
3
with the controlled aqueous polymer solutions [43], we
can evaluate Πmix as
Πmix = Πext +G0c
c01/3
,(2)
where G0and ccan be experimentally measured. We
assume Eq. (2) throughout this study. In this study, we
consider only the equilibrium states of the polymer gels
in external polymer solutions, such as the as-prepared
(Q= 1 and Πext >0), swelling (Q > 1 and Πext >0),
and equilibrium swollen (Q > 1 and Πext = 0) states.
Here, Q≡V/V0is the volume swelling ratio [see
Fig. 1(b)].
In a solution with sufficiently long polymer chains, the
Flory–Huggins (FH) mean-field theory [26–28] predicts
that
Πmix =−kBT
vφ+ ln (1 −φ) + χφ2,(3)
where kB,T,v, and φare the Boltzmann constant, abso-
lute temperature, volume of solvent molecules, and poly-
mer volume fraction, respectively. The FH interaction
parameter χquantifies the energies of the interactions
between the polymer units and the solvent molecules per
lattice site as follows: χ≡z(pp +ss −2ps)/(2kBT),
where zis the coordination number of the lattice, and
pp,ss, and ps are the polymer–polymer, solvent–
solvent, and polymer–solvent interaction energies,
respectively. Although the FH theory is applicable only
to concentrated polymer solutions [29], it has been in-
appropriately applied to polymer gels containing a large
amount of solvent, by suitably adjusting χ[5–8]. Conse-
quently, significant inconsistencies in χwere reported in
previous studies, although χshould be constant for the
same polymer–solvent system and temperature [45,46].
For example, χ= 0.426 for the poly(ethylene glycol)
(PEG) hydrogel [30] and χ= 0.366, 0.441, and 0.465
for the aqueous PEG solution with the molar masses
M= 4.6 [31], 10 [32], and 33 kg/mol [33], respectively,
were reported at T= 298 K, using v= 18×10−6m3/mol.
By contrast, we propose that the semidilute scaling law
of polymer solutions [2,37] governs Πmix in polymer gels,
based on a recent observation [44]. In Ref. [44], it was ob-
served that a universal osmotic equation of state of poly-
mer solutions [2,37,47–49] describes Πmix throughout
the gelation process. For polymer solutions, the semidi-
lute scaling law is expressed as ΠmixM/(cNAkBT) =
K(c/c∗)1/(3ν−1), where K'1.1 is the universal constant
for the semidilute polymer solutions, NAis the Avogadro
constant, c∗is the overlap concentration, and ν'0.588
is the universal critical exponent of the SAW [38–42].
However, for polymer gels, c∗→0 and c/c∗→ ∞ [44],
because the molar mass of a polymer network is infinite.
Thus, we introduce the regulator Mseg to propose the
following semidilute scaling law:
Πmix
nkBT=Kn
n∗
seg
1
3ν−1
,(4)
as a fundamental principle for polymer networks that
contain a large amount of solvent. Here, both sides of
Eq. (4) are dimensionless; n≡cNA/Mseg is the number
density of a “segment” of polymers, where Mseg is
the molar mass of the segment. In this study, we use
Mseg = 44 ×10−3kg/mol, which is the molar mass of an
ethylene glycol unit as a segment. In Eq. (4), n∗
seg is a
parameter that depends on Tand the type of polymer
network and solvent considered. We determined n∗
seg by
adjusting to the experimental data. This procedure is
similar to the determination of χin the FH mean-field
theory from the experimental data.
Notably, a few pioneering studies [11,16–20] attempted
to describe the equilibrium swelling of polymer gels using
the semidilute scaling law such as Πmix ∼φ9/4. However,
they did not focus on demonstrating the superiority of
the semidilute scaling law over the FH mean-field theory,
owing to the limited accuracy of the gel experiments and
the lack of theoretical comprehension of polymer gels. By
contrast, in this study, we experimentally demonstrate
the superiority of Eq. (4) over the FH mean-field the-
ory [Eq. (3)] in terms of predicting the equilibrium state
of polymer gels throughout the quasistatic swelling pro-
cess. We verify the superiority of the semidilute scaling
law using precisely controlled homogeneous networks [34]
and a different definition of the polymer mass concentra-
tion [44] that enables us to extend the universality of the
osmotic equation of state of polymer solutions.
III. MATERIALS AND METHODS
A. Fabrication of model gels
As a model system to examine the quasistatic swelling
process of chemically crosslinked polymer gels, we used
a tetra-branched PEG hydrogel [34], synthesized via the
AB-type cross-end coupling of two prepolymer (tetra-
arm PEG) units of equal size. Each end of the tetra-arm
PEG was modified with a mutually reactive maleimide
(tetra-PEG MA) and thiol (tetra-PEG SH). We dissolved
tetra-PEG MA and tetra-PEG SH (NOF Co., Japan &
XIAMEN SINOPEG BIOTECH Co., Ltd., China) in
phosphate citrate buffer with an ionic strength of 100 mM
and a pH of 3.8. For gelation, we mixed these two so-
lutions with equal molar masses Mfor equal prepoly-
mer mass concentrations c0at various mixing fractions
s. Here, sis the molar fraction of the minor prepolymers
(tetra-PEG SH) to all prepolymers (0 ≤s≤1/2). By
tuning s, the desired connectivity pcan be obtained in
accordance with p= 2s[50,51], where pis defined as
the fraction of reacted terminal functional groups to all