal cycling experiments; all of these experiments were done on Day 0. For 50 cycles (Figure 4-10), it can be seen that the maximum temperature reached during RF heating relatively remained constant. The conductivity of the samples, pre and post thermal cycling are reported in Table 4-3. There is a drop in conductivity for all the samples; however, the conductivity drop after thermal cycling for 10 and 50 wt.% composites are within the measurement error. Even though the conductivity drop for 5 wt.% composite is roughly 50%, it is still less than an order of magnitude. More importantly, in terms of absolute magnitude, the decrease in conductivity for the 5 wt.% composite is ~ 0.12 S m-1.

a b c

Temperature (˚C) Temperature

Temperature (˚C) Temperature Temperature (˚C) Temperature

Time (s) Time (s) Time (s)

d e f

Temperature (˚C) Temperature (˚C) Temperature Temperature (˚C) Temperature

Time (s) Time (s) Time (s)

Figure 4-10: Thermal cycling on Day 0 at 3W of a) 5 wt.% sample, b) 10 wt.% sample, c) 50 wt.% sample. The first five cycles of d) 5 wt.% sample, e) 10 wt.% sample, f) 50 wt.% sample

Table 4-3: Conductivities before and after thermal cycling on Day 0 at 3W of 5 wt.%, 10 wt.%, 50 wt.% sample. The % drop in conductivity is not very significant

MXene composition Pre - thermal cycling Post - thermal cycling % drop in conductivity 50 wt.% 2 2 4.32 ± 0.57x10 3.59 ± 0.81x10 - 16.73%

10 wt.% 0 0 5.01 ± 1.51x10 4.80 ± 0.99x10 - 4.17%

5 wt.% -1 -1 2.41 ± 0.18x10 1.26 ± 0.05x10 - 47.89%

Conclusion

We have shown that Ti3C2Tx MXene composites heat under RF waves but the amount of heating is dependent on the composition, and thus the conductivity of the sample. This is a new feature that can be exploited to induce remote heating with low power RF. Utilizing RF heating provides a method to probe oxidation of MXenes based on how much they heat. RF heating of

Ti3C2Tx polymer composites may provide an alternative pathway to cure thermosets to prepare electrically conductive composites. The heating of Ti3C2Tx MXene composites hints at the possibility of RF field heating for other types of MXenes. With the theoretical existence of over > 200 stable MXene phases, there is an immense range of MXenes that may be useful for RF based applications ranging from food processing to cancerous tumor ablation.

CHAPTER 5 – CONCLUSION In this thesis, the processing challenges and solutions of two nanomaterials were presented:

BNNSs and Ti3C2Tx MXenes. These two materials are on the opposite ends of the spectrum in terms of processability. BNNSs are chemically inert, electrically insulating, difficult to exfoliate, and hydrophobic. Ti3C2Tx MXene nanosheets, on the other hand, are prone to chemical degradation (oxidation), electrically conductive, and hydrophilic. However, the end goal for both nanomaterials are the same; successful processing at the nanoscale to be utilized for suitable end applications.

Boron nitride nanosheets Here, we have shown that processing BNNSs can be done in co-solvent systems, and we augment the experimental data with molecular dynamics simulations to explore the co-solvent stratification. Through this processing route, we are able to avoid using solid dispersants to stabilize BNNSs; avoiding such dispersants provides more flexibility in terms of the end application (the dispersant may not be compatible with the application). Additionally, BNNSs from the co-solvent can be harvested via freeze drying to obtain high quality nanosheets which are re-dispersible in the co-solvent system. The BNNSs can be used as fillers for polymer composites or could be used with paint for coating purposes.

The t-butanol in the co-solvent system behaves as a liquid dispersant to stabilize BNNSs giving us higher yields without sacrificing quality (able to obtain single to few-layered sheets). The carbon tail of t-butanol orients on the sheet and the –OH group interacts with water; the simultaneous interaction with water molecules and BNNSs leads to a stable dispersion. Longer carbon chained alcohols are more hydrophobic which increases interaction between BNNSs and alcohol molecules, but the lack of interaction with water molecules leads to an unstable dispersion. The shorter carbon chained alcohols are more hydrophilic leading to strong interaction between alcohol and water molecules, but the weak interaction between the t-butanol and BNNSs also leads to an unstable dispersion. We aim that this new understanding will lead to design of new co-solvent systems that will lead to BNNSs dispersions with higher stability, which hopefully in turn will lead to higher BNNS yields.

Moving forward, our future work will involve designing of co-solvent systems without alcohol. We plan to use carbon chain compounds with amine groups instead of hydroxyl (e.g. t-

78 butylamine instead of t-butanol). We want to investigate the viability of these amine based co- solvents in dispersing BNNSs. The viability of a wide variety of amine based co-solvents will be tested using molecular dynamics simulations, then the successful co-solvents will then be evaluated experimentally in terms of yield and quality.

In addition to the yield and quality, another important parameter is redispersibility. The amine based co-solvents will be lyophilized to obtain pure BNNSs powder, which will then be redispersed back into 1) amine based co-solvents and also into 2) the previous alcohol based co- solvents. The importance of doing this is to determine redispersibility in the different co-solvent systems.

Ti3C2Tx MXene nanosheets

After studying the Ti3C2Tx oxidation behavior of Ti3C2Tx MXenes in air, liquid, and solid media, we concluded oxidation occurs the fastest in liquid media and the slowest in solid media.

In Ti3C2Tx/polymer composites, oxidation still occurs and our data suggests that the polymer does not act as a barrier against oxidation. Additionally, it was observed that exposing Ti3C2Tx MXenes to UV accelerates oxidation. With this new understanding, our hope is that it will lead to newer processing methods with anti-oxidations that may hinder the chemical degradation of Ti3C2Tx MXenes. After all, with the important work done with MXenes, it is important to acknowledge this obvious problem and design solutions to overcome such challenges.

Although quantitative methods (such as XPS composition measurements) are helpful to evaluate Ti3C2Tx oxidation, these methods are time intensive and requires measurements at intervals of times comparable to timescale of oxidation itself. Here, we presented an alternative and indirect method: conductivity measurements on vacuum-filtered films. Conductivity measurements are quick and provides a quantitative measurement about the usability of Ti3C2Tx MXenes.

Moving forward, the next step is to search for a compound that will decrease oxidation and increase the shelf-life of Ti3C2Tx MXenes. Our approach will be to investigate anti-oxidants and their viability in preventing oxidation. It is quite plausible that introducing anti-oxidants may change the Ti3C2Tx MXene material properties, which may limit the potential applications. The experimental results will be augmented with molecular dynamics simulations to validate the data.

We are also expanding our oxidation study to other MXene types, specifically V2CTx (etched from

V2AlC) and analyzing its degradation in different environments.

In terms of application, we utilized Ti3C2Tx MXenes to fabricate composites for sensing purposes. The two sensing applications we targeted were 1) deformation sensing, and 2) humidity sensing. Deformation sensing composites were fabricated via layer-by-layer asembly of Ti3C2Tx MXene on a flexible polymer substrate. Mechanically deforming the composite changed its electrical resistance; these changes were repeatable and consistent. The target application for such composites are wearable electronics.98 Similarly, these composites can be also be used for humidity sensing; changing the amount of water content on the surface changes the electrical resistance.97 Moving forward, we plan to investigate the viability of these MXene composites for the purposes of pH detection (change in electrical resistance based on concentration of –OH) and also as gas sensors (change in electrical resistance based on adsorption of gases). There is also interest in creating composites with Ti3C2Tx MXenes and other mechanical tough nanomaterials. The goal is to fabricate composites that are both mechanical tough and electrically conductive.

We have shown that Ti3C2Tx/polymer nanocomposites are susceptible to low RF fields and the heating under RF fields depends on conductivity. This also provides another method to assess the usability of Ti3C2Tx MXenes, similar to probing conductivity. Additionally, this opens the door for other MXenes types to be used for RF-based applications which can range from simple synthesis of electrically conductivity composites by using MXenes to cure epoxy or to something more sophisticated by utilizing MXenes for cancerous tumor ablation.

Moving forward, we want to investigate the viability of Ti3C2Tx MXene composites as sorbents. Literature on this topic suggests Ti3C2Tx MXenes can be used to adsorb wide variety of pollutants from NH3 to dangerous pollutants like heavy metals due to the abundance of hydroxyl surface groups.82,171-174 Once the target molecules are adsorbed, we want to implement RF heating for desorption. RF induced desorption may be an improvement over the current desorption methods (volumetric heating) because it is a surface heating method; it is less energy intensive with a quicker response time.

As Ti3C2Tx MXenes become a more promising material for commercialization, a study on the process safety and potential scale up was conducted. There are major safety concerns with the processing for MXenes from the initial synthesis of MAX phase to delamination into MXene

80 nanosheets. There are concerns about dust ignition, runaway reactions, and toxic chemical exposure. To address these issues and educate potential industry suitors, we did a safety analysis for all the processing steps. The breadth of the study covers laboratory hazards (important for researchers who are new to MXene processing) to dangers arising from scaling up MXene production.73

We emphasize that future MXene research needs to fall into three categories, 1) synthesis (etching, delmaniation), 2) stability (protection against oxidation), and 3) applications utilizing MXene properties, including RF response. Most of the MXene literature is focused on applications, with some focus on synthesis. However, MXene chemical stability remains an ongoing problem, and the MXene community should strongly focus on this challenge moving forward.

NOMENCLATURE 2D Two dimension

BNNS boron nitride nanosheet

FLIR forward looking infrared [thermal] camera hBN hexagonal boron nitride

HF hydrofluoric acid

MD molecular dynamics

MHz mega hertz

PTFE polytetrafluoroethylene

PVA polyvinyl alcohol

RF radio frequency

SEM scanning electron microscope

TEM transmission electron microscope wt.% weight percent

XPS X-ray photoelectron spectroscopy

XRD X-ray powder diffraction

REFERENCES (1) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Electric Field Effect in Atomically Thin Carbon Films. Science 2004, 306, 666-669. (2) Bonaccorso, F.; Colombo, L.; Yu, G.; Stoller, M.; Tozzini, V.; Ferrari, A. C.; Ruoff, R. S.; Pellegrini, V. Graphene, related two-dimensional crystals, and hybrid systems for energy conversion and storage. Science 2015, 347. (3) Krupka, J.; Strupinski, W. Measurements of the sheet resistance and conductivity of thin epitaxial graphene and SiC films. Applied Physics Letters 2010, 96, 082101. (4) Balandin, A. A.; Ghosh, S.; Bao, W.; Calizo, I.; Teweldebrhan, D.; Miao, F.; Lau, C. N. Superior Thermal Conductivity of Single-Layer Graphene. Nano Letters 2008, 8, 902-907. (5) Ganz, E.; Ganz, A. B.; Yang, L.-M.; Dornfeld, M. The initial stages of melting of graphene between 4000 K and 6000 K. Physical Chemistry Chemical Physics 2017, 19, 3756- 3762. (6) Lee, C.; Wei, X.; Kysar, J. W.; Hone, J. Measurement of the Elastic Properties and Intrinsic Strength of Monolayer Graphene. Science 2008, 321, 385-388. (7) Kawaguchi, M.; Kuroda, S.; Muramatsu, Y. Electronic structure and intercalation chemistry of graphite-like layered material with a composition of BC6N. Journal of Physics and Chemistry of Solids 2008, 69, 1171-1178. (8) Watanabe, K.; Taniguchi, T.; Kanda, H. Direct-bandgap properties and evidence for ultraviolet lasing of hexagonal boron nitride single crystal. Nature Materials 2004, 3, 404. (9) Li, L. H.; Cervenka, J.; Watanabe, K.; Taniguchi, T.; Chen, Y. Strong Oxidation Resistance of Atomically Thin Boron Nitride Nanosheets. ACS Nano 2014, 8, 1457-1462. (10) Yi, M.; Shen, Z.; Zhao, X.; Liang, S.; Liu, L. Boron nitride nanosheets as oxygen- atom corrosion protective coatings. Applied Physics Letters 2014, 104, 143101. (11) Li, L. H.; Xing, T.; Chen, Y.; Jones, R. Boron Nitride Nanosheets for Metal Protection. Advanced Materials Interfaces 2014, 1, 1300132-n/a. (12) Lindsay, L.; Broido, D. A. Enhanced thermal conductivity and isotope effect in single-layer hexagonal boron nitride. Physical Review B 2011, 84, 155421.

(13) Bosak, A.; Serrano, J.; Krisch, M.; Watanabe, K.; Taniguchi, T.; Kanda, H. Elasticity of hexagonal boron nitride: Inelastic x-ray scattering measurements. Physical Review B 2006, 73, 041402. (14) Pacilé, D.; Meyer, J. C.; Girit, Ç. Ö.; Zettl, A. The two-dimensional phase of boron nitride: Few-atomic-layer sheets and suspended membranes. Applied Physics Letters 2008, 92, 133107. (15) Li, L. H.; Chen, Y.; Behan, G.; Zhang, H.; Petravic, M.; Glushenkov, A. M. Large- scale mechanical peeling of boron nitride nanosheets by low-energy ball milling. Journal of Materials Chemistry 2011, 21, 11862-11866. (16) Deepika; Li, L. H.; Glushenkov, A. M.; Hait, S. K.; Hodgson, P.; Chen, Y. High- Efficient Production of Boron Nitride Nanosheets via an Optimized Ball Milling Process for Lubrication in Oil. Scientific Reports 2014, 4, 7288. (17) Lee, D.; Lee, B.; Park, K. H.; Ryu, H. J.; Jeon, S.; Hong, S. H. Scalable Exfoliation Process for Highly Soluble Boron Nitride Nanoplatelets by Hydroxide-Assisted Ball Milling. Nano Letters 2015, 15, 1238-1244. (18) Ikuno, T.; Sainsbury, T.; Okawa, D.; Fréchet, J. M. J.; Zettl, A. Amine- functionalized boron nitride nanotubes. Solid State Communications 2007, 142, 643-646. (19) Cui, Z.; Oyer, A. J.; Glover, A. J.; Schniepp, H. C.; Adamson, D. H. Large Scale Thermal Exfoliation and Functionalization of Boron Nitride. Small 2014, 10, 2352-2355. (20) Xiao, F.; Naficy, S.; Casillas, G.; Khan, M. H.; Katkus, T.; Jiang, L.; Liu, H.; Li, H.; Huang, Z. Edge-Hydroxylated Boron Nitride Nanosheets as an Effective Additive to Improve the Thermal Response of Hydrogels. Advanced Materials 2015, 27, 7196-7203. (21) Lin, Y.; Williams, T. V.; Connell, J. W. Soluble, Exfoliated Hexagonal Boron Nitride Nanosheets. The Journal of Physical Chemistry Letters 2010, 1, 277-283. (22) Nag, A.; Raidongia, K.; Hembram, K. P. S. S.; Datta, R.; Waghmare, U. V.; Rao, C. N. R. Graphene Analogues of BN: Novel Synthesis and Properties. ACS Nano 2010, 4, 1539- 1544. (23) Sainsbury, T.; Satti, A.; May, P.; Wang, Z.; McGovern, I.; Gun’ko, Y. K.; Coleman, J. Oxygen Radical Functionalization of Boron Nitride Nanosheets. Journal of the American Chemical Society 2012, 134, 18758-18771.

(24) Sainsbury, T.; Satti, A.; May, P.; O'Neill, A.; Nicolosi, V.; Gun'ko, Y. K.; Coleman, J. N. Covalently Functionalized Hexagonal Boron Nitride Nanosheets by Nitrene Addition. Chemistry – A European Journal 2012, 18, 10808-10812. (25) Lee, D.; Song, S. H.; Hwang, J.; Jin, S. H.; Park, K. H.; Kim, B. H.; Hong, S. H.; Jeon, S. Enhanced Mechanical Properties of Epoxy Nanocomposites by Mixing Noncovalently Functionalized Boron Nitride Nanoflakes. Small 2013, 9, 2602-2610. (26) Bhimanapati, G. R.; Kozuch, D.; Robinson, J. A. Large-scale synthesis and functionalization of hexagonal boron nitride nanosheets. Nanoscale 2014, 6, 11671-11675. (27) Bari, R.; Parviz, D.; Khabaz, F.; Klaassen, C. D.; Metzler, S. D.; Hansen, M. J.; Khare, R.; Green, M. J. Liquid phase exfoliation and crumpling of inorganic nanosheets. Physical Chemistry Chemical Physics 2015, 17, 9383-9393. (28) Bari, R.; Parviz, D.; Khabaz, F.; Klaassen, C. D.; Metzler, S. D.; Hansen, M. J.; Khare, R.; Green, M. J. Liquid phase exfoliation and crumpling of inorganic nanosheets. Physical chemistry chemical physics : PCCP 2015, 17, 9383-9393. (29) Coleman, J. N.; Lotya, M.; O’Neill, A.; Bergin, S. D.; King, P. J.; Khan, U.; Young, K.; Gaucher, A.; De, S.; Smith, R. J.; Shvets, I. V.; Arora, S. K.; Stanton, G.; Kim, H.-Y.; Lee, K.; Kim, G. T.; Duesberg, G. S.; Hallam, T.; Boland, J. J.; Wang, J. J.; Donegan, J. F.; Grunlan, J. C.; Moriarty, G.; Shmeliov, A.; Nicholls, R. J.; Perkins, J. M.; Grieveson, E. M.; Theuwissen, K.; McComb, D. W.; Nellist, P. D.; Nicolosi, V. Two-Dimensional Nanosheets Produced by Liquid Exfoliation of Layered Materials. Science 2011, 331, 568-571. (30) Smith, R. J.; King, P. J.; Lotya, M.; Wirtz, C.; Khan, U.; De, S.; O'Neill, A.; Duesberg, G. S.; Grunlan, J. C.; Moriarty, G.; Chen, J.; Wang, J.; Minett, A. I.; Nicolosi, V.; Coleman, J. N. Large‐Scale Exfoliation of Inorganic Layered Compounds in Aqueous Surfactant Solutions. Advanced Materials 2011, 23, 3944-3948. (31) Liu, Y.-T.; Xie, X.-M.; Ye, X.-Y. Tuning the solubility of boron nitride nanosheets in organic solvents by using block copolymer as a "Janus" modifier. Chemical Communications 2013, 49, 388-390. (32) Pierson, H. O. Boron Nitride Composites By Chemical Vapor Deposition. Journal of Composite Materials 1975, 9, 228-240. (33) Rozenberg, A. S.; Sinenko, Y. A.; Chukanov, N. V. Regularities of pyrolytic boron nitride coating formation on a graphite matrix. Journal of Materials Science 1993, 28, 5528-5533.

(34) Middleman, S. The role of gas-phase reactions in boron nitride growth by chemical vapor deposition. Materials Science and Engineering: A 1993, 163, 135-140. (35) Chatterjee, S.; Luo, Z.; Acerce, M.; Yates, D. M.; Johnson, A. T. C.; Sneddon, L. G. Chemical Vapor Deposition of Boron Nitride Nanosheets on Metallic Substrates via Decaborane/Ammonia Reactions. Chemistry of Materials 2011, 23, 4414-4416. (36) Haolin, W.; Yajuan, Z.; Yong, X.; Xiaohua, M.; Xingwang, Z. Recent progress in synthesis of two-dimensional hexagonal boron nitride. Journal of Semiconductors 2017, 38, 031003. (37) Auwärter, W.; Suter, H. U.; Sachdev, H.; Greber, T. Synthesis of One Monolayer of Hexagonal Boron Nitride on Ni(111) from B-Trichloroborazine (ClBNH)3. Chemistry of Materials 2004, 16, 343-345. (38) Müller, F.; Stöwe, K.; Sachdev, H. Symmetry versus Commensurability: Epitaxial Growth of Hexagonal Boron Nitride on Pt(111) From B-Trichloroborazine (ClBNH)3. Chemistry of Materials 2005, 17, 3464-3467. (39) Constant, G.; Feurer, R. Preparation and characterization of thin protective films in silica tubes by thermal decomposition of hexachloroborazine. Journal of the Less Common Metals 1981, 82, 113-118. (40) Bao, J.; Jeppson, K.; Edwards, M.; Fu, Y.; Ye, L.; Lu, X.; Liu, J. Synthesis and applications of two-dimensional hexagonal boron nitride in electronics manufacturing. Electronic Materials Letters 2016, 12, 1-16. (41) Gao, Y.; Ren, W.; Ma, T.; Liu, Z.; Zhang, Y.; Liu, W.-B.; Ma, L.-P.; Ma, X.; Cheng, H.-M. Repeated and Controlled Growth of Monolayer, Bilayer and Few-Layer Hexagonal Boron Nitride on Pt Foils. ACS Nano 2013, 7, 5199-5206. (42) Kim, S. M.; Hsu, A.; Park, M. H.; Chae, S. H.; Yun, S. J.; Lee, J. S.; Cho, D.-H.; Fang, W.; Lee, C.; Palacios, T.; Dresselhaus, M.; Kim, K. K.; Lee, Y. H.; Kong, J. Synthesis of large-area multilayer hexagonal boron nitride for high material performance. Nature Communications 2015, 6, 8662. (43) Song, L.; Ci, L.; Lu, H.; Sorokin, P. B.; Jin, C.; Ni, J.; Kvashnin, A. G.; Kvashnin, D. G.; Lou, J.; Yakobson, B. I.; Ajayan, P. M. Large Scale Growth and Characterization of Atomic Hexagonal Boron Nitride Layers. Nano Letters 2010, 10, 3209-3215.

(44) Weng, Q.; Wang, B.; Wang, X.; Hanagata, N.; Li, X.; Liu, D.; Wang, X.; Jiang, X.; Bando, Y.; Golberg, D. Highly Water-Soluble, Porous, and Biocompatible Boron Nitrides for Anticancer Drug Delivery. ACS Nano 2014, 8, 6123-6130. (45) Lei, W.; Portehault, D.; Liu, D.; Qin, S.; Chen, Y. Porous boron nitride nanosheets for effective water cleaning. Nature Communications 2013, 4, 1777. (46) Sun, W.; Wang, L.; Wu, T.; Pan, Y.; Liu, G. Communication—Multi-Layer Boron Nitride Nanosheets as Corrosion-Protective Coating Fillers. Journal of The Electrochemical Society 2016, 163, C16-C18. (47) Chilkoor, G.; Karanam, S. P.; Star, S.; Shrestha, N.; Sani, R. K.; Upadhyayula, V. K. K.; Ghoshal, D.; Koratkar, N. A.; Meyyappan, M.; Gadhamshetty, V. Hexagonal Boron Nitride: The Thinnest Insulating Barrier to Microbial Corrosion. ACS Nano 2018, 12, 2242-2252. (48) Zulkurnain, E. S.; Ahmad, F.; Gillani, Q. F. Effects of nano-sized boron nitride (BN) reinforcement in expandable graphite based in-tumescent fire retardant coating. IOP Conference Series: Materials Science and Engineering 2016, 146, 012037. (49) Pakdel, A.; Zhi, C.; Bando, Y.; Nakayama, T.; Golberg, D. Boron Nitride Nanosheet Coatings with Controllable Water Repellency. ACS Nano 2011, 5, 6507-6515. (50) Huang, C.; Chen, C.; Ye, X.; Ye, W.; Hu, J.; Xu, C.; Qiu, X. Stable colloidal boron nitride nanosheet dispersion and its potential application in catalysis. Journal of Materials Chemistry A 2013, 1, 12192-12197. (51) Duan, Z. Q.; Liu, Y. T.; Xie, X. M.; Ye, X. Y.; Zhu, X. D. h‐BN Nanosheets as 2D Substrates to Load 0D Fe3O4 Nanoparticles: A Hybrid Anode Material for Lithium‐Ion Batteries. Chemistry – An Asian Journal 2016, 11, 828-833. (52) Khan, U.; May, P.; O'Neill, A.; Bell, A. P.; Boussac, E.; Martin, A.; Semple, J.; Coleman, J. N. Polymer reinforcement using liquid-exfoliated boron nitride nanosheets. Nanoscale 2013, 5, 581-587. (53) Wu, X.; Liu, H.; Tang, Z.; Guo, B. Scalable fabrication of thermally conductive elastomer/boron nitride nanosheets composites by slurry compounding. Composites Science and Technology 2016, 123, 179-186. (54) Xie, S.; Istrate, O. M.; May, P.; Barwich, S.; Bell, A. P.; Khan, U.; Coleman, J. N. Boron nitride nanosheets as barrier enhancing fillers in melt processed composites. Nanoscale 2015, 7, 4443-4450.

(55) Su, J.; Xiao, Y.; Ren, M. Enhanced thermal conductivity in epoxy nanocomposites with hybrid boron nitride nanotubes and nanosheets. physica status solidi (a) 2013, 210, 2699- 2705. (56) Yan, H.; Tang, Y.; Su, J.; Yang, X. Enhanced thermal–mechanical properties of polymer composites with hybrid boron nitride nanofillers. Applied Physics A 2013, 114, 331-337. (57) Hu, X.; Liu, J.; He, Q.; Meng, Y.; Cao, L.; Sun, Y.-P.; Chen, J.; Lu, F. Aqueous compatible boron nitride nanosheets for high-performance hydrogels. Nanoscale 2016, 8, 4260- 4266. (58) Eklund, P.; Beckers, M.; Jansson, U.; Högberg, H.; Hultman, L. The Mn + 1AXn phases: Materials science and thin-film processing. Thin Solid Films 2010, 518, 1851-1878. (59) Magnuson, M.; Mattesini, M. Chemical bonding and electronic-structure in MAX phases as viewed by X-ray spectroscopy and density functional theory. Thin Solid Films 2017, 621, 108-130. (60) Barsoum, M. W.; El-Raghy, T. Synthesis and Characterization of a Remarkable Ceramic: Ti3SiC2. Journal of the American Ceramic Society 1996, 79, 1953-1956. (61) Peng, C.; Wang, C.-A.; Song, Y.; Huang, Y. A novel simple method to stably synthesize Ti3AlC2 powder with high purity. Materials Science and Engineering: A 2006, 428, 54-58. (62) Naguib, M.; Kurtoglu, M.; Presser, V.; Lu, J.; Niu, J.; Heon, M.; Hultman, L.; Gogotsi, Y.; Barsoum, M. W. Two-Dimensional Nanocrystals Produced by Exfoliation of Ti3AlC2. Advanced Materials 2011, 23, 4248-4253. (63) Naguib, M.; Mashtalir, O.; Carle, J.; Presser, V.; Lu, J.; Hultman, L.; Gogotsi, Y.; Barsoum, M. W. Two-Dimensional Transition Metal Carbides. ACS Nano 2012, 6, 1322-1331. (64) Naguib, M.; Halim, J.; Lu, J.; Cook, K. M.; Hultman, L.; Gogotsi, Y.; Barsoum, M. W. New Two-Dimensional Niobium and Vanadium Carbides as Promising Materials for Li-Ion Batteries. Journal of the American Chemical Society 2013, 135, 15966-15969. (65) Meshkian, R.; Näslund, L.-Å.; Halim, J.; Lu, J.; Barsoum, M. W.; Rosen, J. Synthesis of two-dimensional molybdenum carbide, Mo2C, from the gallium based atomic laminate Mo2Ga2C. Scripta Materialia 2015, 108, 147-150.

(66) Urbankowski, P.; Anasori, B.; Makaryan, T.; Er, D.; Kota, S.; Walsh, P. L.; Zhao, M.; Shenoy, V. B.; Barsoum, M. W.; Gogotsi, Y. Synthesis of two-dimensional titanium nitride Ti4N3 (MXene). Nanoscale 2016, 8, 11385-11391. (67) Zhou, J.; Zha, X.; Chen, F. Y.; Ye, Q.; Eklund, P.; Du, S.; Huang, Q. A Two- Dimensional Zirconium Carbide by Selective Etching of Al3C3 from Nanolaminated Zr3Al3C5. Angewandte Chemie 2016, 128, 5092-5097. (68) Zhou, J.; Zha, X.; Zhou, X.; Chen, F.; Gao, G.; Wang, S.; Shen, C.; Chen, T.; Zhi, C.; Eklund, P.; Du, S.; Xue, J.; Shi, W.; Chai, Z.; Huang, Q. Synthesis and Electrochemical Properties of Two-Dimensional Hafnium Carbide. ACS Nano 2017. (69) Ghidiu, M.; Naguib, M.; Shi, C.; Mashtalir, O.; Pan, L. M.; Zhang, B.; Yang, J.; Gogotsi, Y.; Billinge, S. J. L.; Barsoum, M. W. Synthesis and characterization of two-dimensional Nb4C3 (MXene). Chemical Communications 2014, 50, 9517-9520. (70) Anasori, B.; Xie, Y.; Beidaghi, M.; Lu, J.; Hosler, B. C.; Hultman, L.; Kent, P. R. C.; Gogotsi, Y.; Barsoum, M. W. Two-Dimensional, Ordered, Double Transition Metals Carbides (MXenes). ACS Nano 2015, 9, 9507-9516. (71) Anasori, B.; Lukatskaya, M. R.; Gogotsi, Y. 2D metal carbides and nitrides (MXenes) for energy storage. Nature Reviews Materials 2017, 2, 16098. (72) Naguib, M.; Mochalin, V. N.; Barsoum, M. W.; Gogotsi, Y. 25th Anniversary Article: MXenes: A New Family of Two-Dimensional Materials. Advanced Materials 2014, 26, 992-1005. (73) Lakhe, P.; Prehn, E. M.; Habib, T.; Lutkenhaus, J. L.; Radovic, M.; Mannan, M. S.; Green, M. J. Process Safety Analysis for Ti3C2Tx MXene Synthesis and Processing. Industrial & Engineering Chemistry Research 2019, 58, 1570-1579. (74) Ghidiu, M.; Lukatskaya, M. R.; Zhao, M.-Q.; Gogotsi, Y.; Barsoum, M. W. Conductive two-dimensional titanium carbide /`clay/' with high volumetric capacitance. Nature 2014, 516, 78-81. (75) Ghidiu, M.; Halim, J.; Kota, S.; Bish, D.; Gogotsi, Y.; Barsoum, M. W. Ion- Exchange and Cation Solvation Reactions in Ti3C2 MXene. Chemistry of Materials 2016, 28, 3507-3514.

(76) Sun, W.; Shah, S. A.; Chen, Y.; Tan, Z.; Gao, H.; Habib, T.; Radovic, M.; Green, M. J. Electrochemical etching of Ti2AlC to Ti2CTx (MXene) in low-concentration hydrochloric acid solution. Journal of Materials Chemistry A 2017, 5, 21663-21668. (77) Song, J.; Yang, H. Fabrication of high-purity ternary carbide Ti3AlC2 by spark plasma sintering (SPS) technique. Journal of Wuhan University of Technology-Mater. Sci. Ed. 2008, 23, 177-180. (78) Mashtalir, O.; Cook, K. M.; Mochalin, V. N.; Crowe, M.; Barsoum, M. W.; Gogotsi, Y. Dye adsorption and decomposition on two-dimensional titanium carbide in aqueous media. Journal of Materials Chemistry A 2014, 2, 14334-14338. (79) Shah, S. A.; Habib, T.; Gao, H.; Gao, P.; Sun, W.; Green, M. J.; Radovic, M. Template-free 3D titanium carbide (Ti3C2Tx) MXene particles crumpled by capillary forces. Chemical Communications 2017, 53, 400-403. (80) Sun, D.; Wang, M.; Li, Z.; Fan, G.; Fan, L.-Z.; Zhou, A. Two-dimensional Ti3C2 as anode material for Li-ion batteries. Electrochemistry Communications 2014, 47, 80-83. (81) Luo, J.; Tao, X.; Zhang, J.; Xia, Y.; Huang, H.; Zhang, L.; Gan, Y.; Liang, C.; Zhang, W. Sn4+ Ion Decorated Highly Conductive Ti3C2 MXene: Promising Lithium-Ion Anodes with Enhanced Volumetric Capacity and Cyclic Performance. ACS Nano 2016, 10, 2491-2499. (82) Ran, J.; Gao, G.; Li, F.-T.; Ma, T.-Y.; Du, A.; Qiao, S.-Z. Ti3C2 MXene co-catalyst on metal sulfide photo-absorbers for enhanced visible-light photocatalytic hydrogen production. Nature Communications 2017, 8, 13907. (83) Lu, Y.; Yao, M.; Zhou, A.; Hu, Q.; Wang, L. Preparation and Photocatalytic Performance of Ti3C2/TiO2/CuO Ternary Nanocomposites. Journal of Nanomaterials 2017, 2017, 5. (84) Xie, X.; Chen, S.; Ding, W.; Nie, Y.; Wei, Z. An extraordinarily stable catalyst: Pt NPs supported on two-dimensional Ti3C2X2 (X = OH, F) nanosheets for oxygen reduction reaction. Chemical Communications 2013, 49, 10112-10114. (85) Liu, Y.; Du, H.; Zhang, X.; Yang, Y.; Gao, M.; Pan, H. Superior catalytic activity derived from a two-dimensional Ti3C2 precursor towards the hydrogen storage reaction of magnesium hydride. Chemical Communications 2016, 52, 705-708.

(86) Ren, C. E.; Hatzell, K. B.; Alhabeb, M.; Ling, Z.; Mahmoud, K. A.; Gogotsi, Y. Charge- and Size-Selective Ion Sieving Through Ti3C2Tx MXene Membranes. The Journal of Physical Chemistry Letters 2015, 6, 4026-4031. (87) Shahzad, F.; Alhabeb, M.; Hatter, C. B.; Anasori, B.; Man Hong, S.; Koo, C. M.; Gogotsi, Y. Electromagnetic interference shielding with 2D transition metal carbides (MXenes). Science 2016, 353, 1137-1140. (88) Han, M.; Yin, X.; Wu, H.; Hou, Z.; Song, C.; Li, X.; Zhang, L.; Cheng, L. Ti3C2 MXenes with Modified Surface for High-Performance Electromagnetic Absorption and Shielding in the X-Band. ACS Applied Materials & Interfaces 2016, 8, 21011-21019. (89) Rasool, K.; Helal, M.; Ali, A.; Ren, C. E.; Gogotsi, Y.; Mahmoud, K. A. Antibacterial Activity of Ti3C2Tx MXene. ACS Nano 2016, 10, 3674-3684. (90) Zhang, Q.; Wang, F.; Zhang, H.; Zhang, Y.; Liu, M.; Liu, Y. Universal Ti3C2 MXenes Based Self-Standard Ratiometric Fluorescence Resonance Energy Transfer Platform for Highly Sensitive Detection of Exosomes. Analytical Chemistry 2018, 90, 12737-12744. (91) Zhang, H.; Wang, Z.; Zhang, Q.; Wang, F.; Liu, Y. Ti3C2 MXenes nanosheets catalyzed highly efficient electrogenerated chemiluminescence biosensor for the detection of exosomes. Biosensors and Bioelectronics 2019, 124-125, 184-190. (92) Kumar, S.; Lei, Y.; Alshareef, N. H.; Quevedo-Lopez, M. A.; Salama, K. N. Biofunctionalized two-dimensional Ti3C2 MXenes for ultrasensitive detection of cancer biomarker. Biosensors and Bioelectronics 2018, 121, 243-249. (93) Dillon, A. D.; Ghidiu, M. J.; Krick, A. L.; Griggs, J.; May, S. J.; Gogotsi, Y.; Barsoum, M. W.; Fafarman, A. T. Highly Conductive Optical Quality Solution-Processed Films of 2D Titanium Carbide. Advanced Functional Materials 2016, 26, 4162-4168. (94) Hantanasirisakul, K.; Zhao, M.-Q.; Urbankowski, P.; Halim, J.; Anasori, B.; Kota, S.; Ren, C. E.; Barsoum, M. W.; Gogotsi, Y. Fabrication of Ti3C2Tx MXene Transparent Thin Films with Tunable Optoelectronic Properties. Advanced Electronic Materials 2016, 2, 1600050- n/a. (95) Xu, S.; Wei, G.; Li, J.; Ji, Y.; Klyui, N.; Izotov, V.; Han, W. Binder-free Ti3C2Tx MXene electrode film for supercapacitor produced by electrophoretic deposition method. Chemical Engineering Journal 2017, 317, 1026-1036.

(96) Romer, F. M.; Wiedwald, U.; Strusch, T.; Halim, J.; Mayerberger, E.; Barsoum, M. W.; Farle, M. Controlling the conductivity of Ti3C2 MXenes by inductively coupled oxygen and hydrogen plasma treatment and humidity. RSC Advances 2017, 7, 13097-13103. (97) An, H.; Habib, T.; Shah, S.; Gao, H.; Patel, A.; Echols, I.; Zhao, X.; Radovic, M.; Green, M. J.; Lutkenhaus, J. L. Water Sorption in MXene/Polyelectrolyte Multilayers for Ultrafast Humidity Sensing. ACS Applied Nano Materials 2019. (98) An, H.; Habib, T.; Shah, S.; Gao, H.; Radovic, M.; Green, M. J.; Lutkenhaus, J. L. Surface-agnostic highly stretchable and bendable conductive MXene multilayers. Science Advances 2018, 4. (99) Ling, Z.; Ren, C. E.; Zhao, M.-Q.; Yang, J.; Giammarco, J. M.; Qiu, J.; Barsoum, M. W.; Gogotsi, Y. Flexible and conductive MXene films and nanocomposites with high capacitance. Proceedings of the National Academy of Sciences 2014, 111, 16676-16681. (100) Zhang, H.; Wang, L.; Chen, Q.; Li, P.; Zhou, A.; Cao, X.; Hu, Q. Preparation, mechanical and anti-friction performance of MXene/polymer composites. Materials & Design 2016, 92, 682-689. (101) Suzuki, S.; Pallares, R. M.; Hibino, H. Growth of atomically thin hexagonal boron nitride films by diffusion through a metal film and precipitation. J. Phys. D: Appl. Phys. 2012, 45. (102) Sun, W.; Wang, L.; Wu, T.; Pan, Y.; Liu, G. Multi-Layer Boron Nitride Nanosheets as Corrosion-Protective Coating Fillers. Journal of the Electrochemical Society 2015, 163. (103) Pakdel, A.; Zhi, C.; Bando, Y.; Nakayama, T.; Golber, D. Boron Nitride Nanosheet Coatings with Controllable Water Repellency. ACS Nano 2011, 5, 6507–6515. (104) Lei, W.; Portehault, D.; Liu, D.; Qin, S.; Chen, Y. Porous boron nitride nanosheets for effective water cleaning. Nature Communications 2013, 4. (105) Huang, C.; Chen, C.; Ye, X.; Ye, W.; Hu, J.; Xua, C.; Qiu, X. Stable colloidal boron nitride nanosheet dispersion and its potential application in catalysis. J. Mater. Chem. A 2013, 1, 12192-12197. (106) Zhi, C.; Bando, Y.; Tang, C.; Kuwahara, H.; Golber, D. Large-Scale Fabrication of Boron Nitride Nanosheets and Their Utilization in Polymeric Composites with Improved Thermal and Mechanical Properties. Advanced Materials 2009, 21, 2889–2893.

(107) Wang, X.; Zhi, C.; Weng, Q.; Bando, Y.; Golberg, D. Boron Nitride Nanosheets: novel Syntheses and Applications in polymeric Composites. Journal of Physics: Conference Series 2013, 417. (108) Liu, F.; Mo, X.; Gan, H.; Guo, T.; Wang, X.; Chen, B.; Chen, J.; Deng, S.; Xu, N.; Sekiguchi, T.; Golberg, D.; Bando, Y. Cheap, Gram-Scale Fabrication of BN Nanosheets via Substitution Reaction of Graphite Powders and Their Use for Mechanical Reinforcement of Polymers. Scientific Reports 2014, 4. (109) Duan, Z.-Q.; Liu, Y.-T.; Xie, X.-M.; Ye, X.-Y.; Zhu, X.-D. h-BN Nanosheets as 2D Substrates to Load 0D Fe3O4 Nanoparticles: A Hybrid Anode Material for Lithium-Ion Batteries. Chemistry – An Asian Journal 2016, 11, 828-833. (110) Lin, Y.; Williams, T. V.; Xu, T.-B.; Cao, W.; Elsayed-Ali, H. E.; Connell, J. W. Aqueous Dispersions of Few-Layered and Monolayered Hexagonal Boron Nitride Nanosheets from Sonication-Assisted Hydrolysis: Critical Role of Water. The Journal of Physical Chemistry C 2011, 115, 2679-2685. (111) Parviz, D.; Irin, F.; Shah, S. A.; Das, S.; Sweeney, C. B.; Green, M. J. Challenges in liquid phase exfoliation, processing and assembly of pristine graphene. Advanced Materials 2016, in press. (112) Halim, U.; Zheng, C. R.; Chen, Y.; Lin, Z.; Jiang, S.; Cheng, R.; Huang, Y.; Duan, X. A rational design of cosolvent exfoliation of layered materials by directly probing liquid–solid interaction. Nat Commun 2013, 4. (113) Tiano, A. L.; Gibbons, L.; Tsui, M.; Applin, S. I.; Silva, R.; Park, C.; Fay, C. C. Thermodynamic approach to boron nitride nanotube solubility and dispersion. Nanoscale 2016, 8, 4348-4359. (114) Shen, J.; Wu, J.; Wang, M.; Dong, P.; Xu, J.; Li, X.; Zhang, X.; Yuan, J.; Wang, X.; Ye, M.; Vajtai, R.; Lou, J.; Ajayan, P. M. Surface Tension Components Based Selection of Cosolvents for Efficient Liquid Phase Exfoliation of 2D Materials. Small 2016, 12, 2741-2749. (115) Marsh, K. L.; Souliman, M.; Kaner, R. B. Co-solvent exfoliation and suspension of hexagonal boron nitride. Chemical Communications 2015, 51, 187-190. (116) Ni, N.; Tesconi, M.; Tabibi, S. E.; Gupta, S.; Yalkowsky, S. H. Use of pure t- butanol as a solvent for freeze-drying: a case study. International Journal of Pharmaceutics 2001, 226, 39-46.

(117) Wajid, A. S.; Das, S.; Irin, F.; Ahmed, H. S. T.; Shelburne, J. L.; Parviz, D.; Fullerton, R. J.; Jankowski, A. F.; Hedden, R. C.; Green, M. J. Polymer-stabilized graphene dispersions at high concentrations in organic solvents for composite production. Carbon 2012, 50, 526-534. (118) Tersoff, J. New empirical approach for the structure and energy of covalent systems. Physical Review B 1988, 37, 6991-7000. (119) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of simple potential functions for simulating liquid water. The Journal of Chemical Physics 1983, 79, 926-935. (120) Wang, J.; Wang, W.; Kollman, P. A.; Case, D. A. Automatic atom type and bond type perception in molecular mechanical calculations. J. Mol. Graphics Modell. 2006, 25, 247- 260. (121) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and testing of a general amber force field. J. Comput. Chem. 2004, 25, 1157-1174. (122) Jakalian, A.; Bush, B. L.; Jack, D. B.; Bayly, C. I. Fast, efficient generation of high- quality atomic charges. AM1-BCC model: I. Method. Journal of Computational Chemistry 2000, 21, 132-146. (123) Jakalian, A.; Jack, D. B.; Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. Parameterization and validation. Journal of Computational Chemistry 2002, 23, 1623-1641. (124) Hod, O. Graphite and Hexagonal Boron-Nitride have the Same Interlayer Distance. Why? Journal of Chemical Theory and Computation 2012, 8, 1360-1369. (125) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. Journal of Computational Physics 1977, 23, 327-341. (126) Hockney, R. W.; Eastwood, J. W.: Particle-Particle-Particle-Mesh (P3m) Algorithms. In Computer Simulation Using Particles; Taylor & Francis, 1988; pp 267-304. (127) Shinoda, W.; Shiga, M.; Mikami, M. Rapid estimation of elastic constants by molecular dynamics simulation under constant stress. Physical Review B 2004, 69, 134103. (128) Plimpton, S. Fast Parallel Algorithms for Short-Range Molecular Dynamics. J. Comput. Phys. 1995, 117, 1-19.

(129) Lin, P. H.; Khare, R. Local Chain Dynamics and Dynamic Heterogeneity in Cross- Linked Epoxy in the Vicinity of Glass Transition. Macromolecules 2010, 43, 6505-6510. (130) Mani, S.; Khabaz, F.; Godbole, R. V.; Hedden, R. C.; Khare, R. Structure and Hydrogen Bonding of Water in Polyacrylate Gels: Effects of Polymer Hydrophilicity and Water Concentration. J. Phys. Chem. B 2015, 119, 15381-15393. (131) Mahajan, C. V.; Ganesan, V. Atomistic Simulations of Structure of Solvated Sulfonated Poly(ether ether ketone) Membranes and Their Comparisons to Nafion: I. Nanophase Segregation and Hydrophilic Domains (vol 114, pg 8357, 2010). Journal of Physical Chemistry B 2012, 116, 5575-5576. (132) Mashtalir, O.; Naguib, M.; Mochalin, V. N.; Dall’Agnese, Y.; Heon, M.; Barsoum, M. W.; Gogotsi, Y. Intercalation and delamination of layered carbides and carbonitrides. Nature Communications 2013, 4, 1716. (133) Er, D.; Li, J.; Naguib, M.; Gogotsi, Y.; Shenoy, V. B. Ti3C2 MXene as a High Capacity Electrode Material for Metal (Li, Na, K, Ca) Ion Batteries. ACS Applied Materials & Interfaces 2014, 6, 11173-11179. (134) Hu, M.; Li, Z.; Zhang, H.; Hu, T.; Zhang, C.; Wu, Z.; Wang, X. Self-assembled Ti3C2Tx MXene film with high gravimetric capacitance. Chemical Communications 2015, 51, 13531-13533. (135) Hu, Q.; Sun, D.; Wu, Q.; Wang, H.; Wang, L.; Liu, B.; Zhou, A.; He, J. MXene: A New Family of Promising Hydrogen Storage Medium. The Journal of Physical Chemistry A 2013, 117, 14253-14260. (136) Guo, J.; Fu, H.; Zou, G.; Zhang, Q.; Zhang, Z.; Peng, Q. Theoretical interpretation on lead adsorption behavior of new two-dimensional transition metal carbides and nitrides. Journal of Alloys and Compounds 2016, 684, 504-509. (137) Levi, M. D.; Lukatskaya, M. R.; Sigalov, S.; Beidaghi, M.; Shpigel, N.; Daikhin, L.; Aurbach, D.; Barsoum, M. W.; Gogotsi, Y. Solving the Capacitive Paradox of 2D MXene using Electrochemical Quartz-Crystal Admittance and In Situ Electronic Conductance Measurements. Advanced Energy Materials 2015, 5, 1400815-n/a. (138) Luo, J.; Zhang, W.; Yuan, H.; Jin, C.; Zhang, L.; Huang, H.; Liang, C.; Xia, Y.; Zhang, J.; Gan, Y.; Tao, X. Pillared Structure Design of MXene with Ultralarge Interlayer Spacing for High-Performance Lithium-Ion Capacitors. ACS Nano 2017, 11, 2459-2469.

(139) Wang, R.; Wang, S.; Zhang, Y.; Jin, D.; Tao, X.; Zhang, L. Graphene-coupled Ti3C2 MXenes-derived TiO2 mesostructure: promising sodium-ion capacitor anode with fast ion storage and long-term cycling. Journal of Materials Chemistry A 2018, 6, 1017-1027. (140) Tang, Q.; Zhou, Z.; Shen, P. Are MXenes Promising Anode Materials for Li Ion Batteries? Computational Studies on Electronic Properties and Li Storage Capability of Ti3C2 and Ti3C2X2 (X = F, OH) Monolayer. Journal of the American Chemical Society 2012, 134, 16909- 16916. (141) Sun, D.; Wang, M.; Li, Z.; Guangxin, F.; Fan, L.-Z.; Zhou, A.: Two-Dimensional Ti3C2 as Anode Material for Li-Ion Batteries, 2014; Vol. 47. (142) Ghassemi, H.; Harlow, W.; Mashtalir, O.; Beidaghi, M.; Lukatskaya, M. R.; Gogotsi, Y.; Taheri, M. L. In situ environmental transmission electron microscopy study of oxidation of two-dimensional Ti3C2 and formation of carbon-supported TiO2. Journal of Materials Chemistry A 2014, 2, 14339-14343. (143) Halim, J.; Cook, K. M.; Naguib, M.; Eklund, P.; Gogotsi, Y.; Rosen, J.; Barsoum, M. W. X-ray photoelectron spectroscopy of select multi-layered transition metal carbides (MXenes). Applied Surface Science 2016, 362, 406-417. (144) Maleski, K.; Mochalin, V. N.; Gogotsi, Y. Dispersions of Two-Dimensional Titanium Carbide MXene in Organic Solvents. Chemistry of Materials 2017, 29, 1632-1640. (145) Zhang, C. J.; Pinilla, S.; McEvoy, N.; Cullen, C. P.; Anasori, B.; Long, E.; Park, S.-H.; Seral-Ascaso, A.; Shmeliov, A.; Krishnan, D.; Morant, C.; Liu, X.; Duesberg, G. S.; Gogotsi, Y.; Nicolosi, V. Oxidation Stability of Colloidal Two-Dimensional Titanium Carbides (MXenes). Chemistry of Materials 2017, 29, 4848-4856. (146) Ghidiu, M.; Lukatskaya, M. R.; Zhao, M.-Q.; Gogotsi, Y.; Barsoum, M. W. Conductive two-dimensional titanium carbide ‘clay’ with high volumetric capacitance. Nature 2014, 516, 78. (147) Ju, Y.; Wang, M.; Wang, Y.; Wang, S.; Fu, C. Electrical Properties of Amorphous Titanium Oxide Thin Films for Bolometric Application. Advances in Condensed Matter Physics 2013, 2013, 5. (148) Marcano, D. C.; Kosynkin, D. V.; Berlin, J. M.; Sinitskii, A.; Sun, Z.; Slesarev, A.; Alemany, L. B.; Lu, W.; Tour, J. M. Improved Synthesis of Graphene Oxide. ACS Nano 2010, 4, 4806-4814.

(149) Jung, I.; Dikin, D. A.; Piner, R. D.; Ruoff, R. S. Tunable Electrical Conductivity of Individual Graphene Oxide Sheets Reduced at “Low” Temperatures. Nano Letters 2008, 8, 4283- 4287. (150) Sobolčiak, P.; Ali, A.; Hassan, M. K.; Helal, M. I.; Tanvir, A.; Popelka, A.; Al- Maadeed, M. A.; Krupa, I.; Mahmoud, K. A. 2D Ti(3)C(2)T(x) (MXene)-reinforced polyvinyl alcohol (PVA) nanofibers with enhanced mechanical and electrical properties. PLoS One 2017, 12, e0183705. (151) Polyvinyl Compounds, Others. In Ullmann's Encyclopedia of Industrial Chemistry. (152) Li, R.; Zhang, L.; Shi, L.; Wang, P. MXene Ti3C2: An Effective 2D Light-to-Heat Conversion Material. ACS Nano 2017, 11, 3752-3759. (153) Hirakawa, T.; Nosaka, Y. Properties of O2•- and OH• Formed in TiO2 Aqueous Suspensions by Photocatalytic Reaction and the Influence of H2O2 and Some Ions. Langmuir 2002, 18, 3247-3254. (154) Gannon, C. J.; Cherukuri, P.; Yakobson, B. I.; Cognet, L.; Kanzius, J. S.; Kittrell, C.; Weisman, R. B.; Pasquali, M.; Schmidt, H. K.; Smalley, R. E.; Curley, S. A. Carbon nanotube‐ enhanced thermal destruction of cancer cells in a noninvasive radiofrequency field. Cancer 2007, 110, 2654-2665. (155) Pavlovich, C. P.; Walther, M. M.; Choyke, P. L.; Pautler, S. E.; Chang, R.; Linehan, W. M.; Wood, B. J. PERCUTANEOUS RADIO FREQUENCY ABLATION OF SMALL RENAL TUMORS: INITIAL RESULTS. The Journal of urology 2002, 167, 10-15. (156) Goldberg, S. N.; Ahmed, M.; Gazelle, G. S.; Kruskal, J. B.; Huertas, J. C.; Halpern, E. F.; Oliver, B. S.; Lenkinski, R. E. Radio-Frequency Thermal Ablation with NaCl Solution Injection: Effect of Electrical Conductivity on Tissue Heating and Coagulation—Phantom and Porcine Liver Study. Radiology 2001, 219, 157-165. (157) Gazelle, G. S.; Goldberg, S. N.; Solbiati, L.; Livraghi, T. Tumor Ablation with Radio-frequency Energy. Radiology 2000, 217, 633-646. (158) Piyasena, P.; Dussault, C.; Koutchma, T.; Ramaswamy, H. S.; Awuah, G. B. Radio Frequency Heating of Foods: Principles, Applications and Related Properties—A Review. Critical Reviews in Food Science and Nutrition 2003, 43, 587-606. (159) Marra, F.; Zhang, L.; Lyng, J. G. Radio frequency treatment of foods: Review of recent advances. Journal of Food Engineering 2009, 91, 497-508.

(160) Sano, M.; Oguma, H.; Sekine, M.; Sekiguchi, Y.; Sato, C. High-frequency welding of glass–fibre-reinforced polypropylene with a thermoplastic adhesive layer: Effects of ceramic type and long-term exposure on lap shear strength. International Journal of Adhesion and Adhesives 2015, 59, 7-13. (161) Li, C.; Dickie, R. A. Bonding adhesive joints with radio-frequency dielectric heating. International Journal of Adhesion and Adhesives 1991, 11, 241-246. (162) Leighton, J.; Brantley, T.; Szabo, E. RF welding of PVC and other thermoplastic compounds. Journal of Vinyl Technology 1993, 15, 188-192. (163) Corr, S. J.; Raoof, M.; Cisneros, B. T.; Orbaek, A. W.; Cheney, M. A.; Law, J. J.; Lara, N. C.; Barron, A. R.; Wilson, L. J.; Curley, S. A. Radiofrequency electric-field heating behaviors of highly enriched semiconducting and metallic single-walled carbon nanotubes. Nano Research 2015, 8, 2859-2870. (164) Satarkar, N. S.; Johnson, D.; Marrs, B.; Andrews, R.; Poh, C.; Gharaibeh, B.; Saito, K.; Anderson, K. W.; Hilt, J. Z. Hydrogel‐MWCNT nanocomposites: Synthesis, characterization, and heating with radiofrequency fields. Journal of Applied Polymer Science 2010, 117, 1813-1819. (165) Sweeney, C. B.; Moran, A. G.; Gruener, J. T.; Strasser, A. M.; Pospisil, M. J.; Saed, M. A.; Green, M. J. Radio Frequency Heating of Carbon Nanotube Composite Materials. ACS Applied Materials & Interfaces 2018, 10, 27252-27259. (166) Cao, M.-S.; Cai, Y.-Z.; He, P.; Shu, J.-C.; Cao, W.-Q.; Yuan, J. 2D MXenes: Electromagnetic property for microwave absorption and electromagnetic interference shielding. Chemical Engineering Journal 2019, 359, 1265-1302. (167) Liu, R.; Li, W. High-Thermal-Stability and High-Thermal-Conductivity Ti3C2Tx MXene/Poly(vinyl alcohol) (PVA) Composites. ACS Omega 2018, 3, 2609-2617. (168) Sobolčiak, P.; Ali, A.; Hassan, M. K.; Helal, M. I.; Tanvir, A.; Popelka, A.; Al- Maadeed, M. A.; Krupa, I.; Mahmoud, K. A. 2D Ti3C2Tx (MXene)-reinforced polyvinyl alcohol (PVA) nanofibers with enhanced mechanical and electrical properties. PLoS One 2017, 12, e0183705. (169) Sweeney, C. B.; Lackey, B. A.; Pospisil, M. J.; Achee, T. C.; Hicks, V. K.; Moran, A. G.; Teipel, B. R.; Saed, M. A.; Green, M. J. Welding of 3D-printed carbon nanotube–polymer composites by locally induced microwave heating. Science Advances 2017, 3, e1700262.

(170) Habib, T.; Zhao, X.; Shah, S. A.; Chen, Y.; Sun, W.; An, H.; Lutkenhaus, J. L.; Radovic, M.; Green, M. J. Oxidation stability of Ti3C2Tx MXene nanosheets in solvents and composite films. npj 2D Materials and Applications 2019, 3, 8. (171) Peng, Q.; Guo, J.; Zhang, Q.; Xiang, J.; Liu, B.; Zhou, A.; Liu, R.; Tian, Y. Unique Lead Adsorption Behavior of Activated Hydroxyl Group in Two-Dimensional Titanium Carbide. Journal of the American Chemical Society 2014, 136, 4113-4116. (172) Guo, J.; Peng, Q.; Fu, H.; Zou, G.; Zhang, Q. Heavy-Metal Adsorption Behavior of Two-Dimensional Alkalization-Intercalated MXene by First-Principles Calculations. The Journal of Physical Chemistry C 2015, 119, 20923-20930. (173) Guo, X.; Zhang, X.; Zhao, S.; Huang, Q.; Xue, J. High adsorption capacity of heavy metals on two-dimensional MXenes: an ab initio study with molecular dynamics simulation. Physical Chemistry Chemical Physics 2016, 18, 228-233. (174) Ying, Y.; Liu, Y.; Wang, X.; Mao, Y.; Cao, W.; Hu, P.; Peng, X. Two-Dimensional Titanium Carbide for Efficiently Reductive Removal of Highly Toxic Chromium(VI) from Water. ACS Applied Materials & Interfaces 2015, 7, 1795-1803. (175) Hantanasirisakul, K.; Zhao, M.-Q.; Urbankowski, P.; Halim, J.; Anasori, B.; Kota, S.; Ren, C. E.; Barsoum, M. W.; Gogotsi, Y. Fabrication of Ti3C2Tx MXene Transparent Thin Films with Tunable Optoelectronic Properties. Advanced Electronic Materials 2016, 2, n/a-n/a. (176) Peng, X.; Peng, L.; Wu, C.; Xie, Y. Two dimensional nanomaterials for flexible supercapacitors. Chemical Society Reviews 2014, 43, 3303-3323. (177) Halim, J.; Lukatskaya, M. R.; Cook, K. M.; Lu, J.; Smith, C. R.; Näslund, L.-Å.; May, S. J.; Hultman, L.; Gogotsi, Y.; Eklund, P. Transparent conductive two-dimensional titanium carbide epitaxial thin films. Chemistry of Materials 2014, 26, 2374-2381. (178) Lukatskaya, M. R.; Mashtalir, O.; Ren, C. E.; Dall’Agnese, Y.; Rozier, P.; Taberna, P. L.; Naguib, M.; Simon, P.; Barsoum, M. W.; Gogotsi, Y. Cation Intercalation and High Volumetric Capacitance of Two-Dimensional Titanium Carbide. Science 2013, 341, 1502-1505. (179) Dall'Agnese, Y.; Lukatskaya, M. R.; Cook, K. M.; Taberna, P.-L.; Gogotsi, Y.; Simon, P. High capacitance of surface-modified 2D titanium carbide in acidic electrolyte. Electrochemistry Communications 2014, 48, 118-122.

(180) Xie, X.; Xue, Y.; Li, L.; Chen, S.; Nie, Y.; Ding, W.; Wei, Z. Surface Al leached Ti3AlC2 as a substitute for carbon for use as a catalyst support in a harsh corrosive electrochemical system. Nanoscale 2014, 6, 11035-11040. (181) Shahzad, F.; Alhabeb, M.; Hatter, C. B.; Anasori, B.; Hong, S. M.; Koo, C. M.; Gogotsi, Y. Electromagnetic interference shielding with 2D transition metal carbides (MXenes). Science 2016, 353, 1137-1140. (182) Lei, J.-C.; Zhang, X.; Zhou, Z. Recent advances in MXene: Preparation, properties, and applications. Frontiers of Physics 2015, 10, 276-286. (183) Enyashin, A. N.; Ivanovskii, A. L. Atomic structure, comparative stability and electronic properties of hydroxylated Ti2C and Ti3C2 nanotubes. Computational and Theoretical Chemistry 2012, 989, 27-32. (184) Parviz, D.; Metzler, S. D.; Das, S.; Irin, F.; Green, M. J. Tailored Crumpling and Unfolding of Spray-Dried Pristine Graphene and Graphene Oxide Sheets. Small 2015, 11, 2661- 2668. (185) Luo, J.; Jang, H. D.; Huang, J. Effect of Sheet Morphology on the Scalability of Graphene-Based Ultracapacitors. ACS Nano 2013, 7, 1464-1471. (186) Zou, Y.; Kinloch, I. A.; Dryfe, R. A. W. Nitrogen-doped and crumpled graphene sheets with improved supercapacitance. Journal of Materials Chemistry A 2014, 2, 19495-19499. (187) Li, Y.; Lv, X.; Lu, J.; Li, J. Preparation of SnO2-Nanocrystal/Graphene-Nanosheets Composites and Their Lithium Storage Ability. The Journal of Physical Chemistry C 2010, 114, 21770-21774. (188) Zhou, G.; Wang, D.-W.; Li, F.; Zhang, L.; Li, N.; Wu, Z.-S.; Wen, L.; Lu, G. Q.; Cheng, H.-M. Graphene-Wrapped Fe3O4 Anode Material with Improved Reversible Capacity and Cyclic Stability for Lithium Ion Batteries. Chemistry of Materials 2010, 22, 5306-5313. (189) Mashtalir, O.; Naguib, M.; Mochalin, V. N.; Dall’Agnese, Y.; Heon, M.; Barsoum, M. W.; Gogotsi, Y. Intercalation and delamination of layered carbides and carbonitrides. Nat Commun 2013, 4, 1716. (190) Naguib, M.; Gogotsi, Y. Synthesis of Two-Dimensional Materials by Selective Extraction. Accounts of Chemical Research 2015, 48, 128-135.

100

(191) Gao, H.; Benitez, R.; Son, W.; Arroyave, R.; Radovic, M. Structural, physical and mechanical properties of Ti3(Al1−xSix)C2 solid solution with x=0–1. Materials Science and Engineering: A 2016, 676, 197-208. (192) Li, Z.; Wang, L.; Sun, D.; Zhang, Y.; Liu, B.; Hu, Q.; Zhou, A. Synthesis and thermal stability of two-dimensional carbide MXene Ti3C2. Materials Science and Engineering: B 2015, 191, 33-40. (193) Halim, J.; Lukatskaya, M. R.; Cook, K. M.; Lu, J.; Smith, C. R.; Näslund, L.-Å.; May, S. J.; Hultman, L.; Gogotsi, Y.; Eklund, P.; Barsoum, M. W. Transparent Conductive Two- Dimensional Titanium Carbide Epitaxial Thin Films. Chemistry of Materials 2014, 26, 2374-2381. (194) Wang, X.; Shen, X.; Gao, Y.; Wang, Z.; Yu, R.; Chen, L. Atomic-Scale Recognition of Surface Structure and Intercalation Mechanism of Ti3C2X. Journal of the American Chemical Society 2015, 137, 2715-2721. (195) Lukatskaya, M. R.; Halim, J.; Dyatkin, B.; Naguib, M.; Buranova, Y. S.; Barsoum, M. W.; Gogotsi, Y. Room-Temperature Carbide-Derived Carbon Synthesis by Electrochemical Etching of MAX Phases. Angewandte Chemie International Edition 2014, 53, 4877-4880.

101

APPENDIX – TEMPLATE-FREE 3D TITANIUM CARBIDE (TI3C2TX) MXENE PARTICLES CRUMPLED BY CAPILLARY FORCES§** Summary

MXenes, such as Ti3C2Tx, are an exciting new class of 2D materials. However, little has been reported on manipulating shape of MXene nanosheets. Herein, we have demonstrated that the flat Ti3C2Tx encapsulated within spray-dried droplets could be scrolled, bent, and folded into a 3D crumpled structure. This morphological change was observed to be reversible upon rehydration.

Introduction

MXenes are layered, two-dimensional structures with general formula of Mn+1XnTx where M is an early transition metal, X is carbon and/or nitrogen, T is terminal group (-F, -OH, -O, etc.), x is number of terminal groups and n=1-3. Their metallic structure makes them highly electrically conductive, but they are also hydrophilic due to terminal groups on their surface layers. The latter, makes them easily dispersible in water, unlike many other nanosheet types.72 Since the first MXenes were discovered in 2011, they have showed great promise for a wide range of applications including energy storage devices, batteries and supercapacitors, transparent electronics,93,99,175-179 catalyst support,84,180 lead absorption,171 electromagnetic interference shielding88,181 and water desalination86 due to their unique dielectric, transport and chemical properties.62,182 There has been a significant interest in producing MXene-based 3D structures and 3D particles. Macroscale 3D structures have been formed by drop casting MXene nanosheet dispersions on a nickel template.134 However, little is known about the possibility of directly altering MXene morphology to produce bent, scrolled, or crumpled structures at the microscale, although there have been isolated reports of MXene nanoscrolls.62,72,183 One means to create such structures is to use capillary forces to locally bend nanosheets through spray drying.27,184; this has been recently demonstrated for graphene, boron nitride, and transition metal dichalcogenides. Dispersions of nanosheets are

§ Reprinted with permission from S.A. Shah, T. Habib, H. Gao, P. Gao, W. Sun, M.J. Green, and M. Radovic. Chem. Commun., 2017,53, 400-403 . Copyright © The Royal Society of Chemistry 2017 ** T. Habib and S. A. Shah contributed equally

102 aerosolized, and the nanosheets are entrapped and compressed by capillary forces at the interface of the evaporating droplets.27,184 Our prior work has shown that the mechanism of the nanosheet crumpling process is heavily influenced by the bending modulus of the nanosheets. This crumpled morphology prevents restacking and increases the porosity and accessible surface area of the resulting powder consisting of crumpled nanosheets. This, in turn, can highly improve electrochemical energy storage properties of 2D materials.185-188 In this paper, we utilized a spray drying method to successfully demonstrate scrolling and crumpling of Ti3C2Tx MXene nanosheets (Figure A-1). We further demonstrate that the crumpling mechanism is strongly affected by (1) high bending stiffness of MXenes and (2) concentration of dispersed MXenes.

Figure A-1: Schematic showing synthesis of Ti3C2Tx nanosheets from parent MAX phase to nanosheet crumpling via spray drying. Layered MAX phase (Ti3AlC2) is etched using HCl + LiF to obtain Ti3C2Tx clay. This clay is intercalated with DMSO and sonicated to obtain delaminated Ti3C2Tx nanosheet dispersion. This dispersion is spray dried to obtain crumpled Ti3C2Tx nanosheets. The crumpling process on the far right of the figure shows possible crumping mechanism for Ti3C2Tx nanosheets

Materials & Methods

Synthesis of Ti3AlC2 MAX phase

103

Commercial Ti (44 μm average particle size, 99.5% purity), Al (44 μm average particle size, 99.5% purity) and TiC powders (2 – 3 μm average particle size, 99.5% purity), (all from Alfa

(PECS). To fabricate high-purity Ti3AlC2 powder, the PECSed sample was first drill-milled and sieved to obtain powder with particle sizes below 44 μm.

Synthesis of Ti3C2Tx MXene clay

Ti3C2Tx MXene clay was synthesized by etching aluminum from the MAX phase using technique described by Ghidiu et.al.74 Concentrated hydrochloric acid (HCl, ACS reagent, 37% w/w Sigma-Aldrich) was diluted with DI water to obtain 30 mL of 6 M HCl solution. This solution was transferred to a polypropylene (PP) beaker and 1.98 gm of lithium fluoride (LiF, 98+% purity, Alfa Aesar) was added to it. This dispersion was stirred for 5 minutes using a Teflon (PTFE) magnetic stirrer at room temperature. Ti3AlC2 MAX phase powder was slowly added to the HCl+LiF solution to prevent overheating as the reaction is exothermic. The PP beaker was capped to prevent evaporation of water and a hole was made in the cap to avoid buildup of hydrogen gas. The reaction mixture was stirred at 40 ºC for about 45 hours. The slurry product was filtered and washed with deionized (DI) water in a polyvinyl-difluoride (PVDF) filtration unit with pore size of 0.22 µm (Millipore® SCGVU10RE Stericup™ GV) to remove the unreacted HF and water soluble salts. This washing process was repeated until pH of the filtrate reached a value of about

6. Reaction product collected over the PVDF filter is extracted as Ti3C2Tx MXene clay.

Intercalation and delamination of Ti3C2Tx MXene clay

Aldrich) was added to Ti3C2Tx MXene clay (dried in vacuum oven for about 24 hours at 40 ˚C) to form a 60 mg/ml suspension followed by about 18 hours of stirring at room temperature. After intercalation, excess DMSO was removed by several cycles of washing with DI water and

104 centrifugation at 5000 rpm for 4 hours. The intercalated Ti3C2Tx MXene clay suspension in DI water was bath sonicated for 1 hour at room temperature followed by centrifugation at 3500 rpm for 1 hour to separate the heavier components.

Crumpling of Ti3C2Tx MXene nanosheet dispersion

Crumpling of Ti3C2Tx MXene nanosheets was achieved by spray drying delaminated 184 MXene dispersion. Aqueous dispersion of delaminated Ti3C2Tx MXene was diluted to a concentration of 1 mg/ml, and processed in Buchi B-290 mini spray dryer, Figure A-1. The spray drying procedure involved conveying the dispersion using a peristaltic pump to an atomizer where it was mixed with in house air to form micrometer sized droplets. For our experiments, we used a pump flow rate of 10 % (of maximum possible flow rate) and atomizer air pressure of 60 psi. The droplets formed by the atomizer were carried and dried by co-currently flowing hot air in the drying chamber. This airflow was created by an aspirator pump by induced draft mechanism and heating occurred via a heating coil. For our experiments, the highest drying air flow rate was used by operating aspirator at 100% power, and inlet air temperature was maintained at 220 ˚C. Carrier gas loses its heat to droplets causing water to evaporate and the dispersion to dry. The dried particles are eventually collected in a cyclone separator and are stored for further analysis. Above mentioned procedure was also carried out at a starting concentration of 0.1 mg/ml to analyze the effect of concentration on morphology of dried MXenes.

Freeze Drying

Delaminated Ti3C2Tx MXene dispersion was frozen in a freezer overnight and freeze dried for roughly 48 hours in Labconco FreeZone benchtop freeze dryer to obtain dry MXene nanosheet powder. Transmission Electron Microscopy (TEM) TEM images were obtained using FEI Tecnai G2 F20 field emission transmission electron microscope (FE-TEM). All samples were deposited on 200 mesh holey carbon-coated copper grids of 100 µm (HC200-CU-100, Electron Microscopy Sciences) for imaging. Powdered crumpled

Ti3C2Tx MXene samples from spray dryer were directly placed on the grid for imaging. Water rehydrated crumpled MXenes were drop casted on the grid followed by air drying for about 5 minutes.

Scanning Electron Microscopy (SEM)

105

SEM images were obtained using FEI Quanta 600 field emission scanning electron microscope (FE-SEM). For imaging, delaminated native Ti3C2Tx MXene samples were prepared by drop casting their dispersion on a silicon wafer. Crumpled MXene samples were prepared by directly placing spray dried powder on carbon tape.

X-Ray Photoelectron Spectroscopy (XPS)

XPS measurements were performed using an Omicron XPS system with Mg x-ray source.

Sample preparation for XPS was done by drop casting moistened crumpled Ti3C2Tx MXene powder on hydrophilic (oxygen plasma treated) silicon wafer followed by drying in a vacuum oven overnight at 40 °C. Deconvolution was performed using CasaXPS software version 2.3.16.

X-Ray Diffraction (XRD)

Bruker D8 powder X-Ray diffractometer fitted with LynxEye detector, in a Bragg Brentano geometry with CuKα (λ: 1.5418 Å) radiation was used to obtain XRD patterns of powder samples.

Freeze dried Ti3C2Tx MXene powder was placed on a zero background holder to obtain its XRD pattern. Similarly, crumpled MXene powder was also placed on the same holder to measure its XRD pattern. The X-ray scan was performed with a step size of 0.02º and a scan rate of 1 s per step.

UV-Visible Spectroscopy

Absorbance spectra of Ti3C2Tx MXene dispersions were measured using Shimazdu UV- vis spectrophotometer 2550 (Wavelength range: 200 - 800 nm). Samples were placed in a quartz cell with path length of 1 cm, and DI water was used as a blank. Concentrations of MXene dispersions were determined using Beer-Lambert law. The extinction coefficient of MXene dispersions was calculated to be 1167.2 ml.(mg.m)-1 at 580 nm. See Figure A-2 for more detail.

106

Figure A-2: UV-Vis absorbance spectra are shown for Ti3C2Tx dispersions at different concentrations (measured using vacuum filtration). Calibration curve (inset) was made using the absorbance values at 580 nm to measure concentration of subsequent Ti3C2Tx dispersions.

Zeta Potential

Malvern Zetasizer ZS90 and the appropriate capillary cell (DTS1070) was used to measure the zeta potential of the stable Ti3C2Tx MXene dispersions.

Results & Discussion MXenes are derived from MAX phases by etching (using acid) out the “A” layer (a group 13 or 14 elements) from the layered MAX phases. Many MXenes have been reported experimentally and many more are expected to be stable but are yet to be synthesized.190

For this study, the Ti3C2Tx (MXene phase) was derived from Ti3AlC2 (parent MAX 191 phase ) by etching Ti3AlC2 to remove Al. This is typically done using concentrated HF aqueous solution; however, mixtures of LiF and HCl aqueous solution were recently shown to be effective as well.74 We utilize this latter method in this study. After etching the A 72 layer away, the Ti3C2Tx structures possess a multi-layer accordion-like morphology. We then intercalate this structure with dimethylsulfoxide (DMSO) to aid in delamination. The DMSO was then solvent-exchanged to water and bath sonicated; the resulting

107 dispersion was then centrifuged to eliminate unexfoliated material. The supernatant was collected as a colloidally stable few-layered Ti3C2Tx nanosheet dispersion, with typical concentrations of ~1.5 mg/mL as measured by UV-vis absorbance (Figure A-2). These colloids are stabilized electrostatically by the terminating groups, with typical ζ (zeta) potential values of approximately -32 mV (Figure A-3). More details on etching, delamination and dispersion of Ti3C2Tx nanosheets are provided in supplementary information.

Figure A-3: (a) Zeta potential of (b) aqueous Ti3C2Tx nanosheet dispersion being stored in glass container shows its colloidal stability. (c) Colloidal nature of the dispersion can be verified by Tyndall effect (laser being scattered by dispersed MXene nanosheets).

The stable Ti3C2Tx dispersion was diluted to a concentration of 0.1 mg/ml and fed to the spray dryer. The dispersion is aerosolized at an aspirator pressure of 60 psi and dried using in-house air as a carrier gas at 220 oC. The resulting dry powder is collected using a cyclone separator. As described above, the capillary forces on the evaporating droplets crumple the nanosheets during droplet evaporation. Scanning electron microscopy (SEM) images show clear distinctions in morphology between the native, flat MXene nanosheets (Figure A-4a) and the crumpled nanosheet powder (Figure A-4b).

X-ray diffraction (XRD) patterns of these crumpled Ti3C2Tx MXenes (spray dried) are compared to those of the flat, native Ti3C2Tx MXenes (freeze dried) and the parent

Ti3AlC2 MAX phase in Figure A-4c. The initial etching process alters Ti3AlC2 to Ti3C2Tx,

108 and the XRD spectrum for Ti3C2Tx is consistent with prior studies, showing pronounced diffraction peak at around 2θ = 6.3o - 7.1o.74,192 Comparison of position of XRD peaks for native and crumpled MXenes in Figure A-4c, suggests that crumpling process does not appreciably affect the crystal structure of the nanosheets. However, the first peak (7.1°) for the crumpled MXene is broadened, most likely due to stresses induced by the crumpling process.

Figure A-4: SEM images of Ti3C2Tx MXene nanosheets in their (a) native and (b) crumpled morphology (spray dried at a concentration of 1 mg/ml), and (c) XRD spectra of Ti3AlC2 MAX phase powder, flat native Ti3C2Tx MXene nanosheet powder, and crumpled Ti3C2Tx MXene nanosheet powder.

The composition of crumpled MXene powder was probed using X-ray photoelectron spectroscopy (XPS) analysis. Survey spectra (Figure A-5) shows presence of titanium, carbon, oxygen and fluorine and the percentage atomic compositions are reported in Table

109

A-1. Peaks corresponding to sulfur, lithium or chlorine were not detected in the XPS spectrum, indicating that these species were removed in the final product. As shown in Figure A-6, region corresponding to each element was deconvoluted and the results are listed in Table A-1. Binding energy values of components were found to be in accordance 88,143,193,194 with previous XPS studies on Ti3C2Tx MXenes. Our results indicated the presence of three type of surface functional groups on crumpled Ti3C2Tx MXenes: oxide (- O-), hydroxyl (-OH) and fluoride (-F). 56% of carbon was composed of graphitic C-C which may have formed due to Ti etching during the MXene synthesis process.195 About

25% of titanium was estimated to be oxidized to TiO2.

Figure A-5: XPS spectrum of crumpled Ti3C2Tx MXenes labelled with characteristic peaks of Ti, C, O and F

110

Table A-1: XPS peak fitting results for crumpled Ti3C2Tx MXenes

Overall Component Component Element BE (eV) FWHM (eV) Atomic % Name Atomic % Ti-C 6.30 454.9 (461) 0.9 (2) 2+ Ti 51.94 455.7 (461.5) 2.1 (2.4) Ti 2p 3+ 3/2 18.21 Ti 13.04 457.4 (463.2) 1.6 (1.7) (2p1/2) TiO 2 24.49 459 (464.7) 1.6 (2) C-Ti-F x 4.23 460.4 (466.2) 1.8 (1.7) (†) C-Ti-T 281.6 0.8 x 23.97 (†) C-Ti-T x 282.4 1.3 C 1s 43.88 C-C 56.23 284.7 1.6 CH /CO x 15.37 286.3 1.6 COO 4.43 288.7 1.4 TiO 2 24.45 529.7 1.5 C-Ti-O x 25.46 530.3 1.6 C-Ti-(OH) O 1s 25.33 x 21.89 531.3 1.4 Al O 2 3 11.62 532.2 1.1 (*)H O 2 16.58 533.1 1.6 AlF x 12.29 686.2 1.5 F 1s 12.58 C-Ti-F x 87.71 685.1 1.8

111

Figure A-6: XPS spectra of crumpled Ti3C2Tx MXenes. (a) Ti 2p (b) C 1s (c) O 1s (d) F 1s. Binding energy values of each bond associated with deconvoluted peaks are listed in Table A-1

At higher MXene concentration in the feed dispersion, the degree of crumpling in the as-obtained powder is significantly decreased, as seen in Figure A-7a-b. This trend may be explained as follows: with increased concentration, the number of nanosheets per droplet increases, such that multi-nanosheet shells may form during evaporation. These thicker nanosheet shells would have a higher effective bending modulus and crumple less than shells with fewer nanosheets, resulting in a less compact structure. Inspection of the crumpled nanosheet edges reveal far fewer nanosheets in Figure A-7a (0.1 mg/mL) than in Figure A-7b (1 mg/mL). Similarly, at the edges of the structure, there is a significant

112 degree of scrolling rather than crumpling in Figure A-7b. This presents a possible route to control the morphology of dry MXene nanosheets. We also compared the effect of out-of- plane bending modulus of Ti3C2Tx MXenes with other nanosheets such graphene oxide (GO) (Figure A-8). The increased number of atomic layers (5) would result in a higher stiffness of MXenes when compared to the single-layer GO.

Figure A-7: TEM images of crumpled Ti3C2Tx MXene spray dried at a concentration of (a) 0.1 mg/ml, (b) 1mg/ml [zoomed in inset shows few layers and scrolling] and (c) “rehydrated” crumpled MXene from dispersion spray dried at 1 mg/ml.

113

Figure A-8: SEM/TEM images of (a) crumpled GO and (b) crumpled Ti3C2Tx. Crumpled GO shows significant amount of ridges whereas crumpled Ti3C2Tx shows folds with relatively larger local radius of curvature

To determine whether this crumpled morphology is reversible, crumpled Ti3C2Tx was rehydrated in water using a vortex mixer and allowed to settle for two hours. The supernatant was drop cast on a TEM grid and analyzed. From Figure A-7c, the lack of folding, bending, and scrolling sites indicates that this structure has “uncrumpled” to a state closer to the pristine MXenes. This indicates (i) that the capillary forces induce few permanent defects or covalent bonds in the MXene structure and (ii) that the MXene structure remains hydrophilic due to terminal groups on the nanosheet surface, even after crumpling. The latter point is confirmed by the XPS data showing presence of C-Ti-OH terminal groups and C-Ti-Ox sites after rehydration of crumpled MXenes. Conclusion

In this study, we have successfully converted Ti3C2Tx from 2D flat nanosheet to 3D crumpled structure via spray drying method, without the assistance of any templates. The extent of crumpling can be controlled by changing the feed concentration; higher concentration leads to decreased crumpling. The results also showed that crumpled nanosheet can be “uncrumpled” if rehydrated, indicating the change in structure is a reversible process.

114