The characterization of Pluronic P123 micelles in the presence of sunscreen agents

The triblock copolymer Pluronic® is widely used in the personal care industry, including sun protection, for its film‐forming and solubilization capabilities. In this study, the effect of three commonly used organic UV filters (ethylhexyl methoxycinnamate [EMC], ethylhexyl triazone [EHT], and avobenzone [AVB]) on the structure of Pluronic P123 micelles was investigated.


INTRODUCTION
Sun protection formulations, also commonly known as sunscreens, are widely used in the personal care industry by consumers to protect against harmful UV radiations (280-400 nm) [1]. Prolonged exposure to these harmful radiations may result in acute or chronic effects on the skin, such as; erythema (sunburn), premature skin ageing and skin cancer [2,3]. The rate of incidence of non-melanoma skin cancer has been on the rise in recent years [4]. For example, in the UK, over 100 000 new cases of nonmelanoma skin cancer are diagnosed every year [5]. It is therefore imperative that proper measures are in place to protect the skin against these harmful UV radiations. One of which, is the frequent and correct use of sunscreens.
Sunscreens comprise a range of active ingredients that act as UV filters, providing a balanced protection approach from both UVA and UVB radiations. Additionally, sunscreens also contain a wide range of inactive ingredients, such as sensory enhancers (silicones and powders), film formers (silicone-polyurethane based waxes, triblock copolymers), emollients (oils and esters) and emulsifiers (ionic or non-ionic).
Moreover, the presence of Pluronic® copolymers in many formulations including sunscreens in different concentrations affects the existence, nature and types of these aggregates. This is even further complicated in the presence of other ingredients which may result in competitive or synergistic interactions [14], affecting their critical micelle concentration (CMC) and eventually their aggregation behaviour. These changes would highly affect the feel of these formulations, and more importantly, the 'accessibility' of the UV filters in the formulation once they are applied on the skin. For example, Hanson et al. have investigated the effect of both the surfactant (emulsifier) concentration and type on the photostability of the UVA filter Avobenzone (AVB). Simple surfactants such as sodium dodecylsulphate (SDS), cetyl trimethylammonium bromide (CTAB), trimethylammonium bromide (TTAB) and sodium dodecylbenzenesulphonate (SDBS) were used by the authors to solubilize AVB in water [15]. They concluded that the samples containing SDS at concentrations higher than its critical micelle concentration (CMC) showed the best results in terms of AVB solubilization and photostability, followed by the TTAB, SDBS and CTAB. This stronger performance by the SDS in comparison to the other surfactants was attributed to the partitioning of the UV filter within the SDS micelle palisade, allowing UV filter-water interactions, essential for photostability. These observations strongly show that the surfactant structure and micelle shape/size play a significant role in the photostability of UV filters, which will most certainly influence the formulation's SPF. Furthermore, the performance of sunscreens is highly dependent on the surface tension during the film formation phase, their distribution on the skin and their rheological profile, all of which would be sensitive to the emulsifier(s)' structure present in a formulation [16,17]. The sensitivity of the SPF to the rheological properties of sunscreen lotions was also highlighted by Hanno et al. where a range of commercially available surfactants were used. It was noted that varying the surfactant had a significant impact on the rheological profile of the formulations, affecting the formed film thickness and consequently the SPF [16], echoing the findings from Liu et al. where the thickness of the film formed by sunscreen lotions has been shown to directly influence the SPF [18].
The solubilization of hydrophobic molecules such as drugs, dyes and oils, by Pluronic® has been investigated before. Several fragrances were also solubilized in Pluronic® F127 by Grillo et al. [19] where the structure of the F127 micelles was explored by differential scanning calorimetry (DSC), small-angle X-ray scattering (SAXS) and smallangle neutron scattering (SANS). The authors concluded that the insertion of the fragrances into the Pluronic® micelles has introduced variable but sharp changes to the micelles' morphology. Han et al. [8] investigated the interactions between antimetabolites and Pluronic® L62 micelles using SANS where changes in the micelle core hydration level in the presence of the antimetabolites had a significant impact on the micelle size and the aggregation number. Given the frequent use of Pluronic® in sun protection lotions and sprays, for example, F127 in Kiehl's Ultra Light Daily UV Defense SPF50+ lotion or F108 in La Roche-Posay's Anthelios Age Correct SPF50+ lotion, a better understanding of the effect of UV filters on the Pluronic® aggregates morphology is needed, something we believe has not been attempted yet.

Preparation of Pluronic-UV filters solutions
The different UVA and UVB filters were added to 5 wt% aqueous Pluronic P123 solution containing 10 wt% ethanol (deuterated form for the NMR and SANS measurements). The solutions were left in a HulaMixer for 24 h at room temperature.

Tensiometry
The surface tension measurements were performed on the samples using a SITA bubble tensiometer (pro line t15, Germany), at 25°C, calibrated by reference to deionized water. To ensure reproducibility, the measurements were repeated five times. The uncertainties in the data were estimated from the minimum and maximum surface tension values recorded.
Nuclear magnetic resonance (NMR) 1 H NMR spectra were recorded on a JEOL ECZ-600R (600 MHz) at 25°C, operating at 1 H frequency of 600 MHz and equipped with Royal™ probe 5 mm combined Broadband & Inverse probe and an ACS 64-position sample changer.

Small-angle neutron scattering (SANS)
The SANS experiments were performed on the fixedgeometry, time of flight ZOOM diffractometer (ISIS Spallation Neutron Source). The Q range explored on ZOOM was between 0.0025 and 0.5 Å −1 . The samples were contained in a 1 mm path length, UV-spectrophotometer grade and quartz cylindrical cuvettes (Hellma). The cuvettes were mounted in aluminium holders on top of an enclosed, computer-controlled, sample chamber. All measurements were performed at 25°C (±0.5°C). Temperature control was achieved using a thermostatted circulating bath pumping fluid through the base of the sample changer.
Experimental measuring times were ~30 min. All scattering data were normalized for the sample transmission and the incident wavelength distribution, corrected for instrumental and sample backgrounds using a quartz cell filled with D 2 O (this also removes the instrumental background arising from vacuum windows, etc.), and corrected for the linearity and efficiency of the detector response using the instrument-specific software package. The data were put onto an absolute scale using a well characterized partially deuterated polystyrene blend standard sample.

SANS data analysis
SANS data from Pluronic solutions were fitted using SASview v5.0.4 to a core-shell spherical model [22] or a core-shell cylindrical model [23]. The total volume of the spherical micelles was estimated from: where R B is the total radius of the micelle (core + shell), while the total volume of the cylindrical micelle was estimated from: where R B is the total radius of the cylindrical micelle (core + shell) and h is the overall length of the cylinder. The density number of the micelles could be obtained by dividing the volume fraction of the micelles, Φ m by the micellar volume (V sphere or V cylinder ): The aggregation number (N agg ) could be calculated by dividing the volume fraction (Φ P123 ) of the P123 sample prepared (5%) by the density number of the micelles, N p , and the molecular volume (V P123 ) of one P123 molecule:

Dynamic surface tension (DST)
The DST behaviour from a pure P123 sample at concentrations well above the CMC, Figure 1, shows a gradual decrease in the surface tension values from ≈40 mN.m −1 to 33 mN.m −1 as the bubble lifetime increases, reaching an equilibrium state at approximately 10 s. This is in very good agreement with previous investigations on P123 at different concentrations above the CMC [20,24].
The DST behaviour from surfactant solutions is sensitive to several factors, such as the nature of the surfactant, the chain length and more importantly here, the presence of any additives. In the pure P123 solution case, the trend in the DST data is consistent with the presence of a significant number of surfactant species that are available to quickly diffuse and adsorb at the surface of the bubble being introduced in the solution. Goswami et al. [25] discussed the complexity of the interface in the presence of different P123 concentrations, where the different arrangement of the triblock copolymer, in comparison to the non-ionic surfactant C 12 E 6 was found to poorly reflect on the P123 data modelling to the Ward and Tordai model.
As the UV filters are introduced, at relatively low concentrations, 0.25 wt%, we see some changes in the DST behaviour. For example, in the presence of 0.25 wt% EMC, an increase, albeit small, in the surface tension value is observed at the both 'young' and 'old' surface ages, Figure 1. This small increase could be attributed to the low concentrations of the UV filters in the system, especially when comparing it to the 1 wt% case presented later. This overall DST behaviour is indicative of a decrease in the kinetics of the distribution of monomers and micelles. This decrease is presumably attributed to the molecular interactions between the UV filters and the Pluronic molecule, namely the PPO core. The interactions between the hydrophobic salicylic acid molecule and the P123 hydrophobic PPO core have been investigated before by Shah et al. [26], where it was postulated that an increase in the lifetime of the salicylic acid solubilized micelles, led to the disruption of the water molecules around the PPO core. These factors might be expected to also significantly impact the micelle shape and size.
The introduction of the UV filters at higher concentrations, namely EMC and EHT at 1 wt%, has resulted in significant changes in the DST behaviour, namely the slope of the data and the stability of the newly formed interface. The surface tension values at a young surface age (30 ms) in the 1 wt% EHT case for example is ≈46 mN.m −1 (±0.15), while for the pure P123 sample is ≈41 mN.m −1 (±0.09). The lower surface activity exhibited by the P123 + EHT is indeed very interesting. One would expect that the presence of the large hydrophobic UV filter would not lower the surface activity, especially as the aggregation number increases evident by the SANS results (Section Smallangle neutron scattering (SANS)). However, the increase in the size and aggregation number could explain this decrease in surface activity as these larger structures would be slow in migrating to the newly formed interface in bulk [26,27]. By replotting the tensiometry results to show surface tension versus t −1/2 , inset in Figure 1, one is able to highlight the changes in the slopes of the data. By looking at the long times (t −1/2 < 1), the data are linear which is often related to a diffusion-controlled process. However, it should be noted that drawing further conclusions from this representation is limited by the large molecular size of the Pluronic P123, and how it could change its configuration at the air-water interface. The magnitude of these slopes has been previously linked to the surfactant's hydrophobicity in a given environment, that is, the steeper the slope, the higher the hydrophobicity [28,29], which in our case could be attributed to the variation in the size of the P123 micelles in the presence of the various UV filters.

Nuclear magnetic resonance (NMR)
Our NMR data further highlight the interactions between P123 and the UV filters (for reference), full-width NMR spectra from the UV filters and P123 have been included in the electronic supporting information, (Figures 1S-4S). The peaks from the P123 + EHT sample (red line, Figure 2) at ≈1.0 ppm from the PO-CH 3 group and at ≈3.5-3.2 ppm from the PO-CH 2 have shown a change in their position and width, pointing towards the presence of strong molecular interactions between the PPO groups and EHT. The upfield shift of the peaks from the mixture strongly points to the creation of a further hydrophobic environment at the PPO core. In comparison, the peak from the EO-CH 2 group at ≈3.6 ppm shows little change in its width and chemical shift, indicative of weaker interactions between EHT and this group where it remains in contact with water.
Generally, the NMR data also present these strong interactions through sharp changes in the aromatic peaks from the UV filters at the ≈8.0-6.0 ppm range. The example presented here, Figure 3, is from a sample comprising P123 and EMC. The spectra show that the intensity of the EMC aromatic peaks in the presence of the P123 micelles has become weaker and that the peaks have become broader. The effect of the EMC on the chemical shift of the P123 PO-CH 2 , PO-CH 3 and the EO-CH 2 chemical groups is also shown, Figure 5S. A short T 2 relaxation time is likely behind these changes in the peaks' resolution.
If one is to hypothesize, based on the spectra, that the EMC is interacting with both the PPO and the PEO groups, this would mean that a percentage of the EMC is present at the palisade of the micelle. As a result, this would cause some broadening in the NMR signal, as there will be a quick relaxation (in comparison to the time of the experiment) of the NMR resonance from the aromatic group present at this area as noted above. This broadening in the signal could be further highlighted by the NMR spectra from samples comprising EHT/EHT + P123 and AVB/AVB + P123, Figures 6S and 7S.

Small-angle neutron scattering (SANS)
In this section, we examine the effect of the three UV filters, at different concentrations, on the Pluronic P123 micelle structure. First, we investigate the effect of EMC at three different concentrations, 0.1, 0.5, and 1 wt%, Figure 4. SANS from Pluronic P123 micelles has been reported before [20,30], where in general, core-shell F I G U R E 2 1 H NMR spectra from P123 (black) and P123 + EHT (red) in 10 wt% d 6  spherical micelles were observed. Similar structures were present here, evident by a peak or a 'bump' in the data at ≈0.09 Å −1 (d ≈ 69 Å), corresponding to a well-defined coreshell structure.
Our data fitting to the core-shell model as discussed in Methods section, showed the presence of these aggregates with a core radius of ≈49 Å, shell thickness of ≈11 Å and an aggregation number of 50, Table 1, in good agreement with previously published work from P123 solutions containing ethanol [14,24]. It must be noted though that a limitation of this method estimating the aggregation number is that it does not consider the degree of hydration of the PEO in the micelle's shell and its effect on the thickness of the shell. The data fitting routine to the core-shell model however provides information to the sensitivity of both the radius of the core and the shell thickness. We estimate a relative error of ≈5% on the aggregation number as the accuracy on both terms is 1 Å, Tables 1 and 2.
The introduction of low concentrations of EMC, namely 0.1 wt%, has resulted in an increase in the volume of the core-shell micelle from 900 to 1200 nm 3 . This strongly points to the preferential solubilization of the EMC by the P123, also evident from the NMR results as reported earlier. A significant shift in the position of the bump at mid Q towards low Q is observed in the presence of 0.5 wt% EMC, where now an overall radius (core + shell) of ≈75 Å was obtained from the data modelling, in comparison to ≈60 Å from the pure micelles. This increase could also be extended to the micellar volume and the aggregation number calculated from this sample, Table 1. Spherical core-shell micelles were also observed in mixtures of P123 and the AVB molecule, Figure 2S, where similar conclusions to the P123 + EMC system (based on the scattering behaviour and fittings parameters, Table 1S) could be drawn. These similarities could be attributed to their molecular weight and more importantly, their log pvalues [1,31].
The most significant impact on the P123 micelle could be observed in the presence of 1 wt% EMC. The data feature a Q −2 slope at low Q, indicative of a cylindrical structure, in addition to a bump at ≈0.065 Å, suggesting it also F I G U R E 4 SANS from P123 micelles and P123 micelles + EMC at different concentrations, at 25°C in 10 wt% d 6 -ethanol and D 2 O. Data have been staggered for clarity. Solid lines correspond to fits to core-shell spheres (0.1 wt%, 0.5 wt% EMC) and core-shell cylinders (1 wt% EMC). is a well-defined core-shell cylinder. This is indeed also evident from our data modelling, where a core-shell cylinder model revealed a cylinder length of 800 Å and an aggregation number of 735. The large size of these cylinders has also resulted in a significant decrease in the number density micelles, Table 1. The modelling however does not seem to capture some of the data at low Q. This is most likely due to aggregation of the micelles at that particular length scale. The transition from spherical to cylindrical micelles is most likely attributed to the high concentration of the EMC rather than its presence only, as an increase in the spherical micelle size was also observed at lower EMC concentrations. The introduction of 1 wt% EMC to the micelle, and its interaction with both the PEO shell and the PPO core as evident from the NMR spectra, could mean that the EMC has resulted in a core and/or shell dehydration, prompting the micellar transition.

Q/Å -
Sphere to cylinder transition in Pluronic aggregates was previously reported by Shah et al. [26] where the presence of salicylic acid was found to cause core dehydration, resulting in the structural transition. Pluronic P85 spherical micelles transition to cylindrical ones was also reported as a result of shell dehydration by Parekh et al. [32] However, it is not yet clear which of the two dehydration processes is the main driving force behind this transition. This is also further complicated when the hydrophobic molecule is solubilized within the Pluronic micelle interacting with both PPO and PEO.
On the other hand, the introduction of the larger EHT molecule to the P123 micelles has resulted in changes in its aggregation and scattering behaviour, Figure 5. However, we must note that the degree of change does not correlate to the EHT molecular size and concentration. This is further clarified by comparing the effect of the EMC and the EHT on the P123 micelle. For example, at 0.5 wt% EHT, the SANS fitting routine, Table 2, reveals a spherical coreshell micelle with an overall radius of ≈72 Ǻ-largely similar to the P123 + EMC profiles. Furthermore, if the EHT has promoted the presence of more micelles rather than increase its size, one would expect a significant increase in the number density of micelles, as well as an increase in the I(Q) in the unstaggered data, Figures 9S and 10S, however, that is not the case here, where both terms remain very similar to the ones obtained from the P123 + EMC micelles.
The similarity in the aggregation behaviour between the P123/EMC, and P123 /EHT systems strongly suggests weaker interactions between the P123 micelles and the EHT. This could be explained by a preferential interaction between EHT and EO and PO in solution, in addition to the solubilization by the Pluronic micelle-the hydrotope effect. Nguyen-Kim et al. have investigated the interactions between carbamazepine, fenofibrate and a range of Pluronic®, where weaker interactions between Pluronic and carbamazepine were noted, evidenced by no significant changes in the micellar size. It was noted that in this case, the Pluronic® acted as a double agent, once as a micellar solubilizer, and also as a hydrotope [33].

CONCLUSIONS
By combining tensiometry, NMR and SANS, we were able to investigate the effect of three commonly used organic UV filters; AVB, EMC and EHT, at different concentrations, on the micellar properties of Pluronic® P123. First, the introduction of the UV filters to the P123 micelles has resulted in high dynamic surface tension values, strongly suggesting that larger aggregates, with reduced availability of free monomers, are present in bulk. NMR has allowed us to probe the molecular interactions between the UV filters and P123, where the strongest interactions were observed between the core PPO groups and the UV filters. This was followed by the shell PEO groups and the UV filters, suggesting that the UV filter is indeed present at both the core and the palisade of the micelle. Future work involving 2D NMR techniques such as PGSE-NMR would be beneficial as it would allow us to quantify the partitioning of the UV filters across the micelle. The findings from the SANS measurements could allow us to postulate that: (a) F I G U R E 5 SANS from P123 micelles and P123 micelles + EHT at different concentrations, at 25°C in 10 wt% d 6 -ethanol and D 2 O. Data have been staggered for clarity. Solid lines correspond to fits to core-shell spheres (0.1 wt%, 0.5 wt% EHT) and core-shell cylinders (1 wt% EHT). 5 wt% P123 5 wt% P123 + 0.1 wt% EHT 5 wt% P123 + 0.5 wt% EHT 5 wt% P123 + 1 wt% EHT spherical to cylindrical transitions are present in these systems, triggered by the UV filter and its concentration, (b) preferential solubilization of the UV filters in the P123 core and (c) the significant increase in the size of the hydrophobic UV filter molecule does not necessarily translate to an equally significant increase in the micellar size, as evident in the EHT case, where the hydrotope effect might be present. These results provide insights for formulators utilizing the many benefits of Pluronic® triblock copolymers in sun protection formulations, highlighting the sensitivity of these materials to the presence of UV filters, and the subsequent effects it could have on the latter's accessibility and performance. This could be evaluated in the future using ISO approved in vitro SPF measurements and skin permeation studies.