Effect of polymer-polymer interactions on the flow behavior of some polysaccharide-based hydrogel blends

Carbohydrate Polymers 287 (2022) 119352

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

Effect of polymer-polymer interactions on the flow behavior of some
polysaccharide-based hydrogel blends
Tilen Kopaˇc , Aleˇs Ruˇcigaj , Matjaˇz Krajnc *
University of Ljubljana, Faculty of Chemistry and Chemical Technology, Veˇcna pot 113, SI-1000 Ljubljana, Slovenia

A R T I C L E I N F O

A B S T R A C T

Keywords:
Zero-shear viscosity
Yield stress
Shear thinning
Hydrogel interactions
Blending effects
Mathematical modeling

This study is the continuation of our previous work (Kopaˇc, Abrami, et al., 2021) where the theoretical approach
of polymer-polymer interaction to predict the crosslink density of hydrogels was introduced. This theory is
further extended to the flow properties of hydrogels that allow the analysis of synergistic effect in hydrogel
systems and the understanding of possible anomalous behavior of certain mixtures. Various hydrogel structures
were prepared accordingly by blending scleroglucan, anionic nanocellulose, Laponite dispersions and alginate
solution. The relationship between mechanical and flow properties of the hydrogel network was carefully studied
and eventually described by mathematical model. The linear model equation to predict yield stress of hydrogels

in relation to the crosslink density was designed showing a satisfactory agreement between experimental data
and model predictions. The correlation was adjusted by defining a proportionality coefficient, representing the
energy defined per moles of crosslinks that can be used to restore the deformation.

1. Introduction
Rheology plays an essential part in the field of scientific research as
well as in industrial development (Boger, 2009), particularly in the case
of complex systems such as hydrogels whose deformation behavior and
flow properties cannot be described by the classical approaches
belonging to continuum mechanics (the elasticity theory and fluid dy­
namics) (Lei, Li, Xu, & Liu, 2021). The rheological characterization
represents an important tool also to investigate hydrogels' structural
characteristics (Cao, Duan, Zhang, Cao, & Zhang, 2021) in order to
evaluate their potential use and performance in various biomedical
´ & Khutoryanskiy, 2015). Particularly in the case of
applications (Calo
carbohydrate polymer systems, which present several advantages over
the synthetic polymers (Farrukh, Mustafa, Hussain, & Ayoub, 2018), the
rheology enables a detailed investigation of the mechanical (crosslink
density ρx and average mesh size) and flow properties (zero shear vis­
cosity η0 and yield stress τ0) of hydrogels (Kopaˇc, Abrami, Grassi,
Ruˇcigaj, & Krajnc, 2021; Kopaˇc, Krajnc, & Ruˇcigaj, 2021; Kopaˇc,
ˇ
Ruˇcigaj, & Krajnc, 2020; Lapasin, Abrami, Grassi, & Sebenik,
2017;
ˇ
ˇ
Sebenik,
Krajnc, Aliˇc, & Lapasin, 2018; Sebenik,
Lapasin, & Krajnc,

2020).
So far, theoretical approaches, that can satisfactorily predict the flow
behavior of complex hydrogels, are undersupplied in literature, except

for very dilute fluids. In our previous work (Kopaˇc, Abrami, et al., 2021)
a mathematical model that can satisfactory predict the ρx of hydrogel
network on the basis of the number of crosslinks formed due to the
polymer-polymer interactions (ionic, hydrogen, electrostatic, and other
covalent interactions) was developed and validated for polysaccharide
hydrogels. Motivated by such ρx prediction, we hypothesize that equa­
tion can be leveled up to predict the η0 and τ0 of similar hydrogel sys­
ˇ
tems. The literature (Hoti et al., 2021; Lapasin et al., 2017; Sebenik
et al., 2020; Zhao, Mayumi, Creton, & Narita, 2017) reports the effect of

ρx on flow behavior of gels based on rheological rotational tests where
for the most part, η0 and τ0 increase with the ρx of gels. The detailed
study of correlations between ρx and η0 or τ0 can be the starting point for

further mathematical modeling. For this purpose, the comprehensive
rheological characterization of (complex) hydrogel systems based on
rotational tests is needed to be performed and studied in detail. To prove
the high applicability of a new mathematical approach for analyzing the
τ0 and η0 data, the hydrogel systems based on TEMPO (2,2,6,6-tetra­
methylpiperidine-1-oxyl) oxidized nanocellulose (TOCNF), sodium
alginate (ALG), scleroglucan (SCLG) and Laponite (LAP) were prepared.
The baseline of the model development is associated to the under­
standing of polymer-polymer theory in the hydrogel network starting
with the careful study of polymers' chemical structures and their
possible interactions. To fulfil the above orientations, the brief basic

* Corresponding author.
E-mail address: (M. Krajnc).
/>Received 11 January 2022; Received in revised form 3 March 2022; Accepted 9 March 2022
Available online 15 March 2022
0144-8617/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( />

T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

structural description of TOCNF, ALG, SCLG and LAP is needed. ALG is a
linear copolymer composed of (1 → 4)-linked α-L-guluronic (G) and β-Dmannuronic (M) residues of varying sequence. Anionic carboxyl groups
give the hydrogel pH responsivity (Way, Hsu, Shanmuganathan, Weder,
& Rowan, 2012) in an alkaline release environment and also allowing
˜ o-Maso
´
control of ρx (Lin, Bruzzese, & Dufresne, 2012), porosity (Patin
et al., 2019) and flow properties (Kopaˇc et al., 2020) by crosslinking via
divalent ions (e.g., Ca2+) (Curvello & Garnier, 2020; Hecht & Srebnik,
2016; Xu et al., 2019).
TOCNF is linear polysaccharide, anionic constituent biopolymer. The
physical and chemical properties of cellulose fibers (Lundahl, Berta,
Ago, Stading, & Rojas, 2018) and cellulose-based materials (John &
Thomas, 2008; Liang, Bhagia, Li, Huang, & Ragauskas, 2020) are mostly
influenced by the formation of intra- and intermolecular hydrogen
bonding (Lin & Dufresne, 2014) which are allowed by the presence of
three hydroxyl groups on each β-(1–4)-glucopyranosyl unit (Klemm
et al., 2011; Liang et al., 2020). Significant amounts of C6 carboxylate
groups are selectively formed on each cellulose nanofibril surface by

TEMPO-mediated oxidation that gives the anionic character to
biopolymer (Isogai, Saito, & Fukuzumi, 2011) and the possibility of
ionic crosslinking (Kopaˇc, Krajnc, & Ruˇcigaj, 2021).
The nonionic biopolymer SCLG is used as a thickener and suspending
agent (Lapasin et al., 2017; Paolicelli et al., 2017). The β-D-glucopyr­
anosyl units (1,3) linked in the main chain create the internal structure
in addition to the three triplex chains which hold the hydrogen bonds
together in the center of the triplex. Such a conformation is character­
ized by high stiffness, which is reflected in the special properties of
aqueous solutions of SCLG in a wide pH range and even at relatively high
temperatures (Lapasin et al., 2017). Intra- and intermolecular hydrogen
bonds between the main SCLG chains lead to the formation of stable
rigid triple helices in aqueous solutions (Nazmabadi, Shirdast, Sharif, &
Aalaie, 2020). Furthermore, a single β-D-glucopyranosyl bound to the
unit (1,6) on the main chain prevents polymer precipitation and inter­
molecular aggregation (Coviello et al., 2005; Lapasin et al., 2017).
The synthetic hectorite LAP single layer nanoparticles in a form of
rigid, disk-shape crystals were used as a rheological modifier. LAP
hydrogels own a negative charge of faces as a result of Na+ ions disso­
ciation. Besides, the edges of LAP disks are slightly positively charged
due to the protonation of hydroxyl groups (Park, Dawson, Oreffo, Kim, &
ˇ
Hong, 2020; Sebenik
et al., 2020). The face-edge attractions cause a
gelification process, when the system does not contain salt, and depend
on pH, resulting in a gel network also known as a “house of cards”
structure (Becher et al., 2019). Therefore, the LAP addition to weak
polysaccharide-based system could contribute to faster development of
final gel properties (aging), improve the mechanical properties and
ˇ

shear thinning character of the matrix (Sebenik
et al., 2020). At last, the
more detailed analysis of the structural formula of TOCNF (Isogai et al.,
2011; Liang et al., 2020), ALG (Homayouni, Ehsani, Azizi, Yarmand, &
Razavi, 2007), SCLG (Coviello et al., 2005), and LAP (Ruzicka & Zac­
carelli, 2011) can be found elsewhere.
In this study, neutral (SCLG) and anionic (TOCNF and ALG)
biopolymer-based hydrogels in addition to LAP were prepared in various
concentrations and blended in different weight ratios with the addition
of assorted crosslinking agent concentrations to form complex hydrogel
systems (i.e., ionically crosslinked blends of polysaccharide polymers in
addition to the layered inert component). The effect of polymer and
crosslinking agent concentration, respectively, and the polymer fraction
dependence in the preparation of hydrogels on flow behavior of the
complex hydrogel systems was analyzed. Prepared complex hydrogel
systems were subjected to comprehensive rheological rotational tests in
order to broaden the knowledge of the rheological characteristics of
different biopolymer-based systems. The study of hydrogel samples
provided the explanation of polymer-polymer interaction effects on the
flow behavior offering the basis for determination of crucial parameters,
such as η0 and τ0. As a result, the theoretical outlines of the model
development to predict the flow properties were provided following the

polymer-polymer interaction theory correlated to experimental insights.
As an upgrade to a wide range of experimental data, a new mathematical
approach to analyze η0 and τ0 is eventually given, which can be equiv­
alently used to describe the synergistic effect on shear behavior owing to
their interrelationship.
2. Experimental
2.1. Materials

The TEMPO oxidized nanofibrillated cellulose, a linear poly­
saccharide composed of cellobiose repeating units linked by β-(1–4)
glycosidic bonds, was purchased in freeze-dried powder from The Pro­
cess Development Centre, University of Maine (UMaine PDC), USA. The
material specifications given by the producer are: (i) chemical formula
of [(C6H10O5)x(C6H9O4COONa)y], (ii) powder density of 1.5 g/cm3, (iii)
aqueous gel density of 1.0 g/cm3 (1–3 wt%), (iv) carboxylate level is
0.2–2 mmol/g solids, (v) pKa of substituent group is 3.9, and (vi) the
fiber dimensions are classified between 20 and 50 nm in width and
lengths of up to several microns.
Scleroglucan Actigum CS 11 was purchased from Cargill France SAS,
France. The supplier provides 99%min purity, and the average molar
mass is 1.2⋅106 g/mol.
Laponite XLG, synthetic layered silicate with a low heavy metals
content was provided by BYK-Chemie GmbH, Germany. The supplier
specifications provide: (i) disk-shaped crystals with an average diameter
of 30 nm and a thickness of 1 nm, (ii) insolubility in water, (iii) hydrates
and swells give clear and colorless colloidal dispersions in water, and
(iv) obtains a high surface area (BET) of 800 m2/g.
Sodium alginate, a polysaccharide made up of α-L-guluronic acid (G)
content and (1–4)-linked β-d-mannuronic acid (M) monomers, were
supplied by Sigma-Aldrich (St. Louis, USA). The molar mass of
2⋅104–6⋅104 g/mol and the high (70%) of G content is provided by
supplier.
Calcium chloride (CaCl2) was as received from Merck (Darmstadt,
Germany) and sodium azide (NaN3, an assay of ≥99.5%) was supplied
by Sigma-Aldrich (St. Louis, USA).
2.2. Sample preparation
The gels were prepared by slowly added dry TOCNF, ALG, SCLG or
LAP in demineralized water with 0.02 wt% of NaN3 (to prevent micro­

biological contamination) at constant stirring (500 rpm) by using a fourblade propeller stirrer (pH value of sample was 7). After 2 h, the
dispersion was placed in an ultrasonic bath for 1 h (to achieve more
homogeneous dispersion - ALG in water form clear solution and so was
excluded from this process). The dispersions were stirred overnight at
750 rpm at room temperature.
The complex hydrogel systems were prepared by blending single
biopolymer dispersion TOCNF or SCLG and LAP or ALG in different
weight ratios. For homogenization, the same procedure was used as
previously reported. The samples were then quantitatively transferred in
a 3D printed mold to sample achieve the same characteristics as the
rheometer measuring plate (a diameter of 50 mm and thickness of 1
mm).
The crosslinking process including spraying the aqueous solution of
calcium ions onto the surface of samples in the mold. The mold was
designed to have two edges: the first and second corresponding to the
equal volume of biopolymer sample and crosslinking agent solution,
respectively. The crosslinked systems were maintained in a refrigerator
at 4.2 ◦ C for 48 h to establish the final hydrogel structure (time required
for crosslinking process based on Ca2+ ions).
Hydrogels were prepared in 1–3 wt% concentration of a single
polymer since the most important changes in the structure of the
hydrogel network occur in this range. Furthermore, at polymer con­
centration up to 1 wt%, the systems are almost Newtonian and their
2


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

viscoelastic properties are practically undetectable (Lapasin et al., 2017;
ˇ
Sebenik
et al., 2020). Ionic crosslinking was performed by adding a
calcium ion solution to a maximum concentration of 2 wt%. In this state,
almost all carboxyl groups of TOCNF or ALG participate in the cross­
linking process so that a higher concentration of crosslinking agent so­
lution would not (noticeably) affect the rheological properties of the
hydrogels (Kopaˇc et al., 2020; Kopaˇc, Krajnc, & Ruˇcigaj, 2021). The
sample pH value was 7. The prepared hydrogel systems were kept in the
fridge at a temperature of 4.2 ◦ C for 15 days to establish the final
hydrogel structure and to avoid the effect of aging which mainly affect
the rheological properties due to the formation of hydrogen bonds (ac­
ˇ
cording to the literature (Lapasin et al., 2017; Sebenik
et al., 2018)).

and SCLG) and LAP (clay) exhibits pronounced shear thinning character
(Fig. 1). The η0 increases from TOCNF, over SCLG to LAP dispersions.
Especially in the case of SCLG and LAP, the behavior is plastic with a
significant viscosity drop in a narrow stress interval, while TOCNF
hydrogels present a longer stress interval of viscosity decrease which
appears at lower τ0. The 1st Newtonian plateau of TOCNF (η0) is the
consequence of the existence of invisibly small clusters formed during
aging which are enabled by entanglements of nanofibers and non­
covalent and covalent interactions between them (not only hydrogen
bonding, also van der Waals interactions and electrostatic repulsion due
ˇ
to negatively charged carboxyl groups (Sebenik
et al., 2018)). The

clusters arrangement of the TOCNF internal structure in the shear flow
direction is therefore shown in a longer stress interval of viscosity
decrease (effect of clusters concentration, shape, and distribution).
Accordingly, the TOCNF has the most pronounced 2nd Newtonian
plateau and the network-of-forces of the TOCNF internal structure are
weaker than in the case of stable rigid triple helices in SCLG structure
and LAP crystals (η0 values). As opposed to η0, the τ0 occurs in the lower
shear stress in the LAP in comparison with SCLG. The reason may be
attributed to different interactions that predominantly form a structure
in both samples. The intra- and inter-molecular hydrogen bonds among
the main chains of SCLG lead to the formation of a stable structure until
the increased shear stress predominates over hydrogen interactions in
the SCLG structure (τ0). On the other hand, the LAP forms rigid, diskshaped crystals whose random distribution determines the internal
structure (the crystals are folded and occupy the most energy-efficient
state). The viscosity is constant until the crystals start to steer in the
direction of shear (τ0). Fig. 1 shows that the SCLG structure based on
hydrogen interactions between polymer-polymer chains and LAP diskshaped crystals structure (predominantly electrostatic interactions) are
more solid than the TOCNF dispersion. The internal structure in TOCNF
is determined by intra- and intermolecular hydrogen bonding (Lin &
Dufresne, 2014), which are allowed by the presence of three hydroxyl
groups on each β-(1–4)-glucopyranosyl unit. The crosslink density of
TOCNF is significantly lower than in the case of SCLG, due to lower
hydroxyl group content which results in lower number of polymerpolymer hydrogen interactions, which also results in lower η0 and τ0.
Differently, the ALG is soluble in water and does not form gels structure
in the absence of crosslinking agent and therefore exhibits fluid-like
viscoelastic behavior (Fig. 1).
The values of η0 and τ0 are experimentally difficult to determine
uniformly, as they generally depend on the type and experiment per­
formance rate. The η0 values can be determined by extrapolating the
experimental points of the flow curve on the log viscosity–log shear

stress (or shear rate) diagram. The measurements in Fig. 1 show a flow
behavior where even at very low shear stresses the samples “creep” and
become plastically deformed. This is the main reason for the difficult
determination of η0 and τ0. All samples with plastic flow behavior
(SCLG, LAP and TOCNF) are characterized by a very high viscosity due
to a strong internal structure (based on polymer-polymer interactions)
that resists the shear. The η0 values for 2% SCLG, TOCNF and LAP single
dispersion were determined at 101000, 7327 and 520,600 [Pa⋅s],
respectively (Table S1).
Furthermore, yield points are not material constants as they depend
as well on the measuring method as well as on the analysis method used.
The value of τ0 is often determined as a rough approximation by
extrapolating the experimental points of the flow curve on log shear
stress–log shear rate or log shear strain– log shear stress diagram. The
τ0 determination via the “tangent crossover method” is illustrated in
Fig. 2. The τ0 determination via the best straight fitting line (“tangent”)
in the linear–elastic deformation range is also realizable. Shear stress
ramp experiments from 0.1 to 200 Pa were performed in variable
measuring point duration logarithmic profile (initial point duration was
set at 300 s and final point duration was set at 20 s, along with ramp log
+9.63 points/decade). From Figs. 2A–C, the τ0 values for 2% SCLG,
TOCNF and LAP single dispersion were determined at 32, 1.9 and 22

2.3. Rheological characterization
Rheological rotational tests were performed on rheometer Anton
Paar Physica MCR 301 equipped with a crosshatched plate with a
diameter of 50 mm (PP50/P2). The samples were prepared to corre­
spond to a thickness of 1 ± 0.1 mm between the plates and were
measured at a temperature of 25 ◦ C.
Flow curve – viscosity (η) was measured over a range of shear stresses

(σ ) from 0.01 to 3000 Pa. The flow curves were presented in a log vis­
cosity–log shear stress plot. The experimental data were fitted by the
Roberts-Barnes-Carew (RBC) model. The RBC is the modified original
Ellis equation (Roberts, Barnes, & Carew, 2001):
η0

η = η∞ [1 + (σ /σ 2 )s ] +

1+(σ/σ1 )p

− η∞ [1 + (σ/σ 2 )s ]
( )m
1 + σσc

(1)

where η is dynamic viscosity, η0 and η∞ are the asymptotic values of the
viscosity at zero and infinite shear stress, respectively, σ is shear stress,
σ c is critical shear stress, which locates the transition region between
two moderate shear thinning regions, m describes the rate of viscosity
decrease in the transition region, σ 1 and σ 2 are critical shear stresses in
the composite Ellis model, p and s are adjustable parameters (Roberts
et al., 2001).
3. Results and discussion
3.1. Flow behavior
3.1.1. Single polymer hydrogel and yield stress determination
Single polymer hydrogel based on constituent biopolymers (TOCNF

Fig. 1. Flow behavior of single polymer systems based on LAP (squares), SCLG
(circles), TOCNF (triangles) and ALG (diamonds). The lines represent RBC

model fit.
3


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

increase in the mass fraction of SCLG. The shape of flow curves is typical
for shear thinning behavior of hydrogel blends. All the 2 wt% SCLG/
TOCNF systems exhibit an evident plastic behavior (the 1st Newtonian
plateau is distinctive) with a significant viscosity drop more typical for a
single SCLG hydrogel. Furthermore, the 2nd Newtonian plateau is also
detected, as the main contribution of the TOCNF biopolymer (see the
shape of the flow curve for single TOCNF hydrogel - triangle) to the
SCLG/TOCNF hydrogel blends. It may be concluded that 2 wt% SCLG/
TOCNF systems form pronounced viscoelastic properties of hydrogels,
where the addition of SCLG mainly affects the elastic properties of the
hydrogel (hydrogen interactions contributed to stronger network-offorces of internal structure), while the addition of TOCNF gives the
hydrogel a viscous character at high shear stresses (due to the presence
of the small invisible clusters or flocs which are dispersed in a gel-like
ˇ
matrix (Sebenik
et al., 2018)) (also discussed after in Figs. 7A and B).
Based on the calculated data from Table S1, the zero-shear viscosity
interval is between 9.94⋅103 < η0 < 3.19⋅104 [Pa⋅s], the yield stress
interval is determined in the range of 7.84 < τ0 < 30.7 [Pa].
In Figs. 3C–D, the effect of ALG in the hydrogel blend system is
presented. The ALG in water forms a clear solution which is the main
reason that it does not participate in forming polymer-polymer inter­

action in addition to various biopolymer (SCLG) or rheological modifier
(LAP). In Fig. 3C, it is clear that the flow curves of blends containing LAP
in 1.0–0.25 mass fractions have a pronounced plastic behavior evident
in 1st Newtonian plateau with a significant viscosity drop in a narrow
stress interval (same as for single LAP dispersion presented in Fig. 3C –
square symbols), in addition to the decreasing effect of η0 and τ0. This
may be the consequence of diluting effect of ALG in the 2% LAP/2% ALG
systems (see also discussion of Figs. 7A and B). In addition, the ALG
addition gives the more viscous character of systems in high shear
stresses where the 2nd Newtonian plateau is clearer (Fig. 3C) than in
single LAP dispersions (Fig. 1). At the lower mass fractions of LAP in the
mixture, the plastic behavior is not expressed. According to the litera­
ture (Lapasin et al., 2017), the LAP dispersion of less than 0.75 wt% has
properties of fluid or does not show shear thinning and viscoelastic
character, therefore the blending would have only a diluting effect and
not synergistic interaction between LAP nanoparticles and biopolymer
network. The same conclusions (diluting effect) may be given for 2%
SCLG/2% ALG systems (Fig. 3D) where the flow behavior is a function of
SCLG concentration dependence changed by the addition of ALG solu­
tion. While dilution effect of ALG can be observed in non-crosslinked
samples, the drastic change in ALG participation in hydrogel mixtures
on the flow behavior is presented in ionically crosslinked systems
(Section 3.1.3).

Fig. 2. Yield point determination via the tangent crossover method for 2%
SCLG (A), TOCNF (B), LAP (C) and ALG (D) single systems.

[Pa], respectively. Fig. 2D illustrates obvious fluid behavior of ALG.
Additionally, η0 and τ0 can also be defined as a model parameter
when a rheological model is used to describe plastic flow behavior. Such

generally well-known rheological models are the Bingham model, Her­
schel–Bulkley model and Casson model. Both methods for η0 and
τ0 values determination were used. The results of all samples are sum­
marized in Table S1.
3.1.2. Polymer concentration and fraction dependence hydrogel systems
Fig. 3A presents flow curve of SCLG/TOCNF hydrogels systems. The
samples were prepared by blending various (1–3 wt%) concentration of
SCLG and TOCNF dispersions in equal weight ratios. The concentration
of biopolymer plays a crucial role in determining final rheological
properties of SCLG/TOCNF hydrogels, which can be confirmed by a very
wide interval of yield stresses from 1.28 < τ0 < 43.7 [Pa] and also zeroshear viscosity of 4.15⋅102 < η0 < 1.05⋅105 [Pa⋅s] (more data can be
found in Table S1). Shear thinning character increases with increasing
biopolymer concentration which suggests the existence of invisibly
small clusters with significant swell-ability in water. Clusters are formed
during the process of storage (aging) of prepared hydrogels when pre­
dominantly hydrogen, van der Waals and electrostatic repulsive forces
begin to intertwine in SCLG and TOCNF network structure. In this case,
the clusters increase, which reduces their distance from each other and
further enhances the noncovalent interactions among them. After two
weeks, the growth of the clusters slows down reaching the final prop­
ˇ
erties of the hydrogel structure (Sebenik
et al., 2018). In the case of
TOCNF polymers, having modified carboxyl groups on the surface, the
clusters are also formed during the ionic or covalent crosslinking pro­
cess. The experimental data show higher η0 when using a higher con­
centration of SCLG (the effective volume fraction of the dispersed phase
is increased). A less important effect on shear thinning character has the
concentration of TOCNF, which significantly affects especially on τ0 in
low concentration biopolymer systems. The flow curve shape of the

sample with a 1 wt% concentration of both biopolymers is extremely
similar to the single 2 wt% TOCNF hydrogel. It may be concluded that
hydrogel with SCLG below 1 wt% does not have plastic behavior
(Lapasin et al., 2017).
The effect of the mass fraction of different biopolymers on the
rheological properties of hydrogel mixtures was studied in the case of
blending 2 wt% of SCLG and TOCNF (Fig. 3B), LAP and ALG (Fig. 3C) as
well as SCLG and ALG (Fig. 3D) (see single polymer hydrogel properties
from Section 3.1.1 as reference). The mixtures were prepared in mass
fractions of both biopolymers from 0.9 to 0.1. In Fig. 3B, flow curves
show almost negligible change in η0 between 0.1 and 0.25 mass fraction
of SCLG and also 0.5–0.90. On the other hand, τ0 increased with an

3.1.3. Fraction dependence of biopolymer blends on crosslinked hydrogel
systems
Biopolymer blends from Section 3.1.2. were additionally crosslinked
with 2 wt% CaCl2 solution. The rheological properties of such complex
hydrogel systems are illustrated in Fig. 4. Compared to the noncrosslinked systems a η0 increased by up to two decades (see Ta­
ble S1), demonstrating even more plastic behavior of hydrogel systems.
On the opposite, the η0 increases with increasing TOCNF or ALG mass
fraction in the case of crosslinked systems as a result of crosslinking
polymer chains by calcium ions through the carboxylic functional
groups on the polymer surface. As anticipated, the higher it is concen­
tration of the carboxylate level, the more plastic behavior could be
observed. More than that, the non-ionic SCLG cannot contribute to ionic
ˇ
interaction with TOCNF or ALG (Sebenik
et al., 2020) (see also discus­
sion of Figs. 7C and D). Furthermore, the 2nd Newtonian plateau
established in non-crosslinked systems (Fig. 3) was no longer existed in

crosslinked hydrogels (Fig. 4). The effect of increasing the τ0 of cross­
linked systems with increasing mass fraction of TOCNF or ALG
biopolymer is also evident (see Table S1).
Interestingly, the same or even slightly higher τ0 of non-crosslinked
SCLG/TOCNF mixtures with the SCLG mass fraction of 0.9 (29 Pa)
4


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

Fig. 3. Flow behavior of biopolymer concentration and fraction dependence hydrogel systems. In the panel A, the different shape (and color) of symbols represents
the various concentrations of biopolymer (blends of SCLG and TOCNF in equal weight ratio) in single polymer hydrogel preparation (the legend inside panel A).
Panel B (blends of 2% SCLG and TOCNF biopolymers dispersion), C (blends of 2% LAP and ALG) and D (blends of 2% SCLG and ALG) illustrate the flow curve of
biopolymers different weight ratio as shown in the legend below figure. The lines represent RBC model fit. The open symbols correspond to single polymer systems
presented in Fig. 1 (circles – SCLG, triangles – TOCNF, squares – LAP, and diamonds - ALG).

than their analogous crosslinked samples (28 Pa) is measured. In other
words, SCLG forms many hydrogen bonds between polymer chains and
water molecules, which diffuse into the SCLG network in the process of
swelling. Also, due to the lower mass fraction of TOCNF, a smaller
number of ionic bonds during crosslinking are formed because of the
lower concentration of carboxyl functional groups. Hydrogen bonds are
predominant over a few ionic interactions, in this case, reflecting no
effect of increasing τ0 in flow curves. In a previous study (Kopaˇc,
Abrami, et al., 2021) was shown that the hydrogen bonding interaction
arising among SCLG triple helices and TOCNF nanofibrils reduces the
tendency to prone to establish Ca2+ mediated ionic bonds. The opposite
could be clearly seen at higher TOCNF mass fractions (Fig. 4A). This

anomalous behavior of SCLG/TOCNF gels is further studied in detail in
Section 3.4.
Additionally, to prove the predominant effect of electrostatic inter­
action over hydrogen bonds, the crosslinked LAP/TOCNF systems
(Fig. 4B) were prepared. The rheological properties of non-crosslinked
ˇ
LAP/TOCNF are presented in the literature (Sebenik
et al., 2020),
therefore the η0 and τ0 values are shown only in Table S1 and not also in
a graphical form. The effect of hydrogen bonds is lower in exchange for
electrostatic interactions between LAP nanodisks (positively charged
edges of particles) and TOCNF fibrils (anionic side groups). The clay
particles of LAP and its aggregated forms are deposited between the
ˇ
polymer chains of TOCNF (Lapasin et al., 2017; Sebenik
et al., 2020),
which additionally connects the polymer nanofibers and strengthens the
internal structure of hydrogel systems. For that reason, the higher η0 and
τ0 values of crosslinked LAP/TOCNF hydrogels in comparison to SCLG/
TOCNF systems, especially in low TOCNF mass fractions, may possibly
be explained by stronger electrostatic interactions over hydrogen bonds.

3.1.4. Effect of crosslinking agent concentration on complex hydrogel
systems
To study the effect of crosslinking agent concentration on the rheo­
logical properties of hydrogel systems, 2% SCLG/2% TOCNF and 2%
LAP/2% TOCNF samples were blended in the 0.5/0.5 weight ratio and
additionally crosslinked with different concentrations of calcium ion
solution (0.17–2 wt%). Fig. 5 shows almost negligible effect of cross­
linking agent concentration on flow curves of ionically crosslinked

hydrogels. Crosslinked hydrogels have a rigid and stable structure even
at higher shear stresses (over 100 Pa). The ionic bonds between the
TOCNF polymer chains formed during the crosslinking process with
calcium ions (Curvello & Garnier, 2020) are obviously predominant
over other interactions (explained in Section 3.1.3.). Flow curves are
mostly plastic with a drastic drop in viscosity when the structure is
destroyed (see the τ0 data from Table S1). The LAP-containing hydrogel
systems possess higher values of η0 and τ0 (Table S1). As already dis­
cussed in Section 3.1.3, the LAP mainly contributes to electrostatic
interaction in the mixture unlike SCLG-containing systems where the
hydrogen bonds are strengthened. In the crosslinked systems, the
polymer chains are closer to each other which allows the electrostatic
interaction to play a more important role in a strong hydrogel structure
(Fig. 5). In the light of high η0 and τ0 illustrated in Fig. 5. for TOCNFcontaining systems, for ALG-containing systems with SCLG and LAP
only a few measurements at different concentration of crosslinker were
assessed for verification purposes (data not shown). Accordingly, due to
the higher concentration of carboxyl groups in ALG than in TOCNF
(more ionic interactions), the presence of ALG in the crosslinked mixture
exhibit an even stronger internal structure (higher η0 and τ0) compared
to TOCNF gels.

5


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

Fig. 4. Flow behavior of crosslinked (constant – 2 wt% addition of crosslinking agent solution) systems, presented in Fig. 3. The lines represent RBC model fit. The
open symbols correspond to single polymer systems presented in Fig. 1 (squares - LAP and circles – SCLG).

Ottewill, 1982; Chen & Zukoski, 1990) already reports that plateau
modulus (Gp) is proportional to the yield stress τ0:

τ0 ∝Gp

(2)

Further on, the average molecular weight between crosslinks Mc can
be derived from the following equation, including Gp as is clearly pre­
sented in recent paper (Yan et al., 2018):
Mc =

cρRT
Gp

(3)

where c is the polymer concentration, ρ is the density of water, T is the
absolute temperature and R is the ideal gas constant which represent the
product of Avogadro (NA) and Boltzmann (kB) constant. The Eq. (3)
leads to the crosslink density determination:

ρx =

ρg
Mc

(4)

where ρg is hydrogel density. Accordingly, the Eqs. (2)–(4) allow to find
the relation between τ0 and ρx. It is followed that a proportionality exists
between τ0 and ρx:

Fig. 5. Flow curve of crosslinked 2 wt% SCLG/TOCNF (A) and 2 wt% LAP/
TOCNF (B) hydrogel systems blended in the equal weight ratio. The symbols
represent the concentration of the added crosslinking agent solution. The lines
represent RBC model fit.

τ 0 ∝ ρx

(5)

In the previous work (Kopaˇc, Abrami, et al., 2021) we presented an
equation (Eq. (6)) which may be used to predict the crosslink density ρx
of hydrogels and analyze data of shear moduli adopted by the rheo­
logical oscillatory tests. The equation is based on the polymer-polymer
interaction as well as properties of polymers and crosslinking agents
and avoids difficult-to-determine parameters:
(
)
n
n
∑
∑
Nhyd,i mp,i − mp0,i Xhyd,i
Nion,i mp,i Xion,i − aion
ρx =
(6)
e x +

ahyd
2
2
i=0
i=0

3.2. Model development
The previous sections present the results of rheological rotational
tests and detailed characterization of hydrogel flow properties. The
shear thinning behavior was observed since the gels' structure reveals
almost constant viscosity at very small shear stresses (η0) followed by the
structure deformation at the τ0. Accordingly, mathematical model to
predict these flow characteristics of hydrogels is designed based on the
following assumptions. The literature (Buscall, Goodwin, Hawkins, &
6


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

presented in Fig. 6 are adopted from our previous paper (Kopaˇc, Abrami,
et al., 2021) (samples preparation protocol in this research is identical).
On the other hand, the τ0 data were determined based on flow curve
measurements (illustrated in Figs. 3-5) via the tangent crossover method
as presented in Fig. 2. During the action of shear force, hydrogels can
store part of the energy entering the system, and accordingly, after
removing the stress, they can restore part of the deformation. In this
context, parameter S as a coefficient of the energy that can be restored
defined per mole of crosslinks was introduced. Below the τ0, a hydrogel

deforms elastically and returns to its original shape when the applied
stress is removed due to energy stored in the network (described by
parameter S). The Shyd for hydrogen interactions and Sion for ionic in­
teractions were determined by fitting the Eqs. (8) and (9) to single noncrosslinked and crosslinked polymer hydrogels (SCLG, LAP, TOCNF,
ALG), respectively. The predetermined biopolymer individual S co­
efficients values were further endorsed in mixtures as proposed in Eq.
(7) to calculate the yield stress of complex hydrogel of (crosslinked)
polymer mixtures. Furthermore, Sdil was determined in fitting the Eq.
(10) of non-crosslinked ALG-containing mixtures. The results of the
model fitting parameters are presented in Table 1.
The Eq. (7) proved to be a satisfactory model prediction of τ0 (see red
lines in Fig. 6) in most of the samples except for ionically crosslinked
SCLG-containing systems whose unforeseen behavior is detailly dis­
cussed in Section 3.4. It may be observed that the SCLG internal struc­
ture, predominantly formed based on the hydrogen interactions (Shyd),
are able to restore more energy than LAP and TOCNF per moles of
crosslinks. On the other hand, in the case of crosslinked TOCNF and ALG
hydrogels, where the ionic interactions are predominant, the Sion is
significantly lower than Shyd. This means that internal structures based
on ionic interactions have less potential to restore after deformation
compared to hydrogen interactions-based structures. On the other hand,
it has to be noted that ionic interactions are much stronger than
hydrogen interactions along with the usually higher ρx of ionically
crosslinked hydrogels (the affinity to create crosslinks are higher). While
the shear stress should be significantly higher to deform the gels'
structure in crosslinked hydrogels than in non-crosslinked (where the
hydrogen interactions are predominant), the parameter S defined here
only refers to the energy that can be used to restore the deformation
(elastic character). The magnitude of shear stress is therefore hidden in
ionic and hydrogen polymer-polymer interactions (see Eqs. (6) and (7)).

Moreover, not only polymer-polymer interaction but also the (differ­
ently shaped) building blocks that form a hydrogel significantly affects
the flow properties. Therefore, a wide difference between the Shyd of
SCLG and TOCNF appears although the hydrogen interactions are pre­
dominant in both systems (see Table 1). The higher values of Shyd of
SCLG than LAP and TOCNF may indicate that by the additional forma­
tion of the network structure SCLG helices contribute to higher resis­
tance to flow than LAP nanodisks and TOCNF clusters, respectively.
In non-crosslinked ALG solutions the hydrogels' structure is not
formed since ALG is water-soluble and as such non-swellable (see Fig. 1).
In this case, ALG participates in the diluting process in the mixture of
SCLG or LAP dispersion with ALG solution. In ALG polymer solutions,
the internal structure of the liquid is formed by polymer molecules that
are randomly distributed and occupy the most energy-efficient confor­
mation (Lee & Mooney, 2012). Therefore, dilution with such solutions (η
of 2 wt% ALG solution at 25 ◦ C is 29 mPa⋅s) is significantly different
from dilution with water (η of water at 25 ◦ C is 0.89 mPa⋅s). The shear
thinning character of SCLG or LAP dispersions decreases linearly with
increasing the water content, due to the decreasing polymer concen­
tration (as formulated, Sdil for water should be equal to zero). Compared
to water dispersions of SCLG or LAP, the material becomes less elastic by
introducing ALG solution since ALG molecules dispersed in the SCLG
network or LAP hydrogel cannot contribute to the crosslinking process
and may present barriers in the formation of a hydrogel network.
Consequently, Shyd, having an elastic character, is here reduced by the
value of Sdil.

where N is substituent groups content on the polymer surface (OH,
COOH, NH2, etc.) defined per gram of dry polymer [mmol/g], mp and
mp0 are the mass concentration of dry polymer in the solvent and a

minimum mass concentration of polymer with the ability of hydrogel
formation [g/m3], ahyd is the hydrogen bonds functionality affinity [/],
the aion is the functionality affinity [wt%] of the ionic crosslinking agent,
x is the crosslinker concentration [wt%], i and n is the particular poly­
mer and the number of polymers in the hydrogel system, respectively,
and Xion,i is the mass fraction of polymer i in the hydrogel system. The
development of Eq. (6) for particular hydrogel system and the detail
explanation of symbols as well as its determination can be observed in
(Kopaˇc, Abrami, et al., 2021).
According to the relation between τ0 and ρx presented in Eq. (5), the
mathematical model to predict yield stress of complex hydrogel systems
can be proposed by analogy to Eq. (6):
(
)
n
n
∑
∑
Nhyd,i mp,i − mp0,i Xhyd,i
Nion,i mp,i Xion,i − aion
τ0 =
Sion
Shyd
(7)
e x +
ahyd
2
2
i=0
i=0

where S is the coefficient of proportionality in units J/mol and may
represents the energy defined per moles of crosslinks that can be used to
restore the deformation. It is defined as a function of polymer-polymer
interaction type (predominant effect of ionic (Sion) and hydrogen
(Shyd) interactions) and shapes of building blocks (depends on consistent
polymer or clay, etc.) in the gel system. In polymer dispersions or so­
lutions, the internal structure is formed by polymer molecules that are
randomly distributed and occupy the most energy-efficient conforma­
tion. Chains can be intertwined, there may be electrostatic repulsive or
attractive forces between them, and parts of the chains can be inter­
connected (crosslinked). These effects significantly affect the S value
(more detailed discussion of this parameter is further provided in Sec­
tion 3.3.).
Eq. (7) can be further simplified for single: (i) polymer systems giving
origin to hydrogels relaying on inter-chains hydrogen bons (Eq. (8)), and
(ii) ionically crosslinked hydrogel systems (Eq. (9)), and can be written
as analogy to model development in (Kopaˇc, Abrami, et al., 2021):
(
)
Nhyd mp − mp0
ahyd
τ0 = Shyd
(8)
2

τ0 = Sion

Nion mp −
e
2

aion
x

(9)

As it was possible to deduce from experimental observations, other
phenomena can be introduced and discussed in the development of
mathematical model, namely dilution effect. In samples where one of
the polymers does not participate in ionic, hydrogen, or any other in­
teractions process (e.g., ALG solution without crosslinking agent), the
dilution effect described by the Sdil is introduced in Eq. (10). It is spec­
ulated that the internal structure of the polymer solutions plays an
important role in the formation of the hydrogel network.
(
)
(
))
(
Nhyd mp − mp0
(10)
ahyd ∙ Shyd Xhyd − Sdil 1 − Xhyd
τ0 =
2
The Eq. (10) is written for two-component non-crosslinked hydrogel
systems (e.g., SCLG/ALG) where the first term represents the effect of
hydrogen bonding interactions between SCLG chains on τ0 and the
second term is the dilution effect contribution wherein the ALG polymer
molecules inhibit hydrogen interactions between SCLG chains and
reduce the Shyd of SCLG due to the steric hindrances.

3.3. Analysis of zero–shear viscosity and yield stress (model verification)
As an upgrade to a wide range of experimental data, the proposed
mathematical approach to analyze τ0 in a form of Eq. (7) was used to
describe the synergistic effect on shear behavior owing to their inter­
relationship. For this evaluation, the experimentally determined ρx data
7


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

Fig. 6. Correlation between crosslink density and yield stress. The τ0 data were determined by the same procedure as shown in Fig. 2. and refers to hydrogel samples
presented in Figs. 3 and 4. The ρx data were adopted from our publication (Kopaˇc, Abrami, et al., 2021) where the same samples (the same procedure of preparation)
were prepared. The red lines represent the Eq. (7) model prediction and numbers represent a mass fraction of first written polymer or clay in the mixture (in panel C
the numbers represent the concentration of both polymers in 0.5/0.5 mass fractions). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)

formation in a mixture with SCLG. Therefore, the presence of ALG in
SCLG systems has only a diluting effect, to the point, where gel's for­
mation is no longer possible (see also Fig. 3D for reference) which is for
fractions lower than 0.25 of 2 wt% SCLG or 0.5 wt% SCLG (according to
the literature (Lapasin et al., 2017), SCLG water dispersions under 0.75
wt% are almost Newtonian or slightly shear thinning with undetectable
viscoelastic properties). Differently, in SCLG/TOCNF samples, due to
TOCNF nanofibers hydrogen interaction contribution, the hydrogel
structure is observed for all fractions (from 0.9 to 0.1) of 2 wt% SCLG.
Interestingly, although the different effects of ALG and TOCNF on ρx of
SCLG based systems, the same linear trend on log η0 vs ρx plot can be
observed in both cases. It can be figured out that the fraction (concen­

tration) of polymers contributes to the size of invisibly small clusters
which are enabled by entanglements of nanofibers and (non)covalent
interactions between them and the size of beads which are formed due to
triple helical tertiary structure (triplex) in TOCNF and SCLG systems,
respectively.
Similar hydrogel systems as in previous case were prepared by
changing SCLG with LAP (Fig. 7B). The presence of ALG has only
diluting effect also in the non-crosslinked mixture with LAP. This phe­
nomenon is obvious at 50/50 mixture of LAP/ALG (see circle 0.5 in
Fig. 7B), where the trend is out of linearity on log η0 vs. ρx diagram in
comparison to mixtures with higher LAP content. On the other hand, the
TOCNF reduces the effect of the edge-face, edge-edge, and face-face
interactions between LAP particles on the viscoelastic properties,
ˇ
through the steric hindrance (Lapasin et al., 2017; Sebenik
et al., 2020).
It can be observed that LAP/TOCNF system (black circles) possess higher

Table 1
The values of proportionality coefficient S (Eq. (7)) for a specific hydrogel
constituent material. The Shyd was adopted by fitting the Eq. (8) of noncrosslinked single polymer (SCLG, LAP, TOCNF) hydrogels, Sion was adopted
by fitting the Eq. (9) of crosslinked single polymer (ALG, TOCNF) hydrogels, and
Sdil was adopted by fitting the Eq. (10) of non-crosslinked ALG-containing
mixtures. The predetermined values for SCLG, LAP, and TOCNF were fixed in
further model calculations.

SCLG
LAP
TOCNF
ALG

Shyd
[J/mol]

Sion
[J/mol]

Sdil
[J/mol]

317
151
69
/

/
/
26
34

/
/
/
150

Additionally, to analyze in detail the relationship between mechan­
ical and flow properties of hydrogels, the effect of ρx on η0 was also
identified and illustrated in Fig. 7. In absence of (ionic) crosslinking
agent, the predominant effect on the formation of hydrogel network and
consequently on rheological properties have the polymers which in pure

dispersion state forms strong gels' structure due to many interactions
(hydrogen, electrostatic, covalent, other van der Waals bonds…) present
between polymer chains (see Fig. 1 for reference). Fig. 7A reports the
effect of SCLG addition on ρx and η0 of non-crosslinked hydrogel sys­
tems. Due to the previously exposed dilution effect of ALG it can be
presumed that ALG solution has no contribution to hydrogel network
8


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

Fig. 7. Zero-shear viscosity (η0) obtained by RBC model against crosslink density (ρx) of various hydrogel systems (different colors). Panel A and B reports the effect
of SCLG and LAP, respectively, on rheological properties of non-crosslinked hydrogel system. Panel C and D show the predominant effect of ionic polymer (TOCNF
and ALG, respectively) on crosslinked samples. The numbers represent a mass fraction of first written polymer or clay in the mixture. The double color circles
correspond to the single biopolymer or clay dispersion (2% SCLG, 2% LAP, crosslinked 2% TOCNF, and crosslinked 2% ALG in panel A, B, C, and D, respectively)
summarized from Fig. 1.

η0 values at the same ρx of the network than in LAP/ALG system (red

ˇ
polymer nanofibers (Sebenik
et al., 2020) which may be the reason for
higher η0 values for crosslinked TOCNF or ALG systems containing LAP
(unlike nonionic SCLG-containing systems) (see Figs. 7C and D).
Differently, TOCNF–SCLG hydrogen bonds lead to the scarce effect of
calcium ions due to the hydrogen bonding interaction arising among
SCLG triple helices and TOCNF nanofibrils that are no longer so prone to
establish Ca2+ mediated ionic bonds (see Fig. 7C). In the SCLG/ALG

systems (Fig. 7D), the presence of SCLG reduces ρx as a result of steric
hindrances, especially evident in the intermediate region of SCLG frac­
tions (0.25–0.75). The effect of SCLG in complex hydrogel systems on ρx
and τ0 which leads to anomalous rheological behavior is profoundly
discussed in the next section. Here, it may be concluded, that the pres­
ence of SCLG in ionically crosslinked hydrogels reduces the ρx because it
complicates the ionic interactions between ALG or TOCNF polymer
chains (Figs. 7C and D). On the other hand, the addition of SCLG con­
tributes to high values of η0 even at lower ρx (under 2 mol/m3), indi­
cating that by the additional formation of the network structure SCLG
helices contribute to higher resistance to flow.

circles). This is contributed to building blocks that form a hydrogel in
different shapes. The LAP/TOCNF system is made of LAP disks which are
smaller and more homogeneous than TOCNF fibers which form ag­
glomerates (the meshes of invisibly small clusters) between which LAP
ˇ
disks are distributed (Sebenik
et al., 2020). It may be also assumed that
LAP contributes to the rheological response of the system in the same
way as other particulate fillers, although the individual clay nano­
particles are capable of aggregation (squares in Fig. 1). The contact
between the particles is also different due to the various external sur­
faces and therefore the complex LAP/TOCNF building blocks are
accordingly the main reason for higher η0 values.
Figs. 7C and D illustrate the effect of TOCNF and ALG presence on ρx
and η0 in ionically crosslinked hydrogel systems, respectively. It may be
figured out that the ρx and η0 are much higher than in the same noncrosslinked samples (Figs. 7A and B). At this point, the hydrogel for­
mation of ALG systems is stimulated by ionic interactions, omitting the
previously reported dilution effect of ALG in non-crosslinked samples.

This leads to the different dynamics of hydrogel properties, which is why
the ALG-containing crosslinked systems have higher ρx due to more
carboxyl groups on the surface of ALG than in the case of TOCNF.
Generally, the linear (exponential on a normal scale) growth in log
η0 versus ρx can be observed with some deviations marked at lower mass
fractions of ionic polymer. A very similar trend can be observed in LAP/
TOCNF (red circles in Fig. 7C) and LAP/ALG (red circles in Fig. 7D)
samples which may be explained by the noncompetition of LAP disks
with TOCNF or ALG mediated ionic interactions via calcium ions.
Furthermore, LAP nanodisks are distributed (interpenetrate) within the
TOCNF matrix due to electrostatic interactions between positively
charged edges of LAP disks and negatively charged carboxyl groups of
TOCNF nanofibers. The LAP nanodisks act as bridging agents between

3.4. Analysis of synergistic effects in hydrogel systems and anomalous
behavior of gels based on scleroglucan and TEMPO nanofibrillated
cellulose mixture
In Fig. 8 the black and red lines represent the systems where the
hydrogen and ionic (electrostatic) interactions, respectively, play a
pivotal role in the formation of hydrogel network. In Figs. 8B-D, the
predominant effect of ionic interactions over hydrogen bonds (and other
interactions, explained in the introduction) are evident even at low ρx
(the intersection of the black and red curves is at the point where there is
the lowest concentration (fraction) of the polymer with carboxyl groups
- i.e., ALG or TOCNF). Meanwhile, in Fig. 8A, the linear trendline (black)
9


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352

Fig. 8. Correlation between crosslink density and yield stress presented for four different types of polymer mixtures in addition of Laponite. The values in panel: (i) A
refers to Figs. 6A and B, (ii) B refers to Figs. 6F and I, (iii) C refers to Figs. 6 E and H, and (iv) D refers to Figs. 6 D and G. The black and red circles represent the noncrosslinked and crosslinked samples, respectively. The lines correspond to linear fit where black and red color refers to predominant effect of hydrogen and ionic
interaction in hydrogel systems, respectively. In panel A, the gray color represents the linear correlation between τ0 and theoretically calculated ρx by Eq. (6). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

of the SCLG/TOCNF system without a crosslinking agent can be
forwardly forecast into the “red region” where the SCLG/TOCNF sam­
ples were ionically crosslinked. It was proven in Fig. 8A, where the black
circles/lines present the samples without addition of ionic crosslinker
(without ionic interactions). Differently, the red circles/lines present the
same samples which were ionically crosslinked. However, it was spec­
ulated that the ionic interactions are predominant over hydrogen which
results especially in drastic increase in crosslink densities (see different
slopes of red and black lines on panels B, C and D). In panel A, the slope
of black line (predominant hydrogen interactions) is almost the same as
the slope of red line. This clearly shows that in the case of SCLG/TOCNF
systems the hydrogen bonds play a predominant role even at low con­
centrations of ionic interaction (Kopaˇc, Abrami, et al., 2021). As a
matter of fact, in Fig. 6, the (only) nonlinear behavior between τ0 and ρx
is revealed in crosslinked 2% SCLG/2% TOCNF gels in different weight
ratios of both constituent biopolymers (Fig. 6B). A small deviation can
be also observed in crosslinked 2% SCLG/2% ALG system (Fig. 6H).
Apparently, the SCLG biopolymer in mixture with the polymer that
could be ionically crosslinked often leads to anomalous behavior of gel's
rheological properties. Interestingly, a similar deviation was observed in
our previous study (Kopaˇc, Abrami, et al., 2021), where the SCLG/
TOCNF hydrogel systems clearly differ from the ρx prediction of the
developed model equation (Eq. (6)). According to this model prediction,

it can be speculated that ρx is underestimated, since the experimentally
determined values of ρx are lower than the Eq. (6) prediction. To support
this assumption, the model prediction of τ0 by Eq. (7) is added in Fig. 8A
for reference (gray circles). The observed deviations can be explained by
the predominance of hydrogen interactions that probably hinder the
formation of ionic bonds in obtaining the hydrogel network. Therefore,
the competition between TOCNF – Ca2+ − TOCNF mediated ionic in­
teractions and TOCNF–SCLG hydrogen bonds lead to the scarce effect of
calcium ions due to the hydrogen bonding interaction arising among

SCLG triple helices and TOCNF nanofibrils that are no longer so prone to
establish Ca2+ mediated ionic bonds. It is also not negligible that,
theoretically, it is possible that more than twice more hydrogen bonds
(2.8 mmol/g of hydroxyl group content) can be formed between TOCNF
polymer chains than ionic ones (1.2 mmol/g of carboxyl group content).
The SCLG has even more hydroxyl pendant groups (4.2 mmol/g of hy­
droxyl group content) on the surface than TOCNF that are accessible for
hydrogen bond formation (Kopaˇc, Abrami, et al., 2021). Moreover, as
illustrated in Fig. 8C, the presence of SCLG due to the steric hindrance
also reduces the effect of the ionic interactions in the SCLG/ALG system
on the ρx (viscoelastic properties).
On the other hand, except for SCLG, Fig. 8. reports that the addition
of TOCNF, ALG, and LAP to the polymer mixture produces only a
quantitative effect on its viscous behavior, which does not change
qualitatively. In the LAP-containing systems, the synergistic effects of
polymer-clay blending are crucial. The linear response shown in
Figs. 6D, F, G, and I and the apparent switch in the slope of the linear
curve (black to red lines) when the crosslinking agent is added (see
Figs. 8 B and D) may be evidence that the synergistic effect of polymerclay interaction gradually decreases when, despite high concentrations,
nanoparticles, distributed within the polymer network, are no longer

able to develop an extended aggregation structure (Lapasin et al., 2017).
Therefore, an increase of LAP fraction implies an increasing number and
size of nanoparticle aggregates, but polymer-clay interactions do not
disturb possible ionic interactions between crosslinked ALG or TOCNF
chains and thus no anomalous flow behavior occurs as in the case of
SCLG-containing systems. Moreover, addition of LAP in a mixture with
anionic polymer (ALG or TOCNF) contributes to the interpenetration
and electrostatic interaction between TOCNF fibrils or crosslinked ALG
particles and LAP nanodisks.

10


Carbohydrate Polymers 287 (2022) 119352

T. Kopaˇc et al.

4. Conclusions

From Wigner glasses to strong self-healing hydrogels formed by non-covalent
interactions. Soft Matter, 15(6), 1278–1289. />Boger, D. V. (2009). Rheology and the resource industries. Chemical Engineering Science,
64(22), 4525–4536. />Buscall, R., Goodwin, J. W., Hawkins, M. W., & Ottewill, R. H. (1982). Viscoelastic
properties of concentrated latices. Part 1. —Methods of examination. Journal of the
Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 78
(10), 2873–2887. />Cal´
o, E., & Khutoryanskiy, V. V. (2015). Biomedical applications of hydrogels: A review
of patents and commercial products. European Polymer Journal, 65, 252–267. https://
doi.org/10.1016/J.EURPOLYMJ.2014.11.024
Cao, H., Duan, L., Zhang, Y., Cao, J., & Zhang, K. (2021). Current hydrogel advances in
physicochemical and biological response-driven biomedical application diversity.

Signal Transduction and Targeted Therapy, 6(1), 1–31. />s41392-021-00830-x
Chen, L. B., & Zukoski, C. F. (1990). Discontinuous shear thinning in ordered
suspensions. Physical Review Letters, 65(1), 44. />PhysRevLett.65.44
Coviello, T., Palleschi, A., Grassi, M., Matricardi, P., Bocchinfuso, G., & Alhaique, F.
(2005). Scleroglucan: A versatile polysaccharide for modified drug delivery.
Molecules, 10(1), 6–33. />Curvello, R., & Garnier, G. (2020). Cationic cross-linked nanocellulose-based matrices for
the growth and recovery of intestinal organoids. Biomacromolecules, 22(2), 701–709.
/>Farrukh, S., Mustafa, K., Hussain, A., & Ayoub, M. (2018). In Synthesis and applications of
carbohydrate-based hydrogels (pp. 1–24). />Hecht, H., & Srebnik, S. (2016). Structural characterization of sodium alginate and
calcium alginate. Biomacromolecules, 17(6), 2160–2167. />ACS.BIOMAC.6B00378
Homayouni, A., Ehsani, M. R., Azizi, A., Yarmand, M. S., & Razavi, S. H. (2007). Effect of
lecithin and calcium chloride solution on the microencapsulation process yield of
calcium alginate beads. Iranian Polymer Journal, 16(9), 597–606. .
ir/en/journal/ViewPaper.aspx?id=95979 />per.aspx?id=95979.
Hoti, G., Caldera, F., Cecone, C., Pedrazzo, A. R., Anceschi, A., Appleton, S. L.,
Monfared, Y. K., & Trotta, F. (2021). Effect of the cross-linking density on the
swelling and rheological behavior of ester-bridged β-cyclodextrin nanosponges.
Materials (Basel, Switzerland), 14(3), 1–20. />Isogai, A., Saito, T., & Fukuzumi, H. (2011). TEMPO-oxidized cellulose nanofibers.
Nanoscale, 3(1), 71–85. />John, M. J., & Thomas, S. (2008). Biofibres and biocomposites. Carbohydrate Polymers, 71
(3), 343–364. />Klemm, D., Kramer, F., Moritz, S., Lindstră
om, T., Ankerfors, M., Gray, D., & Dorris, A.
(2011). Nanocelluloses: A new family of nature-based materials. Angewandte Chemie
International Edition, 50(24), 5438–5466. />Kopaˇc, T., Ruˇcigaj, A., & Krajnc, M. (2020). The mutual effect of the crosslinker and
biopolymer concentration on the desired hydrogel properties. International Journal of
Biological Macromolecules, 159, 557–569. />IJBIOMAC.2020.05.088
Kopaˇc, T., Abrami, M., Grassi, M., Ruˇcigaj, A., & Krajnc, M. (2021). Polysaccharide-based
hydrogels crosslink density equation: A rheological and LF-NMR study of polymerpolymer interactions. Carbohydrate Polymers, 277, Article 118895. />10.1016/J.CARBPOL.2021.118895
Kopaˇc, T., Krajnc, M., & Ruˇcigaj, A. (2021). A mathematical model for pH-responsive
ionically crosslinked TEMPO nanocellulose hydrogel design in drug delivery
systems. International Journal of Biological Macromolecules, 168, 695–707. https://

doi.org/10.1016/J.IJBIOMAC.2020.11.126
ˇ
Lapasin, R., Abrami, M., Grassi, M., & Sebenik,
U. (2017). Rheology of Laponitescleroglucan hydrogels. Carbohydrate Polymers, 168, 290–300. />10.1016/J.CARBPOL.2017.03.068
Lee, K. Y., & Mooney, D. J. (2012). Alginate: Properties and biomedical applications.
Progress in Polymer Science, 37(1), 106. />PROGPOLYMSCI.2011.06.003
Lei, J., Li, Z., Xu, S., & Liu, Z. (2021). Recent advances of hydrogel network models for
studies on mechanical behaviors. Acta Mechanica Sinica, 37(3), 367–386. https://doi.
org/10.1007/S10409-021-01058-2
Liang, L., Bhagia, S., Li, M., Huang, C., & Ragauskas, A. J. (2020). Cross-linked
nanocellulosic materials and their applications. ChemSusChem, 13(1), 78–87.
/>Lin, N., & Dufresne, A. (2014). Nanocellulose in biomedicine: Current status and future
prospect. European Polymer Journal, 59, 302–325. />EURPOLYMJ.2014.07.025
Lin, N., Bruzzese, C., & Dufresne, A. (2012). TEMPO-oxidized nanocellulose participating
as crosslinking aid for alginate-based sponges. ACS Applied Materials and Interfaces, 4
(9), 4948–4959. />Lundahl, M. J., Berta, M., Ago, M., Stading, M., & Rojas, O. J. (2018). Shear and
extensional rheology of aqueous suspensions of cellulose nanofibrils for biopolymerassisted filament spinning. European Polymer Journal, 109, 367–378. />10.1016/J.EURPOLYMJ.2018.10.006
Nazmabadi, M., Shirdast, A., Sharif, A., & Aalaie, J. (2020). Aqueous/brine solutions
viscosity and surface properties of hydrophobically modified scleroglucans: Role of
grafted chain length. Carbohydrate Polymers, 229, Article 115519. />10.1016/J.CARBPOL.2019.115519
Paolicelli, P., Varani, G., Pacelli, S., Ogliani, E., Nardoni, M., Petralito, S., Adrover, A., &
Casadei, M. A. (2017). Design and characterization of a biocompatible physical

This study presents a comprehensive flow behavior characterization
of the hydrogel systems based on polysaccharide mixtures. In noncrosslinked samples, the shear-thinning behavior is more pronounced
at higher SCLG and LAP fractions in the mixture (higher concentration)
eventually becoming apparently plastic even at the lowest fractions. On
the other hand, TOCNF and ALG content in non-crosslinked samples
contribute to more elastic behavior with pronounced 2nd Newtonian
plateau. In ionically crosslinked samples, however, the anionic bio­

polymers TOCNF and ALG content play a predominant role in deter­
mining rheological properties compared to nonionic polymers, giving
evidentially plastic behavior with significantly higher zero-shear vis­
cosity and yield stress values (extended list is included in Table S1). The
detailed flow behavior examination is performed based on polymerpolymer interaction theory, mainly introduced in our previous work
(Kopaˇc, Abrami, et al., 2021), where the hydrogen interactions play a
pivotal role in hydrogels, until the addition of crosslinking agent, when
ionic interactions take precedence over hydrogen bonds. Following the
rheological observations supported with the theoretical background, the
mathematical prediction of yield stress in relation to crosslink density is
proposed. A detailed analysis of synergetic effect in complex hydrogel
systems that predominantly affect the rheological properties of gels
leads to the conclusion that there is a linear correlation between cross­
link density and yield stress. At this point, the parameter S is introduced,
which serves as the coefficient of proportionality in units [J/mol] and
represents a new way of analyzing the rheological properties of complex
hydrogels. According to the developed model prediction of yield stress,
the higher S represents the more reversible (elastic) hydrogel structures
in the deformation process under the shear force action. The model
correlation provides an insight into the interior of the hydrogel network,
where various anomalous processes occur, which are difficult to detect
with conventional rheological measurements. Thus, the scarce effect of
calcium ions in SCLG/TOCNF systems was discovered. It was found that
the hydrogen interaction can predominantly affect the rheological
properties over ionic interaction due to the hydrogen bonding interac­
tion arising among SCLG triple helices and TOCNF nano-fibrils that are
no longer so prone to establish Ca2+ mediated ionic bonds. Moreover,
the dilution effect of non-crosslinked ALG-containing systems was
detected where dilution with polymer solution mainly contributes to less
elastic hydrogel behavior.

Supplementary data to this article can be found online at https://doi.
org/10.1016/j.carbpol.2022.119352.
CRediT authorship contribution statement
Tilen Kopacˇ: Conceptualization, Methodology, Software, Valida­
tion, Formal analysis, Investigation, Writing – original draft, Visualiza­
tion. Aleˇs Ruˇ
cigaj: Software, Formal analysis, Data curation, Writing –
review & editing, Visualization. Matjaˇ
z Krajnc: Writing – review &
editing, Resources, Supervision, Project administration, Funding
acquisition.
Declaration of competing interest
The authors declared no conflicts of interest.
Acknowledgment
The authors acknowledge the financial support from the Slovenian
Research Agency (research core funding No. P2-0191).
References
Becher, T. B., Braga, C. B., Bertuzzi, D. L., Ramos, M. D., Hassan, A., Crespilho, F. N., &
Ornelas, C. (2019). The structure–property relationship in LAPONITE® materials:

11


T. Kopaˇc et al.

Carbohydrate Polymers 287 (2022) 119352
ˇ
Sebenik,
U., Lapasin, R., & Krajnc, M. (2020). Rheology of aqueous dispersions of
Laponite and TEMPO-oxidized nanofibrillated cellulose. Carbohydrate Polymers, 240,

Article 116330. />Way, A. E., Hsu, L., Shanmuganathan, K., Weder, C., & Rowan, S. J. (2012). pHresponsive cellulose nanocrystal gels and nanocomposites. ACS Macro Letters, 1(8),
1001–1006. />Xu, H., Liu, Y., Xie, Y., Zhu, E., Shi, Z., Yang, Q., & Xiong, C. (2019). Doubly cross-linked
nanocellulose hydrogels with excellent mechanical properties. Cellulose, 26(16),
8645–8654. />Yan, H. J., Casalini, T., Hulsart-Billstră
om, G., Wang, S., Oommen, O. P., Salvalaglio, M.,
Larsson, S., Hilborn, J., & Varghese, O. P. (2018). Synthetic design of growth factor
sequestering extracellular matrix mimetic hydrogel for promoting in vivo bone
formation. Biomaterials, 161, 190–202. />BIOMATERIALS.2018.01.041
Zhao, J., Mayumi, K., Creton, C., & Narita, T. (2017). Rheological properties of tough
hydrogels based on an associating polymer with permanent and transient crosslinks:
Effects of crosslinking density. Journal of Rheology, 61(6), 1371. />10.1122/1.4997589

hydrogel based on scleroglucan for topical drug delivery. Carbohydrate Polymers,
174, 960–969. />Park, K., Dawson, J. I., Oreffo, R. O. C., Kim, Y.-H., & Hong, J. (2020).
Nanoclay–polyamine composite hydrogel for topical delivery of nitric oxide gas via
innate gelation characteristics of Laponite. Biomacromolecules, 21(6), 2096–2103.
/>Pati˜
no-Mas´
o, J., Serra-Parareda, F., Tarr´es, Q., Mutj´
e, P., Espinach, F. X., & DelgadoAguilar, M. (2019). TEMPO-oxidized cellulose nanofibers: a potential bio-based
superabsorbent for diaper production. Nanomaterials, 9(9), 1271. />10.3390/NANO9091271
Roberts, G. P., Barnes, H. A., & Carew, P. (2001). Modelling the flow behaviour of very
shear-thinning liquids. Chemical Engineering Science, 56(19), 5617–5623. https://doi.
org/10.1016/S0009-2509(01)00291-3
Ruzicka, B., & Zaccarelli, E. (2011). A fresh look at the Laponite phase diagram. Soft
Matter, 7(4), 1268–1286. />ˇ
Sebenik,
U., Krajnc, M., Aliˇc, B., & Lapasin, R. (2018). Ageing of aqueous TEMPOoxidized nanofibrillated cellulose dispersions: A rheological study. Cellulose, 26(2),
917–931. />
12