A mobile setup for simultaneous and in situ neutron reflectivity, infrared spectroscopy, and ellipsometry studies

Neutron reflectivity at the solid/liquid interface offers unique opportunities for resolving the structure–function relationships of interfacial layers in soft matter science. It is a non-destructive technique for detailed analysis of layered structures on molecular length scales, providing thickness, density, roughness, and composition of individual layers or components of adsorbed films. However, there are also some well-known limitations of this method, such as the lack of chemical information, the difficulties in determining large layer thicknesses, and the limited time resolution. We have addressed these shortcomings by designing and implementing a portable sample environment for in situ characterization at neutron reflectometry beamlines, integrating infrared spectroscopy under attenuated total reflection for determination of molecular entities and their conformation, and spectroscopic ellipsometry for rapid and independent measurement of layer thicknesses and refractive indices. The utility of this combined setup is demonstrated by two projects investigating (a) pH-dependent swelling of polyelectrolyte layers and (b) the impact of nanoparticles on lipid membranes to identify potential mechanisms of nanotoxicity


INTRODUCTION
Many challenges in science and engineering involve molecular phenomena at interfaces, for example, the interaction of drugs with cell membranes, 1 the compatibility of artificial materials with blood and tissue, 2 catalysis for energy conversion or pollution reduction, 3 coatings for corrosion prevention 4 or lubrication, 5 and atmospheric processes. 6Modern surface chemistry concerns understanding and control of surface reactions and processes at the molecular level, and its role in shaping modern society is hard to overestimate.However, for solid/liquid interfaces, there is still a need for methods providing in situ data on structure, kinetics, and composition, which are required for thorough understanding of surface processes and for rational design of functional coatings.
For the study of interfacial structure in soft matter and the life sciences, neutron scattering offers particular advantages via isotope contrast variation, especially hydrogen-deuterium replacement in organic molecules, which allows selective highlighting of particular system components. 7Via neutron reflectometry (NR), this gives unique access to structural and dynamic information, which are in many cases not available otherwise.However, NR is also associated with certain disadvantages, related to the information provided by neutron experiments and to limitations in beam intensity.For example, NR cannot identify molecular groups

ARTICLE
scitation.org/journal/rsiand entities, their orientation and conformation, or determine the thickness of extended layer structures.For this, the use of complementary techniques is required.That neutron experiments must be carried out at large-scale facilities results in another complex problem, in that samples might be fragile and/or short-lived, and often need to be prepared in the immediate vicinity of the beamline.Carrying out involved sample preparation procedures away from the home laboratories involves a risk that unexpected parameters affect the outcome.Furthermore, unforeseen events during experiments (beam downtime, equipment failure, and operator mistakes) may cast doubt over the validity of the obtained data.In such cases, independent online monitoring of samples would reduce uncertainties about sample condition, allow for more efficient sample modification procedures, lower the risks for mistakes, and might also provide useful constraints on parameters used in subsequent model fitting.Since neutron scattering is usually only one of the many methods used for sample characterization, integrating complementary in situ techniques with neutron scattering experiments is a natural step.This development has probably progressed further for small-angle and inelastic neutron scattering, where a number of sample environments for complementary in situ techniques have been developed. 8For NR experiments, in situ combinations with other auxiliary techniques are less common, but examples include x-ray reflectivity, 9,10 infrared (IR) spectroscopy, [11][12][13] rheology, 14 voltammetry 15 and Brewster angle microscopy. 16R spectroscopy allows identification of specific molecular entities, 17 as well as molecular conformation and orientation at the interfaces. 18As IR radiation is strongly absorbed by water, analysis of soft matter films in contact with aqueous solutions is usually performed in total internal reflection geometry (ATR-IR), 19 since the short penetration depth of the electric field into the fluid phase, typically of the order of micrometers, avoids excessive intensity losses.The used IR wavelengths are typically in the mid-IR regime, corresponding to ∼500-3500 cm −1 .The advantages of combining NR and IR are well known, and have been exploited before.A permanent setup for simultaneous in situ neutron and ATR-IR analysis was implemented in the design of BioRef , a versatile time-of-flight reflectometer, which was especially dedicated to soft matter characterization at solid/liquid interfaces at the neutron source BER II of the Helmholtz Zentrum Berlin (HZB). 11Due to flexible resolution modes and variable addressable wavelength bands, the instrument enabled a broad range of surface and interface investigations: for example, studies of the structure, stability, and phase conditions of solid-supported lipid oligobilayers in contact with hyaluronic acid and polyelectrolyte solutions to elucidate mechanisms of lubrication in mammalian joints, 20,21 the impact of hydrophilic penetration enhancers on lipid models, 22 and the effect of sugar group hydration on the adsorption of saponin. 23Due to the shutdown of BER II at the end of 2019, BioRef was disassembled and shipped to the Australian Nuclear Science and Technology Organisation (ANSTO) in 2017, where it has emerged as the Spatz reflectometer.In a similar approach, Skoda et al. temporarily combined NR and ATR-IR analysis to study the oxidation kinetics of organic monolayers at the air/water interface. 13A slightly different implementation of an integrated NR and IR setup was reported by Topham et al., who used a portable IR spectrometer equipped with optical fiber cables to guide the external IR beam, in combination with NR. 12 The arrangement served to determine the relationship between charge density and structure of polyelectrolyte brushes.The ANSTO Spatz instrument is currently the only reflectometry beamline offering in situ IR analysis as an option to users.
Ellipsometry is a method for optical characterization of the dielectric properties of reflective materials and thin films. 24,25From the change in the polarization of light reflected from an interface, it can provide information about layers far thinner than the wavelength of the probing light and can also be used to elucidate thicknesses and optical constants of multilayer structures, or density gradients in layers up to micrometre thickness.The change in polarization upon reflection is quantified by the ellipsometric angles Δ and Ψ, which describe the phase shift and relative amplitude of the s and p components of light, respectively.For homogeneous, isotropic layers, the reflectance ratio, i.e., the ratio of the normalized p and s reflectivity components, can be written as ρ = rp/rs = tan Ψe iΔ .From the Ψ and Δ values, the refractive index of the surface, or layer thicknesses, can be determined by modeling, using the laws of reflection and Snell's law.Spectroscopic ellipsometry allows more precise models by measuring the wavelength dependence of theΨ and Δ values.In the case of an inhomogeneous sample, effective medium approximations are used to calculate the refractive index of the mixed phase, based on the volume fractions and optical parameters of the components.A more generalized description of the scattering of light at the interface can be obtained by the use of the Mueller matrix formalism. 26Ellipsometers can be operated in multitude ways 24 and adapted to carry out very different tasks; for example, rapid monitoring of adsorption or other kinetic processes by following the change at a single wavelength, or building large bodies of data for detailed characterization of complex interfaces via acquisition over many different wavelengths.In soft matter, common uses concern thickness monitoring of thin adsorbed layers, determination of optical properties of organic layers, and studies of adsorption kinetics.
Ellipsometry is often used off-line in parallel with NR (and X-ray reflectivity, XRR), since these methods are suitable for similar types of samples, but provide complementary information as they all use different contrast mechanisms (often with excellent agreement 27,28 ).Ellipsometers can be made compact and flexible, but there are only few examples of integration of ellipsometry with NR or XRR, or related techniques, in simultaneous measurements, despite the great inherent benefits.Körstgens et al. used an ellipsometer on a grazing incidence small-angle XRR beamline 29 for simultaneous investigation of temperature and humidity effects on thin films in air.Sebastiani et al. designed a chamber for ellipsometry and NR to quantify oxidation kinetics of organic monolayers at the air/water interface, but did not use them simultaneously. 30ccess to shorter (sub-second) timescales for kinetic studies is of particular interest for NR.Currently, NR kinetics at solid/liquid interfaces are limited to the minutes time scale due to neutron flux limitations and the necessity to cover a certain momentum transfer range to extract unambiguous information, but much shorter timescales are accessible with spectroscopic ellipsometry (SE).In addition, ellipsometry allows for accurate film thickness determination of much thicker layers than is possible with NR, and potential benefits of combining SE and NR thus include the possibility of sample monitoring with higher time resolution, monitoring of large sample thicknesses, and ensuring that samples such as swelling polymer layers, have reached an equilibrium thickness before the reflectometry experiment commences.
In the work described here, the goal was to design and implement an integrated sample environment for in situ characterization including (i) neutrons for structural analysis, (ii) ATR-IR spectroscopy for the determination of molecular entities and their conformation, and (iii) spectroscopic ellipsometry (SE) for rapid and independent measurement of adsorbate layer thicknesses and refractive indices.It was also essential that the equipment is portable and flexible, so that it can be brought from the home laboratories and used at different beamlines, or exchanged between reflectometers with little effort, increasing its use and availability to a broader scientific community.The potential impact of this combination is demonstrated by two scientific projects, investigating (a) pH-dependent swelling of polyelectrolyte layers, with applications in antifouling coatings and (b) the effect of nanoparticles (NPs) on lipid membranes to identify potential mechanisms of nanotoxicity.

OPTICAL SETUP
In NR experiments at solid/liquid interfaces, the neutron beam is accessing the interface from the solid side, thus avoiding the considerable intensity loss due to incoherent scattering from hydrogen atoms in the aqueous phase that would result from a beam traversing the liquid reservoir.Neutron sources are intensitylimited, and the reflected intensity is proportional to the illuminated area.The very shallow incidence angles required (∼0.5 ○ -3 ○ ) mean that relatively large solid crystals are necessary to obtain sufficient reflected intensity to keep experiment time within reasonable limits.For parallel ATR-IR analysis, crystals in the form of trapezoidal prisms are required to permit access for the IR beam.The prism needs a certain thickness for stability, typically ∼1 cm or thicker, to keep the interface planar across the entire sample surface, also when used in combination with liquid flow cells where pressure over the surface is uneven as the rim of the prism is pressed against a seal to retain the liquid.Such prisms are commonly made from silicon, which is useful in the wavenumber range 8000 to 1200 cm −1 (corresponding to 1.2-8 μm wavelengths) for IR applications.We use trapezoidal silicon prisms for IR beam paths with three reflections of the IR beam at the solid/liquid interface, and a schematic of the optical arrangement at the sample is shown in Fig. 1.The prisms were round with 60 mm diameter, or square 60 × 60 mm 2 prisms, with 45 ○ faces for entry and exit of the IR beam.In both cases, the prism thicknesses were 10 mm to obtain symmetric three-reflection IR light paths, as shown in Fig. 1.

Liquid flow cells
Flow cells for liquid exchange were prepared from polyether ether ketone (PEEK) and were designed for both round and square silicon prisms, see Figs. 2(a) and 2(b).The flow cells are sealed against the silicon prisms with 1 mm thick O-rings.The requirement that the ellipsometric optical path crosses any windows at normal incidence to minimize their influence on the measurements is a constraint that complicates the cell design.For ellipsometry, it is also of importance to minimize strain in the windows from their mounting, which could otherwise induce birefringence.This was ensured via a design exerting only normal pressure over the window seal and with a knife-edge arrangement that minimizes effects of uneven tightening; see Fig. 2(c).The ellipsometry beam paths need separate drains to allow efficient exchange of the entire liquid volume, and/or to allow removal of air bubbles, irrespective of cell orientation.These flow paths are placed as close to the windows as possible, to minimize the risk of trapping bubbles by the windows [Fig.2(d)].However, if used as liquid inlets, these paths also aid the exchange of the liquid over the entire surface of the prism, as is clear from images showing the liquid exchange (see below, and supplementary material Figs.S 1 and S 2).It is desired to keep the liquid volume small, both to facilitate rapid liquid exchange, and also for economy, since expensive (often deuterated) reagents and solvents are frequently used.For this reason, a design using narrow channels for the ellipsometer beam was used [see section in Fig. 2(d)], in favor of simpler, but larger cell geometries.The total liquid volume of each cell, including channels for liquids, is ∼2.5 ml, and both types were equipped with flow paths designed to distribute the liquid across the sample area and to facilitate efficient liquid exchange (see Fig. 3 for the liquid volume in the cell for square prisms).This is achieved by a series of small (Ø 0.5 mm) holes in the liquid distribution inserts over a supply channel with larger diameter.The pressure drop over these liquid distribution inserts helps distribute the injected liquid in the channel beneath, allowing efficient distribution of liquid across the length of the prism via a single inlet hole.The demountable liquid distribution inserts are held in place with nylon screws and allow variation of the liquid inlets and outlets with minimal additional machining and hence permit fine-tuning of the liquid flow via quick replacement.They also allow visual access and efficient cleaning of the entire flow path.Trapped air bubbles are hard to expel through Ø 0.5 mm orifices; so, one (square cell) or two (round cell) larger Ø 1.5 mm exit holes were added to aid the removal of trapped air bubbles during cell filling.Liquid filling and draining ports are equipped with flat-bottomed UNF 1/4-28-tapped holes for industry-standard fittings and connectors for low-pressure liquid handling.Whenever possible, a bubble trap (Darwin Microfluidics) is used before the inlet ports to minimize the entry of air bubbles into the cell.
A scheme illustrating the liquid handling is shown in Fig. 4. As is customary on soft matter beamlines, a high-performance liquid chromatography (HPLC) pump with a gradient valve for mixing is used to prepare the various liquid contrasts from stocks of hydrogenated and deuterated solvents.A three-way port on the inlet side of the flow system permits manual syringe injections and purging of the fluidic system without affecting the sample.It also facilitates removal of air bubbles upon filling.Keeping the drain in the filling liquid and reversing the flow via withdrawal through the drain and injection port after initial filling aids in removing trapped bubbles.A peristaltic pump facilitates recirculation of the liquid.
The liquid flow cell and the silicon prism are sandwiched between two aluminum clamping plates containing channels for circulating water for temperature control via an external circulator bath; see Figs. 5 and 6.The PEEK flow cell is bolted directly to the lower of these plates, and a rubber sheet is inserted between the silicon prism and the top aluminum plate to minimize the risk of cracking or chipping of the prism.Interleaving with a thin film (e.g., paper tissue and polymer film) between the rubber and the prism could reduce saturation of specific IR absorption bands due to reflections on the upper ("back") side of the prism, if required, and will also reduce sticking of the prism to the rubber upon disassembly.
The H-shaped upper aluminum clamping plate has cutouts for close access of the IR beam path (see details further down).

Spectroscopic ellipsometry
The ellipsometric function is built around a modular J. A. Woollam iSE spectroscopic ellipsometer with dual rotating compensators, forming a PCSCA (polarizer-compensatorsample-compensator-analyzer) configuration.It covers 190 wavelengths from 400 to 1000 nm and completes a measurement of the entire spectral range within 0.3-2 s.The angle of incidence is fixed to 70 ○ from the surface normal, which is near the Brewster angle of the water/silicon interface, to maximize the difference in the reflected s-and p-components of light.Tilt stages supplied with the ellipsometer are included in the design, see Fig. 6, to allow fine adjustment of the beam path.The source and detector units can be mounted on the tilt plates at different angles around the optical axes, independent of each other, in steps of 30 ○ , which is helpful to avoid space constraints at beamlines, allowing output cables from the units to be positioned in suitable directions.The height of the liquid flow cell assembly can be adjusted via three ball-point set screws in the kinematic base bottom plate, forming a flat-groove-cone type kinematic mount, allowing height adjustment of the liquid flow cell assembly relative to the ellipsometer, with retained lateral alignment (see illustration in Fig. S 4).Precise positioning of different liquid flow cell assemblies for rapid sample exchange at a beamline can be achieved with sheet-metal spacers (e.g., ThorLabs BA2Sx) of required thicknesses inserted between the lower clamping plate and the kinematic base, if necessary.All analyses of ellipsometry data were carried out using the CompleteEASE software provided with the ellipsometer, which also includes compensation for window effects.The ellipsometer was calibrated with a 25 nm thick thermal oxide layer on a Si wafer (also supplied with the ellipsometer).

Integrated infrared spectroscopy and ellipsometry
IR spectra are simultaneously recorded during NR and ellipsometry measurements using a Nicolet iG50 Fourier-Transform IR spectrometer (ThermoFisher Scientific, Dreieich, Germany).Being equipped with a liquid nitrogen-cooled mercury cadmium telluride (MCT) detector, as well as a beam splitter and optical windows made from zinc selenide, it covers a spectral range of 720 to 4000 cm −1 .Resolution can be selected between 0.125 and 32 cm −1 .Accordingly, the time required for one complete scan ranges from about 5.7 to 0.3 s.Zinc selenide has been the preferred optical material as it not only provides good transmittance in a wide spectral region, but in contrast to many other materials used for IR optics, it is also resistant to moisture.The latter significantly reduces the risk of damaging the optics upon transport of the mobile setup to the different neutron facilities.Separation of the spectrometer from the detector allows for high flexibility in the overall optical setup, while maintaining good sensitivity (Fig. 7).
Guidance of the IR beam to and from the silicon substrate is achieved with 1 in.diameter gold mirrors (PF10-03-M02, Thorlabs, Bergkirchen, Germany), which were mounted on adjustable mirror holders.They feature a protective coating, yielding a reflectivity of more than 98% in the spectral range used.The IR beam passes through 1 in.tubes held in place with a system constructed from commercially available parts (30 mm cage system, Thorlabs), which were partially modified in-house, and enters and exits the ATR crystal at an angle of 45 ○ .To allow for accurate positioning of the sample relative to the IR beam in all three spatial directions, the ellipsometric setup (see Fig. 6) containing the liquid flow cell with the ATR crystal is mounted on a lab jack with two movable adapter plates.Due to excellent reproducibility of the sample positioning, the lab jack may be replaced with an aluminum block of appropriate dimensions without significant loss of IR intensity if higher mechanical stability is required.The spectrometer, beam path, and detector are purged with dry gas to minimize rotary vibrational bands, originating from ambient moisture and carbon dioxide, from the spectra.To prevent reflections from the inner walls of the tubular system, which may result in spurious interference patterns in the spectra, the beam diameter is reduced to 1 cm via two adjustable pinholes.
Combined NR, IR, and ellipsometry setup The complete system has been tested on the ILL reflectometers FIGARO 31 and D17. 32Since FIGARO and D17 feature vastly different sample stages, two separate optical setups were developed following the design principles detailed above.At FIGARO, built for horizontal sample orientation, a single aluminum base plate was mounted onto the sample stage of the instrument, on which all other parts of the optical setup were fixed via their respective mounting plates [Fig.7(a)].This design allows for fast assembly and disassembly of the whole setup onsite.For the vertical sample orientation at D17, additional spatial restrictions had to be met in order to avoid collision of the combined ATR-IR/ellipsometry setup with the neutron reflectometer upon movement of the sample stage and neutron detector during the measurements.Thus, the spectrometer was mounted on a platform that was situated lower than the sample itself, while the detector was placed above the sample to minimize IR beam path length [Fig.7(b)].To avoid misalignment of the IR beam caused by lateral torque exerted on the lab jack by the weight of the ellipsometry setup, the lab jack was replaced by an aluminum block of appropriate dimensions as described in the previous paragraph.For both configurations of the setup, care was taken to provide as much space as possible between the optomechanical components and the neutron beam in order to minimize activation of the material during experiments.Successful integration of the mobile setup has also been demonstrated for SuperADAM (ILL), 33 another neutron reflectometer with vertical sample orientation, demonstrating the high flexibility of the new equipment.

PROOF-OF-CONCEPT STUDIES
To demonstrate the utility and function of the combined setup, we have carried out two studies, where the benefits of the complementary ellipsometric and IR data are highlighted.

pH-induced swelling of polyelectrolyte layers
Polyelectrolyte films containing both cationic and anionic ionizable residues (polyampholytes) allow for tuning the structure, swelling, and antifouling properties of the films via variations in composition, structure, pKa/pK b of the ionizable residues, solution salinity, or pH.7][38] In protein adsorption studies, the charge-balanced state of these films is the most protein-resistant, 36,37 since the Coulomb attraction is minimized.However, this state is also the least swollen, which is in contrast to the properties observed in a vast number of nominally uncharged polymer brushes and hydrogels, where strong hydration and associated swelling, is essential for fouling resistance. 39Understanding this particular phenomenon, as well as the overall pH-response of such materials, requires knowledge of the polymer composition, the chain segment density distributions, the overall swelling of the film, as well as the degree of ionization.We have investigated the structure of sequentially grafted films where cationic polymers were deposited onto anionic polymers, seeking to relate the swelling to changes in the distribution of the two monomers within the polymer.This is only possible with the contrast obtained by H/D-substitution in NR. 34 Information relating to interactions between charged groups in these films is accessible via ATR-IR spectroscopy, which could also provide information about inner salt formation in the polyions, water content, hydrogen bonding, protein adsorption, and the diffusion of molecular ions.For thick films, the maximum total thicknesses are not accessible by NR, when decreased separation of Kiessig fringes and/or smearing due to roughness smoothens the reflectogram.Swelling kinetics in polyelectrolytes can be slow, but are often difficult and inconvenient to monitor with NR.SE provides fast monitoring of sample thicknesses, to ensure a steady state has been reached after, for example, a pH change.
While the polyampholyte structure and properties remain important targets of study, here we have used less complex single-component polyelectrolyte films to explore the capabilities of this system, i.e., the anionic poly(methacrylic acid) (pMAA) and cationic poly(2-dimethylaminoethyl methacrylate) (pDMAEMA), prepared to considerably different thicknesses.Measurements were carried out in three different solvent contrasts, D 2 O, CM3 (50:50 D 2 O:H 2 O), and H 2 O, with corresponding neutron scattering length densities (SLDs) (in units of 10 −6 Å −2 ): 6.38, 2.91, and −0.56.Each sample was measured at two different pH values, at ∼pH 3.1 and 12.7 (as measured in H 2 O; values for the other contrasts are included in Table S 1), in 10 mM (unbuffered) solutions of HCl and NaOH, respectively.The reflectometry data from the polymers were fitted to a model consisting of three slabs with sigmoidal interfacial roughnesses, corresponding to the Si/SiO 2 /silane layers, an additional slab for the interface between the polymer and the silane layers, and a polymer layer that was modeled using a sliced sigmoidal volume fraction profile.Further details of the sample preparation and model fitting are included in the supplementary material.Neutron reflectivity data obtained from the pDMAEMA film is presented in Fig. 8 (fit parameters and corresponding data for the pMAA layer are included in the supplementary material in Table S 2 and Fig. S 5).
The resulting polymer chain segment density profiles in Fig. 8(b) clearly show the response of the cationic polyelectrolyte to changes in pH.At low pH, the tertiary amines are protonated, and the polymer is stretched due to charge-charge repulsion within the polymer.At high pH, deprotonation leads to a collapse of the layer.This collapse, however, is not complete, in that some water is still trapped within the film.The rate of these processes cannot be followed by NR, but ellipsometry data show that swelling and collapse proceed at different rates.Figure 9 shows the response of the polymer films after pH changes; the swelling upon addition of HCl proceeds relatively fast, and reaches a plateau after ∼10 min.The collapse upon addition of NaOH is slower and associated with a slow decay in thickness and a corresponding increase in polymer volume fraction within the layer.The slower equilibration in the latter case is presumably an effect of hindered water diffusion upon compression of the polymer.
Comparing the ellipsometric thicknesses and volume fractions with the results of the neutron data fitting, reveals clear differences, as are seen in Fig. 8(b).The results for the dry film are similar, but in this case the volume fraction of the polymer was fixed to 100% in the SE analysis since the contrast is not sufficient to distinguish the polymer, air, and any remaining water.The volume fractions for the hydrated films are reasonably similar, but the thicknesses are not.We interpret this as a difference in sensitivity to the location of the diffuse polymer/water interface, where the uncertainty lies primarily in the ellipsometric measurement, since the optical contrast between the bulk water and the film is very low and cannot be adjusted with isotope selection, as in NR.These differences and the potential for co-fitting of NR and SE data are subject to continuing studies and will be presented in forthcoming work.
The characteristic IR bands of amines are generally very weak, and the spectral features of tertiary amines, such as DMAEMA, are particularly featureless in the absence of N-H stretching modes.The C-N stretching is expected at wavenumbers below 1200 cm −1 and falls outside the useful IR-transparent range of silicon prisms.Thus, to monitor the degree of ionization, it is more fruitful to consider other charged residues, and in the following, we show results of IR spectra obtained on the pMAA films.For each polymer film sample, IR background spectra were obtained with the film hydrated with 10 mM HCl in D 2 O, and all spectra show differences in relation to this state.IR spectra obtained at the same pH, but in different Comparing instead different spectra between similar contrasts, but different pH, reveals differences due to variations in the degree of ionization.Figure 10(b) shows the result for the three contrasts, for a change in pH from 3.1 to 12.7.This change is associated with deprotonation of the carboxylic acids, which is observed as negative bands around 1700 cm −1 , and an associated increase in the carboxylate ion band centered at 1550 cm −1 .It is clear from the result in Fig. 10(b), that the degree of deprotonation is isotopedependent, reflecting the somewhat higher pKa of acids in D 2 O than in H 2 O. 40 Carboxylic acids are weaker acids in D 2 O than in H 2 O, which implies that a larger fraction of acidic residues are protonated in D 2 O at a given pH.This is consistent with the greater negative band for D 2 O than for the other contrasts, at the absorption of the protonated carboxylic acid around 1700 cm −1 .The thickness of the studied pMAA film is a mere 40 Å in the dry state (collapsed), and from the signal-to-noise level in Fig. 10(b), it appears that the sensitivity of the IR setup is overall good for this type of study, even with a very thin polymer layer.The increasing noise at lower wavenumbers in Fig. 10(b) originates from the cutoff in the transmittance of the silicon block, just outside the shown wavenumber range.

Interaction of nanoparticles with lipid bilayers
During the last decade, metal and metal oxide NPs have found increasing use in biomedical applications and consumer products.Hence, the question arises whether they may pose risks to human health.In fact, various studies suggest that the interaction of NPs with cell membranes may affect cell function and contribute to toxic effects. 41Beside in vivo investigations, model systems based on solidsupported lipid membranes 42 play an important role in the study of such processes as single experimental parameters can be deliberately controlled.4][45][46] Yet, a detailed understanding of the underlying processes on a molecular level has not been achieved.
Here, we utilize parallel NR, ATR-IR, and ellipsometry measurements to monitor the interaction of gold nanoparticles (AuNPs) with oligobilayers of 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) at temperatures below and above the main phase transition temperature, TM, of the lipid.DMPC differs from 1,2-dipalmitoylsn-glycero-3-phosphocholine (DPPC), one of the most abundant lipids in human cells, only by a two carbon atoms shorter alkyl chain, but has a lower main value of TM, which is advantageous for the experiments.Lipid films have been prepared onto silicon ATR crystals by spin coating following the procedure described by Mennicke and Salditt. 47The use of oligobilayers-instead of single bilayers-not only amplifies the signal, but also minimizes potential spurious lipid/substrate interactions for all bilayers except the one that is in direct contact with the crystal.Accessibility of the lower bilayers to the NP solution is provided by defects in the bilayer stacks. 20igure 11 shows the neutron reflectivity of DMPC oligobilayers exposed to D 2 O at 18 and 26 ○ C, i.e., at temperatures well above and below the main phase transition, which occurs at 24.2 ± 0.2 ○ C as determined by in-house differential scanning calorimetry (DSC) measurements.Liquid exposure and exchange took place with the liquid flow system described above at a flow rate of 0.6 ml/min.The slab model used to fit the data is sketched in Fig. 12.It is composed of the silicon substrate with a silicon oxide (SiOx) layer of thickness d SiOx on top, followed by a sequence of lipid bilayers and interfacial water layers, and the bulk liquid phase.The inner part of the layer system is repeated N − 3 times so that the total number of lipid bilayers is N.All lipid bilayers are subdivided into two head group slabs and one slab representing the chains, with a thickness of d head and d tail , respectively.While the structural parameters related to the lipid heads and chains were assumed to be equal for all lipid bilayers in the fitting routine, the thicknesses of the innermost, d W1 , and outermost, d W3 , interfacial water layers were allowed to differ from the thickness of the interfacial water layer in the core of the film, d W2 .dtot denotes the total thickness of the lipid film.Details of the fitting procedure, together with the derived molecular structure of the films, are presented in the supplementary material, Table S 3. Figure 11 shows that the model accurately describes the experimental data.The major differences between the film structures at 18 and 26 ○ C are the reduced lipid chain length, the increased area per molecule, and the significant reduction of water content in the lipid layers at higher temperatures.While the first two features are in line with a transition from the gel-like P ß ′ to the liquid-crystalline Lα phase, the third points to an annealing process resulting in more compact lipid films with less water inclusions.The improved packing after annealing at higher temperatures has been confirmed by atomic force microscopy (AFM) measurements on DMPC oligobilayers performed in ambient air.
To monitor the interaction of AuNPs with the lipid layers, AuNPs have been added to the flow system.NR measurements have been restricted to the Q-range around the first Bragg peak to provide a temporal resolution of 60 s.As seen from the inset in Fig. 11 for 26 ○ C, the height and the area of the Bragg peak are reduced with time, which suggests successive degradation of the films upon AuNP contact.As the Kiessig oscillations vanished shortly after the oligobilayers started degrading, the loss of layers was derived from the height of the Bragg peak.In short, starting with the model structure for lipid films in contact with D 2 O as determined above, single lipid layers have successively been removed from the model, while the remaining parameters were kept constant.This way, a calibration curve is obtained, which relates the height of the Bragg peak to the number of remaining bilayers.these bands may also be used to determine the relative amount of lipids deposited onto a surface.The calculation is complicated by the fact that the strength of the electric field of the IR beam exponentially decays from the surface into solution.Thus, lipid bilayers located at larger distances from the surfaces contribute less to the CH 2 bands.However, if the original number of bilayers and their interlamellar spacing is taken from NR data, and the decay length of the electric field is calculated from the angle of incidence and the refractive indices of the ATR crystal and the liquid phase, film degradation may also be determined from the reduction in CH 2 peak area.Details of the underlying procedure are given in the supplementary material.The corresponding Δt 50 values of 8.9 min at 18 ○ C and 4.8 min at 26 ○ C (Fig. 13) confirm the faster degradation of the lipid films above T M and are in very good agreement with the values determined by NR.

Review of Scientific Instruments
The third technique to follow film destruction upon AuNP interaction has been ellipsometry.In contrast to NR, ellipsometry also allows for the determination of large film thicknesses.In principle, this approach is straightforward: The lipid stacks including the interfacial water are considered as a single layer attached to the ATR crystal, and the ellipsometric signal is used to quantify the film thickness.However, to obtain reliable results, accurate knowledge of the refractive indices of the film and the liquid phase is required.Also, if the difference of the two refractive indices is small, thickness determination is complicated.Here, Δt 50 values are 9.2 min at 18 ○ C and 1.7 min at 26 ○ C. Figure 13 compares the thickness reduction of the lipid layers as determined by NR, ATR-IR, and ellipsometry at 18 and 26 ○ C. Except for the ellipsometry measurement at 26 ○ C, all Δt 50 values at a given temperature lie in a close range.We suspect that the small difference in the refractive index between the water containing the DMPC oligobilayers and the bulk water phase may be the reason for the observed deviation.While at 18 ○ C the final film thickness is very similar, some deviation is observed at 26 ○ C depending on the technique used.The reason for this behavior is still being investigated.Yet, all three techniques confirm the lower stability of the Lα phase.

ARTICLE
scitation.org/journal/rsiATR-IR spectroscopy may not only be used to prove the presence of lipids at the interfaces and quantify their relative amounts, but also to monitor the phase condition of the lipids via the exact band positions of the CH 2 stretching modes.In the Lα phase, the modes appear at about 2 cm −1 higher wavenumbers than in the P ß ′ phase.This effect is seen in Fig. 14(b) for the starting point of the experiment.Interestingly, the interaction with AuNPs seems to disturb the structure of both lipid phases, as indicated by the shift of the anti-symmetric [Fig.14

SUMMARY AND CONCLUSIONS
We have successfully demonstrated that online, in situ monitoring of both, infrared spectroscopy and spectroscopic ellipsometry data, during neutron reflectometry experiments is feasible.This was realized with a mobile and modular setup, which allows either of the methods to be used independent of the other.We have demonstrated the utility of the combined setup in two different soft matter systems, namely, the pH-dependent swelling of polyelectrolyte layers and the interaction of nanoparticles with lipid oligobilayers, showing how complementary data, such as chemical information and rapid thickness monitoring, can be added to the neutron reflectometry data, providing additional information, or to support or validate information obtained by neutron reflectometry.The flexibility of the setup is demonstrated by its implementation at beamlines with both horizontal and vertical sample geometries.

SUPPLEMENTARY MATERIAL
See the supplementary material for images showing the liquid exchange in the flow cells, additional details about the design of the hardware, experimental descriptions, and further analysis details of the proof-of-concept experiments and the associated data analyses.

FIG. 1 .
FIG. 1. Schematic view of the optical arrangement.(a) Looking along the neutron beam direction.The infrared and ellipsometric optics share the same incidence plane, but access the sample prism from opposite sides, and the neutron beam is perpendicular to this incidence plane.(b) Looking along the IR/ellipsometry incidence planes.

FIG. 2 .
FIG. 2. Liquid flow cell for (a) round and (b) square silicon prisms.(c) Ellipsometry entry and exit window assembly.The assembly rests in a flat-bottomed recess, and is designed to minimize strain during mounting, which could induce birefringence.The knife edge ensures even pressure across the window also if tightening of the retaining screws is not completely symmetric.(d) Section through the cell in the incidence plane of the ellipsometric and IR beam paths, including also the components in (c) on the right-side port.The section is identical for both the cell types in (a) and (b).

FIG. 3 .
FIG. 3. Schematic view of the liquid flow path in the setup with a square prism, showing only the liquid volume and the prism beneath the liquid.The large diameter exit port on the right facilitates removal of air bubbles during filling.

FIG. 4 .
FIG. 4. Schematic of the flow system, with two of the three injection lines (green) supplying the ellipsometry beam paths.For recirculation (dashed line) a peristaltic pump is used to circulate liquid.The flow cell for round prisms has only one output port (see Fig. S 2), but otherwise the setups are similar.

FIG. 5 .FIG. 6 .
FIG. 5.The liquid flow cell assembly components.For clarity, connecting screws, the window assemblies, and tube fittings are not shown.

FIG. 7 .
FIG. 7. Setups for parallel NR, ATR-IR, and SE measurements, including designs for horizontal (a) and vertical (b) sample orientations.The ellipsometer source and detector units are indicated in yellow and the liquid flow cell assembly in blue.(c) Closeup of the liquid cell and the ATR crystal interfaced to the IR and ellipsometry pathways.

FIG. 8 .
FIG. 8. (a) NR data (markers) obtained on the pDMAEMA film, in three contrasts, at two pH values, and in the dry state, along with the corresponding model fits to the data (solid lines).The dataset for the dry film (black markers, bottom) is correctly positioned on the vertical axis; subsequent datasets have been scaled ×100 relative to the previous dataset for clarity.The large errors at high Q occur because of errors that extend to negative values, and that cannot be drawn accurately on a logarithmic scale.(b) Polymer volume fraction profiles obtained from model fitting to the data in (a).The boxes indicate the thicknesses and volume fractions obtained from the simultaneously acquired ellipsometry data, for the curves with the corresponding color.

FIG. 9 .
FIG. 9. (a) Ellipsometric thicknesses and (b) volume fractions of the pDMAEMA film after pH changes (red and blue markers) or after removal of water for the dry measurements (gray markers).

FIG. 10 .
FIG. 10.(a) Spectra obtained at pH 3.1 in CM3 and H 2 O contrasts, reflecting differences in the bulk composition.(b) Differences between spectra obtained at high and low pH, but for the same bulk contrasts, reflecting changes in the polymer film with increasing pH, are shown for the three tested contrasts.

FIG. 11 .FIG. 12 .
FIG. 11.NR of DMPC oligobilayer films exposed to D 2 O at 18 and 26 ○ C. Symbols represent the experimental data, and solid lines the model fit.The inset shows the temporal development of the Bragg peak after addition of AuNPs at 26 ○ C.
(b)] and symmetric (Fig. S 11) CH 2 band positions with exposure time.Based on the observation that the difference in band positions is reduced with increasing incubation time, one may conclude that the structural features of the lipid oligobilayers become more similar upon nanoparticle interaction.