Over the last decade, one of the most intriguing advances in surface science has been the emergence of slippery, anti-adhesive surfaces obtained by simple, covalent attachment of liquid molecules to solid surfaces [1], [2], [3]. These slippery covalently-attached liquid surfaces (SCALS) consist of a thin (1 to 5 nm) tethered layer of flexible molecules (generally polymers or oligomers) and can repel most liquids and many solids. The identifying feature of SCALS is that the difference between the advancing and receding contact angles (the contact angle hysteresis, CAH) is extremely low (<5°) for one or more solvents. CAH is directly related to the friction experienced by a droplet [4], [5], [6], and is consequently more significant than the equilibrium contact angle in the design of antifouling or condensation-enhancing surfaces [3], [7]. SCALS possess self-cleaning, antifouling and anti-icing properties and have demonstrated promise for applications such as improving heat transfer efficiency and water capture.

The exceptional properties of SCALS coatings are typically ascribed to the fact that the tethered molecules retain some of their liquid properties, behaving similarly to liquid-infused surfaces (LIS) — structured or network materials infused with liquid lubricant [8], [9]. Undoubtedly, the recent interest in SCALS has been spurred on by the success of LIS, which have attracted significant attention (the seminal work of Wong et al. [10] has been cited over 3000 times) since their inception in 2011 due to their exceptional antifouling [11], anti-thrombosis [12] and drag reduction [13] properties. However, LIS suffer from the inescapable defect of lubricant depletion, which results in the loss of function [8]. Furthermore, there is a growing realisation that the functional lubricant coating in LIS is generally thin [14], held in place only by van der Waals forces [15]. Consequently, the promise of SCALS systems, that possess desirable ‘liquid-like’ properties and do not suffer from depletion, is of great academic and commercial interest [10], [16], [8], [17].

However, this ‘liquid-like’ property supposedly possessed by tethered, nanoscale layers, has never been confirmed or quantitatively defined [18], [19]. This review collates measurements of CAH on various SCALS, aiming to sharpen the definition of ‘liquid-like’ in the context of SCALS and highlight questions for future research. While the majority of SCALS literature concerns grafted PDMS chains [20], [21], the existence and properties of polyethylene oxide (PEO) [22], [23], perfluorinated polyether (PFPE) [24], short-chain alkane [25], [26] and perfluoroalkane [27] SCALS will also be highlighted. Current work with PDMS identifies the length of the tethered polymer chains as a crucial parameter in determining their efficacy [28], [29], with CAH decreasing with increasing chain length up until a critical point (close to the entanglement length in the polymer melt) and growing with increasing chain length thereafter. This chain-length dependence is hypothesised to be due to entanglement effects [30]. Similarly, the grafting density of the tethered chains is also hypothesised to play a role in SCALS behaviour [17]. This review quantitatively summarises these properties, and their relationship with CAH for SCALS surfaces, identifying trends in the reported data. The picture that emerges is clearer and provides a framework for future research.

The present review is concerned with surfaces that:

exhibit a CAH of <5° for one or more liquids;

derive their low CAH from covalently-attached molecules that would be liquid in the employed conditions, were they not grafted.

Several terms have been used in the literature to refer to layers that meet these criteria; the most notable is ‘slippery omniphobic covalently-attached liquid’ (SOCAL), introduced by Wang and McCarthy [31] and widely used for PDMS SCALS since [32], [33], [2]. The additional ‘omniphobic’ descriptor is omitted here, as while it applies to PDMS SCALS (with the notable exception of silicone oil), not all SCALS exhibit low CAH for all solvents (for instance, PEO SCALS are wetted by low surface-tension liquids [24]). Conversely, other common terms, such as quasi-liquid surface (QLS) [34], [35] or simply referring to these layers as brushes [18] are not specific enough; being ‘quasi-liquid’ or in the brush regime does not guarantee a low CAH. The term slippery, covalently-attached liquid surfaces (SCALS) introduced recently by Chen et al. [26] strikes a balance. The term SCALS captures both the macroscopic behaviour of these surfaces (slipperiness) and the microscopic reasons for their durability and efficacy (covalent attachment and liquid-like nature, respectively).

It is important to stress that SCALS are distinct from LIS [8]. While there has been speculation that the reported properties of SCALS were due to a failure to remove residual untethered lubricant [30], the body of evidence presented in the rest of this review indicates that SCALS are intrinsically slippery even when any residual untethered solvent or monomers has been removed.

A number of measurements are commonly used to quantify a surface’s slipperiness (here defined by droplet mobility). This review focuses on contact angle hysteresis (CAH), as it is the accepted characterisation method for droplet mobility. CAH is the difference between the advancing and receding contact angle of a given liquid on a given surface at a given contact line velocity, and it is a vital parameter in understanding wetting behaviour [3], [6], [7], [38]. In addition to characterising droplet adhesion and friction, CAH is a better predictor of several functional properties of SCALS (ice adhesion, scale deposition, solid friction, Fig. 1) than any other trivially measured parameter (e.g., thickness or roughness). While the low CAH of these surfaces is not directly responsible for these functional properties, the correlation implies that the same fundamental mechanism is responsible for both behaviours. Hence, characterising and understanding the low CAH of SCALS is of fundamental importance to comprehending their efficacy in a range of applications.

CAH hysteresis occurs because contact lines are generally not free to move. Butt et al. [3] provide four fundamental mechanisms that induce CAH: surface heterogeneity, roughness, adaptation, and deformation. Heterogeneity refers to chemical patchiness and causes hysteresis by inducing contact-line pinning at solvophilic patches; low CAH is often seen as an indicator of surface homogeneity [39], [40], [38]. Gao and McCarthy [7] emphasise that all length scales of heterogeneity are important, with Angstrom-scale defects able to induce significant CAH [41]. Roughness induces contact line pinning through geometric inhomogeneities and variations in the local slope of the surface. Adaptation refers to the droplet-induced changes in the wetted region of the substrate, which increase the solvophilicity of patches that are already wetted. Energy is hence required to reconfigure the surface as the droplet moves across it, resulting in friction at the contact line and corresponding CAH. Adaptation is relevant for SCALS where the wetting liquid is a solvent for the tethered liquid. Deformation refers to the mechanical deformation of the solid substrate due to the droplet’s surface tension and/or the contact line’s capillary pressure. While deformation can be ignored for hard substrates, it becomes important for CAH of softer substrates, particularly viscoelastic substrates which can undergo elastic (generally due to surface tension) or plastic (generally due to capillary pressure) deformation [42]. It has been hypothesised that the deformation of SCALS around the contact line of a sessile droplet explains the dependence of CAH on the length of the tethered chains [28], due to viscous dissipation within the layer [29]. In each of the above cases, the contact line becomes trapped in a pinned state — a local Gibbs energy minima — away from the equilibrium contact angle [38], [3]. Energy is required for the contact line to depin, hence resulting in droplet friction [43].

The present review focuses on CAH as the primary measure of slipperiness because of its ubiquity and relative ease of comparison. There are many issues related to the accurate measurement of CAH, as outlined in the recent work of Huhtamaki et al. [44], who also document best practice for CAH measurements. One relevent problem is the velocity of the contact line at which the advancing and receding contact angles are measured; while CAH is a function of contact line velocity on most surfaces [45], [46], this dependence is more pronounced for SCALS. For instance, the CAH on PDMS SCALS only reaches equilibrium values at extremely low capillary numbers — around 10 × 10−8 for water — which corresponds to contact line velocity of around 1 μm s−1[33], [29]. One open question is the effect of small vibrations in the measurement apparatus on the contact line velocity and, consequently, measured CAH [38]. There is also an effect of ambient vapour pressure (either of the wetting liquid or a third solvent) on CAH. While this is negligible when the vapour is a poor solvent for the tethered layer [32], high vapour pressures of a good solvent for the SCALS may result in solvent adsorption, which effectively acts as an additional lubricant [47]. It has been reported that the CAH of PDMS SCALS is not temperature dependent (0 to 100 °C) [16], [30].

In this review it is assumed that CAH measurements are made in accordance with best practice [44] in the limit of zero contact line velocity. Roll-off/sliding angle (the tilt angle at which a droplet starts to move on a surface) is not considered, as the velocity of the contact line cannot be controlled.

The explanation offered for the properties of SCALS (droplet shedding, anti-icing, anti-scaling) is generally that these surfaces are ‘liquid-like’ [19]. From a phenomenological perspective, this description is helpful - as these surfaces behave similarly to conventional liquid interfaces in many regards. Beyond the similarities in properties to liquid-infused surfaces, there are several more quantitative aspects in which SCALS are comparable to liquids. Firstly, Daniel et al. [46] show that the relationship between CAH and contact line velocity is similar for SCALS and liquid-infused surfaces; this behaviour is distinct from conventional superhydrophobic surfaces, for which CAH is independent of velocity. Secondly, Zhao et al. [18] find that the same scaling relation can be used to describe ice adhesion on both thin liquid films and SCALS.

However, explaining the SCALS slippery properties simply as a ‘liquid-like’ surface is unsatisfying from a mechanistic standpoint, as there is no consensus as to what ‘liquid-like’ means at the microscale from either a structural or dynamic perspective. The term ‘liquid-like’ is used in the relevant literature to describe surfaces that are ‘defect-free’, ‘mobile’, or both. These terms — outlined in more detail below — are more helpful, but nonetheless leave significant gaps in our mechanistic understanding of SCALS.

The term defect-free follows logically from the origins of CAH outlined in the previous section, reasoning that a perfectly homogeneous smooth surface that does not adapt or deform should exhibit zero CAH. Surfaces must be defect-free at the nanoscale (and below, depending on the size of the solvent molecules) [7]; the access of solvent molecules to nanoscale defects has been shown to increase CAH significantly [41]. Liquid surfaces have the capacity to self-level at the nanoscale, presenting a flat and homogeneous surface to impinging droplets. Presumably, this makes SCALS less sensitive to nanoscale point defects than conventional self-assembled monolayers (SAMs) as the surface can reconfigure to shield such defects (Fig. 2b vs. e). While being defect-free is a necessary condition to exhibit low CAH, it is not sufficient, as even homogeneous, smooth, non-deforming and non-adapting surfaces can exhibit a significant CAH - for example, the case of a smooth fluoropolymer [48] and silica [49].

The mobile description for SCALS is the most popular contemporary explanation of the SCALS phenomena, having been first offered as an explanation by Chen et al. [50] (flexible) and popularised by the letter of Wooh and Vollmer [2]. Mobile refers to the continuous nanoscale rearrangements of the tethered molecules at the interface. These ‘liquid-like’ fluctuations of mobile tethered moieties could reduce the energy barriers between pinned contact lines that may only differ spatially by a nanometer or less (Fig. 2c vs. f). As angstrom-scale pinning events must be considered [7], the hypothesized reduction in the depth of the energy well of the pinned state is a valid explanation for the exceptionally low CAH of SCALS. Similar to how mechanical vibrations can overcome macroscopic pinning [51], [52], thermal fluctuation of tethered liquids could reduce nanoscale pinning [53], [3]. The obvious question is then: what length and time scales are relevant when considering whether a tethered moiety has ‘liquid-like’ mobility?

Both the qualities of being defect-free and made of mobile molecules explain SCALS behaviour in the case where the wetting liquid is a poor solvent for the tethered molecules; however, as detailed below, SCALS remain slippery even when in contact with a good solvent [22], [23], [54], [31] (the ‘blended liquid–liquid interface’ of Cheng et al. [55]). A blended interface would allow interactions between the wetting liquid and the (imperfect) substrate. Furthermore, the three-phase contact line is more difficult to conceptualise at a blended interface, as the substrate phase has been replaced by a mixed solvent-SCALS layer; interactions within the mixed layer, between the bulk liquid and the mixed layer, and the mixed layer and the substrate now need to be considered. The simulations of Cohen Stuart et al. [56] demonstrate that the interactions between the solvated polymer and the vapour interface are vital to explaining the finite contact angles of good solvents on brushes. It is still an open question whether the mechanism underpinning SCALS behaviour is shared between solvated and unsolvated systems. Going forward, the term ‘liquid-like’ will refer to surfaces that are both free of nanoscale defects and possess the flexibility to undergo nanoscale rearrangement.

The existing literature on SCALS focuses almost entirely on the attachment of molecules through the silanisation of silica substrates, silicon wafers or glass. As such, the SCALS discussed in this review were prepared by a reaction between surface silanols and silicones or silanes through a silanol intermediary (Fig. 3), with some notable exceptions [28], [55], [57]. SCALS are either prepared via grafting-to or grafting-from processes. In grafting-to, pre-formed chains are attached to the surface; grafted-to layers are typically low in grafting density and molecular weight due to steric hindrance, but are simple to characterise and their preparation is relatively insensitive to environmental conditions. In grafting-from, chains are grown from the surface one repeat unit at a time; these methods are not limited in the theoretical grafting density or molecular weight they can achieve but are more difficult to characterise [58] and are very sensitive to environmental conditions [32]. Water catalyses all reactions in Fig. 3, and hence reaction rates can generally be controlled via the quantity of water in the system. The leaving group also plays a key role in kinetics, as it dictates how aggressively it will react with water to form the silanol intermediary.

In the literature on SAMs obtained by attaching silane molecules to silica, a number of chemical factors have been shown to affect wetting. The first of these is the tendency of tri-functional silanes, i.e., silanes with three leaving groups, L, as shown in Fig. 3, to polymerise along the silane backbone [20], leading to the formation of aggregates on the surface and layers that are thicker and rougher than a monolayer. The second of these is that the wetting properties of the substrate influence the wetting properties of the SAM when the layers are a few nanometers thin, due to the van der Waals forces extending through the SAM [59], [60]. Both of these factors are also expected to affect the wetting properties of SCALS.