A comparative study of the rheological properties of three different nanofibrillated cellulose systems

The

enzymatically pre-treated NFC systems are characterized by close-to-uncharged bundles of nanofibrils (width < 100 nm), exhibiting a broad size distribution.On the other hand, TEMPO-based or carboxymethylated NFC systems are highly charged systems, exhibiting a much narrower size distribution.The measured widths of the NFC strands of these systems have been reported to range from 3-20 nm (Wågberg et al. 2008;Isogai et al. 2011).
Rheological characterization of NFC-based systems is understandably complicated by the diverse characteristics of the different NFC systems, which requires the development of different measuring protocols.For example, the authors employed pre-shearing of the samples in the analysis of a CMC-grafted NFC system (Naderi et al. 2015c) with phase-separation tendencies.A similar protocol was also found to be necessary in the study of an enzymatically pre-treated NFC (Naderi, Lindström 2015), which is also an unstable system; otherwise erroneous results would have been obtained.On the other hand, pre-shearing of samples was not viewed as necessary when the same authors investigated a highly charged carboxymethylated NFC (Naderi et al. 2014), which displayed a high colloidal stability.
Rheological measurements are further complicated by the fact that the recorded rheological properties obtained by standard rotational viscometers are the net result of all the inter-particle and particle-measuring chamber interactions.The results are hence apparent, which may have ramifications on the analysis of the results.Hence, it is necessary to combine rheological measurements with other methodologies, something that is often difficult, due to instrumental limitations.It is noted, though, that contributions exist which have addressed this inadequacy.However, the value of the approaches remains to be independently verified.For example, (Nechyporchuk et al. 2014) combined rotational viscometry measurements with camera recordings.The authors observed that a cone-plate geometry with roughened surfaces (when compared with a geometry with smooth surfaces) diminished wall-slip effects and flow instabilities.The investigations further revealed the existence of shearbanding phenomena in certain shearing conditions.However, the importance of flow instabilities and shear banding in the interpretation of the recorded rheological properties was not obvious from authors' report.On the other hand, (Karppinen et al. 2012) combined a rotational viscometry (employing a bob and cup geometry -with the latter being produced in a transparent material) with photography to investigate a commercial, low-charged NFC system.The authors reported changes in the NFC floc structure when altering the shear rate.However, the importance of the observations is ambiguous considering the macro-size of the flocs (contrasting with the nano size of NFC systems) and the afore-mentioned wall-slip effects.
As touched upon, wall-slip (also known as wall depletion) effects constitute another major challenge in the study of the rheology of NFC-based systems.Slip effects occur due to the depletion of the dispersed particles from the solid boundaries (measuring-chamber walls) leaving a liquid layer that has a lower viscosity than the bulk viscosity.Slip effects are often observed for less stable systems (Buscall 2010); for example Nechyporchuk et al. (2014) reported of slip effects for a (presumably) low-charged enzymatically pre-treated NFC, but none for a highly charged TEMPO-based NFC system.The slip effects for less stable systems might, however, be reduced by "proper" shearing of the systems prior to the measurements (Kelessidis et al. 2007;Bonn, Denn 2009).
Wall depletion effects occur also more readily at smooth boundaries, which can lead to significant errors when the relative viscosity (the ratio between the viscosity of the dispersion and the liquid phase) of the suspension is high ( 100, (Buscall 2010)).The wall depletion effects can, however, be significantly reduced by using serrated surfaces; the scale of surface roughness should then be of the same order as the size of the dispersed particles (Buscall 2010).In this context, it is noted that a widely publicized method based on a wide-gap bucket vane viscometer was recently proposed as an attractive methodology for performing rheological measurements with no wall depletion effects (Mohtaschemi et al. 2014).However, as has been pointed out by Nechyporchuk et al. (2015) the methodology is probably most useful in the study of dilute NFC systems (which can be effectively stirred).Furthermore, the methodology of Mohtaschemi et al. was also argued to be affected by flow localization effects.It is clear that awareness of the existence and the impact of the slip effects on the rheological properties of any investigated system is important for the analysis of the measurements.
This report aims to elucidate the impact of temperature and ionic strength (altered by a monovalent salt) on the rheological properties of enzymatic pre-treated, carboxymethylated and CMC-grafted NFC systems (that have been produced by the same amount of mechanical shearing) with the expectation of finding different properties due to the varying chemistries of the systems.The studies are further expanded by investigating the thixotropic behaviour of the systems.The results are rationalized by taking into account the impact of eventual slip effects, which are distinguished by a route that has been proposed by Kelessidis et al. (2010).
The following materials were used in the enzymatic pretreatment process.KH 2 PO 4 and Na 2 HPO 4 (GPR Rectapur) were purchased from VWR (Sweden).These chemicals constitute the phosphate buffer in the enzymatic treatment process.A mono-component endoglucanase enzyme (FiberCare®R) was obtained from Novozym A/S (Denmark).A microbiocide, 5-chloro-2-ethyl-4-isothiazolin-3-one was purchased from Nalco AB (Sweden); the material is used to suppress the bacterial growth in the fiber suspension.
The following materials were employed in the carboxymethylation pre-treatment process.
The below-listed materials were used in the pretreatment process of NFC CMC.
Carboxymethyl cellulose (Aquasorb A500), with a degree of substitution of 0.4 and a molecular weight of 1000 kg/mole, was obtained from Ashland Inc. (USA).Calcium chloride (CaCl 2 , ACS reagent) was purchased from Sigma Aldrich.
All chemicals were used as received, and deionized water was used throughout the studies.

Production of NFC systems
Enzymatically pre-treated nanofibrillated cellulose (NFCEnz) The manufacturing process is similar to that reported by (Naderi et al. 2015b).In brief, the (never-dried) pulp is refined prior to the enzymatic pre-treatment1 as a means to increase the accessibility of the cell wall to the enzyme.Thereafter, after a second refining and the addition of the microbiocide2 , the pulp suspension ( 2% (w/w)) is passed one time through a microfluidizer (M-110EH, Microfluidics Corp., USA) at 1700 bar through two Z-shaped chambers with diameters of 200 m and 100 m, respectively.
It is noted that conductometric measurements as described by Katz et al. (1984) on the pre-treated pulp (in its sodium counter-ion form) revealed a charge density of about 30 eq/g.

Carboxymethylated nanofibrillated cellulose (NFCCarb)
The never dried fibers were dispersed in water at 10000 revolutions using an ordinary laboratory pulper.This was conducted in smaller batches of 30 grams of fibers in two liters of water.The fibers were then exchanged to ethanol by washing the fibers in one liter of ethanol four times with a filtering step in between.
The fibers (110 grams) were then impregnated for 30 min with a solution of 10 g of monochloroacetic acid in 500 ml of isopropanol.The fibers were then added in portions to a solution of 16.2 g of NaOH in 500 ml methanol and mixed with two litres of isopropanol that had been heated just below its boiling temperature in a five-litre reaction vessel fitted with a condenser.The carboxymethylation reaction was allowed to continue for one hour.
Following the carboxymethylation step, the fibers were filtrated and washed in three steps.First, the fibers were washed with 20 l of water.Thereafter, the fibers were washed with two liters of acetic acid (0.1 M) and finally with 10 l of water.The fibers were then impregnated with two liters of NaHCO 3 -solution (4% w/w solution) for 60 min in order to convert the carboxyl groups to their sodium form.Then, the fibers were washed with 15 l of water and drained on a Büchner funnel.
The total charge of the pulp (and hence the resulting NFC), in its sodium counterion form, was determined by means of conductometric titration to be ca 590 μeq/g (degree of substitution (D.S.)  0.1).
The pre-treated fibers ( 2% (w/w)) in their sodium form were microfluidized once, by using the same settings as when NFC Enz was produced.

Carboxymethyl cellulose grafted nanofibrillated cellulose (NFCCMC)
A detailed description of the manufacturing process of NFC CMC is reported in (Naderi et al. 2015c).In brief, a 0.5% (w/w) stock solution of CMC was added at a dosage of 80 mg CMC/g pulp and stirred for one hour, upon which CaCl 2 was added.The pulp and CaCl 2 concentrations were 2.5% (w/w) and 50 mM respectively.The mixture was then heated in an oil bath for two hours at 120C, after which the excess chemicals were removed by washing the pulp with water as described above.The pulp was then ion-exchanged to its hydrogen counterion form (with acetic acid), and washed.Conductometric measurements on the modified pulp (in its sodium counter-ion form) revealed a charge density of about 160 eq/g.The CMC-modified pulp was exchanged to its sodium counter-ion form, and was subsequently delaminated ( 2% (w/w)) in the same manner as NFC Enz and NFC Carb .

Rheological investigations
The rheological studies were conducted on samples that had been stored in a fridge ( 5C) for at least three days after their manufacture, and then equilibrated overnight at room temperature.
The investigations were performed using a Kinexus stress-controlled rotational rheometer (Malvern Instruments, UK) together with the software rSpace (Malvern Instruments, UK).Two standardized (ISO 3219/DIN 53019) metal concentric cylinder (bob and cup) geometries with serrated surfaces were used in the studies.The height and distance between the serrations were 300 µm and 1000 µm, respectively.The diameters of the two bobs were 25 and 23 mm respectively; the length of the bobs was 37.5 mm.The diameter and wall height of the cup were 27.5 and 62.5 mm, respectively.A metal concentric cylinder geometry with smooth surfaces was also employed in the studies.The diameter and length of the bob were 25 and 37.5 mm, respectively; the diameter and height of the smooth cup were the same as the serrated equivalent.
Serrated bob and cups are labelled B23 and B25, where the numbers relate to the cylinder diameter; the smooth bob and cups are labelled as BS followed by the diameter of the cylinder.A working gap of 9.15 mm was employed in all measurements (irrespective of the nature of the smoothness of the geometries).
The NFC samples were sheared at 100 s -1 for one minute in the measuring chamber, as a mean to even out the heterogeneities, and were then left to equilibrate for two minutes before conducting the measurements.
The experimental temperature was set, if not stated otherwise, to 25C.The samples were covered during the studies with a protective hood to minimize their drying.
Each measurement was repeated at least two times; the experimental error was in the order of 10%.

Identification of slip effects
Slip effects in highly viscous systems viz.concentrated NFC can be easily detected by performing experiments with smooth and serrated geometries, as the viscous properties that are obtained by the former geometries are (as will be shown) lower than those obtained by using serrated geometries However, even though slip effects can be diminished by roughening the geometries (Buscall 2010), their absence cannot automatically be presumed.In this communication a method proposed by Kelessidis et al. (2010) has been employed to point to the eventual presence of slip effects in the study of different NFC systems, under different conditions.The proposed method assumes that the rheological property (viscosity) of a system that is marred with slip effects changes with the alteration of the distance between the shearing surfaces.In this communication, the width of the shearing zone was changed by using two (serrated) bobs: B23 and B25.

Impact of temperature on the rheological properties
The impact of temperature was investigated with a serrated bob and cup geometry (B25).
After the insertion of the NFC in the measuring chamber, the samples were allowed to reach the set temperature (5, 25 or 50C).The samples were thereafter equilibrated for another two minutes before pre-shearing and the start of the measurements.

Impact of ionic strength on the rheological properties
The impact of ionic strength was investigated with a serrated bob and cup geometry (B25).Sodium chloride (a monovalent salt) was chosen to alter the ionic strength of the different NFC systems.The studies were conducted in the range of 0-10 mM NaCl; the samples were prepared by a method detailed in an earlier contribution (Naderi, Lindström 2014).In brief, the  2% (w/w) NFC suspensions were diluted with appropriate amounts of deionized water and 100 mM NaCl solutions to a total dry content of 1%  0.1 (w/w).The suspensions were thereafter microfluidized at 300 bar, by employing Z-

Thixotropic properties
The samples were pre-sheared at 100 s -1 for one minute to even out the heterogeneities.The viscosity was subsequently recorded at ̇ = 0.1 s -1 over 20 min.Thereafter, the samples were sheared during 30 s at ̇ = 1000 s -1 , upon which the viscosity was measured again at ̇ = 0.1 s -1 , over 10 min.It is noted that the viscosity was sampled every second during the measurements.It is further noted that ̇ = 1000 s -1 lies within the range for many industrial applications (Metzger 2002).

Results and discussion
Figs 1a-c display the visual appearances of NFC CMC , NFC Enz and NFC Carb .The NFC CMC and NFC Enz systems are characterized by an opaque appearance, indicating systems with a broad size distribution.NFC Enz displays a coherent gel, but with a "rough" surface.However, NFC CMC is characterized by a watery appearance; also, the fibrous flocs do not seem to make a coherent 3D structure.On the other hand, NFC Carb is characterized by a more transparent and stiff gel structure.As evident, the shear viscosity at low shear rates (γ̇  5 s -1 ) measured by the smooth geometry is lower than the viscosity measured by a serrated geometry.Furthermore, the different viscosity curves do not superimpose until high shear rates (γ̇  50 s -1 ) have been reached.The kink in the viscosity curve is a common feature for many NFC systems (see e.g.Lasseuguette et al. 2008;Iotti et al. 2011;Karppinen et al. 2012;Naderi et al. 2015a).However, there are conflicting explanations for the origin of the phenomenon.For example Iotti et al. (2011) proposed the organization of the nanofibrils into "structures" to be a cause behind the observed kink in the flow curve of a low-charged NFC system.A similar explanation was also provided by Lasseuguette et al. (2008) The combined observations indicate that rheological measurements on stable NFC systems are little affected by slip effects, when the measurements are conducted with serrated geometries.However, the impact of wall slip effects increases with the increasing instability of the system, as is evident in Figure 2  The rheological properties of NFC Enz as a function of salt content are summarized in Fig 5 ; Table 1 contains the stiffness (G'/G'') of the different systems.As expected, NFC Enz , having a low amount of charges, is little affected by the ambient ionic strength.However, the observations contrast those of Agoda-Tandjawa et al. ( 2010), who reported significant increases in viscosity and G' when a (presumably) low-charged fibrillated NFC based on sugar-beet pulp was studied; the pulp contained about 1.2% (w/w) galacturonic acid.The reasons behind the discrepancy between the results of the authors and those reported in this communication are not clear.However, it is interesting to note that Agoda-Tandjawa et al. employed a plate-plate geometry with Teflon-covered surfaces (presumably with no serrations).In this context, the importance of choosing the proper geometry can also be highlighted, especially when dealing with low-charged NFC.For example, Naderi and Lindström (2015) showed that the water that is released upon the insertion of a NFC Enz sample in a plate-plate geometry leads to the overestimation of the rheological properties, and erroneous results.This process could, however, be overcome by employing a bob and cup geometry.
Fig 6 displays the rheological properties of NFC CMC ; the gel stiffness of the system as a function of NaCl concentration is summarized in Table 1.The results in Fig 6 imply some change in the magnitude of the rheological properties when the ionic strength is increased; this is most clearly seen in Fig 6b .The impact of the ionic strength is, however, more visible when the gel stiffness of the system (Table 1) is studied, revealing increasing stiffening when the ionic strength of the system is increased.The stiffening effect might imply a more intimate contact between the nanofibrils, but the exact mechanism behind this notion is not clear.It could be hypothesized that screening of the system charges allows for a more intimate fibril-fibril contact.However, it could also be hypothesized that increasing the ionic strength of the system further destabilizes the already unstable NFC CMC system (see Fig 1d).Subsequently, water is released from the network, leading to the formation of a stronger network.(Naderi and Lindström 2014).The storage and loss moduli for NFCEnz and NFCCMC were measured at  = 0.05%.The G' and G'' of NFCCarb were measured at 0.5 Hz. ), as a function of ionic strength of the system; B25 was employed.a) Storage (G') and loss modulus (G'').b) Shear viscosity.

G'/G''
The low impact of ionic strength on the magnitude of the rheological properties (G' and viscosity) of NFC CMC is puzzling, considering the higher amount of charge in the system ( 160 eq/g) as compared to the charge density of NFC Enz ( 30 eq/g).The reasons behind this phenomenon are not understood by the authors.It may be that the amount of charge in the system is too low to have a major impact on the properties of the system.In the literature, the bulk of investigations have been conducted on highly charged NFC systems.For example, major changes in the rheology of TEMPO-based NFCs (< 2000 eq/g) with the addition of salts have been reported by several groups (Jowkarderis, van de Ven 2014;Tanaka et al. 2014).The impact of salt on the rheology of a NFC Carb system (similar to the system of this report,  590 eq/g) has also recently been reported by the authors (Naderi, Lindström 2014); significant lowering in the viscosity and storage modulus of the system were observed already at 1 mM NaCl addition.
The stiffness of NFC Carb systems as a function of ionic strength, reproduced from (Naderi, Lindström 2014), is summarized in Table 1.Interestingly, gel stiffness decreases with increasing ionic strength; it is recapitulated that the viscosity of the system also reportedly decreased with increasing ionic strength of the system (Naderi and Lindström 2014).The observations can be explained by the screening of the surface charges, which gives the system a more fluid-like character when water is released from the fibril-network.
The impact of temperature on the rheological properties of the different NFCs is displayed in Fig 7 ; the gel stiffness of the systems is summarized in Table 2.The results (Figs 7a-b, Table 2) indicate that the viscosity and G' decrease and the gel stiffness of NFC Enz increases with increasing temperature.Similar effects are also observed for NFC CMC .However, the impact of temperature on the magnitude of the rheological properties of NFC Carb is lower (Fig 7e-f); even though a significant increase in gel stiffness is still observed (Table 2).
Oscillatory measurements conducted in the linear viscoelastic regime, on NFC systems provide insight about the properties of intact systems.On the other hand, shear rate measurement provide information about flowing systems, where the networks are in the process of breaking.Hence, conflicting trends in the properties might occur when oscillatory and shear rate results are analysed.
The impact of temperature on the viscosity of nanofibrillated systems has already been studied by Herrick et al. (1983), who found that increasing the temperature (T ≤ 80C) leads to lowering of the shear viscosity.A similar finding has also been reported by Iotti et al. (2011).
The thixotropic behaviour of the different nanofibrillated systems is presented in Figures 8a-b.The viscosity of NFC Carb is rapidly restored, and most of the G' is regained quickly when the high shearing is stopped.The observations can be rationalized by using earlier discussions and considering the visual appearances of the systems.Carboxymethylated NFC is a highly fibrillated system with a strong networking ability.Hence, the viscosity of the system is rapidly regained due to the rapid networking of the system constituents.The storage modulus of the system is, however, not fully restored during the experimental time.It is postulated that prolonged times are needed for the NFC network to reach optimum strength, as noted by among others Rudraraju and Wyandt (2005).
The trends of the viscosity and G' curves of NFC Enz might at first glance be challenging to rationalize, as they indicate that stronger networks are built after severe shearing of the system.However, similar observations on an NFC Enz system were recently reported by the authors (Naderi et al. 2015b), in which the observations were rationalized by a more homogeneous distribution of the NFC flocs when the system was severely sheared.The severe shearing was proposed to break the larger flocs into smaller ones, which then interact more strongly with each other, due to the larger surface area of the smaller entities.This hypothesis finds support in Fig 8c, in which the NFC Enz system has been pre-sheared for 10 min at 1000 s -1 .Evidently, the G' of the system after the severe shearing process is close to the G'-value that was recorded before the severe shearing process.The thixotropic behaviour of NFC CMC has elements of both NFC Carb and NFC Enz .The viscosity of the system can be rationalized in a similar manner as to NFC Enz .The lower G'-value of NFC CMC , after the severe shearing process, is postulated to be caused by the lower networking ability of the system compared to NFC Enz , due to the lower fibrillated state of the former system (see

Concluding remarks
A wide range of dispersed systems are prone to slip effects; many nanofibrillated cellulose systems belong to this category of materials.The occurrence of slip effects in NFC systems is dependent on the inherent stability of the systems, as was shown in this communication.The highly charged and highly fibrillated NFC Carb was found not to be significantly affected by slip effects, but the unstable NFC Enz and NFC CMC displayed wall depletion effects.Interestingly, the magnitude of the slip effects in the latter systems, recorded by employing the methodology of Kelessidis et al. (2010), were found to be small.This is postulated to be due to the efficient mixing of the systems (Kelessidis et al. 2007;Bonn and Denn 2009) by the pre-shearing process of the measuring protocol.The low magnitude of the slip effects was used to draw the following conclusions on the impact of ionic strength and temperature on the rheological properties of the different NFC systems, and the thixotropic properties of the same.
It was shown that the rheological properties of NFC Enz are little affected by the ionic strength of the system (at least in the investigated range).However, an increasing gel stiffening effect was observed for NFC CMC with increasing ionic strength of the system; NFC Carb displayed, on the other hand, an opposite trend with regard to the effect of the ionic strength of the system.Gel stiffening effects as a function of increasing ambient temperature were observed for all systems.A lowering viscosity with increasing temperature was also observed for NFC Enz and NFC CMC ; however, the viscosity of NFC Carb was little affected by temperature.
Finally, it was found that all the investigated nanofibrillated systems could rapidly regain their original rheological properties, once their severe shearing is stopped.
The major conclusion of this work is, however, that different NFC systems can have completely different rheological properties; the awareness of this realization is important for the implementation of NFCs in the industry.
Figs 1a-c display the visual appearances of NFC CMC , NFC Enz and NFC Carb .The NFC CMC and NFC Enz systems are characterized by an opaque appearance, indicating systems with a broad size distribution.NFC Enz displays a coherent gel, but with a "rough" surface.However, NFC CMC is characterized by a watery appearance; also, the fibrous flocs do not seem to make a coherent 3D structure.On the other hand, NFC Carb is characterized by a more transparent and stiff gel structure.Fig 1d displays the state of the fibrillated systems recorded after one week.The carboxymethylated NFC does not show any phase-separation, although phaseseparation tendencies are observed for NFC Enz and NFC CMC .These are most visibly seen for NFC CMC , where a water layer on the top of the container can be discerned.The shear viscosity of NFC Enz measured by serrated and smooth configurations (B25 and BS25) is displayed in Fig 2a.As evident, the shear viscosity at low shear rates (γ̇  5 s -1 ) measured by the smooth geometry is lower than the viscosity measured by a serrated geometry.Furthermore, the different viscosity curves do not superimpose until high shear rates (γ̇  50 s -1 ) have been reached.The kink in the viscosity curve is a common feature for many NFC systems (see e.g.Lasseuguette et al. 2008;Iotti et al. 2011;Karppinen et al. 2012;Naderi et al. 2015a).However, there are conflicting explanations for the origin of the phenomenon.For exampleIotti et al. (2011) proposed the organization of the nanofibrils into "structures" to be a cause behind the observed kink in the flow curve of a low-charged NFC system.A similar explanation was also provided byLasseuguette et al. (2008), who investigated the rheological properties of a TEMPO-based NFC.On the other hand,Karppinen et al. (2012), who combined rheological and imaging techniques (to study a presumably low-charged NFC), have argued that the kink in the flow curve is caused by the increase in the floc size and the broadening of the floc size distribution.
Figs 1a-c display the visual appearances of NFC CMC , NFC Enz and NFC Carb .The NFC CMC and NFC Enz systems are characterized by an opaque appearance, indicating systems with a broad size distribution.NFC Enz displays a coherent gel, but with a "rough" surface.However, NFC CMC is characterized by a watery appearance; also, the fibrous flocs do not seem to make a coherent 3D structure.On the other hand, NFC Carb is characterized by a more transparent and stiff gel structure.Fig 1d displays the state of the fibrillated systems recorded after one week.The carboxymethylated NFC does not show any phase-separation, although phaseseparation tendencies are observed for NFC Enz and NFC CMC .These are most visibly seen for NFC CMC , where a water layer on the top of the container can be discerned.The shear viscosity of NFC Enz measured by serrated and smooth configurations (B25 and BS25) is displayed in Fig 2a.As evident, the shear viscosity at low shear rates (γ̇  5 s -1 ) measured by the smooth geometry is lower than the viscosity measured by a serrated geometry.Furthermore, the different viscosity curves do not superimpose until high shear rates (γ̇  50 s -1 ) have been reached.The kink in the viscosity curve is a common feature for many NFC systems (see e.g.Lasseuguette et al. 2008;Iotti et al. 2011;Karppinen et al. 2012;Naderi et al. 2015a).However, there are conflicting explanations for the origin of the phenomenon.For exampleIotti et al. (2011) proposed the organization of the nanofibrils into "structures" to be a cause behind the observed kink in the flow curve of a low-charged NFC system.A similar explanation was also provided byLasseuguette et al. (2008), who investigated the rheological properties of a TEMPO-based NFC.On the other hand,Karppinen et al. (2012), who combined rheological and imaging techniques (to study a presumably low-charged NFC), have argued that the kink in the flow curve is caused by the increase in the floc size and the broadening of the floc size distribution.
and Figure 4; the latter displays the rheological properties of NFC CMC .In this context, attention is drawn to the highly differing trends in the flow curves of NFC CMC (Fig 4) that were obtained by B23 and B25 compared to what was observed for NFC Enz , which showed similar but shifted trends.The observations in Fig 4 correlate well with the visual appearance of NFC CMC in Fig 1.

Fig 1 -
Fig 1 -The visual appearance of 2%  0.1 (w/w) suspensions of NFCCMC (a), NFCEnz (b) and NFCCarb (c).d) The state of the different NFC systems recorded after one week.From left to right: NFCCMC, NFCEnz and NFCCarb.
Fig 2 -Shear viscosity of NFCEnz measured by different configurations.a) Serrated (square) and smooth (star) bob and cup geometries in deionized water at 25C; B25 and BS25 were employed in the studies.Shear viscosities of NFCEnz measured at 25C b), 50C c) and in 10 mM NaCl d).The measurements (b-d) were conducted by two different serrated bobs: B25 (filled symbol) and B23 (unfilled symbol).

)
Fig -4 Shear viscosity of NFCCMC measured by different configurations.a) Serrated (square) and smooth (star) bob and cup geometries; B25 and BS25 were employed in the measurements.Shear viscosities of NFCCMC measured at 25C (b), 50C (c) and in 10 mM NaCl (d).The measurements (b-d) were conducted by two different serrated bobs: B25 (filled symbol) and B23 (unfilled symbol).

Figure - 7
Figure -7 Rheological properties of NFCEnz (a,b),NFCCMC (c,d) and NFCCarb (e,f) as a function of temperature.The dry content of the different systems is 2.0  0.1% (w/w).B25 was employed in the studies.

Table 1 -
Gel stiffness (G'/G'') of different NFC system (measured at 25C).The values for NFCCarb have been reproduced from