Contents lists available at ScienceDirect

Food Hydrocolloids

journal homepage: www.elsevier.com/locate/foodhyd

Cellulose nanofibrils from Miscanthus floridulus straw as green particleemulsifier for O/W Pickering emulsion

Qi Lia,b, Bing Xiea, Yixiang Wangc,∗∗, Yanting Wangd,∗∗∗, Liangcai Pengd, Yan Lia, Bin Lia,Shilin Liua,b,∗

a College of Food Science & Technology, Huazhong Agricultural University, Wuhan, Hubei, 430070, ChinabNational R&D Center for Citrus Preservation, Huazhong Agricultural University, Wuhan, Hubei, 430070, Chinac Department of Food Science and Agricultural Chemistry, McGill University, Ste Anne de Bellevue, Quebec, H9X 3V9, Canadad College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, 430070, China

A R T I C L E I N F O

Keywords:Cellulose nanofibrilsMiscanthusPickering emulsionParticle stabilizer

A B S T R A C T

For high value utilization of energy crops, in this work, a high stable O/W Pickering emulsion had been preparedsuccessfully by using cellulose nanofibrils (CNFs) obtained from Miscanthus floridulus straw (MFS). The effect ofhigh pressure homogenization time on the physicochemical properties of the resulted CNFs had been in-vestigated. The obtained CNFs had high surface activity and could spontaneously diffuse to the interface of theO/W phase. The particle adsorption kinetics of CNFs was characteristic of soft particles or Pickering emulsifiers,and the adsorption was irreversible. Therefore, stable oil-in-water Pickering emulsion was obtained by using theCNFs as particle emulsifiers, and the emulsion stabilized by CNFs with only 0.15 wt% exhibited outstandingstability. Moreover, the environmental factors, such as pH values, ionic strength, and temperature had littleinfluence on the stability of the emulsion. All the results indicated that CNFs exhibited promising interfacialactivity and powerful emulsifying ability. Owing to the flexible shape of the CNFs, it was concluded that MFS asa kind of energy crop was a promising source to obtain the CNFs for the fabrication of novel food-grade Pickeringemulsions.

1. Introduction

Miscanthus, a fast-growing perennial C4 rhizomatous grass, has ahigh yield of biomass per unit planted area (Peng, Li, Liu, Yi, & Han,2017). Historically, it was widely used as forage, roofing materials, andclothing. Until the end of 20th century, it was gradually considered as ahigh-potential energy crop owing to its high tolerance for various cli-mate conditions and even for an infertile and poor land (Si et al., 2015;Brosse, Dufour, Meng, Sun, & Ragauskas, 2012). Among all the species,M. floridulus has received extensive attention in the biofuel applicationslike enzymatic saccharification and bioethanol (Lee & Kuan, 2015).However, the multi-step biochemical conversion processes were nor-mally consisted of various pretreatments. Enzymatic hydrolysis andyeast fermentation were often combined with high cost and eco-un-friendly biomass conversion (Wang, Xu, & Peng, 2014). The usage ofthe M. floridulus was still limited (Cheng et al., 2018). M. floridulus ismainly composed of cellulose, hemicelluloses and lignin. As the

pretreatment is the initial step for biomass process, it becomes essentialto find out the optimal pretreatments that enhance biomass conversion.Using conventional pretreatments method to obtain cellulosic particles,additional energy or chemicals were needed, which further increasedcapital cost or environmental implication (Dahunsi, Oranusi, &Efeovbokhan, 2017; Janke et al., 2017; Kalyani, Zamanzadeh, Mueller,& Horn, 2017; Peng et al., 2019; Xu, Mao, Peng, Luo, & Chang, 2018).As a result, it is constructive and meaningful to widen the application ofM. floridulus.

Emulsion is one of the major systems in the field of food technology,to meet with the basic requirements of food emulsion technology,natural and sustainable emulsifiers are needed (Dickinson, 2017;McClements, Bai, & Chung, 2017; Xia et al., 2018; Zhang, Cheng, Ye, &Chang, 2017). Previous works about food-grade Pickering emulsionstabilizers were focused on the proteins, lipids, and starch based par-ticles (Tavernier, Wijaya, Van der Meeren, Dewettinck, & Patel, 2016;Xiao, Li, & Huang, 2016a; Yang et al., 2017). In order to improve the

https://doi.org/10.1016/j.foodhyd.2019.105214Received 27 February 2019; Received in revised form 5 July 2019; Accepted 5 July 2019

∗ Corresponding author. College of Food Science & Technology, Huazhong Agricultural University, Wuhan, Hubei, 430070, China.∗∗ Corresponding author.∗∗∗ Corresponding author.E-mail addresses: yixiang.wang@mcgill.ca (Y. Wang), wyt@mail.hzau.edu.cn (Y. Wang), slliu2013@mail.hzau.edu.cn (S. Liu).

Food Hydrocolloids 97 (2019) 105214

Available online 07 July 20190268-005X/ © 2019 Elsevier Ltd. All rights reserved.

T

stability and the resistance to digest of these emulsions, lots of workshave been performed based on cellulosic biomass, which was preparedfrom conventional plants or through fermentation (bacterial cellulose)(Bai, Huan, Xiang, & Rojas, 2018; Fujisawa, Togawa, & Kuroda, 2017;Grishkewich, Mohammed, Tang, & Tam, 2017; Khan, Wen, Huq, & Ni,2018; Tang, Sisler, Grishkewich, & Tam, 2017). However, due to theresurgent interest of cellulosic biomass, the growing demands for cel-lulose-based materials have been attracted an increasing interest(Isogai, 2013; Klemm et al., 2018). Considering the components of MFS,based on our research works, it can be used an ideal candidate for thepreparation of cellulose, which can be used in the food system as water-insoluble dietary fiber. However, little research work had been con-ducted on the utilization of cellulosic particle from energy crop, likeMFS.

Inspired by our previous works on the physical modification ofbacterial cellulose (Li et al., 2019), in this work, we prepared CNFs fromMFS via high pressure homogenization (HPH) process, and developedan effective strategy to fabricate a stable Pickering emulsion. The effectof the homogenization time on the physicochemical and emulsifyingproperties of CNFs has been clarified, and the stability and morphologyof the resultant emulsions have been investigated. With the econom-ically feasible and eco-friendly pretreatment method, MFS is a potentialsource of CNFs, which could be used as Pickering emulsion stabilizers infood applications.

2. Material and methods

2.1. Materials

Dodecane and sodium hydrochloride (AR grade) were purchasedfrom Aladdin Reagent Database Co. (Shanghai, China). All other che-micals with analytical grade were purchased from Sinopharm ChemicalReagent Co., Ltd (SCRC).

2.2. Preparation of cellulose nanofibrils (CNFs)

In order to remove other components, like lignin or hemicellulose, acombination of hot alkaline treatment followed by a treatment withhigh pressure homogenziation was applied to MFS (Seibert-Ludwig,Hahn, Hirth, & Zibek, 2019). The MFS was firstly ground into powders,and then treated with NaOH (5 wt%) and NaClO (0.5 wt%) alternatelyat 60 °C for 5 h and 3 times. The solid components were filtered andwashed with distilled water to neutral, and then it was treated with ahigh pressure homogenization (HPH, AH-1500, ATS, Canada.) for dif-ferent cycle times. The obtained CNFs treated with HPH for 20, 30, 50,and 80 times were coded as HPH-20t, HPH-30t, HPH-50t, and HPH-80tCNFs, respectively.

2.3. Preparation of pickering emulsions

In order to investigate the influence of the particle size of the CNFson the stability of Pickering emulsions, CNFs obtained from differenthomogenization times were used. Firstly, coarse oil-in-water emulsionswere prepared by mixing dodecane (10 v%) with water (90 v%) con-taining with 0.2 wt% of HPH-20t, HPH-30t, HPH-50t, and HPH-80tCNFs, respectively. Then the coarse emulsions were treated by using ahigh shear blender at a speed of 11,000 rpm for 4min to producePickering emulsions. In order to investigate the effect of CNFs contenton the emulsion stability, different amounts of HPH-30t CNFs (0.05,0.1, 0.15, and 0.2 wt%) were used to fabricate Pickering emulsions.

2.4. Characterization

Atomic Force Microscopy (MultiMode8, Bruker, USA) was used toobserve the topography of CNFs. Fourier transform infrared (FT-IR)tests were carried out with a FT-IR analyzer (470-Nexus, Nicolet, USA)

in wavenumbers ranged from 400 to 4000 cm−1. Wide-angle X-raydiffraction (WAXRD) measurements were performed with an XRD dif-fractometer (D8-Advance, Bruker, USA), and the patterns were re-corded in the region of 2θ from 5 to 40° with Cu Kα radiation at 40 kVand 30mA. The crystallinity index (CrI %) of CNFs was calculated byusing Eq. (1) (French, 2014):

=−

×Crl I II

(%) 100%am002

002 (1)

where CrI was calculated as the ratio of height between the maximumintensity of the crystalline peak close to 2θ=22.2° (I002) and the in-tensity of the non-crystalline material diffraction peak close to 2θ=18°(Iam). The surface charge of the CNFs was determined by using a dy-namic light scattering/electrophoresis instrument (Zetasizer Nano ZS,Malvern Instruments, Worcestershire, UK). The tests were performed intriplicate and the results were presented as an average value. Waterholding capacity (WHC) of CNFs with different physical size wasmeasured, and it was calculated by using the following formula(Shezad, Khan, Khan, & Park, 2010):

=−

⋅WHCM M

M(%) 100%wet dry

dry (2)

where the Mwet was the wet weight of the CNFs, the Mdry was the dryweight of CNFs.

The stability of Pickering emulsions was characterized throughevaluating creaming index (CI %) and droplet size. CI was calculated byusing the following formula (Ines et al., 2016):

= ⋅CI HH

(%) 100%s

t (3)

where Hs was the height of the serum layer and Ht was the total heightof the emulsion. The droplet polydispersity was measured by usingMasterSizer2000 (Malvern Instruments, Worcestershire, UK). The dro-plet polydispersity was expressed as the surface-weighted mean dia-meter:

= ∑∑d n d

n di i ii i i3,2

32

(4)

where ni was the number of droplets of diameter di. All the tests wereperformed in triplicate and the results were exhibited as the averagevalue.

The interfacial tension of CNFs (0.01 w%) treated with differenthomogenized times was tested by using a drop shape analyzer rhe-ometer (Tracker Teclis/IT Concept, France) at 25 °C. Briefly, the aqu-eous phase was placed in an optical glass cuvette, and a syringe filledwith dodecane as the oil phase was then submerged into the opticalglass cuvette. During the whole experiment, the initial volume of the oildrop was all ∼10 μL. Distilled water was used as the control. Therheological properties of the emulsions were characterized by using aDiscovery HR-2 Hybrid Rheometer (TA Instruments, New Castle, DE,USA) equipped with concentric cylinder geometry. The temperaturewas set to 25.0 ± 0.1 °C. A dynamic frequency sweep was carried outby applying a constant strain of 2% (selected by measuring the lineardomain), and the storage modulus (G′) and loss modulus (G″) as afunction of frequency were obtained. All the experiments were per-formed in triplicate and the results were exhibited as the average value.

2.5. Statistical analyses

All measurements were carried out in triplicate unless otherwisestated. Results were expressed as the mean value ± standard devia-tion. Statistical analysis was performed using origin 9.0.

3. Results and discussion

Fig. 1 showed the AFM images of CNFs obtained from different

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

2

homogenization cycle times. The cellulose nanofibrils were entangledwith each other, showing a relatively loose network structure. Theaverage width of the CNFs was decreased from 49 to 33 nm with theincreasing the homogenization times from 20 to 30 times, but theaverage height of the CNFs was changed slightly with the increasing the

homogenized times, as it was shown in Table 1. Comparing with ourprevious works on bacterial cellulose nanofibrils under the sametreatment conditions, the CNFs that obtained from MFS exhibitedsmaller in average height (Li et al., 2019). The average height of CNFswould further influence the physical stability of the suspension. It wasworth noting that the aqueous CNFs dispersions had a good stabilitywith solid content up to 0.6 wt%, and they could be stored at ambientconditions for about 2 months without precipitation, suggesting thatthe HPH process had a significant influence on the physical size ofCNFs, it could be confirmed that the physical size of CNFs could besuccessfully reduced by HPH process for 30 cycle times.

FT-IR spectra of CNFs treated with different homogenized timeswere shown in Fig. 2a. The broad absorption band located at3349 cm−1 was ascribed to the stretching of –OH groups and an ab-sorption peak at 2904 cm−1 was related to C–H stretching vibration.The peak centered at 1427 cm−1 was resulted from the intermolecularhydrogen bonding at the C6 group, and the absorption peak at around1370 cm−1 was ascribed to the bending vibration of the C–H and C–Obonds of aromatic ring in polysaccharides (Haafiz, Hassan, Zakaria, &Inuwa, 2014). In addition, the absorption peak at 1100 - 1108 cm−1

was attributed to the C–O–C glycosidic ether bond. Meanwhile, thepeak around 1039 cm−1 was associated with the stretching vibration ofC–O–C pyranose ring (Mandal & Chakrabarty, 2011). These absorptionpeaks at around 3349, 2904, 1427, 1370, 898 cm−1 were associatedwith the characteristics of native cellulose (Rosa, Rehman, de Miranda,Nachtigall, & Bica, 2012). It indicated that the native crystallinestructure of cellulose obtained from MFS was not affected by the HPHprocess. The crystallinity of the CNFs obtained from MFS was char-acterized by using XRD. As shown in Fig. 2b, the main diffraction peaksat 2θ values of 16.3°, 22.4°, 34.6° were ascribed to the typical diffractionpeaks of cellulose I. The crystallinity of CNFs was varied from55.88–64.13%. With the removal of the main components of lignin andhemicellulose or amorphous cellulose from the MFS by treatment withalkali and bleaching agents, the purified cellulose with increasedcrystallinity was obtained, and the crystallinity of CNFs was increasedslightly with increasing the homogenization cycles.

The zeta potential of CNFs homogenized for different times wasshown in Fig. 3a. With the increase of homogenization times from 20 to

Fig. 1. AFM tapping-mode height images of CNFs homogenized with differenttime, (a) was for HPH-20t, (b) was for HPH-30t, (c) was for HPH-50t, and (d)was for HPH-80t, respectively.

Table 1The average width and height of CNFs obtained by AFM.

average width (nm) average height (nm)

HPH-20t CNFs 49.79 ± 17.62 7.55 ± 2.76HPH-30t CNFs 33.27 ± 7.30 3.68 ± 1.17HPH-50t CNFs 40.52 ± 11.03 5.12 ± 1.22HPH-80t CNFs 38.64 ± 10.17 3.77 ± 0.71

Fig. 2. (a) FT-IR spectra of CNFs treated for different homogenized times, and (b) XRD spectra of CNFs treated for different homogenized times.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

3

80, the zeta potential of CNFs was slightly varied from −15 to −19mV, and the average width of the CNFs was decreased slightly, as it wasshown in Table 1. It indicated that it was easy to extract cellulose fromMFS based energy crop, which would reduce the economic costs for thedevelopment of cellulose as functional ingredient from MFS. The WHCof the CNFs was shown in Fig. 3b, and it was increased from 90% toabout 118% when the homogenization times increased from 20 to 80. Itwas well known that the WHC was highly dependent on fiber structure.With the increasing the homogenized times, the average width andheight of the CNFs was decreased, it indicated that CNFs with the sameweight, the number of the CNFs was different, therefore, the combinedwater with the CNFs was different. The high WHC made CNFs as apotential thickener or particle emulsifier that could be used in foodsystems.

The influence of the CNFs on the dynamic interfacial tension at oil-water interface was shown in Fig. 4a. The interfacial tension decreasedand gradually leveled-off. The equilibrium values depended on theCNFs with different physical morphology and were lower than that ofblank. It suggested that the relaxation time (adsorption and re-arrangement) was decreased when the CNFs were used, which were thekey factors affecting the emulsion formation. When the CNFs were usedfor the stabilization of the emulsions, both adsorption and rearrange-ment of the CNFs at the interface had appreciable contributions to thedynamic interfacial tension. The effective diffusion coefficients at longtime (t→∞) was further used to evaluate the adsorption process, whichcould be obtained by following equation (Fainerman, Makievski, &

Miller, 1994):

=∞

→∞−

D πRTc

( )

( ( ))

drd c

dr

dtt

ln4

21

2 (5)

where γ was the interfacial tension, R was the universal gas constant, Twas the temperature, c was the particle concentration in the bulk phase,and t was the adsorption time. In Fig. 4b, it could be seen that the γ-t−1/

2 was linear dependencies, which indicating the equation was applic-able. There was no appreciable difference between the different CNFsamples. Moreover, the CNFs had low rearrangement and equilibriumtime, which was typical for Pickering solid particle stabilizers (Milleret al., 2010). The relative soft fibrils with larger and heterogeneouswetting radius, which was caused by the deformable nature of the fi-brils, displayed stronger attractive capillary than other rigid particles(Chen et al., 2017). This result indicated that the CNFs at the interfaceshowed strong interactions during rearrangement.

Fig. 5 showed the images of the emulsions stabilized with CNFs.Obviously, all the emulsions had a typical unimodal distribution ofdroplet size. Consistent with the distribution of droplet size, emulsionscontaining the HPH-30t CNFs or HPH-50t CNFs were more stable overtime against coalescence or phase separation (as shown in Fig. 6). Thisphenomenon could be associated with the change of the wettability ofthe cellulose particles. From Fig. 3b, with the increase of homogenizedtimes, the WHC of the CNFs was gradually increased. If the particleswere too hydrophilic, they would tend to diffuse into the water phase.Therefore, the HPH-30t and HPH-50t CNFs with adequate hydro-philicity had increased emulsifying capacity than that of the others.From the perspective of energy conservation and environmental pro-tection, the HPH-30t CNFs were selected for the following research. Asall the emulsions displayed different degrees of sedimentation due tothe density difference between oil and water phase, HPH-30t CNFs withdifferent solid contents was used. The effect of solid contents of HPH-30t CNFs on the stability of the emulsions was subsequently in-vestigated. As shown in Figs. 7 and 8, the emulsion prepared with 0.2%HPH-30t CNFs didn't show any phase separation, which indicated ex-cellent physical stability against coalescence. However, with the in-crease of the content of CNF, the excess CNF in the aqueous phase couldlead to aggregation of droplets, causing a slight increase on the emul-sion droplet size. As a result, it could be seen that the droplet size ofemulsions stabilized with 0.2 wt% CNFs was larger than others(Table 2). The long-term stability of emulsions stabilized with differentcontents of HPH-30t CNFs were also investigated. When the content ofCNF was low, a large excess of oil-water interface compared with theamount that could be covered by the presence of solid particles wereproduced during agitation. When the agitation process was stopped, thepartially protected droplets were aggregated, thus reducing the totalamount of oil-water interface until the interface was sufficiently cov-ered. This phenomenon was shown in Table 2. It indicated that thedroplet size of emulsion was slightly increased with the increase of thestorage time. The droplet size of the emulsions was measured by using alaser particle size analyzer and the majority of the droplet had smallsize around 10 μm. Moreover, the emulsions stabilized with highercontents of CNFs appeared to be more stable with smaller droplet sizesafter 14 days. However, the emulsion stabilized with 0.2% HPH-30tCNFs showed a bimodal distribution of droplet size (Fig. 8a). Thisphenomenon could be explained as the smaller peak was attributed tothe minor fraction of larger CNFs particles, which had sizes around100 μm (Tenorio, Gieteling, Nikiforidis, Boom, & Van Der Goot, 2017).Moreover, the creaming index was reduced with the increase of thesolid content of CNFs. It indicated that the increased number of CNFsparticles might contribute to the network formation in the continuousphase (Zhai, Lin, Liu, & Yang, 2018; Zhu et al., 2016).

The total droplet surface area (Sd) was calculated according to theoil volume and the average droplet size (Capron & Cathala, 2013):

Fig. 3. (a) Effect of homogenized times on the zeta potential and WHC (b) ofthe CNFs treated for different homogenized times.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

4

= × =S πR VπR

VR

4 34

3d

2 oil3

oil(6)

The theoretical maximum surface, Sp, was defined by:

= =S N LIMhρp p

p

(7)

where Np was the amount of CNFs, L and I were represented for thelength and width, Mp was the mass of CNFs included in the emulsion(g), h was the thickness which was defined by AFM, and ρ was the CNFsdensity (1.6 g/cm3).

The theoretical surface coverage ratio (C) was calculated by thefollowing equation:

= = ×CSS

M RhρV

(%)3

100%p

d

p

oil (8)

Based on our previous work, the adsorption behavior of cellulose at

the oil/water interface was irreversible (Li et al., 2019). Using differentcontents of HPH-30t CNFs as the emulsion stabilizer, the surface cov-erage (128%) gradually increased with the increase of the solid contentof CNFs, and reached to 578% with the 0.20 wt% CNFs, which meantthat lots of free CNFs were dispersed in aqueous phase. Consistent withFig. 8a, the emulsion stabilized with 0.2% HPH-30t CNFs showed abimodal distribution of droplet size. This could be due to the furtheraggregation of extra CNFs particles, which had sizes around 100 μm.This could further cause the coalescence, yielding larger droplets withhigher polydispersity, as it was listed in Table 3. It could be concludedthat the smaller size of emulsion droplets would further contributed to arelatively limited coalescence, exhibiting the formation of a highlystable emulsion. In fact, the decrease of the emulsion droplet size wasalso associated with larger amounts of cellulose nanofibrils adsorbedinto oil-water interface. As a result, the solid content of CNFs played acrucial role in the formation of the high stable emulsion system via

Fig. 4. (a) Dynamic interfacial tension measurements at the oil-water interface as a function of time with 0.01 wt% CNFs, and (b) Plot of γ - t−1/2 at long time (t→∞)for 0.01 wt% CNFs with different homogenized times.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

5

Fig. 5. Optical Microscopy images of the Pickering emulsions stabilized with0.1 wt% CNFs homogenized for different times, (a) HPH-20t, (b) HPH-30t, (c)HPH-50t, and (d) HPH-80t, and the volume ratio of oil to water was 1:9.

Fig. 6. (a) Particle size distribution and (b) creaming index of the emulsionsstabilized with 0.2 wt% CNFs homogenized for different times, the particle sizedistribution of emulsions was measured after 1 d, and the volume ratio of oil towater was 1:9.

Fig. 7. Micrographs of emulsions stabilized with different content of HPH-30tCNFs: (a) 0.05 wt%, (b) 0.10 wt%, (c) 0.15 wt%, and (d) 0.20 wt%, and thevolume ratio of oil to water was 1:9.

Fig. 8. (a) Particle size distribution and (b) creaming index of emulsions sta-bilized with different contents of HPH-30t CNFs, the particle size distribution ofemulsions was measured after 1 d and the volume ratio of oil to water was 1:9.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

6

Table 2The droplet size of Pickering emulsions stabilized with different contents(0.05–0.20 wt%) of HPH-30t CNFs.

Sample D3,2 (μm)

1d 7d 14d

0.05% 10.88 ± 0.02 11.10 ± 0.05 11.14 ± 0.060.10% 10.62 ± 0.06 10.94 ± 0.26 11.20 ± 0.140.15% 10.72 ± 0.01 11.02 ± 0.18 10.91 ± 0.120.20% 12.33 ± 0.04 13.59 ± 0.06 13.42 ± 0.08

Table 3The surface coverage of Pickering emulsions stabilized with different contents(0.05–0.20 wt%) of HPH-30t CNFs after different storage times.

CNF contents Surface coverage

1d 7d 14d

0.05% 1.28 1.30 1.310.10% 2.49 2.56 2.630.15% 3.77 3.87 3.840.20% 5.78 6.37 6.29

Fig. 9. Rheological properties of freshly prepared emulsion stabilized withdifferent contents of HPH-30t CNFs: (a) apparent viscoelastic as a function ofshear rate and (b) G′ (closed symbols), and G″ (open symbols) as a function ofangular frequency measured at 2% strain, and the volume ratio of oil to waterwas 1:9.

Fig. 10. Effect of pH on the morphology of emulsion stabilized with 0.2 wt%HPH-30t CNFs, (a) pH 3, (b) pH 4.5, (c) pH 7, and (d) pH 11, and the volumeratio of oil to water was 1:9.

Fig. 11. (a) Droplet size of emulsions stabilized with HPH-30t CNFs (0.2 wt%)after different storage times at 25 °C, and (b) particle size distribution ofemulsions stabilized HPH-30t CNFs (0.2 wt%) under different pH values, theparticle size distribution of emulsions was measured after 1 d and the volumeratio of oil to water was 1:9.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

7

inhibiting the free moving of the emulsion droplet.The rheological behavior of the emulsion was significantly influ-

enced by the contents of CNFs. In Fig. 9a, with the increase of CNFscontent, the viscosity of the emulsion was increased, indicating thestrengthened attractive forces between the emulsions (Pandey et al.,2018). An overshoot behavior was observed at low-shear rate region, itwas mainly caused by the orientation of the emulsion droplets when thestress was applied. Because the surface of the emulsions was coveredwith CNFs, the interaction between the CNFs contributed to the beha-vior. When the shearing force was increased, the 3D network formed byCNFs was gradually destroyed, revealing a typical shear thinning be-havior (Lin, Li, Lopez-Sanchez, & Li, 2015). As it was shown in Fig. 9b,all the emulsions exhibited a higher value of storage modulus (G′) thanits corresponding loss modules (G″), indicating the elastic behavior ofparticle-stabilized emulsions (Paximada, Tsouko, Kopsahelis, Koutinas,& Mandala, 2016). In all cases, both the G′ and G″ were progressivelyincreased with the increasing the content of CNFs. This phenomenonfurther supported the conclusion that higher particle fraction was vitalfor preparing the stable Pickering emulsion system (Xiao, Wang,Gonzalez, & Huang, 2016b). Hence, it could be concluded that thehighly physically stable emulsion was obtained by increasing the

particle content. With the increasing the content of CNFs, the cross-linking behavior of the CNFs was enhanced, which strengthened thesteady three-dimensional network structure.

The emulsions were prepared under different pH values (3–11),which were controlled by using 0.1 M citrate/phosphate buffer solu-tion. The effect of pH on the droplet size of the emulsions was shown inFig. 10a. Based on our previous study, the zeta potential of BCNFs wasinfluenced by pH, and the stability of the BCNFs suspensions was im-proved when the pH value was slightly increased (Li et al., 2019).Consistent with this result, it could be found in Fig. 11a that smalleremulsion droplet size was formed under pH=4.5–7.0. All of theemulsions revealed excellent storage stability, regardless of pH values.However, the zeta-potential of CNFs suspensions was decreased withthe increasing pH, indicating a relative poor stability. For the emulsionsystem, when the pH increasing, the stability of the emulsions woulddecrease which further leading to an aggregation. As a result, theemulsion droplet size would increase at pH 9.5 and 11. Different frompH=4.5–7.0, the emulsions prepared under pH 9.5 and 11 exhibitedbimodal distribution of droplet size, which showed a small peak ataround 100 μm (Fig. 10b). The results were also consistent with themicrostructure of the emulsions (Fig. 11). In order to clarify the ob-served trend, zeta potential of HPH-30t CNFs under different pH valueswas further measured (as shown in Fig. 12). Obviously, when the pHvalue of the suspensions was 4.5 or 7, zeta potential of the suspensionswas closed to −35 mV, indicating a more stable system (Mikulcova,Bordes, Minarik, & Kasparkova, 2018). However, the emulsions tendedto aggregate when the pH was 4.5 (Fig. 10). As a result, the droplet sizeand distribution of showed distinct difference under pH 7. The influ-ence of temperature on the droplet size and stability of the emulsionswas shown in Fig. 13. The emulsions stabilized with 0.2 wt% HPH-30tCNFs exhibited almost the same droplet size (10–13 μm) after differenttemperature treatments. Moreover, the increase in the temperatureshowed insignificant decrease on the stability of the emulsion, and thedroplet size distribution was unimodal (Fig. 13d). Further, no phaseseparation was visible for the Pickering emulsions after long-term sto-rage, making them suitable for food-grade emulsion formulations. Re-markably, oil-in-water Pickering emulsions containing concentratededible oil that was stable against oil coalescence during storage wasinteresting. To our knowledge, the results in this study would open anopportunity for the application of CNFs from energy crops in food sci-ence and technology.

4. Conclusions

Cellulose nanofibrils obtained from MFS through HPH process couldbe used to stabilize O/W Pickering emulsions at a relatively low solidcontent of CNFs. The water holding capacity of CNFs was significantlyenhanced with the increase of the homogenized times. As a result, theemulsion stabilized by CNFs exhibited the small droplet size and thehigh stability. Moreover, higher content of the CNFs could also mark-edly improve the stability of the emulsions. The rheology results re-vealed that the obtained emulsions stabilized by CNFs exhibited a ty-pical shear thinning behavior. For all the emulsions, the value ofstorage modulus was higher than loss modulus, indicating a typicalelastic behavior of particle-stabilized emulsions. Based on the aboveresults, it could be concluded that the content of the CNFs had sig-nificant influence on the O/W interface associating with the strength-ened intermolecular cross-linking and increased steric hindrance. As thewhole physical modification process conducted without using any or-ganic solvents, it provided a facile, “green” method to fabricate a po-tential CNFs-based eco-friendly food-grade stabilizer for Pickeringemulsion processing.

Notes

The authors declare no competing financial interest.

Fig. 12. Effect of pH on the zeta potential of 0.01 wt% HPH-30t CNFs.

Fig. 13. Morphology of emulsion stabilized with 0.2 wt% HPH-30t CNFstreated at different temperature, (a) 4 °C, (b) 25 °C, and (c) 50 °C. (d) Particlesize distribution of emulsions stabilized HPH-30t CNFs (0.2 wt%) treated underdifferent temperature for 5 h, and the volume ratio of oil to water was 1:9.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

8

Acknowledgements

This work was supported by the project of the FundamentalResearch Funds for the Central Universities (2662018PY060), and thefund of the Beijing Engineering and Technology Research Center ofFood Additives, Beijing Technology and Business University (BTBU).

References

Bai, L., Huan, S., Xiang, W., & Rojas, O. J. (2018). Pickering emulsions by combiningcellulose nanofibrils and nanocrystals: Phase behavior and depletion stabilization.Green Chemistry, 20(7), 1571–1582.

Brosse, N., Dufour, A., Meng, X., Sun, Q., & Ragauskas, A. (2012). Miscanthus: A fast-growing crop for biofuels and chemicals production. Biofuel Bioprodution &Biorefining-Biofpr, 6(5), 580–598.

Capron, I., & Cathala, B. (2013). Surfactant-free high internal phase emulsions stabilizedby cellulose nanocrystals. Biomacromolecules, 14(2), 291–296.

Cheng, S., Yu, H., Hu, M., Wu, Y., Cheng, L., Cai, Q., et al. (2018). Miscanthus accessionsdistinctively accumulate cadmium for largely enhanced biomass enzymatic sac-charification by increasing hemicellulose and pectin and reducing cellulose CrI andDP. Bioresource Technology, 263, 67–74.

Chen, H., McClements, D. J., Chen, E., Liu, S., Li, B., & Li, Y. (2017). In situ interfacialconjugation of chitosan with cinnamaldehyde during homogenization improves theformation and stability of chitosan-stabilized emulsions. Langmuir, 33(51),14608–14617.

Dahunsi, S. O., Oranusi, S., & Efeovbokhan, V. E. (2017). Optimization of pretreatment,process performance, mass and energy balance in the anaerobic digestion of Arachishypogaea (Peanut) hull. Energy Conversion and Management, 139, 260–275.

Dickinson, E. (2017). Biopolymer-based particles as stabilizing agents for emulsions andfoams. Food Hydrocolloids, 68, 219–231.

Fainerman, V. B., Makievski, A. V., & Miller, R. (1994). The analysis of dynamic surfacetension of sodium alkyl sulphate solutions, based on asymptotic equations of ad-sroption kinetic theory. Colloids and Surface A, 87, 61–75.

French, A. D. (2014). Idealized powder diffraction patterns for cellulose polymorphs.Cellulose, 21(2), 885–896.

Fujisawa, S., Togawa, E., & Kuroda, K. (2017). Nanocellulose-stabilized Pickering emul-sions and their applications. Science and Technology of Advanced Materials, 18(1),959–971.

Grishkewich, N., Mohammed, N., Tang, J., & Tam, K. C. (2017). Recent advances in theapplication of cellulose nanocrystals. Current Opinion in Colloid & Interface Science, 29,32–45.

Haafiz, M. K., Hassan, A., Zakaria, Z., & Inuwa, I. M. (2014). Isolation and character-ization of cellulose nanowhiskers from oil palm biomass microcrystalline cellulose.Carbohydrate Polymers, 103, 119–125.

Ines Guerra-Rosas, M., Morales-Castro, J., Araceli Ochoa-Martinez, L., Salvia-Trujillo, L.,& Martin-Belloso, O. (2016). Long-term stability of food-grade nanoemulsions fromhigh methoxyl pectin containing essential oils. Food Hydrocolloids, 52, 438–446.

Isogai, A. (2013). Wood nanocelluloses: Fundamentals and applications as new bio-basednanomaterials. Journal of Wood Science, 59(6), 449–459.

Janke, L., Weinrich, S., Leite, A. F., Terzariol, F. K., Nikolausz, M., Nelles, M., et al.(2017). Improving anaerobic digestion of sugarcane straw for methane production:Combined benefits of mechanical and sodium hydroxide pretreatment for processdesigning. Energy Conversion and Management, 141, 378–389.

Kalyani, D. C., Zamanzadeh, M., Mueller, G., & Horn, S. J. (2017). Biofuel productionfrom birch wood by combining high solid loading simultaneous saccharification andfermentation and anaerobic digestion. Applied Energy, 193, 210–219.

Khan, A., Wen, Y., Huq, T., & Ni, Y. (2018). Cellulosic nanomaterials in food and nu-traceutical applications: A review. Journal of Agricultural and Food Chemistry, 66(1),8–19.

Klemm, D., Cranston, E. D., Fischer, D., Gama, M., Kedzior, S. A., & Kralisch, D. (2018).Nanoceullose as a natural source from groundbreaking applications in materialsscience: Today's state. Materials Today, 21(7), 720–748.

Lee, W. C., & Kuan, W. C. (2015). Miscanthus as cellulosic biomass for bioethanol pro-duction. Biotechnology Journal, 10(6), 840–854.

Lin, D., Li, R., Lopez-Sanchez, P., & Li, Z. (2015). Physical properties of bacterial celluloseaqueous suspensions treated by high pressure homogenizer. Food Hydrocolloids, 44,435–442.

Li, Q., Wang, Y., Wu, Y., He, K., Li, Y., Luo, X., et al. (2019). Flexible cellulose nanofibrilsas novel pickering stabilizers: The emulsifying property and packing behavior. Food

Hydrocolloids, 88, 180–189.Mandal, A., & Chakrabarty, D. (2011). Isolation of nanocellulose from waste sugarcane

bagasse (SCB) and its characterization. Carbohydrate Polymers, 86(3), 1291–1299.McClements, D. J., Bai, L., & Chung, C. (2017). Recent advances in the utilization of

natural emulsifiers to form and stabilize emulsions. Annual Review of Food Science &Technology, 8, 205–236.

Mikulcova, V., Bordes, R., Minarik, A., & Kasparkova, V. (2018). Pickering oil-in-wateremulsions stabilized by carboxylated cellulose nanocrystals - effect of the pH. FoodHydrocolloids, 80, 60–67.

Miller, R., Ferri, J. K., Javadi, A., Kraegel, J., Mucic, N., & Wuestneck, R. (2010).Rheology of interfacial layers. Colloid & Polymer Science, 288(9), 937–950.

Pandey, A., Derakhshandeh, M., Kedzior, S. A., Pilapil, B., Shomrat, N., Segal-Peretz, T.,et al. (2018). Role of interparticle interactions on microstructural and rheologicalproperties of cellulose nanocrystal stabilized emulsions. Journal of Colloid andInterface Science, 532, 808–818.

Paximada, P., Tsouko, E., Kopsahelis, N., Koutinas, A. A., & Mandala, I. (2016). Bacterialcellulose as stabilizer of o/w emulsions. Food Hydrocolloids, 53, 225–232.

Peng, N., Ding, X., Wang, Z., Cheng, Y., Gong, Z., Xu, X., et al. (2019). Novel dual re-sponsive alginate-based magnetic nanogels for oncotheranostics. CarbohydratePolymers, 204, 32–41.

Peng, X., Li, C., Liu, J., Yi, Z., & Han, Y. (2017). Changes in composition, cellulose de-gradability and biochemical methane potential of Miscanthus species during thegrowing season. Bioresource Technology, 235, 389–395.

Rosa, S. M. L., Rehman, N., de Miranda, M. I. G., Nachtigall, S. M. B., & Bica, C. I. D.(2012). Chlorine-free extraction of cellulose from rice husk and whisker isolation.Carbohydrate Polymers, 87(2), 1131–1138.

Seibert-Ludwig, D., Hahn, T., Hirth, T., & Zibek, S. (2019). Selection and optimization of asuitable pretreatment method for miscanthus and poplar raw material. Global ChangeBiology Bioenergy, 11(1), 171–180.

Shezad, O., Khan, S., Khan, T., & Park, J. K. (2010). Physicochemical and mechanicalcharacterization of bacterial cellulose produced with an excellent productivity instatic conditions using a simple fed-batch cultivation strategy. Carbohydrate Polymers,82(1), 173–180.

Si, S., Chen, Y., Fan, C., Hu, H., Li, Y., Huang, J., et al. (2015). Lignin extraction dis-tinctively enhances biomass enzymatic saccharification in hemicelluloses-richMiscanthus species under various alkali and acid pretreatments. BioresourceTechnology, 183, 248–254.

Tang, J., Sisler, J., Grishkewich, N., & Tam, K. C. (2017). Functionalization of cellulosenanocrystals for advanced applications. Journal of Colloid and Interface Science, 494,397–409.

Tavernier, I., Wijaya, W., Van der Meeren, P., Dewettinck, K., & Patel, A. R. (2016). Food-grade particles for emulsion stabilization. Trends in Food Science & Technology, 50,159–174.

Tenorio, A. T., Gieteling, J., Nikiforidis, C. V., Boom, R. M., & Van Der Goot, A. (2017).Interfacial properties of green leaf cellulosic particles. Food Hydrocolloids, 71, 8–16.

Wang, Y., Xu, Z., & Peng, L. (2014). Research progress in the groove structures of plantcell walls and biomass utilizations. Scientia Sinica, 44, 766–774.

Xiao, J., Li, Y., & Huang, Q. (2016a). Recent advances on food-grade particles stabilizedPickering emulsions: Fabrication, characterization and research trends. Trends in FoodScience & Technology, 55, 48–60.

Xiao, J., Wang, X. a., Gonzalez, A. J. P., & Huang, Q. (2016b). Kafirin nanoparticles-stabilized Pickering emulsions: Microstructure and rheological behavior. FoodHydrocolloids, 54, 30–39.

Xia, Y., Wu, J., Wei, W., Du, Y., Wan, T., Ma, X., et al. (2018). Exploiting the pliability andlateral mobility of Pickering emulsion for enhanced vaccination. Nature Materials,17(2), 187–194.

Xu, R., Mao, J., Peng, N., Luo, X., & Chang, C. (2018). Chitin/clay microspheres withhierarchical architecture for highly efficient removal of organic dyes. CarbohydratePolymers, 188, 143–150.

Yang, Y., Fang, Z., Chen, X., Zhang, W., Xie, Y., Chen, Y., et al. (2017). An overview ofpickering emulsions: Solid-particle materials, classification, morphology, and appli-cations. Frontiers in Pharmacology, 8, 287.

Zhai, X., Lin, D., Liu, D., & Yang, X. (2018). Emulsions stabilized by nanofibers frombacterial cellulose: New potential food-grade Pickering emulsions. Food ResearchInternational, 103, 12–20.

Zhang, T., Cheng, Q., Ye, D., & Chang, C. (2017). Dual physically cross-linked nano-composite hydrogels reinforced by tunicate cellulose nanocrystals with high tough-ness and good self-recoverability. ACS Applied Materials & Interfaces, 9, 24230–24237.

Zhu, Y., Luo, X., Wu, X., Li, W., Li, B., Lu, A., et al. (2016). Cellulose gel dispersions:Fascinating green particles for the stabilization of oil/water pickering emulsion.Cellulose, 24(1), 207–217.

Q. Li, et al. Food Hydrocolloids 97 (2019) 105214

9