Freezing Of Supercooled Water Research Articles

Effect of the porous electrode geometry on the freezing of supercooled water

The supercooled water characteristics in porous electrodes, affect the proton exchange membrane fuel cell (PEMFC) cold starting. This study analyzed the uneven substrate geometry, specifically, the influence of the contact angle and substrate angle on the solidification nucleation rate. For a given temperature, the nucleation rate decreases with increasing contact angle, with less effect as the contact angle exceeds 150°. Experiments observing the freezing of supercooled water in a detached catalyst layer confirmed this result. The simulation also showed that the nucleation rate decreases exponentially with increasing substrate angle: at the 15 K supercooling, with substrate angles increasing by 20°, nucleation rates have a difference of about 10−11 times. Thus, the effect of the substrate geometry cannot be ignored during the freezing of supercooled water in a porous electrode.

Quantitative observation of water phase transition in gas diffusion layers utilizing advanced in situ differential scanning calorimetry

The study investigates the freezing behavior of supercooled water in gas diffusion layers (GDLs) of polymer electrolyte fuel cells (PEFCs). Using an innovative advanced in situ differential scanning calorimetry (advanced DSC) technique, we analyze phase transitions within GDLs at a single-cell scale, excluding impacts of catalyst layers and membranes. This method overcomes the limitations of traditional DSC, allowing variable water saturation and distribution. We find that freezing-thawing and drying-saturating cycles contribute to noticeable variations in the freezing temperature (Tonset) of supercooled water, ranging from −6 °C to −15 °C, attributed to the heterogeneous surface of GDL and water redistribution. Repeated drying-saturating cycles remove nucleation seeds, lowering the freezing point of water − a "cleaning effect." The freezing probability reduces at lower saturation levels, influenced by waterproof treatment uniformity on GDLs. Residual ice can be detrimental to the cold-start capability of PEFCs, as it promotes freezing of supercooled water around −1 °C. Our findings enhance understanding of water behavior in GDLs and its impact on PEFC cold-start performance. Applying this technique to other cell components, such as the catalyst layer, could further elucidate water freezing dynamics in PEFCs and guide the innovation of more reliable cell components for PEFC applications in broader climates.

Towards Improved Cold-Start Capability of PEFCs: Patterned Gas Diffusion Layers

Polymer electrolyte fuel cells (PEFCs) are electrochemical energy converters with high efficiency and zero pollutant emissions. The unassisted cold-start of PEFCs at -30 °C is one of the major challenges for automotive applications due to the freezing of supercooled water generated by the cathodic oxygen reduction reaction (ORR). Once triggered by the nucleus in cathode gas diffusion layers (GDLs), ice propagates through the active area, leading to oxidant gas blockage and cold-start failure1.The cold-start capability of PEFCs can be improved by preventing ice propagation, especially in GDLs. In our previous work2,3, after the GDLs were segmented by polytetrafluoroethylene (PTFE) into two regions, the PEFC kept operating for another 30 minutes after the first freezing event in one region. Nevertheless, the PTFE barrier limited the active area as well as the heat and electric conductivity of PEFCs. In contrast, patterned GDLs allow water cluster separation without the abovementioned limitations. The improvement of water management and cell performance under technical operating conditions was demonstrated4. Yet, the cold-start capability with patterned GDL needs to be further investigated.This work studies the effect of patterned GDLs on the water-ice phase transition and cold-start capability of PEFCs. First, non-isothermal differential scanning calorimetry (DSC) measurements were conducted to determine the onset freezing temperature distribution of supercooled water in patterned GDLs, taking into account various properties such as coating load and pattern size. It was observed that independent freezing events occurred in patterned GDLs in all cycles, as several freezing/exothermic peaks can be seen in Figure 1b. In contrast, ice propagated immediately in untreated GDL, as only one freezing peak was observed in all cycles (Figure 1d). By computing the integral of the heat flux curve, we obtained the corresponding cumulative icing ratio over 60 cycles, as presented in Figures 1a and 1c. The effective prevention of ice propagation is only observed in patterned GDLs. Second, to assess the cold-start capability, statistical isothermal cold-start measurements were conducted using a 4.4 cm2 differential cell, with the cell temperature held at -7.5 °C, -10 °C, -15 °C, and 20 °C. In summary, nonconcurrent freezing events in patterned GDLs indicate that the use of patterned GDLs significantly improved the cold-start capability of PEFCs under isothermal subfreezing conditions.The findings of this work enhance the understanding of the cold-start dependence on the GDL and contribute to the development of unassisted freeze-start technology for PEFCs and the expansion of PEFCs applications. Figure 1

Effects of Hydrophobicity Treatment of Gas Diffusion Layers on Ice Crystallization in Polymer Electrolyte Fuel Cells.

The unassisted cold-start capability of polymer electrolyte fuel cells (PEFCs) remains challenging for large-scale automotive applications. Various studies have shown that the freezing of produced water at the cathode catalyst layer (CL) and gas diffusion layer (GDL) interface blocks the oxidant gas and leads to a cold-start failure. However, the impact of GDL properties, including substrate, size, and hydrophobicity, on the freezing behavior of supercooled water is yet to be thoroughly investigated. We use differential scanning calorimetry to perform non-isothermal calorimetric measurements on untreated and waterproofed GDLs (Toray TGP-H-060, Freudenberg H23). By conducting a large number of experiments (>100) for each type of GDL, we obtained the corresponding distribution of onset freezing temperature (Tonset) and found noticeable sample-to-sample variations in both untreated and waterproofed GDLs. Furthermore, ice crystallization is affected by GDL wettability, coating load, coating distribution, and GDL size, whereas the impact of the GDL substrate and saturation level is not apparent. The Tonset distribution allows for predicting the capability of PEFC freeze-start and the freezing probability of residual water at a given subzero temperature. Our work paves the way for GDL modifications toward the improved cold-start capability of PEFC by identifying and avoiding the features that systematically trigger the freezing of supercooled water with high probability.

Electron microscopy and calorimetry of proteins in supercooled water

Some of the best nucleating agents in nature are ice-nucleating proteins, which boost ice growth better than any other material. They can induce immersion freezing of supercooled water only a few degrees below 0 °C. An open question is whether this ability also extends to the deposition mode, i.e., to water vapor. In this work, we used three proteins, apoferritin, InaZ (ice nucleation active protein Z), and myoglobin, of which the first two are classified as ice-nucleating proteins for the immersion freezing mode. We studied the ice nucleation ability of these proteins by differential scanning calorimetry (immersion freezing) and by environmental scanning electron microscopy (deposition freezing). Our data show that InaZ crystallizes water directly from the vapor phase, while apoferritin first condenses water in the supercooled state, and subsequently crystallizes it, just as myoglobin, which is unable to nucleate ice.

Consistent and flexible thermodynamics in atmospheric models using internal energy as a thermodynamic potential. Part II: Non‐equilibrium regime

In numerical models of the atmosphere, the non‐equilibrium thermodynamic processes involving moisture are not always treated consistently—possibly leading to inconsistencies and errors in the energy budget. Therefore, a more consistent formulation of (moist) thermodynamics is important, for short‐timescale weather models and long‐timescale climate models. In Part I of this work, we derived a thermodynamically consistent framework, describing condensation, evaporation, freezing, and melting of cloud droplets, in which all thermodynamic quantities of interest were derived from an internal energy potential and with the moist thermodynamics coupled to a 2D semi‐implicit semi‐Lagrangian dynamical core. While this framework was primed to express non‐equilibrium processes, it was solved for the equilibrium regime only. Here, we follow the methods in Part I, but with the expression for the non‐equilibrium processes “turned on”, for example, allowing freezing of supercooled water or evaporation into subsaturated air. To implement the proposed approach, it is necessary to translate conventional atmospheric microphysics expressions for transfer rates of matter and entropy in and around a cloud droplet into the formalism of non‐equilibrium thermodynamics. This procedure is first derived for some simple idealised cases, beginning with liquid droplet growth by vapour diffusion, and proceeding to more complex three‐phase cases. To demonstrate the approach, we then simulate some idealised cloudy thermals, comparing the equilibrium and non‐equilibrium regimes—finding a robust decrease in the vertical velocity in the non‐equilibrium regime, as expected. Thus, this work demonstrates the feasibility of building a numerical model that includes a framework for modelling the moist non‐equilibrium thermodynamics of an atmospheric system consistently and provides a step towards this type of more consistent atmospheric modelling.

Consistent and flexible thermodynamics in atmospheric models using internal energy as a thermodynamic potential. Part I: Equilibrium regime

Approximations in the moist thermodynamics of atmospheric models can often be inconsistent; different parts of numerical models may handle the thermodynamics in different ways, or the approximations may disagree with the laws of thermodynamics. To address these problems, all relevant thermodynamic quantities may be derived from a defined thermodynamic potential; approximations are instead made to the potential itself—this guarantees self‐consistency, as well as flexibility. Previous work showed that this concept is viable for vapour and liquid water mixtures in a moist atmospheric system using the Gibbs potential. However, on extension to include the ice phase, an ambiguity is encountered at the triple point. To resolve this ambiguity, here the internal energy is used instead. Constrained maximisation methods on the entropy can then be used to solve for the system equilibrium state. Nevertheless, a further extension is necessary for realistic atmospheric systems, where many importantnon‐equilibriumprocesses take place; for example, freezing of supercooled water and evaporation into subsaturated air. To capture processes such as these fully, the equilibrium method must be reformulated to involve finite rates of approach towards equilibrium. Here the principles of non‐equilibrium thermodynamics are used, beginning with a set of phenomenological equations, to show how non‐equilibrium moist processes may be coupled to a semi‐implicit semi‐Lagrangian dynamical core. Standard bubble test cases and simulations of idealised cloudy thermals are presented to demonstrate the viability of the approach for the equilibrium regime. Further details and results for non‐equilibrium regimes are presented in Part II.

Fundamental study on the thermodynamic non-equilibrium freezing probability prediction model taking into account the super cooling phenomena in porous material

It is commonly considered that frost damage is caused by sudden freezing of supercooled water, which is a random phenomenon. Therefore, the aims of this study are to establish a prediction model for the probability of freezing until any lowest reached temperature, and to obtain the probability distribution function of the freezing point for the proposed analytical prediction model. First, theoretical prediction model for the probability of the instantaneous increment of ice content when lowest achieving temperature was known was derived based on these assumptions that building structure is an aggregation of small elements. Next, the freezing point measurement was carried out by using saturated mortar samples as the small element. As the results, it could be found that the first freezing due to supercooling occurred from -4 to -11 deg. C and the maximum probability was appeared at -7.5 deg. C. The average increment of ice content at every temperature closed to the 40 % volume of pore water until the thermodynamically-based freezing point. Moreover, the proposed method that can calculate the probability distributions of the instantaneous increment of ice content for any lowest achieving temperature from pore size distribution had good agreements with the measurement results.

Investigation into the hygrothermal behavior of fired clay materials during the freezing of supercooled water using experiments and numerical simulations

In this study, supercooling effects on the hygrothermal behavior of fired clay materials under various experimental conditions, such as water content, cooling rates, and size of specimens were investigated using experimental methods and hygrothermal simulations. We report results of the differential scanning calorimetry (DSC) and temperature distribution changes during a freeze–thaw (FT) experiment using unsaturated specimens. Also, we developed a numerical model of the freezing and thawing processes including the supercooling processes. The DSC results show the freezing of the supercooled water in a fired clay material is considerably faster than that in cement-based materials. It was also found that the dependency of the supercooling effects on the cooling rates seemed to be small. When the water saturation of a material decreases, the rate of the ice saturation increase during the freezing of the supercooled water is decreased while the freezing points of the supercooled water was not changed considerably. The comparison of the results of the FT experiment and hygrothermal simulations show that the combination of the existed hygrothermal model and a modified kinetic equation can reproduce the rapid temperature rise during the freezing of the supercooling water in the FT experiment. Finally, the size effects of specimens on the supercooling phenomenon was discussed based on the experimental and calculation results. The freezing points got higher when a specimen was larger. Due to differences in the ratio of the surface area to the volume, hygrothermal behavior in small specimens and relatively large specimens like that of the DSC and the FT experiment, respectively were markedly different. Water in a relatively large specimen with a small ratio of surface area to volume can achieve the thermodynamic equilibrium in a short period after the freezing starts.

Measurement of solidification heat from supercooled water freezing during PEFC cold start and visualization of ice distribution

The water in polymer electrolyte fuel cells (PEFCs) operated below the freezing point interferes with the gas supply, shutting down the operation. Therefore, understanding the freezing mechanism is pivotal to achieving high-performance cold-start. In this paper, a novel method of classifying two different freezing modes using 6 μm thickness thin-film thermocouples (TFTCs) was proposed. We conducted in situ measurement of the solidification heat from supercooled water using originally developed TFTCs and further visualized the ice distribution using X-ray CT. In a mode when TFTCs measure rapid temperature rises in the −10 °C ~ −5 °C, water freezes after diffusing into the GDL, while freezing occurs immediately inside the CL without reaching the GDL in the other mode when TFTCs measure no rapid temperature rise in the −30 °C ~ −10 °C. These modes should be considered for future researches on components improvement, and optimization of current density sweeping to succeed the cold start.

On-chip density-based sorting of supercooled droplets and frozen droplets in continuous flow.

The freezing of supercooled water to ice and the materials which catalyse this process are of fundamental interest to a wide range of fields. At present, our ability to control, predict or monitor ice formation processes is poor. The isolation and characterisation of frozen droplets from supercooled liquid droplets would provide a means of improving our understanding and control of these processes. Here, we have developed a microfluidic platform for the continuous flow separation of frozen from unfrozen picolitre droplets based on differences in their density, thus allowing the sorting of ice crystals and supercooled water droplets into different outlet channels with 94 ± 2% efficiency. This will, in future, facilitate downstream or off-chip processing of the frozen and unfrozen populations, which could include the analysis and characterisation of ice-active materials or the selection of droplets with a particular ice-nucleating activity.

The Role of the Electric Field in Electrofreezing

Electrofreezing studies date back to the late 19th century. Since then it is intensively investigated, yet the mechanism of this phenomenon is still under dispute. In the presented work, we use a device composed of electrodes covered by a thin protective dielectric layer in order to generate a large electric field\surface charge and separating their effects from those of the electric current\electrochemical reactions. We demonstrate that a surface charge density of up to ~75 nC/mm2 and an electric field of up to ~1×108 V/m, maximum attainable without water decomposition, do not have an effect on the freezing of super-cooled water. These results prove that at supercooling below 10 degrees, even a very large electric field does not order water molecules into an ice-like configuration.

Extended stable equilibrium invaded by an unstable state

Coexistence of states is an indispensable feature in the observation of domain walls, interfaces, shock waves or fronts in macroscopic systems. The propagation of these nonlinear waves depends on the relative stability of the connected equilibria. In particular, one expects a stable equilibrium to invade an unstable one, such as occur in combustion, in the spread of permanent contagious diseases, or in the freezing of supercooled water. Here, we show that an unstable state generically can invade a locally stable one in the context of the pattern forming systems. The origin of this phenomenon is related to the lower energy unstable state invading the locally stable but higher energy state. Based on a one-dimensional model we reveal the necessary features to observe this phenomenon. This scenario is fulfilled in the case of a first order spatial instability. A photo-isomerization experiment of a dye-dopant nematic liquid crystal, allow us to observe the front propagation from an unstable state.

Enhanced ice nucleation activity of coal fly ash aerosol particles initiated by ice-filled pores

Abstract. Ice-nucleating particles (INPs), which are precursors for ice formation in clouds, can alter the microphysical and optical properties of clouds, thereby impacting the cloud lifetimes and hydrological cycles. However, the mechanisms with which these INPs nucleate ice when exposed to different atmospheric conditions are still unclear for some particles. Recently, some INPs with pores or permanent surface defects of regular or irregular geometries have been reported to initiate ice formation at cirrus temperatures via the liquid phase in a two-step process, involving the condensation and freezing of supercooled water inside these pores. This mechanism has therefore been labelled pore condensation and freezing (PCF). The PCF mechanism allows formation and stabilization of ice germs in the particle without the formation of macroscopic ice. Coal fly ash (CFA) aerosol particles are known to nucleate ice in the immersion freezing mode and may play a significant role in cloud formation. In our current ice nucleation experiments with a particular CFA sample (CFA_UK), which we conducted in the Aerosol Interaction and Dynamics in the Atmosphere (AIDA) aerosol and cloud simulation chamber at the Karlsruhe Institute of Technology (KIT), Germany, we observed a strong increase (at a threshold relative humidity with respect to ice of 101 %–105 %) in the ice-active fraction for experiments performed at temperatures just below the homogeneous freezing of pure water. This observed strong increase in the ice-active fraction could be related to the PCF mechanism. To further investigate the potential of CFA particles undergoing the PCF mechanism, we performed a series of temperature-cycling experiments in AIDA. The temperature-cycling experiments involve exposing CFA particles to lower temperatures (down to ∼228 K), then warming them up to higher temperatures (238–273 K) before investigating their ice nucleation properties. For the first time, we report the enhancement of the ice nucleation activity of the CFA particles for temperatures up to 263 K, from which we conclude that it is most likely due to the PCF mechanism. This indicates that ice germs formed in the CFA particles' pores during cooling remain in the pores during warming and induce ice crystallization as soon as the pre-activated particles experience ice-supersaturated conditions at higher temperatures; hence, these pre-activated particles show an enhancement in their ice-nucleating ability compared with the scenario where the CFA particles are directly probed at higher temperatures without temporary cooling. The enhancement in the ice nucleation ability showed a positive correlation with the specific surface area and porosity of the particles. On the one hand, the PCF mechanism can play a significant role in mixed-phase cloud formation in a case where the CFA particles are injected from higher altitudes and then transported to lower altitudes after being exposed to lower temperatures. On the other hand, the PCF mechanism could be the prevalent nucleation mode for ice formation at cirrus temperatures rather than the previously acclaimed deposition mode.

Alternate Hydrogen Pump Method Enables Start-Up from −30°C for Graphite-Bipolar-Plate Proton Exchange Membrane Fuel Cells

The sub-zero start-up capability remains as a technical barrier for large-scale adoption of proton exchange membrane fuel cell (PEMFC) vehicles; the situation is even worse when carbon-based (i.e. graphite) bipolar plates, with a larger heat capacity, are used. Herein, we develop an alternate hydrogen pump (AHP) method, which enables successful start-up of a graphite-bipolar-plate (GBP) PEMFC from −30°C. By applying an alternate voltage to both sides of the PEMFC under H2 atmosphere, the AHP method warms up the PEMFC using heat generated mainly from ohmic polarization and eliminates the hazard of super-cooled water freezing. Additionally, The AHP method causes minor changes of the PEMFC system. The energy efficiencies of the AHP method are analyzed and further improvements are proposed.

Nucleation- and Emergence-Limited Growth of Ice from Pores.

Nucleation of ice from vapor on atmospheric aerosols has been attributed to the condensation and freezing of supercooled water in small pores. Here we use wedge pores on mica to directly observe the growth of ice in confinement prior to the growth of bulk crystals. We report a transition in behavior with a decreasing temperature: At low temperatures, the limiting step is not nucleation but a free energy barrier associated with the growth of ice through a narrow pore mouth to become a bulk phase.

Nonthermal ice nucleation observed at distorted contact lines of supercooled water drops.

Ice nucleation is the crucial step for ice formation in atmospheric clouds and therefore underlies climatologically relevant precipitation and radiative properties. Progress has been made in understanding the roles of temperature, supersaturation, and material properties, but an explanation for the efficient ice nucleation occurring when a particle contacts a supercooled water drop has been elusive for over half a century. Here, we explore ice nucleation initiated at constant temperature and observe that mechanical agitation induces freezing of supercooled water drops at distorted contact lines. Results show that symmetric motion of supercooled water on a vertically oscillating substrate does not freeze, no matter how we agitate it. However, when the moving contact line is distorted with the help of trace amounts of oil or inhomogeneous pinning on the substrate, freezing can occur at temperatures much higher than in a static droplet, equivalent to ∼10^{10} increase in nucleation rate. Several possible mechanisms are proposed to explain the observations. One plausible explanation among them, decreased pressure due to interface curvature, is explored theoretically and compared with the observational results quasiquantitatively. Indeed,theobserved freezing-temperature increase scales with contact line speed in a manner consistent with the pressure hypothesis. Whatever the mechanism, the experiments demonstrate a strong preference for ice nucleation at three-phase contact lines compared to the two-phase interface, and they also show that movement and distortion of the contact line are necessary contributions to stimulating the nucleation process.

Delay in the Freezing of Supercooled Water Drops on Superhydrophobic Surfaces of Silicone Rubber at Negative Temperatures

A method is proposed for fabricating textured superhydrophobic surfaces of silicone rubber with mechanical resistance toward liquid or freezing aqueous solutions. The anti-icing characteristics of silicone rubber samples that differ in the wetting characteristics and mechanical stability of their micro- and nanotextures are derived by analyzing the delays in the freezing of supercooled sessile water drops deposited on the sample surface. The longest delay in freezings are observed for sessile water drops on superhydrophobic surfaces prepared by laser texturing with subsequent application of a layer of a hydrophobic agent to consolidate the textural elements. Delay in freezings can be as long as tens of hours on such surfaces at T = −18°C. The prepared superhydrophobic surfaces exhibit greater anti-icing ability with respect to aqueous salt solutions than to deionized water.

A sudden-melting event during water freezing inside a copper well.

We studied the freezing of super-cooled water inside a millimeter-sized copper well by confocal microscopy. During freezing, we surprisingly observed a novel melting scenario, which we call a ‘sudden-melting event’: the ice directly above the bottom substrate suddenly melts in the late stage of the freezing process, while the system is continuously being cooled. After this event, an empty gap around 10 μm to 20 μm between the substrate and the bulk ice is formed. Because this gap occupies the majority of the area of the bottom substrate, the adhesion between the bulk ice and the substrate is greatly reduced: the adhesion force decreases by more than 50% compared with the flat-substrate situation. We further discovered that air dissolved in water plays a crucial role in this melting event: the air excluded by water freezing produces inter-connecting channels in the bulk ice, which transport the warm water produced by latent heat to the substrate which causes the sudden melting event. Because this event makes the contact between ice and substrate very poor, and greatly reduces ice adhesion, our observation may lead to a promising anti-icing method on solid substrates. Compared to the prevalent super-hydrophobic surface technique, our approach only requires millimeter-sized wells instead of complex microscopic textures. Therefore, it is much easier and cheaper to produce, as well as much more robust for large-scale practical applications.

Determining Interface Temperature During Rapid Freezing of Supercooled Water

In the classical Stefan solution of solidification in liquids, the solid–liquid interface temperature is considered to be constantly equal to the freezing point. However, if solidification occurred in supercooled liquids, then the interface temperature would be lower than the freezing point due to the nonequilibrium condition. In this study, the interface temperature during rapid freezing of supercooled water is investigated by molecular dynamics simulations. The relationship of interface temperature and velocity is analyzed and fitted into a quadratic function. In most aviation ice accretion situations, the degree of nonequilibrium condition at the ice–water interface falls in a range of interfacial supercooling from 0 to 4 K. Given this condition, the relationship of interface velocity and temperature could be approximated by a constant interfacial kinetic coefficient of . Both regularities are applicable in numerical simulation of dendrite ice growth.