betaFeedback

WorksPhysics and Astronomy

Response of Hydrogels and Microgels to Nanoparticle Crowding

Abstract

Internal degrees of freedom of cross-linked polymer networks enable compressible hydrogels and microgels to swell or deswell by absorbing or expelling solvent in response to environmental changes. Macromolecules in solution can influence the swelling response by acting as crowding agents, with practical relevance for drug delivery and biosensing. We extend the Flory-Rehner theory of polymer network swelling to incorporate nanoparticle crowders as another implicit species that can perturb a gel through volume exclusion, including the entropic cost of crowder penetration from polymer field theory. Within the free energy landscape of asymmetric ternary mixtures, we numerically solve equilibrium relations equating solvent and crowder chemical potentials inside and outside the gel. Independently, we perform Monte Carlo simulations, including trial moves that allow gels to swell and take up crowders. We investigate how crowder size, solution concentration, cross-link density, and solvent quality influence swelling of polymeric gels, partitioning of crowders between inside and outside, and structural properties of microgel suspensions governed by Hertzian elastic pair interactions. Our results show that nanoparticle crowding promotes deswelling, thereby weakening interparticle correlations, and that, with increasing the concentration of crowders, polymeric gels reach a maximum loading capacity, dependent on crowder size and polymer cross-link density.

Introduction

Hydrogels are cross-linked polymer networks, whose internal degrees of freedom enable them to swell to many times their dry size by absorbing solvent (e.g., water) [1]. Microgels are soft colloidal hydrogel particles, typically tens of nanometers to a micron in size, whose translational degrees of freedom allow them to self-assemble into ordered structures [2, 3, 4, 5]. Such polymeric gels swell and deswell in response to changes in solution properties, such as temperature, pH, and solvent composition. Their inherent capacity to undergo significant volume changes in response to external stimuli qualifies microgels as smart colloidal particles and enables a wide range of applications, from targeted drug delivery systems to advanced biosensors, where precise control over material properties can enhance functionality and performance [6, 7, 8].

Changing solvent composition by adding cosolvents or components of higher molecular weight, such as macromolecules or nanoparticles, can influence polymer conformations and swelling behavior. Mixing cosolvents in solution, for instance, can alter solvent quality, thus modifying effective interactions between polymer segments and the statistical distribution of segments, e.g., changing a polymer coil from a random walk to a self-avoiding walk. Adding nanoscale particles can crowd other species through excluded-volume interactions [9].

The phenomenon of macromolecular crowding manifests when the conformations of biopolymers (e.g., proteins, RNA, DNA) or other macromolecules in biological cells are constrained by interactions with other species or by confinement [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]. Similarly, the conformations of polymer chains in a swollen gel can collectively respond to the presence of crowding agents (crowders) in solution.

Importantly, since gels are penetrable, the crowding here comes from within. The main purpose of this paper is to model and analyze the swelling response of hydrogels and microgels to macromolecular crowding and, conversely, the influence of swelling on uptake and release of crowders, which could represent drug or dye molecules.

The swelling behavior of polymer networks in pure solvents has been traditionally modeled using the classical Flory-Huggins and Flory-Rehner theories [21, 22, 23, 24], which attribute swelling to a balance between osmotic pressure of a binary polymer-solvent mixture and elastic pressure originating from the entropic cost of deforming polymer chains. Previous studies have extended the Flory-Rehner theory to ternary mixtures and have modeled polydimethylsiloxane (PDMS) [25], cross-linked polymer membranes [26], copolymer networks in mixed solvents [27], and polydisperse mixtures of polymers [28], as well as microgels mixed with free polymers [29].

Several recent experimental studies have measured properties of microgels immersed in macromolecular solutions. For example, Bradley et al. Bradley et al. [29, 30] investigated swelling and uptake properties of copolymer microgels of N-isopropylacrylamide (NIPAM) and acrylic acid in aqueous solutions of poly(ethylene oxide) (PEO) of varying molecular weights. In a relevant drug delivery application, lipid nanoparticles were encapsulated in dextran microgels and used as a chemotherapeutic treatment against colon cancer [31]. In another study, nanogels were used as sacrificial pH-degradable templates to control the porosity of functional dextran-based microgels [32].

Despite progress, a comprehensive theoretical framework for modeling the interplay between network swelling and the loading and expulsion of a third volume-excluding component is lacking. Our research aims to bridge this gap by extending the Flory-Rehner theory to mixtures of polymer, solvent, and substantially larger particles (crowders) that significantly exclude volume, with the goal of providing deeper insight into the swelling behavior and functional capabilities of polymeric gels. As a practical implementation of our theory, we also generalize a Monte Carlo simulation approach previously developed for modeling suspensions of interacting microgels in pure solvents [33, 34, 35]. The simulations include novel trial moves that allow microgels to change both size and composition, i.e., interior concentration of crowders. In this way, we investigate how the swelling behavior of microgels depends on cross-link density, solvent quality, and crowder size and concentration.

The remainder of the paper is organized as follows. In § 2, we describe our model of soft, compressible hydrogels and colloidal microgels dispersed in a solution of hard, nanoscale crowding agents. We then derive our theory of polymeric gel swelling and absorption (uptake) of crowders from solution by extending the Flory-Rehner theory of cross-linked polymer networks to predict the free energy of ternary polymer-solvent-crowder mixtures. A distinguishing feature of our model and theory is the incorporation of excluded volume of the crowders. From the approximate free energy, we derive expressions for the chemical potentials of the solvent and crowders. In Sec. III, we describe our computational methods for equating the chemical potentials of the respective species inside and outside the gel. By numerically solving two nonlinear equations, we obtain the equilibrium swelling ratio and interior crowder volume fraction. We also develop and perform Monte Carlo simulations of one or multiple microgels, including elastic interparticle interactions, and compute single-microgel equilibrium properties, as well as structural properties of a suspension of microgels. In Sec. IV, we present and compare our numerical results and simulation data and discuss our interpretation of the impact of crowders on polymeric gels, and vice versa, of gels on the distribution of crowders. Finally, we summarize and conclude.

Models and Theories

Model of Polymeric Gel-Crowder Solutions

The system of interest consists of a polymeric gel and crowding agents (crowders) dispersed in a solvent at constant volume VV and temperature TT, as depicted in Figure 1. The term “crowder” is used generically here to describe any nanoscale object that excludes volume to the microgels, e.g., small proteins, nanoparticles, or other macromolecules smaller than the pores of the gel. We assume a suspension of NmN_m microgels can exchange solvent and crowders via a semi-permeable membrane with a reservoir of crowder solution. The number of solvent molecules NsN_s and of crowders NcN_c vary through exchange with the reservoir, which imposes chemical potentials of those species. The microgels themselves are assumed permeable to solvent, but only selectively permeable to crowders, limited by an entropic cost of perturbing the polymer network (see Figure 2).

A swollen microgel immersed in a crowder solution

Figure 1. A swollen microgel, consisting of a cross-linked polymer network (red chains, bounded by outer dashed circle), immersed in a crowder solution, deswells to a collapsed state (inner dotted circle) by expelling solvent (yellow background) and crowders (green particles). Created with BioRender.com/1kz4x5k.

Microscopic depiction of the interior of a microgel

Figure 2. Microscopic depiction of the interior of a microgel, consisting of a network of polymer chains (red), connected by cross-linkers (blue spheres), immersed in a solution of solvent (yellow background) and crowders (green spheres). Created with BioRender.com/b8o1719.

We model microgels as spherical, compressible, soft colloidal particles, each composed of a cross-linked network of polymer chains, consisting of NmonN_{\mathrm{mon}} monomers, NchN_{\mathrm{ch}} chains, and tetravalent cross-links with fraction x=0.5Nch/Nmonx = 0.5N_{\mathrm{ch}}/N_{\mathrm{mon}}. In experiments, microgels typically have a core-shell structure, consisting of a densely cross-linked core surrounded by a more loosely cross-linked shell [36, 37]. In the present work, however, we adopt a simplified description with a uniform distribution of cross-links. Such homogeneous microgels can be synthesized under appropriate conditions [4, 38, 39]. Polymer chains comprise statistically independent segments (each a few monomers) of length ll (Kuhn length), a measure of the chain stiffness [24, 40]. In a good solvent, the segments avoid one another, but otherwise rotate freely [24, 41]. This flexibility allows a chain to adopt a vast number of conformations, whose statistics closely follow a self-avoiding random walk (SAW) and whose shapes can be modeled as fluctuating, penetrable ellipsoids [42, 43, 44, 45, 46, 47, 48, 49] (see Figure 3). An uncrowded SAW chain with Nseg=Nmon/NchN_{\mathrm{seg}} = N_{\mathrm{mon}}/N_{\mathrm{ch}} segments, spanning neighboring cross-links, has a root-mean-square radius of gyration [49, 50, 51]

Spanning cross-links, polymer-chain segments, and crowders

Figure 3. Spanning cross-links (blue spheres), the segments of a polymer chain (red spheres) form a random walk, whose shape can be approximated by an ellipsoid. Assuming affine swelling, the polymer chain expands or contracts by a factor proportional to the network’s swelling ratio, which is influenced by the presence of crowders (green spheres). Created with BioRender.com/4ywc5yl.

Rg=CNsegγl,(1)R_{\mathrm{g}} = C N_{\mathrm{seg}}^{\gamma} l, \tag*{(1)}

with Flory exponent γ=0.588\gamma= 0.588 and amplitude C=0.441C = 0.441. Internal degrees of freedom allow a microgel to absorb/expel surrounding solution and swell/deswell between a swollen state, with radius aa and volume VmV_{\mathrm{m}}, and collapsed (dry or unswollen) state, with radius a0a_0 and volume V0V_0. The degree of swelling is characterized by the linear swelling ratio, α=a/a0=(Vm/V0)1/3=ϕp−1/3\alpha= a/a_0 = (V_{\mathrm{m}}/V_0)^{1/3} = \phi_{\mathrm{p}}^{-1/3} where ϕp=V0/Vm\phi_{\mathrm{p}} = V_0/V_{\mathrm{m}} defines the fraction of the total volume occupied by polymer monomers.

To isolate the influence of excluded volume on microgel swelling, we model crowders as mutually impenetrable hard spheres of radius RcR_{\mathrm{c}} interacting via a pair potential

vHS(r)={∞,r<2Rc0,r≥2Rc.(2)v_{\mathrm{HS}}(r) = \begin{cases} \infty, & r < 2R_{\mathrm{c}} \\ 0, & r \ge2R_{\mathrm{c}}. \end{cases} \tag*{(2)}

Assuming polymer monomers and solvent molecules to be much smaller than the crowders, the latter thus exclude to the former a volume vc=(4π/3)Rc3v_{\mathrm{c}} = (4\pi/3)R_{\mathrm{c}}^3 and to other crowders a volume 8vc8v_{\mathrm{c}}. As we shall see, excluded volume affects swelling in two independent ways – via mixing entropy and polymer conformational entropy.

Theory of Polymeric Gel-Crowder Solutions

To model polymeric gels suspended in a solution of crowders, we adapt the Flory-Rehner theory of polymer networks [21, 22, 23] to approximate the free energy of a ternary mixture of polymer, solvent, and crowders. Broad applications to swelling of polymeric gels have demonstrated the utility and qualitative accuracy of the Flory-Rehner theory, as well as some limitations [52]. We initially focus on a bulk hydrogel or single isolated microgel and then generalize to a suspension of interacting microgels. Starting from the Flory-Huggins theory of polymer solutions [24], we consider a lattice model of a ternary mixture, in which the lattice sites may be singly occupied by polymer monomers or solvent molecules or multiply occupied by crowders. To account for the potentially much larger volume of crowders, relative to polymer monomers and solvent molecules, we allow the crowders to multiply occupy a number of sites n=vc/vp>1n = v_{\mathrm{c}}/v_{\mathrm{p}} > 1, where vcv_{\mathrm{c}}, vpv_{\mathrm{p}}, and vsv_{\mathrm{s}} are respective volumes of crowders, polymer monomers, and solvent molecules.

The free energy of a polymeric gel in a solution of crowders is a combination of mixing entropy, associated with freedom of the different species of particle to assemble on the lattice, elastic entropy, associated with swelling of the polymer network, pair interaction energy of mixing of the various species, and a free energy of penetration of the polymer network by crowders. The mixing entropy SmixS_{\mathrm{mix}} depends, through the Boltzmann relation, Smix=kBln⁡ΩmixS_{\mathrm{mix}} = k_{\mathrm{B}} \ln\Omega_{\mathrm{mix}}, on the total number of configurations Ωmix\Omega_{\mathrm{mix}} available to all of the particles (kBk_{\mathrm{B}} is the Boltzmann constant). For a sufficiently dilute solution, the placement of crowders can be decoupled from the placement of the other species, allowing Ωmix\Omega_{\mathrm{mix}} to be factorized, according to Ωmix=ΩpΩsΩc\Omega_{\mathrm{mix}} = \Omega_{\mathrm{p}}\Omega_{\mathrm{s}}\Omega_{\mathrm{c}}, into the respective numbers of ways of placing NcN_{\mathrm{c}} crowders, NpN_{\mathrm{p}} polymer monomers, and NsN_{\mathrm{s}} solvent molecules. Assuming full occupancy of the lattice, these numbers are not all independent, but subject to the constraint

Np+Ns+nNc=M,(3)N_{\mathrm{p}} + N_{\mathrm{s}} + nN_{\mathrm{c}} = M, \tag*{(3)}

where MM is the total number of lattice sites, which changes as the gel swells or deswells. Taking the polymer monomer volume as a unit of volume, and expressing the gel volume as

Vm=Mvp=Npvp+Nsvp+Ncvc,(4)V_{\mathrm{m}} = Mv_{\mathrm{p}} = N_{\mathrm{p}}v_{\mathrm{p}} + N_{\mathrm{s}}v_{\mathrm{p}} + N_{\mathrm{c}}v_{\mathrm{c}}, \tag*{(4)}

the site occupation (volume) fractions of the various species are related via

ϕp+ϕ~s+ϕc=1,(5)\phi_{\mathrm{p}} + \tilde{\phi}_{\mathrm{s}} + \phi_{\mathrm{c}} = 1, \tag*{(5)}

where ϕp=Np/M=Npvp/Vm\phi_{\mathrm{p}} = N_{\mathrm{p}}/M = N_{\mathrm{p}}v_{\mathrm{p}}/V_{\mathrm{m}}, ϕs=Nsvs/Vm\phi_{\mathrm{s}} = N_{\mathrm{s}}v_{\mathrm{s}}/V_{\mathrm{m}}, and ϕc=Ncvc/Vm\phi_{\mathrm{c}} = N_{\mathrm{c}}v_{\mathrm{c}}/V_{\mathrm{m}} are the polymer, solvent, and crowder volume fractions, respectively, and for convenience we define ϕ~s≡Ns/M=(vp/vs)ϕs\tilde{\phi}_{\mathrm{s}} \equiv N_{\mathrm{s}}/M = (v_{\mathrm{p}}/v_{\mathrm{s}})\phi_{\mathrm{s}} as an effective solvent volume fraction that accounts for different polymer and solvent volumes. In the reservoir of crowder solution (absent polymer), Eq. (5) reduces to

ϕ~s0+ϕc0=1,(6)\tilde{\phi}_{s}^{0}+\phi_{c}^{0}=1, \tag*{(6)}

where ϕ~s0\tilde{\phi}_{s}^{0} and ϕc0\phi_{c}^{0} are the respective volume fractions of solvent and crowders in the reservoir.

To incorporate mutual excluded volume, we assume that each crowder excludes to solvent and polymer nn sites, but to other crowders a number of sites nex>nn_{\mathrm{ex}}>n. For a solution sufficiently dilute that overlaps of crowder excluded-volume spheres may be neglected in the counting of microstates, the number of ways of placing NcN_{c} indistinguishable crowders is approximated as

Ωc0=MNc!(M−nex)(M−2nex)⋯[M−(Nc−1)nex].(7)\Omega_{c}^{0}=\frac{M}{N_{c}!}(M-n_{\mathrm{ex}})(M-2n_{\mathrm{ex}})\cdots[M-(N_{c}-1)n_{\mathrm{ex}}]. \tag*{(7)}

Assuming Nc≫1N_{c}\gg1 and applying the Stirling approximation, the associated crowder entropy is given by

Sc0kB=−Ncln⁡(NcM)−(Mnex−Nc)ln⁡(1−nexNcM)=−Vmvp[ϕcnln⁡(ϕcn)+1−ϕc′nexln⁡(1−ϕc′)],(8)\begin{aligned} \frac{S_{c}^{0}}{k_{B}}=-N_{c}\ln\left(\frac{N_{c}}{M}\right)-\left(\frac{M}{n_{\mathrm{ex}}}-N_{c}\right)\ln\left(1-\frac{n_{\mathrm{ex}}N_{c}}{M}\right) \\ &=-\frac{V_{m}}{v_{p}}\left[\frac{\phi_{c}}{n}\ln\left(\frac{\phi_{c}}{n}\right)+\frac{1-\phi_{c}^{\prime}}{n_{\mathrm{ex}}}\ln(1-\phi_{c}^{\prime})\right], \tag*{(8)} \end{aligned}

where ϕc′=(nex/n)ϕc\phi_{c}^{\prime}=(n_{\mathrm{ex}}/n)\phi_{c} is the fraction of volume excluded by crowders to one another, which exceeds the volume fraction ϕc\phi_{c} excluded to polymer and solvent.

Going a step further, we approximately account for correlations between pairs of crowders due to overlap of crowder excluded-volume spheres by subtracting from the number of sites excluded to a crowder an estimate of the mean number of overlap sites (which otherwise would be doubly-counted). Still assuming a sufficiently dilute crowder solution that multi-particle clusters contribute negligibly to the microstate count, thus avoiding the complication of triplets and many-particle clusters of crowders, the number of ways of placing NcN_{c} indistinguishable crowders now can be written as

Ωc=MNc!(M−nex)(M−2nex+nˉov)(M−3nex+3nˉov)\Omega_{c}=\frac{M}{N_{c}!}(M-n_{\mathrm{ex}})(M-2n_{\mathrm{ex}}+\bar{n}_{\mathrm{ov}})(M-3n_{\mathrm{ex}}+3\bar{n}_{\mathrm{ov}})
⋯[M−(Nc−1)nex+Nc(Nc−1)2nˉov],(9)\cdots\left[M-(N_{c}-1)n_{\mathrm{ex}}+\frac{N_{c}(N_{c}-1)}{2}\bar{n}_{\mathrm{ov}}\right], \tag*{(9)}

where nˉov\bar{n}_{\mathrm{ov}} is the mean number of sites in the pair overlap regions of crowder excluded-volume spheres, averaged over crowder pair configurations. Again applying the Stirling approximation, we have

ln⁡Ωc≃−Ncln⁡(NcM)+Nc+∫0Ncdmln⁡(1−nexMm+nˉov2Mm2).(10)\ln\Omega_{c}\simeq-N_{c}\ln\left(\frac{N_{c}}{M}\right)+N_{c}+\int_{0}^{N_{c}}dm\ln\left(1-\frac{n_{\mathrm{ex}}}{M}m+\frac{\bar{n}_{\mathrm{ov}}}{2M}m^{2}\right). \tag*{(10)}

We can calculate nˉov\bar{n}_{\mathrm{ov}} by integrating the intersection volume vov(r)v_{\mathrm{ov}}(r) of two excluded-volume spheres, of radius 2Rc2R_{c} and center-center separation rr, over the overlap volume, being a spherical shell of inner radius 2Rc2R_{c} and outer radius 4Rc4R_{c}, and multiplying by the density of sites 1/vp1/v_{p} and the probability 1/M1/M of a site being occupied:

nˉov=4πMvp2∫2Rc4Rcdr r2vov(r).(11)\bar{n}_{\mathrm{ov}}=\frac{4\pi}{Mv_{p}^{2}}\int_{2R_{c}}^{4R_{c}}dr\,r^{2}v_{\mathrm{ov}}(r). \tag*{(11)}

where vov(r)=(π/12)(8Rc+r)(4Rc−r)2v_{\mathrm{ov}}(r)=(\pi/12)(8R_{c}+r)(4R_{c}-r)^{2}. The exact result, nˉov=(17/32)nex2/M\bar{n}_{\mathrm{ov}}=(17/32)n_{\mathrm{ex}}^{2}/M, is closely approximated by nˉov≃nex2/(2M)\bar{n}_{\mathrm{ov}}\simeq n_{\mathrm{ex}}^{2}/(2M). Remarkably (or fortuitously), this approximation allows completing the square in the argument of the logarithm on the right side of Eq. (10), greatly simplifying the integral and leading to

ln⁡Ωc=−Ncln⁡(NcM)+Nc+2∫0Ncdmln⁡(1−nex2Mm).(12)\ln\Omega_{c}=-N_{c}\ln\left(\frac{N_{c}}{M}\right)+N_{c}+2\int_{0}^{N_{c}}dm\ln\left(1-\frac{n_{\mathrm{ex}}}{2M}m\right). \tag*{(12)}

Evaluating the remaining integral, we obtain for the crowder entropy

SckB=Vmvp[ϕcnln⁡(ϕcn)+ϕcn+4nex(1−ϕc′2)ln⁡(1−ϕc′2)].(13)\frac{S_{c}}{k_{B}}=\frac{V_{m}}{v_{p}}\left[\frac{\phi_{c}}{n}\ln\left(\frac{\phi_{c}}{n}\right)+\frac{\phi_{c}}{n}+\frac{4}{n_{\mathrm{ex}}}\left(1-\frac{\phi_{c}^{\prime}}{2}\right)\ln\left(1-\frac{\phi_{c}^{\prime}}{2}\right)\right]. \tag*{(13)}

to be compared with Eq. (8), which neglects pair correlations. Modeling the crowders as hard spheres, each excluding eight times its own volume to others, we take nex=8nn_{\mathrm{ex}}=8n and ϕc′=8ϕc\phi_{c}^{\prime}=8\phi_{c}, finally yielding

SckB=−Vmvp[ϕcnln⁡(ϕcn)+ϕcn+1−4ϕc2nln⁡(1−4ϕc)].(14)\frac{S_{c}}{k_{B}}=-\frac{V_{m}}{v_{p}}\left[\frac{\phi_{c}}{n}\ln\left(\frac{\phi_{c}}{n}\right)+\frac{\phi_{c}}{n}+\frac{1-4\phi_{c}}{2n}\ln(1-4\phi_{c})\right]. \tag*{(14)}

The associated crowder free energy density is given by

βFcVm=ρcln⁡(ρcvp)+ρc+1−4ϕc2vcln⁡(1−4ϕc),(15)\beta\frac{F_{c}}{V_{m}}=\rho_{c}\ln(\rho_{c}v_{p})+\rho_{c}+\frac{1-4\phi_{c}}{2v_{c}}\ln(1-4\phi_{c}), \tag*{(15)}

where β=1/(kBT)\beta=1/(k_{B}T) and ρc=ϕc/vc\rho_{c}=\phi_{c}/v_{c} is the crowder number density. Comparing the low-density expansion,

βFcVm=ρcln⁡(ρcvp)−ρc+4vcρc2+163vc2ρc3+⋯ ,(16)\beta\frac{F_{c}}{V_{m}}=\rho_{c}\ln(\rho_{c}v_{p})-\rho_{c}+4v_{c}\rho_{c}^{2}+\frac{16}{3}v_{c}^{2}\rho_{c}^{3}+\cdots, \tag*{(16)}

with the corresponding virial expansion [53] for a fluid of hard spheres of volume vcv_{c},

βFcVm=ρcln⁡(ρcvp)−ρc+B2vcρc2+12B3vc2ρc3+⋯ ,(17)\beta\frac{F_{c}}{V_{m}}=\rho_{c}\ln(\rho_{c}v_{p})-\rho_{c}+B_{2}v_{c}\rho_{c}^{2}+\frac{1}{2}B_{3}v_{c}^{2}\rho_{c}^{3}+\cdots, \tag*{(17)}

we find that our approximate free energy gives the correct ideal-gas free energy, βFid/Vm=ρcln⁡(ρcvp)−ρc\beta F_{\mathrm{id}}/V_{m}=\rho_{c}\ln(\rho_{c}v_{p})-\rho_{c}, the exact hard-sphere second virial coefficient, B2=4B_{2}=4, and a third virial coefficient, B3=32/3B_{3}=32/3, in very close agreement with the hard-sphere value of B3=10B_{3}=10.

In passing, we note that our approach for incorporating excluded volume in a lattice model of a solution of crowder particles may be compared with the theory of Marques Fernandes et al. [54] for modeling the thermodynamics of a lattice gas of hard cubes, each occupying nn sites. When translated into our symbols, the latter theory (with vc=λdv_c = \lambda^d) predicts a free energy density

βFcVm=ρcln⁡ϕc+1+ln⁡nvc(1−ϕc)ln⁡(1−ϕc).(18)\beta\frac{F_c}{V_m} = \rho_c \ln\phi_c + \frac{1+\ln n}{v_c}(1-\phi_c)\ln(1-\phi_c). \tag*{(18)}

Since it is constructed to interpolate between the low- and high-density limits, Eq. (18) yields the correct ideal-gas free energy (as does our theory). In contrast with our approach, however, the theory of Marques Fernandes et al. applies at all densities, while our theory focuses on and is limited to dilute crowder concentrations.

Considering now the polymer monomers and solvent molecules, for a given configuration of NcN_c crowders that exclude nNcnN_c sites, the number of distinct arrangements of these other components is given by

Ωps=(M−nNc)!Ns!Np!.(19)\Omega_{ps} = \frac{(M-nN_c)!}{N_s!N_p!}. \tag*{(19)}

The associated polymer-solvent entropy is given by

SpskB=Nsln⁡(M−nNcNs)−Npln⁡Np=−Vmvp[ϕ~sln⁡ϕ~s−(1−ϕc)ln⁡(1−ϕc)+ϕpln⁡ϕp].(20)\begin{aligned} \frac{S_{ps}}{k_B} = N_s \ln\left(\frac{M-nN_c}{N_s}\right)-N_p\ln N_p \\ &= -\frac{V_m}{v_p}\left[\tilde{\phi}_s\ln\tilde{\phi}_s-(1-\phi_c)\ln(1-\phi_c)+\phi_p\ln\phi_p\right]. \tag*{(20)} \end{aligned}

where we have invoked the volume fraction sum rule [Eq. (5)]. The last term on the right side represents the entropy of the polymer monomers if considered as independent particles. However, connecting the monomers into chains and cross-linking the chains into a network reduces the polymer entropy by a factor of one over the microgel molecular weight NmonN_{\mathrm{mon}} [24, 40]. Assuming that Nmon≫1N_{\mathrm{mon}} \gg1, the polymer contribution is then completely negligible. Combining Eqs. (14) and (20), the final mixing entropy takes the form

SmixkB=−Vmvp[ϕ~sln⁡ϕ~s−(1−ϕc)ln⁡(1−ϕc)+ϕcnln⁡(ϕcn)+ϕcn+1−4ϕc2nln⁡(1−4ϕc)].(21)\frac{S_{\mathrm{mix}}}{k_B} = -\frac{V_m}{v_p}\left[\tilde{\phi}_s\ln\tilde{\phi}_s-(1-\phi_c)\ln(1-\phi_c) +\frac{\phi_c}{n}\ln\left(\frac{\phi_c}{n}\right)+\frac{\phi_c}{n}+\frac{1-4\phi_c}{2n}\ln(1-4\phi_c)\right]. \tag*{(21)}

The form of the last term on the right side makes clear that our model is limited to the range ϕc<0.25\phi_c < 0.25, such that the argument of the logarithm remains positive. Note that in the limit in which the crowders shrink to a size comparable to that of monomers, and then occupy only single sites (nex=n=1n_{\mathrm{ex}} = n = 1), the crowders would no longer exclude extra volume. In this case, the crowder entropy would be described by Eq. (8) and the mixing entropy would reduce simply to

SmixkB=−Nsln⁡ϕ~s−Ncln⁡ϕc(nex=n=1).(22)\frac{S_{\mathrm{mix}}}{k_B}=-N_s\ln\tilde{\phi}_s-N_c\ln\phi_c\quad(n_{\mathrm{ex}}=n=1). \tag*{(22)}

In the Flory-Huggins lattice model, which includes only nearest-neighbor interactions, the mean energy of the mixture can be expressed as a sum of pair energies,

Eˉmix=ϵppNˉpp+ϵssNˉss+ϵccNˉcc+ϵpsNˉps+ϵpcNˉpc+ϵscNˉsc,(23)\bar{E}_{\mathrm{mix}}=\epsilon_{pp}\bar{N}_{pp}+\epsilon_{ss}\bar{N}_{ss}+\epsilon_{cc}\bar{N}_{cc} +\epsilon_{ps}\bar{N}_{ps}+\epsilon_{pc}\bar{N}_{pc}+\epsilon_{sc}\bar{N}_{sc}, \tag*{(23)}

where ϵij\epsilon_{ij} is the interaction energy between nearest-neighbor particles of species ii and jj and Nˉij\bar{N}_{ij} is the mean number of ijij nearest-neighbor pairs. In the associated mean-field theory, in which correlations are neglected and the neighboring sites of each particle are simply replaced by average occupation fractions for the whole lattice, Nˉij\bar{N}_{ij} is approximated by zijNiNj/Mz_{ij}N_iN_j/M, where zijz_{ij} is the appropriate coordination number, i.e., number of jj particles that can be neighbors of an ii particle. The mean mixing energy EmixE_{\mathrm{mix}}, defined as the difference in mean energy of the mixture and the individual pure components, is then approximated as

βEˉmix=M(χpsϕpϕ~s+χpcϕpϕc+χscϕ~sϕc)=χpsNpϕ~s+χpcNpϕc+χscNsϕc.(24)\begin{aligned} \beta\bar{E}_{\mathrm{mix}}=M\left(\chi_{ps}\phi_p\tilde{\phi}_s+\chi_{pc}\phi_p\phi_c+\chi_{sc}\tilde{\phi}_s\phi_c\right) \\ &=\chi_{ps}N_p\tilde{\phi}_s+\chi_{pc}N_p\phi_c+\chi_{sc}N_s\phi_c. \tag*{(24)} \end{aligned}

where the Flory solvency parameters, χps\chi_{ps}, χpc\chi_{pc}, and χsc\chi_{sc}, are combinations of the ϵij\epsilon_{ij} and zijz_{ij} parameters. As an example, the polymer-solvent parameter is given by

χps=zps(ϵps−ϵpp+ϵss2).(25)\chi_{ps}=z_{ps}\left(\epsilon_{ps}-\frac{\epsilon_{pp}+\epsilon_{ss}}{2}\right). \tag*{(25)}

Note that, because crowders occupy multiple sites, their associated coordination numbers (zpcz_{pc}, zscz_{sc}, zccz_{cc}) exceed those involving only polymer or solvent, although these factors are ultimately absorbed into the solvency parameters. The mixing free energy is the sum of the mixing entropy and mixing energy.

A further contribution to the free energy comes from the conformational entropy of the polymer chains and the freedom of a polymeric gel to swell and deswell. As in our previous work [33, 34], we model this contribution using the mean-field Flory-Rehner theory of polymer networks [21, 22, 23], which treats the polymers as Gaussian coils, assumes affine swelling of the cross-linked network, and predicts an elastic free energy,

βFel=32Nch(α2−1−ln⁡α).(26)\beta F_{\mathrm{el}}=\frac{3}{2}N_{\mathrm{ch}}(\alpha^2-1-\ln\alpha). \tag*{(26)}

While the excluded volume of crowders may certainly affect the swelling ratio in equilibrium, we assume that the basic functional form of FelF_{\mathrm{el}} remains unchanged.

The final contribution to the free energy of a polymeric gel in a solution of crowders comes from the entropy loss resulting from penetrating a polymer network by a crowder. Modeling the polymer chains as fluctuating ellipsoids whose principal radii follow the statistics of a random walk [42, 43, 44, 45, 46, 47, 48, 49], consistent with the Gaussian approximation of the Flory-Rehner theory, and assuming the crowders are small enough as to perturb the conformation of only one chain at a time, we estimate the penetration free energy from polymer field theory [55, 56, 57, 58] as the entropic cost of inserting a hard sphere into a polymer coil in a good solvent:

βεpen(α)=4.416q1.29932(27)\beta\varepsilon_{\mathrm{pen}}(\alpha)=\frac{4.416}{q^{1.29932}} \tag*{(27)}

where q=(α/αres)Rg/Rcq=(\alpha/\alpha_{\mathrm{res}})R_g/R_c is the polymer-to-crowder size ratio in a gel of swelling ratio α\alpha relative to the reservoir swelling ratio αres\alpha_{\mathrm{res}} in pure solvent. Note that this definition of the size ratio, which incorporates the effect of swelling/deswelling on the mean polymer size, is consistent with the affine swelling assumption. In the limit n→1n \to1, in which the crowders become indistinguishable from solvent molecules, q≫1q \gg1 and we can set εpen=0\varepsilon_{\mathrm{pen}}=0.

In the current modified Flory-Rehner approximation, the total Helmholtz free energy of a gel in the solution is finally given by

F=Eˉmix−TSmix+Fel+Ncεpen,(28)F=\bar{E}_{\mathrm{mix}}-TS_{\mathrm{mix}}+F_{\mathrm{el}}+N_c\varepsilon_{\mathrm{pen}}, \tag*{(28)}

where NcN_c here represents the number of crowders inside the gel. We emphasize that, although microscopic species are represented by density fields in this coarse-grained level of modeling, microgels in a suspension must be rendered explicitly, as discussed below.

In equilibrium, the chemical potentials of solvent and crowders inside the gel, μs\mu_s and μc\mu_c, must equal the respective chemical potentials outside, μs0\mu_s^0 and μc0\mu_c^0, which are fixed via exchange of both species with the reservoir:

μs(α,ϕc,ϕc0)=μs0(ϕc0),μc(α,ϕc,ϕc0)=μc0(ϕc0).(29)\mu_s(\alpha,\phi_c,\phi_c^0)=\mu_s^0(\phi_c^0), \qquad\mu_c(\alpha,\phi_c,\phi_c^0)=\mu_c^0(\phi_c^0). \tag*{(29)}

These relations constitute a system of two nonlinear equations, which, for a given reservoir crowder concentration ϕc0\phi_c^0, can be numerically solved for the two unknowns, namely, the crowder and polymer concentrations inside the gel, ϕc\phi_c and ϕp\phi_p (or equivalently the swelling ratio α\alpha). Given ϕc\phi_c and ϕp\phi_p, the solvent concentration inside the gel ϕs\phi_s follows from Eq. (5).

The solvent and crowder chemical potentials in Eqs. (29) are derived from the total free energy:

μs=(∂F∂Ns)Np,Nc,μc=(∂F∂Nc)Np,Ns,(30)\mu_s=\left(\frac{\partial F}{\partial N_s}\right)_{N_p,N_c}, \qquad\mu_c=\left(\frac{\partial F}{\partial N_c}\right)_{N_p,N_s}, \tag*{(30)}

where the partial derivatives are evaluated subject to Eqs. (3)–(5), but allowing for a change of volume of the gel (and thus number of lattice sites) as solvent and crowders are absorbed or expelled [24]. The separation of the free energy into mixing, elastic, and penetration terms implies a corresponding separation of the chemical potentials. Substituting the mixing free energy [Eqs. (21) and (24)] into Eq. (30), we obtain the mixing contributions to the solvent and crowder chemical potentials:

βμs,mix=ln⁡ϕ~s+ϕp+ϕcn−ln⁡(1−ϕc)+12nln⁡(1−4ϕc)+χpsϕp(1−ϕ~s)−χpcϕpϕc+χscϕc(1−ϕ~s)(31)\beta\mu_{s,\mathrm{mix}}=\ln\tilde{\phi}_s+\phi_p+\frac{\phi_c}{n}-\ln(1-\phi_c)+\frac{1}{2n}\ln(1-4\phi_c) +\chi_{ps}\phi_p(1-\tilde{\phi}_s)-\chi_{pc}\phi_p\phi_c+\chi_{sc}\phi_c(1-\tilde{\phi}_s) \tag*{(31)}

and

βμc,mix=ln⁡(ϕcn)+ϕc+nϕp−32ln⁡(1−4ϕc)−n[χpsϕpϕ~s−χpcϕp(1−ϕc)−χscϕ~s(1−ϕc)].(32)\beta\mu_{c,\mathrm{mix}}=\ln\left(\frac{\phi_c}{n}\right)+\phi_c+n\phi_p-\frac{3}{2}\ln(1-4\phi_c) -n\left[\chi_{ps}\phi_p\tilde{\phi}_s-\chi_{pc}\phi_p(1-\phi_c)-\chi_{sc}\tilde{\phi}_s(1-\phi_c)\right]. \tag*{(32)}

From the elastic free energy [Eq. (26)], we obtain the elastic contributions:

βμs,el=NchNs(α2−1/2)ϕs(33)\beta\mu_{s,\mathrm{el}}=\frac{N_{\mathrm{ch}}}{N_s}(\alpha^2-1/2)\phi_s \tag*{(33)}

and

βμc,el=NchNc(α2−1/2)ϕc,(34)\beta\mu_{c,\mathrm{el}}=\frac{N_{\mathrm{ch}}}{N_c}(\alpha^2-1/2)\phi_c, \tag*{(34)}

which result from [Eq. (4)] and the dependence of the swelling ratio α\alpha on volume VmV_m. Finally, from the penetration free energy [Eq. (27)], we obtain the penetration contributions:

μs,pen=−γ3vsvcϕcεpen,μc,pen=(1−γ3ϕc)εpen.(35)\mu_{s,\mathrm{pen}}=-\frac{\gamma}{3}\frac{v_s}{v_c}\phi_c\varepsilon_{\mathrm{pen}}, \qquad\mu_{c,\mathrm{pen}}=\left(1-\frac{\gamma}{3}\phi_c\right)\varepsilon_{\mathrm{pen}}. \tag*{(35)}

From Eqs. (31) and (32), the corresponding solvent and crowder chemical potentials in the reservoir (absent polymer) are given by

μs0=ϕc0n+12nln⁡(1−4ϕc0)+χsc(ϕc0)2(36)\mu_s^0=\frac{\phi_c^0}{n}+\frac{1}{2n}\ln(1-4\phi_c^0)+\chi_{sc}(\phi_c^0)^2 \tag*{(36)}

and

μc0=ln⁡(ϕc0n)+ϕc0−32ln⁡(1−4ϕc0)+χscn(1−ϕc0)2.(37)\mu_c^0=\ln\left(\frac{\phi_c^0}{n}\right)+\phi_c^0-\frac{3}{2}\ln(1-4\phi_c^0)+\chi_{sc}n(1-\phi_c^0)^2. \tag*{(37)}

Substituting Eqs. (31)–(37) into Eqs. (29), we can solve for the equilibrium values of α\alpha and ϕc\phi_c for a polymeric gel immersed in a solution of crowders of given concentration ϕc0\phi_c^0. Alternatively, and equivalently, we can minimize with respect to variations in NsN_s and NcN_c the appropriate thermodynamic potential in the semigrand canonical ensemble, namely, the Landau free energy,

L=F−μs0Ns−μc0Nc,(38)L=F-\mu_s^0N_s-\mu_c^0N_c, \tag*{(38)}

a Legendre transformation of the Helmholtz free energy. For a suspension of NmN_m interacting microgels, the total free energy includes contributions accounting for the entropy associated with microgel center-of-mass configurations and the total energy of interparticle interactions. Since microgels have polymeric degrees of freedom, the interaction energy depends on the internal states of the particles and may have many-body contributions. Assuming strictly pairwise interactions, however, and preaveraging over most internal degrees of freedom, we approximate the interaction energy by a pairwise sum

UH=∑i=1Nm∑j>i=1NmvH(rij),(39)U_{\mathrm{H}}=\sum_{i=1}^{N_{\mathrm{m}}}\sum_{j>i=1}^{N_{\mathrm{m}}}v_{\mathrm{H}}(r_{ij}), \tag*{(39)}

where

vH(rij)=Bij(1−rijai+aj)5/2(40)v_{\mathrm{H}}(r_{ij})=B_{ij}\left(1-\frac{r_{ij}}{a_i+a_j}\right)^{5/2} \tag*{(40)}

is the Hertz elastic potential between microgels ii and jj with swollen radii aia_i and aja_j and center-to-center separation rijr_{ij}. The amplitude depends on the particle radii, Young’s moduli YiY_i, and Poisson’s ratios νi\nu_i via [59]

Bij=815(1−νi2Yi+1−νj2Yj)−1(ai+aj)2aiaj.(41)B_{ij}=\frac{8}{15}\left(\frac{1-\nu_i^2}{Y_i}+\frac{1-\nu_j^2}{Y_j}\right)^{-1}(a_i+a_j)^2\sqrt{a_i a_j}. \tag*{(41)}

Since the Young’s modulus of a gel is proportional to kBTk_{\mathrm{B}}T and the cross-link density [40], we take

βBij=YcNch(ai+aj)2aiajai3+aj3,(42)\beta B_{ij}=Y_{\mathrm{c}}N_{\mathrm{ch}}\frac{(a_i+a_j)^2\sqrt{a_i a_j}}{a_i^3+a_j^3}, \tag*{(42)}

where YcY_{\mathrm{c}} is a dimensionless factor introduced to calibrate the model with experiment. We note that heterogeneously cross-linked microgels can be more accurately described by a multi-Hertzian model [37, 60] that accounts for different cross-linker densities in the core and outer shells.

To determine the equilibrium state of a suspension of microgels interacting via Hertz pair potentials, we perform Monte Carlo simulations with trial moves combining displacements, size changes, and crowder concentration changes of all microgels. The acceptance probabilities for the trial moves can be derived from the Hamiltonian HH of the system and the partition function,

Z=⟨⟨e−βH⟩m⟩p⟩c,(43)Z=\langle\langle e^{-\beta H}\rangle_{\mathrm{m}}\rangle_{\mathrm{p}}\rangle_{\mathrm{c}}, \tag*{(43)}

where ⟨ ⟩m\langle\ \rangle_{\mathrm{m}}, ⟨ ⟩p\langle\ \rangle_{\mathrm{p}}, and ⟨ ⟩c\langle\ \rangle_{\mathrm{c}} denote respective traces over microgel center-of-mass, polymeric, and crowder degrees of freedom, the latter governing partitioning of crowders between inside and outside the microgels. Neglecting many-body interactions, the Hamiltonian separates into one-body and pair terms, H0H_0 and HpairH_{\mathrm{pair}}:

H=H0+Hpair.(44)H=H_0+H_{\mathrm{pair}}. \tag*{(44)}

Assuming that, of all the polymeric degrees of freedom, swelling is governed by a relatively small subset (denoted collectively by aa), and denoting the remaining degrees of freedom by p′p', we may write the partition function as

Z=⟨⟨e−βHeff⟩m⟩a⟩c,(45)Z=\langle\langle e^{-\beta H_{\mathrm{eff}}}\rangle_{\mathrm{m}}\rangle_{a}\rangle_{\mathrm{c}}, \tag*{(45)}

where ⟨ ⟩a\langle\ \rangle_{a} and ⟨ ⟩c\langle\ \rangle_{\mathrm{c}} denote averages over all microgel radii and interior numbers of crowders and

Heff=−kBTln⁡⟨e−β(H0+Hpair)⟩p′(46)H_{\mathrm{eff}}=-k_{\mathrm{B}}T\ln\left\langle e^{-\beta(H_0+H_{\mathrm{pair}})}\right\rangle_{p'} \tag*{(46)}

is an effective Hamiltonian resulting from the trace ⟨ ⟩p′\langle\ \rangle_{p'} over p′p' coordinates. Now making a mean-field approximation that neglects correlations between microgel swelling and pair interactions, and thus decouples one-body (single-microgel) and pair coordinates, we write

Heff=−kBTln⁡(⟨e−βH0⟩p′⟨e−βHpair⟩p′)=F0−kBTln⁡⟨e−βUeff⟩p′.(47)\begin{aligned} H_{\mathrm{eff}}=-k_{\mathrm{B}}T\ln\left(\left\langle e^{-\beta H_0}\right\rangle_{p'}\left\langle e^{-\beta H_{\mathrm{pair}}}\right\rangle_{p'}\right) \\ &=F_0-k_{\mathrm{B}}T\ln\left\langle e^{-\beta U_{\mathrm{eff}}}\right\rangle_{p'}. \tag*{(47)} \end{aligned}

Here we define a single-microgel polymeric free energy,

F0(a,ϕc)=−kBTln⁡⟨e−βH0⟩p′=∑i=1NmF(ai,ϕc,i),(48)F_0(a,\phi_{\mathrm{c}})=-k_{\mathrm{B}}T\ln\left\langle e^{-\beta H_0}\right\rangle_{p'}=\sum_{i=1}^{N_{\mathrm{m}}}F(a_i,\phi_{\mathrm{c},i}), \tag*{(48)}

which depends only on the swollen microgel radii and interior crowder volume fractions, and an effective pair interaction energy,

Ueff(a)=−kBTln⁡⟨e−βHpair⟩p′,(49)U_{\mathrm{eff}}(a)=-k_{\mathrm{B}}T\ln\left\langle e^{-\beta H_{\mathrm{pair}}}\right\rangle_{p'}, \tag*{(49)}

which depends on the swollen radii and center-of-mass coordinates of the microgels. We associate F0F_0 with the Flory-Rehner free energy in (28), summed over all microgels, and UeffU_{\mathrm{eff}} with the Hertz potential energy in (39). After tracing over a subset of internal degrees of freedom of the microgels, the resulting canonical partition function in this coarse-grained model reduces to

Z=⟨⟨e−βF0(a)⟩m⟨e−βUH(a)⟩m⟩a⟩c.(50)Z=\langle\langle e^{-\beta F_0(a)}\rangle_{\mathrm{m}}\langle e^{-\beta U_{\mathrm{H}}(a)}\rangle_{\mathrm{m}}\rangle_{a}\rangle_{\mathrm{c}}. \tag*{(50)}

In the semigrand canonical ensemble, which allows for exchange of solvent and crowders with a reservoir solution of crowders, the semigrand partition function then takes the form

Ξ=⟨⟨e−βL0(a,ϕc)⟨e−βUH(a)⟩m⟩a⟩c,(51)\Xi=\left\langle\left\langle e^{-\beta L_0(a,\phi_{\mathrm{c}})}\left\langle e^{-\beta U_{\mathrm{H}}(a)}\right\rangle_{\mathrm{m}}\right\rangle_{a}\right\rangle_{\mathrm{c}}, \tag*{(51)}

where L0L_0 represents a sum over all microgels of the single-microgel Landau free energy of (38),

L0(a,ϕc)=∑i=1NmL(ai,ϕc,i),(52)L_0(a,\phi_{\mathrm{c}})=\sum_{i=1}^{N_{\mathrm{m}}}L(a_i,\phi_{\mathrm{c},i}), \tag*{(52)}

which depends on the radii and interior crowder volume fractions of all microgels.

Snapshot of microgels in an implicit solution of crowders

Figure 4. Snapshot from an MC simulation of a suspension of microgels (red spheres) in an implicit solution of crowders. The microgels are confined to a cubic cell with periodic boundary conditions and initialized on an FCC lattice.

Computational Methods

To predict the equilibrium state, i.e., mean swelling ratio and interior crowder volume fraction, of a polymeric gel immersed in a solution of crowders, we numerically solved the coupled nonlinear (29), equating chemical potentials inside and outside the gel. In practice, we used the fsolve function from the optimize module in the SciPy library of Python to simultaneously solve for α\alpha and ϕc\phi_{\mathrm{c}}, looping over ϕc0\phi_{\mathrm{c}}^{0} and using the solution for each iteration as an initial estimate for the next.

As a consistency check on our numerical computations, as well as to extend our study to microgel suspensions, we also performed Monte Carlo (MC) simulations [61, 62, 63] of either a single microgel or many interacting microgels. We emphasize that, while our model includes explicit microgel particles, the polymer, solvent, and crowder components are implicit, entering only as densities in the one-body Landau free energy [Eq. (52)]. The implicit reservoir solution of crowders fixes the chemical potentials, μs0\mu_{\mathrm{s}}^{0} and μc0\mu_{\mathrm{c}}^{0}, that appear in L0L_{0}.

A simulation proceeds through a series of trial moves, each comprising three independent random changes applied to every particle: a displacement of the microgel center (r→r+Δr\mathbf{r} \to\mathbf{r} + \Delta\mathbf{r}) and changes (increase or decrease) of the microgel radius (a→a+Δaa \to a + \Delta a) and interior crowder volume fraction (ϕc→ϕc+Δϕc\phi_{\mathrm{c}} \to\phi_{\mathrm{c}} + \Delta\phi_{\mathrm{c}}). While trial displacements are irrelevant for a single microgel, they affect the pair interaction energy for multiple microgels. The trial changes, Δr\Delta\mathbf{r}, Δa\Delta a, and Δϕc\Delta\phi_{\mathrm{c}}, are randomly generated within tolerances set to optimize efficiency and accuracy.

A single MC step consists of attempting, then either accepting or rejecting, a trial move, in which the positions and internal states of all particles are changed. In the Metropolis algorithm [61, 62, 63], the acceptance probability is determined by the ratio of the integrands of the partition function [Eq. (51)] after and before the move:

Pacc=min⁡{e−β(ΔL0+ΔUH),1},(53)P_{\mathrm{acc}} = \min\left\{e^{-\beta(\Delta L_{0}+\Delta U_{\mathrm{H}})},1\right\}, \tag*{(53)}

where ΔL0\Delta L_{0} and ΔUH\Delta U_{\mathrm{H}} are the resulting changes in the one-body Landau free energy [Eqs. (38) and (52)] and the effective pair interaction energy [Eq. (39)], respectively. Any trial move that decreases L0+UHL_{0}+U_{\mathrm{H}} is automatically accepted, while a move that increases L0+UHL_{0}+U_{\mathrm{H}} is accepted only with a probability equal to the Boltzmann factor for the change, ΔL0+ΔUH\Delta L_{0}+\Delta U_{\mathrm{H}}. In case of a rejection, the configuration is unchanged, but counts again in the statistics of the sampled ensemble. The Metropolis algorithm satisfies the condition of detailed balance, ensuring sampling from the correct semigrand canonical ensemble of configurations.

It is through the acceptance probability [Eq. (53)] that the total Helmholtz free energy [Eq. (28)] predicted by our theory governs the evolution of the system toward equilibrium. Agreement between our simulations and numerical calculations (Sec. IV) ultimately demonstrates that the Landau free energy is the appropriate thermodynamic potential. Furthermore, any inconsistency between the free energy and the chemical potentials of solvent and crowders would manifest in a discrepancy between results from the two methods, providing powerful consistency and validation checks.

Figure 4 shows a typical snapshot from a multi-particle simulation, in which the microgels are confined to a cubic cell with periodic boundary conditions applied to opposing faces. The particles are initialized on the sites of an FCC lattice, with each assigned an initial radius equal to its fully swollen radius in pure solvent and initial interior crowder volume fraction equal to the reservoir value. We custom-coded all simulations in Java within the Open Source Physics Library [64, 65].

Results and Discussion

Implementing the methods discussed in Sec. III, we numerically solved (29) and independently performed MC simulations to compute equilibrium swelling ratios and interior crowder volume fractions of single microgels immersed in reservoir solutions of crowders of varying concentration. For illustrative purposes, we chose system parameters as follows: microgel dry radius, a0=50a_{0}=50 nm, polymer monomer radius Rmon=0.3R_{\mathrm{mon}}=0.3 nm, segment length l=0.6l=0.6 nm, and cross-link fraction x=0.0005x=0.0005. For suspensions, we chose the dry microgel volume fraction as ϕ0=NmV0/V=0.03\phi_{0}=N_{\mathrm{m}}V_{0}/V=0.03. Although these parameters are representative of many experimental systems, our qualitative conclusions are insensitive to these choices.

In their fully swollen state, such microgels have a mesh (pore) size, lm≃(vp/x)1/3αl_m \simeq(v_p/x)^{1/3}\alpha, and polymer chain radius of gyration Eq. (1) both of roughly 15 nm.

For these particular system parameters, to ensure that the crowders are always smaller than the pores of a deswollen microgel, thus allowing for significant penetration of the network, we chose relatively small crowder radii Rc=1.0,1.25,1.5R_c = 1.0, 1.25, 1.5 nm, corresponding to crowder-to-monomer volume ratios n=37,72,125n = 37, 72, 125. In the case of larger pores, however, which are achievable by reducing the cross-link fraction or synthesizing the gel from polymers composed of longer segments, the crowders may be correspondingly larger. It should be noted that our results for a single microgel can apply equally well to a bulk hydrogel having the same internal architecture (i.e., polymer type and cross-link fraction). Only for a suspension, in which interparticle interactions affect swelling and uptake, do the size and shape of the particles (spherical microgels) become relevant.

In our simulations, we first equilibrated the system for 10510^5 steps before collecting statistics over an additional 5×1055 \times10^5 steps, sampling at intervals of 100 steps, and computing averages to obtain mean values, ⟨α⟩\langle\alpha\rangle and ⟨ϕc⟩\langle\phi_c\rangle. As diagnostics, we computed on the fly the swelling ratio and chemical potentials to ensure that on approaching equilibrium α→constant\alpha\to\text{constant}, μs→μs0\mu_s \to\mu_s^0, and μc→μc0\mu_c \to\mu_c^0.

Figures 5 and 6 show our predictions for the equilibrium swelling ratio and crowder uptake (volume fraction and number) of a bulk hydrogel, or single microgel, immersed in a solution of crowders as a function of crowder concentration in the reservoir. Initially assuming good solvent conditions, we set all of the Flory solvency (χ\chi) parameters to zero. Results are shown for three different crowder radii, roughly an order of magnitude larger than typical solvent molecules, yet small enough to be accommodated within the pores of the gel. To justify our neglect of crowder-crowder correlations, we restrict the reservoir volume fraction to ϕc0≤0.1\phi_c^0 \leq0.1. The near-exact agreement between our numerical calculations and independent simulation data validates our results and provides an important consistency check on our implementation of both methods. Moreover, the experiments of Bradley et al. [30] on PNIPAM microgels in solutions of poly(ethylene oxide) also demonstrate significant dependence of swelling and uptake on crowder (PEO) concentration and molecular weight. However, pH-dependent chemical (H-bonding) interactions between PEO and acrylic acid, as well as the freedom of PEO chains to change conformation, complicate a direct comparison with predictions from our modeling.

We note that, within our current model, our numerical solution of Eqs. (29) and our simulations always yield unique values of ϕs\phi_s and ϕc\phi_c (and thus a single α\alpha value), as confirmed by the observation that contour plots, with ϕc\phi_c as parameter, exhibit only one crossing point. It is conceivable, however, that a model of microgels having a heterogeneous distribution of cross-linkers might exhibit two solutions, which would reflect an interesting bistable response to crowding. Even if such microgels would have only one stable state, however, the interior distribution of crowders would be nonuniform. For example, core-shell microgels would likely harbor a lower concentration of crowders in the core, which is characterized by higher cross-link fraction and smaller pore size, consistent with the trend seen in Figure 8 below.

Figure 8. Numerical results for equilibrium crowder uptake by a single microgel: interior crowder (a) mean volume fraction ⟨ϕc⟩\langle\phi_c\rangle and (b) mean number ⟨Nc⟩\langle N_c\rangle versus reservoir volume fraction ϕc0\phi_c^0 for cross-link fractions x=0.0005,0.001,0.0015x = 0.0005, 0.001, 0.0015. Other system parameters are same as in Fig. 7.

As seen in Figure 5, with increasing reservoir crowder concentration, polymeric gels progressively deswell, the more so the smaller the crowders. This response to crowding is qualitatively akin to that observed in previous studies of linear homopolymers in a good solvent [49], in which case a decrease in radius of gyration was attributed to the entropic cost of crowder penetration, as predicted by polymer field theory. A key difference, however, is that the polymers here are not isolated, but cross-linked into a network.

Equilibrium mean swelling ratio of a bulk hydrogel or single microgel versus reservoir crowder volume fraction

Figure 5. Equilibrium mean swelling ratio ⟨α⟩\langle\alpha\rangle of a bulk hydrogel or single microgel immersed in a solution of crowders versus reservoir volume fraction ϕc0\phi_c^0 for crowder radii, Rc=1.0,1.25,1.5R_c = 1.0, 1.25, 1.5 nm. Other system parameters are microgel dry radius a0=50a_0 = 50 nm and cross-link fraction x=0.0005x = 0.0005. Flory solvency parameters are all set to zero for good solvent conditions. Numerical results (curves) and MC simulation data (symbols) are in near exact agreement. Black dashed curve is for ϵpen=0\epsilon_{\mathrm{pen}} = 0, Rc=1.25R_c = 1.25 nm.

Evidently, the entropic benefit to the network of absorbing solution and swelling, as described by the mixing part of the modified Flory-Rehner free energy [Eq. (28)], becomes suppressed by the entropic cost of accommodating the excluded volume of the crowders. The suppression increases with crowder size and concentration, consistent with the fact that the excluded-volume parameter nn, which appears in the mixing entropy [Eq. (21)] and chemical potentials [Eqs. (31) and (32)], varies with crowder size as Rc3R_c^3. While the penetration free energy contributes to deswelling, mixing entropy proves to be the dominant factor. In fact, setting εpen=0\varepsilon_{\mathrm{pen}}=0 only slightly reduces the deswelling (black dashed curve in Figure 5). From a mechanical perspective, by excluding volume, crowders modify the pressure of both the interior and exterior solutions, perturbing the delicate balance of osmotic pressure that determines the equilibrium swelling of a gel. From a solvency perspective, the crowders act as if an especially poor cosolvent for the polymer. The net result is that crowding-induced contraction of polymers in the network drives deswelling.

As polymeric gels steadily deswell upon concentrating crowders in the reservoir solution, they also absorb crowders. As seen in Figure 6, with increasing ϕc0\phi_c^0, as the chemical potential gradient drives crowders into the gel, the interior crowder volume fraction and number both initially increase. It now becomes apparent that the calculated numbers of crowders inside the gel are so high as to render explicit modeling computationally impractical, thus justifying our implicit model.

Figure 6. Crowder uptake by a single microgel: (a) mean volume fraction ⟨ϕc⟩\langle\phi_c\rangle and (b) mean number ⟨Nc⟩\langle N_c\rangle versus reservoir volume fraction ϕc0\phi_c^0 for crowder radii, Rc=1.0,1.25,1.5R_c = 1.0, 1.25, 1.5 nm. Other system parameters are same as in Fig. 5. Numerical results (curves) and MC simulation data (symbols) are in near exact agreement. Black dashed curves are for ϵpen=0\epsilon_{\mathrm{pen}} = 0, Rc=1.25R_c = 1.25 nm.

Upon further increase in ϕc0\phi_c^0, the number of crowders inside peaks and then decreases, as crowders are expelled into the surrounding solution. Similarly, the volume fraction inside reaches a maximum (ϕc,max⁡<ϕc0\phi_{c,\max}<\phi_c^0), albeit at higher ϕc0\phi_c^0 owing to the decrease in gel volume. Especially notable is the strong sensitivity to crowder size displayed in Figure 6. With increasing RcR_c, the uptake increases more gradually and peaks at lower values of ϕc0\phi_c^0. This trend is again consistent with the excluded-volume factor nn in the mixing entropy varying as Rc3R_c^3. Thus, absorption of larger crowders is suppressed by excluded-volume penalties in both the penetration free energy and mixing entropy.

The remarkable nonmonotonic response of polymeric gels to crowding reflects a compromise between mixing, elastic, and penetration contributions to the free energy. At lower reservoir concentrations, mixing entropy strongly favors uptake, while at higher concentrations, mixing entropy and penetration free energy both drive expulsion. Further analysis reveals that competition between the solvent and crowder terms in the mixing free energy dominates the uptake and swelling responses. Neglecting the penetration free energy (setting εpen=0\varepsilon_{\mathrm{pen}}=0) still leads to an uptake maximum, but at a higher concentration (black dashed curves in Figure 6). We have also checked that omitting the swelling ratio factor cc from the polymer-to-crowder size ratio qq in Eq. (27) would produce only a slight enhancement of swelling and uptake, without qualitatively changing the behavior.

The observed swelling and uptake behavior highlights the complex nonlinear relationship between reservoir concentration and capacity for uptake by polymeric gels. Tuning this relationship may be important for engineering and controlling the uptake and release of nanoscale particles by such gels. We emphasize that our model of macromolecular crowding highlights the role of excluded volume in the mixing entropy [Eq. (21)]. However, attractive or repulsive interactions between crowders and the gel can be incorporated through the Flory solvency parameters (χij)(\chi_{ij}) in the mean mixing energy [Eq. (24)].

To illustrate the dependence of swelling and uptake on cross-link fraction and solvent quality, we show in Figure 7-Figure 12 our numerical results for {α,ϕc}\{\alpha,\phi_c\} and {Nc}\{N_c\} over ranges of xx and χ\chi. With increasing xx, as the polymer network stiffens, the equilibrium mean swelling ratio decreases (Figure 7), implying smaller mean pore size. Correspondingly, the uptake of crowders is reduced, as indicated by lower values of interior crowder mean volume fraction and number (Figure 8). These trends are consistent with a shrinking polymer chain RgR_g and rising penetration free energy with increasing cross-link fraction.

Numerical results for equilibrium mean swelling ratio

Figure 7. Numerical results for equilibrium mean swelling ratio ⟨α⟩\langle\alpha\rangle of a bulk hydrogel or single microgel immersed in a solution of crowders versus reservoir volume fraction ϕc0\phi_c^0 for microgel dry radius a0=50a_0 = 50 nm, crowder radius Rc=1R_c = 1 nm, and cross-link fractions x=0.0005,0.001,0.0015x = 0.0005, 0.001, 0.0015. Flory solvency parameters are all set to zero for good solvent conditions.

Figure 12. Numerical results for equilibrium crowder uptake by a single microgel: interior crowder (a) mean volume fraction ⟨ϕc⟩\langle\phi_c \rangle and (b) mean number ⟨Nc⟩\langle N_c \rangle versus reservoir volume fraction ϕc0\phi_c^0 for same system parameters as in Figure 11.

Numerical results for equilibrium mean swelling ratio versus reservoir crowder volume fraction

Figure 11. Numerical results for equilibrium mean swelling ratio ⟨α⟩\langle\alpha\rangle of a bulk hydrogel or single microgel immersed in a solution of crowders versus reservoir volume fraction ϕc0\phi_c^0 for same system parameters as in Figure 9. Solvency parameters are set to χps=χsc=0\chi_{ps} = \chi_{sc} = 0 and χpc=0,−0.2,0.2\chi_{pc} = 0, -0.2, 0.2 to mimic effective polymer-crowder attractive and repulsive interactions.

Upon reducing solvent quality for all species, achieved by increasing all solvency parameters (assuming χps=χpc=χsc\chi_{ps}=\chi_{pc}=\chi_{sc}), similar trends are observed for swelling and uptake as the polymer chains compactify (Figure 9 and Figure 10). However, if only the polymer-crowder solvency parameter is varied [Fig. 12], then for χpc<0\chi_{pc}<0 (mimicking effective polymer-crowder attraction) deswelling is reduced and uptake is enhanced, while for χpc>0\chi_{pc}>0 (effective repulsion) the trends are reversed—deswelling is enhanced and uptake is reduced.

Numerical results for equilibrium mean swelling ratio versus reservoir crowder volume fraction

Figure 9. Numerical results for equilibrium mean swelling ratio ⟨α⟩\langle\alpha\rangle of a bulk hydrogel or single microgel immersed in a solution of crowders versus reservoir volume fraction ϕc0\phi_c^0 for microgel dry radius a0=50a_0 = 50 nm, cross-link fraction x=0.0005x = 0.0005, and crowder radius Rc=1R_c = 1 nm. All three solvency parameters are assumed equal (χps=χpc=χsc\chi_{ps} = \chi_{pc} = \chi_{sc}) and set to χij=0,0.1,0.2\chi_{ij} = 0, 0.1, 0.2.

Figure 10. Numerical results for equilibrium crowder uptake by a single microgel: interior crowder mean (a) volume fraction ⟨ϕc⟩\langle\phi_c \rangle and (b) number ⟨Nc⟩\langle N_c \rangle versus reservoir volume fraction ϕc0\phi_c^0 for same system parameters as in Figure 9.

Within our model, we can also mimic mutual effective attractions between crowders [ϵcc<0[\epsilon_{cc} < 0 in Eq. (23)], added to excluded-volume repulsion, by assigning positive values to both χsc\chi_{sc} and χpc\chi_{pc}. As shown in Figure 13, with increasing χsc\chi_{sc} and χpc\chi_{pc}, uptake of crowders is steadily suppressed. This predicted trend is consistent with the energetic cost incurred by crowders mixing with polymer inside the microgel instead of with one another outside.

Figure 13. Numerical results for equilibrium crowder uptake by a single microgel: interior crowder (a) mean volume fraction ⟨ϕc⟩\langle\phi_c\rangle and (b) mean number ⟨Nc⟩\langle N_c\rangle versus reservoir volume fraction ϕc0\phi_c^0 for same system parameters as in Figure 9. Solvency parameters are set to χps=0\chi_{ps}=0 and χsc=χpc=0,0.1,0.2\chi_{sc}=\chi_{pc}=0, 0.1, 0.2 to mimic crowder-crowder effective attraction.

To investigate the influence of interparticle interactions on single-microgel and bulk structural properties in a suspension, we performed simulations of Nm=108N_m = 108 interacting microgels dispersed in a solution of crowders. In the Hertz potential [Eqs. (40)–(42)], we set the calibration factor to Yc=10Y_c = 10, yielding a Young’s modulus of Y≃50Y \simeq50 kPa for relatively stiff microgels [66]. From our simulations, we averaged over microgels and MC steps to compute equilibrium values of the mean swelling ratio and interior number of crowders. Figure 14 and Figure 15 compare our data for an isolated microgel with corresponding data for a suspension. At lower reservoir concentrations (ϕc0<0.04\phi_c^0 < 0.04), where for these system parameters the particles are so swollen that they remain in an FCC crystal configuration, Hertzian interactions between neighbors evidently contribute additional deswelling, especially when the microgel density exceeds that of close-packed spheres. Concomitantly, interparticle interactions reduce uptake of crowders. At higher reservoir concentrations (ϕc0>0.05\phi_c^0 > 0.05), the microgels deswell to the point that interparticle correlations become sufficiently weak as to destabilize (melt) the crystal. In this range, microgels in a suspension have essentially the same swelling ratio and crowder uptake as an isolated microgel, whose deswelling is purely crowder-induced.

Figure 14. MC simulation data for equilibrium mean swelling ratio ⟨α⟩\langle\alpha\rangle of a single isolated microgel (red circles) and of Nm=108N_m=108 microgels (black squares) dispersed in a solution of crowders of radius Rc=1R_c=1 nm versus reservoir volume fraction ϕc0\phi_c^0. The dry volume fraction of microgels in the suspension is ϕ0=0.03\phi_0=0.03. Numerical results are also shown for Nm=1N_m=1 (curve). Other system parameters are same as in Figure 5.

Figure 15. MC simulation data for equilibrium mean number of crowders ⟨Nc⟩\langle N_c\rangle inside an isolated microgel (red circles) and for Nm=108N_m=108 microgels (black squares) dispersed in a solution of crowders of radius Rc=1R_c=1 nm versus reservoir volume fraction ϕc0\phi_c^0. Numerical results are also shown for Nm=1N_m=1 (curve). Other system parameters are same as in Figure 14.

From our simulations, we further computed the radial distribution function g(r)g(r) by histogramming the center-

to-center separation rr between pairs of particles and averaging over configurations. A measure of pair correlations, g(r)g(r) is proportional to the conditional probability of finding a particle at radial distance rr away from another particle [53]. We also computed the orientationally averaged static structure factor,

S(q)=1+2Nm∑i<jNm⟨sin⁡(qrij)qrij⟩,(54)S(q) = 1 + \frac{2}{N_m}\sum_{i<j}^{N_m}\left\langle\frac{\sin(qr_{ij})}{qr_{ij}}\right\rangle, \tag*{(54)}

which is proportional to the intensity of scattered light, qq being the magnitude of the scattered wave vector, and to the Fourier transform of g(r)g(r).

The phenomenon of crowding-induced melting is confirmed by comparing radial distribution functions and static structure factors of suspensions at different reservoir concentrations. As shown in Figure 16 and Figure 17, at ϕc0=0.05\phi_c^0 = 0.05, g(r)g(r) and S(q)S(q) exhibit significant structure, reflecting strong pair correlations indicative of long-range periodic ordering in a stable or metastable crystalline phase. At higher ϕc0\phi_c^0, these correlations weaken as the

Figure 16. MC simulation data for radial distribution function g(r)g(r) vs. radial distance rr (units of microgel dry radius) for Nm=108N_m = 108 microgels dispersed in crowder solutions of reservoir concentrations ϕc0=0.05\phi_c^0 = 0.05 (red), 0.07 (blue), and 0.1 (black). The volume fraction of the microgels in their dry state is ϕ0=0.03\phi_0 = 0.03. Other system parameters are same as in Figure 5. With increasing ϕc0\phi_c^0, the particles deswell, as reflected by the shift of the main peak toward shorter distances. As a result, pair correlations weaken, indicated by lower and broader peaks, and the suspension transitions from a solid (FCC crystal) to a fluid phase.

Figure 17. MC simulation data for static structure factor S(q)S(q) as a function of scattered wave vector qq (units of inverse microgel dry radius) for Nm=108N_m = 108 microgels dispersed in crowder solutions of reservoir concentrations ϕc0=0.05\phi_c^0 = 0.05 (solid red curve), 0.07 (long-dashed blue), and 0.1 (short-dashed black). The volume fraction of the microgels in their dry state is ϕ0=0.03\phi_0 = 0.03. Other system parameters are same as in Figure 5. Consistent with the results for g(r)g(r) shown in Figure 16, with increasing ϕc0\phi_c^0, pair correlations become weaker and the solid phase is destabilized. Thus, crowder uptake weakens correlations and induces melting.

crystal gives way to the fluid phase. We emphasize that the melting transition observed here, at constant temperature and microgel density, is driven entirely by crowding-induced deswelling, and not by any crowder-microgel correlations, which are absent from our coarse-grained model.

Conclusions

By developing and implementing a theory of cross-linked polymer networks immersed in a solution of macromolecular crowding agents, we modeled the response of hydrogels and microgels to volume-excluding crowders. From a free energy composed of mixing, elastic, and penetration entropy contributions, as well as mixing energy in a mean field approximation, we derived the chemical potentials of solvent and crowders. By equating chemical potentials of both species inside and outside the gel, we computed the equilibrium swelling ratio and crowder uptake concentration over a range of reservoir solution concentration, crowder size, cross-link fraction, and solvency parameters.

To validate our numerical methods, we also performed Monte Carlo simulations with trial changes in swelling ratio and interior crowder concentration. The simulations effectively minimize the Landau free energy of the system, which is thermodynamically equivalent to equating chemical potentials in the interior and exterior phases. By adding a Hertzian elastic pair potential to our model, we simulated suspensions of microgels and computed the influence of interparticle interactions on swelling and uptake and of crowding on microgel pair correlations.

Our results demonstrate how macromolecular crowders that exclude volume, when absorbed from solution by polymeric gels, can induce deswelling, ever more strongly with decreasing crowder size. Our analysis reveals that mixing entropy favors uptake of crowders at low reservoir concentrations, but that with increasing concentration, uptake first peaks and then declines. The peak value of the uptake concentration decreases with increasing crowder size. We further conclude that hard crowders, by driving deswelling, weaken pair correlations in microgel suspensions and can induce melting of a crystal.

Finally, we summarize the scope and limitations of our theoretical framework. First, our coarse-grained model of polymeric gels neglects microscopic details in favor of capturing the dominant physics. Second, our theory of microgel-crowder solutions is based on a lattice model and a Flory-Huggins mean-field approximation for the mixing free energy. While our theory approximately accounts for crowder-crowder pair correlations, it neglects many-body correlations and so applies only to low crowder concentrations. Third, we model crowders as idealized hard spheres that exclude volume and represent them by a continuous density field rather than as discrete particles. Fourth, we describe swelling of gels by the Flory-Rehner mean-field theory, which assumes affine swelling and Gaussian polymer chains, and interactions between microgels by a simple elastic Hertzian potential. Although we focus here on uniformly cross-linked microgels, our theory can be naturally extended to heterogeneous microgels, for example, by means of multi-Hertzian interaction models [37, 60]. Lastly, our use of polymer field theory to account for the free energy cost of crowder penetration neglects molecular details of polymers.

Our computationally efficient modeling approach captures the key physical features of these complex materials and our results highlight and demonstrate the important role of excluded-volume effects in macromolecular crowding of polymeric gels. The accuracy of our approximations and predictions may be tested via simulations of molecular-scale models [36, 67, 68] and by experiments that add crowding agents to gel solutions and measure swelling and uptake response. A natural extension of our model will be to incorporate flexible crowders with freedom to change conformation in response to confinement within the pores of a gel. Future work may also include adapting the modeling to describe ionic microgels [69, 70, 71, 72, 73, 74] and study the interplay between crowding-induced and charge-induced swelling.

Acknowledgments

This work was supported by the National Science Foundation under Grant No. DMR-1928073.

Author Declaration

Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

Mahesh Aryal: formal analysis (equal); methodology (equal); calculations (equal); software (equal); data curation (equal); visualization (equal); validation (equal); writing – original draft (equal); writing – review and editing (supporting).

Alan R. Denton: conceptualization (lead); project administration (lead); resources (lead); supervision (lead); formal analysis (equal); methodology (equal); calculations (equal); software (equal); data curation (equal); visualization (equal); validation (equal); writing – original draft (equal); writing – review and editing (lead).

Data Availability

Data supporting the findings of this study are available from the corresponding author upon reasonable request.

References

  1. [1]L. A. Lyon and M. J. Serpe, eds., Hydrogel Micro and Nanoparticles (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2012).
  2. [2]A. Fernández-Nieves, H. Wyss, J. Mattsson, and D. A. Weitz, eds., Microgel Suspensions: Fundamentals and Applications (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2011).
  3. [3]J. Riest, P. Mohanty, P. Schurtenberger, and C. N. Likos, Z. Phys. Chem. 226, 711 (2012).
  4. [4]M. Karg, A. Pich, T. Hellweg, T. Hoare, L. A. Lyon, J. J. Crassous, D. Suzuki, R. A. Gumerov, S. Schneider, I. I. Potemkin, and W. Richtering, Langmuir 35, 6231 (2019).
  5. [5]A. Scotti, M. F. Schulte, C. G. Lopez, J. J. Crassous, S. Bochenek, and W. Richtering, Chem. Rev. 122, 11675 (2022).
  6. [6]M. Malmsten, in Microgel Suspensions: Fundamentals and Applications, edited by A. Fernández-Nieves, H. Wyss, J. Mattsson, and D. A. Weitz (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2011) pp. 375–405.
  7. [7]F. A. Plamper and W. Richtering, Acc. Chem. Res. 50, 131 (2017).
  8. [8]F. Scheffold, Nat. Comm. 11, 4315 (2020).
  9. [9]L. Schäfer, Excluded Volume Effects in Polymer Solutions as Explained by the Renormalization Group (Springer, Berlin, 1999).
  10. [10]A. P. Minton, Biopolymers 20, 2093 (1981).
  11. [11]A. P. Minton, G. C. Colclasure, and J. C. Parker, Proc. Natl. Acad. Sci. 89, 10504 (1992).
  12. [12]A. P. Minton, Current Opin. Biotechnol. 8, 65 (1997).
  13. [13]A. P. Minton, Current Opin. Struct. Biol. 10, 34 (2000).
  14. [14]A. P. Minton, Biophys. J. 88, 971 (2005).
  15. [15]R. J. Ellis, Current Opin. Struct. Biol. 11, 114 (2001).
  16. [16]M. Cheung, D. Klimov, and D. Thirumalai, Proc. Natl. Acad. Sci. 102, 4753 (2005).
  17. [17]R. Phillips, J. Kondev, and J. Theriot, Physical Biology of the Cell (Garland Science, New York, 2009).
  18. [18]A. H. Elcock, Current Opin. Struct. Biol. 20, 1 (2010).
  19. [19]R. Hancock, in Genome Organization and Function in the Cell Nucleus, edited by K. Rippe (Wiley-VCH, Weinheim, 2012) pp. 169–184.
  20. [20]A. R. Denton, in New Models of the Cell Nucleus: Crowding and Entropic Forces and Phase Separation and Fractals, edited by R. Hancock and K. W. Jeon (Academic Press, UK, 2013) pp. 27–72.
  21. [21]P. J. Flory and J. Rehner, J. Chem. Phys. 11, 512 (1943).
  22. [22]P. J. Flory and J. Rehner, J. Chem. Phys. 11, 521 (1943).
  23. [23]P. J. Flory, Principles of Polymer Chemistry (Cornell University Press, 1953).
  24. [24]M. Doi, Soft Matter Physics (Oxford University Press, 2013).
  25. [25]E. Favre, Q. Nguyen, D. Sacco, A. Moncuy, and R. Clement, Chem. Eng. Comm. 140, 193 (1995).
  26. [26]O. Okeowo and J. R. Dorgan, Macromol. 39, 8193 (2006).
  27. [27]R. V. Godbole, F. Khabaz, R. Khare, and R. C. Hedden, J. Phys. Chem. B 121, 7963 (2017).
  28. [28]J. P. de Souza and H. A. Stone, J. Chem. Phys. 161, 044902 (2024).
  29. [29]A. F. Routh, A. Fernandez-Nieves, M. Bradley, and B. Vincent, J. Phys. Chem. B 110, 12721 (2006).
  30. [30]M. Bradley, J. Ramos, and B. Vincent, Langmuir 21, 1209 (2005).
  31. [31]I.-L. Lu, T.-W. Yu, T.-I. Liu, H.-H. Chen, Y.-C. Yang, C.-L. Lo, C.-Y. Wang, and H.-C. Chiu, Adv. Healthc. Mater. 11, 2201140 (2022).
  32. [32]S. Bulut, S.-H. Jung, T. Bissing, F. Schmitt, M. Bund, S. Braun, and A. Pich, Small 19, 2303783 (2023).
  33. [33]M. Urich and A. R. Denton, Soft Matter 12, 9086 (2016).
  34. [34]T. J. Weyer and A. R. Denton, Soft Matter 14, 4530 (2018).
  35. [35]M. E. Brito, A. R. Denton, and G. Nägele, J. Chem. Phys. 151 (2019).
  36. [36]G. Del Monte and E. Zaccarelli, Phys. Rev. X 14, 041067 (2024).
  37. [37]A. Moncho-Jordá, A. Cuetos, M. A. Fernandez-Rodriguez, J. Dzubiella, and A. Patti, Macromolecules 58, 10659 (2025).
  38. [38]R. Pelton and T. Hoare, in Microgel Suspensions: Fundamentals and Applications, edited by A. Fernández-Nieves, H. Wyss, J. Mattsson, and D. A. Weitz (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2011) pp. 3–32.
  39. [39]R. Acciaro, T. Gilányi, and I. Varga, Langmuir 27, 7917 (2011).
  40. [40]P.-G. de Gennes, Scaling Concepts in Polymer Physics (Cornell, Ithaca, 1979).DOI
  41. [41]M. Rubinstein and R. H. Colby, Polymer Physics (Oxford University Press, 2003).DOI
  42. [42]W. Kuhn, Kolloid-Zeitschrift 68 (1934).
  43. [43]J. Rudnick and G. Gaspari, Science 237, 384 (1987).
  44. [44]M. Murat and K. Kremer, J. Chem. Phys. 108, 4340 (1998).
  45. [45]E. Eurich and P. Maass, J. Chem. Phys. 114, 7655 (2001).
  46. [46]W. K. Lim and A. R. Denton, J. Chem. Phys. 141, 114909 (2014).
  47. [47]W. K. Lim and A. R. Denton, J. Chem. Phys. 144, 024904 (2016).
  48. [48]W. K. Lim and A. R. Denton, Soft Matter 12, 2247 (2016).
  49. [49]W. J. Davis and A. R. Denton, J. Chem. Phys. 149, 2018.
  50. [50]S. J. Sciutto, J. Phys. A: Math. Gen. 27, 7015 (1994).
  51. [51]S. J. Sciutto, J. Phys. A: Math. Gen. 29, 5455 (1996).
  52. [52]C. G. Lopez and W. Richtering, Soft Matter 13, 8271 (2017).
  53. [53]J. P. Hansen and I. R. McDonald, Theory of Simple Liquids, 2nd ed. (Academic, London, 1986).
  54. [54]H. C. Marques Fernandes, Y. Levin, and J. J. Arenzon, Phys. Rev. E 75, 052101 (2007).
  55. [55]E. Eisenriegler, A. Hanke, and S. Dietrich, Phys. Rev. E 54, 1134 (1996).
  56. [56]A. Hanke, E. Eisenriegler, and S. Dietrich, Phys. Rev. E 59, 6853 (1999).
  57. [57]E. Eisenriegler, J. Chem. Phys. 113, 5091 (2000).
  58. [58]E. Eisenriegler, A. Bringer, and R. Maassen, J. Chem. Phys. 118, 9083 (2003).
  59. [59]L. D. Landau and E. M. Lifshitz, Theory of Elasticity, 3rd ed. (Elsevier, Amsterdam, 1986).
  60. [60]J.-M. Meijer, L. Rోవigatti, E. Zaccarelli, and P. Schurtenberger, Nat. Commun. 9, 5039 (2018).
  61. [61]D. Frenkel and B. Smit, Understanding Molecular Simulation: From Algorithms to Applications, 2nd ed. (Academic, London, 2001).
  62. [62]K. Binder and D. W. Heermann, Monte Carlo Simulation in Statistical Physics, 5th ed. (Elsevier, 2010).
  63. [63]D. P. Landau and K. Binder, A Guide to Monte Carlo Simulations in Statistical Physics, 5th ed. (Cambridge University Press, 2021).
  64. [64]H. Gould, J. Tobochnik, and W. Christian, Introduction to Computer Simulation Methods, 3rd ed. (Addison Wesley, 2007).
  65. [65]W. Christian, Open Source Physics: A User’s Guide with Examples (Addison Wesley, 2006).
  66. [66]J. J. Liétor-Santos, B. Sierra-Martín, and A. Fernández-Nieves, Phys. Rev. E 84, 060402(R) (2011).
  67. [67]N. Gnan, L. Rოვigatti, M. Bergman, and E. Zaccarelli, Macromol. 50, 8777 (2017).
  68. [68]S. Marín-Aguilar and E. Zaccarelli, Soft Matter 21, 8669 (2025).
  69. [69]Y. Levin, Phys. Rev. E 65, 036143 (2002).
  70. [70]A. R. Denton, in Electrostatics of Soft and Disordered Matter, edited by D. S. Dean, J. Dobnikar, A. Naji, and R. Podgornik (Pan Stanford, Singapore, 2014) pp. 201–215.
  71. [71]T. Colla, C. N. Likos, and Y. Levin, J. Chem. Phys. 141, 234902 (2014).
  72. [72]A. Moncho-Jordá, J. Chem. Phys. 139, 064906 (2013).
  73. [73]I. Adroher-Benítez, S. Ahualli, A. Martín-Molina, M. Quesada-Pérez, and A. Moncho-Jordá, Macromol. 48, 4645 (2015).
  74. [74]A. Moncho-Jordá and J. D ज़ubiella, Phys. Chem. Chem. Phys. 18, 5372 (2016).

Paper details

vlnc:W_btfzejhw
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents