Lauric arginate/cellulose nanocrystal nanorods-stabilized alkenyl succinic anhydride pickering emulsion: enhancement of stabilization and paper sizing performance

Biodegradable colloidal particle materials are becoming attractive candidates as eco-friendly chemical additives in the low-carbon economy era. However, developing cheap, stable, and efficient paper-sizing agents is still a challenging issue for both the paper-making academic community and industry. Here, an easy-fabricating, stable, and high-performance alkenyl succinic anhydride (ASA) paper-sizing emulsion that is stabilized by lauric arginate (LAE)/cellulose nanocrystals (CNCs) nanorods is developed. Furthermore, the partial hydrophobization between ASA and LAE/CNCs can be adjusted due to the partial hydrophobization between ASA and LAE/CNC nanorods, resulting in improved stability of the ASA-sizing emulsion. This novel paper-sizing emulsion is shown to have a small droplet size (0.8 μm), high hydrolysis resistance, and a high paper-sizing degree (300 s) along with a remarkable hydrophobicity contact angle of 110° for long-term storage. This work enables the realization of an interfacial self-assembled Pickering-stabilizer, which leads to an environmentally friendly, pervasive and cost-effective emulsification technique for next-generation paper-sizing additives.


Introduction
Emulsions are vital in the production of food, medicine, paper and cosmetics industries, because of their specific physicochemical and biological properties, such as their excellent stability and resistance to liquid penetration. However, the traditional emulsions have disadvantages in terms of long-term stability. Therefore, solid particles are applied to stabilize the emulsion and named as Pickering emulsion, in which they are adsorbed on the oil-water interface, and form a mechanical barrier with densely packed particles (Pickering 1907). In particular, Pickering emulsions stabilized by biodegradable colloidal particles are attracting increasing interest because the demand for merchandise made of environmentally friendly ingredients is growing (Dickinson 2017;McClements and Gumus 2016). Therefore, many types of proteins and/ or polysaccharide-based colloidal particles have been developed as environmentally friendly Pickering emulsifiers (Hong et al. 2018;Sun et al. 2021). Pickering emulsions show better stability than traditional surfactant-stabilized emulsions due to the irreversible adsorption of particles even at a low particle concentration (Xia et al. 2018).
It is well known that sizing agents are important functional additives that impart paper with resistance to liquid penetration. Among them, alkenyl succinic anhydride (ASA) is a commonly applied neutral sizing agent, and needs to be prepared as an oil-in-water emulsion before sizing (Guodong Li et al. 2017;Yu et al. 2020). In the traditional emulsifying process, the excessive use of surfactants can facilitate emulsification, but that will have an adverse effect on sizing (Yu et al. 2013). In addition, ASA is easily hydrolysed, and the hydrolysate will hinder the sizing effect and paper-making process (Mohit et al. 2007;Robert 1990). Based on the problems mentioned above, the preparation and application of ASA emulsions are complex and difficult to control, causing an unstable sizing effect (Hubbe 2007). In practical production, ASA requires on-site emulsification preparation and immediate usage. Meanwhile, nanoclay and the modified particle-stabilized ASA emulsion exhibit good performance and are expected to solve the current issues (Yucheng Li et al. 2021;Yu et al. 2020). However, clays and their modified products are not degradable and are prone to creating an emulsion with high viscosity by forming a continuous phase gel system Yuanyuan Zhang et al. 2021a, b). Therefore, it is necessary to explore more efficient and environmentally friendly solid emulsifiers, and provide reliable methods for obtaining green paper additives. Furthermore, the Pickering emulsion system is expected to offer a potential alternative method for paper sizing emulsions, for large-scale manufacturing by paper companies and the rapid development of paper machines (Bai et al. 2021).
Cellulose nanocrystals (CNCs) are an effective Pickering stabilizer for oil-in-water (O/W) emulsions due to their superior properties in terms of reusability, sustainability, nontoxicity and biocompatibility (Chen et al. 2019;Chu et al. 2020;Habibi et al. 2010;Kedzior et al. 2020). Because of their wettability and rod-like nanostructure, CNCs can adsorb at the water-oil interface and form a dense interface network (Capron et al. 2017;Lu et al. 2020). Nonetheless, the high hydrophilicity of CNCs limits their application in Pickering emulsions, and CNCs are generally modified to improve their emulsification performance (Du et al. 2017;Patel et al. 2021;Saidane et al. 2016). Chemical modification of CNCs is a widely used strategy (Cunha et al. 2014;Lu et al. 2020;Tang et al. 2014;Zoppe et al. 2012), but physical adsorption with surfactants or macromolecules is a more promising strategy (Ben Azouz et al. 2012;Gong et al. 2017;Huang et al. 2016;Patel et al. 2021;Tardy et al. 2017) because of the nontoxicity, biocompatibility, and easy processing of adsorption methods. Since CNCs are negatively charged, cationic organic polymers can bind with CNCs through electrostatic adsorption. Lauric arginate (LAE) is a biological and amino acid-based cationic surfactant that is synthesized from lauric acid, arginine and ethanol (Bai et al. 2018). It can adsorb onto the surface of CNCs, improving their emulsification ability, while retaining their properties (Bai et al. 2018;Chi and Catchmark 2017). The colloid behaviour of the LAE/CNC complex is controllable and regulatable at the oil-water interface, which plays a very unique and vital role in stabilizing the Pickering emulsion system (Li et al. 2018).
Herein, an ASA sizing Pickering emulsion stabilized with LAE-modified CNCs was prepared. The emulsification mechanism was systematically studied by analysing the partial hydrophobization between LAE/CNCs and ASA and the adsorption of LAE/ CNCs at the water-oil interface. Scheme 1 shows that an ASA Pickering emulsion could be obtained by adding LAE/CNCs. The composition and structure of the obtained emulsions as well as the sized paper were analysed by scanning electron microscopy (SEM), atomic force microscopy (AFM), transmission electron microscopy (TEM), laser scanning confocal microscopy (CLSM), Fourier transform infrared (FTIR), and X-ray nanocomputed tomography (nano-CT). Subsequently, the water-repellent and hydrolysis resistance capacity of the sized paper was evaluated. It is expected to prepare an environmentally friendly ASA papermaking sizing emulsion with enhanced sizing performance and higher stability. The introduction of functionalized cellulose nanocrystals into ASA emulsions to prepare superior sizing emulsions is a long-term project that will enrich the colloid interface chemistry and facilitate the development of technology for applying emulsion additives in papermaking.

Materials
The cationic, food-grade surfactant ethyl-N αlauroyl-L-arginate hydrochloride (≥ 98% purity, C 20 H 41 N 4 O 3 Cl, MW ~ 421.02 g/mol) was purchased from Maclin (Shanghai, China). Sulfonate cellulose nanocrystal (CNC, ScienceK Co., Ltd., Hu Zhou, China) is a needle-like nanomaterial obtained by sulfuric acid treatment of natural plant cellulose. The surface contains sulfonic acid groups and hydroxyl groups. ASA oil with octadecenyl succinic anhydride as the main ingredient was supplied by Kamira (Shanghai, China). Nile red and Nile blue dyes were purchased from Maclean (Shanghai, China). Aspen alkaline peroxide mechanical pulp (APMP) was provided in a pulp board form by Shandong Zhongmao Shengyuan Pulping Co., Ltd. Ultrapure water with a resistivity of 18.2 MΩ·cm was produced by a Milli-Q water system attached to an ELGA reverse osmosis unit. All other chemicals used were all analytical reagents.
Preparation and characterization of LAE/CNC nanorods Concentrated LAE was diluted to 1.0 wt% with Milli-Q water. The CNC hydrogel was first diluted and then LAE was added to obtain the desired suspended CNC concentrate. LAE was first added to the CNC suspension and gently shaken for 20 s, followed by sonicating with an ultrasonic generator (YQ-620C, Shanghai Yi Jing Ultrasonic Instrument Co., Ltd., China) at ambient temperature (25 °C) for 10 min. The prepared nanorods were dialysed by a dialysis bag (~ 5000WM, Maclin, China) to remove the excess LAE and held at ambient temperature (25 °C) before analysis.
The Fourier transform infrared (FT-IR) spectra of the LAE/CNCs were measured by an FTIR spectrophotometer (Bruker Vertex V70, USA) in absorbance mode with an attenuated total reflection (ATR) diamond sensor. Spectra were collected over 600-4000 cm −1 regions under ambient conditions at a resolution of 4 cm −1 . All the LAE/CNC suspension samples were filtered by a B-GLM22 syringe filter (BKMAN, China) with a 0.22 μm membrane pore size before measurement. The particle size distribution and zeta potential of the samples were determined by dynamic light scattering measurements Scheme 1 Schematic diagram of interfacial adsorption of LAE/CNCs as ASA Pickering paper-sizing emulsion stabilizer using a Nano-ZS Zetasizer (Malvern, U.K.) angle with a He-Ne laser (laser wavelength 633 nm) at 178° scattering and a temperature of 298 K. To prevent multiple scattering effects, the samples were diluted to a concentration of 0.005 wt% with MQ water before measuring and this process was repeated three times. The three-phase contact angles of the LAE/ CNCs were measured with an OCA-50 micro contact angle measuring tool (Dataphysics, Germany) using a conventional particle-platelet method (Yucheng Li et al. 2021). Before starting the test, the samples were modulated as 2 mm-thick circular pieces with an HY-12 tablet machine (Tianguang, China). Protractor software was employed to determine the contact angle via axisymmetric drop shape analysis. The contact angle of each sample was measured three times, and the average of each sample was reported. Atomic force microscopy (AFM, Bruker, USA) was employed to measure the morphology of LAE/CNC nanorods in tapping mode. Before inspection, a 10 μL drop of LAE/CNC suspension with a concentration of 5 μg/ mL was deposited onto a mica substrate (20 × 20 mm) and dried at 25 °C overnight. A video optical contact angle measurement instrument (OCA-50, Data Physics Instruments, Germany) was used to study the surface tension and interface tension of the LAE/CNC suspension using the pendant drop method. ASA oil was added to a cuvette, and the transmissivity of the oil container was checked prior to measurement. Pendant droplets of 5 μL of the solution containing LAE/CNCs, were formed at the end of a low retention pipet tip immersed in the oil phase. The droplets were left for 10 s at 25 °C before measurement. The optical turbidity of the LAE/CNC suspension was characterized by a UV-VIS spectrophotometer (Agilent 8453, U.S) at a wavelength of 600 nm operated at 25 °C.

Preparation and characterization of ASA Pickering emulsions stabilized by LAE/CNC NRs
The LAE/CNC NR-stabilized Pickering emulsions were prepared as follows: the LAE/CNC NR suspension was prepared and placed in a plastic tube, followed by the addition of ASA oil to the suspension and sonication using an ultrasonic generator (YQ-620C, Shanghai Yi Jing Ultrasonic Instrument Co., Ltd., China) with an immersed titanium microtip. In this study, the oil-water ratio of LAE/CNC-stabilized ASA Pickering emulsions is 1:2 (v:v).
The stability of the emulsions with respect to creaming and coalescence was evaluated by counting the volume of discharged oil and water in 24 h via a FinePix S4050 digital camera (Fujifilm, Japan). The stability is expressed as the emulsifying index (EI), as follows: where H E is the total emulsion height and H e is the serum layer height. The values of H E and H e for different samples were measured via visual analysis of the photographs of the samples using ImageJ software. A DM750 M optical microscope (Leica, Germany) was applied to check the morphology of the ASA Pickering emulsion without further dilution. The average size and distribution of the droplets in the prepared emulsions were analysed by ImageJ. To visualize and distinguish LAE/CNCs and ASA oil in the emulsion system, the ASA oil and LAE/CNC suspension were dyed with Nile blue and Nile red respectively before emulsion preparation. Then, the emulsion droplets were observed under a laser confocal microscope. The adsorption of particles on the surfaces of the emulsion droplets was analysed by confocal laser scanning microscopy (CLSM, Leica, SP8, Germany). The FT-IR spectra of ASA emulsions stabilized by LAE/CNCs were measured using a FT-IR spectrophotometer (Vertex V70, Bruker, USA) under absorbance mode via an attenuated total reflection (ATR) diamond sensor.
Dried ASA droplets loaded with LAE/CNCs were inspected under scanning electronic microscopy (SEM, Hitachi, Regulus8220, Japan) and transmission electron microscopy (TEM, Hitachi Koki Co., Ltd., JEM 2100, Japan). A 10 μL drop of the ASA emulsion with a concentration of 0.02 g/mL was deposited onto a carbon membrane support (T11023, Beijing XXBR Technology Co., Ltd.) and dried at 25 °C overnight before inspection. Additionally, the morphology of the ASA emulsions was measured by atomic force microscopy (AFM, Bruker, USA) in tapping mode. A 10 μL drop of the ASA emulsion with a concentration of 0.02 g/mL was deposited onto a mica substrate (20 × 20 mm) and dried at 25 °C overnight before inspection. The emulsion creaming stability data were logged by a Turbiscan device (LUMiSizer, LUM Jiangsu Instruments Co., Ltd., China). Dynamic changes were evaluated by laser light backscattering intensity as a function of height during creaming. After high-speed centrifugation (895 × g, 10 min at 25 °C), the samples were prepared in rectangular plastic tubes (10 mm × 10 mm × 100 mm). The viscosities of the samples were measured by a rheometer (MCR 300, Anton Paar, Germany) using parallel plates (PP25) with the gap fixed at 1.0 mm. The corresponding viscosity was recorded with a shear rate ranging from 0.01 s −1 to 100 s −1 . The dynamic viscoelastic range was derived by a strain sweep from 0.01% to 100% at a constant frequency of 10 rad/s. A dynamic frequency sweep was conducted at a constant strain of 1.0% and a wide frequency range of 0.1 to 100 rad/s. All measurements were made at 25 °C.
Handsheet making and sizing performance of the ASA emulsion A PTI-type paper former (ISO 5269-2, Austria) was used to prepare ASA-sized paper handsheets with a base weight of 60 g/m 2 . The pulp was first soaked in water overnight and then dispersed at 20 g/L by a standard disintegrator (Lorentzen & Wettre, Sweden) at 20,000 rpm, followed by purification using a PFI mill (KRK Company, Japan). ASA-sized paper was prepared using a CUF5/200 laboratory coating unit (Sumet Technologies Ltd., Germany), The as-prepared ASA emulsion was diluted to 1.0 wt% before paper sizing, which was quantitatively determined by the liquid permeation method using GB/ T460-2008 (Tan et al. 2014), and further used for quantitative determination. An automatic video optical contact angle measurement instrument (OCA-50, Dataphysics, Germany) was used to measure the static water contact angle of the sized paper at various time points. The 3D structural images of the sized paper were derived by a multiscale X-ray nanocomputed tomography (nano-CT) system (Skyscan 2211, Bruker, Germany). The size of the tested sample paper was 2 mm × 2 mm.

Characterization of LAE/CNC nanorods
The FT-IR spectra of the CNCs and LAE/CNCs are displayed and compared in Fig. 1a. Two new bands at 2915 and 2851 cm −1 were found in the spectra of the LAE/CNCs, which were assigned to the asymmetric and symmetric stretching of CH 2 in the long alkyl chain of LAE. Meanwhile, the peaks of 3420 cm −1 and 1756 cm −1 corresponded to the vibration of O-H and carbonyl group on side chains from LAE. In addition, the signals at 1643 cm −1 and 1530 cm −1 were correlated with amide I (C = O stretching) and amide II (C-N stretching and N-H deformation) from the LAE backbone respectively, which was consistent with the results of Johnson (Johnson et al. 2011). Overall, LAE has been successfully connected to CNCs.
The surface and interface tensions of CNCs (2.0 wt%) with LAE solutions were derived using the pendant drop method (Fig. 1b). The results showed that the surficial and interfacial tensions fell rapidly with the addition of LAE, and then remained flat until the dosage of LAE was 0.015 wt%, since the interaction among the hydrophobic LAE tail groups and the electrostatic interaction with CNC can induce local micellization, facilitating LAE binding to the surface of CNC nanorods. When the amount of LAE was 0.015 ~ 0.025 wt%, the surface tension remained constant approximately 45 mN/m, which could be caused by the formation of free unbound micelles in the native solution, an indication of surfactant saturation binding to CNC NRs. The hydrophobicity of pristine CNC NRs and LAE/CNC NRs was measured with three-phase contact angle (Fig. 1c). The LAE/ CNC NRs were observed to have a contact angle that increased to 80° with the increasing dosage of LAE, and then decreased rapidly. However, the CNC NRs had a contact angle of 60° and showed relative hydrophilicity. Therefore, the addition of LAE increased the contact angle and hydrophobicity of the CNC NRs. The LAE/CNC NRs reached a maximum contact angle of approximately 87.6° with 0.016% LAE, and then decreased to 36.5° with 0.02% LAE, which could be due to the formation of a hydrophilic micellar structure with excessive LAE and the reduced hydrophobicity of LAE/CNC NRs. Following LAE modification, the hydrophobicity of CNCs increased, which may facilitate their attachment to the surface of ASA oil droplets, as well as the formation of stable oil-in-water emulsions with uniform and small spherical droplets.
To better understand the aggregation of LAE/CNC NRs, the microstructural changes of primary CNCs and their nanorods with different LAE concentrations can be studied by AFM imaging (Fig. 1d). It was obvious that the dispersion of CNC NRs was gradually improved with an increasing amount of LAE, which was mainly because of the electrostatic adsorption and bridging traction action between the positively charged LAE and negatively charged CNC NRs. The LAE/CNC NRs showed the best dispersion and stability with 0.02 wt% LAE. Furthermore, the LAE/CNC NRs formed bundle structures with LAE concentration above 0.02 wt%. The zeta potential and macromorphology of the LAE/CNC NR suspension are recorded in Fig. 1e. The absolute value of the zeta potential first increased to − 70 mV with 0.02 wt% CNC NR, and then decreased with the increasing dosage of LAE, which illustrated the optimal dispersion and stability of LAE/ CNC NR suspension with 0.02 wt% CNC NRs. With excessive LAE (> 0.02%), the absolute value of the zeta potential decreased again, which confirmed the connection between LAE and CNC NRs. Otherwise, the decreased absolute value of zeta potential would result in microflocculation of the LAE/CNC NRs, which would be beneficial for the stability of Pickering emulsion (Bollhorst et al. 2017). Meanwhile, this variation tendency was further confirmed by the size distribution results (Fig. 1f), and the size distribution curve of LAE/ CNC NRs was the narrowest with 0.02 wt% LAE.  uniform diameter were clearly observed, and their diameters tended to decrease with the increasing dosage of CNC NRs. The corresponding data of the size distribution were analysed, as shown in Fig. 2b; the emulsion droplets had an average diameter of 5.3 μm with 0.2 wt% CNCs, while it was reduced to 0.8 μm at 2.0 wt% CNCs with a narrow size distribution curve. All freshly prepared ASA emulsions had a homogeneous appearance. They were observed to become stable within 24 h, and the EI values were recorded.  Figure 2c shows that the emulsion stability was improved by increasing the dosage of CNCs, and the emulsion still maintained a homogeneous appearance with 2.0 wt% CNCs after 24 h. The ASA emulsions stabilized by 0.2 wt%, 0.6 wt% and 1.0 wt% CNCs all had separated phases and decreased EI values of 92%, 85%, and 64%, respectively. The emulsion consistently showed a similar tendency with centrifugal stability. The microphotographs of the ASA emulsion are listed in Fig. S1, and all samples were recorded to have good homogeneity after 3 h. The emulsion stabilized by low CNC concentration (< 1.8 wt%) showed an uneven distribution of emulsion droplets, and ungraded droplets were also found in the images, which may be the agglomerated emulsion droplets caused by incomplete coverage. With a higher dosage of CNCs at 1.8 wt% and 2.0 wt%, the emulsion droplets showed a decreased diameter and a more uniform distribution. However, the agglomerated emulsion droplets and their uneven distribution appeared again with excessive CNCs (> 2.0 wt%). Therefore, the LAE/CNC NRs have an improved emulsifying capacity and can adjust the size of Pickering emulsions.

Effect of pH
The pH value of the system is an important parameter for the stability of nanocellulose-stabilized Pickering emulsions (Kedzior et al. 2020). Figure 3a shows that the pH influenced the size distribution and morphology of the ASA emulsion. Especially at pH ~ 7, the emulsion had large droplets and a large size distribution. However, the emulsion had small droplets and good homogeneity at other pH values. The appearance of the LAE/CNC-stabilized ASA emulsion under different pH values is shown in Fig. 3b. The emulsification index of the freshly prepared emulsion is 100%, and there is no phase precipitation. After being left for 24 h, a large amount of turbid water phases were precipitated in the emulsion at pH > 7. This may be due to the precipitation of phases caused by the polymerization of emulsion droplets, which is consistent with the results in Fig. 3a. Given that the carbonyl lactone ring groups of ASA will be in the form of -COOH in a high pH environment, it leads to a hydrolysis reaction and reduces the stability of the emulsion (Fig. 3c). Furthermore, when the pH > 7, the hydroxy groups of the LAE/CNC NRs interacted with the carboxyl groups from the ASA-acid at the o/w interface, leading to the oriented movement of the weakly hydrogen-bonded water molecules in the interfacial region, which could hinder the penetration of water into an oil phase and/or ASA molecular transport (Scatena et al. 2001).
The effect of a pH ranging from 3 to 11 on the droplet size distribution of the Pickering emulsions is given in Fig. 3a and c. The ASA emulsion stabilized by LAE/CNCs under different pH values for 24 h, showed almost no phase precipitation (Fig. 3b). It was determined that H + would reduce the repulsion between the LAE/CNC NRs within the acid circumstance (Patel et al. 2021). Hence, microgels were formed to resist the flocculation of the emulsion. The droplet size distribution and optical micrographs of emulsions with different pH values after 24 h are listed in Figs. S3 and S4. Similar trends can be observed in Figs. 3c and S4, where the average droplet size of the prepared emulsion is steady except at pH ~ 9, which implies that the weakly alkaline condition is not conducive to the stability of the LAE/ CNC-stabilized ASA emulsion.

Rheological behaviour of the ASA Pickering emulsions stabilized by LAE/CNC NRs
To better understand the gel or liquid properties of the as-prepared ASA Pickering emulsions, the storage modulus (G')/loss modulus (G") and viscosity with different C CNC and pH values were investigated ( Fig. 3e-g). The result shows that the value of G' was higher than that of G" in the whole range of angular frequencies when the C CNC was above 1.8%, indicating that the linear rheological behaviour of the emulsions was mainly viscoelastic and that well-developed elastic networks exsied, which were formed by the dispersed droplets in the emulsion system (Qiao et al. 2015). The enlarged difference between G' and G" indicated the high resistance to flocculation (Zhang et al. 2021a, b), especially when the C CNC was 2.0 wt%. Moreover, the maximum values of G' and G" were found in the ASA emulsion stabilized by the LAE/CNC NRs with C CNC ~ 2.0 wt%, which indicates the high resistance to flocculation and explains the better stability. Setting the pH value of the emulsion as a variable, the value of G' was higher than that of G" when the pH value was 3-9 over the whole angular frequency range. Elasticity dominated the rheological behaviour of the emulsions stabilized by LAE/ CNC NRs. It is worth noting that the as-prepared emulsion showed greater G' and G" values at pH ~ 5 than those at other pH values, suggesting that the LAE/CNC-stabilized emulsion at pH ~ 5 shows strong resistance to flocculation (Binks et al. 2017). Figure 3f shows the relationship between the apparent viscosity of the ASA emulsion and the shear rate. All emulsions have a reduced viscosity and then tend to a constant with an increasing shear rate, so the emulsion showed a shearing-thinning behaviour, which was due to the deformation of the emulsion droplets and their ordered arrangement, as well as the continuous phase (Thaiphanit et al. 2016). The results showed that the C CNC obviously influenced the rheological apparent viscosity behaviour of the emulsion. Since the emulsion system network was mostly destroyed by the shearing force, the repulsive forces were predominant at the high shear rate  (Niu et al. 2018). In addition, that the result show that when the shear rate is low, the apparent viscosity of the emulsion decreases at the fastest rate. As the shear rate increases, the viscosity reduction range is gently decreases. Therefore, the ASA Pickering emulsion prepared by using modified CNCs as emulsifiers has the typical characteristics of a pseudoplastic fluid. The concentration of CNCs affects the overall viscosity of the emulsion, and its overall trend increases in proportion to the concentration. The dependence of apparent viscosity on the shear rate of emulsions prepared at different pH values is shown in Fig. 3g. The flow curves of the emulsion samples under all pH value systems exhibited shear thinning, with the viscosity decreasing with increasing pH value. A lower pH can decrease the surface charge of CNC/LAE NRs. CNC/LAE NRs are more likely to be attracted to and connect with each other to form a stronger network, either by surface-surface or edge-surface interactions. A higher pH can increase the OH − and decrease the H + , leading to a more negative surface charge of CNC/LAE NRs, which results in a larger electrical repulsive force among particles. In this case, the viscoelasticity of the interface layer is weak. Hence, the particles would connect with each other by dominated edge-edge particle interactions and form a weaker network. In addition, the highest viscosity among all the emulsions is exhibited by the as-prepared Pickering emulsions at pH ~ 3. These data indicate that interfacial packing behaviours and interactions of particles at the interface are affected by the pH conditions, which could result in a weaker network architecture and vitrified emulsions with poorer viscoelasticity at higher pH (i.e., pH ~ 7, 9 and 12) . Thus, ASA emulsions prepared at a pH range between 3 and 5 possessed a more favourable structure with gel-like elastic network.
Interfacial adsorption of LAE/CNC NRs Figure 4a, b shows the CLSM of the ASA emulsion, where red refers to the ASA oil dyed with Nile red, and blue refers to the emulsifier of LAE/CNC NRs. In addition, the blue colour around the droplets indicates flocculation consisting of LAE/CNC NRs, which are encircled by deep orange and white colour. Figure 4c is the merged CLSM image. Interestingly, an incompletely covered ASA droplet with a black loophole (Fig. 4b) and a red oil phase inside (Fig. 4c) can be observed, suggesting that the LAE/CNC NRs remained on the ASA-water interface instead of entering the oil phase. To observe the distribution of LAE/CNC NRs at the interface, the dosage of CNC NRs was reduced to 0.2%, and the result is shown in Fig. 4d. The LAE/CNC NR microgels with a "friedegg" structure tended to swell in the oil phase and partially collapse in the water phase with a contact angle (θ) favouring O/W emulsion-type stabilization (Rey et al. 2020). Under low concentrations of CNC NRs, such microgels would decrease the interfacial stability and lead to the aggregation of adjacent droplets, which would result in large droplets and accelerated hydrolysis of ASA. The microflocculation formed by LAE/CNC NRs caused bridging flocculation between the droplets, further preventing coalescence and oil-phase transference between oil droplets (the elliptical area heightened by the orange dotted in Fig. 4b and 4e), facilitating the stability of the emulsion obtained.
The total coverage rate C (%) of LAE/CNC NRs on emulsion droplets was calculated by the following equation (Bai et al. 2018): where m p is the mass of LAE/CNC NRs, D is the average diameter of ASA droplets (D 43 ), h is the thickness of LAE/CNC NRs (measured from the AFM image), ρ is the density of LAE/CNC NRs (1.6 g/cm 3 ), and V oil is the ASA volume. According to the calculation, C was 64.75%, with 0.02 wt% LAE and 0.2 wt% CNCs. When the concentration of CNCs increased by 2.0%, C was as high as 84.25%, suggesting good coverage and stabilization of LAE/CNCs on the ASA oil.
ASA is a specific oil phase with esterification reaction ability and is bound to react with polyhydric CNCs, thus modifying the morphology of particles and altering the performance of the emulsion. The centrifugal stability of the ASA Pickering emulsion stabilized by LAE/CNCs is shown in Fig. 5a. The results show that the centrifugal stability of the emulsion first increases and then decreases with the increasing dosage of LAE, with the maximum stability achieved at 0.02 wt% LAE. This may be caused by the protective effect of LAE on the hydroxyl groups of CNCs, which provided LAE/CNC NRs with C = m p D 6h V oil moderate wettability at the oil-water interface. However, excessive LAE may be an obstacle to the reaction, and the high hydrophobicity of CNCs is not conducive to the adsorption of LAE/CNC NRs, which ultimately reduces the stability of the emulsion. AFM images of the emulsion stabilized by LAE/CNC NRs are shown in Fig. 5b, which shows that the LAE/CNC NRs form denser arrangement around the emulsion droplets with an increasing LAE dosage. The emulsion with 0.02% LAE has smaller droplet size and a more uniform longitudinal distribution (− 201.4 to 165.7 nm), which improves the adsorption stability of LAE/CNC NRs at the water-oil interface. Figure 5c displays the SEM image of the ASA emulsion stabilized by CNCs alone. A granular film structure with a honeycomb-like shape is revealed, which may be caused by the aligned arrangement of CNCs. This arrangement mode may be related to the decreased repulsion between ASA-CNCs and CNCs caused by the reduced number of hydroxyl groups.
Otherwise, such an alignment is more obvious without LAE in the TEM image (Fig. 5d), where CNCs sufficiently react with ASA and form an NRs matrix with an oriented arrangement (white dotted box). However, the ASA emulsion stabilized by LAE/CNC NRs presents a circular arc granular film in Fig. 5e due to the effective stabilizing and protecting function of LAE/CNCs on the ASA droplets by providing valid coverage. Moreover, uniformly distributed nanorods on the emulsion surface can be clearly observed in Fig. 5f. We assume that LAE protects the hydroxyl groups of CNCs from being completely esterified, thereby presenting NR film with an anisotropic arrangement, allowing a graft reaction to occur between some of the hydroxyl groups on LAE/ CNCs and carbonyl lactone on ASA at the o/w interface. The LAE/CNC NRs with moderate wettability induced by esterification could form microgels at the o/w interface, which is consistent with the results observed in Fig. 3d. A schematic diagram of partial hydrophobization and the stabilizing effect of LAE/CNC NRs at the ASA-water interface is drawn and displayed in Fig. 5g, h. CNCs can sufficiently react with ASA and generate hydrophobic ASA-CNCs (blue) at the ASA-water interface without LAE. Then the ASA is emulsified as an oil-in-water emulsion by interfacial nanoparticles along with CNC NRs in the continuous phase. However, the ASA emulsion shows poor stability and easily forms honeycomb-like walls and oriented CNCs because of the coalescence and compression of the droplets. To some extent, the introduction of LAE restricts the reaction between ASA and CNCs, and more LAE/CNC NRs form a stable interfacial granular film, which leads to the formation of gel state with moderate wettability of the LAE/CNC NRs. The LAE/CNC-ASA NRs can efficiently prevent droplet coalescence induced by phase migration, hinder hydrolysation of ASA and improve the stability of the emulsion.
In view of the characteristic functional groups, the ASA emulsion shows high intensity at 1863 cm −1 and 1778 cm −1 and an essentially unchanged intensity at 1663 cm −1 after 6 h, as shown in Fig. 6a, which are assigned to the carbonyl lactone ring of ASA and the carboxyl group of ASA-acid, respectively. These peaks show significant changes until 12 h later. After approximately 144 h, the ASA began to lose most of its activity (Fig. S6). It is concluded that LAE/CNC NRs provided lasting protection against ASA and significantly inhibited the hydrolysis reaction of ASA within 12 h, which will promote the sizing application to cellulose-fibre paper.
Furthermore, the sizing performance of the ASA sizing Pickering emulsion is evaluated with the sizing degree and the contact angle of the sized paper. The dynamic water contact angle of the sized paper is shown in Fig. 6b, and the contact angle gradually increases with the increasing amount of CNC added to the ASA emulsion, which indicates improved hydrolysis resistance. The ASA-sized paper has a contact angle of 115° after retaining water for 1 s with 2% CNC NRs, while the contact angle is 103° for the sized paper with 0.5% CNC NRs. Even after 900 s, the contact angle of the paper sized by the ASA emulsion containing 0.5 wt% and 2.0 wt% LAE/CNC NRs decreased slowly to 72° and 93°, respectively, which further confirms the reinforcement of LAE/ CNC NRs. Figure 6c shows that the sizing degree improves with the increasing dosage of the ASA emulsion, and the sizing degree increases by 152 s with 0.5 wt% ASA (relative to oven-dried stock), and then increases by approximately 300 s with 1.0 wt% ASA. However, the static contact angle of the Fig. 6 a FT-IR spectra of the ASA emulsion stabilized by LAE/CNC NRs at different storage time. b Optical images revealing the dynamic wetting behaviors of a water droplet (≈25 μL) atop sized paper with 0.5 wt% (top), 1.0 wt% (middle) and 2.0 wt% (bottom) CNCs at room temperature. The LAE loading is fixed to 0.02 wt%. c Sizing degree and contact angle of the sized paper by ASA based on LAE/CNC NRs with different ASA's dosage (wt%, based on dry paper pulp d Top view of a micro-CT image of the sized paper with CNC-stabilized ASA and LAE/CNC-stabilized ASA, respectively sized paper first increases and then decreases with the increasing dosage of ASA, which reaches a maximum value of 113° with the ASA emulsion containing 1.0 wt% CNC NRs. Overall, the ASA emulsion has good internal sizing performance. The micro-CT of the ASA emulsion sized paper shows great differences with different emulsifiers, as shown in Fig. 6d. The blue and green colours indicate that many hydrophilic hydroxyl groups and hydrolysed ASA acid remained with the ASA emulsion stabilized by CNC NRs. However, the golden colour indicates that many ASA celluloses are present, which further explains the enhancement of LAE on the sizing performance of ASA emulsions. Figure 7 illustrates the adsorption of LAE/CNC NRs, the partial hydrophobization between LAE/ CNC NRs and ASA at the ASA-water interface, and the stabilization mechanism of Pickering emulsions. Otherwise, the interfacial granular film composed of amphiphilic LAE/CNC NRs prevents the migration of the oil phase and the coalescence of droplets, reducing the contact and hydrolysis reaction between water and ASA (Fig. 7a). Two main reactions occur during the course of stabilization of the ASA Pickering emulsion based on LAE/CNCs (Fig. 7b). The ASA molecule near the oil-water interface will be partially hydrolysed to form ASA acid (reaction A). Moreover, LAE/CNC NRs could graft with ASA or/and ASAacid through the esterification reaction (reaction B) between the hydroxide from the CNCs and the carbonyl from the lactonic ringlactone of ASA or the carboxyl from ASA-acid, generating a more compact interface granular film with moderate wettability and,thus, improving the stability of the emulsion. At present, CNCs have been applied to many areas where cationic surfactants are commonly involved, so it is very important to understand this binding interaction. The binding interactions between CNCs and these oppositely charged surfactants could have a profound influence on the colloidal stability and functionality of CNCs in these systems, as well as the bulk properties of the systems. We provide a sustainable and ecologically compatible surface modification strategy for CNCs that combines a novel biologically-derived surfactant and a simple environmentally benign mixing approach. The successful formulation of LAE/CNCstabilized ASA Pickering emulsion provides a basis for the application of ASA paper sizing systems with tuneable and functional properties based on Pickering systems.

Conclusion
This study investigated the dual-enhancement effect of interfacial adsorption and partial hydrophobization on the emulsification properties of LAE/CNC NRs in ASA Pickering emulsions. Moderate wettability of CNCs with a contact angle of 87.6° can be achieved by the introduction of LAE, which facilitates attachment on the surface of ASA droplets, and generates stable emulsions with uniform, small and spherical droplets. The LAE/CNC-ASA droplets exhibit a small droplet size, long-term storage stability and Fig. 7 Conceptual schematic of the a adsorption and b the reaction during the course of partial hydrophobization of LAE/CNC NRs at ASA-water interfaces strong resistance to flocculation when the pH is below 5. CLSM images showed that the LAE/CNC NRs are adsorbed at the oil-water interface. SEM and TEM images revealed that the ASA Pickering emulsion droplets were encircled and stabilized by a matrix of LAE/CNC NRs, which were densely immobilized at the oil-water interface. LAE/CNC-ASA can efficiently prevent droplet coalescence induced by phase migration, hinder hydrolysation of ASA and improve the stability of the emulsion. LAE/CNC NRs had a lasting protective ability against ASA and significantly inhibited the hydrolysis reaction of ASA within 12 h. Cellulose paper seized with an LAE/CNC-ASA emulsion resulted in a 300 s of sizing-degree with a 113° water contact angle. These findings offer significant insights into the fabrication of surfactant-modified CNCs and the commercial application of LAE/ CNC NR-stabilized ASA Pickering emulsions in the papermaking industry.