论文部分内容阅读
Abstract： To reduce the adverse effects of non-cellulose materials on subsequent homogenization， the effects of a high-pressure homogenization treatment on the structure and properties of cellulose nanofibers （CNF） prepared by acid treatment of soybean residue were studied. The effects of the number of homogenization step on the microfibrillation degree， crystalline structure and mechanical properties of the soybean residue were analyzed by SEM， FT-IR， XRD， TG and DTG. The results showed that an increase in the number of homogenization steps led to an increase in the degree of microfibrillation， a more uniform distribution of the CNF diameter， and an increase in the crystallinity of CNF. However， but when the number of homogenization steps exceeded 15， the rate of change decreased， and the crystallinity of CNF decreased. As the number of homogenization steps increased， the average degree of polymerization and average molecular weight of CNF decreased continuously， and after 15 homogenization steps， their rate of change also decreased. Therefore， 15 steps of high-pressure homogenization represented a suitable number of steps to prepare the soybean residue CNF with an average diameter of 15 nm.
Keywords： soybean residue; cellulose nanofibers; high pressure homogenization; characteristics
1 Introduction
Soybean residue （SR）， an agricultural waste， is produced during the processing of soybean products. SR is a low cost， renewable resource， which is produced in large quantities. Presently， only a small part of SR is used for food consumption， while most of it is used as animal feed or is directly discarded， which has led to a great waste of resources as well as environmental pollution. SR is represented by the parenchymal tissue， with a thin cell wall and a weak binding force between the microfibrils， which is more conducive to the preparation of SR cellulose nanofibers. Its main components include dietary fiber （60% to 70%）， protein （13% to 20%）， fat （6% to 19%）， and ash （3.5% to 5%）， As such， it is useful for the preparation of cellulose nanofibers （CNF） by mechanical treatment.
Mechanical processing is the most energy-intensive part of CNF preparation. However， the amount of energy consumed is closely related to the number of mechanical treatments. It is well known that the number of mechanical treatments is directly proportional to the amount of energy consumed; however， the quality of CNF is not a simple linear relationship to the number of mechanical treatments. A study by Lee et al[1] showed that the degree of microfibrillation and the uniformity of diameter distribution of CNF increased when the number of homogenization steps increased. However， the mechanical properties of CNF remained homogeneous. Moreover， when the number of homogenization steps was between 5 and 10， the degree of microfibrillation and diameter distribution uniformity of CNF increased significantly. Another study by Leitner et al[2] prepared CNF by using beet pulp， also showed that the degree of microfibrillation increased proportionally to the number of homogenization steps. Moreover， after reaching a certain threshold， the degree of microfibrillation gradually decreased while the performance of CNF remained constant[3-4]. However， energy consumption was also increased. As such， this suggests that it is particularly important to control the number of mechanical treatments to reduce the energy consumption of CNF production while also obtaining CNF with good performance. This study uses agricultural waste soybean residue as raw materials to explore the effects of different homogenization times on the structure and properties of cellulose nanofibers （CNF） The effects of the number of homogenization steps on the degree of microfibrillation and the crystalline structure of CNF were analyzed by SEM， FT-IR， XRD， TG and DTG， which provided some references for the preparation of CNF from soybean residue cellulose. The preparation process， and structure and performance characteristics of soybean residue CNF were obtained. As such， the preparation materials of CNF were enriched， and the application field of soybean residue was broadened. The results of this study provide a theoretical basis for further research on nanocellulose.
2 Experimental
2.1 Materials and reagents
Raw materials： HCl-treated soybean residue （SR） was used as raw material and is recorded as HCl-S. The main components of SR include dietary fiber （60% to 70%）， protein （13% to 20%）， fat （6% to 19%）， and ash （3.5% to 5%）. The optimum process for acid treatment is HCl concentration 8%， reaction time of 80 min， and reaction temperature of 75℃， the yield of SR cellulose was 78.75%[5-6].
Hydrochloric acid （HCl）， analytical grade， Sinopharm Chemical Reagent Co.， Ltd.
2.2 Methods
2.2.1 Soybean residue CNF preparation
The SR was vacuum dried at 50℃ for 12 h， and was continuously kneaded to prevent it from being kneaded， then dispersed in a certain concentration of HCl solution， and reacted for a while in a temperature-controlled water bath shaker. After the reaction was completed， it was continuously washed with distilled water until the filtrate was neutral， and the solid matter was collected and freeze-dried. The freeze-dried sample was formulated into a 1% to 2% solution， homogenized under a pressure of 60 MPa， and freeze-dried to obtain CNF. The acid treated sample was recorded as HCl-S.
An HCl-S suspension with concentration of 1% to 2% was prepared and subsequently mixed using a mixer to ensure complete dispersion. Finally， high-pressure homogenization was applied. The homogenization pressure was set to 60 MPa and the suspensions passed through the high-pressure homogenizers 5， 10， 15 and 25 times and were recorded and stored as HGT-5， HGT-10， HGT-15， HGT-25， respectively.
2.2.2 Scanning electron microscopy analysis
CNF suspensions with different homogenization times was taken and diluted with distilled water respectively， followed by constant stirring to disperse them thoroughly. Subsequently， the samples were freeze-dried and an appropriate amount of freeze-dried floc sample was taken for sample preparation and observation after gold spraying. SEM pictures are taken by S4800， Hitachi， Japan. 2.2.3 Infrared spectroscopy
The dried sample was mixed with KBr in a ratio of 1：100 and thoroughly ground in an agate grinder. A suitable amount of the mixture was tableted， kept at a pressure of 10 MPa for 2 min， then placed in an infrared spectrometer （Vertex70， Bruker， Germany） for scanning. Detection conditions were as follows： wave number resolution of 1 cm-1， wave number 500～4000 cm-1.
2.2.4 X-ray diffraction analysis
The dried sample was analyzed to determine the crystalline structure of the sample by a D/max 2200PC type X-ray diffractometer. Using a Cu target Ka ray， the tube pressure was set at 40 kV， tube flow was 40 mA， the scanning range was 5°～60°， the scanning speed was 2 s/step， and the angle of each step was 0.02°.
The crystallinity of the sample was expressed by the crystallinity index （CrI）， which is determined by the empirical formula of Segal et al[7] and the specific formula is shown in Formula （1）.
Where， I200 is the diffraction intensity of the （002） crystal plane， Iam is the diffraction intensity of the amorphous region， and for the cellulose I， 2q=18° diffraction intensity.
2.2.5 Thermogravimetric analysis
The thermal stability of the samples was analyzed using a STA449F3-1053-M synchronous TG-DSC thermal analyzer. Detection conditions were as follows： 3 to 5 mg of sample （oven dry）， temperature of 25 to 600℃， temperature increase rate of 10℃/min， nitrogen flow rate of 40 mL/min.
2.2.6 Degree of polymerization
In this experiment， the degree of polymerization （DP） of the soybean residue cellulose was measured by using the copper ethylenediamine viscosity method. In the sample bottle， an appropriate amount of the dry sample， 25 mL of distilled water， 25 mL of copper ethylenediamine solution， and 2 to 3 pieces of copper pieces were sequentially added and then repeatedly shaken until the sample was completely dissolved. Finally， the average DP of the soybean residue cellulose was calculated by measuring the viscosity of the solution. The calculation method used is shown in the following Formula （2）：
Where， DP is the average degree of polymerization of soybean residue cellulose; [h] is the intrinsic viscosity of the solution， mL/g.
3 Results and discussion
3.1 SEM analysis of soybean residue CNF
The morphologies of CNF with different homogenization times ars shown in Fig.1. In the early stages of homogenization， the cell wall of soybean residue was damaged and CNF was obtained. However， some cell fragments were still present， as shown in Fig.1（a）. As the number of homogenization steps increased， the cellular debris gradually decreased and more CNF were isolated[8-10]. By comparing the SEM images representing 15 and 25 homogenization steps， it can be seen that by the 15th homogenization step the cell wall is almost destroyed. Moreover， when the number of homogenization steps is increased， the effect of mechanical treatment can no longer be observed. As such， this suggests that when the number of homogenization steps exceeds 15， the efficiency of the mechanical treatment is reduced. Considering the energy consumption of mechanical treatments， it is suggest that the most optimal number of homogenization steps for CNF preparation is 15[11-12]. Due to the presence of numerous hydrogen bonds on the surface of the CNF， it has strong interaction forces， thereby easily forming aggregates which are difficult to separate[13]. Therefore， CNF is not a homogeneous solution. Since CNF is generally a solution where single CNF and CNF aggregates coexist， the SEM images representing CNF are a network structure.
3.2 FT-IR analyses of soybean residue CNF
The infrared spectra of CNF prepared in a different number of homogenization treatments are shown in Fig.2. Compared with the infrared spectrogram of soybean residue cellulose， the CNF characteristic absorption peak did not change significantly， indicating that the change in the number of homogenizations does not change the chemical structure of cellulose molecules， and therefore， would not introduce new chemical functional groups. As such， the prepared CNF still has the chemical structure of natural cellulose[14-15]. Among them， the absorption peak at 3600 to 3200 cm-1 is wide， indicating that there are a large number of O—H stretching vibration groups in the soybean residue， and they mainly exist in the form of self-hydrogen bonding. As the non-cellulosic material was removed during the acid pretreatment process， more hydroxyl groups were released， so it gradually moved toward higher frequencies. The absorption peak at 2900 cm-1 is a stretching vibration absorption peak of a C—H bond of a methyl group or a methylene group in cellulose and hemicellulose. The absorption peak at 1730 cm-1 is mainly formed by vibration of acetyl groups in hemicellulose or ferulic acid in hemicellulose and carbonyl groups in hydroxyphenylacrylic acid. The absorption peak at 1373 cm-1 is caused by the bending vibration of the methyl group （—CH3）; the absorption peak at 1049 cm-1 is caused by the stretching vibration of the C—O bond in the cellulose and the bending vibration of the O—H bond. The absorption peak at 897 cm-1 is a bending vibration of an O—H bond in cellulose and a deformation vibration absorption peak of a C—H bond.
3.3 XRD analyses of soybean residue CNF
X-ray diffraction spectra of CNF with different numbers of homogenization steps are shown in Fig.3. As seen in Fig.3， the number of homogenization steps did not significantly change the X-ray diffraction spectra of CNF， affecting only the absorption peak intensity， thereby indicating that the number of homogenization steps did not change the crystalline structure of cellulose， with the CNF still having the cellulose I crystal structure. The crystallinity of CNF with different homogenization times is shown in Table 1. As shown in Table 1， the crystallinity of CNF first increases and then subsequently decreases with the increase in homogenization steps. When the number of homogenization steps exceeds 10， CNF crystallinity increases and then subsequently decreases when the number of homogenization steps was more than 15. In the early stages of homogenization， the cell wall of soybean residue was damaged， and amorphous substances， such as proteins and hemicellulose， were dissolved， which reduced the amorphous area and increased the diffraction intensity. However， as underwent homogenization， shearing， and other mechanical forces， the suspension of soybean residue fiber led to a degradation of the cellulose amorphous area， thereby exposing the crystalline region. As such， the initial stages of homogenization led to an increase in CNF crystallinity and diffraction intensity. The crystalline regions of cellulose were arranged in a regular manner， with strong intermolecular forces which are difficult to damage. Therefore， in the middle stages of homogenization， the crystallinity of CNF remained constant. With the further increase of the number of homogenization steps， the strong mechanical force began to damage the crystallization areas， causing the dense structure of the cellulose crystallization area to become loose， leading to a reduction in the crystallinity of CNF and thereby the reduction of diffraction strength. 3.4 TGA analyses of soybean residue CNF
The TG （a） and DTG （b） curves of CNF with different homogenization times are shown in Fig.4 and the main pyrolysis parameters are shown in Table 2. When compared with HCl-S， the characteristic peak of the DTG curve did not change significantly.
As shown in Fig.4（b）， the temperature of CNF’s maximum weight loss peak at a different number of homogenization steps did not change significantly. However， the maximum mass loss rate was significantly different. It can be seen in Fig.4（a） that the residual quantity of CNF with a different number of homogenization steps was also significantly different. As shown in Table 2， the maximum mass-loss rates of HGT-10 and HGT-15 are 11.41% and 11.51%， respectively， remaining constant. However， both values are smaller with respect to the maximum mass-loss rates of HGT-5 （13.12%） and HGT-25 （13.45%）， while the maximum loss rate of HGT-25 is slightly higher than that of HGT-5. At the same time， the residual quantities of HGT-10 （30.10%）， HGT-15 （31.24%） are all larger than those of HGT-5 （18.74%） and HGT-25 （18.85%）， but the differences between them are small. According to the analysis， the crystallinities of HGT-10 and HGT-15 are greater than those of HGT-5 and HGT-25， indicating that there are more crystalline regions in HGT-10 and HGT-15， that the rate of solution at the same temperature drop is smaller， and that they require more energy to be decomposed. Therefore， this suggests that their maximum loss rate is small， thereby leading to more residues.
3.5 Effect of homogenization on the CNF aggregation structure of soybean residue
The DP of cellulose indicates the number of glucose units connected in the cellulose molecular chain and indirectly indicates the length of the cellulose molecular chain. As such， cellulose degradation under the action of chemical and physical factors will change the DP of cellulose[16]. The effect of homogenization times on the DP of CNF aggregation is shown in Table 3.
As shown in Table 3， the DP of CNF aggregation decreases with the increase in the number of homogenization steps. This suggests that in the initial stages of the homogenization treatment， the DP of CNF significantly changes， as the cellulose has undergone drastic degradation. Moreover， this indicates that the homogenization efficiency is high[17]. When the number of homogenization steps exceeds 15， the DP of CNF tends to be stable， indicating that the homogeneity efficiency begins to decrease. In the initial stages of homogenization， the cellulose molecular chain was relatively long. However， when passing through the homogenization valve of the homogenizer， it was subjected to intense shear， cavitation and impact forces. As a result， the molecular chain was rapidly reduced and subsequently， the DP was reduced. When the number of homogenization steps increases， the number of small cellulose molecular chains also increases. These are easier to pass through the homogenization valve， and as such the effects of shearing， cavitation， and impact are decreased. Therefore， when the number of homogenization steps exceeds a certain number， the DP of cellulose is no longer significant. 4 Conclusions
In this paper， the effects of a high-pressure homogenization treatment on the structure and properties of CNF prepared by acid treatment of soybean residue were studied. The results showed that when the number of homogenization steps was varied within the range of 5 to 25， the increase in the number of homogenization steps led to an increase in the degree of microfibrillation. The crystallinity firstly increased and then decreased， but the crystal structure of cellulose did not change， and the cellulose I remained. However， when the number of homogenization steps exceeded 15， the average degree of polymerization and average molecular weight of CNF gradually decreased， and the thermal stability of CNF was less affected. The pyrolysis temperature of CNF with different steps of homogenization corresponded to the maximum weight loss peak. The temperature difference was small， only the weight loss rate and the residual amount were different， which was related to the crystallinity of the CNF. The greater the crystallinity of the CNF， the more energy was required for pyrolysis， and the smaller the weight loss rate， the more the residual amount. Therefore， the 15 steps of high pressure homogenization was suitable to prepare the soybean residue CNF with an average diameter of 15 nm.
Acknowledgments
This work was supported by State Key Laboratory of Pulp and Paper Engineering （201819）， the project of Shaanxi Provincial Department of Education Key Laboratory Research Open Fund （Grant No. 17JS017）， and the project of Shaanxi University of Science and Technology Research Initial Fund （Grant No. BJ15-29）.
References
[1] Lee S Y， Chun S J， Kang I A， et al. Preparation of cellulose nanofibrils by high-pressure homogenizer and cellulose-based composite films[J]. Journal of Industrial & Engineering Chemistry， 2009， 15（1）： 50-55.
[2] Leitner J， Hinterstoisser B， Wastyn M， et al. Sugar beet cellulose nanofibril-reinforced composites[J]. Cellulose， 2007， 14（5）： 419-425.
[3] Donsì F， Annunziata M， Ferrari G. Microbial inactivation by high pressure homogenization： Effect of the disruption valve geometry[J]. Journal of Food Engineering， 2013， 115（3）： 362-370.
[4] Carre?o J M， Gurrea M C， Sampedro F， et al. Effect of high hydrostatic pressure and high-pressure homogenisation on Lactobacillus plantarum inactivation kinetics and quality parameters of mandarin juice[J]. European Food Research & Technology， 2010， 232（2）： 265-274. [5] Li P， Wang Y， Hou Q， et al. Effect of Pretreatment on the Structure and Properties of Nanofibrillated Cellulose from Soybean Residues[J]. BioResources， 2018， 14（1）： 554-560.
[6] Li P， Wang Y， Hou Q， et al. Isolation and characterization of microfibrillated cellulose from agro-industrial soybean residue （okara）[J]. BioResources， 2018， 13（4）： 7944-7956.
[7] Segal L， Creely J J， Martin A E， et al. An Empirical Method for Estimating the Degree of Crystallinity of Native Cellulose Using the X-Ray Diffractometer[J]. Textile Research Journal， 1959， 29（10）： 786-794.
[8] Galvis-Sánchez A C， Castro M C R， Biernacki K， et al. Natural deep eutectic solvents as Green plasticizers for chitosan thermoplastic production with controlled/desired mechanical and barrier properties[J]. Food Hydrocolloids， 2018， 82：S0268005X18301206.
[9] Zhao Z， Chen X， Ali M F， et al. Pretreatment of wheat straw using basic ethanolamine-based deep eutectic solvents for improving enzymatic hydrolysis[J]. Bioresource Technology， 2018， 263： S0960852418306655.
[10] Huang Y ， Feng F ， Chen Z G ， et al. Green and efficient removal of cadmium from rice flour using natural deep eutectic solvents[J]. Food Chemistry， 2018， 244： 260-265.
[11] Diels A M J， Callewaert L， Wuytack E Y， et al. Moderate Temperatures Affect Escherichia coli Inactivation by High‐Pressure Homogenization Only through Fluid Viscosity[J]. Biotechnology Progress， 2010， 20（5）： 1512-1517.
[12] Dong P， Georget E S， Aganovic K， et al. Ultra high pressure homogenization （UHPH） inactivation of Bacillus amyloliquefaciens spores in phosphate buffered saline （PBS） and milk[J]. Frontiers in Microbiology， 2015， 6（712）： 712.
[13] Ramamoorthy S K， Skrifvars M， Persson A. A Review of Natural Fibers Used in Biocomposites： Plant， Animal and Regenerated Cellulose Fibers[J]. Polymer Reviews， 2015， 55（1）： 107-162.
[14] Deepa B， Abraham E， Cordeiro N， et al. Utilization of various lignocellulosic biomass for the production of nanocellulose： a comparative study[J]. Cellulose， 2015， 22（2）： 1075-1090.
[15] Laitinen O， Ojala J， Sirvi? J A， et al. Sustainable stabilization of oil in water emulsions by cellulose nanocrystals synthesized from deep eutectic solvents[J]. Cellulose， 2017， 24（4）： 1679-1689.
[16] Leitner J， Hinterstoisser B， Wastyn M， et al. Sugar beet cellulose nanofibril-reinforced composites[J]. Cellulose， 2007， 14（5）： 419-425.
[17] Lee S Y， Chun S J， Kang I A， et al. Preparation of cellulose nanofibrils by high-pressure homogenizer and cellulose-based composite films[J]. Journal of Industrial & Engineering Chemistry， 2009， 15（1）： 50-55.
Keywords： soybean residue; cellulose nanofibers; high pressure homogenization; characteristics
1 Introduction
Soybean residue （SR）， an agricultural waste， is produced during the processing of soybean products. SR is a low cost， renewable resource， which is produced in large quantities. Presently， only a small part of SR is used for food consumption， while most of it is used as animal feed or is directly discarded， which has led to a great waste of resources as well as environmental pollution. SR is represented by the parenchymal tissue， with a thin cell wall and a weak binding force between the microfibrils， which is more conducive to the preparation of SR cellulose nanofibers. Its main components include dietary fiber （60% to 70%）， protein （13% to 20%）， fat （6% to 19%）， and ash （3.5% to 5%）， As such， it is useful for the preparation of cellulose nanofibers （CNF） by mechanical treatment.
Mechanical processing is the most energy-intensive part of CNF preparation. However， the amount of energy consumed is closely related to the number of mechanical treatments. It is well known that the number of mechanical treatments is directly proportional to the amount of energy consumed; however， the quality of CNF is not a simple linear relationship to the number of mechanical treatments. A study by Lee et al[1] showed that the degree of microfibrillation and the uniformity of diameter distribution of CNF increased when the number of homogenization steps increased. However， the mechanical properties of CNF remained homogeneous. Moreover， when the number of homogenization steps was between 5 and 10， the degree of microfibrillation and diameter distribution uniformity of CNF increased significantly. Another study by Leitner et al[2] prepared CNF by using beet pulp， also showed that the degree of microfibrillation increased proportionally to the number of homogenization steps. Moreover， after reaching a certain threshold， the degree of microfibrillation gradually decreased while the performance of CNF remained constant[3-4]. However， energy consumption was also increased. As such， this suggests that it is particularly important to control the number of mechanical treatments to reduce the energy consumption of CNF production while also obtaining CNF with good performance. This study uses agricultural waste soybean residue as raw materials to explore the effects of different homogenization times on the structure and properties of cellulose nanofibers （CNF） The effects of the number of homogenization steps on the degree of microfibrillation and the crystalline structure of CNF were analyzed by SEM， FT-IR， XRD， TG and DTG， which provided some references for the preparation of CNF from soybean residue cellulose. The preparation process， and structure and performance characteristics of soybean residue CNF were obtained. As such， the preparation materials of CNF were enriched， and the application field of soybean residue was broadened. The results of this study provide a theoretical basis for further research on nanocellulose.
2 Experimental
2.1 Materials and reagents
Raw materials： HCl-treated soybean residue （SR） was used as raw material and is recorded as HCl-S. The main components of SR include dietary fiber （60% to 70%）， protein （13% to 20%）， fat （6% to 19%）， and ash （3.5% to 5%）. The optimum process for acid treatment is HCl concentration 8%， reaction time of 80 min， and reaction temperature of 75℃， the yield of SR cellulose was 78.75%[5-6].
Hydrochloric acid （HCl）， analytical grade， Sinopharm Chemical Reagent Co.， Ltd.
2.2 Methods
2.2.1 Soybean residue CNF preparation
The SR was vacuum dried at 50℃ for 12 h， and was continuously kneaded to prevent it from being kneaded， then dispersed in a certain concentration of HCl solution， and reacted for a while in a temperature-controlled water bath shaker. After the reaction was completed， it was continuously washed with distilled water until the filtrate was neutral， and the solid matter was collected and freeze-dried. The freeze-dried sample was formulated into a 1% to 2% solution， homogenized under a pressure of 60 MPa， and freeze-dried to obtain CNF. The acid treated sample was recorded as HCl-S.
An HCl-S suspension with concentration of 1% to 2% was prepared and subsequently mixed using a mixer to ensure complete dispersion. Finally， high-pressure homogenization was applied. The homogenization pressure was set to 60 MPa and the suspensions passed through the high-pressure homogenizers 5， 10， 15 and 25 times and were recorded and stored as HGT-5， HGT-10， HGT-15， HGT-25， respectively.
2.2.2 Scanning electron microscopy analysis
CNF suspensions with different homogenization times was taken and diluted with distilled water respectively， followed by constant stirring to disperse them thoroughly. Subsequently， the samples were freeze-dried and an appropriate amount of freeze-dried floc sample was taken for sample preparation and observation after gold spraying. SEM pictures are taken by S4800， Hitachi， Japan. 2.2.3 Infrared spectroscopy
The dried sample was mixed with KBr in a ratio of 1：100 and thoroughly ground in an agate grinder. A suitable amount of the mixture was tableted， kept at a pressure of 10 MPa for 2 min， then placed in an infrared spectrometer （Vertex70， Bruker， Germany） for scanning. Detection conditions were as follows： wave number resolution of 1 cm-1， wave number 500～4000 cm-1.
2.2.4 X-ray diffraction analysis
The dried sample was analyzed to determine the crystalline structure of the sample by a D/max 2200PC type X-ray diffractometer. Using a Cu target Ka ray， the tube pressure was set at 40 kV， tube flow was 40 mA， the scanning range was 5°～60°， the scanning speed was 2 s/step， and the angle of each step was 0.02°.
The crystallinity of the sample was expressed by the crystallinity index （CrI）， which is determined by the empirical formula of Segal et al[7] and the specific formula is shown in Formula （1）.
Where， I200 is the diffraction intensity of the （002） crystal plane， Iam is the diffraction intensity of the amorphous region， and for the cellulose I， 2q=18° diffraction intensity.
2.2.5 Thermogravimetric analysis
The thermal stability of the samples was analyzed using a STA449F3-1053-M synchronous TG-DSC thermal analyzer. Detection conditions were as follows： 3 to 5 mg of sample （oven dry）， temperature of 25 to 600℃， temperature increase rate of 10℃/min， nitrogen flow rate of 40 mL/min.
2.2.6 Degree of polymerization
In this experiment， the degree of polymerization （DP） of the soybean residue cellulose was measured by using the copper ethylenediamine viscosity method. In the sample bottle， an appropriate amount of the dry sample， 25 mL of distilled water， 25 mL of copper ethylenediamine solution， and 2 to 3 pieces of copper pieces were sequentially added and then repeatedly shaken until the sample was completely dissolved. Finally， the average DP of the soybean residue cellulose was calculated by measuring the viscosity of the solution. The calculation method used is shown in the following Formula （2）：
Where， DP is the average degree of polymerization of soybean residue cellulose; [h] is the intrinsic viscosity of the solution， mL/g.
3 Results and discussion
3.1 SEM analysis of soybean residue CNF
The morphologies of CNF with different homogenization times ars shown in Fig.1. In the early stages of homogenization， the cell wall of soybean residue was damaged and CNF was obtained. However， some cell fragments were still present， as shown in Fig.1（a）. As the number of homogenization steps increased， the cellular debris gradually decreased and more CNF were isolated[8-10]. By comparing the SEM images representing 15 and 25 homogenization steps， it can be seen that by the 15th homogenization step the cell wall is almost destroyed. Moreover， when the number of homogenization steps is increased， the effect of mechanical treatment can no longer be observed. As such， this suggests that when the number of homogenization steps exceeds 15， the efficiency of the mechanical treatment is reduced. Considering the energy consumption of mechanical treatments， it is suggest that the most optimal number of homogenization steps for CNF preparation is 15[11-12]. Due to the presence of numerous hydrogen bonds on the surface of the CNF， it has strong interaction forces， thereby easily forming aggregates which are difficult to separate[13]. Therefore， CNF is not a homogeneous solution. Since CNF is generally a solution where single CNF and CNF aggregates coexist， the SEM images representing CNF are a network structure.
3.2 FT-IR analyses of soybean residue CNF
The infrared spectra of CNF prepared in a different number of homogenization treatments are shown in Fig.2. Compared with the infrared spectrogram of soybean residue cellulose， the CNF characteristic absorption peak did not change significantly， indicating that the change in the number of homogenizations does not change the chemical structure of cellulose molecules， and therefore， would not introduce new chemical functional groups. As such， the prepared CNF still has the chemical structure of natural cellulose[14-15]. Among them， the absorption peak at 3600 to 3200 cm-1 is wide， indicating that there are a large number of O—H stretching vibration groups in the soybean residue， and they mainly exist in the form of self-hydrogen bonding. As the non-cellulosic material was removed during the acid pretreatment process， more hydroxyl groups were released， so it gradually moved toward higher frequencies. The absorption peak at 2900 cm-1 is a stretching vibration absorption peak of a C—H bond of a methyl group or a methylene group in cellulose and hemicellulose. The absorption peak at 1730 cm-1 is mainly formed by vibration of acetyl groups in hemicellulose or ferulic acid in hemicellulose and carbonyl groups in hydroxyphenylacrylic acid. The absorption peak at 1373 cm-1 is caused by the bending vibration of the methyl group （—CH3）; the absorption peak at 1049 cm-1 is caused by the stretching vibration of the C—O bond in the cellulose and the bending vibration of the O—H bond. The absorption peak at 897 cm-1 is a bending vibration of an O—H bond in cellulose and a deformation vibration absorption peak of a C—H bond.
3.3 XRD analyses of soybean residue CNF
X-ray diffraction spectra of CNF with different numbers of homogenization steps are shown in Fig.3. As seen in Fig.3， the number of homogenization steps did not significantly change the X-ray diffraction spectra of CNF， affecting only the absorption peak intensity， thereby indicating that the number of homogenization steps did not change the crystalline structure of cellulose， with the CNF still having the cellulose I crystal structure. The crystallinity of CNF with different homogenization times is shown in Table 1. As shown in Table 1， the crystallinity of CNF first increases and then subsequently decreases with the increase in homogenization steps. When the number of homogenization steps exceeds 10， CNF crystallinity increases and then subsequently decreases when the number of homogenization steps was more than 15. In the early stages of homogenization， the cell wall of soybean residue was damaged， and amorphous substances， such as proteins and hemicellulose， were dissolved， which reduced the amorphous area and increased the diffraction intensity. However， as underwent homogenization， shearing， and other mechanical forces， the suspension of soybean residue fiber led to a degradation of the cellulose amorphous area， thereby exposing the crystalline region. As such， the initial stages of homogenization led to an increase in CNF crystallinity and diffraction intensity. The crystalline regions of cellulose were arranged in a regular manner， with strong intermolecular forces which are difficult to damage. Therefore， in the middle stages of homogenization， the crystallinity of CNF remained constant. With the further increase of the number of homogenization steps， the strong mechanical force began to damage the crystallization areas， causing the dense structure of the cellulose crystallization area to become loose， leading to a reduction in the crystallinity of CNF and thereby the reduction of diffraction strength. 3.4 TGA analyses of soybean residue CNF
The TG （a） and DTG （b） curves of CNF with different homogenization times are shown in Fig.4 and the main pyrolysis parameters are shown in Table 2. When compared with HCl-S， the characteristic peak of the DTG curve did not change significantly.
As shown in Fig.4（b）， the temperature of CNF’s maximum weight loss peak at a different number of homogenization steps did not change significantly. However， the maximum mass loss rate was significantly different. It can be seen in Fig.4（a） that the residual quantity of CNF with a different number of homogenization steps was also significantly different. As shown in Table 2， the maximum mass-loss rates of HGT-10 and HGT-15 are 11.41% and 11.51%， respectively， remaining constant. However， both values are smaller with respect to the maximum mass-loss rates of HGT-5 （13.12%） and HGT-25 （13.45%）， while the maximum loss rate of HGT-25 is slightly higher than that of HGT-5. At the same time， the residual quantities of HGT-10 （30.10%）， HGT-15 （31.24%） are all larger than those of HGT-5 （18.74%） and HGT-25 （18.85%）， but the differences between them are small. According to the analysis， the crystallinities of HGT-10 and HGT-15 are greater than those of HGT-5 and HGT-25， indicating that there are more crystalline regions in HGT-10 and HGT-15， that the rate of solution at the same temperature drop is smaller， and that they require more energy to be decomposed. Therefore， this suggests that their maximum loss rate is small， thereby leading to more residues.
3.5 Effect of homogenization on the CNF aggregation structure of soybean residue
The DP of cellulose indicates the number of glucose units connected in the cellulose molecular chain and indirectly indicates the length of the cellulose molecular chain. As such， cellulose degradation under the action of chemical and physical factors will change the DP of cellulose[16]. The effect of homogenization times on the DP of CNF aggregation is shown in Table 3.
As shown in Table 3， the DP of CNF aggregation decreases with the increase in the number of homogenization steps. This suggests that in the initial stages of the homogenization treatment， the DP of CNF significantly changes， as the cellulose has undergone drastic degradation. Moreover， this indicates that the homogenization efficiency is high[17]. When the number of homogenization steps exceeds 15， the DP of CNF tends to be stable， indicating that the homogeneity efficiency begins to decrease. In the initial stages of homogenization， the cellulose molecular chain was relatively long. However， when passing through the homogenization valve of the homogenizer， it was subjected to intense shear， cavitation and impact forces. As a result， the molecular chain was rapidly reduced and subsequently， the DP was reduced. When the number of homogenization steps increases， the number of small cellulose molecular chains also increases. These are easier to pass through the homogenization valve， and as such the effects of shearing， cavitation， and impact are decreased. Therefore， when the number of homogenization steps exceeds a certain number， the DP of cellulose is no longer significant. 4 Conclusions
In this paper， the effects of a high-pressure homogenization treatment on the structure and properties of CNF prepared by acid treatment of soybean residue were studied. The results showed that when the number of homogenization steps was varied within the range of 5 to 25， the increase in the number of homogenization steps led to an increase in the degree of microfibrillation. The crystallinity firstly increased and then decreased， but the crystal structure of cellulose did not change， and the cellulose I remained. However， when the number of homogenization steps exceeded 15， the average degree of polymerization and average molecular weight of CNF gradually decreased， and the thermal stability of CNF was less affected. The pyrolysis temperature of CNF with different steps of homogenization corresponded to the maximum weight loss peak. The temperature difference was small， only the weight loss rate and the residual amount were different， which was related to the crystallinity of the CNF. The greater the crystallinity of the CNF， the more energy was required for pyrolysis， and the smaller the weight loss rate， the more the residual amount. Therefore， the 15 steps of high pressure homogenization was suitable to prepare the soybean residue CNF with an average diameter of 15 nm.
Acknowledgments
This work was supported by State Key Laboratory of Pulp and Paper Engineering （201819）， the project of Shaanxi Provincial Department of Education Key Laboratory Research Open Fund （Grant No. 17JS017）， and the project of Shaanxi University of Science and Technology Research Initial Fund （Grant No. BJ15-29）.
References
[1] Lee S Y， Chun S J， Kang I A， et al. Preparation of cellulose nanofibrils by high-pressure homogenizer and cellulose-based composite films[J]. Journal of Industrial & Engineering Chemistry， 2009， 15（1）： 50-55.
[2] Leitner J， Hinterstoisser B， Wastyn M， et al. Sugar beet cellulose nanofibril-reinforced composites[J]. Cellulose， 2007， 14（5）： 419-425.
[3] Donsì F， Annunziata M， Ferrari G. Microbial inactivation by high pressure homogenization： Effect of the disruption valve geometry[J]. Journal of Food Engineering， 2013， 115（3）： 362-370.
[4] Carre?o J M， Gurrea M C， Sampedro F， et al. Effect of high hydrostatic pressure and high-pressure homogenisation on Lactobacillus plantarum inactivation kinetics and quality parameters of mandarin juice[J]. European Food Research & Technology， 2010， 232（2）： 265-274. [5] Li P， Wang Y， Hou Q， et al. Effect of Pretreatment on the Structure and Properties of Nanofibrillated Cellulose from Soybean Residues[J]. BioResources， 2018， 14（1）： 554-560.
[6] Li P， Wang Y， Hou Q， et al. Isolation and characterization of microfibrillated cellulose from agro-industrial soybean residue （okara）[J]. BioResources， 2018， 13（4）： 7944-7956.
[7] Segal L， Creely J J， Martin A E， et al. An Empirical Method for Estimating the Degree of Crystallinity of Native Cellulose Using the X-Ray Diffractometer[J]. Textile Research Journal， 1959， 29（10）： 786-794.
[8] Galvis-Sánchez A C， Castro M C R， Biernacki K， et al. Natural deep eutectic solvents as Green plasticizers for chitosan thermoplastic production with controlled/desired mechanical and barrier properties[J]. Food Hydrocolloids， 2018， 82：S0268005X18301206.
[9] Zhao Z， Chen X， Ali M F， et al. Pretreatment of wheat straw using basic ethanolamine-based deep eutectic solvents for improving enzymatic hydrolysis[J]. Bioresource Technology， 2018， 263： S0960852418306655.
[10] Huang Y ， Feng F ， Chen Z G ， et al. Green and efficient removal of cadmium from rice flour using natural deep eutectic solvents[J]. Food Chemistry， 2018， 244： 260-265.
[11] Diels A M J， Callewaert L， Wuytack E Y， et al. Moderate Temperatures Affect Escherichia coli Inactivation by High‐Pressure Homogenization Only through Fluid Viscosity[J]. Biotechnology Progress， 2010， 20（5）： 1512-1517.
[12] Dong P， Georget E S， Aganovic K， et al. Ultra high pressure homogenization （UHPH） inactivation of Bacillus amyloliquefaciens spores in phosphate buffered saline （PBS） and milk[J]. Frontiers in Microbiology， 2015， 6（712）： 712.
[13] Ramamoorthy S K， Skrifvars M， Persson A. A Review of Natural Fibers Used in Biocomposites： Plant， Animal and Regenerated Cellulose Fibers[J]. Polymer Reviews， 2015， 55（1）： 107-162.
[14] Deepa B， Abraham E， Cordeiro N， et al. Utilization of various lignocellulosic biomass for the production of nanocellulose： a comparative study[J]. Cellulose， 2015， 22（2）： 1075-1090.
[15] Laitinen O， Ojala J， Sirvi? J A， et al. Sustainable stabilization of oil in water emulsions by cellulose nanocrystals synthesized from deep eutectic solvents[J]. Cellulose， 2017， 24（4）： 1679-1689.
[16] Leitner J， Hinterstoisser B， Wastyn M， et al. Sugar beet cellulose nanofibril-reinforced composites[J]. Cellulose， 2007， 14（5）： 419-425.
[17] Lee S Y， Chun S J， Kang I A， et al. Preparation of cellulose nanofibrils by high-pressure homogenizer and cellulose-based composite films[J]. Journal of Industrial & Engineering Chemistry， 2009， 15（1）： 50-55.