Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Modeling of stimuli-responsive nanoreactors: rational rate control towards the design of colloidal enzymes

Modeling of stimuli-responsive nanoreactors: rational rate control towards the design of... MSDE View Article Online PAPER View Journal Repertoire Builder: high-throughput structural Modeling of stimuli-responsive nanoreactors: rational modeling of B and T cell receptors† Cite this: DOI: 10.1039/c9me00020h rate control towards the design of colloidal enzymes a ab ab ab Dimitri Schritt,‡§ Songling Li,§ John Rozewicki,§ Kazutaka Katoh, a b c de Matej Kanduc, ˇ Won Kyu Kim, Rafael Roa and Joachim Dzubiella a c c ab Kazuo Yamashita,¶ Wayne Volkmuth, Guy Cavet and Daron M. Standley Repertoire Builder (https://sysimm.org/rep_builder/) is a method for generating atomic-resolution, three- dimensional models of B cell receptors (BCRs) or T cell receptors (TCRs) from their amino acid sequences. It is currently capable of handling batches of up to 10 sequences in approximately 30 minutes. This per- In modern applications of heterogeneous liquid-phase nanocatalysis, the catalysts (e.g., metal formance was achieved by applying a multiple sequence alignment extension technique originally devel- nanoparticles) need to be typically affixed to a colloidal carrier system for stability and easy han- oped for phylogenetic analysis to the template selection problem of complementarity determining region dling. “Passive carriers” (e.g., simple polyelectrolytes) serve for a controlled synthesis of the (CDR) loops. Under comparable conditions, average all-atom root-mean square deviations (RMSDs) from nanoparticles and prevent coagulation during catalysis. Recently, however, hybrid conjugates of experimentally-determined structures of CDRH3 loops in BCRs were significantly lower than tested third- nanoparticles and synthetic thermosensitive polymers have been developed that enable to change party high-throughput modeling methods, including ABodyBuilder, PigsPro, and LYRA. For TCRs, similar the catalytic activity of the nanoparticles by external triggers. In particular, nanoparticles embed- trends were observed when Repertoire Builder was compared with TCRmodel and LYRA. We also found ded in a stimuli-responsive network made from poly(N-isopropylacrylamide) (PNIPAM) have be- that Repertoire Builder model errors were, in general, lower than those produced by our earlier Kotai Anti- come the most-studied examples of such hybrids. It has been demonstrated that the permeability body Builder, even when CDRH3 loop refinement was used. However, in a subset of cases, which could be of the polymer network and thus the reactant flux can be switched and controlled by external distinguished by poor Repertoire Builder scores, refinement by Kotai Antibody Builder or Rosetta Antibody, stimuli. Such “active carriers” may thus be viewed as true nanoreactors that open up new design Received 12th February 2019, both of which utilize extensive structural sampling, improved the third heavy chain CDR (CDRH3) RMSD on routes in nano-catalysis and elevate synthesis to create highly selective, programmable “colloidal Accepted 21st May 2019 average. Taken together, these results indicate that the MSA extension approach used by Repertoire Builder enzymes”. However, only a comprehensive understanding of these materials on all time and resulted in a favorable balance between speed and accuracy when compared to alternative methods. Fur- DOI: 10.1039/c9me00020h length scales can lead to a rational design of future, highly functional materials. Here we review thermore, we conclude that more sensitive scoring, rather than extended structural sampling, is needed to the current state of the theoretical and multi-scale simulation approaches, aiming at a fundamental rsc.li/molecular-engineering further improve the accuracy of BCR and TCR modeling. understanding of these nanoreactors. In particular, we summarize a theoretical approach for re- action rates of surface-catalyzed bimolecular reactions in responsive nanoreactors in terms of the key material parameters, the polymer shell permeability P and the reactant partition ratio K . We Design, System, Application discuss recent computer simulation studies of both atomistic and coarse-grained polymer models Repertoire Builder is a tool for building 3D models of B cell receptors (BCRs) or T cell receptors (TCRs) to atomic resolution. The strategy used by Repertoire Builder is an application of the multiple in which sequence thesealignment quantities (MSA) haextension ve beenfeature character of the iz MAFFT ed insoftware. some detail. The particula Wer conclude application h with ere isan outlook to represent structural templates for each complementarity-determining region (CDR) of a given length by a single MSA. By repeatedly applying the MSA ex- on selected open questions and future theoretical challenges in nanoreactor modeling. tension method, a complete set of templates that covers the 3 CDRs and 1 framework for each chain can be obtained. The input must be either paired (heavy and light) or unpaired (heavy or light) chain amino acid sequence of the variable region for the receptor in question (BCR or TCR). The immediate application of Repertoire Builder is to render 3D models in a high-throughput and accurate manner, in order to allow structure-based analyses for BCR or 1 TCR Intr sequ oduction ence data. Because the volume of such data is currently growing exponentially, Repertoire Builder represents a unique approach to large-scale 1–12 10–14 reactions to biosensors for the diagnosis of diseases. Ex- BCR or TCR repertoire data analysis. Synthetic nanoreactors are an emerging and promising new nan- amples of natural nanoreactors are lipid-based membranes (e.g., otechnology for liquid-phase heterogenous catalysis. In these liposomes), cage-like proteins (e.g., ferritins), protein-based bac- 11–13,15 nanoreactors, the catalysts are confined in a permeable nanos- terial microcompartments, and viruses. Artificial nanore- Introduction tructure, which acts as a carrier and can be used to shelter and actors (based on spherical polyelectrolyte brushes, dendrimers, control the catalytic processes. In particular, the catalysis can be Recent ligands, single-cell or even resolution DNA) aresequencing simpler than technologies the naturalcan ones elu- and thus 4–12,16,17 Immunology Frontier Research Center, Osaka University, 3-1 Yamadaoka, Suita, made selective and responsive if the nanoreactor permeability dif- cidate easier the to natively controlpaired for targeted (heavy– applications. light) B cell receptor (BCR) Osaka 565-0871, Japan. E-mail: standley@biken.osaka-u.ac.jp ferentiates among molecules and can be modulated by external and T cell receptor (TCR) sequences in a high-throughput Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, 1–14 In particular, nanoreactors containing metal nanoparticles have stimuli. These nanoreactors can be used for a large variety manner. Although the coverage has not yet reached that of Suita, Osaka 565-0871, Japan 4–9,18–21 emerged as a promising catalytic system. For exam- bulk sequencing methods, state-of-the-art paired sequencing of applications, ranging from analytical tools to study chemical Atreca Inc, 500 Saginaw Drive, Redwood City, CA, 94063-4750, USA ple, gold becomes an active catalyst when divided down to the † Electronic supplementary information (ESI) available. See DOI: 10.1039/ platforms can yield thousands of unique receptor sequences 22–27 c9me00020h nanophase. However, the handling of the particles in the in a single experiment. Because of the critical role of B and T ‡ Current address: Department of Biosciences and Nutrition, Karolinska cellsliquid in the phase prevention is an important or progression problem: The of disease, surface ofnew the parti- Jožef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia Institutet. methods cles should to functionally be easilyanalyze accessible emerging for the sequence mixture ofdata theare reactants. Korea Institute for Advanced Study, 85 Hoegiro, Seoul 02455, Republic of Korea. § Equal contribution. Departamento de Física Aplicada I, Facultad de Ciencias, Universidad de Málaga, needed. Structural modeling can, in principle, contribute to This condition would require the nanoparticles to be freely sus- ¶ Current address: KOTAI Biotechnologies Inc, 3-1 Yamadaoka, Suita, Osaka Campus de Teatinos s/n, E-29071 Málaga, Spain such functional analysis, since structures allow physical and 565-0871. pended in the solution, and coagulation or any type of Ostwald Research Group for Simulations of Energy Materials, Helmholtz-Zentrum Berlin, ripening of the nanoparticles should not occur during the cat- Hahn-Meitner-Platz 1, D-14109 Berlin, Germany alytic reaction. Also, leaching of metal or loss of nanoparticles Institut für Physik, Albert-Ludwigs-Universität Freiburg, Hermann-Herder-Str. 3, D- This journal is © The Royal Society of Chemistry 2019 Mol. Syst. Des. Eng. 79104 Freiburg, Germany from the carrier must be prevented to ensure a meaningful and Journal Name, [year], [vol.], 1–17 | 1 Open Access Article. Published on 21 June 2019. Downloaded on 6/27/2019 2:27:17 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. arXiv:2003.09001v1 [physics.chem-ph] 19 Mar 2020 repeated use of the catalyst. The latter requirements necessitate bulk solution a suitable carrier that ensures a safe and repeated handling of 28,29 the nanoparticles. It was demonstrated that suitable carrier 30,31 32–42 systems include colloidal particles, dendrimers, meso- 43–45 46,47 porous materials, spherical polyelectrolyte brushes, and C other systems structured on a length scale between one and a polymer few hundred nm. shell In recent years, the concept of such carrier systems has been further advanced with the synthesis of hydrogel-based nanoreac- tors, for which rate control by external stimuli has become pos- B ΔG(r ) 6,7,10,12,15,18,49–57 sible. Thermosensitive hydrogels made from ΔG a network of poly(N-isopropylacrylamide) (PNIPAM) and its 4–8,18–21,58–67 copolymers provide a good example: Typical col- loidal carrier architectures are of core-shell or yolk-shell type where the polymer gel constitutes a permeable shell around a in solid core (core-shell) or around a hollow void (yolk-shell). The core can be the nanoparticle itself, cf. Fig. 1, or, for example, a NR polystyrene core. The catalytic nanoparticles can be located dur- ing synthesis in a well controlled fashion, e.g., into the voids, Fig. 1 Bimolecular reactions in core-shell nanoreactors. (A) Two reac- onto the cores, or distributed within the polymer shell. Some- tants, A and B, diffusing from a bulk solution, generate a product, C, in the times simply a pure hydrogel (nano- or microgel) particle is the proximity of a catalyst nanoparticle (central yellow sphere) embedded in carrier for the catalysts. In this case carrier and polymer shell in a PNIPAM polymer network. (B) Schematic representation of a core-shell our context are essentially the same. A survey of selected but very nanoreactor. A nanoparticle of radius R is embedded in a spherical shell of outer radius R . The shell permeability depends on the diffusivity, typical experimental architectures and results for polymer-based NR D(r), and on the transfer free energy profiles, DG(r). We model both as nanoreactor carriers is provided in Tab. 1. step functions with values D and DG inside, and D and zero reference in 0 The responsive polymer shell is in a swollen hydrophilic state at outside the shell, respectively. low temperature, but sharply collapses into a rather hydrophobic state above the critical solution temperature. The sharp volume 69–75 transition of the gel is reversible and depends on the temper- 76,77 ature, or more general, solvent quality. This has substantial responsive nanoreactors can be described by combining a thermo- influence on reactant partitioning close to the catalysts as well as dynamic two-state model for the polymer volume transition with reactant transport towards it. Hence, there are two key roles of the appropriate reaction–diffusion equations. In particular, the the polymer shell. On the one hand, the shell acts as a integral important effect of a change in the shell permeability on the reac- part of the whole carrier that protects nanoparticles from aggre- tants approach to the catalyst’s surface can be described by theory 7,78,82,83 gation and hinders chemical degradation processes, e.g., oxida- of diffusion through an energy landscape, in the spirit of 84–88 tion. On the other hand, the polymer ability to switch between Debye–Smoluchowski diffusion-controlled rate theory. Im- states with different physicochemical properties upon changes in portantly, the latter also takes into account the local reactant environmental parameters, e.g., temperature, pH, or concentra- concentration, i.e., the partitioning inside the polymer shell close tion of certain solutes, provides a handle to actively control the to the catalyst. Recently, we have presented an extended theory nanoreactor’s catalytic properties. of diffusion-limited bimolecular reactions in nanoreactors, which A quantitative study and understanding of a nanoreactor re- can be employed to rationally design the activity and selectivity 78,83 quires kinetic data measured with the highest precision possible. of a nanoreactor. The main result of our consideration was Up to now, most of the testing of the catalytic activity of nanopar- the following formula for the total catalytic rate in bimolecular ticles in aqueous phase has been done using the reduction of 4- reactions in core-shell nanoreactors (cf. Fig. 1): 79 36 nitrophenol by borohydride. Pal et al. and Esumi et al. have 1 k k A B been the first who have demonstrated the usefulness of this re- D D k = + k A + k B tot D D 2 k action. In the meantime, the reduction of 4-nitrophenol has be- R come the most used model reaction for the quantitative testing and analyzing of the catalytic activity of nanoparticles in the liq- k Ak B D D + k A + k B 4k Ak B : (1) 80,81 D D D D uid phase. Further examples of catalytic reactions in aqueous solution studied in this system are the reductions of nitrobenzene 6,7,18,19 A B and hexacyanoferrate (III) by borohydride ions and the Here, k A(P ) and k B(P ) are the diffusion rates of the reactants D D 8 i decomposition of methyl orange under visible light. A and B, which explicitly depend on the shell permeability P , and A B All the aforementioned examples deal with surface-catalyzed k (K ;K ) is the surface reaction rate, explicitly depending on partitioning K as defined below. bimolecular reactions, being a very common type. As pointed out 5–7,78 before, pseudo-unimolecular surface-catalyzed reactions in In general, permeability of a material defines the ability of the Journal Name, [year], [vol.], 2 | 1–17 Table 1 Survey of selected publications on responsive nanoreactor catalytic experiments with different architectures. Polymer abbreviations: poly(N-isopropylacrylamide) (PNIPAM), polysterene (PS), maleated carboxymethylchitosan (MACACS), poly(N-vinylcaprolactam) (PVCL), poly(styrene-NIPAM) (P(S-NIPAM)), poly(NIPAM-co-methacrylic acid) (P(NIPAM-co-MAA)), poly(NIPAM-co-2-(dimethylamino)ethyl methacrylate) (P(NIPAM-co-AMPS)), poly(4-vinylphenylboronic acid-co-NIPAM-co-acrylamide) (P(VPBA-NIPAM-AAm)). Solute abbreviations: 4-nitrophenol (NP), 4-aminophenol (AP), nitrobenzene (NB), aminobenzene (AB), hexacyanoferrate(III) (HCF), hexacyanoferrate(II) (HCF2), o-nitrophenyl-D-glucopyranoside (oNPG). Ref. Architecture Core - Polymer Catalyst Reaction Result 7 Yolk-shell Au - PNIPAM Au reduction: NP ! AP & NB ! AB T -dependence on gel swelling and reaction rate 5,6 Core-shell Au - PNIPAM Au reduction: HCF ! HCF2 T -dependence on gel swelling and reaction rate, rate dependence on nanoreactor concentration and cross-linking density 6,67 Core-shell Pt/Au - PNIPAM Pt/Au reduction: NP ! AP rate dependence on reactant concentration 8 Core-shell Cu O - PNIPAM Cu O decomposition by visible light: methyl orange T -dependence on gel swelling and reaction rate 2 2 21 Core-shell Au - PNIPAM Ag reduction: NP ! AP photoresponsive gel size and reaction rate 65 Core-shell Au - P(VPBA-NIPAM- Au reduction: NP ! AP & NB ! AB glucose concentration dependence on gel swelling and reaction AAm) rate 4,51 Core-shell PS - PNIPAM Au/Pt/Rh oxidation: benzyl alcohol ! benzaldehyde T -dependence on gel swelling and reaction rate 4,56 Core-shell PS - PNIPAM b-D- hydrolysis: oNPG! D-glucose + o-nitrophenol T -dependence on gel swelling and reaction rate glucosidase 57 Core-shell PS - PNIPAM Ag reduction: NP ! AP T -dependence on gel swelling and reaction rate 59 Core-shell P(S-NIPAM) - Ag reduction: NP ! AP & NB ! AB T -dependence on gel swelling and reaction rate P(NIPAM-co-MAA) 19 Microgel PNIPAM/MACACS Ag reduction: NP ! AP T -dependence on gel swelling and reaction rate 55 Microgel PVCL-a-cyclodextrin Au reduction: NP ! AP T -dependence on gel swelling and reaction rate 60 Hydrogel P(NIPAM-co-MAA) Au reduction: NP ! AP T -dependence on gel swelling and reaction rate 61 Microgel cellulose cellulase hydrolysis: cellulose ! glucose T -dependence on gel swelling and time-dependent product con- (enzyme) centration 62 Microgel P(NIPAM-co-AMPS) Ni reduction: NP ! AP T -dependence on gel swelling and reaction rate 63 Microgel P(NIPAM-co-AMPS) Ag reduction: methylene blue T -dependence on gel swelling and reaction rate, pH-dependence on gel swelling penetrating molecules (e.g., gas, ligands, reactants, etc.) to per- responsive nanoreactors to reach the high recognition, selectiv- 115,116 meate and flow through a given medium under the action of an ity, and feedback control as found for enzymes, to create external field or chemical gradient. It is thus without doubt one of “colloidal enzymes”. Here we review the state-of-the-art of the the most fundamental transport descriptors employed in the phys- current understanding of the intricate links between nanoreactor ical sciences and material engineering. In the standard ‘solution– reaction rate and polymer permeability. Most of the results pre- diffusion’ picture for permeable membranes, it is commonly de- sented here are based on our recent research endeavor of multi- 89–99 fined on the linear response level by scale modeling schemes of hydrogel systems in order to establish rational design principles of responsive nanoreactors. We start P = K D ; (2) in in Section 2 by summarizing the rate theory for nanoparticle- catalyzed bimolecular reactions including partitioning and per- where meability of the polymer shell. The key property to be tuned in K = (3) and ‘programmed’ during the synthesis in order to select and switch catalytic activity is the permeability. In Section 3 we thus is the partition ratio, in this work simply referred to as partition- proceed with mesoscale coarse-grained computer models, which ing, defined as the ratio of number densities of the solutes in- give fundamental insights on partition–diffusion correlations in side (c ) and outside (c ) the medium in equilibrium, and D is in 0 in the permeability and how they can be tuned qualitatively by mi- the diffusion coefficient of those inside. Permeability can be thus croscopic interactions. In the last part we turn to atomistically- defined as the inverse of a diffusional resistance of a medium re- resolved molecular simulations of the PNIPAM hydrogel models garding the total mass transport (flux) towards the catalyst driven in swollen and collapsed state. Here we address the question of by the reaction. The optimization of permeability, especially the influence of the ‘chemistry’ of the interactions, e.g., role of for being highly selective among different solutes, has been a 95,99,100 (temperature-dependent) hydration, polarity, reactant type and grand challenge in material design over the last decades. size, etc. Obviously, there are many open questions, missing con- Prominent applications revolve around gas separation and re- 90,95,99,101–103 nections, and remaining challenges to overcome to obtain a com- covery, desalination and nanofiltration (‘molecular 104–106 prehensive multi-scale model. We will briefly discuss those and sieving’), medical treatments by dialysis or selective drug 107,108 give an outlook in the final, concluding section. transport, hydrogel-based soft sensors, and the nanoreac- 2,4,83,93 tors. We have studied partitioning and permeability of 2 Bimolecular reactions in nanoreactors polymer networks and PNIPAM polymers recently on the molec- 109,110 ular level by coarse-grained as well as all-atom molecular 2.1 Macroscopic rates and dependence on permeability 111–114 dynamics computer simulations. We review the rate theory for nanoreactors for the case of surface- Hence, in the last couple of years many quantitative concepts catalyzed bimolecular reactions in one of the simplest nanoreac- have emerged both on the continuum and the microscopic level tor geometries, a core-shell nanoreactor, depicted in Fig. 1A, that will eventually lead to a more fundamental understanding where a catalytically active metal nanoparticle of radius R is em- of nanoreactors in the future. The possibility arises to optimize bedded in a thermoresponsive hydrogel matrix of outer radius Journal Name, [year], [vol.], 1–17 | 3 R . In this spherically symmetric system, we consider that two the width equal to the polymer shell width (see Fig. 1B), i.e. NR species A and B diffuse from a bulk solution with respective (ini- A B D R r R ; NR tial) concentration c and c through the polymer shell towards i in 0 0 D (r) = (7) the catalyst nanoparticle. A fraction of the reactants arriving at D elsewhere; the surface combines with each other to produce a third molec- and ular species C. Assuming a total concentration of nanoreactors ( DG R r R ; NR c , the experimentalist would measure the transformation of a NR DG (r) = (8) 0 elsewhere: reactant (say reactant A) per time, according to i i A Here, D and D stand for the diffusion coefficients in the poly- dc (t) in 0 A B =k (c (t);c (t))c (4) i tot NR mer shell and solution, respectively. DG represents the transfer dt free energy from bulk water into the shell and as such strongly A B with instantaneous bulk concentrations c (t) and c (t), and k tot depends on the state (swollen/collapsed) of the nanoreactor. Us- has the units of inverse time and is a non-trivial function of the ing standard thermodynamic relations, we connect the flux of the reactant concentrations. In general, and as we will see below, the i species i to its local concentration c (r) chemical reaction has no well-defined order. In some limits, e.g., i i i i in an abundance of species B, it may reduce to pseudo unimolec- J =D c Ñbμ ; (9) ular or even pseudo first-order kinetics (see also Section 2.2 where μ (r) is the chemical potential of the species i, and b = later). 1=k T , with k denoting the Boltzmann’s constant and T the ab- B B To derive the functional form of the total catalytic rate, we as- solute temperature of the system. The chemical potential of a sume k (number of molecules reacting per unit of time) is equal tot molecule interacting with an external environment with a spa- to the radial flux of reactants at the nanoparticle surface. In bi- tially dependent concentration and free energy is molecular reactions, then the fraction of molecules A reacting is proportional to the number of molecules B at the same location, i i bμ = ln +bDG ; (10) and vice versa. Thus, k can be estimated through the standard tot 78,117 ref mean-field relation where c is a reference concentration whose value can be chosen A B ref k = K c (R)c (R); (5) tot vol arbitrarily. Equation (10) can now be used to relate the transfer A B free energy DG and partitioning, eqn (3), where c (R) and c (R) are the reactant concentrations at the nanoparticle surface, and K the probability that the two species vol i i K = exp bDG : (11) react on the surface (with units per time and per concentration squared). To calculate k , we solve the stationary continuity tot With the aforementioned definitions, the shell permeability to the equation for the density fields of reactants, species i is calculated as ÑJ = 0; (6) i i i P = K D (12) in with J (r) being the radial flux of the species i = A, B, C as a func- i i where for r > R we have P = D . NR tion of the distance from the nanoparticle. We make the station- arity assumption that the system is always in a steady-state and there is no explicit time-dependence of the fluxes. In other words, we assume the microscopic relaxation of the system, roughly given by the time of reactants to diffuse through the nanoreac- We found that the total catalytic rate for bimolecular reac- tor R =D , is faster than the reaction time as defined in eqn (4). NR tions in responsive nanoreactors is obtained as in eqn (1). In this If we use the fastest, diffusion-controlled (Smoluchowski) rate expression, k ' k = 4pRD c as the reaction rate scale, we find the condi- tot 0 0 A B 2 5 A bDG (R) B bDG (R) A B A B tion for stationarity that c  1=(4pR R) ' 10 mol/l for typ- k = K c e c e = K c c K K (13) NR R vol vol 0 0 0 0 ical geometries where R ' 1 nm and R ' 10 nm. In experi- NR stands for the surface-part of the reaction rate, which is explicitly ments, typically sub-micromolar reactant concentrations are used partitioning-dependent, and and the reaction rate is at least 1–2 orders slower than the fastest, fully diffusion-controlled limit, so that the condition is in most Z cases very well satisfied. k i = 4pc dr (14) D 0 i 2 P (r)r In their diffusive approach to the catalyst nanoparticle, reac- tants have to permeate the shell. The kinetics of this process is is the permeability-dependent diffusion part of the reaction rate i i thus governed by the shell permeability, which depends on the of the reactant i. In the absence of the shell, P (r) = D , and the i 0 i i diffusivity profile, D (r), and on the thermodynamic barrier, i.e., diffusion rate turns into the Smoluchowski rate k = 4pRD c . 0 0 the transfer free energy between bulk and shell, DG (r). For sim- For the core-shell configuration depicted in Fig. 1B the step pro- plicity, we take both profiles to be shell-centered step functions of files in eqn (7) and eqn (8) apply and the relation between the Journal Name, [year], [vol.], 4 | 1–17 treated as pseudo-unimolecular (see also next Section 2.2), that 0.08 HCF HCF is, dc (t)=dt =k c (t). obs The temperature dependence of the measured pseudo-first- 0.06 order constant is shown by blue filled squares in Fig. 2A. We observe that the reaction rate decreases by one order of mag- 0.04 nitude when the temperature of the solution exceeds the lower critical solution temperature (LCST) of the PNIPAM polymer. The measured nanoreactor hydrodynamic radius data, displayed by 0.02 orange open circles in Fig. 2A, exhibit the well-known volume transition between the swollen and the collapsed states below and above the LCST, respectively. As we pointed out before, this 280 280 290 290 300 300 310 310 320 320 transition changes the physicochemical properties of the polymer, TT (K) (K) which leads to different reactant diffusivity and transfer free en- 0.25 ergy values resulting in a nanoreactor permeability switch at the LCST. In the diffusion-controlled limit for pseudo-unimolecular 0.2 HCF HCF HCF kinetics we can identify k = k c =c . The c is the obs NR D 0 0 initial bulk concentration, which can be replaced by the instan- 0.15 HCF taneous c (t) in the equations during the reaction because of the stationarity assumption. Using eqn (15) together with the 0.1 experimental data from Fig. 2A, we estimate the temperature de- pendence of the nanoreactor shell permeability for HCF (Fig. 2B) 0.05 and clearly observe the aforementioned permeability switch be- low and above the LCST. The permeability decreases around one 280 290 300 310 320 order of magnitude from the swollen to the collapsed state. By T (K) comparing Figs. 2A and B we clearly see that the nanoreactor shell permeability is the essential ingredient to understand the Fig. 2 (A) Temperature dependence of the measured pseudo-first-order reaction rate response of nanoreactors in diffusion-controlled re- constant k (blue filled squares) of the electron-transfer reaction be- obs tween HCF and borohydride ions in Au-PNIPAM nanoreactors. The mea- actions. Detailed mesoscopic and microscopic insights on the in- sured temperature dependence of the nanoreactor hydrodynamic radius fluence of effective interaction potentials, hydrogel density, and is shown by orange opened circles. All data were taken from Ref. 5. chemistry on the permeability of polymer shells is presented in (B) Temperature dependent nanoreactor shell permeability for HCF esti- Sections 3 and 4. mated using eqn (15). 2.2 Pseudo-unimolecular reactions in nanoreactors Bimolecular reactions are typically treated as pseudo- shell permeability and the diffusion rate, eqn (14), simplifies to unimolecular when one of the reactants is in large excess " ! # with respect to the other. The reasoning behind this assumption k i R D 0 = 1+ 1 1 : (15) i is that, according to the simple Smoluchowski rate, the reactant P R i NR in larger concentration would diffuse towards the nanoparticle Equation (1) is the main analytical result for nanoparticle surface at a much larger rate than the other one. Therefore, surface-catalyzed bimolecular reactions. It shows that, in the when the reactant in limiting concentration arrives to the catalyst, it will always find a reactant of the other species to fully bimolecular case, the diffusional fluxes of the different re- actants are coupled. Thus, k depends in a non-trivial way on the combine with. However, this is not always true when considering tot surface and the diffusion rates and nanoreactor shell permeabil- nanoreactors. In this case, the diffusion rate, eqn (14), not only ity, in contrast to the simple unimolecular case (i.e., in general depends on the bulk reactant concentration but also on the shell 1 1 1 permeability and thus on the molecular interactions of reactants k 6= k + k in bimolecular reactions). tot D R Equation (1) together with eqs. (13) and (15) can be used with the shell. It is thus the combination of both quantities that determines whether a bimolecular reaction can be treated as to predict the total catalytic rate once the nanoreactor shell pseudo-unimolecular or not. permeability and the reactant partition ratios are known (e.g., from experiments, or from simulations, see Sections 3 and 4), If one of the reactants has a much larger diffusion rate than or, conversely, to extract the parameters by fitting to experi- the other one, e.g., k B  k A , the total reaction rate, eqn (1), D D mental data. Carregal-Romero et al. investigated the bimolec- reduces to (see Supporting Information in Ref. 83) ular electron-transfer reaction between hexacyanoferrate (III), 1 1 1 Fe(CN) (HCF), and borohydride BH ions in Au-PNIPAM core- k ! k = k + k ; (16) 6 4 tot A tot R shell nanoreactors. In a previous work we demonstrated that that bimolecular reaction is diffusion-controlled and can be which is the well-known expression of the total reaction rate Journal Name, [year], [vol.], 1–17 | 5 -1 PP / D k (s ) <latexit sha1_base64="q1hBcpw8Oxqa2w7zfIic8raScM8=">AAAB8nicbVDLSsNAFL3xWeur6tLNYBFclaQKiquCG5cV7APaUCbTSTt0MgkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeIJHCoOt+O2vrG5tb26Wd8u7e/sFh5ei4beJUM95isYx1N6CGS6F4CwVK3k00p1EgeSeY3OV+54lrI2L1iNOE+xEdKREKRtFKvX5EccyozJqzQaXq1tw5yCrxClKFAs1B5as/jFkacYVMUmN6npugn1GNgkk+K/dTwxPKJnTEe5YqGnHjZ/PIM3JulSEJY22fQjJXf29kNDJmGgV2Mo9olr1c/M/rpRje+JlQSYpcscVHYSoJxiS/nwyF5gzl1BLKtLBZCRtTTRnalsq2BG/55FXSrte8y1r94arauC3qKMEpnMEFeHANDbiHJrSAQQzP8ApvDjovzrvzsRhdc4qdE/gD5/MHhf+RYg==</latexit> 0 obs LCST LCST R (nm) NR control), but the values in the polymer shell, which can be the limiting factor. The latter is defined by the shell permeability P -3 B A P P <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> /<latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> =10 and can thus strongly differ from the bulk value. Because of the -2 B A P /P =10 <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit><latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> 100 responsive nature of the gating shell of nanoreactors, this depen- -1 B A P P <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> /<latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> =10 dence crucially implies that the identity of the limiting reactant B A P P <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> /<latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> =1 can switch upon a change in the external stimulus. Failure to recognize this fact can lead to very large discrepancies between 10 the correct and the approximate rate. This theoretical framework for pseudo-unimolecular reactions qualitatively rationalizes the large and sharp variations in catalytic rate observed in the rele- 6–8,78 vant nanoreactor experiments. 3 Partitioning and diffusion: coarse-grained 0 1 2 3 4 10 10 10 10 10 B A simulations c / c 0 0 As we have just described, key parameters to understand a Fig. 3 Total rate for unimolecular reactions k , eqn (16), divided by the tot nanoreactor’s selectivity and rate response to stimuli are the total reaction rate k for bimolecular reactions, eqn (1), as a function of tot B A permeability of its polymeric shell and the reactant partitioning the relative reactant bulk concentration, c =c . The lines stand for dif- 0 0 B A within. In the following, we review two selected coarse-grained ferent relative nanoreactor shell permeabilities to the reactants, P =P . We assume k = k . A (CG) simulation studies of partitioning, diffusion, and permeabil- 109,110 ity in model membranes. Mesoscopic models, neglecting chemical resolution, play a pivotal role not only as a bridge be- tween the aforementioned macroscopic reaction model and the in unimolecular reactions, k . In this case, the total catalytic tot following microscopic all-atom models but also for the deeper time is the sum of the diffusion time of the slower reactant and understanding of essential physics, e.g., of molecular adsorption the surface reaction time. Hence, in nanoreactors, unimolecu- and transport in polymer systems. Particle-based simulations on lar reactions can be diffusion- or surface-controlled if k  k D R various scales with increasing complexity and chemical detail are or k  k , respectively. If both rates are comparable in magni- D R now emerging. For the convenience of the reader, we have sum- tude, the reaction is termed diffusion-influenced. Analogously, a marized selected relevant simulation efforts in Tab. 2. reaction is diffusion- or surface-controlled if Da  1 or Da  1, II II respectively, where Da = k =k is the second Damköhler num- II tot D 3.1 Influence of gel volume transition on reactant partition- ber. If both reactants diffuse from the bulk solution, according B B A A ing in a model polymeric membrane to eqn (14), this condition is satisfied when c P  c P . This 0 0 Responsive polymers feature a sharp volume transition where the means that one of the reactants should be in a much higher bulk density of the polymer drastically changes. The partitioning of concentration and/or subject to a much larger shell permeability reactants across the volume transition and the feedback of the than the other. polymer to the permeation is complex and poorly understood. We In Fig. 3 we analyze how large should be the excess of re- thus first discuss a CG simulation model with details described actant B for the pseudo-unimolecular reaction limit to be valid. previously consisting of permeating reactants in a polymer- This value depends on the relative nanoreactor shell permeabil- B A based thin membrane (Fig. 4A), where we aim at a qualitative ity, P =P . For simplicity, we consider that the surface rate is study of the effects of structural transitions of gels on the reactant equal to the diffusion rate of the reactant in limiting concentra- partitioning and its back-coupling to the volume transition. In tion (k = k A , diffusion-influenced reaction). When both reac- the following, we refer to the permeating reactants generally as tants have the same permeability (red line), the concentration of ‘solutes’. reactant B should be roughly 10 times larger than the one of A to have a unimolecular reaction. If we then decrease the shell The membrane is constructed as cross-linked semi-flexible net- 118–123 permeability to the reactant B by 10 times, its concentration has work of polymers formed on a regular cubic lattice, and thus to become 100 times higher with respect to that of A to keep the solutes can diffuse throughout the membrane and the bulk this limit. Figure 3 also shows that the catalytic rate predicted regions. This enables a direct sampling of the solute partition- for a pseudo-unimolecular reaction for the reactant in limiting ing from the simulations simply according to eqn (3). We use concentration may differ from the fully bimolecular one by or- the Lennard-Jones (LJ) pair potential and its size unit s as the ders of magnitude. Thus, when dealing with nanoreactors, it is diameter for all particles and the monomer–monomer (bonded) necessary to consider not only the difference between the bulk distance in the polymers. In order to model such a gel in the concentrations of the reactants but also the difference in the shell presence of various solutes, we employ inter- and intra-particle permeability to the reactants. interactions in terms of LJ pair potentials. We focus on two Hence, when considering nanoreactors, care should be taken key interaction parameters: The membrane–membrane interac- since in these systems it is not the bulk mobility and concentra- tion e controls the solvent quality, turning it from good to poor mm tion that determine the reaction type (diffusion versus surface upon the increase of e . The membrane–solute interaction e mm ms Journal Name, [year], [vol.], 6 | 1–17 k / k tot tot solute-induced collapsed swollen 1 4 membrane solute Solute-involved collapsed Solute-induced collapsed 1 5 Solute-adsorbed collapsed Critical 3 6 collapsed 3 6 Collapsed Swollen Fig. 4 (A) Various conformational states and regimes in the mesoscopic network membrane–solute system from CG computer simulations are depicted in the main phase diagram (center) depending on the Lennard-Jones (LJ) interactions e and e : (1) is a ‘solute-induced collapsed’ state, (2) is a mm ms ‘critical’ transition line (yellow contour) between the intrinsic (3) ‘swollen’ and (6) ‘collapsed’ states. (4) is a ‘solute-involved collapsed’ state, while (5) is a ‘solute-adsorbed’ collapsed state where solutes adsorb mostly on the membrane surface. For details see text. Reprinted with permission from Ref. 109, copyright 2017 American Chemical Society. (B) Solute partitioning landscape K depending on e and e . mm ms governs the membrane–solute coupling and thus models different only on the membrane–solute interaction but also significantly on kinds of solutes. For the solute–solute interaction we always use the solvent quality. The partitioning overall becomes large (i.e., e = 0:1 k T , essentially being a steep (r ) repulsion. higher adsorption) as the membrane–solute attraction, e , in- ss B ms creases, while it has large regions of unity in the swollen states In Fig. 4A we show the landscape of the gel structural phases, (light blue-greenish areas). Note that when compared with the depending on both interaction parameters. The red region de- structural landscape in Fig. 4A, both extrema of K (i.e., the min- picts swollen states, where the gel volume is relatively large, imum and the maximum) are in the collapsed regions, indicating while the blue region indicates collapsed states. Without the so- that the collapsed phase can relate to extremely different parti- lutes (e = 0) our model exhibits a collapse transition at around ms tionings and there is no unique mapping. Moreover, at interme- e ' 1:0 k T . In the presence of the solutes, however, the pic- mm B diate values of e , K is a nonmonotonic function of the solvent ms ture becomes more complex: The ‘critical’ transition line (yellow quality, meaning that it can be maximized by an optimal solvent contour line) between swollen and collapsed states depends sub- quality. The maximization of the partitioning is in fact a quite stantially on the membrane–solute interaction, as shown by the general feature in attractive but crowded systems as we will dis- label (2). The stark color contrast around this critical line signifies cuss in the following section. the sharp transition. In addition, one can identify in total five dis- To sum up, the CG simulation model of a polymer network in tinct phase regions (or states), classified into (1) “solute-induced the presence of solutes reveals a rich topology of structure phases collapsed”, (3) “swollen”, (4) “solute-involved collapsed”, (5) and their relation to solute partitioning, entering the rate equa- “solute-adsorbed collapsed", and (6) “collapsed” states, indexed tions eqs. (13) and (15). In particular, for very attractive solute– by the numbers in the colored center panel of Fig. 4A. Inter- membrane interactions (e & 1 k T ) the network structure and ms estingly, the “solute-induced collapsed” state (state 1) can occur B partitioning are coupled. The mesoscopic model thus provides a even in good solvent conditions, where the membrane undergoes landscape of the partitioning, thereby bridging the macroscopic a relatively sharp collapse transition induced by a strong ‘bridg- continuum model and microscopic discrete data in terms of the ing’ attraction between the solutes and the network monomers. generic interaction parameters. In addition, the results will be The effect has been reported in computer simulations before but 124–131 helpful for the interpretation of experiments for certain polymer– only on the single polymer level. The “solute-involved” col- reactant systems and could also be useful to design feedback- lapsed state (state 4) occurs at the intermediate solvent quality systems where the local reactant (or product) concentration may where the membrane collapses with most of the internal solutes couple back to polymer structure in a prescribed way. embedded, yielding a bulkier collapsed gel than the intrinsically collapsed case. The “solute-adsorbed collapsed” state (state 5) is an example for the limiting case, where both of the membrane– 3.2 Partitioning, diffusion, and permeability in a model lat- membrane and membrane–solute attractions are strong, but the tice membrane first one dominates and excludes the solutes, therefore leading to Now we present a related but different CG model of a a strong surface accumulation of those. membrane–solute system (Fig. 5A) in order to study permeability The solute partitioning K (on a log-scale) averaged over the in dense media qualitatively on a generic level. We demonstrate how the permeability can be tuned massively in magnitude by membrane slab is shown in Fig. 4B in a 2D-landscape plot, and is related to the transfer free energy from the bulk into the net- systematically varying the membrane–solute interactions and the work, DG =k TlnK , which quantifies the average transfer free density of the membrane. The study also gives important insights about how partitioning and diffusion are correlated. energy for the solute transfer from bulk to the membrane. The partitioning varies by several orders of magnitude, depending not The permeability is defined following the solution–diffusion Journal Name, [year], [vol.], 1–17 | 7 solute-involved collapsed solute-adsorbed collapsed D A B C membrane solute Fig. 5 (A) Snapshot of the mesoscopic lattice membrane–solute system. The membrane sites (red) are fixed on a face-centered-cubic (fcc) lattice with the volume fraction f , and the penetrating solutes (blue) are diffusing and interacting via the LJ potentials with e . (B) Solute partitioning K (f ) at m ms m different e . The solid lines depict the exact relation, eqn (17), and the dashed line depicts the approximated partitioning K with be = 0:6 (see ms B2 ms text for details). (C) Solute diffusivity D (f )=D with different e . The dashed lines depict the approximation D =D = exp(f ) valid for low e and in m 0 ms in 0 m ms 2=3 f , and the scaling D =D  f valid for high e and low f . (D) Permeability P(f )=D at different e . Reprinted with permission from Ref. 110, m in 0 m ms m m 0 ms copyright 2019 American Physical Society. 89–91,94–99 theory by eqn (2). There have been pioneering the- sion. In addition, a leading order approximation of K on a two- ms oretical models to elucidate the transport phenomena in mem- body level, K (f ;e ) = exp 2c B for be = 0:6 is de- B2 m ms m ms 89–91,94,132–135 branes based on simple theories for either parti- picted by the dashed line, where c µ f is the membrane con- m m ms tioning or diffusion. Recently, a simulation study revealed maxi- centration, and B is the second virial coefficient. Figure 5C mization of partitioning of penetrating solutes in polymer mem- shows the solute diffusivity D =D in the membrane as a func- in 0 branes tuned by the polymer volume fraction. Diffusion in tion of f , rescaled by the free diffusivity in the bulk. As the dense membranes is usually quite complex and highly dependent membrane becomes dense, D tends to exponentially decrease, in 102,103,113,137–141 on details of the interaction potentials. Never- showing more complex behavior with higher couplings. We com- theless, there have been no comprehensive studies on the perme- pare the simulation results with scaling theories for diffusion in ability P , being a product of partitioning and diffusion. two limiting cases. The upper dashed line indicates the limit- ing law D =D = exp(f ) based on the ‘volume-exclusion’ the- In the CG model membrane–solute system as shown in Fig. 5A in 0 m 102,134,135,143,144 the diffusive solutes are ideal (e = s = 0), and the membrane ory. It is indeed valid only for low couplings, ss ss which acts essentially repulsive. For high membrane–solute at- consists of immobile interaction sites, located on a face-centered- tractions and low membrane density, the diffusivity follows the cubic (fcc) lattice with a fixed unit cell size l, variation of which 2=3 power law D =D  f . The scaling relation is derived by the tunes the monomer packing fraction f . The simplicity of such m in 0 m an ordered and rigid model membrane with ideal solutes ren- limiting law from the Kramers’ barrier crossing over the distance 2=3 1=3 2 110 l  f , and therefore D  l =t f . ders the problem easier for interpretation and perhaps theoret- in ically tractable. The ideal solutes diffuse throughout the simu- The resulting permeability, the product of K and D , shown in lation box but interact only with the membrane sites via the LJ in Fig. 5D, exhibits intriguing features. For essentially repulsive potential with the coupling strength e . For the ideal solutes the ms solutes, P decreases monotonically as the membrane density in- partitioning can be exactly computed via the transfer free energy creases, and the overall magnitude is below unity, almost ap- bH 142 ms shown in eqn (11). In our case DG = k T lne , where proaching zero for very dense membranes. We speculate that this H (r) = U (jr r j) is the total Hamiltonian (summing over ms i ms i essentially repulsive case may be the scenario in the experiments all membrane sites i), and x  dVx=V is the volume average, with the highly charged reactant HCF in Section 2.1, which prob- yielding ably does not like to enter the collapsed gel, but this suspicion bH ms K = e : (17) needs further scrutiny. On the other hand, for high couplings (at- which verifies the simulation results (Fig. 5B). traction) the permeability is first minimal around f = 0:1, then The computed partitioning as a function of the membrane vol- maximized at large membrane densities f ' 0:8. The perme- ume fraction f is shown in Fig. 5B for various e . For rel- ability vanishes at the maximum overlapping density (f  1), m ms m atively low membrane–solute couplings (be . 0:3), the LJ in- where no percolating holes for diffusion are present anymore. ms teraction between solutes and membrane sites is essentially re- This demonstrates a clear maximization of permeability when the pulsive (signified by a positive second virial coefficient), and the system is highly attractive and dense. The nonmonotonic behav- partitioning monotonically decreases as the membrane becomes ior of permeability results from drastic nontrivial cancellations dense, owing to the dominant exclusion by the membrane. For between the partitioning and the diffusivity, which exponentially intermediate couplings around be = 0:6, which is moderately increase and decrease, respectively. The massive cancellation be- ms attractive, partitioning reaches a maximum at an optimal mem- tween two largely varying functions over several orders of mag- brane density around f = 0:6. The partitioning maximization nitude yields a permeability of the order of unity, implying a is attributed to a balance between adsorption and steric exclu- high potential for fine-tuning of the permeability behavior in ex- Journal Name, [year], [vol.], 8 | 1–17 periments by small changes in density or interactions. c , G = G N c ; (19) chain m 0 Mesoscopic models of membrane–solute systems demonstrate that the permeability, typically resulting from large cancellation Another setup of swollen hydrogels, shown in Fig. 6.ii, mimics effects of partitioning and diffusion, is very sensitively tuned by the cross-linker unit of a hydrogel network, and thus lends itself the effective interaction potentials and the membrane density. to study the influence of cross-linkers on adsorption of molecules. The results indicate that most drastic selectiviy effects are at high In our previous study , we considered a very common N,N’- membrane densities and significant (& k T ) membrane–solute at- methylenebisacrylamide (BIS) cross-linker, connecting four PNI- tractions. The effective potentials in realistic material design as- PAM chains with their ends tethered in a tetrahedral geometry. semblies can be somewhat controlled by various external param- The solute molecules in general adsorb in different proportions to eters, such as temperature, ionic strength, pH, and possibly var- the chain regions and the cross-linker neighborhood. The overall ious additives in the solution. The results from the mesoscopic adsorption in the radial interval [r ;r ] from the cross-linker is ob- 1 2 models thus provide useful physical insight and may bear impor- tained in a straightforward manner by integrating the (spherical) tant applications in design and engineering of molecular systems RDF g(r) of the solutes to achieve a selective transport by fine-tuning interactions and topologies, particularly in highly attractive membrane systems. 2 G(r ;r ) = c [g(r) 1]4pr dr: (20) 1 2 0 4 Partitioning and diffusion: all-atom simu- With this, we can scan the adsorption in different regions with lations respect to the cross-linker. It also allows us to evaluate G , as in the single-chain geometry. Finally, the total adsorption can be The advantage of the mesoscopic simulations in the previous sec- deconvoluted into two contributions, tion is that we can obtain fundamental and qualitative insights on how permeability depends on basic input parameters such as G = G +G : (21) tot chain xlink interaction energies, lattice geometry, and single solute diffusion. The adsorption on the chains G (unperturbed by the presence However, in experiments we deal with specific, chemical systems, chain of cross-linker) is given by eqn (19), whereas G represents the where the effects of interactions are highly convoluted and solva- xlink effect on the adsorption due to the presence of the cross-linker. tion, polarity, electrostatics, and specific steric constraints come The value of G can be evaluated from known G and G . explicitly into play. Hence, for a more detailed insight and quan- tot xlink chain In the infinite-dilution limit, the adsorption on the cross-linker titative numbers for the continuum approach to reaction rates in is proportional to the bulk solute concentration, G = G c , Section 2, we need to resort to higher resolution, molecular dy- xlink 0 xlink where G is the adsorption coefficient of the cross-linker. namics computer simulations. All-atom simulation studies of par- xlink titioning and diffusion through polymer networks with increasing The resulting adsorption coefficients G are shown in Fig. 7B 111 112 complexity and chemical detail are now emerging and growing (blue shaded bars), from the single-chain and cross-linker in the literature. Selected works in this field are given in Tab. 2. geometries. Quite generally, the adsorption grows with the In the following, we will review our recent efforts to understand molecular size. The effect of the cross-linker, G , is shown in xlink partitioning and diffusion of solutes in swollen and collapsed PNI- Fig. 7B by orange bars: The apolar compounds C , C , and B 4 6 111–114 PAM hydrogels by all-atom (AA) MD simulations. show a low affinity to the cross-linker. In contrast, the adsorption of nitro-aromatic solutes to the cross-linker is significant, in par- 0 112 0 ticular for NP . NB shows more than doubled and NP even an 4.1 Swollen state order of magnitude higher adsorption to the cross-linker region In order to model the swollen state of a hydrogel shell one can fo- than to a monomer of the polymer. Note that the BIS cross-linker cus on one elongated PNIPAM chain, as shown in Fig. 6A.i, where has two amide groups and is slightly more hydrophilic than the the chain is replicated through periodic boundary conditions. The PNIPAM chain, hence favoring polar molecules. cylindrical geometry allows for a simple extraction of adsorption From the known adsorptions on individual chains and cross- properties of solvated molecules in the solution. The first step linkers we can predict the partitioning in extensive hypothetical is to evaluate the cylindrical radial distribution function (RDF) of swollen polymer architectures, such as hydrogels. The partition- the solute molecules from the backbone, g (r), as shown in an 2D ing follows from K = 1+G =(c V), where V is the volume of the tot 0 example for nitrobenzene (NB) in Fig. 7A. The adsorption coef- gel and G the total adsorption of molecules on all the chains tot ficient G per monomer of the polymer is then obtained by inte- and cross-linkers [eqn (21)], which leads to gration along the spatial coordinates, K = 1+ n G + n G ; (22) ¥ m m xlink xlink G = DL [g (r) 1]2prdr; (18) m m 2D where n and n are the monomer and cross-linker number xlink where DL = 0:265 nm is the distance between neighboring densities, respectively. The former can be easily linked to the monomers in the chain. The total adsorbed number of molecules polymer volume fraction f as n = f =(pR DL ), where R = m m m m 0 G on the chain is proportional to the number of monomers N 0:5 nm is an estimated effective radius of the polymer chain. chain and, in the infinite dilution limit, to the bulk solute concentration Assuming f = 0:1 for a typical architecture of a swollen state, we Journal Name, [year], [vol.], 1–17 | 9 Table 2 Survey of selected publications on computer simulations of diffusion D, partitioning K , permeability P , or related adsorption or transport phenomena of (co)solutes in polymers. Abbreviations: All-atom (AA), Coarse-Grained (CG), Molecular Dynamics (MD), Monte-Carlo (MC), Langevin Dynamics (LD), Brownian Dynamics (BD), Dissipative Particle Dynamics (DPD). Ref. Simulation Architecture Polymer Resulting quantity Comment methods 145 AA-MD swollen cross-linked polyethyleneglycole diffusivity of water, ions, water content: 75–91%, mesh size: 2.3–5.5 nm network rhodamine 146 AA-MD collapsed polyethylene diffusivity and partitioning permeability for oxygen 5–6 orders of magnitude larger than for water of oxygen and water 147 AA-MD collapsed poly(vinyl alcohol) diffusivity of O2 water uniformly distributes 148, 149 AA-MD collapsed poly(vinyl alcohol) diffusivity of water hydrogel with 4–40 % water 150 AA-MD collapsed polydimethylsiloxane diffusivity of water and water/ethanol mixtures; water molecules faster than ethanol ethanol 151 AA-MD + tran- collapsed polystyrene and its diffusivity and partitioning sition state ap- copolymers of gas and water molecules proach 152 AA-MD single chain PNIPAM adsorption of urea studying volume phase transition 153 AA-MD single chain PNIPAM adsorption of TMAO, urea studying volume phase transition 154 AA-MD collapsed and sol- PNIPAM partitioning of ions thin core-shell membrane, direct measuring of partitioning vent phase 155 AA-MD cross-linked net- PNIPAM diffusivity of water; vol- studying volume transition; cross-linking inhibits the collapse work ume transition 156 AA-MD collapsed and sol- PNIPAM (3mer) water–polymer coexis- also 30mer of PNIPAM: no conclusions on chain configuration vent phase tence 157 AA-MD swollen & collapsed PNIPAM partitioning of large ions umbrella sampling of the potential of mean force of the ions finite aggregate 158 Gibbs-ensemble cubic network bead-spring solvent sorption and effects of solvents on swelling CG-MD swelling isotherm 119 Two-box– cubic network bead-spring solvent sorption and effects of cross-linkers on swelling particle-transfer swelling isotherm CG-MD 159, 160 CG-MC tetra-functional net- bead-spring solvent sorption and effects of polymer network density and deformation on swelling work swelling isotherm 120 CG-MC cubic network rigid rod cosolute diffusivity effects of cosolute size and polymer density on cosolute diffusivity 121 CG-LD cubic network charged bead-spring energy conversion effects of compression and solvents on energy contribution 122 CG-LD highly swollen cubic charged bead-spring adsorption and conforma- counterion-induced deformation network tional response 123 LD cubic network charged bead-spring ion transport effects of electrostatic coupling between polymer and ions on ion transport 136 CG-MC tetra-functional net- bead-spring cosolute partitioning effects of polymer density on partitioning work 139 CG-MD polymer melt semi-flexible gas partitioning, diffusivity effect of gas size and polymer semi-flexibility on gas transport and permeability 161 CG-BD cubic network bead-spring cosolute diffusivity effects of cosolute density on cosolute diffusivity 162 CG-DPD random network semi-flexible permeability and cosolute effects of porosity and deformation on permeability diffusivity 163–165 CG-BD cubic network rigid rod cosolute diffusivity effects of interactions, hydrodynamics, and network heterogeneity on cosolute diffusivity 166 CG-BD random cubic net- rigid rod cosolute diffusivity effects of network porosity, flexibility, degree of cross linking, and electro- work static interaction on cosolute diffusivity compute K for several solutes from the obtained MD parameters water–polymer spatial heterogeneity is a crucial player in the sol- in Fig. 8. The values range around unity, K 1–3, as also resulted vation of small molecules, whereby the nature of the solute (being from the CG models in Fig. 5B for this polymer fraction range. As polar, nonpolar, or ionic ) is a decisive property. Two represen- we will see in the following, the collapsed state can give rise to tative snapshots in Fig. 9A, showing a benzene (nonpolar) and a much higher partitioning. phenol (polar) molecule, demonstrate that nonpolar solutes are preferentially expelled from water clusters and tend to reside in ‘dryer’ regions of the gel, whereas polar molecules tend to parti- 4.2 Collapsed state tion closer to or inside water clusters. Thus, the “dual” character The collapsed state of the PNIPAM hydrogel can be modeled as a of the gel can favorably accommodate both polar and nonpolar bulk of aggregated polymeric chains (in our case 20 monomeric species. units long) at 340 K (above the LCST), where cross-linkers are ignored. The amount of sorbed water between the polymeric The transfer free energy for a given molecule is obtained as the chains is chosen such that it corresponds to the chemical equi- difference between the solvation free energy in PNIPAM (G ) and librium with bulk water. The amount of water in the col- in water (G ), DG = DG DG , both evaluated via the thermo- w g w lapsed state depends on temperature, and amounts to around dynamic integration procedure. Figure 9B shows DG for var- 20 wt. % (somehow less than experimental estimates of around ious solutes plotted versus the molecular surface area A of the 168–172 30 wt. % ), which roughly correspond to the polymer vol- solutes (defined as the envelope area of the fused union of the ume fraction of f = 0:8. Note that this is in the range of packing m 178 atoms ). The results follow a clear linear trend for the groups fractions for which we observed the most interesting behavior of of nonpolar and aromatic solutes as well as alcohols and water. permeability in the CG simulations in Section 3. The linearity in the very heterogeneous polymer–water medium Water molecules are very non-uniformly distributed through- is rather surprising. The results can be conveniently described in out the phase and tend to flock together into irregular clusters 179,180 terms of an effective molecular surface tension g , of various nanoscopic sizes, which were observed also in simu- 151,173–176 lations of other amorphous polymer structures . This DG = DG +g A : (23) 0 m m Journal Name, [year], [vol.], 10 | 1–17 Fig. 6 (A) Atomistic modeling of PNIPAM hydrogels: i) elongated, infinitely long chain (mimicking a part of a swollen network where the adjacent chains are far apart), ii) a cross-linker connected with four chains in a tetrahedral structure (representing a unit of a swollen network), iii) dense aggregate of PNIPAM polymers at 340 K (a model for a collapsed PNIPAM hydrogel). (B) Solute molecules in our study; polarity is characterized by the hydroxyl (OH) group. polar 21±3 nonpolar 15 Γ (single chain geometry) Γ (cross-linker geometry) Γ (cross-linker geometry) xlink C C C MeOH PrOH C B T NB NP 1 4 6 0 0.5 1 1.5 2 -5 r (nm) Fig. 7 (A) Cylindrical RDF of backbone–NB for an extended PNIPAM chain (see Fig. 6A.i). (B) Adsorption coefficients of various solutes (see 111 112 Fig. 6B) to a PNIPAM monomer G (from single-chain geometry and the OPLS force field, and from cross-linker geometry using the OPLS-QM2 force field) and to a cross-linker G (OPLS-QM2 force field) at 300 K. xlink g is strongly related to the difference in the molecule-PNIPAM collapsed swollen and molecule–water surface tension. Note that the sign de- pends on the transfer direction. The fit of eqn (23) to the NP 3 nonpolar solutes (dashed line in Fig. 9B) gives the value g = NB 1 2 18 kJ mol nm . For the alcohols and water, the transfer free energies are by about 7 kJ/mol above the trend of the nonpo- lar solutes, owing to a slightly different character of the hydroxyl 1 181 group than in alkyl chains. The molecular size is hence the dominant factor that determines its affinity to the hydrogel. In MeOH the CG description (Section 3) the molecular size is therefore re- T = 300 K T = 340 K flected in the interaction parameter e . φ = 0.1 φ = 0.8 ms -1 Using eqn (11), we show the partitioning in the collapsed state Fig. 8 Partitioning of several molecules resulting from the atomistic mod- in Fig. 8. In general, the partitioning of our neutral molecules els of a swollen (at 300 K and polymer fraction of f = 0:1) and collapsed (at 340 K and f = 0:8) state of a PNIPAM gel. The values for the swollen is larger in the collapsed state. Also, the larger the partitioning state are computed from eqn (22) and assuming polymer volume fraction in the swollen state, the progressively larger it is in the collapsed f = 0:1, whereas the values for the collapsed state are computed from state. With some heuristic arguments, we showed that the parti- eqn (3). 2 114 tioning roughly follows the relation K µ K . This collapsed swollen is in line with the universal observation from our CG model (Sec- Journal Name, [year], [vol.], 1–17 | 11 g(r ) Partitioning, ∗ 3 Γ (nm ) 0 500 ns 1000 ns He H O 5 nonpolar Ne MeOH polar Ar PrOH Me -1 He Et Ph 3 Ne -2 B NP H O 10 2 Me NB Et 2 0 -20 MeOH NP -3 nonpolar 10 NB polar (aromatic) -4 polar (alcohols, water) -40 0 0 0.1 0.2 0.3 0 0.5 1 1.5 2 0 1 2 3 4 5 A (nm ) x (nm) a (nm) Fig. 9 (A) Snapshots of benzene and phenol molecules solvated in the PNIPAM phase. Hydrophobic parts of the solutes are shown in yellow, the hydroxyl groups in green, PNIPAM polymers in blue, and water in red–white. (B) Transfer free energies from water to PNIPAM versus the molecular surface area. The dashed line is a fit of eqn (23) to the data points of nonpolar solutes. (C) Microsecond-long trajectory sequence of a NP molecule projected on a 2D plane (color coded from blue at t = 0 to red at t = 1000 ns). The green bubbles schematically depict the hopping transition with a transient pore opening. (D) Diffusion coefficients of molecules in the collapsed PNIPAM polymer versus their Stokes radii in water. The dashed line shows a fit of eqn (24) to the data points. tion 3.1), namely that a collapsed state can have much more ex- lower (i.e., higher DF ) and at the same time the diffusion coeffi- treme effects on partitioning than a swollen state. cients indeed decay faster with solute size. Moving on to the diffusion properties of solutes in the col- In conclusion, all-atom MD simulations offer insights into the lapsed PNIPAM, we first look at the projected trajectory of a NP molecular nature of the transport and solvation properties of molecule in Fig. 9C. Its connected blob-like structure suggests molecules in hydrogels. These mechanisms are not only impor- 182,183 that the diffusion advances via the hopping mechanism : tant for PNIPAM hydrogels, but most probably play important A penetrating solute resides for longer time in a local cavity roles also in other responsive hydrogels, and their understand- and suddenly performs a longer jump into a neighboring cav- ing is important for the rational design of novel materials. No- ity through a transient water channel that forms between the tably, we see drastically larger effects for K and D in the collapsed chains (schematically illustrated in the bubbles in Fig. 9C). We phase than in the swollen states, but apparently also a large anti- plot the diffusion coefficients versus the size of the solutes a (de- correlation between them, like in the coarse-grained simulations fined as the Stokes radius in pure water) in Fig. 9D. As the size in Section 3. The dense, collapsed state is thus more decisive for of a solute increases by a factor of 7, the diffusion coefficients nanoreactor design and control. Very recent studies indicate that in decreases by dramatic 5 orders of magnitude. The diffusion coef- particular for charged molecular reactants, the presence of water ficients depend on the solute size a roughly exponentially, w clusters and resulting substantial interfacial effects within a dense hydrogel may decisively affect their permeability behavior. a =l D = D e : (24) 5 Concluding remarks The fit to the data points yields the decay length l = 0:019 nm. Stimuli-responsive nanoreactors are of high potential for the de- Note that the rate-determining step in the hopping diffusion is sign of programmable and selective nano-devices for controlled the opening of a channel, which is associated with a free energy catalysis and can therefore serve as candidates to create novel barrier DF and can be via Boltzmann probability related to the synthetic enzymes on the colloidal scale. However, they consti- diffusion coefficient as D exp(DF =k T). In conjunction with a B tute complex devices with non-equilibrium processes starting at the empirically obtained diffusion relation [eqn (24)], this implies the electronic scale, defining the chemical surface reactions, cou- k T pled to those at the polymer network scale with all the intrinsic DF (a ) = a : (25) a w w complexity of polymer–reactant interactions, including the feed- back of responsive polymers, up to the device scale where re- That is, the free energy barrier depends linearly on the particle actants diffuse and react in a suspension of colloids. Here we size, and hence represents a special case of possible scenarios pre- reviewed the recent theoretical attempts of understanding some dicted by an assortment of different theories. The majority of the- parts of the processes by focussing mostly on the key roles played ories that are based on activated diffusion predict either square by the permeability of the polymer shell and the reactant parti- or cubic scaling. A possible linear dependence of the free energy tioning in order to control activity and selectivity, and how those barrier has recently been theoretically envisioned in scaling the- 137,184 enter the continuum rate predictions for the nanoreactors. ories for particle mobility in dense polymer solutions and in dense liquids by using a self-consistent cooperative hopping As an important general result, we see substantial variations theory. As also seen from eqn (25) the height of the free en- and correlations among K and D in the dense, collapsed polymer ergy barrier is related to the decay length l in eqn (24). We also phases, in both coarse-grained and atomistically-resolved simu- showed that in a less hydrated gel, the the diffusion of solutes is lations, which are thus more decisive and tuneable for nanore- Journal Name, [year], [vol.], 12 | 1–17 −1 ΔG (kJ mol ) y (nm) 2 −1 D (nm ns ) actor rate control than for the swollen states. Results for the Conflicts of interest temperature-induced rate switch observed in reference experi- There are no conflicts to declare ments, like the HCF reduction briefly discussed in Section 2.1, can be thus traced back to, for example, the large exclusion (low Acknowledgements partition ratio) and significant slowing down (low diffusion) of The authors thank Richard Chudoba, Karol Palczynski, Sebas- reactants in the collapsed state of neutral PNIPAM. However, a tian Milster, Arturo Moncho-Jordá, Stefano Angioletti-Uberti, quantification of partitioning and diffusion of molecular ions by Daniel Besold, Yan Lu, and Matthias Ballauff for inspiring dis- simulation approaches remains a challenge because of the water 177 cussions. This project has received funding from the European heterogeneities in the collapsed states. Research Council (ERC) under the European Union’s Horizon A large number of challenges and questions remain, which 2020 research and innovation programme (Grant Agreement No. we try to tackle currently or leave open for future studies. For 646659-NANOREACTOR). M.K. acknowledges the financial sup- example, continuum approaches to diffusion- and permeability- port from the Slovenian Research Agency (research core funding influenced rates in confinement are often based on mean-field No. P1-0055). W.K.K. acknowledges funding from the Deutsche theories (like presented here), but more elaborate and accurate Forschungsgemeinschaft (DFG) via grant NE 810/11. The sim- treatments, like Green’s-function approaches, are yet to be de- ulations were performed with resources provided by the North- vised. German Supercomputing Alliance (HLRN). The polymer permeability and the knowledge of how reactants Author contributions partition in the polymer are the keys to program the desired func- tion and response into a nanoreactor. Clearly, the number of ex- All authors contributed equally to this manuscript. perimental and chemical ways to synthesize a responsive hydro- gel shell (e.g., with various combinations of copolymerization) is References basically infinite. Modeling the features of diverse polymer sys- 1 S. H. Petrosko, R. Johnson, H. White and C. A. Mirkin, J. tems on various scales is therefore out of reach. Our CG and all- Am. Chem. Soc., 2016, 138, 7443–7445. atom studies so far delivered some basic but important insights 2 M. A. C. Stuart, W. T. S. Huck, J. Genzer, M. Müller, C. Ober, into the physics of these systems. However, we are continuing M. Stamm, G. B. Sukhorukov, I. Szleifer, V. V. Tsukruk, M. Ur- the endeavors towards even more refined notions of the general ban, F. Winnik, S. Zauscher, I. Luzinov and S. Minko, Nature response features of hydrogels both experimentally and theoreti- Materials, 2010, 9, 101–113. 6,177 cally. Some of such features are ions, charged reactants, and 3 S. Campisi, M. Schiavoni, C. Chan-Thaw and A. Villa, even charged (pH-responsive) hydogels, which we ignored in Catalysts, 2016, 6, 185. this review, but are of high practical relevance. Combining all the 4 Y. Lu and M. Ballauff, Prog. Polym. Sci., 2011, 36, 767–792. simulations with continuum-based approaches will help devising 5 S. Carregal-Romero, N. J. Buurma, J. Pérez-Juste, L. M. Liz- models, or at least semi-empirical rules how the hydrogel prop- Marzán and P. Hervés, Chem. Mater., 2010, 22, 3051–3059. erties, in particular the permeability of certain molecular species, 6 P. Hervés, M. Pérez-Lorenzo, L. M. Liz-Marzán, J. Dzubiella, are connected and can be tuned by stimuli. Y. Lu and M. Ballauff, Chem. Soc. Rev., 2012, 41, 5577– In order to formulate improved rate equations that carry more physical information, also the chemical processes on the nanopar- 7 S. Wu, J. Dzubiella, J. Kaiser, M. Drechsler, X. Guo, M. Bal- ticle surface in the solvent/polymer environment have to be bet- lauff and Y. Lu, Angew. Chem. Int. Ed., 2012, 51, 2229– ter understood, which we did not touch in this review. For in- stance, rate-limiting chemical intermediates could be present. 8 H. Jia, R. Roa, S. Angioletti-Uberti, K. Henzler, A. Ott, X. Lin, It would be also important to know whether and how strong the J. Möser, Z. Kochovski, A. Schnegg, J. Dzubiella, M. Ballauff (often charged) reactants and products adsorb and diffuse on the and Y. Lu, J. Mat. Chem. A, 2016, 4, 9677–9684. nanoparticle surface in the crowded polymer environment. In 9 G. Prieto, H. Tüysüz, N. Duyckaerts, J. Knossalla, G.-H. Wang some cases this may lead to steric hindrance and reaction inhibi- and F. Schüth, Chem. Rev., 2016, 116, 14056–14119. tion on the reactive surface by both reactants and products and 10 J. Gaitzsch, X. Huang and B. Voit, Chem. Rev., 2015, 116, highly nonlinear rate behavior. Here, particle-based reaction– 1053–1093. diffusion simulations may also help illuminating dynamical tran- 11 D. M. Vriezema, M. Comellas Aragonès, J. A. A. W. Elemans, sitions and collective effects during the reaction. J. J. L. M. Cornelissen, A. E. Rowan and R. J. M. Nolte, Chem. Only the fundamental understanding on all scales will enable Rev., 2005, 105, 1445–1490. us to reach the high recognition, selectivity, and feedback be- 12 K. Renggli, P. Baumann, K. Langowska, O. Onaca, N. Bruns havior in these colloidal devices as found for the nano-sized en- and W. Meier, Adv. Funct. Mater., 2011, 21, 1241–1259. 115,116 zymes. On the other hand, the large scale and diverse 13 P. Tanner, P. Baumann, R. Enea, O. Onaca, C. Palivan and building blocks that constitute the nanoreactors in various archi- W. Meier, Acc. Chem. Res., 2011, 44, 1039–1049. tectures establish the opportunity to develop many new design 14 Y. Guan and Y. Zhang, Soft Matter, 2011, 7, 6375–6384. directions within the goal of programmable, ‘intelligent’ nanopar- 15 A. Liu, C. H. H. Traulsen and J. J. L. M. Cornelissen, ACS ticle catalysis in the liquid phase. Catal., 2016, 6, 3084–3091. Journal Name, [year], [vol.], 1–17 | 13 16 S. Montolio, C. Vicent, V. Aseyev, I. Alfonso, M. I. Burguete, K. J. Stevenson and G. Henkelman, The Journal of Physical H. Tenhu, E. García-Verdugo and S. V. Luis, ACS Catal., Chemistry C, 2013, 117, 7598–7604. 2016, 6, 7230–7237. 42 J. A. Johnson, J. J. Makis, K. A. Marvin, S. E. Rodenbusch 17 A. Zinchenko, Y. Che, S. Taniguchi, L. I. Lopatina, and K. J. Stevenson, The Journal of Physical Chemistry C, V. Sergeyev and S. Murata, J. Nanopart. Res., 2016, 18, 1–9. 2013, 117, 22644–22651. 18 Y. Lu, Y. Mei, M. Drechsler and M. Ballauff, Angew. Chem. 43 E. Gross and G. A. Somorjai, Topics in Catalysis, 2014, 57, Int. Ed., 2006, 45, 813–816. 812–821. 19 J.-T. Zhang, G. Wei, T. F. Keller, H. Gallagher, C. Stötzel, 44 A. Calvo, M. C. Fuertes, B. Yameen, F. J. Williams, O. Azza- F. A. Müller, M. Gottschaldt, U. S. Schubert and K. D. Jandt, roni and G. J. A. A. Soler-Illia, Langmuir, 2010, 26, 5559– Macromol. Mater. Eng., 2010, 295, 1049–1057. 5567. 20 R. Contreras-Cáceres, A. Sánchez-Iglesias, M. Karg, 45 S. L. Brock, N. Duan, Z. R. Tian, O. Giraldo, H. Zhou and I. Pastoriza-Santos, J. Pérez-Juste, J. Pacifico, T. Hellweg, S. L. Suib, Chemistry of Materials, 1998, 10, 2619–2628. A. Fernández-Barbero and L. M. Liz-Marzán, Adv. Mater., 46 M. Ballauff and O. Borisov, Current Opinion in Colloid & 2008, 20, 1666–1670. Interface Science, 2006, 11, 316 – 323. 21 S. Li, D. Lin, J. Zhou and L. Zha, J. Phys. Chem. C, 2016, 47 M. Ballauff, Progress in Polymer Science, 2007, 32, 1135 – 120, 4902–4908. 1151. 22 M. Haruta, Chem. Rec., 2003, 3, 75–87. 48 J. Cao, S. Mei, H. Jia, A. Ott, M. Ballauff and Y. Lu, 23 G. J. Hutchings and M. Haruta, Appl. Catal., A, 2005, 291, 2 Langmuir, 2015, 31, 9483–9491. – 5. 49 Y. Lu, M. Hoffmann, R. S. Yelamanchili, A. Terrenoire, 24 Y. Zhang, X. Cui, F. Shi and Y. Deng, Chem. Rev., 2012, 112, M. Schrinner, M. Drechsler, M. W. Möller, J. Breu and M. Bal- 2467–2505. lauff, Macromolecular Chemistry and Physics, 2009, 210, 25 D. Astruc, in Nanoparticles and Catalysis, ed. D. Astruc, 377–386. Wiley-VCH Verlag GmbH, Weinheim, Germany, 2008, ch. 1, 50 M. Ballauff and Y. Lu, Polymer, 2007, 48, 1815 – 1823. pp. 1–48. 51 Y. Lu, S. Proch, M. Schrinner, M. Drechsler, R. Kempe and 26 P. Zhao, N. Li and D. Astruc, Coord. Chem. Rev., 2013, 257, M. Ballauff, J. Mater. Chem., 2009, 19, 3955–3961. 638 – 665. 52 S. Wu, J. Kaiser, X. Guo, L. Li, Y. Lu and M. Ballauff, 27 N. Li, P. Zhao and D. Astruc, Angew. Chem.-Int. Ed., 2014, Industrial & Engineering Chemistry Research, 2012, 51, 53, 1756–1789. 5608–5614. 28 E. Boisselier and D. Astruc, Chem. Soc. Rev., 2009, 38, 1759– 53 A. Lu and R. K. O’Reilly, Current Opinion in Biotechnology, 1782. 2013, 24, 639 – 645. 29 U. Taylor, C. Rehbock, C. Streich, D. Rath and S. Barcikowski, 54 M. Resmini, K. Flavin and D. Carboni, in Molecular Nanomedicine, 2014, 9, 1971–1989. Imprinting, Springer, 2010, pp. 307–342. 30 G. Sharma and M. Ballauff, Macromolecular Rapid 55 H. Jia, D. Schmitz, A. Ott, A. Pich and Y. Lu, J. Mater. Chem. Communications, 2004, 25, 547–552. A, 2015, 3, 6187–6195. 31 Y. Mei, G. Sharma, Y. Lu, M. Ballauff, M. Drechsler, T. Irrgang 56 N. Welsch, A. Wittemann and M. Ballauff, J. Phys. Chem. B, and R. Kempe, Langmuir, 2005, 21, 12229–12234. 2009, 113, 16039–16045. 32 N. C. Antonels and R. Meijboom, Langmuir, 2013, 29, 57 Y. Lu, Y. Mei, M. Drechsler and M. Ballauff, Angewandte 13433–13442. Chemie International Edition, 2006, 45, 813–816. 33 R. M. Crooks, M. Zhao, L. Sun, V. Chechik and L. K. Yeung, 58 Y. Lu, Y. Mei, M. Ballauff and M. Drechsler, The Journal of Accounts of Chemical Research, 2001, 34, 181–190. Physical Chemistry B, 2006, 110, 3930–3937. 34 R. M. Anderson, D. F. Yancey, L. Zhang, S. T. Chill, G. Henkel- 59 L.-Q. Yang, M.-M. Hao, H.-Y. Wang and Y. Zhang, Colloid and man and R. M. Crooks, Accounts of Chemical Research, Polymer Science, 2015, 293, 2405–2417. 2015, 48, 1351–1357. 60 S. Shi, Q. Wang, T. Wang, S. Ren, Y. Gao and N. Wang, The 35 C. Deraedt, N. Pinaud and D. Astruc, Journal of the Journal of Physical Chemistry B, 2014, 118, 7177–7186. American Chemical Society, 2014, 136, 12092–12098. 61 A. Chang, Q. Wu, W. Xu, J. Xie and W. Wu, Chem. Commun., 36 K. Esumi, K. Miyamoto and T. Yoshimura, Journal of Colloid 2015, 51, 10502–10505. and Interface Science, 2002, 254, 402 – 405. 62 J. Liu, J. Wang, Y. Wang, C. Liu, M. Jin, Y. Xu, L. Li, X. Guo, 37 N. Bingwa and R. Meijboom, The Journal of Physical A. Hu, T. Liu, S. F. Lincoln and R. K. Prud’homme, Colloids Chemistry C, 2014, 118, 19849–19858. and Interface Science Communications, 2015, 4, 1 – 4. 38 N. Bingwa and R. Meijboom, Journal of Molecular Catalysis 63 Y. Tang, T. Wu, B. Hu, Q. Yang, L. Liu, B. Yu, Y. Ding and A: Chemical, 2015, 396, 1 – 7. S. Ye, Materials Chemistry and Physics, 2015, 149, 460–466. 39 J.-H. Noh and R. Meijboom, Applied Surface Science, 2014, 64 J. Plazas-Tuttle, L. S. Rowles, H. Chen, J. H. Bisesi, T. Sabo- 320, 400 – 413. Attwood and N. B. Saleh, Nanomaterials, 2015, 5, 1102– 40 J.-H. Noh and R. Meijboom, Applied Catalysis A: General, 1123. 2015, 497, 107 – 120. 65 Q. Wu, H. Cheng, A. Chang, W. Xu, F. Lu and W. Wu, Chem. 41 Z. D. Pozun, S. E. Rodenbusch, E. Keller, K. Tran, W. Tang, Commun., 2015, 51, 16068–16071. Journal Name, [year], [vol.], 14 | 1–17 66 F. A. Plamper and W. Richtering, Acc. Chem. Res., 2017, 50, 95 P. Pandey and R. Chauhan, Prog. Polym. Sci., 2001, 26, 853– 131–140. 893. 67 S. Gu, S. Wunder, Y. Lu, M. Ballauff, R. Fenger, K. Rade- 96 S. C. George and S. Thomas, Prog. Polym. Sci., 2001, 26, mann, B. Jaquet and A. Zaccone, J. Phys. Chem. C, 2014, 985–1017. 118, 18618–18625. 97 M. Ulbricht, Polymer, 2006, 47, 2217–2262. 68 R. Pelton, Adv. Colloid Interface Sci., 2000, 85, 1–33. 98 R. W. Baker and B. T. Low, Macromolecules, 2014, 47, 6999– 69 A. Khokhlov, Polymer, 1980, 21, 376–380. 7013. 70 B. Erman and P. Flory, Macromolecules, 1986, 19, 2342– 99 H. B. Park, J. Kamcev, L. M. Robeson, M. Elimelech and B. D. 2353. Freeman, Science, 2017, 356, 1137. 71 A. Khokhlov, S. Starodubtzev and V. Vasilevskaya, in 100 J. K. Guesta and J. H. Prévost, Computer Methods in Applied Responsive gels: Volume transitions I, Springer, 1993, pp. Mechanics and Engineering, 2007, 196, 1006–1017. 123–171. 101 E. Atci, I. Erucar and S. Keskin, J. Phys. Chem. C, 2011, 115, 72 T. M. Barenbrug, J. Smit and D. Bedeaux, Polymer Gels and 6833–6840. Networks, 1995, 3, 331–373. 102 K. Falk, B. Coasne, R. Pellenq, F.-J. Ulm and L. Bocquet, Nat. 73 M. Heskins and J. E. Guillet, J. Macromol. Sci., Chem., 1968, Commun., 2015, 6, 6949. 2, 1441–1455. 103 A. Obliger, R. Pellenq, F.-J. Ulm and B. Coasne, J. Phys. 74 K. Dušek and D. Patterson, J. Polym. Sci. A-2 Polym. Phys, Chem. Lett., 2016, 7, 3712–3717. 1968, 6, 1209–1216. 104 M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, 75 A. Habicht, W. Schmolke, G. Goerigk, F. Lange, K. Saal- B. J. Mariñas and A. M. Mayes, Nature, 2008, 452, 301–310. wächter, M. Ballauff and S. Seiffert, J. Polym. Sci. Part B 105 G. M. Geise, H. B. Park, A. C. Sagle, B. D. Freeman and J. E. Polym. Phys., 2015, 53, 1112–1122. McGrath, J. Memb. Sci., 2011, 369, 130–138. 76 S. Zhou and C. Wu, Macromolecules, 1996, 29, 4998–5001. 106 B. Tansel, J. Sager, T. Rector, J. Garland, R. F. Strayer, 77 C. Wu and S. Zhou, Macromolecules, 1997, 30, 574–576. L. Levine, M. Roberts, M. Hummerick and J. Bauer, Sep. 78 S. Angioletti-Uberti, Y. Lu, M. Ballauff and J. Dzubiella, J. Purif. Technol., 2006, 51, 40–47. Phys. Chem. C, 2015, 119, 15723–15730. 107 C. S. Brazel and N. A. Peppas, Polymer, 1999, 40, 3383– 79 N. Pradhan, A. Pal and T. Pal, Colloids and Surfaces A: 3398. Physicochemical and Engineering Aspects, 2002, 196, 247 108 D. F. Stamatialis, B. J.Papenburg, M. Gironés, S. Saiful, – 257. S. N. M. Bettahalli, S. Schmitmeier and M. Wessling, Journal 80 T. Aditya, A. Pal and T. Pal, Chem. Commun., 2015, 51, of Membrane Science, 2008, 308, 1–34. 9410–9431. 109 W. K. Kim, A. Moncho-Jordá, R. Roa, M. Kanduˇ c and J. Dzu- 81 P. Zhao, X. Feng, D. Huang, G. Yang and D. Astruc, biella, Macromolecules, 2017, 50, 6227–6237. Coordination Chemistry Reviews, 2015, 287, 114 – 136. 110 W. K. Kim, M. Kanduc, R. Roa and J. Dzubiella, Physical 82 M. Galanti, D. Fanelli, S. Angioletti-Uberti, M. Ballauff, Review Letters, 2019, 122, 108001. J. Dzubiella and F. Piazza, Phys. Chem. Chem. Phys., 2016, 111 M. Kanduˇ c, R. Chudoba, K. Palczynski, W. K. Kim, R. Roa 18, 20758–20767. and J. Dzubiella, Phys. Chem. Chem. Phys., 2017, 19, 5906– 83 R. Roa, W. K. Kim, M. Kanduˇ c, J. Dzubiella and S. Angioletti- 5916. Uberti, ACS Catalysis, 2017, 7, 5604–5611. 112 S. Milster, R. Chudoba, M. Kanduˇ c and J. Dzubiella, Phys. 84 M. v. Smoluchowski, Z. Phys. Chem., 1917, 92, 129–168. Chem. Chem. Phys., 2019, 21, 6588. 85 P. Debye, Trans. Electrochem. Soc., 1942, 82, 265–272. 113 M. Kanduˇ c, W. K. Kim, R. Roa and J. Dzubiella, 86 G. Wilemski and M. Fixman, J. Chem. Phys., 1973, 58, Macromolecules, 2018, 51, 4853–4864. 4009–4019. 114 M. Kanduˇ c, W. K. Kim, R. Roa and J. Dzubiella, J. Phys. 87 D. F. Calef and J. M. Deutch, Annu. Rev. Phys. Chem., 1983, Chem. B, 2019, 123, 720–728. 34, 493–524. 115 M. Garcia-Viloca, J. Gao, M. Karplus and D. G. Truhlar, 88 P. Hänggi, P. Talkner and M. Borkovec, Rev. Mod. Phys., Science, 2004, 303, 186–195. 1990, 62, 251–341. 116 K. M. Ramsey, J. Yoshino, C. S. Brace, D. Abrassart, 89 H. Yasuda, A. Peterlin, C. Colton, K. Smith and E. Merrill, Y. Kobayashi, B. Marcheva, H.-K. Hong, J. L. Chong, E. D. Die Makromol. Chemie, 1969, 126, 177–186. Buhr, C. Lee, J. S. Takahashi, S. ichiro Imai and J. Bass, 90 L. M. Robeson, J. Memb. Sci.,, 1991, 62, 165–185. Science, 2009, 324, 651–654. 91 J. Williams and R. W. Baker, Journal of Membrane Science, 117 P. Atkins and J. de Paula, Physical Chemistry, W. H. Freeman 1995, 107, 1–21. and Company, 2010. 92 J. M. Diamond and Y. Katz, J. Membrane Biol., 1974, 17, 118 E. Aydt and R. Hentschke, J. Chem. Phys., 2000, 112, 5480– 121–154. 5487. 93 M. Palasis and S. H. Gehrke, J. Control Release., 1992, 18, 119 Z. Y. Lu and R. Hentschke, Phys. Rev. E, 2002, 66, 1–8. 1–11. 120 P. A. Netz and T. Dorfmüller, J. Chem. Phys., 1997, 107, 94 S. Gehrke, J. Fisher, M. Palasis and M. E. Lund, Ann. N. Y. 9221–9233. Acad. Sci., 1997, 831, 179–207. 121 A. Erba¸ s and M. Olvera de la Cruz, ACS Macro Lett, 2015, 4, Journal Name, [year], [vol.], 1–17 | 15 857–861. 151 E. Kucukpinar and P. Doruker, Polymer, 2003, 44, 3607– 122 A. Erba¸ s and M. Olvera de la Cruz, Macromolecules, 2016, 3620. 49, 9026–9034. 152 F. Rodríguez-Ropero and N. F. van der Vegt, J. Phys. 123 H. Li, A. Erba¸ s, J. Zwanikken and M. Olvera de la Cruz, Chem. B, 2014, 118, 7327–7334. Macromolecules, 2016, 49, 9239–9246. 153 M. A. Schroer, J. Michalowsky, B. Fischer, J. Smiatek and 124 J. Heyda, A. Muzdalo and J. Dzubiella, Macromolecules, G. Grübel, Physical Chemistry Chemical Physics, 2016, 18, 2013, 46, 1231–1238. 31459–31470. 125 D. Mukherji and K. Kremer, Macromolecules, 2013, 46, 154 I. Adroher-Benítez, A. Moncho-Jordá and G. Odriozola, 9158–9163. J. Chem. Phys., 2017, 146, 194905. 126 D. Mukherji, C. M. Marques and K. Kremer, Nat. Commun., 155 S. Deshmukh, D. A. Mooney, T. McDermott, S. Kulkarni and 2014, 5, 4882. J. D. MacElroy, Soft Matter, 2009, 5, 1514–1521. 127 F. Rodríguez-Ropero, T. Hajari and N. F. van der Vegt, J. 156 V. Bo¸ tan, V. Ustach, R. Faller and K. Leonhard, J. Phys. Phys. Chem. B, 2015, 119, 15780–15788. Chem. B, 2016, 120, 3434–3440. 128 F. Rodríguez-Ropero and N. F. van der Vegt, Phys. Chem. 157 L. Pérez-Fuentes, C. Drummond, J. Faraudo and D. Bastos- Chem. Phys., 2015, 17, 8491–8498. González, Soft Matter, 2015, 11, 5077–5086. 129 J. Rika, M. Meewes, R. Nyffenegger and T. Binkert, Phys. 158 E. Aydt and R. Hentschke, J. Chem. Phys., 2000, 112, 5480– Rev. Lett., 1990, 65, 657. 5487. 130 L.-T. Lee and B. Cabane, Macromolecules, 1997, 30, 6559– 159 F. A. Escobedo and J. J. de Pablo, J. Chem. Phys., 1997, 106, 6566. 793–810. 131 J. Heyda, H. I. Okur, J. Hladílková, K. B. Rembert, W. Hunn, 160 F. A. Escobedo and J. J. De Pablo, Phys. Rep., 1999, 318, T. Yang, J. Dzubiella, P. Jungwirth and P. S. Cremer, J. Am. 85–112. Chem. Soc., 2017, 139, 863–870. 161 D. Sandrin, D. Wagner, C. Sitta, R. Thoma, S. Felekyan, 132 H. Yasuda, C. Lamaze and L. D. Ikenberry, Die Makromol. H. Hermes, C. Janiak, N. de Sousa Amadeu, R. Kühnemuth, Chemie, 1968, 118, 19–35. H. Löwen et al., Physical Chemistry Chemical Physics, 2016, 133 H. Yasuda, L. Ikenberry and C. Lamaze, Die Makromol. 18, 12860–12876. Chemie, 1969, 125, 108–118. 162 H. Masoud and A. Alexeev, Macromolecules, 2010, 43, 134 L. Masaro and X. Zhu, Prog. Polym. Sci., 1999, 24, 731–775. 10117–10122. 135 B. Amsden, Macromolecules, 1998, 31, 8382–8395. 163 J. Hansing and R. R. Netz, Macromolecules, 2018, 51, 7608– 136 L. Pérez-Mas, A. Martín-Molina, M. Quesada-Pérez and 7620. A. Moncho-Jordá, Phys. Chem. Chem. Phys., 2018, 20, 164 J. Hansing and R. R. Netz, Biophysical journal, 2018, 114, 2814–2825. 2653–2664. 137 L.-H. Cai, S. Panyukov and M. Rubinstein, Macromolecules, 165 J. Hansing, J. R. Duke III, E. B. Fryman, J. E. DeRouchey and 2015, 48, 847–862. R. R. Netz, Nano letters, 2018, 18, 5248–5256. 138 B. Rotenberg, J.-F. Dufreche, B. Bagchi, E. Giffaut, J.-P. 166 H. Zhou and S. B. Chen, Phys. Rev. E, 2009, 79, 021801. Hansen and P. Turq, J. Chem. Phys., 2006, 124, 154701. 167 D. Horinek and R. R. Netz, J. Phys. Chem. A, 2011, 115, 139 K. Zhang and S. K. Kumar, ACS Macro Lett., 2017, 6, 864– 6125–6136. 868. 168 L.-C. Dong and A. S. Hoffman, J. Control. Release, 1990, 13, 140 R. Zhang and K. S. Schweizer, J. Chem. Phys., 2017, 146, 21–31. 194906. 169 S. Sasaki, S. Koga and H. Maeda, Macromolecules, 1999, 32, 141 P. M. Kekenes-Huskey, C. E. Scott and S. Atalay, J. Phys. 4619–4624. Chem. B, 2016, 120, 8696–8706. 170 R. Raccis, R. Roskamp, I. Hopp, B. Menges, K. Koynov, 142 A. Leo, C. Hansch and D. Elkins, Chem. Rev., 1971, 71, 525– U. Jonas, W. Knoll, H.-J. Butt and G. Fytas, Soft Matter, 616. 2011, 7, 7042–7053. 143 J. Haus and K. Kehr, Physics Reports, 1987, 150, 263–406. 171 Y. Kaneko, R. Yoshida, K. Sakai, Y. Sakurai and T. Okano, 144 S. K. Ghosh, A. G. Cherstvy and R. Metzler, Phys. Chem. J. Membr. Sci., 1995, 101, 13–22. Chem. Phys., 2014, 17, 1847–1858. 172 L. C. Dong and A. S. Hoffman, J. Control. Release, 1986, 4, 145 Y. Wu, S. Joseph and N. R. Aluru, J. Phys. Chem. B, 2009, 223–227. 113, 3512–3520. 173 Y. Tamai, H. Tanaka and K. Nakanishi, Macromolecules, 146 A. Börjesson, E. Erdtman, P. Ahlström, M. Berlin, T. Anders- 1994, 27, 4498–4508. son and K. Bolton, Polymer, 2013, 54, 2988–2998. 174 M. Fukuda, J. Chem. Phys., 1998, 109, 6476–6485. 147 G. E. Karlsson, U. W. Gedde and M. S. Hedenqvist, Polymer, 175 S. Goudeau, M. Charlot, C. Vergelati and F. Müller-Plathe, 2004, 45, 3893–3900. Macromolecules, 2004, 37, 8072–8081. 148 F. Müller-Plathe, Ber. Bunsenges. Phys. Chem., 1998, 102, 176 G. Marque, S. Neyertz, J. Verdu, V. Prunier and D. Brown, 1679–1682. Macromolecules, 2008, 41, 3349–3362. 149 F. Müller-Plathe, J. Membr. Sci., 1998, 141, 147–154. 177 M. Kanduˇ c, W. K. Kim, R. Roa and J. Dzubiella, ACS Nano, 150 L. Fritz and D. Hofmann, Polymer, 1997, 38, 1035–1045. 2019, 13, 11224–11234. Journal Name, [year], [vol.], 16 | 1–17 178 M. Karelson, Molecular Descriptors in QSAR/QSPR, Wiley- Interscience, 2000. 179 C. Tanford, Proc. Natl. Acad. Sci., 1979, 76, 4175–4176. 180 H. S. Ashbaugh and L. R. Pratt, Rev. Mod. Phys., 2006, 78, 181 C. Hansch, A. Leo and R. Taft, Chem. Rev., 1991, 91, 165– 182 H. Takeuchi, J. Chem. Phys., 1990, 93, 2062–2067. 183 F. Müller-Plathe, J. Chem. Phys., 1991, 94, 3192–3199. 184 L.-H. Cai, S. Panyukov and M. Rubinstein, Macromolecules, 2011, 44, 7853–7863. 185 O. Bénichou, M. Coppey, M. Moreau and G. Oshanin, J. Chem. Phys., 2005, 123, 194506. 186 R. Roa, T. Siegl, W. K. Kim and J. Dzubiella, J. Chem. Phys., 2018, 148, 065705. 187 M. J. del Razo, H. Qian and F. Noé, J. Chem. Phys., 2018, 149, 044102. Journal Name, [year], [vol.], 1–17 | 17 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Physics arXiv (Cornell University)

Modeling of stimuli-responsive nanoreactors: rational rate control towards the design of colloidal enzymes

Physics , Volume 2020 (2003) – Mar 19, 2020

Loading next page...
 
/lp/arxiv-cornell-university/modeling-of-stimuli-responsive-nanoreactors-rational-rate-control-1NY7Tt8i3x
ISSN
2058-9689
eISSN
ARCH-3341
DOI
10.1039/C9ME00106A
Publisher site
See Article on Publisher Site

Abstract

MSDE View Article Online PAPER View Journal Repertoire Builder: high-throughput structural Modeling of stimuli-responsive nanoreactors: rational modeling of B and T cell receptors† Cite this: DOI: 10.1039/c9me00020h rate control towards the design of colloidal enzymes a ab ab ab Dimitri Schritt,‡§ Songling Li,§ John Rozewicki,§ Kazutaka Katoh, a b c de Matej Kanduc, ˇ Won Kyu Kim, Rafael Roa and Joachim Dzubiella a c c ab Kazuo Yamashita,¶ Wayne Volkmuth, Guy Cavet and Daron M. Standley Repertoire Builder (https://sysimm.org/rep_builder/) is a method for generating atomic-resolution, three- dimensional models of B cell receptors (BCRs) or T cell receptors (TCRs) from their amino acid sequences. It is currently capable of handling batches of up to 10 sequences in approximately 30 minutes. This per- In modern applications of heterogeneous liquid-phase nanocatalysis, the catalysts (e.g., metal formance was achieved by applying a multiple sequence alignment extension technique originally devel- nanoparticles) need to be typically affixed to a colloidal carrier system for stability and easy han- oped for phylogenetic analysis to the template selection problem of complementarity determining region dling. “Passive carriers” (e.g., simple polyelectrolytes) serve for a controlled synthesis of the (CDR) loops. Under comparable conditions, average all-atom root-mean square deviations (RMSDs) from nanoparticles and prevent coagulation during catalysis. Recently, however, hybrid conjugates of experimentally-determined structures of CDRH3 loops in BCRs were significantly lower than tested third- nanoparticles and synthetic thermosensitive polymers have been developed that enable to change party high-throughput modeling methods, including ABodyBuilder, PigsPro, and LYRA. For TCRs, similar the catalytic activity of the nanoparticles by external triggers. In particular, nanoparticles embed- trends were observed when Repertoire Builder was compared with TCRmodel and LYRA. We also found ded in a stimuli-responsive network made from poly(N-isopropylacrylamide) (PNIPAM) have be- that Repertoire Builder model errors were, in general, lower than those produced by our earlier Kotai Anti- come the most-studied examples of such hybrids. It has been demonstrated that the permeability body Builder, even when CDRH3 loop refinement was used. However, in a subset of cases, which could be of the polymer network and thus the reactant flux can be switched and controlled by external distinguished by poor Repertoire Builder scores, refinement by Kotai Antibody Builder or Rosetta Antibody, stimuli. Such “active carriers” may thus be viewed as true nanoreactors that open up new design Received 12th February 2019, both of which utilize extensive structural sampling, improved the third heavy chain CDR (CDRH3) RMSD on routes in nano-catalysis and elevate synthesis to create highly selective, programmable “colloidal Accepted 21st May 2019 average. Taken together, these results indicate that the MSA extension approach used by Repertoire Builder enzymes”. However, only a comprehensive understanding of these materials on all time and resulted in a favorable balance between speed and accuracy when compared to alternative methods. Fur- DOI: 10.1039/c9me00020h length scales can lead to a rational design of future, highly functional materials. Here we review thermore, we conclude that more sensitive scoring, rather than extended structural sampling, is needed to the current state of the theoretical and multi-scale simulation approaches, aiming at a fundamental rsc.li/molecular-engineering further improve the accuracy of BCR and TCR modeling. understanding of these nanoreactors. In particular, we summarize a theoretical approach for re- action rates of surface-catalyzed bimolecular reactions in responsive nanoreactors in terms of the key material parameters, the polymer shell permeability P and the reactant partition ratio K . We Design, System, Application discuss recent computer simulation studies of both atomistic and coarse-grained polymer models Repertoire Builder is a tool for building 3D models of B cell receptors (BCRs) or T cell receptors (TCRs) to atomic resolution. The strategy used by Repertoire Builder is an application of the multiple in which sequence thesealignment quantities (MSA) haextension ve beenfeature character of the iz MAFFT ed insoftware. some detail. The particula Wer conclude application h with ere isan outlook to represent structural templates for each complementarity-determining region (CDR) of a given length by a single MSA. By repeatedly applying the MSA ex- on selected open questions and future theoretical challenges in nanoreactor modeling. tension method, a complete set of templates that covers the 3 CDRs and 1 framework for each chain can be obtained. The input must be either paired (heavy and light) or unpaired (heavy or light) chain amino acid sequence of the variable region for the receptor in question (BCR or TCR). The immediate application of Repertoire Builder is to render 3D models in a high-throughput and accurate manner, in order to allow structure-based analyses for BCR or 1 TCR Intr sequ oduction ence data. Because the volume of such data is currently growing exponentially, Repertoire Builder represents a unique approach to large-scale 1–12 10–14 reactions to biosensors for the diagnosis of diseases. Ex- BCR or TCR repertoire data analysis. Synthetic nanoreactors are an emerging and promising new nan- amples of natural nanoreactors are lipid-based membranes (e.g., otechnology for liquid-phase heterogenous catalysis. In these liposomes), cage-like proteins (e.g., ferritins), protein-based bac- 11–13,15 nanoreactors, the catalysts are confined in a permeable nanos- terial microcompartments, and viruses. Artificial nanore- Introduction tructure, which acts as a carrier and can be used to shelter and actors (based on spherical polyelectrolyte brushes, dendrimers, control the catalytic processes. In particular, the catalysis can be Recent ligands, single-cell or even resolution DNA) aresequencing simpler than technologies the naturalcan ones elu- and thus 4–12,16,17 Immunology Frontier Research Center, Osaka University, 3-1 Yamadaoka, Suita, made selective and responsive if the nanoreactor permeability dif- cidate easier the to natively controlpaired for targeted (heavy– applications. light) B cell receptor (BCR) Osaka 565-0871, Japan. E-mail: standley@biken.osaka-u.ac.jp ferentiates among molecules and can be modulated by external and T cell receptor (TCR) sequences in a high-throughput Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, 1–14 In particular, nanoreactors containing metal nanoparticles have stimuli. These nanoreactors can be used for a large variety manner. Although the coverage has not yet reached that of Suita, Osaka 565-0871, Japan 4–9,18–21 emerged as a promising catalytic system. For exam- bulk sequencing methods, state-of-the-art paired sequencing of applications, ranging from analytical tools to study chemical Atreca Inc, 500 Saginaw Drive, Redwood City, CA, 94063-4750, USA ple, gold becomes an active catalyst when divided down to the † Electronic supplementary information (ESI) available. See DOI: 10.1039/ platforms can yield thousands of unique receptor sequences 22–27 c9me00020h nanophase. However, the handling of the particles in the in a single experiment. Because of the critical role of B and T ‡ Current address: Department of Biosciences and Nutrition, Karolinska cellsliquid in the phase prevention is an important or progression problem: The of disease, surface ofnew the parti- Jožef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia Institutet. methods cles should to functionally be easilyanalyze accessible emerging for the sequence mixture ofdata theare reactants. Korea Institute for Advanced Study, 85 Hoegiro, Seoul 02455, Republic of Korea. § Equal contribution. Departamento de Física Aplicada I, Facultad de Ciencias, Universidad de Málaga, needed. Structural modeling can, in principle, contribute to This condition would require the nanoparticles to be freely sus- ¶ Current address: KOTAI Biotechnologies Inc, 3-1 Yamadaoka, Suita, Osaka Campus de Teatinos s/n, E-29071 Málaga, Spain such functional analysis, since structures allow physical and 565-0871. pended in the solution, and coagulation or any type of Ostwald Research Group for Simulations of Energy Materials, Helmholtz-Zentrum Berlin, ripening of the nanoparticles should not occur during the cat- Hahn-Meitner-Platz 1, D-14109 Berlin, Germany alytic reaction. Also, leaching of metal or loss of nanoparticles Institut für Physik, Albert-Ludwigs-Universität Freiburg, Hermann-Herder-Str. 3, D- This journal is © The Royal Society of Chemistry 2019 Mol. Syst. Des. Eng. 79104 Freiburg, Germany from the carrier must be prevented to ensure a meaningful and Journal Name, [year], [vol.], 1–17 | 1 Open Access Article. Published on 21 June 2019. Downloaded on 6/27/2019 2:27:17 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence. arXiv:2003.09001v1 [physics.chem-ph] 19 Mar 2020 repeated use of the catalyst. The latter requirements necessitate bulk solution a suitable carrier that ensures a safe and repeated handling of 28,29 the nanoparticles. It was demonstrated that suitable carrier 30,31 32–42 systems include colloidal particles, dendrimers, meso- 43–45 46,47 porous materials, spherical polyelectrolyte brushes, and C other systems structured on a length scale between one and a polymer few hundred nm. shell In recent years, the concept of such carrier systems has been further advanced with the synthesis of hydrogel-based nanoreac- tors, for which rate control by external stimuli has become pos- B ΔG(r ) 6,7,10,12,15,18,49–57 sible. Thermosensitive hydrogels made from ΔG a network of poly(N-isopropylacrylamide) (PNIPAM) and its 4–8,18–21,58–67 copolymers provide a good example: Typical col- loidal carrier architectures are of core-shell or yolk-shell type where the polymer gel constitutes a permeable shell around a in solid core (core-shell) or around a hollow void (yolk-shell). The core can be the nanoparticle itself, cf. Fig. 1, or, for example, a NR polystyrene core. The catalytic nanoparticles can be located dur- ing synthesis in a well controlled fashion, e.g., into the voids, Fig. 1 Bimolecular reactions in core-shell nanoreactors. (A) Two reac- onto the cores, or distributed within the polymer shell. Some- tants, A and B, diffusing from a bulk solution, generate a product, C, in the times simply a pure hydrogel (nano- or microgel) particle is the proximity of a catalyst nanoparticle (central yellow sphere) embedded in carrier for the catalysts. In this case carrier and polymer shell in a PNIPAM polymer network. (B) Schematic representation of a core-shell our context are essentially the same. A survey of selected but very nanoreactor. A nanoparticle of radius R is embedded in a spherical shell of outer radius R . The shell permeability depends on the diffusivity, typical experimental architectures and results for polymer-based NR D(r), and on the transfer free energy profiles, DG(r). We model both as nanoreactor carriers is provided in Tab. 1. step functions with values D and DG inside, and D and zero reference in 0 The responsive polymer shell is in a swollen hydrophilic state at outside the shell, respectively. low temperature, but sharply collapses into a rather hydrophobic state above the critical solution temperature. The sharp volume 69–75 transition of the gel is reversible and depends on the temper- 76,77 ature, or more general, solvent quality. This has substantial responsive nanoreactors can be described by combining a thermo- influence on reactant partitioning close to the catalysts as well as dynamic two-state model for the polymer volume transition with reactant transport towards it. Hence, there are two key roles of the appropriate reaction–diffusion equations. In particular, the the polymer shell. On the one hand, the shell acts as a integral important effect of a change in the shell permeability on the reac- part of the whole carrier that protects nanoparticles from aggre- tants approach to the catalyst’s surface can be described by theory 7,78,82,83 gation and hinders chemical degradation processes, e.g., oxida- of diffusion through an energy landscape, in the spirit of 84–88 tion. On the other hand, the polymer ability to switch between Debye–Smoluchowski diffusion-controlled rate theory. Im- states with different physicochemical properties upon changes in portantly, the latter also takes into account the local reactant environmental parameters, e.g., temperature, pH, or concentra- concentration, i.e., the partitioning inside the polymer shell close tion of certain solutes, provides a handle to actively control the to the catalyst. Recently, we have presented an extended theory nanoreactor’s catalytic properties. of diffusion-limited bimolecular reactions in nanoreactors, which A quantitative study and understanding of a nanoreactor re- can be employed to rationally design the activity and selectivity 78,83 quires kinetic data measured with the highest precision possible. of a nanoreactor. The main result of our consideration was Up to now, most of the testing of the catalytic activity of nanopar- the following formula for the total catalytic rate in bimolecular ticles in aqueous phase has been done using the reduction of 4- reactions in core-shell nanoreactors (cf. Fig. 1): 79 36 nitrophenol by borohydride. Pal et al. and Esumi et al. have 1 k k A B been the first who have demonstrated the usefulness of this re- D D k = + k A + k B tot D D 2 k action. In the meantime, the reduction of 4-nitrophenol has be- R come the most used model reaction for the quantitative testing and analyzing of the catalytic activity of nanoparticles in the liq- k Ak B D D + k A + k B 4k Ak B : (1) 80,81 D D D D uid phase. Further examples of catalytic reactions in aqueous solution studied in this system are the reductions of nitrobenzene 6,7,18,19 A B and hexacyanoferrate (III) by borohydride ions and the Here, k A(P ) and k B(P ) are the diffusion rates of the reactants D D 8 i decomposition of methyl orange under visible light. A and B, which explicitly depend on the shell permeability P , and A B All the aforementioned examples deal with surface-catalyzed k (K ;K ) is the surface reaction rate, explicitly depending on partitioning K as defined below. bimolecular reactions, being a very common type. As pointed out 5–7,78 before, pseudo-unimolecular surface-catalyzed reactions in In general, permeability of a material defines the ability of the Journal Name, [year], [vol.], 2 | 1–17 Table 1 Survey of selected publications on responsive nanoreactor catalytic experiments with different architectures. Polymer abbreviations: poly(N-isopropylacrylamide) (PNIPAM), polysterene (PS), maleated carboxymethylchitosan (MACACS), poly(N-vinylcaprolactam) (PVCL), poly(styrene-NIPAM) (P(S-NIPAM)), poly(NIPAM-co-methacrylic acid) (P(NIPAM-co-MAA)), poly(NIPAM-co-2-(dimethylamino)ethyl methacrylate) (P(NIPAM-co-AMPS)), poly(4-vinylphenylboronic acid-co-NIPAM-co-acrylamide) (P(VPBA-NIPAM-AAm)). Solute abbreviations: 4-nitrophenol (NP), 4-aminophenol (AP), nitrobenzene (NB), aminobenzene (AB), hexacyanoferrate(III) (HCF), hexacyanoferrate(II) (HCF2), o-nitrophenyl-D-glucopyranoside (oNPG). Ref. Architecture Core - Polymer Catalyst Reaction Result 7 Yolk-shell Au - PNIPAM Au reduction: NP ! AP & NB ! AB T -dependence on gel swelling and reaction rate 5,6 Core-shell Au - PNIPAM Au reduction: HCF ! HCF2 T -dependence on gel swelling and reaction rate, rate dependence on nanoreactor concentration and cross-linking density 6,67 Core-shell Pt/Au - PNIPAM Pt/Au reduction: NP ! AP rate dependence on reactant concentration 8 Core-shell Cu O - PNIPAM Cu O decomposition by visible light: methyl orange T -dependence on gel swelling and reaction rate 2 2 21 Core-shell Au - PNIPAM Ag reduction: NP ! AP photoresponsive gel size and reaction rate 65 Core-shell Au - P(VPBA-NIPAM- Au reduction: NP ! AP & NB ! AB glucose concentration dependence on gel swelling and reaction AAm) rate 4,51 Core-shell PS - PNIPAM Au/Pt/Rh oxidation: benzyl alcohol ! benzaldehyde T -dependence on gel swelling and reaction rate 4,56 Core-shell PS - PNIPAM b-D- hydrolysis: oNPG! D-glucose + o-nitrophenol T -dependence on gel swelling and reaction rate glucosidase 57 Core-shell PS - PNIPAM Ag reduction: NP ! AP T -dependence on gel swelling and reaction rate 59 Core-shell P(S-NIPAM) - Ag reduction: NP ! AP & NB ! AB T -dependence on gel swelling and reaction rate P(NIPAM-co-MAA) 19 Microgel PNIPAM/MACACS Ag reduction: NP ! AP T -dependence on gel swelling and reaction rate 55 Microgel PVCL-a-cyclodextrin Au reduction: NP ! AP T -dependence on gel swelling and reaction rate 60 Hydrogel P(NIPAM-co-MAA) Au reduction: NP ! AP T -dependence on gel swelling and reaction rate 61 Microgel cellulose cellulase hydrolysis: cellulose ! glucose T -dependence on gel swelling and time-dependent product con- (enzyme) centration 62 Microgel P(NIPAM-co-AMPS) Ni reduction: NP ! AP T -dependence on gel swelling and reaction rate 63 Microgel P(NIPAM-co-AMPS) Ag reduction: methylene blue T -dependence on gel swelling and reaction rate, pH-dependence on gel swelling penetrating molecules (e.g., gas, ligands, reactants, etc.) to per- responsive nanoreactors to reach the high recognition, selectiv- 115,116 meate and flow through a given medium under the action of an ity, and feedback control as found for enzymes, to create external field or chemical gradient. It is thus without doubt one of “colloidal enzymes”. Here we review the state-of-the-art of the the most fundamental transport descriptors employed in the phys- current understanding of the intricate links between nanoreactor ical sciences and material engineering. In the standard ‘solution– reaction rate and polymer permeability. Most of the results pre- diffusion’ picture for permeable membranes, it is commonly de- sented here are based on our recent research endeavor of multi- 89–99 fined on the linear response level by scale modeling schemes of hydrogel systems in order to establish rational design principles of responsive nanoreactors. We start P = K D ; (2) in in Section 2 by summarizing the rate theory for nanoparticle- catalyzed bimolecular reactions including partitioning and per- where meability of the polymer shell. The key property to be tuned in K = (3) and ‘programmed’ during the synthesis in order to select and switch catalytic activity is the permeability. In Section 3 we thus is the partition ratio, in this work simply referred to as partition- proceed with mesoscale coarse-grained computer models, which ing, defined as the ratio of number densities of the solutes in- give fundamental insights on partition–diffusion correlations in side (c ) and outside (c ) the medium in equilibrium, and D is in 0 in the permeability and how they can be tuned qualitatively by mi- the diffusion coefficient of those inside. Permeability can be thus croscopic interactions. In the last part we turn to atomistically- defined as the inverse of a diffusional resistance of a medium re- resolved molecular simulations of the PNIPAM hydrogel models garding the total mass transport (flux) towards the catalyst driven in swollen and collapsed state. Here we address the question of by the reaction. The optimization of permeability, especially the influence of the ‘chemistry’ of the interactions, e.g., role of for being highly selective among different solutes, has been a 95,99,100 (temperature-dependent) hydration, polarity, reactant type and grand challenge in material design over the last decades. size, etc. Obviously, there are many open questions, missing con- Prominent applications revolve around gas separation and re- 90,95,99,101–103 nections, and remaining challenges to overcome to obtain a com- covery, desalination and nanofiltration (‘molecular 104–106 prehensive multi-scale model. We will briefly discuss those and sieving’), medical treatments by dialysis or selective drug 107,108 give an outlook in the final, concluding section. transport, hydrogel-based soft sensors, and the nanoreac- 2,4,83,93 tors. We have studied partitioning and permeability of 2 Bimolecular reactions in nanoreactors polymer networks and PNIPAM polymers recently on the molec- 109,110 ular level by coarse-grained as well as all-atom molecular 2.1 Macroscopic rates and dependence on permeability 111–114 dynamics computer simulations. We review the rate theory for nanoreactors for the case of surface- Hence, in the last couple of years many quantitative concepts catalyzed bimolecular reactions in one of the simplest nanoreac- have emerged both on the continuum and the microscopic level tor geometries, a core-shell nanoreactor, depicted in Fig. 1A, that will eventually lead to a more fundamental understanding where a catalytically active metal nanoparticle of radius R is em- of nanoreactors in the future. The possibility arises to optimize bedded in a thermoresponsive hydrogel matrix of outer radius Journal Name, [year], [vol.], 1–17 | 3 R . In this spherically symmetric system, we consider that two the width equal to the polymer shell width (see Fig. 1B), i.e. NR species A and B diffuse from a bulk solution with respective (ini- A B D R r R ; NR tial) concentration c and c through the polymer shell towards i in 0 0 D (r) = (7) the catalyst nanoparticle. A fraction of the reactants arriving at D elsewhere; the surface combines with each other to produce a third molec- and ular species C. Assuming a total concentration of nanoreactors ( DG R r R ; NR c , the experimentalist would measure the transformation of a NR DG (r) = (8) 0 elsewhere: reactant (say reactant A) per time, according to i i A Here, D and D stand for the diffusion coefficients in the poly- dc (t) in 0 A B =k (c (t);c (t))c (4) i tot NR mer shell and solution, respectively. DG represents the transfer dt free energy from bulk water into the shell and as such strongly A B with instantaneous bulk concentrations c (t) and c (t), and k tot depends on the state (swollen/collapsed) of the nanoreactor. Us- has the units of inverse time and is a non-trivial function of the ing standard thermodynamic relations, we connect the flux of the reactant concentrations. In general, and as we will see below, the i species i to its local concentration c (r) chemical reaction has no well-defined order. In some limits, e.g., i i i i in an abundance of species B, it may reduce to pseudo unimolec- J =D c Ñbμ ; (9) ular or even pseudo first-order kinetics (see also Section 2.2 where μ (r) is the chemical potential of the species i, and b = later). 1=k T , with k denoting the Boltzmann’s constant and T the ab- B B To derive the functional form of the total catalytic rate, we as- solute temperature of the system. The chemical potential of a sume k (number of molecules reacting per unit of time) is equal tot molecule interacting with an external environment with a spa- to the radial flux of reactants at the nanoparticle surface. In bi- tially dependent concentration and free energy is molecular reactions, then the fraction of molecules A reacting is proportional to the number of molecules B at the same location, i i bμ = ln +bDG ; (10) and vice versa. Thus, k can be estimated through the standard tot 78,117 ref mean-field relation where c is a reference concentration whose value can be chosen A B ref k = K c (R)c (R); (5) tot vol arbitrarily. Equation (10) can now be used to relate the transfer A B free energy DG and partitioning, eqn (3), where c (R) and c (R) are the reactant concentrations at the nanoparticle surface, and K the probability that the two species vol i i K = exp bDG : (11) react on the surface (with units per time and per concentration squared). To calculate k , we solve the stationary continuity tot With the aforementioned definitions, the shell permeability to the equation for the density fields of reactants, species i is calculated as ÑJ = 0; (6) i i i P = K D (12) in with J (r) being the radial flux of the species i = A, B, C as a func- i i where for r > R we have P = D . NR tion of the distance from the nanoparticle. We make the station- arity assumption that the system is always in a steady-state and there is no explicit time-dependence of the fluxes. In other words, we assume the microscopic relaxation of the system, roughly given by the time of reactants to diffuse through the nanoreac- We found that the total catalytic rate for bimolecular reac- tor R =D , is faster than the reaction time as defined in eqn (4). NR tions in responsive nanoreactors is obtained as in eqn (1). In this If we use the fastest, diffusion-controlled (Smoluchowski) rate expression, k ' k = 4pRD c as the reaction rate scale, we find the condi- tot 0 0 A B 2 5 A bDG (R) B bDG (R) A B A B tion for stationarity that c  1=(4pR R) ' 10 mol/l for typ- k = K c e c e = K c c K K (13) NR R vol vol 0 0 0 0 ical geometries where R ' 1 nm and R ' 10 nm. In experi- NR stands for the surface-part of the reaction rate, which is explicitly ments, typically sub-micromolar reactant concentrations are used partitioning-dependent, and and the reaction rate is at least 1–2 orders slower than the fastest, fully diffusion-controlled limit, so that the condition is in most Z cases very well satisfied. k i = 4pc dr (14) D 0 i 2 P (r)r In their diffusive approach to the catalyst nanoparticle, reac- tants have to permeate the shell. The kinetics of this process is is the permeability-dependent diffusion part of the reaction rate i i thus governed by the shell permeability, which depends on the of the reactant i. In the absence of the shell, P (r) = D , and the i 0 i i diffusivity profile, D (r), and on the thermodynamic barrier, i.e., diffusion rate turns into the Smoluchowski rate k = 4pRD c . 0 0 the transfer free energy between bulk and shell, DG (r). For sim- For the core-shell configuration depicted in Fig. 1B the step pro- plicity, we take both profiles to be shell-centered step functions of files in eqn (7) and eqn (8) apply and the relation between the Journal Name, [year], [vol.], 4 | 1–17 treated as pseudo-unimolecular (see also next Section 2.2), that 0.08 HCF HCF is, dc (t)=dt =k c (t). obs The temperature dependence of the measured pseudo-first- 0.06 order constant is shown by blue filled squares in Fig. 2A. We observe that the reaction rate decreases by one order of mag- 0.04 nitude when the temperature of the solution exceeds the lower critical solution temperature (LCST) of the PNIPAM polymer. The measured nanoreactor hydrodynamic radius data, displayed by 0.02 orange open circles in Fig. 2A, exhibit the well-known volume transition between the swollen and the collapsed states below and above the LCST, respectively. As we pointed out before, this 280 280 290 290 300 300 310 310 320 320 transition changes the physicochemical properties of the polymer, TT (K) (K) which leads to different reactant diffusivity and transfer free en- 0.25 ergy values resulting in a nanoreactor permeability switch at the LCST. In the diffusion-controlled limit for pseudo-unimolecular 0.2 HCF HCF HCF kinetics we can identify k = k c =c . The c is the obs NR D 0 0 initial bulk concentration, which can be replaced by the instan- 0.15 HCF taneous c (t) in the equations during the reaction because of the stationarity assumption. Using eqn (15) together with the 0.1 experimental data from Fig. 2A, we estimate the temperature de- pendence of the nanoreactor shell permeability for HCF (Fig. 2B) 0.05 and clearly observe the aforementioned permeability switch be- low and above the LCST. The permeability decreases around one 280 290 300 310 320 order of magnitude from the swollen to the collapsed state. By T (K) comparing Figs. 2A and B we clearly see that the nanoreactor shell permeability is the essential ingredient to understand the Fig. 2 (A) Temperature dependence of the measured pseudo-first-order reaction rate response of nanoreactors in diffusion-controlled re- constant k (blue filled squares) of the electron-transfer reaction be- obs tween HCF and borohydride ions in Au-PNIPAM nanoreactors. The mea- actions. Detailed mesoscopic and microscopic insights on the in- sured temperature dependence of the nanoreactor hydrodynamic radius fluence of effective interaction potentials, hydrogel density, and is shown by orange opened circles. All data were taken from Ref. 5. chemistry on the permeability of polymer shells is presented in (B) Temperature dependent nanoreactor shell permeability for HCF esti- Sections 3 and 4. mated using eqn (15). 2.2 Pseudo-unimolecular reactions in nanoreactors Bimolecular reactions are typically treated as pseudo- shell permeability and the diffusion rate, eqn (14), simplifies to unimolecular when one of the reactants is in large excess " ! # with respect to the other. The reasoning behind this assumption k i R D 0 = 1+ 1 1 : (15) i is that, according to the simple Smoluchowski rate, the reactant P R i NR in larger concentration would diffuse towards the nanoparticle Equation (1) is the main analytical result for nanoparticle surface at a much larger rate than the other one. Therefore, surface-catalyzed bimolecular reactions. It shows that, in the when the reactant in limiting concentration arrives to the catalyst, it will always find a reactant of the other species to fully bimolecular case, the diffusional fluxes of the different re- actants are coupled. Thus, k depends in a non-trivial way on the combine with. However, this is not always true when considering tot surface and the diffusion rates and nanoreactor shell permeabil- nanoreactors. In this case, the diffusion rate, eqn (14), not only ity, in contrast to the simple unimolecular case (i.e., in general depends on the bulk reactant concentration but also on the shell 1 1 1 permeability and thus on the molecular interactions of reactants k 6= k + k in bimolecular reactions). tot D R Equation (1) together with eqs. (13) and (15) can be used with the shell. It is thus the combination of both quantities that determines whether a bimolecular reaction can be treated as to predict the total catalytic rate once the nanoreactor shell pseudo-unimolecular or not. permeability and the reactant partition ratios are known (e.g., from experiments, or from simulations, see Sections 3 and 4), If one of the reactants has a much larger diffusion rate than or, conversely, to extract the parameters by fitting to experi- the other one, e.g., k B  k A , the total reaction rate, eqn (1), D D mental data. Carregal-Romero et al. investigated the bimolec- reduces to (see Supporting Information in Ref. 83) ular electron-transfer reaction between hexacyanoferrate (III), 1 1 1 Fe(CN) (HCF), and borohydride BH ions in Au-PNIPAM core- k ! k = k + k ; (16) 6 4 tot A tot R shell nanoreactors. In a previous work we demonstrated that that bimolecular reaction is diffusion-controlled and can be which is the well-known expression of the total reaction rate Journal Name, [year], [vol.], 1–17 | 5 -1 PP / D k (s ) <latexit sha1_base64="q1hBcpw8Oxqa2w7zfIic8raScM8=">AAAB8nicbVDLSsNAFL3xWeur6tLNYBFclaQKiquCG5cV7APaUCbTSTt0MgkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeIJHCoOt+O2vrG5tb26Wd8u7e/sFh5ei4beJUM95isYx1N6CGS6F4CwVK3k00p1EgeSeY3OV+54lrI2L1iNOE+xEdKREKRtFKvX5EccyozJqzQaXq1tw5yCrxClKFAs1B5as/jFkacYVMUmN6npugn1GNgkk+K/dTwxPKJnTEe5YqGnHjZ/PIM3JulSEJY22fQjJXf29kNDJmGgV2Mo9olr1c/M/rpRje+JlQSYpcscVHYSoJxiS/nwyF5gzl1BLKtLBZCRtTTRnalsq2BG/55FXSrte8y1r94arauC3qKMEpnMEFeHANDbiHJrSAQQzP8ApvDjovzrvzsRhdc4qdE/gD5/MHhf+RYg==</latexit> 0 obs LCST LCST R (nm) NR control), but the values in the polymer shell, which can be the limiting factor. The latter is defined by the shell permeability P -3 B A P P <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> /<latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> =10 and can thus strongly differ from the bulk value. Because of the -2 B A P /P =10 <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit><latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> 100 responsive nature of the gating shell of nanoreactors, this depen- -1 B A P P <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> /<latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> =10 dence crucially implies that the identity of the limiting reactant B A P P <latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> /<latexit sha1_base64="QtbWTbVFh0i4IokLzsgNgkqnRVA=">AAAB8nicbVDLSsNAFL3xWeur6tJNsAgupCRV0GXBjcsK9gFtKJPppB06mQkzN0IJ/Qw3LhRx69e482+ctFlo64GBwzn3MueeMBHcoOd9O2vrG5tb26Wd8u7e/sFh5ei4bVSqKWtRJZTuhsQwwSVrIUfBuolmJA4F64STu9zvPDFtuJKPOE1YEJOR5BGnBK3U68cEx5SIrDkbVKpezZvDXSV+QapQoDmofPWHiqYxk0gFMabnewkGGdHIqWCzcj81LCF0QkasZ6kkMTNBNo88c8+tMnQjpe2T6M7V3xsZiY2ZxqGdzCOaZS8X//N6KUa3QcZlkiKTdPFRlAoXlZvf7w65ZhTF1BJCNbdZXTommlC0LZVtCf7yyaukXa/5V7X6w3W1cVnUUYJTOIML8OEGGnAPTWgBBQXP8ApvDjovzrvzsRhdc4qdE/gD5/MHgcmRVA==</latexit> =1 can switch upon a change in the external stimulus. Failure to recognize this fact can lead to very large discrepancies between 10 the correct and the approximate rate. This theoretical framework for pseudo-unimolecular reactions qualitatively rationalizes the large and sharp variations in catalytic rate observed in the rele- 6–8,78 vant nanoreactor experiments. 3 Partitioning and diffusion: coarse-grained 0 1 2 3 4 10 10 10 10 10 B A simulations c / c 0 0 As we have just described, key parameters to understand a Fig. 3 Total rate for unimolecular reactions k , eqn (16), divided by the tot nanoreactor’s selectivity and rate response to stimuli are the total reaction rate k for bimolecular reactions, eqn (1), as a function of tot B A permeability of its polymeric shell and the reactant partitioning the relative reactant bulk concentration, c =c . The lines stand for dif- 0 0 B A within. In the following, we review two selected coarse-grained ferent relative nanoreactor shell permeabilities to the reactants, P =P . We assume k = k . A (CG) simulation studies of partitioning, diffusion, and permeabil- 109,110 ity in model membranes. Mesoscopic models, neglecting chemical resolution, play a pivotal role not only as a bridge be- tween the aforementioned macroscopic reaction model and the in unimolecular reactions, k . In this case, the total catalytic tot following microscopic all-atom models but also for the deeper time is the sum of the diffusion time of the slower reactant and understanding of essential physics, e.g., of molecular adsorption the surface reaction time. Hence, in nanoreactors, unimolecu- and transport in polymer systems. Particle-based simulations on lar reactions can be diffusion- or surface-controlled if k  k D R various scales with increasing complexity and chemical detail are or k  k , respectively. If both rates are comparable in magni- D R now emerging. For the convenience of the reader, we have sum- tude, the reaction is termed diffusion-influenced. Analogously, a marized selected relevant simulation efforts in Tab. 2. reaction is diffusion- or surface-controlled if Da  1 or Da  1, II II respectively, where Da = k =k is the second Damköhler num- II tot D 3.1 Influence of gel volume transition on reactant partition- ber. If both reactants diffuse from the bulk solution, according B B A A ing in a model polymeric membrane to eqn (14), this condition is satisfied when c P  c P . This 0 0 Responsive polymers feature a sharp volume transition where the means that one of the reactants should be in a much higher bulk density of the polymer drastically changes. The partitioning of concentration and/or subject to a much larger shell permeability reactants across the volume transition and the feedback of the than the other. polymer to the permeation is complex and poorly understood. We In Fig. 3 we analyze how large should be the excess of re- thus first discuss a CG simulation model with details described actant B for the pseudo-unimolecular reaction limit to be valid. previously consisting of permeating reactants in a polymer- This value depends on the relative nanoreactor shell permeabil- B A based thin membrane (Fig. 4A), where we aim at a qualitative ity, P =P . For simplicity, we consider that the surface rate is study of the effects of structural transitions of gels on the reactant equal to the diffusion rate of the reactant in limiting concentra- partitioning and its back-coupling to the volume transition. In tion (k = k A , diffusion-influenced reaction). When both reac- the following, we refer to the permeating reactants generally as tants have the same permeability (red line), the concentration of ‘solutes’. reactant B should be roughly 10 times larger than the one of A to have a unimolecular reaction. If we then decrease the shell The membrane is constructed as cross-linked semi-flexible net- 118–123 permeability to the reactant B by 10 times, its concentration has work of polymers formed on a regular cubic lattice, and thus to become 100 times higher with respect to that of A to keep the solutes can diffuse throughout the membrane and the bulk this limit. Figure 3 also shows that the catalytic rate predicted regions. This enables a direct sampling of the solute partition- for a pseudo-unimolecular reaction for the reactant in limiting ing from the simulations simply according to eqn (3). We use concentration may differ from the fully bimolecular one by or- the Lennard-Jones (LJ) pair potential and its size unit s as the ders of magnitude. Thus, when dealing with nanoreactors, it is diameter for all particles and the monomer–monomer (bonded) necessary to consider not only the difference between the bulk distance in the polymers. In order to model such a gel in the concentrations of the reactants but also the difference in the shell presence of various solutes, we employ inter- and intra-particle permeability to the reactants. interactions in terms of LJ pair potentials. We focus on two Hence, when considering nanoreactors, care should be taken key interaction parameters: The membrane–membrane interac- since in these systems it is not the bulk mobility and concentra- tion e controls the solvent quality, turning it from good to poor mm tion that determine the reaction type (diffusion versus surface upon the increase of e . The membrane–solute interaction e mm ms Journal Name, [year], [vol.], 6 | 1–17 k / k tot tot solute-induced collapsed swollen 1 4 membrane solute Solute-involved collapsed Solute-induced collapsed 1 5 Solute-adsorbed collapsed Critical 3 6 collapsed 3 6 Collapsed Swollen Fig. 4 (A) Various conformational states and regimes in the mesoscopic network membrane–solute system from CG computer simulations are depicted in the main phase diagram (center) depending on the Lennard-Jones (LJ) interactions e and e : (1) is a ‘solute-induced collapsed’ state, (2) is a mm ms ‘critical’ transition line (yellow contour) between the intrinsic (3) ‘swollen’ and (6) ‘collapsed’ states. (4) is a ‘solute-involved collapsed’ state, while (5) is a ‘solute-adsorbed’ collapsed state where solutes adsorb mostly on the membrane surface. For details see text. Reprinted with permission from Ref. 109, copyright 2017 American Chemical Society. (B) Solute partitioning landscape K depending on e and e . mm ms governs the membrane–solute coupling and thus models different only on the membrane–solute interaction but also significantly on kinds of solutes. For the solute–solute interaction we always use the solvent quality. The partitioning overall becomes large (i.e., e = 0:1 k T , essentially being a steep (r ) repulsion. higher adsorption) as the membrane–solute attraction, e , in- ss B ms creases, while it has large regions of unity in the swollen states In Fig. 4A we show the landscape of the gel structural phases, (light blue-greenish areas). Note that when compared with the depending on both interaction parameters. The red region de- structural landscape in Fig. 4A, both extrema of K (i.e., the min- picts swollen states, where the gel volume is relatively large, imum and the maximum) are in the collapsed regions, indicating while the blue region indicates collapsed states. Without the so- that the collapsed phase can relate to extremely different parti- lutes (e = 0) our model exhibits a collapse transition at around ms tionings and there is no unique mapping. Moreover, at interme- e ' 1:0 k T . In the presence of the solutes, however, the pic- mm B diate values of e , K is a nonmonotonic function of the solvent ms ture becomes more complex: The ‘critical’ transition line (yellow quality, meaning that it can be maximized by an optimal solvent contour line) between swollen and collapsed states depends sub- quality. The maximization of the partitioning is in fact a quite stantially on the membrane–solute interaction, as shown by the general feature in attractive but crowded systems as we will dis- label (2). The stark color contrast around this critical line signifies cuss in the following section. the sharp transition. In addition, one can identify in total five dis- To sum up, the CG simulation model of a polymer network in tinct phase regions (or states), classified into (1) “solute-induced the presence of solutes reveals a rich topology of structure phases collapsed”, (3) “swollen”, (4) “solute-involved collapsed”, (5) and their relation to solute partitioning, entering the rate equa- “solute-adsorbed collapsed", and (6) “collapsed” states, indexed tions eqs. (13) and (15). In particular, for very attractive solute– by the numbers in the colored center panel of Fig. 4A. Inter- membrane interactions (e & 1 k T ) the network structure and ms estingly, the “solute-induced collapsed” state (state 1) can occur B partitioning are coupled. The mesoscopic model thus provides a even in good solvent conditions, where the membrane undergoes landscape of the partitioning, thereby bridging the macroscopic a relatively sharp collapse transition induced by a strong ‘bridg- continuum model and microscopic discrete data in terms of the ing’ attraction between the solutes and the network monomers. generic interaction parameters. In addition, the results will be The effect has been reported in computer simulations before but 124–131 helpful for the interpretation of experiments for certain polymer– only on the single polymer level. The “solute-involved” col- reactant systems and could also be useful to design feedback- lapsed state (state 4) occurs at the intermediate solvent quality systems where the local reactant (or product) concentration may where the membrane collapses with most of the internal solutes couple back to polymer structure in a prescribed way. embedded, yielding a bulkier collapsed gel than the intrinsically collapsed case. The “solute-adsorbed collapsed” state (state 5) is an example for the limiting case, where both of the membrane– 3.2 Partitioning, diffusion, and permeability in a model lat- membrane and membrane–solute attractions are strong, but the tice membrane first one dominates and excludes the solutes, therefore leading to Now we present a related but different CG model of a a strong surface accumulation of those. membrane–solute system (Fig. 5A) in order to study permeability The solute partitioning K (on a log-scale) averaged over the in dense media qualitatively on a generic level. We demonstrate how the permeability can be tuned massively in magnitude by membrane slab is shown in Fig. 4B in a 2D-landscape plot, and is related to the transfer free energy from the bulk into the net- systematically varying the membrane–solute interactions and the work, DG =k TlnK , which quantifies the average transfer free density of the membrane. The study also gives important insights about how partitioning and diffusion are correlated. energy for the solute transfer from bulk to the membrane. The partitioning varies by several orders of magnitude, depending not The permeability is defined following the solution–diffusion Journal Name, [year], [vol.], 1–17 | 7 solute-involved collapsed solute-adsorbed collapsed D A B C membrane solute Fig. 5 (A) Snapshot of the mesoscopic lattice membrane–solute system. The membrane sites (red) are fixed on a face-centered-cubic (fcc) lattice with the volume fraction f , and the penetrating solutes (blue) are diffusing and interacting via the LJ potentials with e . (B) Solute partitioning K (f ) at m ms m different e . The solid lines depict the exact relation, eqn (17), and the dashed line depicts the approximated partitioning K with be = 0:6 (see ms B2 ms text for details). (C) Solute diffusivity D (f )=D with different e . The dashed lines depict the approximation D =D = exp(f ) valid for low e and in m 0 ms in 0 m ms 2=3 f , and the scaling D =D  f valid for high e and low f . (D) Permeability P(f )=D at different e . Reprinted with permission from Ref. 110, m in 0 m ms m m 0 ms copyright 2019 American Physical Society. 89–91,94–99 theory by eqn (2). There have been pioneering the- sion. In addition, a leading order approximation of K on a two- ms oretical models to elucidate the transport phenomena in mem- body level, K (f ;e ) = exp 2c B for be = 0:6 is de- B2 m ms m ms 89–91,94,132–135 branes based on simple theories for either parti- picted by the dashed line, where c µ f is the membrane con- m m ms tioning or diffusion. Recently, a simulation study revealed maxi- centration, and B is the second virial coefficient. Figure 5C mization of partitioning of penetrating solutes in polymer mem- shows the solute diffusivity D =D in the membrane as a func- in 0 branes tuned by the polymer volume fraction. Diffusion in tion of f , rescaled by the free diffusivity in the bulk. As the dense membranes is usually quite complex and highly dependent membrane becomes dense, D tends to exponentially decrease, in 102,103,113,137–141 on details of the interaction potentials. Never- showing more complex behavior with higher couplings. We com- theless, there have been no comprehensive studies on the perme- pare the simulation results with scaling theories for diffusion in ability P , being a product of partitioning and diffusion. two limiting cases. The upper dashed line indicates the limit- ing law D =D = exp(f ) based on the ‘volume-exclusion’ the- In the CG model membrane–solute system as shown in Fig. 5A in 0 m 102,134,135,143,144 the diffusive solutes are ideal (e = s = 0), and the membrane ory. It is indeed valid only for low couplings, ss ss which acts essentially repulsive. For high membrane–solute at- consists of immobile interaction sites, located on a face-centered- tractions and low membrane density, the diffusivity follows the cubic (fcc) lattice with a fixed unit cell size l, variation of which 2=3 power law D =D  f . The scaling relation is derived by the tunes the monomer packing fraction f . The simplicity of such m in 0 m an ordered and rigid model membrane with ideal solutes ren- limiting law from the Kramers’ barrier crossing over the distance 2=3 1=3 2 110 l  f , and therefore D  l =t f . ders the problem easier for interpretation and perhaps theoret- in ically tractable. The ideal solutes diffuse throughout the simu- The resulting permeability, the product of K and D , shown in lation box but interact only with the membrane sites via the LJ in Fig. 5D, exhibits intriguing features. For essentially repulsive potential with the coupling strength e . For the ideal solutes the ms solutes, P decreases monotonically as the membrane density in- partitioning can be exactly computed via the transfer free energy creases, and the overall magnitude is below unity, almost ap- bH 142 ms shown in eqn (11). In our case DG = k T lne , where proaching zero for very dense membranes. We speculate that this H (r) = U (jr r j) is the total Hamiltonian (summing over ms i ms i essentially repulsive case may be the scenario in the experiments all membrane sites i), and x  dVx=V is the volume average, with the highly charged reactant HCF in Section 2.1, which prob- yielding ably does not like to enter the collapsed gel, but this suspicion bH ms K = e : (17) needs further scrutiny. On the other hand, for high couplings (at- which verifies the simulation results (Fig. 5B). traction) the permeability is first minimal around f = 0:1, then The computed partitioning as a function of the membrane vol- maximized at large membrane densities f ' 0:8. The perme- ume fraction f is shown in Fig. 5B for various e . For rel- ability vanishes at the maximum overlapping density (f  1), m ms m atively low membrane–solute couplings (be . 0:3), the LJ in- where no percolating holes for diffusion are present anymore. ms teraction between solutes and membrane sites is essentially re- This demonstrates a clear maximization of permeability when the pulsive (signified by a positive second virial coefficient), and the system is highly attractive and dense. The nonmonotonic behav- partitioning monotonically decreases as the membrane becomes ior of permeability results from drastic nontrivial cancellations dense, owing to the dominant exclusion by the membrane. For between the partitioning and the diffusivity, which exponentially intermediate couplings around be = 0:6, which is moderately increase and decrease, respectively. The massive cancellation be- ms attractive, partitioning reaches a maximum at an optimal mem- tween two largely varying functions over several orders of mag- brane density around f = 0:6. The partitioning maximization nitude yields a permeability of the order of unity, implying a is attributed to a balance between adsorption and steric exclu- high potential for fine-tuning of the permeability behavior in ex- Journal Name, [year], [vol.], 8 | 1–17 periments by small changes in density or interactions. c , G = G N c ; (19) chain m 0 Mesoscopic models of membrane–solute systems demonstrate that the permeability, typically resulting from large cancellation Another setup of swollen hydrogels, shown in Fig. 6.ii, mimics effects of partitioning and diffusion, is very sensitively tuned by the cross-linker unit of a hydrogel network, and thus lends itself the effective interaction potentials and the membrane density. to study the influence of cross-linkers on adsorption of molecules. The results indicate that most drastic selectiviy effects are at high In our previous study , we considered a very common N,N’- membrane densities and significant (& k T ) membrane–solute at- methylenebisacrylamide (BIS) cross-linker, connecting four PNI- tractions. The effective potentials in realistic material design as- PAM chains with their ends tethered in a tetrahedral geometry. semblies can be somewhat controlled by various external param- The solute molecules in general adsorb in different proportions to eters, such as temperature, ionic strength, pH, and possibly var- the chain regions and the cross-linker neighborhood. The overall ious additives in the solution. The results from the mesoscopic adsorption in the radial interval [r ;r ] from the cross-linker is ob- 1 2 models thus provide useful physical insight and may bear impor- tained in a straightforward manner by integrating the (spherical) tant applications in design and engineering of molecular systems RDF g(r) of the solutes to achieve a selective transport by fine-tuning interactions and topologies, particularly in highly attractive membrane systems. 2 G(r ;r ) = c [g(r) 1]4pr dr: (20) 1 2 0 4 Partitioning and diffusion: all-atom simu- With this, we can scan the adsorption in different regions with lations respect to the cross-linker. It also allows us to evaluate G , as in the single-chain geometry. Finally, the total adsorption can be The advantage of the mesoscopic simulations in the previous sec- deconvoluted into two contributions, tion is that we can obtain fundamental and qualitative insights on how permeability depends on basic input parameters such as G = G +G : (21) tot chain xlink interaction energies, lattice geometry, and single solute diffusion. The adsorption on the chains G (unperturbed by the presence However, in experiments we deal with specific, chemical systems, chain of cross-linker) is given by eqn (19), whereas G represents the where the effects of interactions are highly convoluted and solva- xlink effect on the adsorption due to the presence of the cross-linker. tion, polarity, electrostatics, and specific steric constraints come The value of G can be evaluated from known G and G . explicitly into play. Hence, for a more detailed insight and quan- tot xlink chain In the infinite-dilution limit, the adsorption on the cross-linker titative numbers for the continuum approach to reaction rates in is proportional to the bulk solute concentration, G = G c , Section 2, we need to resort to higher resolution, molecular dy- xlink 0 xlink where G is the adsorption coefficient of the cross-linker. namics computer simulations. All-atom simulation studies of par- xlink titioning and diffusion through polymer networks with increasing The resulting adsorption coefficients G are shown in Fig. 7B 111 112 complexity and chemical detail are now emerging and growing (blue shaded bars), from the single-chain and cross-linker in the literature. Selected works in this field are given in Tab. 2. geometries. Quite generally, the adsorption grows with the In the following, we will review our recent efforts to understand molecular size. The effect of the cross-linker, G , is shown in xlink partitioning and diffusion of solutes in swollen and collapsed PNI- Fig. 7B by orange bars: The apolar compounds C , C , and B 4 6 111–114 PAM hydrogels by all-atom (AA) MD simulations. show a low affinity to the cross-linker. In contrast, the adsorption of nitro-aromatic solutes to the cross-linker is significant, in par- 0 112 0 ticular for NP . NB shows more than doubled and NP even an 4.1 Swollen state order of magnitude higher adsorption to the cross-linker region In order to model the swollen state of a hydrogel shell one can fo- than to a monomer of the polymer. Note that the BIS cross-linker cus on one elongated PNIPAM chain, as shown in Fig. 6A.i, where has two amide groups and is slightly more hydrophilic than the the chain is replicated through periodic boundary conditions. The PNIPAM chain, hence favoring polar molecules. cylindrical geometry allows for a simple extraction of adsorption From the known adsorptions on individual chains and cross- properties of solvated molecules in the solution. The first step linkers we can predict the partitioning in extensive hypothetical is to evaluate the cylindrical radial distribution function (RDF) of swollen polymer architectures, such as hydrogels. The partition- the solute molecules from the backbone, g (r), as shown in an 2D ing follows from K = 1+G =(c V), where V is the volume of the tot 0 example for nitrobenzene (NB) in Fig. 7A. The adsorption coef- gel and G the total adsorption of molecules on all the chains tot ficient G per monomer of the polymer is then obtained by inte- and cross-linkers [eqn (21)], which leads to gration along the spatial coordinates, K = 1+ n G + n G ; (22) ¥ m m xlink xlink G = DL [g (r) 1]2prdr; (18) m m 2D where n and n are the monomer and cross-linker number xlink where DL = 0:265 nm is the distance between neighboring densities, respectively. The former can be easily linked to the monomers in the chain. The total adsorbed number of molecules polymer volume fraction f as n = f =(pR DL ), where R = m m m m 0 G on the chain is proportional to the number of monomers N 0:5 nm is an estimated effective radius of the polymer chain. chain and, in the infinite dilution limit, to the bulk solute concentration Assuming f = 0:1 for a typical architecture of a swollen state, we Journal Name, [year], [vol.], 1–17 | 9 Table 2 Survey of selected publications on computer simulations of diffusion D, partitioning K , permeability P , or related adsorption or transport phenomena of (co)solutes in polymers. Abbreviations: All-atom (AA), Coarse-Grained (CG), Molecular Dynamics (MD), Monte-Carlo (MC), Langevin Dynamics (LD), Brownian Dynamics (BD), Dissipative Particle Dynamics (DPD). Ref. Simulation Architecture Polymer Resulting quantity Comment methods 145 AA-MD swollen cross-linked polyethyleneglycole diffusivity of water, ions, water content: 75–91%, mesh size: 2.3–5.5 nm network rhodamine 146 AA-MD collapsed polyethylene diffusivity and partitioning permeability for oxygen 5–6 orders of magnitude larger than for water of oxygen and water 147 AA-MD collapsed poly(vinyl alcohol) diffusivity of O2 water uniformly distributes 148, 149 AA-MD collapsed poly(vinyl alcohol) diffusivity of water hydrogel with 4–40 % water 150 AA-MD collapsed polydimethylsiloxane diffusivity of water and water/ethanol mixtures; water molecules faster than ethanol ethanol 151 AA-MD + tran- collapsed polystyrene and its diffusivity and partitioning sition state ap- copolymers of gas and water molecules proach 152 AA-MD single chain PNIPAM adsorption of urea studying volume phase transition 153 AA-MD single chain PNIPAM adsorption of TMAO, urea studying volume phase transition 154 AA-MD collapsed and sol- PNIPAM partitioning of ions thin core-shell membrane, direct measuring of partitioning vent phase 155 AA-MD cross-linked net- PNIPAM diffusivity of water; vol- studying volume transition; cross-linking inhibits the collapse work ume transition 156 AA-MD collapsed and sol- PNIPAM (3mer) water–polymer coexis- also 30mer of PNIPAM: no conclusions on chain configuration vent phase tence 157 AA-MD swollen & collapsed PNIPAM partitioning of large ions umbrella sampling of the potential of mean force of the ions finite aggregate 158 Gibbs-ensemble cubic network bead-spring solvent sorption and effects of solvents on swelling CG-MD swelling isotherm 119 Two-box– cubic network bead-spring solvent sorption and effects of cross-linkers on swelling particle-transfer swelling isotherm CG-MD 159, 160 CG-MC tetra-functional net- bead-spring solvent sorption and effects of polymer network density and deformation on swelling work swelling isotherm 120 CG-MC cubic network rigid rod cosolute diffusivity effects of cosolute size and polymer density on cosolute diffusivity 121 CG-LD cubic network charged bead-spring energy conversion effects of compression and solvents on energy contribution 122 CG-LD highly swollen cubic charged bead-spring adsorption and conforma- counterion-induced deformation network tional response 123 LD cubic network charged bead-spring ion transport effects of electrostatic coupling between polymer and ions on ion transport 136 CG-MC tetra-functional net- bead-spring cosolute partitioning effects of polymer density on partitioning work 139 CG-MD polymer melt semi-flexible gas partitioning, diffusivity effect of gas size and polymer semi-flexibility on gas transport and permeability 161 CG-BD cubic network bead-spring cosolute diffusivity effects of cosolute density on cosolute diffusivity 162 CG-DPD random network semi-flexible permeability and cosolute effects of porosity and deformation on permeability diffusivity 163–165 CG-BD cubic network rigid rod cosolute diffusivity effects of interactions, hydrodynamics, and network heterogeneity on cosolute diffusivity 166 CG-BD random cubic net- rigid rod cosolute diffusivity effects of network porosity, flexibility, degree of cross linking, and electro- work static interaction on cosolute diffusivity compute K for several solutes from the obtained MD parameters water–polymer spatial heterogeneity is a crucial player in the sol- in Fig. 8. The values range around unity, K 1–3, as also resulted vation of small molecules, whereby the nature of the solute (being from the CG models in Fig. 5B for this polymer fraction range. As polar, nonpolar, or ionic ) is a decisive property. Two represen- we will see in the following, the collapsed state can give rise to tative snapshots in Fig. 9A, showing a benzene (nonpolar) and a much higher partitioning. phenol (polar) molecule, demonstrate that nonpolar solutes are preferentially expelled from water clusters and tend to reside in ‘dryer’ regions of the gel, whereas polar molecules tend to parti- 4.2 Collapsed state tion closer to or inside water clusters. Thus, the “dual” character The collapsed state of the PNIPAM hydrogel can be modeled as a of the gel can favorably accommodate both polar and nonpolar bulk of aggregated polymeric chains (in our case 20 monomeric species. units long) at 340 K (above the LCST), where cross-linkers are ignored. The amount of sorbed water between the polymeric The transfer free energy for a given molecule is obtained as the chains is chosen such that it corresponds to the chemical equi- difference between the solvation free energy in PNIPAM (G ) and librium with bulk water. The amount of water in the col- in water (G ), DG = DG DG , both evaluated via the thermo- w g w lapsed state depends on temperature, and amounts to around dynamic integration procedure. Figure 9B shows DG for var- 20 wt. % (somehow less than experimental estimates of around ious solutes plotted versus the molecular surface area A of the 168–172 30 wt. % ), which roughly correspond to the polymer vol- solutes (defined as the envelope area of the fused union of the ume fraction of f = 0:8. Note that this is in the range of packing m 178 atoms ). The results follow a clear linear trend for the groups fractions for which we observed the most interesting behavior of of nonpolar and aromatic solutes as well as alcohols and water. permeability in the CG simulations in Section 3. The linearity in the very heterogeneous polymer–water medium Water molecules are very non-uniformly distributed through- is rather surprising. The results can be conveniently described in out the phase and tend to flock together into irregular clusters 179,180 terms of an effective molecular surface tension g , of various nanoscopic sizes, which were observed also in simu- 151,173–176 lations of other amorphous polymer structures . This DG = DG +g A : (23) 0 m m Journal Name, [year], [vol.], 10 | 1–17 Fig. 6 (A) Atomistic modeling of PNIPAM hydrogels: i) elongated, infinitely long chain (mimicking a part of a swollen network where the adjacent chains are far apart), ii) a cross-linker connected with four chains in a tetrahedral structure (representing a unit of a swollen network), iii) dense aggregate of PNIPAM polymers at 340 K (a model for a collapsed PNIPAM hydrogel). (B) Solute molecules in our study; polarity is characterized by the hydroxyl (OH) group. polar 21±3 nonpolar 15 Γ (single chain geometry) Γ (cross-linker geometry) Γ (cross-linker geometry) xlink C C C MeOH PrOH C B T NB NP 1 4 6 0 0.5 1 1.5 2 -5 r (nm) Fig. 7 (A) Cylindrical RDF of backbone–NB for an extended PNIPAM chain (see Fig. 6A.i). (B) Adsorption coefficients of various solutes (see 111 112 Fig. 6B) to a PNIPAM monomer G (from single-chain geometry and the OPLS force field, and from cross-linker geometry using the OPLS-QM2 force field) and to a cross-linker G (OPLS-QM2 force field) at 300 K. xlink g is strongly related to the difference in the molecule-PNIPAM collapsed swollen and molecule–water surface tension. Note that the sign de- pends on the transfer direction. The fit of eqn (23) to the NP 3 nonpolar solutes (dashed line in Fig. 9B) gives the value g = NB 1 2 18 kJ mol nm . For the alcohols and water, the transfer free energies are by about 7 kJ/mol above the trend of the nonpo- lar solutes, owing to a slightly different character of the hydroxyl 1 181 group than in alkyl chains. The molecular size is hence the dominant factor that determines its affinity to the hydrogel. In MeOH the CG description (Section 3) the molecular size is therefore re- T = 300 K T = 340 K flected in the interaction parameter e . φ = 0.1 φ = 0.8 ms -1 Using eqn (11), we show the partitioning in the collapsed state Fig. 8 Partitioning of several molecules resulting from the atomistic mod- in Fig. 8. In general, the partitioning of our neutral molecules els of a swollen (at 300 K and polymer fraction of f = 0:1) and collapsed (at 340 K and f = 0:8) state of a PNIPAM gel. The values for the swollen is larger in the collapsed state. Also, the larger the partitioning state are computed from eqn (22) and assuming polymer volume fraction in the swollen state, the progressively larger it is in the collapsed f = 0:1, whereas the values for the collapsed state are computed from state. With some heuristic arguments, we showed that the parti- eqn (3). 2 114 tioning roughly follows the relation K µ K . This collapsed swollen is in line with the universal observation from our CG model (Sec- Journal Name, [year], [vol.], 1–17 | 11 g(r ) Partitioning, ∗ 3 Γ (nm ) 0 500 ns 1000 ns He H O 5 nonpolar Ne MeOH polar Ar PrOH Me -1 He Et Ph 3 Ne -2 B NP H O 10 2 Me NB Et 2 0 -20 MeOH NP -3 nonpolar 10 NB polar (aromatic) -4 polar (alcohols, water) -40 0 0 0.1 0.2 0.3 0 0.5 1 1.5 2 0 1 2 3 4 5 A (nm ) x (nm) a (nm) Fig. 9 (A) Snapshots of benzene and phenol molecules solvated in the PNIPAM phase. Hydrophobic parts of the solutes are shown in yellow, the hydroxyl groups in green, PNIPAM polymers in blue, and water in red–white. (B) Transfer free energies from water to PNIPAM versus the molecular surface area. The dashed line is a fit of eqn (23) to the data points of nonpolar solutes. (C) Microsecond-long trajectory sequence of a NP molecule projected on a 2D plane (color coded from blue at t = 0 to red at t = 1000 ns). The green bubbles schematically depict the hopping transition with a transient pore opening. (D) Diffusion coefficients of molecules in the collapsed PNIPAM polymer versus their Stokes radii in water. The dashed line shows a fit of eqn (24) to the data points. tion 3.1), namely that a collapsed state can have much more ex- lower (i.e., higher DF ) and at the same time the diffusion coeffi- treme effects on partitioning than a swollen state. cients indeed decay faster with solute size. Moving on to the diffusion properties of solutes in the col- In conclusion, all-atom MD simulations offer insights into the lapsed PNIPAM, we first look at the projected trajectory of a NP molecular nature of the transport and solvation properties of molecule in Fig. 9C. Its connected blob-like structure suggests molecules in hydrogels. These mechanisms are not only impor- 182,183 that the diffusion advances via the hopping mechanism : tant for PNIPAM hydrogels, but most probably play important A penetrating solute resides for longer time in a local cavity roles also in other responsive hydrogels, and their understand- and suddenly performs a longer jump into a neighboring cav- ing is important for the rational design of novel materials. No- ity through a transient water channel that forms between the tably, we see drastically larger effects for K and D in the collapsed chains (schematically illustrated in the bubbles in Fig. 9C). We phase than in the swollen states, but apparently also a large anti- plot the diffusion coefficients versus the size of the solutes a (de- correlation between them, like in the coarse-grained simulations fined as the Stokes radius in pure water) in Fig. 9D. As the size in Section 3. The dense, collapsed state is thus more decisive for of a solute increases by a factor of 7, the diffusion coefficients nanoreactor design and control. Very recent studies indicate that in decreases by dramatic 5 orders of magnitude. The diffusion coef- particular for charged molecular reactants, the presence of water ficients depend on the solute size a roughly exponentially, w clusters and resulting substantial interfacial effects within a dense hydrogel may decisively affect their permeability behavior. a =l D = D e : (24) 5 Concluding remarks The fit to the data points yields the decay length l = 0:019 nm. Stimuli-responsive nanoreactors are of high potential for the de- Note that the rate-determining step in the hopping diffusion is sign of programmable and selective nano-devices for controlled the opening of a channel, which is associated with a free energy catalysis and can therefore serve as candidates to create novel barrier DF and can be via Boltzmann probability related to the synthetic enzymes on the colloidal scale. However, they consti- diffusion coefficient as D exp(DF =k T). In conjunction with a B tute complex devices with non-equilibrium processes starting at the empirically obtained diffusion relation [eqn (24)], this implies the electronic scale, defining the chemical surface reactions, cou- k T pled to those at the polymer network scale with all the intrinsic DF (a ) = a : (25) a w w complexity of polymer–reactant interactions, including the feed- back of responsive polymers, up to the device scale where re- That is, the free energy barrier depends linearly on the particle actants diffuse and react in a suspension of colloids. Here we size, and hence represents a special case of possible scenarios pre- reviewed the recent theoretical attempts of understanding some dicted by an assortment of different theories. The majority of the- parts of the processes by focussing mostly on the key roles played ories that are based on activated diffusion predict either square by the permeability of the polymer shell and the reactant parti- or cubic scaling. A possible linear dependence of the free energy tioning in order to control activity and selectivity, and how those barrier has recently been theoretically envisioned in scaling the- 137,184 enter the continuum rate predictions for the nanoreactors. ories for particle mobility in dense polymer solutions and in dense liquids by using a self-consistent cooperative hopping As an important general result, we see substantial variations theory. As also seen from eqn (25) the height of the free en- and correlations among K and D in the dense, collapsed polymer ergy barrier is related to the decay length l in eqn (24). We also phases, in both coarse-grained and atomistically-resolved simu- showed that in a less hydrated gel, the the diffusion of solutes is lations, which are thus more decisive and tuneable for nanore- Journal Name, [year], [vol.], 12 | 1–17 −1 ΔG (kJ mol ) y (nm) 2 −1 D (nm ns ) actor rate control than for the swollen states. Results for the Conflicts of interest temperature-induced rate switch observed in reference experi- There are no conflicts to declare ments, like the HCF reduction briefly discussed in Section 2.1, can be thus traced back to, for example, the large exclusion (low Acknowledgements partition ratio) and significant slowing down (low diffusion) of The authors thank Richard Chudoba, Karol Palczynski, Sebas- reactants in the collapsed state of neutral PNIPAM. However, a tian Milster, Arturo Moncho-Jordá, Stefano Angioletti-Uberti, quantification of partitioning and diffusion of molecular ions by Daniel Besold, Yan Lu, and Matthias Ballauff for inspiring dis- simulation approaches remains a challenge because of the water 177 cussions. This project has received funding from the European heterogeneities in the collapsed states. Research Council (ERC) under the European Union’s Horizon A large number of challenges and questions remain, which 2020 research and innovation programme (Grant Agreement No. we try to tackle currently or leave open for future studies. For 646659-NANOREACTOR). M.K. acknowledges the financial sup- example, continuum approaches to diffusion- and permeability- port from the Slovenian Research Agency (research core funding influenced rates in confinement are often based on mean-field No. P1-0055). W.K.K. acknowledges funding from the Deutsche theories (like presented here), but more elaborate and accurate Forschungsgemeinschaft (DFG) via grant NE 810/11. The sim- treatments, like Green’s-function approaches, are yet to be de- ulations were performed with resources provided by the North- vised. German Supercomputing Alliance (HLRN). The polymer permeability and the knowledge of how reactants Author contributions partition in the polymer are the keys to program the desired func- tion and response into a nanoreactor. Clearly, the number of ex- All authors contributed equally to this manuscript. perimental and chemical ways to synthesize a responsive hydro- gel shell (e.g., with various combinations of copolymerization) is References basically infinite. Modeling the features of diverse polymer sys- 1 S. H. Petrosko, R. Johnson, H. White and C. A. Mirkin, J. tems on various scales is therefore out of reach. Our CG and all- Am. Chem. Soc., 2016, 138, 7443–7445. atom studies so far delivered some basic but important insights 2 M. A. C. Stuart, W. T. S. Huck, J. Genzer, M. Müller, C. Ober, into the physics of these systems. However, we are continuing M. Stamm, G. B. Sukhorukov, I. Szleifer, V. V. Tsukruk, M. Ur- the endeavors towards even more refined notions of the general ban, F. Winnik, S. Zauscher, I. Luzinov and S. Minko, Nature response features of hydrogels both experimentally and theoreti- Materials, 2010, 9, 101–113. 6,177 cally. Some of such features are ions, charged reactants, and 3 S. Campisi, M. Schiavoni, C. Chan-Thaw and A. Villa, even charged (pH-responsive) hydogels, which we ignored in Catalysts, 2016, 6, 185. this review, but are of high practical relevance. Combining all the 4 Y. Lu and M. Ballauff, Prog. Polym. Sci., 2011, 36, 767–792. simulations with continuum-based approaches will help devising 5 S. Carregal-Romero, N. J. Buurma, J. Pérez-Juste, L. M. Liz- models, or at least semi-empirical rules how the hydrogel prop- Marzán and P. Hervés, Chem. Mater., 2010, 22, 3051–3059. erties, in particular the permeability of certain molecular species, 6 P. Hervés, M. Pérez-Lorenzo, L. M. Liz-Marzán, J. Dzubiella, are connected and can be tuned by stimuli. Y. Lu and M. Ballauff, Chem. Soc. Rev., 2012, 41, 5577– In order to formulate improved rate equations that carry more physical information, also the chemical processes on the nanopar- 7 S. Wu, J. Dzubiella, J. Kaiser, M. Drechsler, X. Guo, M. Bal- ticle surface in the solvent/polymer environment have to be bet- lauff and Y. Lu, Angew. Chem. Int. Ed., 2012, 51, 2229– ter understood, which we did not touch in this review. For in- stance, rate-limiting chemical intermediates could be present. 8 H. Jia, R. Roa, S. Angioletti-Uberti, K. Henzler, A. Ott, X. Lin, It would be also important to know whether and how strong the J. Möser, Z. Kochovski, A. Schnegg, J. Dzubiella, M. Ballauff (often charged) reactants and products adsorb and diffuse on the and Y. Lu, J. Mat. Chem. A, 2016, 4, 9677–9684. nanoparticle surface in the crowded polymer environment. In 9 G. Prieto, H. Tüysüz, N. Duyckaerts, J. Knossalla, G.-H. Wang some cases this may lead to steric hindrance and reaction inhibi- and F. Schüth, Chem. Rev., 2016, 116, 14056–14119. tion on the reactive surface by both reactants and products and 10 J. Gaitzsch, X. Huang and B. Voit, Chem. Rev., 2015, 116, highly nonlinear rate behavior. Here, particle-based reaction– 1053–1093. diffusion simulations may also help illuminating dynamical tran- 11 D. M. Vriezema, M. Comellas Aragonès, J. A. A. W. Elemans, sitions and collective effects during the reaction. J. J. L. M. Cornelissen, A. E. Rowan and R. J. M. Nolte, Chem. Only the fundamental understanding on all scales will enable Rev., 2005, 105, 1445–1490. us to reach the high recognition, selectivity, and feedback be- 12 K. Renggli, P. Baumann, K. Langowska, O. Onaca, N. Bruns havior in these colloidal devices as found for the nano-sized en- and W. Meier, Adv. Funct. Mater., 2011, 21, 1241–1259. 115,116 zymes. On the other hand, the large scale and diverse 13 P. Tanner, P. Baumann, R. Enea, O. Onaca, C. Palivan and building blocks that constitute the nanoreactors in various archi- W. Meier, Acc. Chem. Res., 2011, 44, 1039–1049. tectures establish the opportunity to develop many new design 14 Y. Guan and Y. Zhang, Soft Matter, 2011, 7, 6375–6384. directions within the goal of programmable, ‘intelligent’ nanopar- 15 A. Liu, C. H. H. Traulsen and J. J. L. M. Cornelissen, ACS ticle catalysis in the liquid phase. Catal., 2016, 6, 3084–3091. Journal Name, [year], [vol.], 1–17 | 13 16 S. Montolio, C. Vicent, V. Aseyev, I. Alfonso, M. I. Burguete, K. J. Stevenson and G. Henkelman, The Journal of Physical H. Tenhu, E. García-Verdugo and S. V. Luis, ACS Catal., Chemistry C, 2013, 117, 7598–7604. 2016, 6, 7230–7237. 42 J. A. Johnson, J. J. Makis, K. A. Marvin, S. E. Rodenbusch 17 A. Zinchenko, Y. Che, S. Taniguchi, L. I. Lopatina, and K. J. Stevenson, The Journal of Physical Chemistry C, V. Sergeyev and S. Murata, J. Nanopart. Res., 2016, 18, 1–9. 2013, 117, 22644–22651. 18 Y. Lu, Y. Mei, M. Drechsler and M. Ballauff, Angew. Chem. 43 E. Gross and G. A. Somorjai, Topics in Catalysis, 2014, 57, Int. Ed., 2006, 45, 813–816. 812–821. 19 J.-T. Zhang, G. Wei, T. F. Keller, H. Gallagher, C. Stötzel, 44 A. Calvo, M. C. Fuertes, B. Yameen, F. J. Williams, O. Azza- F. A. Müller, M. Gottschaldt, U. S. Schubert and K. D. Jandt, roni and G. J. A. A. Soler-Illia, Langmuir, 2010, 26, 5559– Macromol. Mater. Eng., 2010, 295, 1049–1057. 5567. 20 R. Contreras-Cáceres, A. Sánchez-Iglesias, M. Karg, 45 S. L. Brock, N. Duan, Z. R. Tian, O. Giraldo, H. Zhou and I. Pastoriza-Santos, J. Pérez-Juste, J. Pacifico, T. Hellweg, S. L. Suib, Chemistry of Materials, 1998, 10, 2619–2628. A. Fernández-Barbero and L. M. Liz-Marzán, Adv. Mater., 46 M. Ballauff and O. Borisov, Current Opinion in Colloid & 2008, 20, 1666–1670. Interface Science, 2006, 11, 316 – 323. 21 S. Li, D. Lin, J. Zhou and L. Zha, J. Phys. Chem. C, 2016, 47 M. Ballauff, Progress in Polymer Science, 2007, 32, 1135 – 120, 4902–4908. 1151. 22 M. Haruta, Chem. Rec., 2003, 3, 75–87. 48 J. Cao, S. Mei, H. Jia, A. Ott, M. Ballauff and Y. Lu, 23 G. J. Hutchings and M. Haruta, Appl. Catal., A, 2005, 291, 2 Langmuir, 2015, 31, 9483–9491. – 5. 49 Y. Lu, M. Hoffmann, R. S. Yelamanchili, A. Terrenoire, 24 Y. Zhang, X. Cui, F. Shi and Y. Deng, Chem. Rev., 2012, 112, M. Schrinner, M. Drechsler, M. W. Möller, J. Breu and M. Bal- 2467–2505. lauff, Macromolecular Chemistry and Physics, 2009, 210, 25 D. Astruc, in Nanoparticles and Catalysis, ed. D. Astruc, 377–386. Wiley-VCH Verlag GmbH, Weinheim, Germany, 2008, ch. 1, 50 M. Ballauff and Y. Lu, Polymer, 2007, 48, 1815 – 1823. pp. 1–48. 51 Y. Lu, S. Proch, M. Schrinner, M. Drechsler, R. Kempe and 26 P. Zhao, N. Li and D. Astruc, Coord. Chem. Rev., 2013, 257, M. Ballauff, J. Mater. Chem., 2009, 19, 3955–3961. 638 – 665. 52 S. Wu, J. Kaiser, X. Guo, L. Li, Y. Lu and M. Ballauff, 27 N. Li, P. Zhao and D. Astruc, Angew. Chem.-Int. Ed., 2014, Industrial & Engineering Chemistry Research, 2012, 51, 53, 1756–1789. 5608–5614. 28 E. Boisselier and D. Astruc, Chem. Soc. Rev., 2009, 38, 1759– 53 A. Lu and R. K. O’Reilly, Current Opinion in Biotechnology, 1782. 2013, 24, 639 – 645. 29 U. Taylor, C. Rehbock, C. Streich, D. Rath and S. Barcikowski, 54 M. Resmini, K. Flavin and D. Carboni, in Molecular Nanomedicine, 2014, 9, 1971–1989. Imprinting, Springer, 2010, pp. 307–342. 30 G. Sharma and M. Ballauff, Macromolecular Rapid 55 H. Jia, D. Schmitz, A. Ott, A. Pich and Y. Lu, J. Mater. Chem. Communications, 2004, 25, 547–552. A, 2015, 3, 6187–6195. 31 Y. Mei, G. Sharma, Y. Lu, M. Ballauff, M. Drechsler, T. Irrgang 56 N. Welsch, A. Wittemann and M. Ballauff, J. Phys. Chem. B, and R. Kempe, Langmuir, 2005, 21, 12229–12234. 2009, 113, 16039–16045. 32 N. C. Antonels and R. Meijboom, Langmuir, 2013, 29, 57 Y. Lu, Y. Mei, M. Drechsler and M. Ballauff, Angewandte 13433–13442. Chemie International Edition, 2006, 45, 813–816. 33 R. M. Crooks, M. Zhao, L. Sun, V. Chechik and L. K. Yeung, 58 Y. Lu, Y. Mei, M. Ballauff and M. Drechsler, The Journal of Accounts of Chemical Research, 2001, 34, 181–190. Physical Chemistry B, 2006, 110, 3930–3937. 34 R. M. Anderson, D. F. Yancey, L. Zhang, S. T. Chill, G. Henkel- 59 L.-Q. Yang, M.-M. Hao, H.-Y. Wang and Y. Zhang, Colloid and man and R. M. Crooks, Accounts of Chemical Research, Polymer Science, 2015, 293, 2405–2417. 2015, 48, 1351–1357. 60 S. Shi, Q. Wang, T. Wang, S. Ren, Y. Gao and N. Wang, The 35 C. Deraedt, N. Pinaud and D. Astruc, Journal of the Journal of Physical Chemistry B, 2014, 118, 7177–7186. American Chemical Society, 2014, 136, 12092–12098. 61 A. Chang, Q. Wu, W. Xu, J. Xie and W. Wu, Chem. Commun., 36 K. Esumi, K. Miyamoto and T. Yoshimura, Journal of Colloid 2015, 51, 10502–10505. and Interface Science, 2002, 254, 402 – 405. 62 J. Liu, J. Wang, Y. Wang, C. Liu, M. Jin, Y. Xu, L. Li, X. Guo, 37 N. Bingwa and R. Meijboom, The Journal of Physical A. Hu, T. Liu, S. F. Lincoln and R. K. Prud’homme, Colloids Chemistry C, 2014, 118, 19849–19858. and Interface Science Communications, 2015, 4, 1 – 4. 38 N. Bingwa and R. Meijboom, Journal of Molecular Catalysis 63 Y. Tang, T. Wu, B. Hu, Q. Yang, L. Liu, B. Yu, Y. Ding and A: Chemical, 2015, 396, 1 – 7. S. Ye, Materials Chemistry and Physics, 2015, 149, 460–466. 39 J.-H. Noh and R. Meijboom, Applied Surface Science, 2014, 64 J. Plazas-Tuttle, L. S. Rowles, H. Chen, J. H. Bisesi, T. Sabo- 320, 400 – 413. Attwood and N. B. Saleh, Nanomaterials, 2015, 5, 1102– 40 J.-H. Noh and R. Meijboom, Applied Catalysis A: General, 1123. 2015, 497, 107 – 120. 65 Q. Wu, H. Cheng, A. Chang, W. Xu, F. Lu and W. Wu, Chem. 41 Z. D. Pozun, S. E. Rodenbusch, E. Keller, K. Tran, W. Tang, Commun., 2015, 51, 16068–16071. Journal Name, [year], [vol.], 14 | 1–17 66 F. A. Plamper and W. Richtering, Acc. Chem. Res., 2017, 50, 95 P. Pandey and R. Chauhan, Prog. Polym. Sci., 2001, 26, 853– 131–140. 893. 67 S. Gu, S. Wunder, Y. Lu, M. Ballauff, R. Fenger, K. Rade- 96 S. C. George and S. Thomas, Prog. Polym. Sci., 2001, 26, mann, B. Jaquet and A. Zaccone, J. Phys. Chem. C, 2014, 985–1017. 118, 18618–18625. 97 M. Ulbricht, Polymer, 2006, 47, 2217–2262. 68 R. Pelton, Adv. Colloid Interface Sci., 2000, 85, 1–33. 98 R. W. Baker and B. T. Low, Macromolecules, 2014, 47, 6999– 69 A. Khokhlov, Polymer, 1980, 21, 376–380. 7013. 70 B. Erman and P. Flory, Macromolecules, 1986, 19, 2342– 99 H. B. Park, J. Kamcev, L. M. Robeson, M. Elimelech and B. D. 2353. Freeman, Science, 2017, 356, 1137. 71 A. Khokhlov, S. Starodubtzev and V. Vasilevskaya, in 100 J. K. Guesta and J. H. Prévost, Computer Methods in Applied Responsive gels: Volume transitions I, Springer, 1993, pp. Mechanics and Engineering, 2007, 196, 1006–1017. 123–171. 101 E. Atci, I. Erucar and S. Keskin, J. Phys. Chem. C, 2011, 115, 72 T. M. Barenbrug, J. Smit and D. Bedeaux, Polymer Gels and 6833–6840. Networks, 1995, 3, 331–373. 102 K. Falk, B. Coasne, R. Pellenq, F.-J. Ulm and L. Bocquet, Nat. 73 M. Heskins and J. E. Guillet, J. Macromol. Sci., Chem., 1968, Commun., 2015, 6, 6949. 2, 1441–1455. 103 A. Obliger, R. Pellenq, F.-J. Ulm and B. Coasne, J. Phys. 74 K. Dušek and D. Patterson, J. Polym. Sci. A-2 Polym. Phys, Chem. Lett., 2016, 7, 3712–3717. 1968, 6, 1209–1216. 104 M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, 75 A. Habicht, W. Schmolke, G. Goerigk, F. Lange, K. Saal- B. J. Mariñas and A. M. Mayes, Nature, 2008, 452, 301–310. wächter, M. Ballauff and S. Seiffert, J. Polym. Sci. Part B 105 G. M. Geise, H. B. Park, A. C. Sagle, B. D. Freeman and J. E. Polym. Phys., 2015, 53, 1112–1122. McGrath, J. Memb. Sci., 2011, 369, 130–138. 76 S. Zhou and C. Wu, Macromolecules, 1996, 29, 4998–5001. 106 B. Tansel, J. Sager, T. Rector, J. Garland, R. F. Strayer, 77 C. Wu and S. Zhou, Macromolecules, 1997, 30, 574–576. L. Levine, M. Roberts, M. Hummerick and J. Bauer, Sep. 78 S. Angioletti-Uberti, Y. Lu, M. Ballauff and J. Dzubiella, J. Purif. Technol., 2006, 51, 40–47. Phys. Chem. C, 2015, 119, 15723–15730. 107 C. S. Brazel and N. A. Peppas, Polymer, 1999, 40, 3383– 79 N. Pradhan, A. Pal and T. Pal, Colloids and Surfaces A: 3398. Physicochemical and Engineering Aspects, 2002, 196, 247 108 D. F. Stamatialis, B. J.Papenburg, M. Gironés, S. Saiful, – 257. S. N. M. Bettahalli, S. Schmitmeier and M. Wessling, Journal 80 T. Aditya, A. Pal and T. Pal, Chem. Commun., 2015, 51, of Membrane Science, 2008, 308, 1–34. 9410–9431. 109 W. K. Kim, A. Moncho-Jordá, R. Roa, M. Kanduˇ c and J. Dzu- 81 P. Zhao, X. Feng, D. Huang, G. Yang and D. Astruc, biella, Macromolecules, 2017, 50, 6227–6237. Coordination Chemistry Reviews, 2015, 287, 114 – 136. 110 W. K. Kim, M. Kanduc, R. Roa and J. Dzubiella, Physical 82 M. Galanti, D. Fanelli, S. Angioletti-Uberti, M. Ballauff, Review Letters, 2019, 122, 108001. J. Dzubiella and F. Piazza, Phys. Chem. Chem. Phys., 2016, 111 M. Kanduˇ c, R. Chudoba, K. Palczynski, W. K. Kim, R. Roa 18, 20758–20767. and J. Dzubiella, Phys. Chem. Chem. Phys., 2017, 19, 5906– 83 R. Roa, W. K. Kim, M. Kanduˇ c, J. Dzubiella and S. Angioletti- 5916. Uberti, ACS Catalysis, 2017, 7, 5604–5611. 112 S. Milster, R. Chudoba, M. Kanduˇ c and J. Dzubiella, Phys. 84 M. v. Smoluchowski, Z. Phys. Chem., 1917, 92, 129–168. Chem. Chem. Phys., 2019, 21, 6588. 85 P. Debye, Trans. Electrochem. Soc., 1942, 82, 265–272. 113 M. Kanduˇ c, W. K. Kim, R. Roa and J. Dzubiella, 86 G. Wilemski and M. Fixman, J. Chem. Phys., 1973, 58, Macromolecules, 2018, 51, 4853–4864. 4009–4019. 114 M. Kanduˇ c, W. K. Kim, R. Roa and J. Dzubiella, J. Phys. 87 D. F. Calef and J. M. Deutch, Annu. Rev. Phys. Chem., 1983, Chem. B, 2019, 123, 720–728. 34, 493–524. 115 M. Garcia-Viloca, J. Gao, M. Karplus and D. G. Truhlar, 88 P. Hänggi, P. Talkner and M. Borkovec, Rev. Mod. Phys., Science, 2004, 303, 186–195. 1990, 62, 251–341. 116 K. M. Ramsey, J. Yoshino, C. S. Brace, D. Abrassart, 89 H. Yasuda, A. Peterlin, C. Colton, K. Smith and E. Merrill, Y. Kobayashi, B. Marcheva, H.-K. Hong, J. L. Chong, E. D. Die Makromol. Chemie, 1969, 126, 177–186. Buhr, C. Lee, J. S. Takahashi, S. ichiro Imai and J. Bass, 90 L. M. Robeson, J. Memb. Sci.,, 1991, 62, 165–185. Science, 2009, 324, 651–654. 91 J. Williams and R. W. Baker, Journal of Membrane Science, 117 P. Atkins and J. de Paula, Physical Chemistry, W. H. Freeman 1995, 107, 1–21. and Company, 2010. 92 J. M. Diamond and Y. Katz, J. Membrane Biol., 1974, 17, 118 E. Aydt and R. Hentschke, J. Chem. Phys., 2000, 112, 5480– 121–154. 5487. 93 M. Palasis and S. H. Gehrke, J. Control Release., 1992, 18, 119 Z. Y. Lu and R. Hentschke, Phys. Rev. E, 2002, 66, 1–8. 1–11. 120 P. A. Netz and T. Dorfmüller, J. Chem. Phys., 1997, 107, 94 S. Gehrke, J. Fisher, M. Palasis and M. E. Lund, Ann. N. Y. 9221–9233. Acad. Sci., 1997, 831, 179–207. 121 A. Erba¸ s and M. Olvera de la Cruz, ACS Macro Lett, 2015, 4, Journal Name, [year], [vol.], 1–17 | 15 857–861. 151 E. Kucukpinar and P. Doruker, Polymer, 2003, 44, 3607– 122 A. Erba¸ s and M. Olvera de la Cruz, Macromolecules, 2016, 3620. 49, 9026–9034. 152 F. Rodríguez-Ropero and N. F. van der Vegt, J. Phys. 123 H. Li, A. Erba¸ s, J. Zwanikken and M. Olvera de la Cruz, Chem. B, 2014, 118, 7327–7334. Macromolecules, 2016, 49, 9239–9246. 153 M. A. Schroer, J. Michalowsky, B. Fischer, J. Smiatek and 124 J. Heyda, A. Muzdalo and J. Dzubiella, Macromolecules, G. Grübel, Physical Chemistry Chemical Physics, 2016, 18, 2013, 46, 1231–1238. 31459–31470. 125 D. Mukherji and K. Kremer, Macromolecules, 2013, 46, 154 I. Adroher-Benítez, A. Moncho-Jordá and G. Odriozola, 9158–9163. J. Chem. Phys., 2017, 146, 194905. 126 D. Mukherji, C. M. Marques and K. Kremer, Nat. Commun., 155 S. Deshmukh, D. A. Mooney, T. McDermott, S. Kulkarni and 2014, 5, 4882. J. D. MacElroy, Soft Matter, 2009, 5, 1514–1521. 127 F. Rodríguez-Ropero, T. Hajari and N. F. van der Vegt, J. 156 V. Bo¸ tan, V. Ustach, R. Faller and K. Leonhard, J. Phys. Phys. Chem. B, 2015, 119, 15780–15788. Chem. B, 2016, 120, 3434–3440. 128 F. Rodríguez-Ropero and N. F. van der Vegt, Phys. Chem. 157 L. Pérez-Fuentes, C. Drummond, J. Faraudo and D. Bastos- Chem. Phys., 2015, 17, 8491–8498. González, Soft Matter, 2015, 11, 5077–5086. 129 J. Rika, M. Meewes, R. Nyffenegger and T. Binkert, Phys. 158 E. Aydt and R. Hentschke, J. Chem. Phys., 2000, 112, 5480– Rev. Lett., 1990, 65, 657. 5487. 130 L.-T. Lee and B. Cabane, Macromolecules, 1997, 30, 6559– 159 F. A. Escobedo and J. J. de Pablo, J. Chem. Phys., 1997, 106, 6566. 793–810. 131 J. Heyda, H. I. Okur, J. Hladílková, K. B. Rembert, W. Hunn, 160 F. A. Escobedo and J. J. De Pablo, Phys. Rep., 1999, 318, T. Yang, J. Dzubiella, P. Jungwirth and P. S. Cremer, J. Am. 85–112. Chem. Soc., 2017, 139, 863–870. 161 D. Sandrin, D. Wagner, C. Sitta, R. Thoma, S. Felekyan, 132 H. Yasuda, C. Lamaze and L. D. Ikenberry, Die Makromol. H. Hermes, C. Janiak, N. de Sousa Amadeu, R. Kühnemuth, Chemie, 1968, 118, 19–35. H. Löwen et al., Physical Chemistry Chemical Physics, 2016, 133 H. Yasuda, L. Ikenberry and C. Lamaze, Die Makromol. 18, 12860–12876. Chemie, 1969, 125, 108–118. 162 H. Masoud and A. Alexeev, Macromolecules, 2010, 43, 134 L. Masaro and X. Zhu, Prog. Polym. Sci., 1999, 24, 731–775. 10117–10122. 135 B. Amsden, Macromolecules, 1998, 31, 8382–8395. 163 J. Hansing and R. R. Netz, Macromolecules, 2018, 51, 7608– 136 L. Pérez-Mas, A. Martín-Molina, M. Quesada-Pérez and 7620. A. Moncho-Jordá, Phys. Chem. Chem. Phys., 2018, 20, 164 J. Hansing and R. R. Netz, Biophysical journal, 2018, 114, 2814–2825. 2653–2664. 137 L.-H. Cai, S. Panyukov and M. Rubinstein, Macromolecules, 165 J. Hansing, J. R. Duke III, E. B. Fryman, J. E. DeRouchey and 2015, 48, 847–862. R. R. Netz, Nano letters, 2018, 18, 5248–5256. 138 B. Rotenberg, J.-F. Dufreche, B. Bagchi, E. Giffaut, J.-P. 166 H. Zhou and S. B. Chen, Phys. Rev. E, 2009, 79, 021801. Hansen and P. Turq, J. Chem. Phys., 2006, 124, 154701. 167 D. Horinek and R. R. Netz, J. Phys. Chem. A, 2011, 115, 139 K. Zhang and S. K. Kumar, ACS Macro Lett., 2017, 6, 864– 6125–6136. 868. 168 L.-C. Dong and A. S. Hoffman, J. Control. Release, 1990, 13, 140 R. Zhang and K. S. Schweizer, J. Chem. Phys., 2017, 146, 21–31. 194906. 169 S. Sasaki, S. Koga and H. Maeda, Macromolecules, 1999, 32, 141 P. M. Kekenes-Huskey, C. E. Scott and S. Atalay, J. Phys. 4619–4624. Chem. B, 2016, 120, 8696–8706. 170 R. Raccis, R. Roskamp, I. Hopp, B. Menges, K. Koynov, 142 A. Leo, C. Hansch and D. Elkins, Chem. Rev., 1971, 71, 525– U. Jonas, W. Knoll, H.-J. Butt and G. Fytas, Soft Matter, 616. 2011, 7, 7042–7053. 143 J. Haus and K. Kehr, Physics Reports, 1987, 150, 263–406. 171 Y. Kaneko, R. Yoshida, K. Sakai, Y. Sakurai and T. Okano, 144 S. K. Ghosh, A. G. Cherstvy and R. Metzler, Phys. Chem. J. Membr. Sci., 1995, 101, 13–22. Chem. Phys., 2014, 17, 1847–1858. 172 L. C. Dong and A. S. Hoffman, J. Control. Release, 1986, 4, 145 Y. Wu, S. Joseph and N. R. Aluru, J. Phys. Chem. B, 2009, 223–227. 113, 3512–3520. 173 Y. Tamai, H. Tanaka and K. Nakanishi, Macromolecules, 146 A. Börjesson, E. Erdtman, P. Ahlström, M. Berlin, T. Anders- 1994, 27, 4498–4508. son and K. Bolton, Polymer, 2013, 54, 2988–2998. 174 M. Fukuda, J. Chem. Phys., 1998, 109, 6476–6485. 147 G. E. Karlsson, U. W. Gedde and M. S. Hedenqvist, Polymer, 175 S. Goudeau, M. Charlot, C. Vergelati and F. Müller-Plathe, 2004, 45, 3893–3900. Macromolecules, 2004, 37, 8072–8081. 148 F. Müller-Plathe, Ber. Bunsenges. Phys. Chem., 1998, 102, 176 G. Marque, S. Neyertz, J. Verdu, V. Prunier and D. Brown, 1679–1682. Macromolecules, 2008, 41, 3349–3362. 149 F. Müller-Plathe, J. Membr. Sci., 1998, 141, 147–154. 177 M. Kanduˇ c, W. K. Kim, R. Roa and J. Dzubiella, ACS Nano, 150 L. Fritz and D. Hofmann, Polymer, 1997, 38, 1035–1045. 2019, 13, 11224–11234. Journal Name, [year], [vol.], 16 | 1–17 178 M. Karelson, Molecular Descriptors in QSAR/QSPR, Wiley- Interscience, 2000. 179 C. Tanford, Proc. Natl. Acad. Sci., 1979, 76, 4175–4176. 180 H. S. Ashbaugh and L. R. Pratt, Rev. Mod. Phys., 2006, 78, 181 C. Hansch, A. Leo and R. Taft, Chem. Rev., 1991, 91, 165– 182 H. Takeuchi, J. Chem. Phys., 1990, 93, 2062–2067. 183 F. Müller-Plathe, J. Chem. Phys., 1991, 94, 3192–3199. 184 L.-H. Cai, S. Panyukov and M. Rubinstein, Macromolecules, 2011, 44, 7853–7863. 185 O. Bénichou, M. Coppey, M. Moreau and G. Oshanin, J. Chem. Phys., 2005, 123, 194506. 186 R. Roa, T. Siegl, W. K. Kim and J. Dzubiella, J. Chem. Phys., 2018, 148, 065705. 187 M. J. del Razo, H. Qian and F. Noé, J. Chem. Phys., 2018, 149, 044102. Journal Name, [year], [vol.], 1–17 | 17

Journal

PhysicsarXiv (Cornell University)

Published: Mar 19, 2020

There are no references for this article.