pubs.acs.org/Langmuir Article Heterogeneous Rate Constant for Amorphous Silica Nanoparticle Adsorption on Phospholipid Monolayers Alex Vakurov, Rik Drummond-Brydson, Nicola William, Didem Sanver, Neus Bastús, Oscar H. Moriones, V. Puntes, and Andrew L. Nelson* Cite This: Langmuir 2022, 38, 5372−5380 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: The interaction of amorphous silica nanoparticles with phospholipid monolayers and bilayers has received a great deal of interest in recent years and is of importance for assessing potential cellular toxicity of such species, whether natural or synthesized for the purpose of nanomedical drug delivery and other applications. This present communication studies the rate of silica nanoparticle adsorption on to phospholipid monolayers in order to extract a heterogeneous rate constant from the data. This rate constant relates to the initial rate of growth of an adsorbed layer of nanoparticles as SiO2 on a unit area of the monolayer surface from unit concentration in dispersion. Experiments were carried out using the system of dioleoyl phosphatidylcholine (DOPC) monolayers deposited on Pt/Hg electrodes in a flow cell. Additional studies were carried out on the interaction of soluble silica with these layers. Results show that the rate constant is effectively constant with respect to silica nanoparticle size. This is interpreted as indicating that the interaction of hydrated SiO2 molecular species with phospholipid polar groups is the molecular initiating event (MIE) defined as the initial interaction of the silica particle surface with the phospholipid layer surface promoting the adsorption of silica nanoparticles on DOPC. The conclusion is consistent with the observed significant interaction of soluble SiO2 with the DOPC layer and the established properties of the silica−water interface. ■ INTRODUCTION The interaction of silica nanoparticles with biological systems, silica nanoparticles often involves the interaction of the nanoparticle with the cell membrane, many studies have in general, has achieved an increasing amount of interest over investigated the interaction of silica particles and silica surfaces the last decade.1−3 At a more specific cellular level, the contact with phospholipid membrane models11−22 because these are of some types of silica particles with the membrane of cells, the underlying backbone of the biological membrane. particularly erythrocytes, is known to cause its rupture and However, in spite of the fact that silica particles have been cellular death.4−7 In addition, because silica nanoparticles have shown to interact strongly with phospholipid membranes, the been extensively proposed as carriers in nanomedicine,8 precise mechanism of interaction and the species involved assessing their potential toxicity is of utmost importance and therein remain uncertain. Pera and coworkers11 studied the coverage and disruption of a primary aim of fundamental research. Furthermore, toxicity is phospholipid membranes with silica and titania nanoparticles. strongly dependent on the type of silica polymorph and its Their approach was comprehensive in that they used both precise surface chemistry. For instance, the US Food and Drug vesicles and supported lipid bilayers to implement their study Administration (FDA) regards amorphous silica as generally in the same way as completed in a previous study.12 Their recognized as safe (GRAS), whereas crystalline silica nano- studies followed two studies13,14 published by this laboratory, particles can be cytotoxic.5 However, so-called fumed silica which conclusively showed the adsorption of silica nano- (amorphous silica nanoparticles produced by gas-phase particles on, and disruption of, phospholipid monolayers and pyrolysis) does induce cytolysis and is also toxic to certain bilayers. Since then, there have been several studies on this cellular strains. In contrast, colloidal silica such as Stoeber silica9 and so-called mesoporous silica nanoparticles are much less harmful to eukaryotic cells and result in significantly less Received: November 24, 2021 cytolysis than fumed silica.4,7 Notably, the sol−gel synthesis of Revised: March 24, 2022 Stoeber silica nanoparticles happens via a monomer-addition Published: April 26, 2022 process similar to naturally occurring silica biomineraliza- tion.9,10 As a result, owing to the fact that the molecular initiating event (MIE) involved in the biological activity of © 2022 American Chemical Society https://doi.org/10.1021/acs.langmuir.1c03155 5372 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article Figure 1. RCV of DOPC on a Pt/Hg electrode recorded at 40 V s−1 control PBS (black) and after; (a) exposure in flow to 0.169 mol dm−3 (SiO2) 17.5 nm radius Stoeber10 synthesized silica nanoparticles in Milli Q water at pH 8.4 (red); (b) exposure to silica dispersion supernatant from (a) (red) and 0.06 mol dm−3 (TiO2) titanium dioxide dispersion (2.5 to 5 nm radius) supernatant (blue); and (c) recovery from interaction with silica dispersion in (a) (red) and with silica dispersion supernatant in (b) (blue). subject15−19 in addition to reports on the interaction of silica present study, therefore, was to comprehensively extend the surfaces with lipid membranes20−22 in general. One of the adsorption rate experiments and analysis to larger particle sizes main conclusions of Pera et al.11 was that electrostatic forces to derive a heterogeneous rate constant for amorphous SiO2 play a key role in the interaction. To elucidate the interaction nanoparticle adsorption on the DOPC monolayer. From the mechanism, they looked at the rate of adsorption of the silica dependence of this rate constant on the particle size, further nanoparticles on to the supported phospholipid bilayers. insight into the adsorption mechanism of silica particles on However, Pera’s results cannot be directly compared with phospholipid layers might be obtained. the results in this study because different experimental setups An additional aim of this study was to use the derived rate were used and different hypotheses were pursued. On the constant in physiologically based pharmacokinetic (PBPK) other hand, fundamentally identical systems are being studied models23 to define the first step or molecular initiating event in employing the same silica particles (Ludox) and the same the uptake of SiO2 particles by the cell membrane,24 leading to lipids (DOPC) in the supported conformation. Nonetheless, the cell membrane’s ultimate damage. Additional experiments Pera et al. extracted no absolute rate constants from their data were also carried out on the effect of the silica nanoparticle and speculated on a barrier to the adsorption rate. The lack of dispersion supernatant on the DOPC layer. In the text, silica conclusions from their rate experiments in their excellent piece nanoparticles and silica surfaces are referred to as such, whilst of work has prompted this laboratory to revisit this laboratory’s molecular silica and associated derived species are referred to adsorption model13 and to significantly extend the experiments as SiO2. together with the analysis. From working with dioleoyl phosphatidylcholine (DOPC) and dimyristoyl PC (DMPC) monolayers and bilayers, it had been concluded that van der ■ MATERIALS AND METHODS The silica dispersions used for the adsorption rate experiments were Waals forces were responsible for the interaction.13,14 This sourced exactly as described previously.13,14 (see Table S1 in the conclusion had been arrived at because the analysis showed the Supporting Information). The nanoparticle hydrodynamic size, as presence of a “reaction layer” (h) on the surface of the particle determined by dynamic light scattering (DLS), was used in all of about 3.2 nm in thickness, above which no interaction calculations. These dispersions were also purified by gel filtration between the silica nanoparticle and the lipid monolayer occurs. immediately prior to the experiment. The dispersions remained stable Subsequent studies cited above15−22 have characterized a and not agglomerated throughout each rate experiment (see Figure S1 similar interaction between lipid membranes and silica surfaces in the Supporting Information). DOPC was obtained to prepare the monolayers, as detailed in earlier studies.25−27 The equipment and and hypothesized on the forces responsible for the interaction, methods used were as described earlier.13,25−27 In principle, the ranging from electrostatic to van der Waals to hydrogen adsorption of silica particles from a static and flowing dispersion onto bonding and/or a combination of all three. This presented a a supported DOPC monolayer was measured by RCV at 40 V s−1 double motivation for this laboratory to return to the work in with a voltage excursion from −0.4 to −1.2 V, as described more detail and assess if any new results and subsequent new previously.13,26 The monolayer was deposited on a fabricated Pt/Hg analysis thereof could clarify the mechanisms of silica electrode positioned in a flow cell. The silica dispersion was in 0.1 mol nanoparticle/lipid membrane interaction. dm−3 KCl buffered with 0.01 mol dm−3 phosphate buffer at pH 7.5, a Preliminary results from this laboratory showed that the pH at which all rate experiments in this study were carried out. The adsorption of the particles on the DOPC affected the depression of adsorption rate of amorphous silica nanoparticles on Hg- two capacitance current peaks 1 and 2 (see Figure S2 in the supported DOPC monolayers is remarkably well behaved.13 Supporting Information), where the extent of peak depression relates This is because the experiments were carried out in a flow cell to the particle coverage of the layer and the size of particles. This peak where the sample flow rate was 10 cm3 per minute in a flow depression is a result13 of the adsorption of the particles onto the lipid cell with a volume of 0.75 cm3. This minimizes any diffusional layer surface, “freezing” the layer14 and impeding the phospholipid control of the adsorption rate. The adsorption rate, which is lipid reorientations. The particle charge had some effect on this measured as suppression in the capacitance current peak in the adsorption process because when the pH of the particle dispersion rapid cyclic voltammogram (RCV) of the DOPC layer, was was taken to 4, which is closer to the PZC of 2.828 of the silica particles lowering their negative charge, the interaction between found to be linearly related to the concentration of SiO2 as particles and DOPC increased slightly13 (see Figure S3 in the mmol dm−3 in aqueous dispersion. Accordingly, the adsorption Supporting Information). It has been shown13 that the current peak 2 rate was normalized to the SiO2 mmol dispersion concen- at more negative potentials was the most exact indicator of particle tration, and this normalized rate (k’) seemed to vary linearly coverage on the monolayer. The measurement of this peak height was and inversely with the particle size.13 The objective of the taken from the height of the peak to the RCV baseline, which in the 5373 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article case of complete peak suppression, is shown in Figure S2 in the cell (0.1 cm × 7.5 cm2) minimizes any diffusional control of Supporting Information. adsorption and means that adsorption is predominantly under kinetic The adsorption can be followed in “real-time” by continuing the control,31 as justifiably assumed throughout this study. RCV monitoring during the adsorption process. In this manner, the Analysis. From the previous study,13 a specific “reactive area” is peak current plotted against time gives a measure of the extent of the area of the silica particle that interacts with the DOPC layer and adsorption versus time. The initial slope of this plot characterizes the extends from the physical contact of the silica particle with the lipid initial rate of adsorption, which decreases as the DOPC surface layer up to the “reaction layer” thickness and is defined by the becomes occupied with particles. These effects are shown very clearly numerator of eq 5 multiplied by, in this instance, the number of in Figure 1B of a previous study13 (Figure S3 in the Supporting particles in 1 mmol. A can therefore be expressed with units of cm2 Information), where preliminary studies of adsorption rate were mmol−1 as follows: carried out by assessing the slope of the plot between times 0 and 50 s. A parameter, k’, is obtained, which is the initial slope of this plot A = π(2Rh + h2) × [(60.08 g mol−1 × 10−3)/2.196 g cm−3] divided by the silica concentration as SiO2 and is defined as the /(4πR3/3)} (1) normalized rate with units μA mol−1 dm3 s−1. In this present study, extensive adsorption rate determinations were carried out precisely where R is the particle radius in cm, h is the “reaction layer” thickness this way for the full range of particle sizes (6.75 to 85 nm radius) and (=3.23 × 10−7 cm), 60.08 g mol−1 is the molecular mass of SiO2, and SiO2 concentrations (0−12 mmol dm−3). The relation between the 2.196 g cm−3 density of amorphous silica. It should be noted that rate of development of the adsorbed silica layer as defined by the when eq 6 in a previous study13 equals zero, all the particles on the slope of the plot and the relative peak height at equilibrium on the Y DOPC monolayer are close-packed such that their “reactive areas” are axis is shown very clearly in Figure S3 in the Supporting Information confluent. At smaller particle sizes eq 613 is forced to equal zero, and and Figure 1B,C of a previous study.13 We define equilibrium the value of the “reaction layer” thickness, h, and the available adsorption on these figures as that attained when the rate decreases to “reactive area” become smaller. a constant value equivalent to that of the control, as shown in Figure There are two methods for calculating the heterogeneous rate 1B,C of a previous study13 and Figure S3 (curves a, and a1 and a2 constant for silica particle adsorption on DOPC: the “bottom-up” and respectively) in the Supporting Information and is directly linked to the “top-down” procedures, respectively. The “bottom-up” approach the complete coverage of the DOPC layer with nanoparticles, as considers the adsorption of SiO2 molecular species on DOPC layers, evidenced by scanning electron microscopy (SEM) images (Figure whereas the “top-down” approach considers the adsorption of silica S4a−c in the Supporting Information). Indeed, in the previous nanoparticles on DOPC layers. The heterogeneous rate constant study,13 this equilibrium peak suppression was plotted against the described in this study is the rate of growth of an adsorbed layer particle size (Figure S5 in the Supporting Information), and the (molecules or nanoparticles) per unit area divided by unit bulk equilibrium adsorption model as eq 6 in the previous study13 was solution or dispersion concentration of those molecules or nano- developed from this. particles. Accordingly, the division of the normalized rate (k’) by the In order to investigate the interaction of soluble SiO2, as opposed current peak depression corresponding to the full coverage of particles to silica particles, with DOPC layers, 17.5 nm radius particle on the DOPC gives the fraction of full coverage per bulk unit solution dispersions (polydispersity index (PDI) = 0.11) in MilliQ water were concentration growing as an adsorbed layer on the DOPC in 1 s. If used. These dispersions had a pH of 8.5 because the silica this value is multiplied by the full coverage (=reciprocal of area per nanoparticles were synthesized using the Stoeber method9 and mmol), a value for the growth rate of the adsorbed layer per unit area contained traces of ammonia. Ammonia traces had no effect on the of the surface from unit solution/dispersion concentration is obtained. agglomeration of the particles because the transmission electron In the “top-down” approach, the same arguments are used as above, microscopy (TEM) image and the DLS plot showed single particles in but instead, the adsorption of nanoparticles is considered. In this case, the dispersion (see Figures S6 and S7 in the Supporting Information). the normalized rate (k’) is converted to a heterogeneous adsorption The concentration of SiO2 in the dispersion was 0.169 mol dm−3. The rate constant (k2 and k2) by dividing it by the maximum depression of dissolved SiO2 was separated from the dispersion by centrifugation at the capacitance current peak specific to that particle size (Figure S2 or 10,000 to 15,000 g, relative centrifugal force (RCF), for 20 min and Figure 1A13 ) and multiplied by the millimolar close-packed collecting the supernatant. Using small Eppendorf 1.5 mL nanoparticle coverage on the DOPC surface. The maximum centrifugation tubes, centrifugation is efficient and sediments almost experimental depression of the capacitance peak current specific to 100% of the particulate silica. Evidence for this is provided by the fact a particular particle size has been used to calculate k2. The maximum that the derived mean count rate of 100−120 counts per second from depression of the capacitance peak current particular to a nanoparticle DLS measurement and the optical density at 360 nm of the size can also be estimated from eq 6,13 which closely fits the data. The supernatant resembled that of MilliQ water. These values of the silica surface area (SA) on a surface in cm2 occupied by a mmol of dispersion are very much larger (see Figure S8 in the Supporting nanoparticles13 is 6.02 × 1020 × 2 × √3 × R2. This area represents Information). The sedimented silica was resuspended and stored in the total area available for binding of the nanoparticles, and the MilliQ water. These dispersions were not gel-filtered prior to reciprocal of this number is the close-packed nanoparticle coverage screening and contained dissolved SiO2 in equilibrium with the per cm2. A factor 'v' or 'v' (depending on how it is calculated) can be particulate silica. Both the dispersions and centrifuged supernatant identified, which is the maximum depression of the capacitance peak were screened for silica nanoparticles and SiO2 interaction with current specific to a particle size multiplied by the SA and in the case DOPC using the flow system platform, as described in a recent where the maximum depression of the capacitance peak current study.29 A supernatant of 0.06 mol dm−3 (TiO2) titania nanoparticle particular to a particle size is obtained from the experiment, v, can be dispersions, prepared from the same source and similarly centrifuged, directly calculated. v can also be estimated from eq 6,13 where the was also screened as a control. Monolayer recovery experiments were maximum depression of the capacitance peak specific to the particle carried out on the interactions of both the dispersion and the size is 32.1 μA × π(2Rh + h2)/(2 × √3 × R2). Therefore v = 32.1 μA supernatant with the DOPC monolayer by introducing control × π(2Rh + h2) × 6.02 × 1020. This is used in the calculation of k2. phosphate-buffered saline (PBS) into the flow sensor module in place The equations for k2 and k2 are as follows: of the dispersion or supernatant in PBS. Accordingly, following the interaction experiment, PBS was flushed through the flow cell for 400 k 2 = 1000 × C tr × k ’/v (2) s with continuous RCV cycling from −0.4 to −1.2 V to allow for and to monitor any recovery of the DOPC layer’s initial structure to occur. k 2 = 1000 × C tr × k ’/v (3) These experiments have been carried out previously to follow the recovery of DOPC layers, following interaction with organic where, because k’ as the numerator is normalized to 1 mmol of compounds.29,30 The forced convection (10 cm3 min−1) in the flow molecules (per dm3) and the denominator ( v or v) is normalized to a 5374 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article Figure 2. Plots of the rate of decrease in the height of capacitance current peak 2 height (V) versus silica nanoparticle bulk concentration as mmol dm−3 SiO2 (Cnp) derived from RCVs of DOPC on Pt/Hg exposed to (a) 15.05 nm radius and (b) 86.05 nm radius nanoparticles. Figure 3. Plots of (a) “normalized” rate (k’) derived from slopes of V versus Cnp plots versus silica nanoparticle radius (R); (b) k’ versus the “reactive area” of silica nanoparticles (A) specific to each value of R; (c) heterogeneous rate constants, k2 (filled black triangles), k2 (blue triangles), and k2cor (filled red circles) for silica adsorption versus R; and (d) k’/k’cor (red-filled circles) and k2/k2cor (black circles) versus R. Errors in (c) and (d) are propagated from the errors of k’ estimation, and for k2, they include the capacitance current peak measurement error. millimole of particles (per cm2), it was necessary to multiply the Because separate platforms were used to obtain Figures S2 and numerator by the number of SiO2 molecules in one silica nanoparticle 1a−c, the current values in Figures S2 and 1a−c differ, but the (Ctr) and, by 1000 to bring all units to cm and by Ctr= 6.02 × 1023 × RCV profiles are similar. Indeed, the RCV profiles following 4/3πR3 × 2.196 g cm−3/60.08 g mole−1. interactions are interesting because they exhibit similarity in ■ RESULTS AND DISCUSSION Figure S2 in the Supporting Information displays voltammo- terms of the interaction of amorphous and Stoeber9 silica particles with the DOPC layer. Although two very different dispersion concentrations of silica were used, the total coverage grams of the interaction of a gel permeation-purified silica of the DOPC layer with nanoparticles was maintained. nanoparticle dispersion; Figure 1a shows a MilliQ water- Particularly evident in Figure 1b is the influence of the equilibrated silica dispersion, and Figure 1b,its supernatant supernatant interaction with the DOPC layer, which elicits a with DOPC monolayers. Also displayed in Figure 1b is the similar but not identical interaction to that of the silica interaction of the supernatant from titania nanoparticle nanoparticle dispersion. Significantly, as shown in Figure 1b, dispersions as a negative control. Figure 1c shows the recovery the capacitance current peaks are shifted to negative potentials. RCVs following silica nanoparticle dispersion and its super- The almost insignificant effect of the titania nanoparticle natant’s interaction with DOPC and subsequent flushing with dispersion supernatant on the RCV profile (as a negative PBS. Figure S2 is reproduced from Figure 1A,13 including an control) establishes the silica nanoparticle supernatant additional RCV plot, and is displayed to clarify the RCV dispersion interaction as significant. response to the silica nanoparticle/DOPC interaction. Of additional interest is the effect of the recovery Capacitance current peaks 1 and 2 are labeled thereon. experiments shown in Figure 1c. Following the interaction of RCVs in Figure 1a−c were obtained in the current work. the silica nanoparticle dispersion supernatant with the DOPC 5375 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article layer and the DOPC layer’s recovery through flushing with “reactive area” will always occupy the same unit area. PBS, the RCV capacitance current peaks are close to Accordingly, the ratio between total SiO2 coverage per specific resembling those of the control. On the other hand, following “reactive area” becomes smaller with larger particle size, and the interaction of the silica nanoparticle dispersion with the the normalized adsorption rate becomes smaller, as seen in DOPC layer and the DOPC layer’s recovery, the RCV profile Figure 3b. shows a more significant departure from that of the control, in The “bottom-up” approach takes the slope of 1.36 ± 0.025 particular, an increase in the baseline capacitance current at × 10−5 μA cm s−1, as shown in Figure 3b, and divides it by the more positive potentials. Intuitively, molecular SiO2 presum- maximum current depression (32.1 μA). The result, using the ably present in the supernatant can be more readily removed above reasoning, is the heterogeneous rate constant of SiO2 from DOPC into solution than adsorbed silica nanoparticles adsorption (k1) of 4.22 ± 0.077 × 10−4 cm s−1. The constancy can be removed into the dispersion phase. of the slope shown in Figure 3b shows that k1 is independent Figure 2a,b shows representative plots of the rate of of the silica particle size because each point shown in Figure 3b adsorption (V) as measured by the initial rate of decrease in represents a different particle size. The equation for the capacitance current peak 2 height in μA s−1 plotted against calculation can be summarized as follows: silica particle concentration expressed as mmol dm−3 of SiO2 (Cnp) for two respective silica nanoparticle sizes (see Figure S9 k1 = k ’ × 1000/(A × 32.1 μA) (4) in the Supporting Information for all rate versus silica concentration plots obtained in this study). The rates of where k’ is multiplied by 1000 because it has units of mmol decrease in the capacitance current peak height in units μA s−1, dm−3 and requires to be converted to mmol cm−3, and A as measured using the method shown in Figure S3 in the comes from eq 1. Supporting Information, are inversely related to the initial rate Results for k2 and k2 estimated using the “top-down” of growth of an adsorbed silica layer on the DOPC monolayer. approach can be observed in Figure 3c. It is seen that k2 =4.48 All V versus Cnp plots measured are linear. ± 0.44 × 10−4 cm s−1 and is generally constant with respect to The value of the normalized rate, k’, is defined as the slope the silica particle size when the particle size is small but shows of the plots of V versus SiO2 concentration (Cnp), and its a small increase at the largest particle sizes. When the standard deviation (SD) is extracted from the linear fit to the experimental maximum decrease in capacitance current peak plots, as shown in Figure 2a,b. An extension of the k’ height is used in the rate constant estimation, the increase in k2 measurement to larger particle sizes compared to the =4.35 ± 0.34 × 10−4 cm s−1 with the particle size is less preliminary measurements carried out previously13 shows significant. The apparent increase in k2 can be minimized by that the relationship between k’ and the particle size is not multiplying the best fit slope value from the k’ vs A plot linear, as originally derived for small particle sizes but actually (Figure 3b) with the “reactive area” specific to each particle exponential (see Figure 3a). The interesting feature of the radius and substituting this value termed k’cor into eq 3. The normalized adsorption rate is that if it is plotted against the adjusted value of k2 expressed as k2cor is now effectively specific “reactive area” (A) available for adsorption, a linear constant with respect to the particle size with a mean value of relationship is obtained with R2 = 0.99 (as shown in Figure 4.22 ± 0.01 × 10−4 cm s−1, which is not significantly different 3b). from the value of k1 estimated from the slope of k’ versus A. In Conceptually, the slope of the plot in Figure 3b represents fact, the effective equality of the “bottom-up” k1 value to the the normalized rate, k’, at a specific nanoparticle size divided “top-down” k2cor value comes from the fact that the algebraic by the “reactive area” at the same size per mmol SiO2, a ratio treatment of the data in both approaches is identical, as shown which is constant irrespective of the particle size (Figure 2b). It by expanding and comparing eqs 3 and 4, which has been is to be noted that the reciprocal of the “reactive area” is the exemplified in the Supporting Information (eqs S4 and S9, surface coverage, of total SiO2 mmoles in the particles, on a respectively). This arises from the relation between the cm2 of “reactive area,” where, by definition, every silica surface depression of the capacitance current peak and the particle SiO2 unit interacts with the DOPC. Accordingly, as the size being underwritten by the way the particles of increasing particles become larger, this coverage becomes smaller.13 At sizes interact with the DOPC layer through the “reaction layer” the same time, it is known13 that a close-packed coverage of (see eq 613). the DOPC monolayer with silica nanoparticles of such size that It is instructive now to assess whether there is any their “reactive areas” are confluent will lead to the total significance in the increase of k2 with respect to the particle depression to the baseline of the capacitance current peak 2 of size. To do this, the deviations of the experimental values of k’ 32.1 μA.13 A RCV reflecting adsorption of the smallest SiO2 from the best fit slope line of the k’ versus A plot were particles on the DOPC layer is also displayed in Figure S2 in estimated from the value of k’ divided by the value of k’cor. the Supporting Information. It is noted that the capacitance These values are displayed in Figure 3d together with the k2/ current peak 2 is almost totally suppressed to be continuous k2cor ratios corresponding to each particle size. The k’ with the baseline capacitance current. SEM images in the measurement errors are also shown in the diagram and are previous study confirmed complete close-packed coverage of divided by k’cor for a given particle size. Similarly, the k’ the silica layer at equilibrium adsorption (see Figure S4 in the measurement errors that are propagated through in the Supporting Information). This is in contrast to Pera et al.11 calculation of k2 are divided by k2cor and displayed in the results, which show that only incomplete coverage (∼30%) of same diagram. The plots of k’/k’cor and k2/k2cor coincide as silica particles on pure DOPC monolayers is obtained. In this expected. It can be seen from this that the increase in k2 with study, however, whenever a unit area of lipid layer covered by a the particle size exactly mirrors the positive deviation of k’ “reactive area” is considered, there is a total interaction of the from the linear best fit line of k’ versus A. This deviation is SiO2 units with that unit area. Therefore, the silica particles more significant at the larger particle sizes and represents an could cover a much larger area than the unit area, but the adsorption rate overestimation error at low values. Indeed, the 5376 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article increase in k2 is only marginally outside the k’ measurement surrounded by a hydrated layer which, by the previous error and is thus barely significant. argument, could suggest that the interaction of TiO2 with the Several points can be concluded from this analysis. First, the phospholipid layer would be similar to that of SiO2. In fact, the evidence points to the same rate-determining step controlling interaction of TiO2 nanoparticles with lipid layers is quite the baseline silica adsorption which is almost independent of different.27 We believe that one of the reasons for this the silica particle size. This can occur if molecular SiO2 and difference is that SiO2 dissolves slowly in water, and TiO2 is small hydrated SiO2 species promote the adsorption of highly insoluble in water. particulate silica onto the lipid monolayers. Because molecular A series of investigations of ZnO,37 CdS,46 and Au and Ag47 SiO2 and small hydrated SiO2 species such as silicic acid, SiO2 nanoparticle interactions with the phospholipid layer shows polymorphs, and silanol groups have considerable H-bonding that it is primarily the nanoparticle surface properties and potential,32 H-bonding between the phospholipid polar groups, coating that drive the interactions. This leads us to think that in particular, the hydrated phosphate moieties and the the “reaction layer” thickness is composed of an interphase hydrated SiO2 species will play a large part in adsorption.21,22 between amorphous silica and the electrolyte consisting of a Second, the meaning of the “reaction layer” thickness has to be progressively hydrated layer of loose SiO2 molecules and more re-examined. Previously, it was thought that this was the complex hydrated SiO2 species, which promote the silica/ maximum distance over which van der Waals forces could DOPC monolayer interaction. The evidence that amorphous operate to facilitate adsorption. However, the rate constant silica dissolves in water and the electrolyte to 0.012%48,49 or data show that it is unlikely that van der Waals forces are 0.002 mol dm−3 in solution supports this idea. Many further implicated as a main driving force in adsorption. Van der Waals studies50−53 have looked at the rate of silica dissolution, which forces do vary inversely with the particle size,33,34 and if they can be of the order of hours53 and which increases markedly promoted adsorption, a negative dependence of the rate with solution pH.53 Soluble SiO2 has been shown to play a part constant on the particle size would be expected. In addition, in the toxicology of particulate silica to the lung tissue.53 A the particle geometry plays a key role in the adhesion force recent study has indeed confirmed the existence of surface- between particles decreasing in order for plates, cylinders, and bound water stabilized by the silanol-rich groups on spheres of a given mass when van der Waals forces are amorphous silica surfaces.54 This has been preceded by a considered.35 In this study, the particle geometry in the form of study which showed that a silica gel-like structure of 1−2 nm curvature is changed by a factor of >10, but clearly, this has no thickness of silicic acid and silanol grows on hydrophilic silica significant effect on the adsorption rate constant. The extent of surfaces.55 This interfacial “reaction layer” might also diffuse the significance of VdW forces in the adsorption of amorphous more within the phospholipid polar group region. A recent SiO2 onto DOPC monolayers could be investigated directly model of the aqueous SiO2 biomembrane interphase has through force-distance measurements using an atomic force predicted that Si(OH)4 and (HO)3SiO− species can penetrate microscopy (AFM) probe coated in silica in close proximity to deep into the lipid membrane. In addition, hydrated SiO2 a supported DOPC layer. A similar study has been carried out clusters are able to spontaneously enter deep into the region elsewhere.36 This laboratory has investigated the behavior of occupied by the lipid heads. There, they form a strong network several other classes of nanoparticles with respect to their of hydrogen bonds with the negatively charged phosphate interaction with phospholipid layers within the platform set up groups56 and possibly also with the cationic polar heads. used in this study including -TiO227 and ZnO,37 where their This study has shown that soluble SiO2 is active on the polarizability compared to SiO2 is more or less similar. If VdW DOPC surface (see Figure 1c). The silica dispersion’s forces were the main factor promoting adsorption, their supernatant interaction is similar to that of the silica dispersion adsorption behavior would be expected to be more or less but somewhat lessened, and the capacitance current peaks are similar and not greatly different, except showing a particle size shifted to more negative potentials. It has already been dependence. In fact, SiO2 adsorption on DOPC is very shown48,49 that at the solution pH 8.4 of the silica dispersion, a different from that of other main classes of nanoparticles in proportion of the solubilized SiO2 will exist as both the that it fits a straightforward equilibrium model.13 Si(OH)4 molecular species and the (HO)3SiO− ion as well as Where both physical and chemical forces are responsible for more complex condensed species.48 The adsorption of the particle adsorption on a surface, the rate constant of the negatively charged (HO)3SiO− on the DOPC layer can give adsorption of nanoparticles can be dependent on their particle rise to the negative potential shift of the capacitance peaks.57 It size in various ways.38 The facilitation of particle adsorption by is appreciated that a small amount of fine colloidal silica will molecular SiO2 and more complex hydrated SiO2 species have escaped the centrifugation process in this study. On the throws new light on the “reaction layer” thickness and the other hand, DLS and UV absorption evidence indicates that “reactive area” concept. the supernatant more nearly resembles MilliQ water than that Recent observations have indicated that amorphous silica is of a silica dispersion (see the Supporting Information). As a covered by a layer of hydrated SiO2 in various forms39−44 and result, it is most likely that the predominant effect of the silica that the adsorption of lipid vesicles on amorphous silica dispersion supernatant on the DOPC capacitance current surfaces is in fact promoted by water molecules associated with peaks is from soluble Si(OH)4 and (HO)3SiO−. From eq 1, the the surface in layers up to 2.5 nm thick.22 This aforementioned available “reactive area” per 0.169 mol from 0.169 mol dm−3 study22 concluded that in the adsorption of lipid vesicles on a (SiO2) amorphous silica particle of radius 17.5 nm dispersion silica surface, there were two thermodynamically stable is 8 × 105 cm2. If a tetrahedrally coordinated structure of both adsorption sites characterized by different widths of the species of dissolved SiO2 is assumed with a distance between water layer between the membrane and the substrate. It is the O atoms of 0.26 nm,58 a tetrahedral base area of ∼0.03 nm2 reasonable to conclude that it is simply a hydrated layer can be estimated. The 0.002 mol dm−3 dissolved SiO2 promoting the adsorption of silica onto DOPC layers. Recent (0.012%) in equilibrium with the silica dispersion will have a evidence45 has shown that TiO2 nanoparticles are also “reactive area” of 3.6 × 105 cm2 per 0.002 mol SiO2 if every 5377 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article SiO2 tetrahedron is close-packed on the DOPC layer surface. finding is entirely compatible with (a) the occurrence of The adsorption of these species on the DOPC layer will significant H-bonding between hydrated SiO2 species and account for the observed effect of the supernatant on the phospholipid polar in particular phosphate groups; (b) the voltammogram of DOPC. In the silica dispersion, both the continuous slow dissolution of SiO2 into the aqueous phase particulate and dissolved species will compete for the DOPC leading to a well-documented hydrated silica-water interface; surface. Because the “reactive area” of the particulate silica is (c) an observed interaction of soluble SiO2 with the DOPC more than twice that of the dissolved species (8 compared with and; (d) the existence of an interfacial “reactive area” with a 3.5 × 10−5 cm2), the DOPC surface will be mainly occupied by defined maximum “reaction layer” thickness between the silica adsorbed silica particles for whom the adsorption rate will be surface and the DOPC layer. Overall, gaining an understanding twice that of the dissolved species. regarding the adsorption rate of amorphous silica nanoparticles Both the observed partial reversibility of the silica nano- on to phospholipid monolayers is of great importance to particle adsorption and the almost total reversibility of the SiO2 elucidate their anticipated applications for silica-based nano- dissolved species adsorption are interesting. The irreversi- particle medical applications including carrier-based drug ble59,60 and reversible61 adsorption of colloids on surfaces has delivery. The results of this work should also encourage been reported and indeed modeled.61 Generally, irreversible spectroscopic and other studies that aim to elucidate the adsorption involves chemical bond formation,59 whereas chemical processes that are underpinning the adsorption physical processes in adsorption such as van der Waals61 and phenomena. H-bonding62 are reversible. In this study, any change in the size of the particle has no significant effect on the activation step, unlike in other studies,60 and the promotion of the ■ * ASSOCIATED CONTENT sı Supporting Information interaction with the lipid layer by hydrated SiO2 species The Supporting Information is available free of charge at through H-bonding with the lipid polar groups is entirely https://pubs.acs.org/doi/10.1021/acs.langmuir.1c03155. consistent with near-reversible adsorption.61 The partial reversibility effected by silica nanoparticle adsorption is Additional experimental details, materials and methods, probably more because of a mechanical disturbance of the and expanded details of calculations used, providing a DOPC surface15 rather than a strong binding between particles full underpinning of the material described in the and DOPC. A consolidation of the interaction due to binding manuscript (PDF) between the hydrated SiO2 species and the hydrated phosphate moieties following the activation step could cause this. This investigation has therefore established two separate ■ AUTHOR INFORMATION Corresponding Author observations on the adsorption of amorphous silica nano- Andrew L. Nelson − School of Chemistry, University of Leeds, particles on DOPC layers. First, the heterogeneous rate Leeds LS2 9JT, U.K.; orcid.org/0000-0002-5842-7459; constant for adsorption is constant and independent of the Email: a.l.nelson@leeds.ac.uk particle size and points to the adsorption of molecular silica species promoting the initial activation step; second, the ability Authors of molecular SiO2 species to interact with the phospholipid Alex Vakurov − School of Chemistry, University of Leeds, layers is confirmed through the significant interaction of the Leeds LS2 9JT, U.K.; orcid.org/0000-0001-6654-4287 silica dispersion supernatant with the DOPC layer. In this case, Rik Drummond-Brydson − School of Chemical and Process the partial reversibility of the adsorption of silica nanoparticles Engineering, University of Leeds, Leeds LS2 9JT, U.K. and the almost total reversibility of the supernatant’s Nicola William − School of Chemistry, University of Leeds, interaction show that an H-bonding interaction rather than Leeds LS2 9JT, U.K. any chemical binding is a critical factor in the adsorption Didem Sanver − Department of Food Engineering, Faculty of process. The clearest explanation of the activation step in the Engineering, Necmettin Erbakan University, Konya 42050, adsorption process is that a layer of dissolved hydrated silica Turkey; orcid.org/0000-0001-6696-9541 molecules within a liquid structure surrounds the amorphous Neus Bastús − Catalan Institute of Nanoscience and silica particle and that this has a thickness of 3.2 nm. It is the Nanotechnology (ICN2), CSIC, The Barcelona Institute of hydrated silica molecules within this structure, which promote Science and Technology, Barcelona 08193, Spain; the interaction with the DOPC layer through H-bonding. orcid.org/0000-0002-3144-7986 Because the hydrodynamic radius of the particle was used in Oscar H. Moriones − Catalan Institute of Nanoscience and the calculations, the “reactive layer” of hydrated SiO2 and Nanotechnology (ICN2), CSIC, The Barcelona Institute of associated water molecules must fall outside the hydrodynamic Science and Technology, Barcelona 08193, Spain; Universitat radius, as measured by DLS. Autònoma de Barcelona (UAB), Barcelona 08193, Spain; ■ CONCLUSIONS Significantly extending a previous study,13 we have shown that orcid.org/0000-0003-0310-0210 V. Puntes − Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC, The Barcelona Institute of the adsorption rate of amorphous silica nanoparticles on Science and Technology, Barcelona 08193, Spain; Fundacio supported DOPC monolayers is governed by a single Hospital Universitari Vall D’Hebron - Institut De Recerca, heterogeneous rate constant of ∼4.2 × 10−4 cm s−1. This Barcelona 08035, Spain; ICREA, Barcelona 08010, Spain shows that one rate-controlling step independent of the Complete contact information is available at: particle size controls the molecular event preceding adsorption, https://pubs.acs.org/10.1021/acs.langmuir.1c03155 which leads to the conclusion that molecular SiO2 and hydrated more complex SiO2 species are the main drivers for Notes the adsorption of the particles on the phospholipid layer. The The authors declare no competing financial interest. 5378 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir ■ pubs.acs.org/Langmuir Article ACKNOWLEDGMENTS nanoparticles interact differently with lipid monolayers compared to lipid bilayers. Environ. Sci.: Nano 2018, 5, 289−303. We acknowledge EU for supporting ENNSATOX GA No. (20) Rapuano, R.; Carmona-Ribeiro, A. M. Supported bilayers on 229244, HISENTS GA No. 685817, and SABYDOMA GA No silica. J. Colloid Interface Sci. 2000, 226, 299−307. 862296. (21) Schneemilch, M.; Quirke, N. Free energy of adhesion of lipid ■ REFERENCES (1) Pieszka, M.; Bederska-Łojewska, D.; Szczurek, P.; Pieszka, M. bilayers on silica surfaces. J. Chem. Phys. 2018, 148, 194704. (22) Vishnyakov, A.; Li, T.; Neimark, A. V. Adhesion of phospholipid bilayers to hydroxylated silica: Existence of nano- meter-thick water interlayers. Langmuir 2017, 33, 13148−13156. The Membrane Interactions of Nano-Silica and Its Potential Application in Animal Nutrition. Animals 2019, 9, 1041. (23) Dubaj, T.; Kozics, K.; Sramkova, M.; Manova, A.; Bastús, N. G.; (2) El-Shetehy, M.; Moradi, A.; Maceroni, M.; Reinhardt, D.; Petri- Moriones, O. H.; Kohl, Y.; Dusinska, M.; Runden-Pran, E.; Puntes, Fink, A.; Rothen-Rutishauser, B.; Mauch, F.; Schwab, F. Silica V.; Nelson, A.; Gabelova, A.; Simon, P. Pharmacokinetics of nanoparticles enhance disease resistance in Arabidopsis plants. Nat. PEGylated Gold Nanoparticles: In VitroIn Vivo Correlation. Nanotechnol. 2021, 16, 344−353. Nanomaterials 2022, 12, 511. (3) Kim, I.-Y.; Joachim, E.; Choi, H.; Kim, K. Toxicity of silica (24) Allen, T. E. H.; Goodman, J. M.; Gutsell, S.; Russell, P. J. nanoparticles depends on size, dose, and cell type. Nanomed.: Defining molecular initiating events in the adverse outcome pathway Nanotechnol. Biol. Med. 2015, 11, 1407−1416. framework for risk assessment. Chem. Res. Toxicol. 2014, 27, 2100− (4) Pavan, C.; Fubini, B. Unveiling the variability of “quartz hazard” 2112. in light of recent toxicological findings. Chem. Res. Toxicol. 2017, 30, (25) Coldrick, Z.; Steenson, P.; Millner, P.; Davies, M.; Nelson, A. 469−485. Phospholipid monolayer coated microfabricated electrodes to model (5) Pavan, C.; Tomatis, M.; Ghiazza, M.; Rabolli, V.; Bolis, V.; the interaction of molecules with biomembranes. Electrochim. Acta Lison, D.; Fubini, B. In search of the chemical basis of the hemolytic 2009, 54, 4954−4962. potential of silicas. Chem. Res. Toxicol. 2013, 26, 1188−1198. (26) Coldrick, Z.; Penezić, A.; Gašparović, B.; Steenson, P.; (6) Shi, J.; Hedberg, Y.; Lundin, M.; Odnevall Wallinder, I.; Merrifield, J.; Nelson, A. High throughput systems for screening Karlsson, H.; Möller, L. Hemolytic properties of synthetic nano-and biomembrane interactions on fabricated mercury film electrodes. J. porous silica particles: the effect of surface properties and the Appl. Electrochem. 2011, 41, 939−949. protection by the plasma corona. Acta Biomater. 2012, 8, 3478−3490. (27) Vakurov, A.; Drummond-Brydson, R.; Ugwumsinachi, O.; (7) Zhang, H.; Dunphy, D. R.; Jiang, X.; Meng, H.; Sun, B.; Tarn, Nelson, A. Significance of particle size and charge capacity in TiO2 D.; Xue, M.; Wang, X.; Lin, S.; Ji, Z.; Li, R.; Garcia, F. L.; Yang, J.; nanoparticle-lipid interactions. J. Colloid Interface Sci. 2016, 473, 75− Kirk, M. L.; Xia, T.; Zink, J. I.; Nel, A.; Brinker, C. J. Processing 83. pathway dependence of amorphous silica nanoparticle toxicity: (28) Balderas-Hernandez, P.; Ibanez, J. G.; Godinez-Ramirez, J. J.; colloidal vs pyrolytic. J. Am. Chem. Soc. 2012, 134, 15790−15804. Almada-Calvo, F. Microscale environmental chemistry: part 7. (8) Barbé, C.; Bartlett, J.; Kong, L.; Finnie, K.; Lin, H. Q.; Larkin, Estimation of the Point of Zero Charge (pzc) for simple metal M.; Calleja, S.; Bush, A.; Calleja, G. Silica particles: a novel drug- oxides by a simplified potentiometric mass titration method. Chem. delivery system. Adv. Mater. 2004, 16, 1959−1966. Educ. 2006, 10, 267−270. (9) Stöber, W.; Fink, A.; Bohn, E. Controlled growth of (29) William, N.; Nelson, A.; Gutsell, S.; Hodges, G.; Rabone, J.; monodisperse silica spheres in the micron size range. J. Colloid Teixeira, A. Hg-supported phospholipid monolayer as rapid screening Interface Sci. 1968, 26, 62−69. device for low molecular weight narcotic compounds in water. Anal. (10) McIntosh, G. J. Theoretical investigations into the nucleation of Chim. Acta 2019, 1069, 98−107. silica growth in basic solution part I−ab Initio studies of the formation (30) Mohamed, H. A.; Shepherd, S.; William, N.; Blundell, H. A.; of trimers and tetramers. Phys. Chem. Chem. Phys. 2013, 15, 3155− Das, M.; Pask, C. M.; Lake, B. R.; Phillips, R. M.; Nelson, A.; Willans, 3172. C. E. Silver (I) N-Heterocyclic Carbene Complexes Derived from (11) Pera, H.; Nolte, T. M.; Leermakers, F. A. M.; Kleijn, J. M. Clotrimazole: Antiproliferative Activity and Interaction with an Coverage and disruption of phospholipid membranes by oxide Artificial Membrane-Based Biosensor. Organometallics 2020, 39, nanoparticles. Langmuir 2014, 30, 14581−14590. 1318−1331. (12) Pera, H.; Kleijn, J. M.; Leermakers, F. A. M. Interaction of silica (31) Fainerman, V.; Miller, R.; Ferri, J. K.; Watzke, H.; Leser, M.; nanoparticles with phospholipid membranes. Chem. Lett. 2012, 41, Michel, M. Reversibility and irreversibility of adsorption of surfactants 1322−1324. and proteins at liquid interfaces. Adv. Colloid Interface Sci. 2006, 123- (13) Vakurov, A.; Brydson, R.; Nelson, A. Electrochemical modeling of the silica nanoparticle−biomembrane interaction. Langmuir 2012, 126, 163−171. 28, 1246−1255. (32) Musso, F.; Mignon, P.; Ugliengo, P.; Sodupe, M. Cooperative (14) Zhang, S.; Nelson, A.; Beales, P. A. Freezing or wrapping: the effects at water−crystalline silica interfaces strengthen surface silanol role of particle size in the mechanism of nanoparticle−biomembrane hydrogen bonding. An ab initio molecular dynamics study. Phys. interaction. Langmuir 2012, 28, 12831−12837. Chem. Chem. Phys. 2012, 14, 10507−10514. (15) Liu, Y.; Zhang, Z.; Zhang, Q.; Baker, G. L.; Worden, R. M. (33) Jiang, K.; Pinchuk, P. Temperature and size-dependent Biomembrane disruption by silica-core nanoparticles: effect of surface Hamaker constants for metal nanoparticles. Nanotechnology 2016, functional group measured using a tethered bilayer lipid membrane. 27, No. 345710. Biochim. Biophys. Acta 2014, 1838, 429−437. (34) Xu, C.-Y.; Zhou, T.-T.; Wang, C.-I.; Liu, H.-Y.; Zhang, C.-T.; (16) Kettiger, H.; Québatte, G.; Perrone, B.; Huwyler, J. Interactions Hu, F.-N.; Zhao, S.-W.; Geng, Z.-C. Aggregation of polydisperse soil between silica nanoparticles and phospholipid membranes. Biochim. colloidal particles: Dependence of Hamaker constant on particle size. Biophys. Acta 2016, 1858, 2163−2170. Geoderma 2020, 359, No. 113999. (17) Alkhammash, H. I.; Li, N.; Berthier, R.; de Planque, M. R. R. (35) Mullins, M. E.; Michaels, L. P.; Menon, V.; Locke, B.; Ranade, Native silica nanoparticles are powerful membrane disruptors. Phys. M. B. Effect of geometry on particle adhesion. Aerosol Sci. Technol. Chem. Chem. Phys. 2015, 17, 15547−15560. 1992, 17, 105−118. (18) Baowan, D.; Peuschel, H.; Kraegeloh, A.; Helms, V. Energetics (36) Liu, X. Interactions of silver nanoparticles formed in situ on of liposomes encapsulating silica nanoparticles. J. Mol. Model. 2013, AFM tips with supported lipid bilayers. Langmuir 2018, 34, 10774− 19, 2459−2472. 10781. (19) Asghari Adib, A.; Nazemidashtarjandi, S.; Kelly, A.; Kruse, A.; (37) Vakurov, A.; Guillermo Mokry; Drummond-Brydson, R.; Cimatu, K.; David, A. E.; Farnoud, A. M. Engineered silica Wallace, R.; Svendsen, C.; Nelson, A. ZnO nanoparticle interactions 5379 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380 Langmuir pubs.acs.org/Langmuir Article with phospholipid monolayers. J. Colloid Interface Sci. 2013, 404, phosphatidylcholine monolayer on Hg. Electrochim. Acta 2015, 155, 161−168. 458−465. (38) Wang, H.; Shadman, F. Effect of particle size on the adsorption (58) Van Ginhoven, R. M.; Jónsson, H.; Corrales, L. R. Silica glass and desorption properties of oxide nanoparticles. AIChE J. 2013, 59, structure generation for ab initio calculations using small samples of 1502−1510. amorphous silica. Phys. Rev. B 2005, 71, No. 024208. (39) Cimas, Á .; Tielens, F.; Sulpizi, M.; Gaigeot, M.-P.; Costa, D. (59) Senger, B.; Voegel, J.-C.; Schaaf, P. Irreversible adsorption of The amorphous silica−liquid water interface studied by ab initio colloidal particles on solid substrates. Colloids Surf. A Physicochem. molecular dynamics (AIMD): local organization in global disorder. J. Eng. Asp. 2000, 165, 255−285. Phys.: Condens. Matter 2014, 26, No. 244106. (60) Adamczyk, Z.; Jaszczółt, K.; Michna, A.; Siwek, B.; Szyk- (40) Rehl, B.; Gibbs, J. M. Role of Ions on the Surface-Bound Water Warszyńska, L.; Zembala, M. Irreversible adsorption of particles on Structure at the Silica/Water Interface: Identifying the Spectral heterogeneous surfaces. Adv. Colloid Interface Sci. 2005, 118, 25−42. Signature of Stability. J. Phys. Chem. Lett. 2021, 12, 2854−2864. (61) Salipante, P. F.; Hudson, S. D. Reversible adsorption kinetics of (41) Dalstein, L.; Potapova, E.; Tyrode, E. The elusive silica/water near surface dimer colloids. Langmuir 2016, 32, 8565−8573. interface: Isolated silanols under water as revealed by vibrational sum (62) Schlesier, T.; Metzroth, T.; Janshoff, A.; Gauss, J. R.; frequency spectroscopy. Phys. Chem. Chem. Phys. 2017, 19, 10343− Diezemann, G. Reversible hydrogen bond network dynamics: 10349. Molecular dynamics simulations of calix [4] arene-catenanes. J. (42) Shchukarev, A.; Rosenqvist, J.; Sjöberg, S. XPS study of the Phys. Chem. B 2011, 115, 6445−6454. silica−water interface. J. Electron Spectrosc. Relat. Phenom. 2004, 137- 140, 171−176. (43) Sulpizi, M.; Gaigeot, M.-P.; Sprik, M. The silica−water interface: how the silanols determine the surface acidity and modulate the water properties. J. Chem. Theory Comput. 2012, 8, 1037−1047. (44) Rimola, A.; Costa, D.; Sodupe, M.; Lambert, J.-F.; Ugliengo, P. Silica surface features and their role in the adsorption of biomolecules: computational modeling and experiments. Chem. Rev. 2013, 113, 4216−4313. (45) Benkoula, S.; Sublemontier, O.; Patanen, M.; Nicolas, C.; Sirotti, F.; Naitabdi, A.; Gaie-Levrel, F.; Antonsson, E.; Aureau, D.; Ouf, F. X.; Wada, S. I.; Etcheberry, A.; Ueda, K.; Miron, C. Water adsorption on TiO2 surfaces probed by soft X-ray spectroscopies: bulk materials vs. isolated nanoparticles. Sci. Rep. 2015, 5, No. 15088. (46) Zhang, S.; Chen, R.; Malhotra, G.; Critchley, K.; Vakurov, A.; Nelson, A. Electrochemical modelling of QD-phospholipid inter- actions. J. Colloid Interface Sci. 2014, 420, 9−14. (47) William, N.; Bamidoro, F.; Beales, P. A.; Drummond-Brydson, R.; Hondow, N.; Key, S.; Kulak, A.; Walsh, A. C.; Winter, S.; Nelson, L. A. Tuning stable noble metal nanoparticles dispersions to moderate their interaction with model membranes. J. Colloid Interface Sci. 2021, 594, 101−112. (48) Alexander, G. B.; Heston, W. M.; Iler, R. K. The solubility of amorphous silica in water. J. Phys. Chem. 1954, 58, 453−455. (49) Dove, P. M.; Han, N.; Wallace, A. F.; De Yoreo, J. J. Kinetics of amorphous silica dissolution and the paradox of the silica polymorphs. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 9903−9908. (50) Seidel, A.; Löbbus, M.; Vogelsberger, W.; Sonnefeld, J. The kinetics of dissolution of silica ‘Monosphere’ into water at different concentrations of background electrolyte. Solid State Ionics 1997, 101- 103, 713−719. (51) Diedrich, T.; Dybowska, A.; Schott, J.; Valsami-Jones, E.; Oelkers, E. H. The dissolution rates of SiO2 nanoparticles as a function of particle size. Environ. Sci. Technol. 2012, 46, 4909−4915. (52) Rimer, J. D.; Trofymluk, O.; Navrotsky, A.; Lobo, R. F.; Vlachos, D. G. Kinetic and thermodynamic studies of silica nanoparticle dissolution. Chem. Mater. 2007, 19, 4189−4197. (53) Larson, R. R.; Story, S. G.; Hegmann, K. T. Assessing the solubility of silicon dioxide particles using simulated lung fluid. Open Toxicol. J. 2010, 4, 51−55. (54) Schrader, A. M.; Monroe, J. I.; Sheil, R.; Dobbs, H. A.; Keller, T. J.; Li, Y.; Jain, S.; Shell, M. S.; Israelachvili, J. N.; Han, S. Surface chemical heterogeneity modulates silica surface hydration. Proc. Natl. Acad. Sci. U. S. A. 2018, 115, 2890−2895. (55) Vigil, G.; Xu, Z.; Steinberg, S.; Israelachvili, J. Interactions of silica surfaces. J. Colloid Interface Sci. 1994, 165, 367−385. (56) Delle Piane, M.; Potthoff, S.; Brinker, C. J.; Colombi Ciacchi, L. Molecular dynamics simulations of the silica−cell membrane interaction: insights on biomineralization and Nanotoxicity. J. Phys. Chem. C 2018, 122, 21330−21343. (57) Rashid, A.; Vakurov, A.; Nelson, A. Role of electrolyte in the occurrence of the voltage induced phase transitions in a dioleoyl 5380 https://doi.org/10.1021/acs.langmuir.1c03155 Langmuir 2022, 38, 5372−5380