“Nonlinear” pursuit of understanding pollutant accumulation and chemistry at environmental and biological interfaces

At present, public anxieties are rising over nanoplastics (diameter < 1 μm) and perfluoroalkyl substances (PFAS) given their alarming concentrations and pervasiveness within the environment.^1–6 These compounds belong to a larger list of pollutants (e.g., polycyclic aromatic hydrocarbons—PAH, polychlorinated biphenyls—PCB, and pharmaceuticals), and understanding the impact of their presence in our environment is an urgent task.⁷ In response to concerns over adverse health effects by exposure to pollutants, significant attention has been directed toward understanding the fate and transport of pollutants through the environment and the associated risks of exposure to humans and other hosts. Due to the amphiphilic nature of many pollutants, it should be expected that they will accumulate at environmental interfaces, for example, PFAS adsorbing to the air–foam interface.⁸ Other types of interfaces found within our environment include water (air–water, mineral–water, etc.), aerosol, emulsion, and nanoplastic interfaces.^9–11 The accumulation of PFAS at the air–water interface within the aqueous thin films of foams near highly contaminated sites provides an important illustration of how pollutants accumulate at environmental interfaces in ways that have important health implications. At one site, it was found that PFAS concentrations were enriched by over 1000-fold in the foam.¹² This translates into increased exposure risks for humans as the foam collects at the shoreline. Thus, it is critical to develop a fundamental understanding of pollutants at all interfaces, as the molecular structure and activity can drastically change as pollutants adsorb to aqueous interfaces, compared to bulk properties.^13,14 Furthermore, for nanoscale material pollutants, such as nanoplastics, the formation of the biocorona at the nanoparticle interface defines the identity of the nanomaterial in the environment.^4,15

Environmental interfaces, however, are not the only surfaces that should be considered in the study of interfacial chemistry of pollutants. Once a host (human or ecological) is exposed to pollutants, the interaction that occurs at the pollutant nanoparticle or membrane interface becomes important. Understanding how PFAS incorporate into the lipid bilayers of various tissues, how eco- and biocoronas around nanoplastics alter their membrane interactions and transport through the body, and whether certain pollutants are capable of crossing tissue barriers are all examples of important health related questions rooted in the interfacial chemistry of pollutants at membrane surfaces.

But how does one measure reaction kinetics at aqueous surfaces, determine molecular structure, and/or measure the adsorption of pollutants to complex and often buried interfaces? When does the curvature and size of nanomaterials, whether they be particles, emulsion, or lipid vesicles, begin to influence pollutant adsorption and chemistry? Embedded in these questions, and others, is a need to isolate the molecular interactions of interfacial bound pollutants from that of unbound pollutants.

The present challenge for common experimental techniques is a lack of either a chemical specificity or an interface specificity. The latter experimental feature is distinguished from interface-sensitivity because to be interface-sensitive, as opposed to specific, still allows for the convolution of experimental signatures of both “near interfacial” and interfacial chemicals. An example of this characteristic would be the sea-surface microlayer (SML).¹⁶ SML sampling methods need to be interfacially sensitive (e.g., rotating drums or glass plates),¹⁷ but the physical region these methods probe will be much deeper than the first few molecular layers that form the interfacial region. Experimentally probing the chemistry of pollutants at nanoparticle or vesicle surfaces becomes even more challenging due to the buried nature of these interfaces. Current methods of directly probing molecules at extended planar or nanoparticle surfaces include Fourier transform infrared spectroscopy, circular dichroism spectroscopy, two-dimensional infrared (2D-IR) spectroscopy, and intrinsic/extrinsic fluorescence.^18,19 However, these methods do not discriminate between molecules at the interface and those unbound in solution. In addition, these methods will not be able to resolve the molecular structure or orientation at the interface.

To overcome these challenges, second-order nonlinear spectroscopies offer an inherently interface-sensitive approach to probe the chemistry of pollutants adsorbed to environmental and biological interfaces. Second-order nonlinear spectroscopies are surface-specific methods that can be built to detect the chemical signatures of pollutants adsorbed to extended planar interfaces or the surface of nanoparticles dispersed in a solvent. In this Perspective, we present the two most common methods, vibrational sum-frequency generation (VSFG), and second-harmonic generation (SHG) and consider how these techniques could be used to study the chemical behavior of pollutants adsorbed to environmental and biological interfaces.

Vibrational sum-frequency generation and second-harmonic generation spectroscopies are two interface-specific methods capable of detecting changes in the composition and molecular structure at interfaces.²⁰ VSFG measures the vibrational spectrum of adsorbed pollutants and ordered water at interfaces,²¹ while SHG measurements are sensitive to ordered interfacial solvents or the presence of adsorbed chromophores in its nonresonant^22,23 and resonant^24,25 variations, respectively. While other optical spectroscopies, such as infrared reflection adsorption spectroscopy (IRRAS), are considered interface-sensitive, the interface specificity of VSFG and SHG comes from the selection rule that any chemical environment must be asymmetrically ordered to generate a VSFG or SHG signal. Two illustrations of the difference between interface specificity versus sensitivity can be found in the ability of VSFG, but not infrared spectroscopy, to detect unbound water oscillators in the top-most layer of water at an aqueous interface or the surface-most layer of polymer thin films.^26–28 Thus, the use of these spectroscopies to study pollutants directly at aqueous interfaces is a straightforward extension of previous applications.

VSFG and SHG techniques are collectively known as second-order nonlinear spectroscopies. The theory underlying second-order spectroscopic methods has been well established, with both mathematically rigorous^29,30 and introductory^21,31 descriptions available. Herein, we describe the information accessible by both spectroscopic methods and provide examples of model environmental and biological interfaces before turning to consider how these methods can be applied to study the interfacial chemical behavior of pollutants.

In the typical VSFG experiment, the sum-frequency response is generated from an aqueous interface by overlapping a mid-IR pulse and a fixed frequency visible pulse at the interface (Fig. 1, top).^32–34 The mid-IR pulse frequency will vary between 1000 and 4000 cm⁻¹, with the pulse bandwidth dependent on the experimental system used in the VSFG study. The intensity of the sum-frequency response is dependent on the intensity of the incident infrared and visible laser pulses, as well as the second-order surface susceptibility $( χ ( 2 ))$⁠. The second-order susceptibility $( χ ( 2 ) ∝ N ⟨ β ( 2 ) ⟩ )$ includes contributions from the surface concentration (N) and average orientation of the molecular hyperpolarizability $( ⟨ β ( 2 ) ⟩ )$ being measured.^21,29 For a vibrational mode to appear in a sum-frequency spectrum, it must be IR and Raman active and the molecule must have a net orientation (i.e., not isotropically distributed).²¹ Therefore, a VSFG spectrum contains information pertaining to the molecular surface density, the local chemical environment, and molecular structure and orientation. For a comprehensive introduction to VSFG, a tutorial series has recently been assembled that approaches both the experimental and theoretical aspects of this technique.^31,35–42

SHG is governed by similar principles as VSFG, except the two incident excitation pulses are degenerate (Fig. 1, bottom). In an SHG experiment, a single pulsed laser is incident on the interfaces and the intensity of the second harmonic of the incident beam is detected. Similar to VSFG experiments, the intensity of the second-harmonic response is dependent on the second-order susceptibility and the intensity of the incident beam. However, the composition of the second-order susceptibility will change depending on whether the two incident pulses in an SHG experiment are resonant^24,25 or nonresonant²² with an electronic excitation at the surface. For the resonant SHG case, where the energy of the two incident pulses is sufficient to excite an electronic dipole, the experiment is sensitive to the orientational ordering of whichever chromophore (e.g., malachite green) is adsorbed to the surface. In the nonresonant SHG case, where the energy of the two incident pulses is less than that required to excite an electronic dipole, the SHG experiment becomes a probe of the orientational ordering of interfacial water and interfacial electric fields.^{22,23,43–47} It is noted that these solvent contributions will be negligible in the resonant SHG condition due to the resonant enhancement of the electronic excitation being the dominant SHG response.

While VSFG and SHG experiments are commonly used to probe planar interfaces, the techniques have progressed and enabled researchers to study the interface of micro- and nanoparticle dispersions. First demonstrated in the 1990s,^48,49 second-harmonic scattering (SHS) spectroscopy under both the resonant⁴⁹ and nonresonant^44,50 conditions has emerged as a powerful probe to understand molecular adsorption and the electric double layer at particle interfaces. A few years later, in 2003,⁵¹ vibrational sum-frequency scattering (SFS) was first demonstrated, leading to the experimental capability to measure vibrational spectra of nanoemulsion,^52,53 vesicle,⁵⁴ and water droplet^55,56 interfaces. Collectively, SHG and VSFG, and their scattering counterparts, provide the essential interface-specificity experimental capabilities and have the potential to prove invaluable to understanding the pollution chemistry at environmental and biological interfaces.

Interface-specific nonlinear spectroscopies are already being applied to study some environmentally relevant interfaces. For instance, the air–water interface has historically been used as a model for aerosol interfaces. Here, VSFG has been used to study the reaction between CO₂ and monoethanolamine (a benchmark scrubber),⁵⁷ the hydration states and interfacial behavior of secondary organic aerosol precursors,^58,59 and changes in the protonation state of fatty acids,⁶⁰ and have observed the enhanced reaction rates of phenol radicals.¹⁴ At the oil–water interface, a model interface used for oil spills, SHS and SFS studies have revealed the impact of interfacial electric fields on surfactant adsorption to nanoemulsion surfaces^61,62 and provided insight into the stabilization of oil droplets by a component of the oil dispersant Corexit.^53,63 VSFG and SHG have also been used to examine the ion driven surface adsorption of carboxyl-containing polymers as a model for humic acid^64,65 and investigate the surface pH of aqueous–oil interfaces.⁶⁶ Last, nonlinear spectroscopic studies of metal oxide nanoparticles have helped provide insight into the hydration state and interfacial electric fields at their surfaces.^67,68

Models of biological membrane surfaces have also been studied using interface-specific spectroscopies. Lipid monolayers at the air–water interface and supported lipid bilayers at the solid–water interface have been the most common membrane models for surface-specific nonlinear spectroscopy.^51,69–73 Golbek et al. used lipid monolayers as membrane models to show that the antimicrobial peptide WLBU2 adopts a ß-sheet structure at mammalian membranes but an α-helix at bacterial membranes, providing molecular insight into its antimicrobial activity.⁷⁴ Engstrom et al. studied the interaction between silver nanoparticles and lipid monolayers, resulting in a mechanistic model to explain size-dependent nanoparticle toxicity.⁷⁵ Conboy and co-workers found supported lipid bilayers to be invaluable models to measure the lipid flip-flop rates using VSFG.^76,77 These are but a few of the many VSFG and SHG studies at these planar membrane models.

Recent advances with nonlinear scattering spectroscopies SHS and SFS open new possibilities with the use of lipid vesicles and lipid stabilized nanoemulsions as 3D membrane models.^54,78–80 For instance, at the nanoemulsion interface, it was found LK-peptides orient in an upright fashion at the particle surface, as opposed to a flat orientation at the extended planar surface.⁸¹ Protein-lipid binding studies are now possible, exemplified by the use of SFS to study lysozyme-lipid binding where it was shown that the protein inserts itself into the lipid surface, causing a conformational change in the lipid acyl chains and a reorientation of the lipid headgroups.⁸² Using a vesicle model, Dedic et al. demonstrated that SHS can detect α-synuclein binding the vesicle surface through changes in interfacial water ordering.⁸³ 

As the Environmental Protection Agency (USA) tightens limits on PFAS concentrations in drinking water (0.02 PPT),⁸⁴ opportunities arise to understand PFAS adsorption at environmental interfaces. Much of what is known about PFAS adsorption has been determined from laboratory experiments of fluorinated surfactants where control can be exerted over the ionic strength and dissolved organic matter concentrations. From well controlled laboratory experiments, it is known that PFAS bulk properties and surface activity will differ relative to similar hydrogenated surfactants. PFAS micelle formation will be impacted by elevated Krafft temperature (minimum temperature required for micelle formation), resulting in changes in solubility and micelle formation.^85,86 The fluorinated alkyl chains also result in PFAS saturating aqueous interfaces at lower bulk concentrations and a reduced CMC.^86,87 At environmental interfaces, where the composition of the aqueous subphase is not controlled, measuring adsorption becomes more difficult. One method of measuring PFAS concentrations near environmental interfaces, within the SML,⁸⁸ uses glass plates to collect the PFAS near the interface.¹⁶ The PFAS content on these slides is then rinsed off, collected, and concentrations measured using analytical methods.¹⁷ However, this method is not surface specific and will collect the PFAS that reside below the top-most molecular layers that constitute the interface. Thus, to experimentally measuring PFAS surface behavior at environmental interfaces requires methods to detect PFAS chemical signatures with a surface-specificity.

One challenge in using traditional adsorption experiments, such as tensiometry, to determine molecular properties at liquid interfaces is the nonspecific nature of these experiments. If the subphase contains unknown ion composition or unknown surface-active organics, any analysis of surface tensiometry results will be complicated by their presence. The C–F vibrational modes of PFAS species provide vibrational resonances that are often isolated from other resonant modes and, thus, open avenues to study their adsorption behavior at natural aqueous surfaces. Interestingly, published VSFG spectra of the C–F stretching region (∼1250–1450 cm⁻¹) are rare, with fluorinated polymers being the focus of the majority of previous studies.^89–91 Some published works have used VSFG to monitor adsorption and calculate the molecular orientation of perfluorononionic acid (PFNA) and fluorinated myristic acid at the air–water interface,^33,92,93 and it has recently been demonstrated that VSFG spectra of the C–F stretching region can be used to estimate surface concentrations of PFAS adsorbed to river water–air interface.⁹⁴ 

We have recently recorded the VSFG spectra of a model PFAS (polyfluorinated dodecylphosphonic acid, F₂₁-DDPA) at the air–water interface of river water (Willamette River, Oregon, USA). In order to apply VSFG to measure adsorption isotherms, spectral contributions from changes in both molecular orientation and surface concentration need to be accounted for. Calculations of the molecular orientation from VSFG data have previously shown the average molecular orientation of fluorinated surfactant, including F₂₁-DDPA, is invariant to changes in the surface density.^33,94 Thus, it can be assumed that the orientation does not change and that the sum-frequency amplitude $( | E V S F G | = I V S F G )$ will be linearly related to the PFAS interfacial density.²¹ VSFG spectra of the F₂₁-DDPA C–F stretching vibrational modes were recorded in the PPP (P – sum frequency, P – visible, P – infrared) polarization combination at the river water–air interface (Willamette River, pH 7.5). We observe a linear relationship between the PFAS surface density and sum-frequency amplitude (Fig. 2). Details of the VSFG spectrometer used in these experiments can be found elsewhere.⁹⁴ At this pH, F₂₁-DDPA is insoluble in the river water. Therefore, the surface concentration of F₂₁-DDPA was directly adjusted by dissolving in CHCl₃ and adding the solution dropwise to the neutral river water surface. After each addition, the C–F stretching spectrum was recorded [Fig. 2(a)]. From these spectra, the electric field amplitude was calculated from the spectral fits, revealing a linear relationship between the sum-frequency amplitude and the known surface density [Fig. 2(b)].

Combining these VSFG spectra of PFAS at the river–water interface with recently published work,⁹⁴ it is clear that VSFG has the potential to probe changes in PFAS surface concentrations at complex aqueous interfaces. However, work is needed to better understand the C–F vibrational region using VSFG and to push the detection limit toward environmentally relevant concentrations. For instance, it might be expected that the SSP polarization combination would yield higher signal intensities than PPP and, therefore, be better suited for detecting low concentrations of PFAS. However, we have found that the spectral response for this vibrational region is stronger than the PPP polarization combination, consistent with previous VSFG studies involving this spectral region.^33,92 Future work is needed to model the C–F spectral response, which will aid in moving it toward a better probe of environmentally relevant concentrations. This is necessary as the bulk concentrations of the data shown here, and from previous work, are at least an order of magnitude higher than the concentrations found in the most contaminated waters.^95,96 Future applications of VSFG to measure environmentally relevant concentrations should aim to detect the interfacial PFAS at bulk concentrations of 1 μg/l or lower. VSFG methods exist that have a lower detection limit (e.g., heterodyne VSFG and electronic SFG)^97–100 and could possibly be utilized to pursue adsorption studies at concentrations relevant to PFAS river contamination sites.

The capacity of pollutants to adsorb to nanoparticle surfaces, causing the nanoparticle to become a reservoir for pollutants, is a phenomenon currently being investigated.^101,102 The adsorption of pollutants to nanoplastic surfaces can impact the fate and transport of the nanoparticles, and there are concerns over how pollutants adsorbed to the nanoplastics alter the toxicity of both the nanoplastics and pollutants.^{101,103–106} Yet, understanding pollutant adsorption at nanoparticle surfaces requires direct probing of the particle surface, which is extremely challenging as the particle surface is buried and distributed throughout a bulk medium. Adsorption studies to extended planar surfaces can produce relevant results for understanding adsorption at larger microplastic (>10 μm) surfaces, but adsorption to charged nanoplastic surfaces cannot be accurately modeled by studies of planar interfaces due to interfacial electric fields suppressing adsorption^61,62 and altering the interfacial structure of adsorbed materials.^52,107

Nonlinear scattering techniques, such as SHS and SFS, can directly probe nanoparticle surfaces and, thus, could provide critical insight into pollutant adsorption to nanoparticle surfaces. SFS has measured surfactant adsorption to the nanoemulsion surface.^52,63,81 Resonant SHS has been used to malachite green adsorption to nanoparticle surfaces,⁴⁹ and nonresonant SHS has helped understand electric field effects on surfactant adsorption to nanoemulsion surfaces.⁶² Therefore, these experimental methods have the potential to significantly contribute to our understanding of the interaction of pollutants at nanoparticle interfaces.

Here, we illustrate how nonresonant SHS could be used to detect pollutant adsorption at environmentally relevant concentrations (Fig. 3). A nonresonant SHS signal, for these experiments, is generated from ordered water molecules at the nanoparticle surface.¹⁰⁸ As such, the SHS signal will contain contributions from the ordered water dipole and interfacial electric fields through the higher-order $χ ( 3 )$ tensor.^22,109 These effects have been well documented and have been reviewed elsewhere.^108,110,111 As a proof of principle, we used nonresonant SHS to detect the adsorption of acetaminophen to polystyrene nanoparticles (PS-NPs). Details of the SHS system for these experiments have been recently published.¹¹² Without acetaminophen PS-NPs generate an SHS scattering pattern typical of 500 nm PS-NPs [black trace, Fig. 3(b)].¹¹³ At 10 ng/l acetaminophen, an SHS signal intensity is observed to increase slightly [red trace, Fig. 3(b)] and even more at 1 μg/l acetaminophen [green trace, Fig. 3(b)]. The observed increase in the SHS intensity is a direct consequence of acetaminophen adsorption to the PS-NP surfaces. Further experiments are required to untangle whether pollutant adsorption causes a change in the SHS signal through a shift in the interfacial electric fields or a reorientation of water dipoles, but perturbations to either could prove useful in the future as a probe of pollutant adsorption to nanoparticle surfaces. Notably, the concentrations probed here are on the same order of magnitude as pharmaceutical pollution found in wastewater and untreated drinking water.^114,115 If SHS were combined with SFS, adsorption isotherms could be measured to determine free energies of adsorption, maximum surface coverage, and the concentrations of pollutants at nanoparticle surfaces.

While questions over whether micro- and nanoplastics can be pollution reservoirs, these particles will become a potential health problem when humans are exposed to the pollutant coated particles. Nanoparticle toxicity is known to vary with the particle’s physical properties (i.e., size, corona, and charge),^{15,75,116,117} yet the molecular mechanism resulting in particle toxicity is a topic of ongoing investigation. When a nanoparticle enters the body, its fate can be influenced by interactions between the particle surface and biological tissue. Coating of the nanoparticle surface by organic or biological mass (eco- and biocoronas) and the composition of the cellular membrane can all have an effect on nanoparticle interaction and subsequent nanoparticle uptake. Therefore, to develop an understanding of the molecular mechanism of nanoplastic toxicity and how it varies when the surface is modified, it will be necessary to study how nanoparticles interact with membrane interfaces.

At biomimetic membranes, VSFG has been used to discern structural perturbations to the membrane structure as a result of nanoparticle adsorption.^75,118,119 By observing vibrational signatures associated with lipid acyl chains and headgroup, the average structural organization of lipids within the model membrane can be determined and monitored under varying chemical conditions. At model cell membranes, the influence of surface coating and particle size/shape on the lipid membrane structure has been studied for Au nanoparticles.^75,119 Some of these observations were accompanied by toxicology studies, enabling VSFG to provide mechanistic explanation for size-dependent nanoparticle toxicity.⁷⁵ A similar approach can be used to observe nanoplastic interactions with biological membranes and reveal how membrane surfaces respond to adsorbed nanoplastics.

Mammalian cell membrane composition consists of nearly 50% phosphatidylcholine containing lipids.¹²⁰ Thus, a monolayer of DPPC (1,2-dipalmitoyl-sn-glycero-3-phosphocholine) was used as an initial model mammalian membrane. The lipid monolayer was formed by the dropwise addition of a lipid solution to a water surface such that the final lipid density was 53 Ų. VSFG spectra measurements of the DPPC monolayer in the absence and presence of PS-NPs (Fig. 4) reveal that the lipid structure is changed in the presence of the nanoplastics. The C-D stretching region (2000–2300 cm⁻¹) probes the structural organization of the deuterated lipid acyl chains.¹²¹ In the spectra shown here, the presence of PS-NPs results in an increase in the CD₃ intensity and loss of CD₂ intensity. This suggests the PS-NPs have “snorkeled” through the monolayer,¹²² compressing the lipids from a liquid expanded-liquid compressed phase into the liquid compressed phase.⁵¹ Changes in the amplitude ratio between the CD₂ and CD₃ symmetric stretches have been used to discern changes in the conformational arrangement of lipid acyl chains.⁷⁴ An increase in the CD₃ symmetric stretch intensity and a decrease in the CD₂ symmetric stretch intensity indicate that the lipid acyl chains contain less gauche defects on average. This spectral change has been observed in VSFG measurements of changing lipid density¹²³ and has since been used to ascertain nanoparticle induced compression to biomimetic membrane surfaces.⁷⁵ 

Additional VSFG experiments could enable researchers to describe a more complete picture of the nanoplastic–lipid interaction. For instance, lipid vibrational groups such as the phosphate and carbonyl stretching vibrational modes can provide additional insight into the bonding interactions and hydration environment during nanoparticle–lipid interactions.^124–129 Further experiments could explore how modifications of nanoplastic surfaces, modeling different coronas, alter the particle–lipid interaction that could elucidate molecular mechanisms behind variations in nanoplastic toxicity.^104,117

This Perspective has focused on illustrating the application of interface-specific nonlinear spectroscopies in the study of interfacial pollutant chemistry, with the experiments outlined here serving to demonstrate how these spectroscopic methods could be applied to the study of pollutant chemistry at environmental and biological interfaces. These examples have included measuring adsorption isotherms at complex aqueous interfaces, detecting pollutant adsorption to nanoparticle surfaces, and discerning molecular structure and orientation within biomimetic membranes exposed to pollutants.

While these applications are promising, challenges exist that present technological hurdles to application and will be important to address. For instance, calculating molecular orientations relies on accurately measuring the second-order surface susceptibility. For nonlinear scattering experiments of nanoparticle surfaces, the incident beams pass through the bulk continuous phase where light absorption can distort the spectra. The process to correct spectra in absorptive media has already been demonstrated in water–oil systems^130,131 but will need to be expanded for complex aquatic matrices in order to determine molecular orientation at the surface of nanoplastics dispersed in natural aqueous media. Understanding the contributing vibrational modes in complex spectral regions will also have an impact on determining the structure of interfacial species. Resolving the interfacial orientation of PFAS from C–F vibrational modes remains a challenge as computational modeling efforts have varied across studies, leading to inconsistencies in the calculated interfacial structures.^33,92 Additional opportunities exist in ensuring spectrometer detection limits are capable of probing the interfacial behavior of environmentally and biologically relevant pollutant concentrations. Reliable model systems and experimental capabilities are being developed to better approximate the complex chemical environment of environmental and biological interfaces, which will provide solutions to these challenges.

It is worth noting that the illustrative experiments we have presented demonstrate none of the potential to observe chemical reactions at the environmental interface. Since chemical and photochemical reactions will transform pollutants into new chemical species, it would be advantageous if the spectroscopic methods presented here could contribute to our understanding of these chemical transformations occurring within interfacial environments. To that end, recent work has demonstrated that there is much to be excited about with regard to the measurement of reaction kinetics at aqueous interfaces. Recent ultrafast VSFG measurements found that the photochemical reaction of phenol becomes ultrafast at the air–water interface, with the reaction rate increasing ∼10⁴ times over the bulk reaction rate.¹⁴ Subsequent theoretical work showed that the enhanced reaction rates were the consequence of the unique hydrogen bonding environment at the air–water interface stabilizing the excited electronic state, resulting in a reduced reaction barrier.¹³² Since many persistent pollutants (e.g., PAH and PCBs) contain aromatic structures, these studies could hint at an important role interfaces hold in the transformation of pollutants within the environment. The approach demonstrated by Kusaka et al. could provide an avenue to study how the hydrogen bonding environment of the aqueous interfaces impacts the chemical reactivity and transformation of pollutants containing aromatic moieties.

Additional developments have recently successfully applied the suite of nonlinear scattering spectroscopies (SHS and SFS) to the study of liquid droplets, providing an avenue to in situ studies of aerosol interfaces. Some of the earliest studies demonstrated that nonresonant SHS could be used to measure the second-harmonic scattering pattern from a water droplet surface and SFS could be used to measure vibrational spectra of an ethanol droplet surface.¹³³ Resonant SHS was then used to measure the adsorption of a dye to the surface of 100 nm aerosol surfaces, demonstrating that aerosol surfaces were accessible with these techniques.¹³⁴ More recently, electronic SFS and vibrational SFS have been used to measure electronic and vibrational spectra of molecules adsorbed to aerosol surfaces.^55,135 Most intriguingly, when considering probing interfacial pollutant chemistry, vibrational SFS was used to measure adsorption isotherms of small organic molecules to the interface of 40 nm aerosols.⁵⁵ While there is an ongoing discussion as to how aerosol particles of this size are capable of generating a measurable SFS signal,^55,112,136 these studies make it possible to envision future in situ studies of pollutant chemistry at aerosol surfaces.

Here, we have described only some of the exciting possibilities of applying surface-specific nonlinear optical spectroscopy to the study of interfacial pollutant chemistry. With the past several decades of development laying the foundation, these methods are in the position to become powerful tools for probing pollutant chemistry within the unique chemical environments of environmental and biological interfaces.

A.P.C. would like to acknowledge the support from the National Science Foundation Ascend Postdoctoral Research Fellowship under Grant No. 2137997. T.W.G. acknowledges the Lundbeck Foundation for support of this work (Postdoc Grant No. R322-2019-2461). T.W.G. acknowledges the funding from the Novo Nordisk Foundation (New Exploratory Research and Discovery—NERD—Grant No. NNF22OC0074640). This article is part of a project that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 Research and Innovation Programme (Grant Agreement No. 819039 F-BioIce).

The authors have no conflicts to disclose.

No ethics approval was required for this work.

Andrew P. Carpenter: Conceptualization (lead); Formal analysis (lead); Funding acquisition (lead); Investigation (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (equal). Thaddeus W. Golbek: Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Visualization (supporting); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.