Solubilization of active ingredients of different polarity in Pluronic® micellar solutions – Correlations between solubilizate polarity and solubilization site

Journal of Colloid and Interface Science 477 (2016) 94–102

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science journal homepage: www.elsevier.com/locate/jcis

Solubilization of active ingredients of different polarity in PluronicÒ micellar solutions – Correlations between solubilizate polarity and solubilization site Viet Nguyen-Kim a,b, Sylvain Prévost a,c, Karsten Seidel b, Walter Maier b, Ann-Kathrin Marguerre b, Günter Oetter b, Tharwat Tadros d, Michael Gradzielski a,⇑ a

Stranski-Laboratorium für Physikalische und Theoretische Chemie, Institut für Chemie, Straße des 17. Juni 124, Sekr. TC7, Technische Universität Berlin, D-10623 Berlin, Germany BASF SE, Carl-Bosch-Straße 38, D-67056 Ludwigshafen am Rhein, Germany c Helmholtz-Zentrum-Berlin, Lise-Meitner-Campus, Hahn Meitner Platz 1, 14109 Berlin, Germany d 89 Nash Grove Lane, Wokingham, Bershire, UK b

g r a p h i c a l a b s t r a c t

a r t i c l e

i n f o

Article history: Received 17 February 2016 Revised 12 May 2016 Accepted 12 May 2016 Available online 19 May 2016 Keywords: Solubilization mechanisms PluronicsÒ Carbamazepine Fenofibrate Solubilization enhancement

a b s t r a c t The solubilization of two pharmaceutically active ingredients (AI) with significantly different water solubility, namely carbamazepine and fenofibrate (solubility of 150 ppm and 10 ppm, respectively), has been investigated using a series of PluronicsÒ (Poloxamers) containing different ethylene oxide and propylene oxide (EO/PO) units in the molecule. The results show largely enhanced solubilization of fenofibrate by PluronicÒ micelles that increases with the PPO chain length provided the temperature is above the critical micelle temperature (cmt). In contrast the more water-soluble carbamazepine only shows a moderate increase in solubilization upon addition of PluronicsÒ. Small angle neutron scattering (SANS) and pulsed field gradient (PFG) NMR experiments show that the solubilization of fenofibrate occurs in the core of the micelles, whereas carbamazepine shows no direct association with the micelles. These clearly different solubilization mechanisms for the two AIs were confirmed by Nuclear Overhauser Enhancement Spectroscopy (NOESY) experiments, which show that fenofibrate interacts only with the PPO core of the micelle, whereas carbamazepine interacts with both PPO and PEO similarly. Accordingly, the large enhancement of the solubilization of fenofibrate is related to the fact that it is solubilized within the PPO core of the PluronicÒ micelles, while the much more moderate increase of carbamazepine solubility is attributed to the change of solvent quality due to the presence of the amphiphilic copolymer and the interaction with the EO and PO units in solution. Ó 2016 Elsevier Inc. All rights reserved.

⇑ Corresponding author. E-mail address: [email protected] (M. Gradzielski). http://dx.doi.org/10.1016/j.jcis.2016.05.017 0021-9797/Ó 2016 Elsevier Inc. All rights reserved.

95

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

1. Introduction The bioavailability of sparingly soluble pharmaceutical active ingredient (AI) molecules is often enhanced by micellar solubilization due to the presence of surfactants [1–4]. This is an important aspect as a sufficiently high content in aqueous solution and the transport through membranes etc. has to be facilitated in order to have efficiently working pharmaceutical formulations. However, this depends largely on the type of solubilizate that will have specific and individual interactions with a given surfactant [5]. Accordingly, one of the major challenges in pharmaceutical formulations is to develop effective surfactant solubilizer systems [6,7]. So far there has been little fundamental work published relating the effect of the molecular architecture of the solubilizer on its ability to solubilize a given hydrophobic compound and on how it influences their specific interaction with the solubilizate. In the pharmaceutical industry, solubilization is most commonly examined by ‘trial and error’ using high-throughput screening. That method helps to determine a suitable solubilizer without the need of knowledge in solubilization processes. However, an understanding of solubilization mechanisms at a molecular level and the specific interaction with the surfactant would allow the design of effective solubilizers in a more systematic and targeted manner. Here one also has to consider that the relevant solubilizates may differ largely with respect to their water solubility, polarity, and molecular architecture, i.e., one may not expect a general answer for enhanced solubilization but concepts that are adapted to the particular properties of the solubilizate. Accordingly, the main objective of this work was to find a relationship between the surfactant structure and its solubilization capacity for different types of solubilizates. For this purpose we have used two different pharmaceutical AIs, namely fenofibrate and carbamazepine, which differ substantially both in their chemical structure and polarity, with carbamazepine being much more water-soluble than fenofibrate (150 ppm vs. 10 ppm). Several solubilization techniques and surfactants have been applied to solubilize carbamazepine [8–11] and fenofibrate [12–14] in order to enhance their bioavailability and to determine the behavior of these systems at different temperatures and salinities. Such experiments showed that carbamazepine becomes solubilized by the micelles of alkyltrimethylammonium and alkylsulfates in a similar fashion and the molar solubilization capacity increases with increasing alkyl chain length of the surfactant [15]. Similar finding have also been reported for the solubilization capacity of fenofibrate in SDS solution [16]. For the case of SDS also the effect of surfactant purity and electrolyte content on the solubilization capacity and the dissolution rate has been investigated [17]. In general, the dissolution rate of carbamazepine was found to be directly proportional to the SDS concentration [14,18]. The solubilization of carbamazepine could be enhanced in a targeted manner by variation of the chain length of amphiphilic block copolymers and at higher temperatures, whereas the addition of salt had no effect. For the case of nonionic surfactants it was recently found that carbamazepine is typically located within the palisade layer of such nonionic micelles, i.e., it prefers the interaction with the EO groups [10] and also nonionic microemulsions are suitable to improve the solubilization of carbamazepine [19]. Fenofibrate showed good incorporation into microemulsions with a distinct dependence on the hydrophobic chain composition [20]. Prevalent solubilizers are PluronicsÒ (Poloxamers) consisting of various ethylene oxide/propylene oxide EO/PO blocks, whose properties can be varied in a wide range via the EO/PO ratio and their chain lengths [21,22]. In addition, PluronicsÒ are very suitable for such solubilization studies as they are basically approved for formulations in the fields of pharmacy or agriculture. Extensive studies on PluronicÒ structures were carried out at different

temperatures, concentrations and pH values showing a large influence of the molecular architecture on the phase behavior [23–28]. PluronicsÒ are widely applied to form thermodynamically stable micellar systems in water that are able to solubilize insoluble AIs [29,30] and oils [31]. These solubilized hydrophobic compounds can also influence the PluronicÒ structure [32–34], e.g. by increasing the aggregation number, the micellar size and the fraction of polymer micellized. In drug delivery systems, PluronicsÒ are used to enhance bioavailability, metabolic stability and circulation time of the drug [35,36]. These properties also enable PluronicsÒ to act as a scavenger for cardiotoxic drugs [37]. It might also be noted that the presence of compounds which are well solubilized within the PPO core of Pluronic micelles (which typically applies to moderately polar oils) also have a tendency to stimulate the formation of micelles aggregates, so that they will be formed at much lower concentration or temperature [38–40]. We focused our investigation on the experimental determination of the solubilization capacity of various PluronicsÒ with fenofibrate and carbamazepine and to find correlations between the solubilization performance and the molecular structure of solubilizate and PluronicÒ. For a further structural understanding we performed a structural characterization of these systems by means of small angle neutron scattering (SANS) to determine the size and structure of the PluronicÒ micelles. Furthermore pulsed field gradient (PFG)-NMR and nuclear Overhauser effect (NOESY)-NMR were applied to determine the solubilization sites of the AIs. The results of such a comprehensive characterization then shall provide the basis for a thorough understanding of the details of the solubilization process that can be applied to develop more effective solubilization systems in the future. 2. Materials and methods 2.1. Materials 2.1.1. Active ingredients Two pharmaceutical compounds, carbamazepine and fenofibrate (both P99%), were used. Their structure is given in Fig. 1. Both AIs were supplied by Sigma-Aldrich. 2.1.2. Surfactants Different PluronicsÒ were supplied by BASF SE, Ludwigshafen Germany. They are listed in Table 1 which gives the trade name, molecular structure, hydrophilic-lipophilic balance (HLB)numbers and values for the critical micelle temperature (cmt) at different concentrations of some PluronicsÒ.

(a)

N O

NH2

(b)

O

O

O

Cl

O Fig. 1. (a) Carbamazepine and (b) fenofibrate.

96

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

Table 1 List of surfactant name, average molecular weight, average chemical formulas, HLB numbers, and cmt values.

a b c d

PluronicÒ

MW/ g/mola

Average formulab

HLBc

cmt/°Cd 1 wt%/2.5 wt%/ 5 wt%

PE PE PE PE PE PE PE PE PE

1750 2450 2900 8000 3650 4600 4950 5900 6500

EO5-PO19-EO5 EO5-PO30-EO5 EO13-PO30-EO13 EO79-PO30-EO79 EO8-PO47-EO8 EO21-PO47-EO21 EO16-PO56-EO16 EO25-PO56-EO25 EO37-PO56-EO37

12 7 15 29 6 14 9 13 15

–/–/– –/–/– 31.5/28.5/26.5 50.0/46.0/40.0 –/–/– –/–/– –/–/– –/–/– 19.5/–/–

4300 6200 6400 6800 9200 9400 10300 10400 10500

The The The The

molecular weights were provided by BASF SE. average formulas were designated by the given EO/PO ratios. hydrophilic-lipophilic-balance (HLB) values were provided by BASF SE. cmt values were measured by Alexandridis et al. [41].

PluronicÒ solutions were prepared in distilled water or D2O for SANS and NMR experiments. D2O was provided by Euriso-TopÒ at 99.9% isotopic purity.

2.2. Measurement of solubilization capacity The powder of each AI was equilibrated in the aqueous surfactant solution at a given concentration for 24 h at 25 °C. The unsolubilized AI was removed by filtration (Millipore PVDF syringe filter; 0.45 lm pore size) and the concentration of the active ingredient in the micellar system was determined after a ten to hundred-times dilution with ethanol using UV/Vis spectroscopy (Agilent Cary-60; dual beam mode 600–200 nm; one data point per nm; scan rate: 600 nm/min). A calibration curve had been established by measuring different concentrations of the AI dissolved in ethanol.

2.4. Pulsed field gradient (PFG)-NMR PFG-NMR experiments were performed using a Bruker Avance III spectrometer operating at 400 MHz 1H Larmor frequency, 60A gradient amplifier, and micro-5 probe with diff-60 gradient coil, at 25 °C. A gradient stimulated-echo sequence [44] was used with the following parameters: Diffusion time (D) 20 ms, pulsed-field gradient duration (d) 1 ms, gradient strength (g) up to 2000 G/ cm, in 16 linearly spaced steps, up to 256 scans per gradient step, 4 dummy scans, 2 s effective recycle delay, 800 ms acquisition of the free induction decay, 5 kHz spectral width, and spectral Fourier transform processing with 5 Hz exponential line broadening. The gradient-dependent signal decays were analyzed by fitting mono- or bi-exponential functions for selected integral regions with signal intensity I, relative amplitudes a1 and a2, Diffusion coefficients D1 and D2, and the gradient-dependent suppression parameter B = g2c2d2(D
(d/3)), wherein c is the gyromagnetic ratio of 1H, according to the equation IðBÞ ¼ a1 e
D1 B þ a2 e
D2 B . 2.5. Nuclear Overhauser Enhancement Spectroscopy (NOESY) The NOESY experiments were performed on a Varian VNMRS 500 MHz at 25 °C. The samples were prepared in a mixture of D2O and H2O at a volume ratio of 9:1. Zero quantum interferences were filtered and the spectra were recorded with 24 scans per increment with altogether 250 increments. The mixing time was 200 ms. The software MNova von Mestrelab was used in these experiments. 3. Results and discussion In our experiments we systematically varied the molecular architecture of the PluronicsÒ and studied its effect on the solubilization of fenofibrate and carbamazepine into the aqueous solution. For that purpose we investigated a concentration range of 1–5 wt% PluronicÒ and characterized those systems comprehensively, using the techniques described in the introduction.

2.3. Small angle neutron scattering (SANS) 3.1. Solubilizer screening The results for the solubilization capacities of the various PluronicsÒ for the two different AIs are summarized in Fig. 2a. Fig. 2b shows the solubilization enhancement with various PluronicsÒ relative to the water solubility of the respective AI. These results show that the structure of the solubilizer has a much more pronounced influence on the solubilization of fenofibrate than on

2500

solubilization / ppm

Small angle neutron scattering (SANS) experiments were carried-out on the V4 instrument at the Helmholtz-Zentrum Berlin (Germany) in order to obtain size and shape of the PluronicÒ micelles. Samples were prepared in D2O, and the scattering was measured in quartz cuvettes of 1 mm neutron pathway. Measurements were carried out at 22.0 °C using an average neutron wavelength of 0.457 nm (FWHM: 10%) for three configurations. The sample-to-detector distances were 0.84, 3.83, and 14.83 m with respective collimation lengths of 2, 8 and 16 m, thus covering a q-range of 0.06–7 nm
1; q being the magnitude of the scattering vector (q = (4p/k) ⁄ sin(h/2); h being the scattering angle, and k the wavelength); detector offset and wavelength were confirmed by the scattering from silver behenate. The solid angle variation (for 0.84 and 3.83 m S-D distance) and the efficiency of the detector elements were accounted for using a 1 mm H2O sample as an homogeneous scatterer. Data reduction was performed with BerSANS [42], taking into account the electronic background with the data obtained with cadmium at the sample position, sample’s transmission and thickness, subtracting the scattering of D2O (i.e., the incoherent scattering arising from the dissolved compounds is still present in the experimental data, and thereafter a constant background is subtracted for data visualization). Subsequently the data were scaled in absolute units with a 1 mm H2O sample as a secondary standard using water levels calibrated on this instrument. Fits of the SANS data were done using the SASFIT program package [43].

2000

Carbamazepine Fenoibrate

1500 1000

500 0

Fig. 2a. Solubilization capacity for fenofibrate and carbamazepine with various PluronicsÒ (5 wt% at 25 °C; errors for the solubilization are estimated to be <5%).

97

1000

Carbamazepine Fenoibrate

100

10

1

Fig. 2b. Solubilization enhancement for fenofibrate and carbamazepine with various PluronicsÒ (5 wt% at 25 °C; errors for the relative solubilization are estimated to be <5%) in relation to the water solubility of the AIs.

carbamazepine, differing between almost zero and 2000 ppm. In particular, PluronicÒ PE 9200, PE 10300, PE 10400 and PE 10500 with long PO units (50, 60, 61 and 56 respectively; see also Table 1) show very good solubilization for fenofibrate. In contrast, PluronicÒ PE 9400 which also contains a long PO chain (48) but in addition a relatively long EO chain, does not show much enhanced solubilization when compared with PluronicÒ PE 9200 with the same number of PO units. The main reason for this difference can be related to the effect of the molecular structure on the cmt. The results obtained by Alexandridis et al. [41] showed a large dependence of the cmt on PPO block length and size of the PEO headgroup. With PluronicÒ PE 9200 at 5 wt% concentration the cmt is well below 25 °C and hence micellar solubilization is well efficient in this case. In contrast with PluronicÒ PE 9400 that contains a total of 42 EO units and has an HLB-number of 14 the cmt is somewhat above 25 °C and hence there are almost no micelles at the temperature of measurement resulting in much lower solubilization in this case, which confirms that enhanced solubilization here has to be attributed to micelle formation. In general, PluronicsÒ with a shorter PPO-block show much less solubilization. This can be attributed to the higher cmts (Table 1) [41], which explains why PE 6400 and PE 6800 show almost no enhanced solubilization for fenofibrate. In contrast, the enhancement of solubilization for the more polar carbamazepine is very similar for the different PluronicsÒ. It is barely affected by their structure and aggregation behavior, although it may be noted that PE 9200, PE 10300, PE 10400 and PE 10500 are slightly better solubilizers than the lower molecular weight PluronicsÒ. It might be added here that such a rather moderate increase of carbamazepine solubility has been observed before in the presence of different PluronicsÒ [8]. It may be assumed for this case that the carbamazepine molecules are either only weakly bound to the surface of the PEO corona of the micelle [45] or also not at all specifically interacting with the micelles. An indirect evidence for this latter mechanism was obtained by investigating the solubilization behavior of a 5 wt% solution of PEO homopolymer with a molecular weight of 1000 g/mol. This addition of homopolymer also increased the solubilization of carbamazepine to 300 ppm (twice its solubility of 150 ppm), where it might be noted that a concentration dependent increase of solubility of carbamazepine due to the presence of PEG-400 in aqueous solution has been reported before [46]. However, at the same time the presence of PEO did not have any effect on the solubilization of

fenofibrate, confirming for this case that solubilization into PluronicÒ micelles must be taking place. Obviously, for carbamazepine already the very presence of a soluble polymer in aqueous solution is able to enhance solubilization but not much depending on the PluronicÒ structure. Further insights into the solubilization mechanism can be gained by studying its concentration dependence. Fig. 3 shows the solubilization of fenofibrate at different PluronicÒ concentrations of 1.0, 2.5, and 5.0 wt%. For all the different PluronicsÒ employed (PE 9200, PE 10300, PE 10400, PE 10500; i.e., the most interesting PluronicsÒ according to the results shown in Fig. 2) a linear dependence of the solubilization capacity on PluronicÒ concentration is observed. This linear dependence applies also to the solubilization of carbamazepine (Fig. 4) but one observes a much lower slope. For fenofibrate an extrapolation of the linear curves (Fig. 3) predicts the necessity of a minimum concentration of solubilizer in order to observe solubilization, which is summarized in Table 2. This can be attributed to the fact that solubilization requires the presence of micellar aggregates, which is only the case above the critical micelle concentration (cmc) of the respective PluronicÒ. The values allow to extrapolate the cmc and they are summarized in Table 2. Here they are also compared to values of previous publications [41,47], where it should be noted that the cmc of PluronicsÒ depends strongly on temperature as an increase by only 5 °C causes a large decrease of the cmc by one decade. Considering this pronounced temperature dependence of the cmcs our estimates are in quite good agreement with the previously determined cmc values, except for the PE 10300 where a rather large discrepancy is seen and where we observe a surprisingly large value. It can also be concluded that a much enhanced solubilization for fenofibrate is obviously connected to the presence of micelles in solution. For carbamazepine, extrapolation of the results to zero PluronicÒ concentration gives an estimate of the water solubility of the AI which is approximately 170 ppm (Fig. 4). These values (Table S1) are in good agreement with the solubility value of 150 ppm in pure aqueous solution. The presence of the PluronicsÒ causes an enhanced solubility of carbamazepine but not necessarily within the micelle, as apparently the enhancement is simply proportional to amount of added copolymer, irrespective of its ability to form micelles or not. The linear increase of the amount of carbamazepine solubilized with increasing PluronicÒ concentration can be explained by considering the solubility of carbamazepine in aqueous solution from a thermodynamic point of view, with xsol being the mole fraction of the maximum solubility of carbamazepine that is described by:

 0  l
l1 xsol ¼ exp s kT 2500

solubilization / ppm

factor of solubilization enhancement

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

2000 1500

ð1Þ

Pluronic PE 9200

Pluronic PE 10300 Pluronic PE 10400 Pluronic PE 10500

1000 500 0 0%

1%

2%

3%

4%

5%

6%

Pluronic concentration / wt% Fig. 3. Solubilization capacity for fenofibrate with PluronicsÒ at various concentrations.

98

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

800 700

PE 9400

Pluronic PE 9200

PE 6200

PE 10300

600

Pluronic PE 10300

PE 6400

PE 10400

Pluronic PE 10400

PE 6800

500

PE 10500

Pluronic PE 10500

PE 9200

PEO 1000

5

400

4.5

300

4

200

nsol / npol

solubilization / ppm

PE 4300

Pluronic PE 4300

100 0 0%

1%

2%

3%

4%

5%

Pluronic concentration / wt%

3.5 3 2.5 2

Fig. 4. Solubilization capacity of carbamazepine with different PluronicsÒ in aqueous solution at various concentrations.

1.5 1

Table 2 Effective cmc values extrapolated from the solubilization data for fenofibrate for different PluronicsÒ in comparison to ones determined in previous work by Kabanov et al. [47]and Alexandridis et al. [41]. The error of our extrapolation of the cmc is 10– 20%.

PluronicÒ 9200 PluronicÒ 10300 PluronicÒ 10400 PluronicÒ 10500

cmc from extrapolation/ wt% (25 °C)

cmc from Alexandridis et al. [41] (20 °C/25 °C)

PE

0.39

–

PE

0.95

0.7/0.07

PE

0.31

2.0/0.3

PE

0.59



1 1 þ A  xpol

2.2/0.3

 ð2Þ

Putting Eq. (2) into Eq. (1) one arrives at Eq. (3), with x0sol being the mole fraction of AI soluble in pure water and xsol being the amount of solubilized carbamazepine in the presence of polymer. Hence Eq. (3) describes the case of having an otherwise not interfering molecule dissolved in parallel in the solvent, composed of water and polymer.

xsol  x0sol ð1 þ A  xpol Þ

0.005

0.01

0.015

0.02

0.025

xpol Fig. 5. Solubilization of carbamazepine as a function of the amount of PluronicÒ monomer units for different PluronicÒ. In addition, one data point for a pure 5 wt% PE 1000 solution is added.

Table 3 A-values for different PluronicsÒ (number of water molecules affected per monomerunit assuming no interaction to carbamazepine).

where l0s is the standard chemical potential of the solid AI and l1 the chemical potential of the dissolved AI in polymer containing aqueous solution. l1 can be constructed from its value in pure water l01 and a term that in a first approximation depends logarithmically on the mole fraction of monomeric PO and EO units of the polymer xpol , which takes into account the lowering of the chemical potential of the solvent due to the presence of the polymer (here A is a factor that accounts for how many water molecules are affected effectively per polymer monomer unit):

l1 ¼ l01 þ kT ln

0

ð3Þ

In agreement with the experimental finding it predicts a linear increase of the solubility with the amount of polymer added (Eq. (3)). By plotting the factor of solubilization enhancement nsol =n0sol ð¼ xsol =x0sol Þ against xpol (Fig. 5) the values for A can be determined (Table 3). Those values are quite similar for the non-micelle forming PluronicsÒ PE 4300, PE 6200, PE 6400 and PE 9400, and they are only somewhat larger than the value for pure PEO. The block lengths of those PluronicsÒ thus play basically no role for the solubilization capacity, indicating a higher solubility of carbamazepine due to the lowering of the chemical potential of the

PluronicÒ type

A-values

PE PE PE PE PE PE PE PE PE

110 116 107 84 200 116 167 197 165

4300 6200 6400 6800 9200 9400 10300 10400 10500

solvent, which is simply induced by the interaction of PO or EO units with water (and specific interactions between the EO and PO units with the carbamazepine). Interesting in that context is mostly that this analysis indicates that about 100 water molecules would accordingly be affected by the presence of one EO/PO unit. Clearly visible in Fig. 5 is further that the group of PluronicsÒ forming micelles is having almost twice as high A values. The additionally enhanced solubilization here cannot solely be attributed to the lower chemical potential of water arising from the presence of dissolved polymer. It has to be due to additional solubilization of the solute within micelles. Also the scenario of having about 100 water molecules affected by the presence of one EO/PO unit appears to be somewhat unrealistic. The pronounced effect of enhanced solubility by the addition of homopolymer, together with the high A values, leads one to conclude that this general increase in the presence of PEO/PPO polymer is due to a specific interaction between carbamazepine and the PEO and PPO units of the unimeric polymers (and also the EO units of micelles). Such a picture is further corroborated by the NMR experiment of 3.4 that show a pronounced interaction between carbamazepine and the EO and PO units. 3.2. SANS In order to determine the micellar structures present we performed SANS experiments. Here we concentrated on PluronicÒ PE

99

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

10400, as that had been shown to be very efficient in drug solubilization (Fig. 2). PluronicÒ PE 9200 with an even better solubilization capacity was not further characterized due to its formation of large polydisperse aggregates instead of core shell micelles (Fig. S5), which remain stable for only several days and subsequently phase separate. For different concentrations PluronicÒ PE 10400 shows scattering patterns (Fig. S3) characteristic for spherical micelles, and as they are typically found for PluronicsÒ in aqueous solution [48–50]. The pattern does not change in the higher q-range as a function of concentration, thereby indicating that the micellar structure of PE 10400 does not change with concentration. One only finds the development of a correlation peak at 0.2–0.3 nm
1 that arises for increasing concentration from the steric interaction of the micelles. In Fig. 6 the scattering of pure 5 wt% PE 10400 micelle is compared to AI loaded micelles. One finds that fenofibrate leads to a clear increase of the scattering intensity at low q, thereby indicating a growth of the aggregates, while the presence of carbamazepine has basically no effect. The scattering patterns were modeled by means of a model of spherical core-shell micelles, described by a form factor Pcs,pd(q), interacting via a hard-sphere potential, described by a structure factor SHS(q) [51] and a contribution from chain scattering (non micellized Gaussian coils), described by Icoils(q). With such a model the scattering pattern is given as (for details see Supplementary data):

IðqÞ ¼ 1 N  Pcs; pdðqÞ  SHS ðqÞ þ Icoils ðqÞ þ Iinc

ð4Þ

where 1N is the number density of the micelles and Iinc the incoherently scattered background. Fitting our experimental data with such a model yields as relevant fit parameters the micellar aggregation number Nagg and the fraction of micellized polymer. All results are summarized in Table 4. The hard sphere radius of blank and carbamazepine loaded PluronicÒ PE 10400 micelles is 9.0 nm whereas for fenofibrate loading a small increase to 9.4 nm is observed. At the same time the mean aggregation number of Pluronic PE 10400 increases from 101 to 120 upon solubilization of fenofibrate. In contrast with carbamazepine there is basically no change in the aggregation number, i.e., it does not become incorporated into the micelles to any significant extent.

Table 4 Radii and aggregation numbers of Pluronic PE 10400 micelles at a concentration of 5 wt% in D2O as deduced from the SANS analysis. Sample

Radius/nm

Mean aggregation number

PluronicÒ PE 10400 +Fenofibrate +Carbamazepine

9.0 9.4 9.0

101 120 98

Previous SANS experiments analyzing the solubilization of hydrophobic compounds in aqueous PluronicÒ solutions have shown that the presence of hydrophobic substrates enhances the tendency for aggregation, thereby leading to the formation of larger micellar aggregates and also shifting the equilibrium with the unimeric polymer towards the micellar side [39,52–54] for the case of solubilization in the micellar core. Similar results were deduced from NMR and light scattering investigations [55]. This effect of solubilizate enhanced aggregation is rather little pronounced here due to the still relatively small solubilization of the fenofibrate (0.007 mol fenofibrate per mole PO, corresponding to about 40 fenofibrate molecules per aggregate) and correspondingly the rather small amounts solubilized can also not lead to a larger change of the hydration conditions within the PO units in the core. It might be note that the SANS experiments were done at 3 K lower temperature than the solubilization experiments and in D2O instead of H2O (which effectively also corresponds to a lowering of temperature). However, the main conclusion that the fenofibrate leads to an increase of the aggregates, while the carbamazepine does not affect the aggregates, should not depend on these differences. It might also be noted that our SANS analysis yielded a substantial increase of the PE 10400 present in unimeric form, increasing from 0.53 wt% at 1 wt% total concentration to 1.74 wt% for 5 wt% concentration (Table S2), where the value for 5 wt% total concentration is in good agreement with the unimer concentration of 1.5 wt% deduced from the PFG-NMR experiment (Table 5). This increase can be attributed to the repulsive interactions that prevail between the micelles and lead to an increase of the chemical potential of the Pluronic molecules, which then in turn leads to a higher unimer solubility. 3.3. Pulsed field gradient (PFG)-NMR The PFG-NMR method allows one to obtain the self-diffusion coefficients of the different components within an aggregate [44]. Therefore we used this method to determine the localization of the AI to be within or outside of Pluronic micelles. Fig. 7 shows the normalized NMR signal intensity as a function of the gradient suppression parameter for PluronicÒ PE 10400 with solubilized fenofibrate or carbamazepine. From these results the selfdiffusion coefficients of the poloxamer molecules and those of the AIs could be deduced separately. For the PluronicÒ molecules the data were fitted to a bi-exponential function (IðBÞ ¼ a1 e
D1 B þ a2 e
D2 B ) giving a fast relaxation process that arises from the surfactant monomer and a slower one stemming from the micellar aggregates. The slope of the line gives the diffusion coefficients D1 and D2 of the AI and the micelle. The values for a1 and a2 Table 5 Diffusion coefficients and hydrodynamic radii of Pluronic PE 10400 and active ingredients in aqueous 5 wt% polymer-solution. D1/m2/s

Fig. 6. SANS spectra of unloaded PluronicÒ PE 10400 micelles (5 wt%) in comparison with AI loaded micelles.

Pluronic PE 10400 Fenofibrate Carbamazepine


10

1.2 e – 2.1 e
10

Rh1/nm

a1/%

D2/m2/s

Rh2/nm

a2/%

1.98 – 1.13

30 – 100

2.0 e
11 1.5 e
11 –

11.9 15.8 100

70 100 –

NMR signal intensity (normalized)

100

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

1.0E+00

PE 10400 (Carbamazepine) Carbamazepin PE 10400 (Fenoibrate)

Fenoibrat

1.0E-01

1.0E-02

1.0E-03 0.0E+00

1.0E+11

2.0E+11

3.0E+11

Gradient suppression parameter B (m2/s) Fig. 7. Spin-echo signal of fenofibrate and carbamazepine in the presence of PluronicÒ PE 10400 compared to pure PluronicÒ PE 10400 for 5 wt% solutions of PE 10400 in D2O.

give information about the amount of components with D1 respectively D2 (Table 5). By fitting the signal loss of PluronicÒ PE 10400 to the biexponential function two different diffusion coefficients were determined with D1 being about one decade higher than D2. Hence D1 can obviously be referred to free PluronicÒ unimers accounting for 30% of PE 10400. D2 represents the diffusion coefficient of PluronicÒ micelles, which are formed by 70% of PE 10400 in the aqueous phase. From the diffusion coefficients one can deduce the hydrodynamic radii via the Stokes-Einstein relation and these values are also included in Table 5. The larger value Rh2 is in good agreement with the micellar size seen by SANS, just a bit larger than the hard sphere radius, which is expected as here the full extension of the PEO chains will be seen. The smaller value Rh1 then gives the hydrodynamic radius of the single polymer chains of PE10400. For fenofibrate one obtains one diffusion coefficient in the same order of D2 of PE 10400, implying that fenofibrate is located within the PluronicÒ micelles. Actually the fenofibrate diffuses even somewhat more slowly than the pure PE 10400 micelles, which can be explained by the fact that the fenofibrate loaded micelles are somewhat larger (as seen by SANS). In contrast, the selfdiffusion coefficient of carbamazepine is one order of magnitude higher than that of the micelle or fenofibrate. This indicates that carbamazepine diffuses freely in solution and hence it is not solubilized in the core of the micelle. Therefore, the PFG-NMR results give a clear indication of the very different localization of the two different AIs, fenofibrate and carbamazepine, within the micellar solution. Fenofibrate is apparently solubilized into the micelles, while carbamazepine diffusion is not bound to that of the micelles. It might be noted that this does not contradict with the observation derived from Fig. 5 that a certain percentage of the carbamazepine is associated with the micelles as due to its high affinity to water one can expect a rapid exchange and on the NMR time scale one sees the average diffusion, which then is dominated by the faster movement of free carbamazepine in the aqueous phase.

Fig. 8. NOESY spectrum of fenofibrate in a 5 wt% solution of PluronicÒ PE 10400 in D2O. The colors represent the signal intensity. a, b, c being proton signals of the fenofibrate. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The NOESY spectrum of an aqueous PluronicÒ PE 10400 (5 wt%) solution with solubilized fenofibrate is shown in Fig. 8. The resonance signal of the protons of the methylene group (CH2) of PEO is at d  3.82 ppm. The signal of the protons of the methyl group (CH3) of PPO is located at d  1.25 ppm, the methylene group is at d  3.30 ppm and the signal for the methine (CH) group is at d  3.60 ppm. The NOESY spectrum shows coupling of fenofibrate protons to PPO protons, but does not reveal any spatial proximity of fenofibrate to PEO. As the PluronicÒ micelles are known to possess a core (PPO)-shell(PEO) morphology [23,29,35], these results show that fenofibrate is exclusively located within the hydrophobic PPO core. The NOESY spectrum of an aqueous PluronicÒ PE 10400 solution with solubilized carbamazepine is shown in Fig. 9. In contrast to the fenofibrate case this spectrum shows coupling of the protons of carbamazepine to both PPO and PEO protons, which confirms that there is no specific incorporation into micellar aggregates occurring. Considering the previously shown results one may safely assume that carbamazepine is molecularly dissolved and thereby in general it is statistically located in the vicinity of the

3.4. NOESY Further information about the localization of the drug molecules within a given aggregate or its extent of association within such an aggregate can be deduced from NOESY-NMR experiments. These reveal local proximity of the drug molecule to specific groups within a micellar aggregate, in our case in particular, the proximity to PEO and/or PPO groups.

Fig. 9. NOESY-spectrum of carbamazepine in a 5 wt% solution of PluronicÒ PE 10400 in D2O. The NOE signals in the antiphase diagonal are colored red and the COSY signals blue. a, b, c, d being proton signals of the carbamazepine. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

PO or EO groups of free PluronicÒ unimers within the water phase. That there is also an appreciable interaction with the PO further corroborates our picture that carbamazepine is largely contained within the aqueous phase but also to some extent solubilized within the micelles (cf. data in Fig. 5). 4. Conclusions In this study we did a systematic comparison of the solubilization of two AIs of different polarity, carbamazepine and fenofibrate, in which we employed differently built PluronicsÒ (PEO-PPO-PEO triblock copolymers). For the less polar fenofibrate a substantial increase of solubilization is observed using PluronicsÒ with sufficiently long PPO blocks as those have a cmt lower than 25 °C, while shorter PPO blocks show an increasingly higher cmt and therefore less tendency for micelle formation and solubilization. In contrast, for the more polar carbamazepine only a rather small enhancement of solubilization is observed that somewhat depends on the type of surfactant employed. This could be explained such that carbamazepine solubilization is mostly affected by the lowering of the chemical potential of the solvent (as one observes the same effect upon the addition of PEO homopolymer), which is due to the presence of the polymer and presumably by specific interactions between carbamazepine and EO and PO in solution. Such a pronounced interaction may be ascribed to the strong dipoles of the EO and PO units that interact synergetically with the carbamazepine molecule. However, there is also some additional enhancement of solubilization for PluronicsÒ forming micelles, which indicates that about the same amount of carbamazepine that is additionally contained in solution due to the changed chemical potential is also solubilized by incorporation within the micelles. This is a bit of a change of conventional approaches to drug solubilization, as here apparently the PluronicÒ has a double function, working once as a hydrotrope and then also delivering some micellar solubilization, which, however, is only of lower significance. In general, one observes that the solubility of the AIs increases in a fashion directly proportional to the amount of dissolved polymer, but much more pronouncedly for fenofibrate. For fenofibrate longer-chain PluronicsÒ show a significantly better solubilization than their lower molecular weight counterparts, while for carbamazepine only a small dependence on the PluronicÒ architecture is observed. By combining SANS, PFG-NMR and NOESY experiments we achieved a detailed insight into the different solubilization mechanisms of fenofibrate and carbamazepine which we detailed for the case of PluronicÒ PE 10400. These experiments yield a consistent picture, in which fenofibrate is exclusively solubilized within the PPO core leading to a corresponding enhancement of the micelle size (as seen by SANS and PFG-NMR). However, this effect is rather small, in agreement with previous observations on size effects of drug solubilization in Pluronic micelles [56], that also showed only rather small size effects. In contrast, carbamazepine does not show such a pronounced interaction with the PluronicÒ micelles and diffuses effectively almost as free molecules. This is in agreement with the observed strong dependence of the solubilization capacity of fenofibrate on the PluronicÒ structure, whereas the solubilization of carbamazepine is increasing to a much lesser extent. PFGNMR experiments show no separate micellar diffusion of carbamazepine, as apparently it exchanges quickly within the solution between bound and free state, and accordingly diffuses rapidly. Similarly the NOESY experiments show exclusive interactions of the fenofibrate with the PPO core, while carbamazepine interacts with PEO and PPO in a similar fashion. Those findings regarding solubilization site and mechanism for differently polar AIs are of significant relevance when choosing a

101

suitable PluronicÒ system for a given type of AI to be solubilized by considering the polarity of the AIs in order to achieve optimum solubilization. Obviously the polarity and molecular structure of the solubilized drug molecule has a large effect on its interaction with the Pluronic molecules and thereby on the solubilization mechanism. They also show that seemingly similar effects of solubilization enhancement may have to be attributed to quite different molecular mechanisms of solubilization, either by changing the chemical potential of the solvent or by specific incorporation within the micellar aggregates. This is then not only important for controlling the solubilization capacity but also relevant for the availability of an AI in a given formulation. All these points therefore have to be considered accordingly when optimizing such formulations for solubilization. Future research then should extend such explanation and reasoning for drug solubilization in amphiphilic systems. Acknowledgments The beam time at HZB became available due to a long-term cooperation contract between the TU Berlin and HZB. Realization of this work was enabled by BASF SE. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2016.05.017. References [1] A. Prokop, J.M. Davidson, Nanovehicular intracellular delivery systems, J. Pharm. Sci. 97 (2008) 3518–3590. [2] G.S. Kwon, K. Kataoka, Block-copolymer micelles as long-circulating drug vehicles, Adv. Drug Del. Rev. 16 (1995) 295–309. [3] V.P. Torchilin, Structure and design of polymeric surfactant-based drug delivery systems, J. Control. Release 73 (2001) 137–172. [4] C.O. Rangel-Yagui, A. Pessoa, L.C. Tavares, Micellar solubilization of drugs, J. Pharm. Pharmaceut. Sci. 8 (2005) 147–163. [5] A. Parmar, K. Singh, A. Bahadur, G. Marangoni, P. Bahadur, Interaction and solubilization of some phenolic antioxidants in Pluronic (R) micelles, Colloids Surf. B 86 (2011) 319–326. [6] R.G. Strickley, Solubilizing excipients in oral and injectable formulations, Pharm. Res. 21 (2004) 201–230. [7] G. Gaucher, P. Satturwar, M.C. Jones, A. Furtos, J.C. Leroux, Polymeric micelles for oral drug delivery, Eur. J. Pharm. Biopharm. 76 (2010) 147–158. [8] Y. Kadam, U. Yerramilli, A. Bahadur, Solubilization of poorly water-soluble drug carbamazepine in PluronicÒ micelles: effect of molecular characteristics, temperature and added salt on the solubilizing capacity, Colloids Surf. B 72 (2009) 141–147. [9] M. Pavan Kumar, G.Y. Srawan Kumar, S. Apte, Y. Madhusudan Rao, Review of solubilization techniques for a poorly water-soluble drug: carbamazepine, PDA J. Pharm. Sci. Technol. 64 (2010) 264–277. [10] M. Maswal, O.A. Chat, S. Jabeen, U. Ashraf, R. Masrat, R.A. Shah, A.A. Dar, Solubilization and co-solubilization of carbamazepine and nifedipine in mixed micellar systems: insights from surface tension, electronic absorption, fluorescence and HPLC measurements, RSC Adv. 5 (2015) 7697–7712. [11] M.W. Samaha, M.A.F. Gadalla, Solubilization of carbamazepine by different classes of nonionic surfactants and a bile salt, Drug Dev. Ind. Pharm. 13 (1987) 93–112. [12] L.D. Hu, H.Y. Wu, F. Niu, C.H. Yan, X. Yang, Y.H. Jia, Design of fenofibrate microemulsion for improved bioavailability, Int. J. Pharm. 420 (2011) 251– 255. [13] V.P. Sant, D. Smith, J.C. Leroux, Enhancement of oral bioavailability of poorly water-soluble drugs by poly(ethylene glycol)-block-poly(alkyl acrylate-comethacrylic acid) self-assemblies, J. Control. Release 104 (2005) 289–300. [14] G.E. Granero, C. Ramachandran, G.L. Amidon, Dissolution and solubility behavior of fenofibrate in sodium lauryl sulfate solutions, Drug Dev. Ind. Pharm. 31 (2005) 917–922. [15] M. Srabovic, M. Poljakovic, E. Pehlic, Micellar solubilization of carbamazepine, J. Sci. Res. Rep. 3 (2014) 3106–3116. [16] S. Jamzad, R. Fassihi, Role of surfactant and pH on dissolution properties of fenofibrate and glipizide – a technical note, AAPS Pharm. Sci. Tech. 7 (2006) E17–E22. [17] J.R. Crison, N.D. Weiner, G.L. Amidon, Dissolution media for in vitro testing of water-insoluble drugs: effect of surfactant purity and electrolyte on in vitro dissolution of carbamazepine in aqueous solutions of sodium lauryl sulfate, J. Pharm. Sci. 86 (1997) 384–388.

102

V. Nguyen-Kim et al. / Journal of Colloid and Interface Science 477 (2016) 94–102

[18] H. Lee, S.-A. Park, H. Sah, Surfactant effects upon dissolution patterns of carbamazepine immediate release tablet, Arch. Pharm. Res. 28 (2005) 120– 126. [19] A. Kogan, A. Aserin, N. Garti, Improved solubilization of carbamazepine and structural transitions in nonionic microemulsions upon aqueous phase dilution, J. Colloid Interf. Sci. 315 (2007) 637–647. [20] L. Hu, H. Wu, F. Niu, C. Yan, X. Yang, Y. Jia, Design of fenofibrate microemulsion for improved bioavailability, Int. J. Pharm. 420 (2011) 251–255. [21] M.Y. Kozlov, N.S. Melik-Nubarov, E.V. Batrakova, A.V. Kabanov, Relationship between pluronic block copolymer structure, critical micellization concentration and partitioning coefficients of low molecular mass solutes, Macromolecules 33 (2000) 3305–3313. [22] M. Almgren, W. Brown, S. Hvidt, Self-aggregation and phase-behaviour of poly (ethylene oxide) poly(propylene oxide) poly(ethylene oxide) blockcopolymers in aqueous solution, Colloid Polym. Sci. 273 (1995) 2–15. [23] K. Mortensen, J.S. Pedersen, Structural study on the micelle formation of poly (ethylene oxide) poly(propylene oxide) poly(ethylene oxide) triblock copolymer in aqueous solution, Macromolecules 26 (1993) 805–812. [24] P. Linse, M. Malmsten, Temperature dependent micellization in aqueous block copolymer solutions, Macromolecules 25 (1992) 5434–5439. [25] M. Bohorquez, C. Koch, T. Trygstad, N. Pandit, A study of the temperaturedependent micellization of pluronic F127, J. Colloid Interf. Sci. 216 (1999) 34– 40. [26] K. Pandya, P. Bahadur, T.N. Nagar, A. Bahadur, Micellar and solubilizing behaviour of pluronic L-64 in water, Colloids Surf. A 70 (1993) 219–227. [27] R. Basak, R. Bandyopadhyay, Encapsulation of hydrophobic drugs in pluronic F127 micelles: effects of drug hydrophobicity, solution temperature, and pH, Langmuir 29 (2013) 4350–4356. [28] G. Wanka, H. Hoffmann, W. Ulbricht, Phase-diagrams and aggregation behaviour of poly(oxyethylene)-poly(oxypropylene)-poly(oxyethylene) triblock copolymers in aqueous solutions, Macromolecules 27 (1994) 4145– 4159. [29] P. Alexandridis, Poly(ethylene oxide) poly(propylene oxide) block copolymer surfactants, Curr. Opin. Colloid Interf. Sci. 2 (1997) 478–489. [30] D.A. Chiappetta, A. Sosnik, Poly(ethylene oxide)–poly(propylene oxide) block copolymer micelles as drug delivery agents: improved hydrosolubility, stability and bioavailability of drugs, Eur. J. Pharm. Biopharm. 66 (2007) 303–317. [31] H.Y. Liu, S. Prevost, M. Gradzielski, Solubilisation of oils of different polarity in aqueous solutions of pluronic trib lock copolymers, Z. Phys. Chem. (Muenchen, Ger.) 226 (2012) 675–694. [32] P. Holmqvist, P. Alexandridis, B. Lindman, Modification of the microstructure in poloxamer block copolymer-water-‘‘oil” systems by varying the ‘‘oil” type, Macromolecules 30 (1997) 6788–6797. [33] B. Foster, T. Cosgrove, B. Hammouda, Pluronic triblock copolymer systems and their interactions with ibuprofen, Langmuir 25 (2009) 6760–6766. [34] S. Alexander, T. Cosgrove, S.W. Prescott, T.C. Castle, Flurbiprofen encapsulation using pluronic triblock copolymers, Langmuir 27 (2011) 8054–8060. [35] A.V. Kabanov, E.V. Batrakova, V.Y. Alakhov, PluronicÒ block copolymers as novel polymer therapeutics for drug and gene delivery, J. Control. Release 82 (2002) 189–212. [36] V.Y. Alakhov, E.Y. Moskaleva, E.V. Batrakova, A.V. Kabanov, Hypersensitization of multidrug resistant human ovarian carcinoma cells by pluronic P85 block copolymer, Bioconjugate Chem. 7 (1996) 209–216. [37] M. Varshney, T.E. Morey, D.O. Shah, J.A. Flint, B.M. Moudgil, C.N. Seubert, D.M. Dennis, Pluronic microemulsions as nanoreservoirs for extraction of bupivacaine from normal saline, J. Am. Chem. Soc. 126 (2004) 5108–5112. [38] M.A. Sharp, C. Washington, T. Cosgrove, Solubilisation of model adjuvants by Pluronic block copolymers, J. Colloid Interf. Sci. 344 (2010) 438–446.

[39] J. Causse, J. Oberdisse, J. Jestin, S. Lagerge, Small-angle neutron scattering study of solubilization of tributyl phosphate in aqueous solutions of L64 pluronic triblock copolymers, Langmuir 26 (2010) 15745–15753. [40] V. Patel, J. Dey, R. Ganguly, S. Kumar, S. Nath, V.K. Aswal, P. Bahadur, Solubilization of hydrophobic alcohols in aqueous Pluronic solutions: investigating the role of dehydration of the micellar core in tuning the restructuring and growth of Pluronic micelles, Soft Matter 9 (2013) 7583– 7591. [41] P. Alexandridis, J.F. Holzwarth, T.A. Hatton, Micellization of poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) triblocks copolymers in aqueous solutions – thermodynamics of copolymer association, Macromolecules 27 (1994) 2414–2425. [42] U. Keiderling, The new ‘BerSANS-PC’ software for reduction and treatment of small angle neutron scattering data, Appl. Phys. A Mater. Sci. Process. 74 (2002) S1455–S1457. [43] I. Bressler, J. Kohlbrecher, A.F. Thuenemann, SASfit: a tool for small-angle scattering data analysis using a library of analytical expressions, J. Appl. Cryst. 48 (2015) 1587–1598. [44] E.O. Stejskal, J.E. Tanner, Spin diffusion measurements: spin echoes in the presence of a timde-dependent field gradient, J. Chem. Phys. 42 (1965) 288– 292. [45] T. Tadros, Applied Surfactants, Wiley, VCH, Germany, 2005. [46] A. Beig, J.M. Miller, A. Dahan, Accounting for the solubility–permeability interplay in oral formulation development for poor water solubility drugs: the effect of PEG-400 on carbamazepine absorption, Eur. J. Pharm. Biopharm. 81 (2012) 386–391. [47] E. Batrakova, S. Lee, S. Li, A. Venne, V. Alakhov, A. Kabanov, Fundamental relationships between the composition of Pluronic block copolymers and their hypersensitization effect in MDR cancer cells, Pharm. Res. 16 (1999) 1373– 1379. [48] K. Mortensen, J.S. Pedersen, Structural study on the micelle formation of poly (ethylene oxide) poly(propylene oxide) poly(ethylene oxide) triblock copolymer in aqueous solution, Macromolecules 26 (1993) 805–812. [49] K. Mortensen, Y. Talmon, Cryo-TEM and SANS microstructural study of pluronic polymer solutions, Macromolecules 28 (1995) 8829–8834. [50] J.S. Pedersen, M.C. Gerstenberg, The structure of P85 Pluronic block copolymer micelles determined by small-angle neutron scattering, Colloids Surf. A 213 (2003) 175–187. [51] N.W. Ashcroft, J. Lekner, Structure and resistivity of liquid metals, Phys. Rev. 145 (1966) 83–91. [52] R. De Lisi, M. Gradzielski, G. Lazzara, S. Milioto, N. Muratore, S. Prevost, Aqueous block copolymer-surfactant mixtures and their ability in solubilizing chlorinated organic compounds. A thermodynamic and SANS study, J. Phys. Chem. B 110 (2006) 25883–25894. [53] G. Lazzara, S. Milioto, M. Gradzielski, The solubilisation behaviour of some dichloroalkanes in aqueous solutions of PEO–PPO–PEO triblock copolymers: a dynamic light scattering, fluorescence spectroscopy, and SANS study, Phys. Chem. Chem. Phys. 8 (2006) 2299–2312. [54] J.S. Lettow, T.M. Lancaster, C.J. Glinka, J.Y. Ying, Small-angle neutron scattering and theoretical investigation of poly(ethylene oxide)-poly(propylene oxide)poly(ethylene oxide) stabilized oil-in-water microemulsions, Langmuir 21 (2005) 5738–5746. [55] J.H. Ma, Y. Wang, C. Guo, H.Z. Liu, Y.L. Tang, P. Bahadur, Oil-induced aggregation of block copolymer in aqueous solution, J. Phys. Chem. B 111 (2007) 11140–11148. [56] M. Scherlund, M. Malmsten, P. Holmqvist, A. Brodin, Thermosetting microemulsions and mixed micellar solutions as drug delivery systems for periodontal anesthesia, Int. J. Pharm. 194 (2000) 103–116.