Adsorption of soft NIPAM nanogels at hydrophobic and hydrophilic interfaces: Conformation of the interfacial layers determined by neutron reﬂectivity

The application of stimuli-responsive microgels and nanogels in drug delivery, catalysis, sensing, and coatings is restricted currently by the limited understanding of the factors inﬂuencing their adsorption dynamics and structural changes at interfaces. We have used neutron reﬂectivity to resolve, on the Ångström scale, the structure of 5% crosslinked N -isopropylacrylamide nanogels at both hydrophobic and hydrophilic interfaces in situ, as a function of temperature and bulk nanogel concentration. Our results show that the higher ﬂexibility given by the low crosslinker content allows for a more ordered structure and packing. The adsorption of the thermoresponsive nanogels is primarily driven by temperature, more speciﬁcally its proximity to its volume phase transition temperature, while concentration plays a secondary role. Hydrophobic interactions drive the conformation of the ﬁrst layer at the interface, which plays a key role in inﬂuencing the overall nanogel structure. The mobility of the ﬁrst layer at the air–water interface as opposed to the interfacial conﬁnement at the solid (SiC8)-liquid interface, results in a different conformation, a more compact and less deformed packing structure, which ultimately drives the structure of the subsequent layers. The evidence for the different structural conformations determined by the degree of hydrophobicity of the interface provides new knowledge, which is essential


Introduction
The functional activity of biological systems are often dependent on specific interactions that influence their structure and complex formation [1][2][3] and scientists have been developing functional materials that rely on self-assembly to attain specific functions for applications as coatings [4], sensors [5] and drug delivery vehicles [6].Progress in nanotechnology has widened the pool of materials, with chemistries that allow further tailoring of the materials' properties, including the introduction of various stimuli [7][8][9].
Soft colloid-like materials have high potential for applications, due to their intrinsic deformability and elasticity, which can significantly influence their behavior [10][11][12][13].Studying the interactions of these materials at interfaces, understanding the adsorption behavior and kinetics, and identifying their structural features is a key priority in order to develop new tailored materials.Among the various matrices, covalently crosslinked polymeric microgels and nanogels have emerged as versatile materials for multiple applications in biotechnology, drug delivery, electronics and sensor technology [14][15][16][17][18].By changing their chemical structure through a combination of different monomers and cross-linkers, the morphology and physicochemical properties of the particles can be tailored and stimuli-responsive characteristics introduced [19].
N-isopropylacrylamide (NIPAM)-based microgels and nanogels have attracted considerable interest due to their tunable thermoresponsive properties, surface activity, and ability to form stable colloidal solutions [20][21][22][23][24][25][26].The hydration shells of the side chains are the main driver of the gels' phase behavior and interfacial activity as a function of temperature [27][28][29].The soft character of the nanogel's matrix, controlled by the extent of crosslinking, has a significant influence on the degree of the lateral elastic deformation and assembly at the interface [30,31].
The behavior of polymeric surface-active nano/micro-particles at interfaces has been the subject of both experimental and theoretical studies, focusing on their use for emulsion stabilisation [32], internal structure studies [33], stiffness tomography [34], rheology and mechanical properties [35] at interfaces.The flexibility and potential responsiveness of these soft materials offer potential advantages compared to rigid particles [36], such as the ones used for Pickering emulsions.However, a better understanding of the dynamics and structural conformation of these soft nanoparticles, and their influence on the subsequent self-assembly towards the aqueous subphase at the microscopic level, is essential in optimizing their future applications [37].Neutron reflectivity, grazing small angle scattering, and molecular simulation studies of microgels spin-coated on a hydrophilic solid substrate have been reported as a function of crosslinker content and temperature [38,39].The data suggest a strong influence of surface confinement at the solid substrate.A combined imaging and simulation approach to visualize in 3D the deformation of individual NIPAM-based nanogel particles at solid-liquid interfaces has been reported by Alvarez et al. [40].The team identified the key role played by the hydrophilicity of the solid surface in determining the extent of the matrix deformation into the 'fried-egg' conformation of the adsorbed layer.Additionally Harrer et al. [41] investigated the effect of interfacial confinement on the volume phase transition temperature (VPTT) of NIPAM-based microgels at the air-water interface, albeit via scanning electron microscopy and atomic force microscopy (AFM) of dry microgel samples.
In recent years several studies have exploited computational approaches to model interfacial behavior, especially at the oil-water interface [42][43][44].Arismendi et al. [45] reported large-scale molecular dynamics simulations of the impact of particle deformability on solvent permeability when soft nanoparticles were placed at the liquid-liquid interface.They concluded that the packing efficiency of soft nanoparticles under interfacial confinement influenced their solvent uptake and mixing capacities.Recently, Camerin et al. [46] reported a combined quantitative and experimental study based on advanced modelling of microgels at the water-oil interface complemented by cryo-electron microscopy experiments and AFM.Both the computational and experimental results suggested that the microgel's corona plays an important role in determining microgel extension at the plane of the interface.
These important developments in the studies of soft polymeric particles at interfaces using imaging, modelling, and combinations of different experimental techniques, do not provide direct experimental observation and understanding of the structure of buried adsorbed layers away from the interfaces towards the bulk aqueous phase, as they lack appropriate resolution and are only sensitive to the surface monolayer.A detailed nanoscopic description of the utmost first monolayer as well as buried portions of surface-bound soft polymeric materials perpendicular to the interface (''3D" structural information) is essential to understand the interaction mechanisms of these materials with molecules and surfaces, and therefore tailor their applications [47].Recent examples of its importance in applications include patterned NIPAM-based microgels on a solid surface for the capture of targeted cancer cells [48], and a power output of an enzymatic biofuel cell controlled by pNIPAM chains [49].
Specular neutron reflectivity (NR) [50][51][52] is a powerful technique that provides Ångström-scale information on the structure and composition of surface-active materials perpendicular to an interface.Significant advances in kinetic and dynamic studies of soft matter at the air-water interface have been achieved in recent years as a result of the exploitation of this relatively new methodology [53].We previously reported on the relationship between chemical structure and interfacial behavior of NIPAM-based nanogels using a combination of NR with isotopic contrast variation, surface tension, and ellipsometry, complemented by dynamic light scattering (DLS) and transmission electron microscopy [54,55].We demonstrated that the degree of crosslinking (between 10% and 30%) has a profound effect not only on bulk morphology of a nanogel solution but also on adsorption kinetics, steady state adsorbed amount and resulting interfacial structures.The interactions and transport of nanogels across lipid (ceramide) monolayers [56] and multilayers [57] have also been investigated by our group, contributing to the understanding of the relationship between nanogels' structural features and interfacial behavior.
In this work we studied the self-assembly and interfacial behavior of NIPAM nanogels crosslinked with 5% methylenebisacrylamide (MBA) -as a function of concentration and temperature at both air-water and solid-water interfaces, the latter using hydrophilic and hydrophobic modified-silicon substrates.The choice of low crosslinked nanomaterials was driven by the wide range of applications currently reported for these matrices [29,[58][59][60][61].The surface activities of the nanogels at the air-water interface were measured using a force tensiometer.Adsorption and volume fraction profiles perpendicular to the interface were determined from NR in combination with isotopic contrast variations at air-water and solid-water interfaces.Furthermore, the surface conformation of nanogels on both hydrophilic and hydrophobic solid supports at room temperature was also resolved using fluid mode AFM, providing visualization of nanogels at the solid-water interface to complement our NR results.

Nanogel synthesis
NIPAM-based nanogels with 5% crosslinker content were synthesized by high-dilution radical polymerization following our previously reported protocol [19,62].The required amount of monomer NIPAM and crosslinker MBA were dissolved in anhydrous DMSO in a Wheaton bottle to a 1% total monomer molar concentration.The bottle was sealed after the initiator AIBN was added at 1% of total double bond moles.The solution was vigorously bubbled with nitrogen for 30 min to remove dissolved oxygen.After that, the bottle was heated to 70 °C (oven) for 24 hours.The post-polymerization solution was dialyzed (MWCO 3500 Da) against distilled water for at least 3 days with frequent water exchange before being frozen and lyophilized by freeze drier (Lyotrap, LTE Scientific, Oldham, UK).

Nanogel characterizations
Fluffy nanogel flakes were dissolved in water to give a stable colloidal solution prior to characterizations.Gel solutions were sonicated for 5 min before being filtered through GHP filters with a pore size of 0.20 lm (VWR, Radnor, PA) to eliminate any dust or aggregates that could disturb the measurements and induce a large error.Hydrodynamic diameters of nanogels were measured in triplicate by dynamic light scattering (DLS) with a Zetasizer Ultrafrom Malvern Instruments (Malvern, UK).The surface tension measurements were performed using a Krüss K9S (C11227) Wilhelmy plate tensiometer (Hamburg, Germany) as a function of temperature and concentration.The VPTTs of nanogel solutions were determined by monitoring optical transmittance ranging from 25 to 65 °C (0.5 °C min À1 ) using a Cary 100 UV-vis spectrophotometer (Agilent Technologies, Santa Clara, CA) at 500 nm.
The first order derivative of percentage of transmittance against temperature was plotted, which was fitted with the Gaussian peak function to find the point of inflection.The x-axis of this point was then taken as the VPTT value.

Atomic force microscopy (AFM) measurements
AFM was carried out on a Bruker Dimension Icon system (Bruker, Billerica, MA) in PeakForce Quantitative Nanomechanical Mapping (QNM) mode.Samples were imaged at a resolution of 512 samples per line across 512 lines at a rate of 0.5 Hz with Bruker ScanAsyst Fluid + silicon nitride tips (spring constant = 0.7 N/m; tip radius = 2 nm).AFM samples were prepared by submerging hydrophobic SiC8 and hydrophilic Si substrates in glass petri dishes filled with 0.1 mg ml À1 nanogel solution for 20 min prior to AFM imaging.The samples were imaged directly in the nanogel solution submerging the substrates by lowering the AFM probe into the petri dishes.AFM images were subsequently analyzed using Bruker Nanoscope Analysis software.

Neutron reflectivity (NR)
Conformations of nanogels at air-water and solid-water interfaces were characterized on the FIGARO reflectometer [63] at the Institut Laue-Langevin (ILL), Grenoble, France, and on the SURF reflectometer [64] at the Rutherford Appleton Laboratory, Didcot, UK, respectively.
The silicon crystals used as solid substrates in NR experiments were cubic in section with a diameter of 100 mm and a thickness of 10 mm.The silicon surface was initially cleaned using the RCA method [65].The substrates were then processed to chemically remove water from the surface prior to hydrophobic modification with a silane coupling agent (octyltrichlorosilane, C8), which has been proved to render a chemically bound uniform hydrophobic layer on the silica surface [66].
In an NR experiment, a highly collimated neutron beam is reflected from a flat interface.The intensity of the reflected beam was measured as a function of momentum transfer Q = 4psin h/k normal to the interface via the time-of-flight technique, where h is the grazing angle of incident neutrons, and k is the neutron wavelength.The NR profiles were measured at two incident angles (h = 0.624°and 3.78°) on the FIGARO reflectometer, and at three different grazing angles (h = 0.35°, 0.65°, and 1.5°) on the SURF reflectometer to provide a wide Q-range hence the highest sensitivity to the interfacial structure.The sample was illuminated at a resolution dQ/Q % 8%.The water sub-phase was D 2 O or null reflectivity water (NRW, whose scattering length density to neutrons is zero) for the air-water interface.In the case of the SiC8water interface, the aqueous subphase was D 2 O, H 2 O or a mixture of D 2 O and H 2 O, which is contrast matched to the scattering length density of silicon 2.07 Â 10 À6 Å À2 (contrast matched silicon CMSi).
NR data were analyzed using the exact optical matrix method as previously reported [67].The analysis of NR data includes generating model parameters such as scattering length density, thickness, roughness, and number of layers, calculating the reflectivity, and comparing the result to the data obtained from the experiments.This process was repeated until the differences between the global fit and the experimental data in different isotopic contrasts have reached a minimum.The optimal fitted parameters were combined to give the scattering length density profile of the sample, which represents the variation in composition perpendicular to the interface.The nature of model fitting of neutron reflectivity data necessitates finding the simplest model, i.e., the model with the minimum number of fitting parameters.
When resolving the volume fraction and adsorbed amount of nanogel in each layer in NR data recorded the air-water interface, we assumed that the first layer consisted of nanogels, air and water whilst the subsequent layers were made up of nanogels and water.Based on these assumptions, for the first layer.
where u is the volume fraction and Nb is the scattering length density.
As two different isotopic contrasts were used, it was possible to calculate the volume fraction of water, air, and nanogels in each layer.
In the case of solid-water interfaces, we assume no air bubbles exist at the interface (u air ¼ 0Þ, i.e. each layer contained only nanogels and water.Hence the equation ( 3) can be applicable to solidwater interfaces and the additional isotopic contrasts further reduced uncertainty in the fitted parameters.
The adsorbed amount of nanogel at the interface can be calculated from volume fraction of nanogels.
where q is the density of dry nanogels and d is the fitted layer thickness.The density of dry 5% crosslinked MBA-NIPAM nanogels was estimated to be 1.3 g cm À3 using an AccuPyc 1330 helium pycnometer (Micrometric Ltd, Lincoln, UK).

Nanogel synthesis and characterization
The adsorption of nanogels at interfaces, unlike hard particles, is determined by their interfacial activity and conformation, [68,69] the latter depending on the structural flexibility of the matrix and the hydrophobic/hydrophilic properties of the interface.NIPAM-based nanogels with 5% MBA as crosslinker were prepared by high dilution radical polymerization, and obtained with >80% monomer conversion as determined by nuclear magnetic resonance, >80% chemical yield, and an average particle size of $7 nm (by volume distribution) determined by DLS at room temperature (Fig. S1).The VPTT, indicative of thermoresponsive behaviour, was found to vary between 316 K and 310 K depending on nanogel concentration (0.01 mg ml À1 to 1 mg ml À1 in H 2 O; Table S1).Additionally, Table S1 also showed that NIPAM-based nanogels exhibited $1 K higher VPTT values in D 2 O compared to in H 2 O.This difference in VPTT indicates increased affinity of nanogels toward D 2 O as compared to H 2 O, in good agreement with the known deuteration effect on the phase transition behaviour of NIPAM based microgels [70].DLS measurements at 312 K showed that a significant increase in size (Fig. S1), suggesting some degree of association in the bulk, due to increased hydrophobicity.Measurements of VPTT below 0.1 mg ml À1 could not be obtained experimentally, as the concentration is too low to detect change in transmittance using UV-Vis spectroscopy, however a plot of VPTT values vs ln[nanogel] displayed a linear trend that allowed an estimate (Fig. S2).The interfacial behaviour was first characterized using force tensiometry, which provided evidence of a clear variation in surface pressure as a function of temperature and concentration (Fig. S3).The surface pressure of nanogels (0.01 mg ml À1 ) increased with temperature from 27.9 mN m À1 at 294 K to 29.9 mN m À1 at 312 K, while a 3% increase in surface pressure was observed when the nanogel concentration was increased ten-fold to 0.10 mg ml À1 at room temperature.The morphology and behavior of these nanogels were consistent with previous data [71].

Behavior at the air-water interface
We first examined the interfacial behavior of 5 %MBA-NIPAM nanogels as a function of concentration and temperature by NR.The chosen concentrations (0.01, 0.05, and 0.10 mg ml À1 ) were consistent with our previous work, while the different temperatures (294, 300, 307, and 312 ± 0.5 K) were all below the corresponding VPTTs.
It is important to note the experimental NR data are averaged across the whole sample footprint ($30 cm 2 ).The data are modelled using a well-established optical matrix method.A model is first proposed, and the corresponding reflectivity profile is then calculated and compared with the experimental one.To obtain a unique model to represent the NR data, the model is constrained by using the contrast variation technique where a single model is required to simultaneously fit a series of contrast profiles.
A diffuse layer structure model is used to represent the interface given the soft nature and spherical shape of the nanogels [72].
The normalized NR profiles in subphases of D 2 O and NRW are shown in Fig. 1, plotted as RQ 4 as a function of the scattering wave-vector Q = 4psin h/k (where h is the angle of incidence and k is the wavelength) for clarity.The solid lines in Fig. 1(a)-(f) show the fits to the data.Both contrasts (D 2 O and NRW) are fitted simultaneously to a single model.The fitted parameters and the SLD for the nanogel are presented in Table S2 and S3, respectively).It is worth to mention that these nanogels display a slight increased affinity towards D 2 O as compared to H 2 O.In other words, the interfacial behaviour of the nanogel in D 2 O was not fully identical to that in H 2 O when the experimental temperature was approaching the VPTT.For example, the two contrasts could not perfectly be fitted simultaneously to a single model for 0.1 mg ml -1 at 312 K.However, two similar models only with a slight difference in the thickness or volume fraction of the second layer could adequately fit both data sets.As a result, fitting parameters at 312 K in Table S3 were average results for both H 2 O and D 2 O contrasts.The NR profiles did not change significantly between 294 and 307 K for all three concentrations; however, when the temperature was increased to 312 K, substantial changes in the NR profiles were observed at 0.05 mg ml À1 and 0.10 mg ml À1 .A set of welldefined Kiessig fringes (dips in the Q-range 0.03-0.07Å À1 ) were observed for both contrasts at 312 K, indicating the presence of layered structure at the air-water interface.The layer thickness increased when the nanogel concentration was doubled from 0.05 to 0.10 mg ml À1 , as evidenced by the shift in Q of the dips to lower values.Furthermore, an increased nanogel concentration resulted in a higher volume fraction of nanogels found in the layers, as demonstrated by enhancement of the fringe intensities.While we previously reported NR data for MBA-NIPAM nanogels at air-water interfaces with 10%, 20%, and 30% MBA as a function of concentration or temperature, [54][55][56] we did not observe any fringes in the NR profiles, indicating a less organized packing in the layers compared to the structurally more flexible 5% MBA nanogels used here.This suggests that the degree of flexibility of nanogels, controlled by the amount of crosslinking, together with the nanogel concentration and the proximity to VPTT all play a role in determining the interfacial packing towards the bulk solution.
The fitted scattering length density profiles were then used to calculate the volume fraction profiles perpendicular to the interface as shown in Fig. 2(a)-(c) as a function of temperature at the three concentrations.In the first monolayer region, the spread of nanogels by surface tension is opposed by the structural rigidity of the densely crosslinked core together with weak (van der Waals) attractive forces between the isopropyl groups [73].Hence, the nanogel structure at the air-water interface shown schematically in Fig. 3(a) is frequently referred to as a ''fried egg" structure [74].
At room temperature (294 K), the thickness of the first layer was similar at all three concentrations ($15 Å), with nanogel volume fractions of between 80% and 90%.As the temperature increased to 312 K, the volume fraction of nanogels in the first layer increased, with the highest value observed for 0.10 mg ml À1 .The second layer increased both in thickness and nanogel volume fraction, extending into the aqueous phase as the temperature of the solution approached the VPTT values (NR fitting parameters in Table S3).
As seen in the volume fraction profiles of the nanogels in Fig. 2  (a)-(c), the first two layers consisted mostly of nanogel.The combined thickness of the first two layers at the air-water interface at 0.01 mg ml À1 bulk concentration increased from 40 Å at 307 K to 52 Å at 312 K, with the adsorbed amount changing from 2.43 to 3.33 mg m À2 .When the nanogel concentration was raised tenfold to 0.1 mg ml À1 , both the thickness and adsorbed amount of the first two layers more or less doubled (51 to 87 Å and 3.08 to 7.45 mg m À2 , respectively) with temperature changes from 307 to 312 K. Interestingly, the adsorbed amount of nanogels at 0.01 mg ml À1 and 312 K (À4 K from VPTT; see Table S1 and Fig. S4) was 3.74 mg m À2 , comparable to the adsorbed amount of 3.53 mg m À2 obtained for the same material at 0.10 mg ml À1 and 307 K (À6 K from VPTT).
The adsorbed amount of nanogel at the layered interface is a small fraction of the total present in the bulk solution.In the concentration range studied here, the adsorbed amount of nanogel at the interface is primarily driven by experimental temperature (i.e., surface activity) and its proximity to its corresponding VPTT, with concentration appearing play a minor role.Nevertheless, the nanogel concentration has a specific influence on its interfacial structure, significantly influencing the adsorbed amount when the nanogel is closer to its VPTT (Fig. S4).This has significant implications, since the VPTT of nanogels can be tailored by the choice of functional monomers and degree of crosslinking, therefore offering an additional tool for controlling the interfacial behavior.

Behavior at the hydrophobic SiC8-water interface
To further explore the findings relating to the structure of nanogels at the air-water interface and the nature/role of the interface itself, we next conducted experiments at the solid-water interface using silicon (Si) substrate modified with C8 to create a highly hydrophobic surface.Nanogel adsorption at the SiC8-water interface was studied using a temperature controlled solid-liquid cell  ($2 ml).A Si surface was synthesized by chemisorption of octyltrichlorosilane (H-C8, the hydrophobic layer) following a well-established protocol [75].The H-C8 layer on the silicon substrate measured 14 ± 2 Å by NR, slightly thinner than the previously reported thickness for a C18 chain as expected [76].NR profiles were obtained at three nanogel concentrations (0.05, 0.10, and 0.50 mg ml À1 ), two temperatures (300 and 312 K), and for three contrasts (D 2 O, H 2 O, and CMSi (0.05 mg ml À1 only)).The NR profiles are shown in Fig. 4 and Fig. S5.To fit all three contrasts simultaneously to a single coherent model, a minimum of four layers was required.The fitted parameters are shown in Table S4, and a comparison of nanogel structures in the first two layers at both the air-water and solid-water interfaces is summarized in Table 1.
As evident in Fig. 4, the fringe (the dip in the Q-range 0.07-0.08Å À1 ) was clearly present in the NR profile for the D 2 O contrast at 312 K and 0.10 mg ml À1 , although less intense than those seen at the less constrained air-water interface.The NR profiles were not significantly different for 0.05 mg ml À1 nanogels when the temperature was changed from 300 K to 312 K for D 2 O, H 2 O, and CMSi water contrasts (Fig. S5).Interestingly, when the same experiment was carried out at a higher concentration (0.5 mg ml À1 ; Fig. S5) at 312 K, the fringes were significantly enhanced in raw NR profiles.However, the data could not be fitted simultaneously to a single coherent model for both contrasts, possibly due to the temperature being above the (311 K) at this concentration.The fitted scattering length density profiles were used to calculate the volume fraction profiles perpendicular to the SiC8-water interface.A comparison of the volume fraction profiles of nanogels at air-water and SiC8-water interfaces is presented in Fig. 5 (at 0.10 mg ml À1 ) and Fig. S6 (at 0.05 mg ml À1 ).Note that the SiC8 data show a $15 Å gap due to the presence of the grafted C8 layer on the solid substrate.
At 300 K, the first layer thickness, volume fraction, and adsorbed amount at both the air-water and SiC8-water interfaces were very similar.However, the second layer was clearly different: in the case of C8 resulting in a much denser second layer containing nearly three times more adsorbed nanogel compared to the air-water interface.This suggests that the presence of the hydrophobic chains in SiC8 provides a first layer conformation that templates the subsequent layer, with hydrophobic interactions driving the adsorption process.The difference in the packing of the layers at the two interfaces could also be further influenced by the very small nanogel particles intercalating the C8 chains, a phenomena that cannot be quantified by NR experiments with limited contrast variations.
Raising the temperature to 312 K, close to the VPTT, significantly altered the volume fraction profiles.In the case of the SiC8-water interface, the total adsorbed amount increased by 18% from 4.82 mg m À2 to 5.71 mg m À2 , although the total layer thickness did not change significantly.In contrast, self-assembled nanogels at the air-water interface changed from a three-to a five-layer structure.The total adsorbed amount more than tripled from 2.68 mg m À2 to 10.52 mg m À2 and the total layer thickness increased by $50 Å.As the temperature in the bulk approaches the VPTT, the structural feature of the first layer at the SiC8 water interface was identical to that at the free air-liquid interface.However, the second layer shows a very different profile, with air-water data showing a thicker second layer of over 70 Å with a 59% volume fraction and adsorbed amount of 5.52 mg m À2 .
The different behavior and interfacial conformation of the nanogels at 300 K and 312 K in response to the change of interface is significant (Fig. 3(a)-(b)).When considering the adsorbed structure of the nanogels, we hypothesized that the mobility of the first layer at the air-water interface as opposed to the interfacial confinement at the solid-liquid interface in addition to a higher degree of hydrophobicity provided by the C8 layer, results in a changed conformation, a more compact packing structure, and ultimately is the driving factor that controls the structure of the subsequent layers.Having identified hydrophobic interactions as key features driving nanogels' interfacial structure and conformation at air-water and SiC8-water interfaces, we next studied the interfacial behavior towards a hydrophilic, unmodified Si substrate.

Behavior at the hydrophilic Si-water interface
NR experiments were carried out at three temperatures (281, 300, and 312 K) with 0.1 mg ml À1 nanogel in the D 2 O contrast only at a hydrophilic Si-water interface, and the obtained profiles are presented in Fig. 6.Fits are shown with solid lines, and the NR profiles at 300 and 312 K are both shifted by a factor of 10 for clarity with the inset showing unshifted data.The fit parameters are given in Table S5.There was no evidence of nanogel adsorption onto the hydrophilic Si surface at 300 K, and there were no discernible changes at this hydrophilic surface as the temperature was further increased to 312 K. Therefore, even as the temperature increased closer to the VPTT, adsorption did not occur, as expected.However, NIPAM-based nanogels are known to be more hydrophilic at lower temperatures [77].As a control experiment, we repeated the experiment at 281 K. NR data showed evidence of adsorption at the lower 281 K now driven by hydrophilichydrophilic interactions between the nanogel and the bare hydrophilic Si surface (Fig. 3(c)).A two-layer model was used to fit the data, and the first layer was 8 ± 1 Å thick with a low water content of 70% (Table S5).Interestingly, this behavior bears some resemblance to the phenomena observed with the cold denaturation of proteins occurring at low temperatures [78].

Visualization by atomic force microscopy
In order to gain additional evidence of the interfacial behavior of nanogels with solid interfaces, i.e. to confirm that nanogels bind to a hydrophobic SiC8 surface than a hydrophilic bare Si surface, atomic force microscopy (AFM) was carried out in fluid.SiC8 and Si substrates were incubated in 0.1 mg ml À1 nanogel solutions for 20 min at 298 K before being directly imaged by AFM in the nanogel solution.
In the case of SiC8 (see Fig. 7(a) and Fig. S7(a)), AFM imaging provides evidence the nanogel adhered to the surface in a collapsed state.The nanogel layer was 21 ± 5 Å thick based on AFM height profile measurements, in agreement with the NR data of a 18 ± 1 Å thickness with a 4 Å layer roughness.Representative height profile measurements are shown in Fig. S8.The roughness values as measured by AFM (Table S6) were higher on treated SiC8 substrates (mean RMS roughness = 0.80 nm) compared to treated Si substrates (mean RMS roughness = 0.47 nm).The substrate was imaged in the presence of only water (Fig. S9) to confirm that it was indeed the nanogel seen by AFM on the SiC8 substrate, and no discernible nanogel structures were observed.In contrast to the SiC8 substrate with nanogel, the AFM images of the hydrophilic bare Si substrate incubated with nanogel did not show physiosorbed nanogel (Fig. 7(b) and Fig. S7(b)), suggesting that the material remained free in solution.This is in agreement with the NR data, confirming that the nanogel does not interact strongly with the hydrophilic surface.In addition, Fig. S10 shows adhesion and DMT (Young's) Modulus channels for Fig. 7 images, where darker regions indicate higher tip to sample adhesion and lower elasticity in each channel, respectively.These measurements suggest a clear difference in the physical properties of the adsorbed nanogel and the exposed SiC8 surface.In contrast, the hydrophilic substrates appear uniform in all channels.
Fig. 6.NR data (markers) and fitting (solid lines) of 5% crosslinked MBA-NIPAM nanogels at the hydrophilic Si-water interface at 0.10 mg ml À1 as a function of temperature (281, 300, and 312 K) in D 2 O contrast.For clarity, the profiles both at 300 and 312 K are both shifted up by factors of 10 and the inset shows the original unshifted data.

Conclusion
Neutron reflectivity allows to acquire specific knowledge of the structure and conformation of soft nanogels at interfaces from a bulk nanogel solution.We synthesized 5% MBA-NIPAM nanogels by high dilution radical polymerization and we used neutron reflectivity and surface tensiometry to study their structure and conformation at interfaces, providing an understanding of the driving forces for the adsorption and conformational changes.The first interesting result was the observation of fringes in the NR data, that were never observed in our previously work on nanogels with higher crosslinking content (10, 20 and 30%).This provided evidence that the rigidity of the nanogel matrix plays a key role in determining the interfacial behavior and structure, with the higher flexibility given by lower cross-linker content providing a more ordered structure and packing at the interface.
The NR data provide evidence that the adsorption of MBA-NIPAM nanogels at the air-water interface is primarily driven by temperature, more specifically its proximity to the corresponding VPTT.The nanogel concentration appears to play a minor role, influencing the adsorbed amount only when the experimental temperature is closer to the VPTT.This is an important finding that provides an additional tool for tailoring the interfacial behavior, given the ability to alter VPTT by changing the nanogel formulation, e.g.monomer structure and crosslinking content.
By comparing data obtained at air-water and solid (SiC8)-water interface we were able to show that the conformation of the first layer, driven by the strength of the hydrophobic interactions with the interface, is responsible for the nanogel packing.The mobility of the first layer at the air-water interface as opposed to the interfacial confinement at the solid-liquid interface, in addition to a higher degree of hydrophobicity provided by the C8 layer, results in a different conformation, a more compact packing structure, which ultimately drives the structure of the subsequent layers.This was particularly evident when the temperature was raised to 312 K and the adsorbed amount for the air-water interface increased considerably more than with the solid-water.
On a hydrophilic surface weak adsorption was observed only below room temperature (281 K) because of entropic changes and a weakening of hydrophobicity of the nanogels analogous to cold denaturation of proteins.Finally, nanogel adsorption inferred by NR on the hydrophobic SiC8 surface was confirmed and visualized by fluid AFM experiments, which showed in real space that nanogels adhered to the substrate.
These results provide important contributions to our understanding of the interfacial behavior of thermoresponsive NIPAMbased nanogels, with regards to their conformation and structure, and their changes in response to variations in temperature, concentrations, and the degree of hydrophilicity of the interface.This knowledge will aid the development of materials with properties tailored for specific applications but also provide a better understanding of the mechanisms of interactions with biological units such as membranes and proteins.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig.1.NR data (markers) and fitting (solid lines) of 5% crosslinked MBA-NIPAM nanogels at air-water interfaces in D 2 O and NRW contrasts: (a) and (d) 0.01 mg ml À1 , (b) and (e) 0.05 mg ml À1 , and (c) and (f) 0.10 mg ml À1 as a function of temperature.For clarity, the profiles are shifted up by factors of 10.

Fig. 3 .Fig. 4 .
Fig. 3. Schematic diagram of the adsorption of nanogels at different interfaces (a) air-water, (b) hydrophobic Si-water, and (c) hydrophilic Si-water interface as a function of temperature.

Table 1
Comparison of nanogel layer thickness, volume fraction and adsorbed amount for the first two layers at different interfaces are shown.The total adsorbed amount at the air-water interface is shown graphically in Fig.S4.U = volume fraction of the nanogel; C = adsorbed amount of nanogels.The nanogel concentration in the bulk was 0.10 mg ml À1 .Fig.5.Volume fraction comparisons of nanogels at the air-water and SiC8-water interfaces (0.10 mg ml À1 ).