Counterion Size and Nature Control Structural and Mechanical

Subscriber access provided by Iowa State University | Library

Article

Counterion size and nature control structural and mechanical response in cellulose nanofibril nanopapers Alejandro J. Benitez, and Andreas Walther Biomacromolecules, Just Accepted Manuscript • DOI: 10.1021/acs.biomac.7b00263 • Publication Date (Web): 28 Mar 2017 Downloaded from http://pubs.acs.org on March 29, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biomacromolecules is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

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

Biomacromolecules

Counterion size and nature control structural and mechanical response in cellulose nanofibril nanopapers

Alejandro J. Benítez, Andreas Walther Institute for Macromolecular Chemistry, Stefan-Meier-Str. 31, Albert-Ludwigs-University Freiburg, 79104 Freiburg, Germany. & Freiburg Materials Research Center, Stefan-Meier-Str. 21, Albert-Ludwigs-University Freiburg, 79104 Freiburg, Germany. & Freiburg Center for Interactive Materials and Bioinspired Technologies, Georges-KöhlerAllee 105, Albert-Ludwigs-University Freiburg, 79110 Freiburg, Germany. E-mail: [email protected]

KEYWORDS: nanocellulose, nanopaper, colloid stability, counterion, toughness, porosity.

1 ACS Paragon Plus Environment

Biomacromolecules

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

Page 2 of 35

Abstract: Nanopapers formed from aqueous dispersions of cellulose nanofibrils (CNFs) combine stiffness, strength and toughness. Yet, delicate interactions operate between the CNFs during nanopaper formation and mechanical deformation. We unravel in detail how counterions, being either of the organic alkyl ammonium kind (NR4+) or of the earth metal series (Li+, Na+, Cs+), need to be chosen to achieve outstanding combinations of stiffness, strength and toughness, extending to previously unreached territories. We relate structure formation processes in terms of colloidal stabilization to nanostructural details such as porosity and ability for swelling, as well as to interfibrillar interactions in bulk and macroscale mechanical properties. We demonstrate that our understanding also leads to new levels of ductility in bioinspired CNF/polymer nanocomposites at high levels of reinforcements. These results contribute to future rational design of CNF-based high-performance materials.


Page 3 of 35

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

Biomacromolecules

INTRODUCTION Cellulose nanofibrils (CNFs) emerge as an attractive building block for sustainable nanoscience applications due their renewable origin, simple preparation, easy functionalization, and facile processing into films, fibers, nanocomposites and 3D-printed materials.1-11 CNFs can form nanopapers with excellent mechanical properties by drying from aqueous suspension, whereupon the nanofibrils form an entangled network (colloidal glass) with strong interfibrillar hydrogen bonding.11-16 Various factors influence the final structure and interactions. For instance, the fibrillation method controls surface chemistry, aspect ratio, size distribution and crystallinity of the nanofibers.17 The film preparation method, i.e. film casting or filtration, addition of stabilizing polymers, and the pH strongly influence the nanopaper structure beyond simple entangled networks.11,

16, 18-20

Preventing excess flocculation during the processing aids in

realizing better mechanical properties (in particular strain-to-failure, toughness).12, 16, 18 While there are increasing articles addressing mechanical properties, it is still a challenge to comprehensively understand the design criteria for a rational mechanical performance increase of nanopapers. For example, rationally designed enhancements in toughness by combining CNFs with polymers can be brought forward by changing the glass transitions temperatures or the volume fraction of the polymers, as long as coagulation or ill-defined structures are prevented.11, 21, 22

The general intention is to introduce soft components capable of improving frictional

interfibrillar sliding. It was also clearly shown that ill-defined nanopaper structures formed from coagulated dispersions lead to inferior properties, while tailored porosities in defined networks can enhance ductility, yet on expense of stiffness and strength.11-13, 16, 19, 23 Increase in strength and stiffness by combining CNFs with multivalent metal ions, clay or carbon nanotubes has been a viable approach.24-29 Furthermore, orientation by cold drawing or wet stretching has been a key approach to enhance mechanical properties in a unidirectional way by alignment of the CNF.7, 2830

Two key aspects emerge when aiming at improved mechanical properties, that is well-defined structure formation and providing components to manage the interfibrillar interactions.4, 12, 18, 31 For both factors it is important to control the colloidal solution properties, that primarily depend upon the chemical pretreatment during pulp processing.32 TEMPO oxidation is one of the most common and straightforward pretreatments, leading to very thin CNFs having carboxylate 3 ACS Paragon Plus Environment

Biomacromolecules

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

Page 4 of 35

groups on the surface that introduce pH-dependent electric double layer repulsion in dispersions and improve compatibility in bulk materials.5,

33, 34

Nanopapers derived from TEMPO-CNF

exhibit remarkable mechanical properties with a Young’s modulus between 13-15 GPa and strength of 230-300 MPa.5, 16 Recently, Shimizu et al. showed that ammonium counterions with different alkyl chain length (methyl to butyl) can lead to elongation enhancements by providing interfibrillar slippage. However, the results indicated a non-linear relationship between the alkyl chain length and the mechanical properties. Maximum elongation was found for ethyl (and not butyl), which is counterintuitive and remains difficult to understand.35 Hence, while the toughness enhancements are promising, more understanding on the origins of such non-linear behavior is very desirable. At the same time, we hypothesized that the modification of the ionic binding strength between the carboxylic groups on the CNF surface and their counterions may be a crucial aspect to modify the systems behavior beyond interfibrillar interactions in bulk. It also determines CNF suspension stability and hence structure formation during filtration. In a previous study, we demonstrated the importance of the colloidal state by showing that CNF suspensions at low pH form aggregates and flocs that are translated to grains inside the nanopapers reducing density, strength and toughness.16 Alkali metal counterions (Li+, Na+, K+, Cs+) are known to have a complex adsorption behavior that often depends on the detailed surface under study, whether hydrophilic or hydrophobic and what other chemical groups are present. Hofmeister series on hydrophilic nanoparticles in suspensions show higher affinity for the smaller counterions.36, 37 Since colloidal stability is decisive in defining structures and porosity during nanopaper formation or filtration, it is fundamentally important to comprehensively understand mechanical behavior beyond reduction to interfibrillar interactions. Herein, we present a detailed study on the effect of different counterions - a full series of alkali metal ions and alkyl ammonium ions - on the colloidal solution properties, the structural and mechanical response, and deformation mechanisms of bulk nanopapers prepared by vacuum filtration. From thermoporosity, swollen nanopapers present a pore formation capability depending on the counterion size, showing that pores act as potential energy dissipators during deformation to improve toughness. We present a systematic study of the colloid stability by checking light transmittance, total interaction potential energy from DLVO theory, aggregation 4 ACS Paragon Plus Environment

Page 5 of 35

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

Biomacromolecules

by salinity, and sedimentation speed to differentiate dispersion stabilities and explain the observed nanostructures and properties from different angles. Finally, a comparison with CNF/polymer bionanocomposites illustrates advantages of the counterion exchange and link advances in nanopapers to advantages in nanocomposite science. EXPERIMENTAL PART Materials and Counterion Exchange Never dried TEMPO-oxidized Kraft pulp oxidized under neutral conditions was diluted to 1.4 wt% and set to pH 9 with NaOH. The content of carboxyl groups is 0.44 mmol/g, the degree of polymerization measured by viscosimetry (DPv) is 725

and crystallinity is 77%.11 The

suspension is called CNF-Na+, and was homogenized in a microfluidizer MRT CR5 by applying four shear cycles (1400, 1000, 1000 and 1000 bar). Standard sodium hydroxide (NaOH, 1 M), standard hydrochloric acid (HCl 1 M), lithium hydroxide monohydrate (LiOH•H2O, solid), cesium hydroxide monohydrate (CsOH•H2O, solid), ammonium hydroxide (AOH, 25 wt% in H2O), tetraethyl ammonium hydroxide (TEAOH, 40 wt% in H2O), tetrabutyl ammonium hydroxide (TBAOH, 40 wt% in H2O), tetrahexyl ammonium hydroxide (THAOH, 40 wt% in H2O), sodium chloride (NaCl), lithium chloride (LiCl), cesium chloride (CsCl), ammonium chloride (ACl), tetraethyl ammonium chloride (TEACl), tetrabutyl ammonium chloride (TBACl) and tetrahexyl ammonium chloride (THACl) were purchased from Sigma-Aldrich in the highest purity available. The Na+ ions in the original CNF-Na+ suspension were exchanged as follows: (i) 0.25 wt% of CNF-Na+ dispersions were prepared by addition of MilliQ water, (ii) adjustment to pH = 3 with HCl to form CNF-COOH, (iii) centrifugation for 60 min at 8000 rpm to collect a pellet of protonated CNF-COOH and remove excess NaCl (this step was done twice), (iv) redispersion of CNF-COOH to pH 9 with hydroxide solutions of the different counterions forming CNF-Me+ or CNF-QA+ for metals and quaternized ammonium counterions, respectively. Then, (v) the dispersions were placed in dialysis at pH 9 (during 72 hours) with the corresponding hydroxide solution to assure maximum ion exchange and removal of impurities. Finally, (vi) the CNFs dispersions were filtered using a 50 µm pore size mesh sieve to remove dust and other large particles, followed by passing the dispersions through a 20 µm pore size mesh. 5 ACS Paragon Plus Environment

Biomacromolecules

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

Page 6 of 35

Atomic Force Microscopy (AFM) A NanoScope V AFM (Digital Instruments Veeco Instruments, Santa Barbara, CA) operating in tapping mode was used to record the micrographs. The samples were obtained by dip-coating from diluted suspensions in water (ca. 0.005 wt%) onto freshly cleaved mica. Preparation of the nanopapers Vacuum filtration (nanopapers): Typically, 250 mL of a 0.25 wt% dispersions were filtered through 1.0 µm Omnipore filters during ca. 24 hours. Afterwards, the hydrogel films were dried in an oven at 60 °C for 5 days under pressure (approx. 2.5 kPa) to yield films with thicknesses of 30-40 µm. The densities were measured by weighing a piece of nanopapers of known dimensions at 50% RH. Adsorbed water is included in the density value. Solvent casting (polymer/CNF nanocomposites): 60 mL of the 0.25 wt% suspensions were added to petri dishes at room temperature to give films of ca. 22 µm thickness. Elemental Analysis Alkali earth metals content was determined by inductively coupled plasma optical emission spectrometry (ICP-OES) using a Perkin Elmer Plasma 400 spectrometer. Nitrogen content was calculated using a Vario EL Element analyzer. Tensile Testing Mechanical tests were carried out on a DEBEN minitester equipped with a 200 N load cell at room temperature. The specimen sizes used were typically in the range of 10 mm × 1.25 mm × 40−50 µm. At least five specimens were tested for each condition. A nominal strain rate of 1.5 mm/min was used. The slope of the linear region of the stress−strain curves was used to determine the Young’s modulus (E). The yield points were determined by the intersection of the two lines of the elastic region with the strain hardening region.11 The samples were conditioned at different relative humidities (RH) in the range of 20 to 99% for at least 2 days and measured at this RH. Scanning Electron Microscopy (SEM) and Energy-dispersive X-ray spectroscopy (EDX) 6 ACS Paragon Plus Environment

Page 7 of 35

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

Biomacromolecules

The cross-sections of fractured samples were imaged using a Hitachi S-4800 field emission microscope (1-1.5 kV) after sputter-coating a thin Au/Pd layer. EDX were recorded in a Hitachi S-3000N SEM at 30 kV acceleration voltage. Fourier transform infrared spectroscopy (FTIR) FTIR spectra were recorded using a Thermo Nicolet Nexus 470 spectrometer with a smart split ATR single reflection Si crystal from 1800 to 1500 cm−1 with a resolution of 4 cm−1. Water Sorption Water sorption isotherms were recorded with a HIDEN ISOCHEMA IGAsorp device at 25 °C in a range of 5−95% of RH. Samples (⁓ 20 mg) were exhaustively dried at 60 °C for 24 h before starting the test. Thermoporosity Nanopapers were cut into ca. 4 x 4 mm and placed in aluminum pans with MilliQ water with a moisture content between 1 - 1.5 g/g for 24 h until reaching equilibrium swelling. Samples (ca. 10 mg) were encapsulated in aluminum sealed pans and ultra-dry high purity dry nitrogen was used as an inert atmosphere for all tests in a PerkinElmer DSC8500 calibrated with cyclohexane, indium and zinc. The range span from -35 to 20 ˚C with scanning rates of 1 and 10 ˚C/min for the isothermal step and dynamic (heating and cooling) tests, respectively.38 The following procedure was carried out twice for each sample: Isothermal steps for porosity size distribution (PSD). 1. The sample was cooled from 20 ˚C, at 10 ˚C/min, to - 35˚C to crystallize all water in the sample. 2. Isothermal melting step: the sample was heated at 1 ˚C/min from -35 ˚C to the selected isothermal temperature (Tm). Then the sample was held for enough time to complete the melting transition. This step was repeated for the isothermal temperatures listed in Table S1. The full program is presented in Figure S1a with CNF-Na+ as example. The melting enthalpy of each step (Hm) was calculated for the area under the curve and used to determine the amount of water via the specific enthalpy of water (334 J/g). Dynamic step for bound water content determination.


Biomacromolecules

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

Page 8 of 35

1. After the last isothermal step at -0.2 ˚C, the sample was frozen to -35 ˚C, held for 5 min for stabilization, and then ramped to 25 ˚C to completely melt the water contained in the pan. The program is presented in the Figure S1b using CNF-Na+ as example. Two overlapping peaks were obtained (related to freeze bound and free water) and separated by deconvolution using a Gaussian method (Figure S2), and giving two parameters: (i) the area of the complete overlapped peak provides total amount of freezing water (TFW), and (ii) the integration of the peak molten at lower temperature represents the freeze water within the pores. Specific Surface Area An ASAP 2020 Plus-Physisorption device was employed for Kr gas adsorption experiments. CNF nanopapers were cut into ca. 4 x 4 mm, dried at 60 °C in a convection oven for 24 hours and degassed at 95 °C during 24 h to ensure no moisture effects.39 Adsorption isotherms were recorded at 77 K with relative pressures (P/P0) in the range of 0.05─0.3. Stability of CNF/water dispersions. The stability of the CNFs in suspension was evaluated by adding suitable amounts of the corresponding salt solutions to 0.25 wt% of CNF dispersions. The added salts were in correspondence with the counterion previously exchanged, that is, solutions of MeCl or QACl were added to dispersions of CNF-Me+ or CNF-QA+, respectively. Optical Transmittance Light transmittance of the dispersions (0.25 wt%) was measured at 600 nm using an Ocean Optics UV-Vis spectrophotometer with a HL-2000-FHSA light source at 25 ºC. Zeta Potential Zeta potentials of CNF dispersions were measured using a Zetasizer Nano ZS from Malvern at 25 ºC. Measurements were performed three times in disposable capillary cells. pH Titration Titration of CNF-COOH were recorded on a Methrom Titrando 90 at 25 ºC. The pH was adjusted in the range of 3-10 with NaOH (0.1 M). Sedimentation Speed


Page 9 of 35

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

Biomacromolecules

An analytical photocentrifuge LUMiSizer® with a wavelength of 865 nm was used to evaluate sedimentation velocities of CNF dispersions with salt solutions. The dispersions were place in PC 110-132XX cells and centrifuged at 2000 rpm and 25 ºC.


Biomacromolecules

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

Page 10 of 35

RESULTS AND DISCUSSIONS Ion Exchange Efficiency of the Nanopapers The first step in our study is to understand the counterion exchange efficiency of the CNFs in dispersion and their final content after nanopaper preparation (filtration), which we analyzed using Fourier transform infrared spectroscopy (FTIR), energy-dispersive X-ray spectroscopy (EDX) and elemental analysis. The CNFs are prepared by TEMPO-mediated oxidation at neutral conditions with a charge density of 0.44 mmol/g, and subsequent microfluidization, yielding micrometer-long CNFs with diameters of 2.0 - 2.5 nm (Figure S3a).11 We applied a stringent and highly standardized protocol to ensure the counterion exchange. First, we purposely induced aggregation by setting 0.25 wt% dispersions of CNF-Na+ to pH = 3 with HCl. This protonates the COO- groups on the CNF surface, and the CNF-COOH dispersion tends to coagulate and can be pelletized by centrifugation.40 This pellet can be redispersed easily using the appropriate metal and alkyl ammonium hydroxide solutions set to pH = 9 (Scheme 1). Repeating the precipitation and redispersion step and additional dialysis against the hydroxide solutions (pH = 9) ensure complete counterion exchange. The corresponding samples are called CNF-Me+ (Me+ = Li+, Na+, Cs+) for the alkali metal ions, and CNF-QA+ (QA+ = ammonium (A+), tetraethyl ammonium (TEA+), tetrabutyl ammonium (TBA+) and tetrahexyl ammonium (THA+)) for nanopapers with quaternized ammonium counterions. Finally, we placed the different dispersions into vacuum filtration units to obtain transparent nanopapers (Figure S3). From previous studies it is known that well-stabilized CNFs with homogeneous and clear dispersions produce better structured nanopapers.16 This necessitates to perform the vacuum filtration from high pH (pH = 9). We chose vacuum filtration over film casting, because there is no concentration increase (only at the filter cake), preventing agglomeration of the CNFs in the bulk solution, and because excess hydroxide solution can be washed out.


Page 11 of 35

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

Biomacromolecules

Scheme 1. Preparation of CNF nanopapers with different counterions. Washed and dialyzed CNFCOOH suspensions are neutralized to pH = 9 using alkali metal (Li+, Na+, Cs+) or quaternized ammonium (A+, TEA+, TBA+, THA+) hydroxide solutions. The counterions locate as a sheath around the CNF in bulk after filtration.

After vacuum filtration, we trace the sodium content with EDX spectroscopy (Figure 1a-d). Sodium (Kα = 1.041 keV) is detectable only for CNF-Na+ and eliminated for the other nanopapers after ion exchange. The cesium signal (Lα = 4.2865) for CNF-Cs+ indicates proper substitution. Since the signals for lithium and nitrogen (Kα of 0.0544 and 0.392, respectively) overlap with carbon and oxygen, it is only possible to confirm the sodium removal for the other samples. Subsequently, FTIR spectroscopy reveals a high ion exchange efficiency as most of the carboxyl groups are present as carboxylate, COO- (Figure 1c-d). Differences in ionization can be followed via the absorption bands (C=O stretching) for COOH groups at 1720−1740 cm−1 (grey curve for CNF-COOH) and COO- at approximately 1600 cm−1 (shifting from pH 3 to pH 9, respectively). Once the pH is increased (using Me+OH and QA+OH), the relative intensity of the COOH band inverts and the COO- functionality is dominant (Figure 1c). Nanopapers in the form of CNF-Me+ present no shift for the COO- band as a function of the metal counterion, hence they seem unaffected by the counterion size in bulk (Figure 1c). In contrast, CNF-QA+ nanopapers show a subtle shift from 1600 to 1615 cm−1 in the COO- band for longer alkyl chains (Figure 1d). This


Biomacromolecules

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

Page 12 of 35

shift is evidence of weakened ionic bonds between the anionic carboxylates and the cationic QA+ as the alkyl chain length is increased.41

Figure 1. Analysis of the ion exchange of CNF-Me+ and CNF-QA+ nanopapers. (a-b) EDX spectra and (c-d) FTIR spectra, respectively. CNF-COOH is used as reference.

Finally, the ion substitution can be followed by elemental analysis: for instance, the used TEMPO-CNF possesses 0.44 mmol(COOH)/g. A titration showed that 2.5 mmol/L (⁓1 mmol/gCNF) of NaOH is necessary for reaching to pH = 9 when titrating from pH = 3 (Figure S4a). After vacuum filtration, its nanopaper shows 0.30 mmol/g of Na+ content by elemental analysis with inductively coupled plasma optical emission spectrometry (ICP-OES), which demonstrates that excess counterions and base are indeed removed during vacuum filtration. The dried films of CNF-Li+, and all CNF-QA+ have a similar exchange efficiency as judged from the Li content (0.31 mmol/g) and the nitrogen content for QA+ counterions (0.29─0.31 mmol/g; see


Page 13 of 35

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

Biomacromolecules

Table S2; Cs cannot be measured due to lack of sensitivity). Therefore, these protocols demonstrate similarly efficient and near quantitative counterion exchange.


Biomacromolecules

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

Page 14 of 35

Mechanical Properties and Deformation Mechanisms Deformation mechanisms in CNF nanopapers are complex as we deal with a material structured on several length scales, and the presence of counterions on the surface modifies the interfibrillar interactions. We start our analysis by performing tensile tests at a relative humidity (RH) of 50% to break down the effect of each counterion. Second, we observe cross-sections of the fractured samples to illustrate further the deformation mechanisms and interfibrillar interactions, and differentiate and correlate which features depend on the nanopaper structure, and which depend on the chemical make-up applied on the surface of the nanofibrils. After ion substitution, the variation in mechanical response of the nanopapers should be discussed in terms of chemical interactions and the nanostructure. In general, increasing the size of the counterion reduces the bond strength and the ability of forming interfibrillar hydrogen bonds (Me+ and QA+). Interfibrillar hydrogen bonds are further hindered by the introduction of bulky hydrophobic components like the alkyl chains (QA+). The stiffness, Young’s modulus (E), shows low values for CNF-COOH (⁓10 GPa; prepared from the aggregated state at pH = 3), that increase drastically when the nanopaper dispersions are ion-exchanged and processed from high pH to reduce floc formation (Figure 2c and Table S3). The highest values correspond to the smallest ions (Li+ and A+) producing stronger interfibrillar bonds. A subtle decrease of E is obtained as the counterion size increases for the alkali metals. This drop is stronger for CNF-QA+ nanopapers due to the larger size (in comparison with Me+) and further hindrance of interfibrillar hydrogen bonding as the alkyl chains grow and shield surfaces. This agrees with data on large cationic surfactants such as cetyltrimethylammonium bromide, which lead to reduced hydrogen bonding of CNF in bulk and add hydrophobicity as the alkyl length is increased.42-44 CNF-Me+ nanopapers show a correlation of the yield point, the post yield behavior and strength at break (σy, σb, respectively) with the size of the ion in the order of Li+ > Na+ > Cs+ (Figure 2a and Table S3). The smaller ion (Li+) allows higher interfibrillar hydrogen bonding and ultimate strength of ⁓301 MPa (See Table S3). This is visible in particular in the higher yield points and the larger slope of the post-yield behavior. The larger ions (Na+ and Cs+) allow for higher deformability and increase the strain (11-13%, respectively), while leading to lower tensile strength. 14 ACS Paragon Plus Environment

Page 15 of 35

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

Biomacromolecules

Increasing the length of the alkyl chains in CNF-QA+ nanopapers shows a similar softening but with further enhancement in deformation (Figure 2b and Table S3): The yield points, σy, are reduced from 125 to 105 MPa by increasing the alkyl chain length (A+ to THA+), while σb is maintained (⁓300 MPa). The alkyl chains ease interfibrillar sliding allowing to reach an ultimate deformation up to 22% for CNF-THA+ (hexyl chain substituent) without losing mechanical strength. It is worth mentioning, that deformation at yielding (εy) is around ⁓1% for all nanopapers confirming to be a structural parameter, while yield stress depends on the interfibrillar bonding (type of counterion). The alkyl chains reduce the hydrogen bonding and act as plasticizers. Strikingly, we can observe a second yield point (σy2, εy2) for the largest metal counterion, CNFCs+, and for nanopapers substituted with QA+ with alkyl chains (TEA+, TBA+ and THA+). A second yield point represents the activation of motion of a second order structure with lower dynamics (grains and mesoscopic layers).16 This is prevented for the smallest ions (Li+, Na+ and A+) due to the strong combination of hydrogen bonds and ionic bonds. This result clearly shows that increasing the counterion size improves interfibrillar sliding, allowing further deformation without losing strength. Moreover, εy2 is similar for CNF-Cs+, CNF-TEA+ and CNF-TBA+ indicating again its relation to a similar nanostructure (Table S4). Interestingly, CNF-THA+ is the only nanopaper able to translate the second yield point, εy2, to higher values. This phenomenon is an indication for a slightly different nanostructure. It is important to realize that our results for the QA+ nanopapers do not show non-linearities, as found in an earlier study for alkyl ammonium-modified nanopapers (methyl to butyl).35 We discuss it further below.


Biomacromolecules

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

Page 16 of 35

Figure 2. Mechanical tensile properties of nanopapers at RH 50%. Tensile curves for nanopapers with (a) metal CNF-Me+ and (b) quaternized amines CNF-QA+ counterions. (c) Young’s modulus (E, empty symbols) and work of fracture (Ut, filled symbols) as function of the different counterions. Black arrows in a-b show yield points.

The work of fracture or toughness (Ut), calculated from the area under the σ─ε curve, shows a continuous increase with ion size. (Figure 2c). CNF-COOH shows drastically lower Ut corroborating that a poor dispersion leads to structures with low mechanical performance.16 The introduction of alkyl chains in the counterions improves the inelastic deformation, yielding tougher nanopapers with outstanding ductility up to a strain-to-failure of above 20% and close to 300 MPa in tensile strength. To comprehensively compare to previous data, we plot the E vs σb and E vs Ut in two Ashby plots to provide an instructive comparison between stiffness, strength and ductility (Figure 3 and Table S5-6). In general, the best materials are located on the top right of the graphs. Literature contains a wide range of CNF nanopapers (green data) substituted with small ions (mainly with sodium, see Table S5), where it is hardly useful to go into further details because of the large differences in preparation, carboxylate content, fibril size and aspect ratio. The strongest nanopapers are typically obtained in combination with nanoclays (blue data, 340-510 MPa) and span the toughness between 4-30 MJ/m3. Finally, CNF/polymer nanocomposites show that promoting deformation with a softer response (lower E and σb) depends on the polymer content (CNF/HEC) or thermal glass transitions (CNF/EGxDMAmy). Our new materials maintain the highest combination of strength and toughness in comparison to any other nanopapers with small ions. CNF-Me+ and CNF-A+ have a competitive stiffness (E = 19-17 GPa) with the best CNF/clay nanopapers and the CNF-THA+ nanopaper has the highest toughness reported (38 MJ/m3). Specifically, A+ (i.e. NH4+) cations can interact with the carboxylic groups via directional bonding and can form multiple hydrogen bonds.45 The most related work by Shimizu et al. has found a non-linear evolution of properties in CNFQA+ nanopapers prepared by solvent casting.35 Specifically, an increase in strain to failure up to 14-18% with a strength of 240-150 MPa for methyl and ethyl substituents (CNF-TMA and CNFTEA marked with white borders in Figure 3), respectively, while propyl- (CNF-TPA) and butylammonium (CNF-TBA) nanopapers drop to 8% and 12% elongation, respectively, with a 16 ACS Paragon Plus Environment

Page 17 of 35

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

Biomacromolecules

strength of 130-140 MPa (see Table S5). This is not the case for our samples, which follow a trend in all properties. We believe that the differences to our CNF-QA+ samples arise mainly from the preparation method. Longer preparation times by solvent casting for hydrophobically modified CNFs (e.g. CNF-TBA) may lead to flocculation in the dispersion due to constant concentration increase. Additionally, excess counterions are not washed out.

Figure 3. Comparison of stiffness (Young’s modulus E), (a) strength (σb) and (b) toughness (Ut) of CNF nanopapers in Ashby plots: CNF with small ions (green), CNF/clay (blue), CNF/polymer nanocomposites (CNF/EGxDMAmy (purple) and CNF/HEC (magenta)). Our samples are circled in black lines and Shimizu et al.35 samples are in white lines.

Next, we analyze the fracture surfaces to understand differences in deformation mechanisms (Figure 4). In general, a layered structure is observed with different levels of pull-out phenomena. Stronger nanofibril and mesoscale layer pullout phenomena occur as the counterion size increases. The

CNF-Me+ nanopapers with the smallest counterions present a similar

structure (CNF-Li+ and CNF-Na+ in Figure 4a and b, respectively), but CNF-Cs+ shows pull-out of mesoscopic layers (50-100 mm thick layers) and some bundles of individual fibrils being pull out (Figure 4c). The effect is more evident for alkyl substituted nanopapers (CNF-QA+ in Figure 4d-g). CNF-A+ shows a layered structure similar to small metal ions (Figure 4d), while CNFTEA+, CNF-TBA+ and CNF-THA+ present pull-out of the mesoscale layers and individual fibrils (Figure 4e-g). These results strongly correlate with the presence of a second yield point, εy2, 17 ACS Paragon Plus Environment

Biomacromolecules

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

Page 18 of 35

observed only in tough nanopapers with Cs+ and QA+ alkyl substituted counterions. The crosssection images of a strongly toughened CNF-Na+ nanopaper plasticized by water sorption at RH 99% present pull-out of fibrils and mesoscale layers (Figure S5), confirming the relation of nano/mesostructure deformation and toughness with weakened interfibrillar interactions. These observations are an interesting complement to earlier studies, where it could be shown that increased pullout on several length scales originated from weakened interfibrillar interactions by (a) increased humidity or (b) for changing the thermomechanical properties of the polymers from glassy to ductile in CNF/polymer nanocomposites.11,

16, 46

The small ions advantageously

preserve the inner strength of the CNF and promote deformability at the same time, while nanocomposites with polymers or clay typically lose strength or strain (except for nanocomposites with optimized conditions of small amounts of clay content (5-10%))25,

47, 48

Overall, this allows to reach new areas in the structure/property chart above (Figure 3).


Page 19 of 35

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

Biomacromolecules

Figure 4. Cross-section micrographs of nanopapers fractured at RH 50%. CNF-Me+: (a) Li+, (b) Na+ and (c) Cs+. CNF-QA+: (d) A+, (e) TEA+, (f) TBA+ and (g) THA+ (Scale bar = 5 µm).


Biomacromolecules

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

Page 20 of 35

Nanostructure Analysis It is known that mechanical response is not only related to chemical makeup in such CNF-based hierarchically structured entangled glasses, but also intimately connected to the nanostructure formed during the film preparation, and to how the structure responds during exposure to humidity and mechanical deformation. This all connects to interfibrillar interactions modified by the counterions (Scheme 1). Consequently, to provide a better understanding we access the nanostructural details with a range of complementary methods: specific surface area (SSA), apparent density (ρ), water uptake and porosity during swelling (Table S7 for a summary of the obtained parameters). First, measurements on the specific surface area (SSA), measured under fully dry conditions, shows a very low SSA of ⁓0.04 m2/g for all nanopapers prepared with counterions while CNFCOOH has 0.17 m2/g (Figure 6a). In agreement, the apparent density (ρ) at RH 50% is fairly constant around ⁓1.42 g/cm3 for CNF-Me+ and CNF-QA+, while it is 1.35 g/cm3 for CNFCOOH. This corresponds to a porosity of around 10% and 15%, respectively, and a significantly more porous structure for CNF-COOH (1.58 g/cm3 was used for crystalline CNF and the porosity is estimated as 1- ρ/ρcrystalline). These results indicate the formation of a tightly packed fibril network independent of the counterion for dispersions set to pH = 9. On the contrary, the uncontrolled aggregation at low pH in CNF-COOH leads to entrapping of porous structures and the typical mechanical response of nanopapers composed of poorly connected flocs, that is lower mechanical strength. Other porous CNF papers with more controlled pore formation processes contain better connected porous fibrillar networks and partly show larger deformations.12, 13, 16 Going away from fully dry conditions, we first turn to water sorption experiments. Interestingly, the amount of water uptake is similar for all counterion-containing nanopapers (Figure 5). Hence, differences in the water content can be excluded as a strong factor (H2O-induced weakening of interfibrillar interactions) influencing the mechanical response.


Page 21 of 35

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

Biomacromolecules

Figure 5. Water uptake of the nanopapers as a function of relative humidity.

Finally, we employ thermoporosimetry to gain more information about the interfibrillar interactions in hydrated state by the ability to swell and form nanopores (ratio 1:1). In thermoporosimetry, the nanopapers swell until equilibrium, and a series of isothermal melting steps allows to identify depressed melting temperatures of water within nanopores, enabling to evaluate the pore size distribution (Table S1). The water content is divided into three categories: free water (FW), non-freezing water (NFW) and total freeze bound water (TBW). Free water is the unbound water with thermodynamic properties similar to bulk water. NFW is adsorbed on the cellulose surfaces and bound to the cellulose chains by strong hydrogen bonds restricting mobility and ability to melt (2-3 layers of water molecules).38 TBW is inside the nanopores, and has depressed melting temperatures as a function of pore size. The content of the NFW has an average of ⁓0.45 g/g and is independent of the counterion presence, size or nature (Table S7). Regarding the most interesting parameter, TBW, the CNFCOOH nanopaper has the highest TBW (porosity) arising from the aggregate structure with lower density (Figure 6a). For both, CNF-Me+ and CNF-QA+, a proportional increase in TBW is observed for larger ion sizes (Li+

Counterion Size and Nature Control Structural and Mechanical

Recommend Documents