Network-forming System Research Articles

Liquid–Crystal Structure Inheritance in Machine Learning Potentials for Network-Forming Systems

It has been studied whether machine learning interatomic potentials parameterized with only disordered configurations corresponding to liquid can describe the properties of crystalline phases and predict their structure. The study has been performed for a network-forming system SiO2, which has numerous polymorphic phases significantly different in structure and density. Using only high-temperature disordered configurations, a machine learning interatomic potential based on artificial neural networks (DeePMD model) has been parameterized. The potential reproduces well ab initio dependences of the energy on the volume and the vibrational density of states for all considered tetra- and octahedral crystalline phases of SiO2. Furthermore, the combination of the evolutionary algorithm and the developed DeePMD potential has made it possible to reproduce the really observed crystalline structures of SiO2. Such a good liquid–crystal portability of the machine learning interatomic potential opens prospects for the simulation of the structure and properties of new systems for which experimental information on crystalline phases is absent.

The progression of thermodynamic anomalies in MX2 networks with local tetrahedral geometries

Key thermodynamic anomalies in density and compressibility, as well as the related stability limits, are determined using an ionic model for BeF2 which includes many-body polarization terms. BeF2 is chosen as an example of an archetypal network-forming system whose structure can be rationalised in terms of connected local tetrahedral coordination polyhedra. The anion dipole polarizability (which effectively controls the bond angles linking neighbouring tetrahedra) is used as a single free parameter in order to help rationalise the changes in the anomaly locations in phase space, whilst all other potential parameters remain fixed. The anomalies and stability limits systematically shift to lower temperature and higher pressure as the anion polarizability is increased. At high dipole polarizabilities the temperature of maximum density anomaly locus becomes suppressed into the supercooled regime of the phase space. The movements of the anomaly loci are analysed in terms of the network structure and the correlation with the inter-tetrahedral bond angles is considered. The high sensitivity of the anomalies to the details of the potential models applied is discussed with reference to previous works on related systems. The relationship to analogous studies on Stillinger–Weber liquids is discussed.

Microscopic dynamics and failure precursors of a gel under mechanical load

Material failure is ubiquitous, with implications from geology to everyday life and material science. It often involves sudden, unpredictable events, with little or no macroscopically detectable precursors. A deeper understanding of the microscopic mechanisms eventually leading to failure is clearly required, but experiments remain scarce. Here, we show that the microscopic dynamics of a colloidal gel, a model network-forming system, exhibit dramatic changes that precede its macroscopic failure by thousands of seconds. Using an original setup coupling light scattering and rheology, we simultaneously measure the macroscopic deformation and the microscopic dynamics of the gel, while applying a constant shear stress. We show that the network failure is preceded by qualitative and quantitative changes of the dynamics, from reversible particle displacements to a burst of irreversible plastic rearrangements.

The heat capacity signal from modulated temperature DSC in non-isothermal conditions as a tool to obtain morphological information during reaction-induced phase separation

Thermal properties of materials formed during reaction-induced phase separation (RIPS) of modified epoxy–amines can be obtained from the derivative of the heat capacity signal (dCp/dT) in non-isothermal conditions following a certain cure schedule. A method is proposed in which the dCp/dT signal is deconvoluted, resulting in values of Tg and ΔCp at Tg, which can be used as probes for the composition and fraction of each phase, respectively. The following modified epoxy–amine systems, showing LCST-type demixing behavior, are presented: (i) the linearly polymerizing system diglycidyl ether of bisphenol A (DGEBA)+aniline; (ii) the network-forming system DGEBA+methylenedianiline (MDA) both modified with the high-Tg thermoplastic poly(ether sulphone) (PES: Tg=223°C) and (iii) DGEBA+MDA modified with the low-Tg copolymer poly(ethylene oxide)-block-poly(propylene oxide)-block-poly(ethylene oxide) (triblock: Tg=−70°C). The onset of RIPS can be detected before the cloud point for systems (i) and (ii), indicating that the dCp/dT signal is sensitive to phases in the sub-micron scale. Evolutions of the composition and fraction as a function of conversion during RIPS in isothermal conditions contribute to the understanding of the relation between the cure schedule and the morphological development. For the network-forming systems (ii) and (iii), overlapping peaks in the dCp/dT signal reveal complex phase structures with interphases arising in certain cure conditions. The network structure in combination with limited interdiffusion rates results in fixation of the morphology in the former system. On the contrary, higher interdiffusion rates gives rise to a better segregation between the PES-rich and epoxy–amine-rich phases in system (i).

Rheological Behavior and Gel-Point Determination for a Model Lewis Acid-Initiated Chain Growth Epoxy Resin

The sol−gel transition, or gelation, in a polymer network-forming system is defined as the critical point in the polymerization at which an infinite macromolecule is formed. There are many studies in the literature concerning the exact and accurate determination of the gel point and its interpretation in terms of the critical conversions of reactant functional groups. However, these have mainly involved step-growth polymerizations with few studies of chain-growth polymerizations such as that described in this study. Chemorheological studies on a cationically polymerized epoxy network-forming system are used to determine accurate gel times and critical conversions, which are interpreted in terms of classical branching and chemical bond percolation models for nonlinear polymerizations. Rheological studies during the polymerization of a trifunctional epoxide monomer, using BF3 as the Lewis acid initiator, yielded a gel conversion of 0.17 ± 0.02 and a power law dependence for the development of shear modulus, with a critical exponent n = 0.64 ± 0.03, which is shown to be independent of temperature and BF3 concentration. The viscosity at conversions close to the gel point was shown to vary according to a power law with a critical exponent k = 1.33 ± 0.03 in the pregel region. The critical behavior of this complex network-forming system is shown to be closely predicted by percolation theory.

Molecular Dynamics in Cyanate Ester Resin Networks and Model Cyanurate Compounds

Molecular dynamics via broad-band dielectric relaxation spectroscopy (DRS) are obtained on a class 3 network-forming system: a generic cyanate ester resin. The potential for in-situ monitoring of the network-forming reaction is explored. The segmental dynamics of the growing network are investigated, and a hitherto unknown secondary relaxation process, β*, is found. This relaxation is distinctly different from the Johari−Goldstein slow β process. Model compounds and conformational strain analysis are used to elucidate the β* relaxation molecular mechanism. The α process in selected cyanate networks was analyzed with the Adam−Gibbs thermodynamics-based relations (AG model). The AG model provides a link between the relaxation time of a cooperatively rearranging domain (CRD) and configurational entropy. The model is explored via DRS and calorimetrically for various network structures. A good agreement is found between independent quantities derived from both techniques when examined in the AG framework. These results lend support to the CRD concept used in the AG model and, more generally, toward the establishment of the heterogeneous view of relaxations in glass-formers.

On the observation of propagating sound modes at high momentum transfer in viscous liquids and glasses

The dynamic structure factors, S(Q,ω), at a range of scattering vectors (Q) in the range 0.2<Q<3.0 Å−1 are calculated in a molecular dynamics (MD) simulation of the viscous, network-forming liquid, ZnCl2. At intermediate Q (0.2<Q<0.5 Å−1), Brillouin features, with shifts and widths proportional to Q and Q2, respectively, are observed. These features persist above the frequency of the “boson peak,” in apparent discord with some suggestions for its origin. At high values of Q, close to the first peak of the static structure factor, a Q-dependent feature is seen in the inelastic part of S(Q,ω). This is analogous to a recent experimental observation on the similar, network-forming system B2O3 which was assigned to a “sound mode” contribution. An instantaneous normal mode (INM) analysis was applied to understand the relationship between these observations and the nature of the underlying modes of the liquid. The INM analysis accurately reproduces the observed (in MD) behavior of S(Q,ω) but does not support the existence of well-defined, plane-wave-like sound modes in the fluid in the relevant régimes of Q and ω. Alternative explanations of the origin of the Brillouin and S(Q)-dependent features are proposed on the basis of the INM results, and results related to the origin of the boson peak in the density of states are presented.

Influence of Pendant Chains on Mechanical Properties of Model Poly(dimethylsiloxane) Networks. 1. Analysis of the Molecular Structure of the Network

A recursive approach is used to evaluate the resulting molecular structure of a network-forming system composed of a mixture of bifunctional and monofunctional prepolymer chains reacting with a polyfunctional cross-linker (Af + B2 + B1). This system can be used to find formulations for model networks with pendant chains composed of linear molecules of known and uniform molecular weight. Several important molecular parameters, such as the weight fraction of pendant chains, number- and weight-average molecular weights, and polydispersity of the pendant chains in the network, are calculated as a function of the weight fraction of monofunctional chains (B1) added to the reaction mixture. In a naive analysis of the system Af + B2 + B1, it may be wrongly assumed that networks obtained from the reaction of these components are formed by elastically active chains comprising B2 molecules and pendant chains comprising B1 molecules. A careful analysis shows that this hypothetical situation would take place only unde...

Transient network theory for shear-thickening fluids and physically crosslinked networks

A model is introduced to treat the dynamics of coexistence between a physically cross-linked network with temporary junctions and unbound chains. Focusing on unentangled networks in which the molecular weight between neighboring cross-links a smaller than the entanglement molecular weight, three types of chains are introduced in a network-forming system: the elastically effective chains in the network that undergo affine deformation, dangling chains with one and sticking to junctions that experience deformation only through viscous interactions, and unbound free chains. Kinetic exchange between these chains is described by deriving three rate equations