Article pubs.acs.org/JPCB
Liquid Crystalline Phase Formation in Suspensions of Solid Trimyristin Nanoparticles Simone Gehrer,† Martin Schmiele,† Martin Westermann,‡ Frank Steiniger,‡ and Tobias Unruh*,† †
Physik Department, Friedrich-Alexander-Universität Erlangen−Nürnberg, Staudtstrasse 3, 91058 Erlangen, Germany Center for Electron Microscopy, Jena University Hospital, Ziegelmühlenweg 1, 07743 Jena, Germany
‡
S Supporting Information *
ABSTRACT: The presence of liquid crystalline phases in suspensions of solid lipid nanoparticles can increase the risk of their gelling upon administration through fine needles. Here we study the formation of liquid crystalline phases in aqueous suspensions of platelet-like shaped solid lipid nanoparticles. A native lecithin-stabilized trimyristin (20 wt %) suspension was investigated at different dilution levels by small-angle X-ray scattering (SAXS) and visual inspection of their birefringence between two crossed polarizers. For trimyristin concentrations φMMM < 6 wt %, the dispersed platelets are well separated from each other whereas they start to self-assemble into stacked lamellae for 6 wt % ≤ φMMM < 12 wt %. For φMMM ≥ 12 wt %, the SAXS patterns become increasingly anisotropic, which is a signature of an evolving formation of a preferred orientation of the platelets on a microscopic scale. Simultaneously, the suspensions become birefringent, which proves the existence of an anisotropic liquid crystalline phase formed in the still low viscous liquid suspensions. Spatially resolved SAXS scans and polarization microscopy indicate rather small domains in the (sub)micrometer size range in the nematic liquid crystalline phase and the presence of birefringent droplets (tactoids). The observed critical concentrations for the formation of stacks and the liquid crystalline phase are significantly higher as for equivalent suspensions prepared from triglycerides with longer chains. This can be explained with the lower aspect ratio of trimyristin platelets. Special emphasis is put on the isotropic−liquid crystalline phase transition as a function of the ionic strength of the dispersion medium and φMMM. Higher salt concentrations allow shifting of the phase transition to higher trimyristin concentrations. This can be attributed to a partial screening of the repulsive forces between the platelets, which allows higher packing densities within the platelet stacks and of remaining isolated platelets.
■
INTRODUCTION Platelet-like shaped colloidal particles can form different types of liquid crystalline phases in suspension. Theoretical calculations and simulations predicted the existence of nematic and columnar phases for suspensions of hard colloidal platelets.1−3 Experimentally, nematic and columnar phases were observed for inorganic Gibbsite4−10 and nickel hydroxide11 suspensions. In phase-separated Gibbsite suspensions with coexisting isotropic and nematic phases, droplets with a nematic order of the platelets (tactoids) have been observed.4−12 On a microscopic scale, these droplets can be regarded as the precursor of a macroscopic nematic phase. At sufficiently high concentrations also columnar tactoids were found, where the platelets are ordered in hexagonally arranged columns.13 Aqueous suspensions of lecithin-stabilized triglyceride nanocrystals have been studied in the context of drug carrier systems for parenteral administration of poorly water-soluble drugs.14−19 The nanoparticles are produced by high pressure melt homogenization. Upon cooling, the emulsion droplets crystallize typically into platelet-like nanocrystals in the stable β-modification of the triglycerides.19,20 © 2014 American Chemical Society
In previous SAXS and TEM studies, it was demonstrated that the platelet-like triglyceride nanoparticles can self-assemble at sufficiently high concentrations into stacked lamellae21−24 whereby the stack formation is completely reversible upon dilution. Recently, it has been demonstrated for lecithinstabilized tripalmitin (PPP) nanoparticles that at even higher concentrations the stacks can transform into a nematic phase.25 The formation of the liquid crystalline phases could adversely affect the rheological properties and the stability of such suspensions upon administration through fine needles and storage. For native PPP suspensions (10 wt % PPP) stabilized with the lecithins 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC) and soybean lecithin (S100), the isotropic−nematic phase transitions started at PPP concentrations of only 7 and 9 wt %, respectively,25 and thus in the range relevant for pharmaceutical applications. The difference of the critical concentrations of the phase transition could be attributed to Received: July 8, 2014 Revised: September 7, 2014 Published: September 8, 2014 11387
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
until a count rate in the range between 50 and 100 kcps was reached. Because the crystallization temperature of the MMM emulsion is significantly lower than 20 °C (cf. Results), the measurements of the sample were performed at room temperature (21 °C) in the state of a suspension and emulsion (after heating to 60 °C and subsequent cooling to 21 °C). Micro Differential Scanning Calorimetry (μDSC). The μDSC measurements were performed with a Micro DSC III microcalorimeter (Setaram, Caluire-et-Cuire, France). The sample cell was filled with 150 mg of the native MMM suspension, the reference cell with the same amount of purified water. The sample was scanned at a rate of 0.1 K/min from 6 to 60 °C and afterward cooled at the same rate back to 6 °C. The enthalpies of fusion and crystallization were determined by an integration of the heating and cooling curves with respect to their baselines. Freeze-Fracture Transmission Electron Microscopy (TEM). A small droplet of the native MMM suspension was embedded between two copper sandwich profiles. The sandwich was rapidly frozen in a liquid propane/ethane mixture (1:1) cooled by liquid nitrogen, placed in a freeze-fracture unit (BAF 400D, BAL-TEC, Liechtenstein), and fractured in a vacuum chamber at −140 °C and 10−6 mbar. The fractured samples were shadowed under an angle of 35° with platinum/carbon covering the fracture with a 2 nm layer of Pt/C. Subsequently 15−20 nm of carbon was deposited from the top in order to stabilize the replica. The replicas were detached from the copper profiles in deionized water, cleaned for 1 h in a mixture of methanol and chloroform (1:1), fished up on copper grids, and transferred to a transmission electron microscope (Zeiss CEM 902A, Carl Zeiss AG, Oberkochen, Germany) operated at 80 kV. Polarization Microscopy. Polarization micrographs were collected at 21 °C for the native MMM suspension, placed between a microscope slide and a coverslip with an Eclipse LV100D microscope (Nikon, Tokyo, Japan) equipped with a 2 mega pixel CCD camera (Nikon DS-2Mv). Small-Angle X-ray Scattering (SAXS). The SAXS patterns were collected with a highly customized Ganesha 300 XL + SAXS instrument (SAXSLAB ApS, Skovlunde, Denmark) at our institute. The system was equipped with a Cu Kα (λ = 1.5418 Å) Microfocus X-ray source (GeniX, 30 W, Xenocs, Sassenage, France) and a FOX 3D multilayer optics. The beam was collimated by two automated double slit systems (aperture sizes g1 and g2) with a distance of about 1.2 m. The second slit system consists of four “scatterless” silicon single crystal blades. The sample position was located directly between the second slit system and the evacuated detector tube. A 2D Pilatus 300 K detector (Dectris Ltd., Baden, Switzerland) was used to collect the scattered radiation. The 3D motion of the detector inside the detector tube including the adjustment of the sample−detector distance was motorized. All measurements were performed at 21 °C. The sample− detector distance (sdd) and the beam center were determined using a silver behenate standard. With these parameters, λ and the size of the quadratic detector pixels of 172 μm, the scattering angle 2θ of each pixel, can be converted to an s-scale (s = Q/(2π) = (2/λ) sin (2θ/2) in units of (1/nm)). The collected scattering patterns of the samples and water were corrected for transmission, acquisition time, and sample thickness and were put on an absolute scale (in units of (1/cm/ sr)) using a glassy carbon standard,32 kindly provided by the 15ID-D USAXS beamline at the Advanced Photon Source,
different aspect ratios of the platelets induced by the different lecithin stabilizers. Here, the critical concentrations for the formation of stacks and for the isotropic−liquid crystalline phase transition are studied for a lecithin-stabilized trimyristin (MMM) suspension in a broad concentration range of MMM between 4 and 20 wt %. In contrast to the previous study, MMM is used here to study the influence of the matrix triglyceride on the critical concentrations. The critical concentration for the isotropic− liquid crystalline phase transition is determined in a dilution series by visual inspection of the samples between crossed polarizers and by the manifestation of an orientation in the collected 2D SAXS patterns. Using a two-dimensional SAXS scan and polarization microscopy, we estimate the domain size of the liquid crystalline phase. The influence of different ionic strengths on the isotropic−liquid crystalline phase transition is investigated.
■
EXPERIMENTAL METHODS Sample Preparation. A native trimyristin suspension (20 wt % MMM, 4.8 wt % stabilizer, and 1.2 wt % costabilizer) was prepared by high pressure melt homogenization as follows: MMM (Dynasan 114, 98% purity,26 Sasol GmbH, Witten, Germany) and the stabilizer S100, a purified soybean lecithin (≥94% phosphatidylcholine with mostly unsaturated C18 fatty acids,27 Lipoid GmbH, Ludwigshafen, Germany) were heated until a clear yellowish melt was obtained. Sodium glycocholate (NaGC, >98% purity, TCI Europe N.V., Zwijndrecht, Belgium) was dissolved in purified water and heated to the same temperature. The costabilizer NaGC is required to prevent the gelation of the dispersion upon crystallization.28,29 The hot mixture of both liquids was predispersed by using an Ultra-Turrax T25 Basic disperser (IKA-Werke GmbH & Co, Staufen, Germany) with a speed of 25000 rpm for 3 min. The thus obtained pre-emulsion was passed through a heated (65 °C) high pressure melt homogenizer APV-2000 (APV Deutschland GmbH, Unna, Germany) at 1 kbar for 4 min. The hot emulsion was allowed to cool down to 6 °C and stored at this temperature. Upon cooling, the emulsion droplets crystallize to platelet-like nanocrystals.19 For a dilution series, the native suspension was diluted with purified water to reach final MMM concentrations in the range between 4 and 20 wt %. Similarly, for different NaCl concentrations (10−4, 10−3, 10−2, and 10−1 mol/L), appropriate aliquots of the native suspension were mixed with adequate volumes of NaCl solutions. In the following, except for the salt concentrations, all sample compositions are given in wt %. Because the aqueous phase and the lipid ingredients possess weight densities close to 1 g/cm3 (1.036 g/cm3 for β-MMM,30 about 1−1.05 g/cm3 for phosphatidylcholines31), the concentrations given in wt % are nearly the same as in vol %. Thus, the native suspension has a total lipid volume fraction of about 26 vol %. Photon Correlation Spectroscopy (PCS). The particle size and the polydispersity index were determined for the native MMM dispersion in the state of an emulsion and suspension by PCS, using the cumulant method to analyze the measured correlation function. The correlation function was measured with a photon correlation spectrometer (Brookhaven Instruments Corporation, Holtsville, NY, USA), consisting of a MiniL 30 compact diode laser (30 mW, 637.6 nm) and a BI-200 SM goniometer with a photomultiplier, at a scattering angle of 90°. A few droplets of the dispersion were diluted in purified water 11388
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
method. For this analysis, the scattering patterns of the dilute dispersion are simulated by N = 10 ensembles, each consisting of 500 single platelets with a parallelepipedal shape. All platelets in the ith ensemble (i = 1,...N) possess a thickness of i MMM unit cells along the G001 direction. Thus, the thickness hi of the nanocrystals in ensemble i reads as hi = i·d001 = i·3.57 nm (excluding the lecithin stabilizer layer). The two lateral diameters of the parallelepipedal platelets in the ensembles are subject to a Gaussian distribution with a mean and variance of 200 and 50 nm, respectively. The lecithin stabilizer layer covering the MMM nanocrystals is modeled by two concentric shells representing the hydrophobic acyl chains of the lecithin molecules (inner layer) and their hydrophilic headgroups (outer layer), respectively. Both the thicknesses (disl, dosl) and the electron densities (ρisl, ρosl) of the inner and outer stabilizer layers (isl, osl), can be determined later by a fit of the simulation data to the SAXS pattern. The scattering of the particle core is calculated on the basis of the triclinic structure of β-MMM with Z = 2 molecules per unit cell, which has been determined by van Langevelde et al.30 Here we use the convention a < b < c for the three lattice parameters a, b, and c. The rearranged lattice parameters from ref 30 for the triclinic unit cell read as a = 5.4588 Å, b = 12.0626 Å, c = 41.7140 Å, α = 118.274°, β = 73.388°, and γ = 100.408°. To ensure two full MMM molecules within the unit cell, the transformed fractional coordinates x′, y′, and z′ must be used: x′ = z + 1/2, y ′ = x + 1/2, and z′ = y (first molecule) and x′ = 1/2 − z, y′ = 1/2−x, and z′ = 1 − y (second molecule), where x, y, and z are the fractional coordinates reported by van Langevelde et al.30 In the next step, the macroscopic scattering cross section (dΣ/dΩ) of the measured SAXS pattern can be fitted by a linear combination of the simulated scattering patterns (dΣ/ dΩ)i including a constant background term A > 0,
Argonne, IL, USA. Subsequently, from the sample patterns, the pattern of the water reference was subtracted. One-dimensional SAXS patterns were obtained by an angular average of the 2D patterns. A dilution series of the native MMM suspension was measured in the concentration range between 4 and 20% MMM with the configuration g1 = 0.7 mm, g2 = 0.4 mm and sdd = 761 mm, providing a beam size of about 0.5 mm × 0.5 mm at the sample position. Acquisition times varied between 1800 and 9000 s depending on concentration. The samples were contained between two 5−7 μm thick mica windows in circular cells with a diameter of 5 mm and a sample thickness of 1 mm. To investigate the domain sizes and the spatial distribution of the favored orientations of the platelet stacks in the native suspension, the native suspension was measured between two 20−25 μm thick mica windows in a circular sample cell with a diameter of 20 mm and a thickness of 1.25 mm. SAXS patterns were recorded for 480 s at 81 different positions of a quadratic grid centered in the middle of the sample cell (nine positions each in steps of 0.7 mm along the horizontal and vertical direction, corresponding to a total scanned area of 5.6 × 5.6 mm2). The same configuration for the collimation system was used as above (beam diameter of about 0.5 mm), and the sdd was set to 810 mm. A single SAXS pattern was recorded with higher angular resolution using a configuration with g1 = 0.35 mm, g2 = 0.27 mm, and sdd = 2078 mm for a measuring time of 1200 s. To study the formation of the liquid crystalline phases in their native form, it was attempted to avoid shear-stress as far as possible. All sample holders were filled using pipettes or syringes without needles. However, when sealing the sample cells with a mica window, shear effects cannot be fully avoided. Additional shear stress, caused by, e.g., pushing the suspension through fine needles, could promote the formation of a liquid crystalline phase.
⎛ dΣ ⎞ ⎜ ⎟ = ⎝ dΩ ⎠
■
THEORETICAL METHODS Determination of the Platelet Thicknesses. With the Xray powder pattern simulation analysis (XPPSA) method, it is possible to reproduce the SAXS patterns of suspensions of triglyceride nanocrystals on an absolute scale.23,27 MMM emulsions stabilized with S100 and NaGC crystallize predominantly into the stable β-modification where the nanocrystals adopt a platelet-like shape (cf. Results).21,33,34 The β-modification of triglycerides possess a triclinic crystal structure.20,30 Regarding their thickness, the platelets consist of only a few molecular layers of MMM molecules.35 Assuming the c-axis to be the longest axis of the unit cell, the height of one molecular layer is given by the crystallographic d001 value, which represents the height of one unit cell along the direction of the reciprocal lattice vector G001. Because of the small thickness of the platelets (i.e., a low number of MMM unit cells along the G001 direction) and a fairly large d001 spacing of 3.57 nm for β-MMM,30 a broad 001 Bragg peak can be observed in the SAXS patterns. This broadened Bragg peak represents a key feature in the SAXS patterns and allows determination of the distribution function of the platelet thicknesses and thus in combination with the platelet diameters to estimate the aspect ratios of the platelets. For that purpose, the SAXS pattern of a dilute MMM suspension which possesses essentially no self-assembled but only well separated particles is analyzed with the XPPSA
N
⎛ dΣ ⎞i ⎟ + A dΩ ⎠
∑ ci⎝⎜ i=1
N
∑ ci = 1 i=1
(1)
The fitted linear coefficients ci can be interpreted as the volume fractions of all platelets in the suspension with a thickness of i MMM unit cells.27 In the fitting procedure, they must obey the completeness relation given on the right-hand side of eq 1.27 As mentioned above also the stabilizer shell thicknesses disl and dosl as well as the electron densities ρisl and ρosl, which are implicitly included in the (dΣ/dΩ)i, are fitted, too. A scaling factor 0.8 < χ < 1.2 for the simulated SAXS pattern is permitted to account for possible errors in the determination of the absolute scale. A volume averaged thickness of the nanocrystals, ⟨i⟩V, in units of d001 = 3.57 nm, can be calculated as27 N
⟨i⟩V =
∑i ci·i N
∑i ci
N
=
∑ ci·i i
(2)
Thus, the mean platelet thickness ⟨h⟩V is then given by ⟨h⟩V = ⟨i⟩V ·d001 + 2·(d isl + dosl)
(3)
The computer simulations were performed with the C++ program XNDiff36 on the high performance computing cluster Lima at the Regionales Rechenzentrum Erlangen (RRZE). Details on the computer simulations and the fitting routine can be found elsewhere.27 11389
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
■
Article
RESULTS AND DISCUSSION Sample Characterization. The native sample was characterized by PCS to get an estimate for the platelet diameters and by mDSC and freeze-fracture TEM to characterize the polymorphic state of the nanoparticles. The PCS measured correlation functions (Figures S1 and S2 in the Supporting Information) can be fitted for both the emulsion and suspension state with a monomodal distribution with moderate polydispersities, respectively. The z-averaged mean particle diameter in the case of the suspension is 111.5 nm and the polydispersity index 0.165. However, for platelets, the z-average provides only a crude approximation of the platelet diameters. For the spherical emulsion droplets, the determination of the mean diameters is more reliable. The determined z-average of dPCS = 100.3 nm and polydispersity index of 0.185 for the emulsion droplets indicate that the platelet diameters for their crystallized counterparts are significantly higher than the measured 111.5 nm. A better estimation for the platelet diameters can be obtained under the assumption of equal volumes for the emulsion droplets and the nanocrystals with the mean platelet thickness ⟨h⟩V = 18.9 nm, which is determined with the XPPSA method below. A mean platelet diameter ⟨D⟩ can be estimated as ⟨D⟩ =
3 2 dPCS = 188 nm 3 ⟨h⟩V
Figure 2. Freeze-fracture TEM micrograph taken from the native MMM suspension.
platelet-like shape which is characteristic for the β-modification of the triglycerides.39,40 The platelet diameters typically range between 50 and 300 nm. Spherical nanoparticles which are characteristic for the thermodynamically instable α-modification39 could not be observed. Dilution Series. A dilution series of the native MMM suspension was studied in the concentration range between 4 and 20% MMM on a microscopic level by SAXS and on a macroscopic scale by the observation of the samples placed between crossed polarizers. The aim was to examine the formation of stacks and in particular liquid crystalline phases and to determine their respective critical concentrations. Observation under Crossed Polarizers. For the visual inspection, the suspensions were filled at a temperature of 21 °C in quartz cuvettes with a path length of 0.5 mm, placed between two crossed polarizers, illuminated through one of the polarizers by a standard tungsten halogen lamp with an optical diffuser (cf. Figure S3 in the Supporting Information), and photographed using a standard digital camera (cf. Figure 3). In the range between 4 and 12% MMM, no birefringence can be observed. At 13% MMM, a weak lighted strip can be seen in the middle of the cuvette (might be visible in the electronic version
(4)
The μDSC heating and cooling runs are shown in Figure 1. The melting curve shows the characteristic fractionated melting
Figure 1. Microcalorimetric heating (bottom) and cooling (top) scans of the native MMM suspension. The gray shaded areas determine the enthalpy of fusion and crystallization in the heating and cooling run, respectively.
pattern of β-MMM platelets in the temperature range between about 35 and 55 °C. In this range, the melting curve consists of many partially overlapping melting peaks, where each peak belongs to the melting of platelets with a particular thickness (i.e., a certain number of molecular layers of β-MMM).33,37 The MMM emulsion crystallizes in the subsequent cooling run only at 10.5 °C due to a strong supercooling effect.38 The determined enthalpies of fusion and crystallization for the native suspension (20% MMM) are 33.4 and −30.3 J/g, respectively. This corresponds to 167 and −151.5 J/g for 100% MMM, respectively. The comparison with reported enthalpy of fusion for bulk β-MMM of 181 J/g38 demonstrates that the platelets mostly crystallized in the stable β-modification. Freeze-fracture TEM micrographs (cf. Figure 2) taken from a native suspension support this finding. The particles exhibit a
Figure 3. Photographs of cuvettes filled with MMM suspensions of different dilutions (4−20% MMM), demonstrating the birefringence of the dispersions. For reference, a corresponding photograph of a water sample is displayed. The white cross indicates the orientations of the polarizers. Starting at a concentration of about 7% MMM, the images become brighter. Above 13% MMM, the samples exhibit an inhomogeneous pattern of birefringent areas. 11390
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
only), this effect gets stronger at 14%, and at 15% MMM large bright spots appear which further grow with rising concentration. In contrast to the Gibbsite suspensions, no phase separation could be observed, probably because the weight density difference between MMM and the aqueous phase is too small and the viscosity of the dispersions are significant at high concentrations. It is worth noting that at concentrations ≥7% a diffuse brightening over the complete cuvette can be observed. For concentrations lower than 7%, this happens only in the middle of the cuvette (maximum intensity of the light source). It is assumed that this brightening is caused by the stacked lamellae formed in the dispersions at the corresponding concentrations. SAXS Measurements. The circularly averaged SAXS patterns of the dilution series are visualized in Figure 4. The
Figure 5. Azimuthal distribution of the intensity along the 001 Debye−Scherrer ring for MMM nanosuspensions at different MMM concentrations. The curves were shifted vertically for better visualization.
ranges 0−180° and 180−360°, respectively, were averaged for getting better statistics. Because the reciprocal lattice vector G001 is parallel to the director of the platelets (i.e., perpendicular to the large (001) faces of the platelets) and, furthermore, also parallel to the stacking direction of the platelets, an elevated scattering contribution at a particular azimuthal angle of the 001 Bragg powder ring and the stackrelated maxima corresponds to a preferential orientation of platelets and platelet stacks in a direction under this angle. Up to a concentration of 11%, the intensity profiles are completely flat, i.e., the 2D scattering patterns are fully isotropic. At approximately 12%, the intensity shows a tiny anisotropy around 75°. At 13% and above, the anisotropy becomes clearly visible and the magnitudes of the amplitudes grow with rising concentration. From the SAXS data, it is found that the formation of stacks starts at the critical concentration of 6% MMM. This agrees well with the diffuse brightening found for samples with ≥7% MMM in the optical examination. The brightening is most probably caused by the platelet stacks which cause small birefringent contributions. Also the deformation of the 001 Bragg reflection with the scattering contributions from stacked lamellae and the pronounced rise in the amplitudes of the stack-related peaks for MMM concentrations ≥13% coincides with the optically observed phase transition. This is caused by a higher degree of order of the platelets in the liquid crystalline phase. An increasing proportion of the platelets in the liquid crystalline phase orientates over length scales of at least several hundred nanometers in a preferred direction. Below the critical concentration of the phase transition, the stacks are rather randomly orientated isolated objects with a poorer order of the platelets in the stacks.
Figure 4. Representation of the azimuthally averaged 2D-SAXS patterns of a MMM dispersion at different concentrations. The curves were shifted vertically by a factor of 2.5 for a better visualization. For the suspension with 4% MMM, a fit obtained with the XPPSA method is drawn in red. The inset in the lower left corner shows the fitted platelet thickness distribution, where the bars represent the volume fractions ci of platelets with a thickness of i = 1···10 unit cells MMM.
SAXS pattern for the suspension with 4% MMM features the 001 Bragg reflection besides the diffuse small-angle scattering at small s-values. Starting at about 6% MMM, additional interference maxima in the low s-range become visible (labeled with numbers in Figure 4 in the case of the native suspension). They can be assigned to the lamellar order of self-assembled platelets. With rising MMM concentration, the intensity of the maxima grows (in particular for ≥13% MMM) and their positions shift to larger s, i.e., the amount of self-assembled stacks rises, while the repeat distance of the platelets within the stacks decreases. At higher concentrations, higher orders of the interferences become apparent, indicating a higher degree of order of the platelets within the stacks. At 13% MMM, the maxima also start to interfere with the 001 Bragg reflection, resulting in its strong deformation. From the 2D SAXS patterns, intensity profiles of the 001 Bragg peak as a function of the azimuthal angle (along the 001 Bragg powder ring) could be extracted and are displayed in Figure 5. The two intensity profiles achieved from the angular 11391
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
Figure 6. (a) Assembly of the 2D detector images from a SAXS scan over different postions of the native MMM suspension (20% MMM). (b) Extracted preferential platelet orientations at the different sample positions. At a particular position, the direction of the dashes represents the preferential platelet orientations. The lengths of the dashes describe the magnitude of the anisotropy. Areas with similar dash patterns are outlined with (red) solid lines.
PPP platelets from a previous study (dPCS = 84.2 nm, disl = 7.0 Å and dosl = 3.5 Å, ⟨i⟩V = 2.79·d001, d001 = 4.02 nm).25 The lower critical concentrations of 4% PPP for the stack formation in PPP suspensions22,25 is in good agreement with the observation that the propensity of triglyceride platelets to self-assemble in stacks rises with the length of the fatty acids of the triglyceride in the sequence trilaurin (C12) < trimyristin (C14) < tripalmitin (C16) < tristearin (C18).21 Triglycerides with longer fatty acids produce platelets with higher aspect ratios that promote the formation of stacks and increase the viscosity of the suspensions.26,41 Accordingly, the significantly higher critical concentration of 12% MMM for the isotropic−liquid crystalline phase transition found here for the MMM suspension with respect to the reported 9% PPP for PPP suspensions can be attributed to the lower aspect ratio for MMM platelets with respect to PPP platelets. Domain Size. 2D SAXS Scan for the Native Suspension. We performed a two-dimensional SAXS scan for the native suspension to investigate how the spatial distribution of the preferred stack orientations varies in order to estimate the domain sizes of the liquid crystalline phase. The 81 positions of a quadratic grid (5.6 × 5.6 mm2), centered in the middle of the circular sample holder with a diameter of 20 mm, were scanned. The step size of 0.7 mm in horizontal and vertical directions was larger than the beam size of 0.5 mm. The assembled 2D SAXS patterns of all positions are visualized in Figure 6a. For all measured positions, the 2D SAXS patterns are anisotropic, and thus the total scanned range is liquid crystalline. At many positions, two preferential orientations are found. Exemplary Figure 7 shows the 2D SAXS pattern for such a position. The intensity profile along the 001 powder ring of such a 2D pattern has been extracted as described above and is shown in Figure S4 in the Supporting Information. The intensity profiles I(ϕ) were fitted with up to two Gaussian functions with amplitudes Ai, mean values ϕi (i = 1,2), variances σi, and a constant A0:
Both the SAXS measurements and visual inspection with crossed polarizers provide similar critical concentrations for the stack formation (6%) and the phase transition (12%). The concentration-dependent transition from randomly oriented isolated platelets via randomly oriented platelet stacks to preferentially oriented platelets and stacks in the liquid crystalline phase is schematically visualized for the probed concentration range in the graphical abstract. The slightly lower values for the critical concentrations determined by SAXS can be attributed to the fact that SAXS probes changes in the sample on a microscopic level. Eventually, these microscopic changes become apparent at slightly higher concentrations on a macroscopic scale (photographs). Determination of the Platelet Thicknesses. From the SAXS pattern of the dilute suspension containing 4% of MMM, the thickness distribution of the platelets in the MMM suspensions has been determined by using the XPPSA method. The fitted SAXS curve and its fitted distribution of the volume fractions ci of platelets with a thickness of i molecular layers of MMM are visualized in Figure 4. Two-thirds of the nanocrystals possess a thickness of 3−5 unit cells. On average, the thickness of the MMM core of the platelets is ⟨i⟩V = 4.74 in units of d001. The fitted inner and outer shell thicknesses are disl = 6.0 Å and dosl = 3.5 Å, respectively, are in good agreement with disl = 7.0 Å and dosl = 3.0 Å found in a previous study on S100-stabilized tripalmitin (PPP) suspensions.24 The fitted scaling factor χ = 0.88 for the SAXS pattern deviates by 12% from 1, which is suggested to be a reasonable value considering possible errors in the determination of the absolute scale (sample thickness, transmission, and calibration of the absolute intensity scale with the glassy carbon standard). The mean platelet thickness including the stabilizer layer becomes with eq 3 ⟨h⟩V = 18.9 nm. Together with the estimated mean platelet diameter ⟨D⟩ = 188 nm (cf. eq 4), the mean platelet aspect ratio ⟨D⟩/⟨h⟩V for the S100-stabilized MMM platelets is 9.9. This is significantly lower than the aspect ratio ⟨D⟩/⟨h⟩V = 13.0 that can be calculated in a similar way for S100-stabilized 11392
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
orientations over length scales of a few millimeters, explaining the large fairly homogeneous birefringent areas optically observed. Polarization Microscopy. To verify the model proposed above, the native sample has also been studied by polarization microscopy (cf. Figure 8). The micrographs in Figure 8a,b exhibit some birefringent droplets of spherical shape and diameters typically ranging between 5 and 30 μm beside some diffuse birefringent background. Some droplets appear as bright spheres with a dark cross (cf. Figure 8c). This birefringence pattern is comparable to the patterns of the nematic tactoids found in suspensions of Gibbsite platelets. The micrographs support the finding from the SAXS data that the domain sizes are rather small and do not exceed several micrometers. Most probably the observed birefringent droplets represent the largest domain sizes in the suspensions. Liquid Crystalline Order. To study the liquid crystalline phase in more detail, an additional high resolved SAXS pattern has been recorded at the center position of the 2D SAXS scan. The corresponding SAXS pattern is displayed in Figure 9.
Figure 7. 2D SAXS pattern of a native MMM nanosuspension exhibiting two main preferential orientations of the liquid crystalline phase (azimuth angles ϕ1 and ϕ2). The 001 Bragg peak and the stackrelated interference maxima (white arrows) are indicated. 2
I(ϕ) = A 0 +
⎛ (ϕ − ϕ)2 ⎞ i ⎟ ⎟ 2σi2 ⎠ ⎝
∑ Ai exp⎜⎜− i=1
(5)
In Figure 6, the favored orientations ϕi of the liquid crystalline phase and their magnitudes Ai are visualized for all measured positions in the sample. The dashes represent the preferred orientations of the normals of the platelets which are oriented along the G001 direction. The lengths of the dashes represent the magnitudes Ai of the anisotropy. Generally, ϕi and Ai vary from position to position. However, areas with similar patterns for the favored orientations can be identified and are framed with (red) solid lines. These areas vary in size and shape and can stretch over a few millimeters in space. A0 is a measure for a remaining fraction of randomly orientated stacks and single platelets which cause an isotropic scattering contribution to the 001 Bragg ring. A0 is within a few percent of a constant independent of the investigated position in the sample. In most cases the amplitudes A1 and A2 do not exceed A0 by a factor of 3 or more. The lack of very strong orientations in the 2D SAXS patterns and the presence of a non-negligible amount of randomly oriented stacks indicate that the liquid crystalline phase consists of many small domains with sizes in the (sub)micrometer size range rather than of one (or two consecutive) macroscopic domains. Within the domains, the platelets share the same directors. The directors of the small domains can probably vary significantly but obviously share one or two similar preferential
Figure 9. SAXS pattern of a native MMM suspension (20% MMM). The curve is assembled of two measurements at two different sample− detector distances. The stack-related interference maxima are labeled with numbers and the position of the 001 Bragg reflection is indicated. P1 marks a point that might be related to side−side correlations between the platelets.
Besides the peaks that can be assigned to the lamellar order of the platelets, the SAXS pattern exhibits a small broadened bump at s = 0.0057 nm−1. The corresponding length scale of about 175 nm agrees well with the estimated platelet diameter ⟨D⟩ = 188 nm. However, because the bump is very weak, there exist if at all only weak side−side correlations between the
Figure 8. Polarization microscopy images taken from the native MMM suspension at different magnifications (scale bars correspond to lengths of 100 μm (a,b) and 10 μm (c)). The white crosses indicate the polarizer orientations. 11393
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
ionic strength up to 10−2 mol/L NaCl, the viscosity of triglyceride suspensions dropped significantly without altering the stability of the suspensions significantly.41 Flocculation was also found at 10−1 mol/L NaCl. Thus, the formation of the liquid crystalline phase at elevated concentrations of the dispersed nanoparticles adversely affects the flow properties of the suspensions. Formulations exhibiting liquid crystalline phases already in their native state should be considered critical with regard to a potential application because they expose the risk of gelling upon administration through narrow needles. The addition of salts can increase the critical concentration for the isotropic−liquid crystalline phase transformation and by this improve the flow properties of the suspensions.
platelets, and thus, the liquid crystalline phase has only a nematic order rather than a nematic columnar one. This is also reasonable with regard to the significant polydispersity in size and shape of the MMM platelets (cf. TEM micrograph in Figure 2), which hinders the formation of liquid crystalline phases with a higher degree of order. Salt Concentration. For different ionic strengths (0, 10−4, −3 10 , 10−2, 10−1 mol/L NaCl) and MMM concentrations φMMM in the range between 2% and 18%, we established a phase diagram. For that purpose, a freshly prepared native MMM suspension was used. As it was found in the dilution series, it is sufficient to examine the samples between crossed polarizers because it provides nearly the same critical concentration for the isotropic−nematic phase transition as the SAXS measurements. We can distinguish three cases: isotropic phase, liquid crystalline phase, and flocculation. Samples are considered as flocculated when several larger macroscopic particles are observed. The phase diagram is shown in Figure 10.
■
CONCLUSION The physical state of suspensions of lecithin-stabilized trimyristin (MMM) platelets was studied over a large concentration range of the dispersed phase. In good agreement, the critical concentrations for the stack formation (6−7 wt % MMM) and a isotropic−nematic phase transition (12−13 wt % MMM) can be identified by SAXS and optical birefringence experiments. The critical concentrations for the stack formation and the phase transition are significantly higher than the reported values for similar tripalmitin (PPP) suspensions. This can be attributed to the lower aspect ratio of MMM platelets with respect to PPP platelets in equivalent S100-stabilized PPP suspensions. Thus, using monoacid triglycerides with shorter acyl chains (e.g., trimyristin and trilaurin) as matrix lipids allows shifting of the isotropic−nematic phase transition to higher triglyceride concentrations while those with longer chains (e.g., tripalmitin and tristearin) promote the phase transition. Raising the ionic strength can also shift the isotropic−nematic phase transition to higher MMM concentrations and thus improve the flow properties of the suspensions. This can be explained by a denser packing of the platelets as a result of a decrease of the repulsive forces between them. Spatial SAXS scans and polarization micrographs indicate that the liquid crystalline phase consists of rather small domains with sizes of a few micrometers. In the micrographs, birefringent droplets of several micrometers in diameter are found that resemble the nematic droplets (tactoids) previously found in suspensions of Gibbsite platelets. Optical birefringence measurements of such triglyceride nanosuspensions (in their native state or under shear stress) could complement rheological measurements and serve as an easy-to-use method to evaluate their applicability for an intravenous administration (injectability, syringeability).
Figure 10. Phase diagram for MMM suspensions at different MMM concentrations φMMM (or total lipid concentration φ = 1.3 φMMM) and ionic strengths.
Without salt, the observations are in good agreement with the previous dilution series and the suspensions become increasingly birefringent starting at 14% MMM. For low salt concentrations of 10−4 mol/L, the same results are obtained. As the salt concentration further rises, the isotropic−liquid crystalline phase transition is shifted to higher MMM concentrations. At 10−3 mol/L NaCl, the phase transition occurs at 15% MMM. For 10−2 mol/L NaCl, all studied samples are isotropic in the studied concentration range. Thus, adding moderate amounts of salt allows inhibition of a phase transition in the formulations. At the highest studied salt concentration, 10−1 mol/L NaCl, the suspensions with 14% MMM and above begin to flocculate. The triglyceride nanocrystals stabilized by lecithins and NaGC possess a negative ζ potential42 which causes repulsive forces between them. The shift of the phase transition to higher MMM concentrations with rising ionic strength can be explained by a reduction of the repulsive forces between the platelets. At higher ionic strengths, the platelets and stacks can be more densely packed. At sufficiently high ionic strengths and platelet concentrations, the stabilization of the platelets collapses and the suspension starts to flocculate. The effect of salts on the viscosity of suspensions of solid lipid nanoparticles shows a similar impact.41,43 When rising the
■
ASSOCIATED CONTENT
S Supporting Information *
Figures showing the PCS measured correlations functions of the native suspension and emulsion, the experimental setup of the optical birefringence measurements, and the evaluation of the scattered intensity along the 001 Debye−Scherrer rings. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Phone: +49-(0)9131-8525189. Fax: +49-(0)9131-8525182. 11394
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
Notes
(20) van Langevelde, A.; van Malssen, K.; Hollander, F.; Peschar, R.; Schenk, H. Structure of Mono-Acid Even-Numbered β-Triacylglycerols. Acta Crystallogr., Sect. B: Struct. Sci. 1999, 55, 114−122. (21) Unruh, T.; Westesen, K.; Bösecke, P.; Lindner, P.; Koch, M. H. J. Self-Assembly of Triglyceride Nanocrystals in Suspension. Langmuir 2002, 18, 1796−1800. (22) Illing, A.; Unruh, T.; Koch, M. Investigation on Particle SelfAssembly in Solid Lipid-Based Colloidal Drug Carrier Systems. Pharm. Res. 2004, 21, 592−597. (23) Unruh, T. Interpretation of Small-Angle X-Ray Scattering Patterns of Crystalline Triglyceride Nanoparticles in Dispersion. J. Appl. Crystallogr. 2007, 40, 1008−1018. (24) Schmiele, M.; Schindler, T.; Westermann, M.; Steiniger, F.; Radulescu, A.; Kriele, A.; Gilles, R.; Unruh, T. Mesoscopic Structures of Triglyceride Nanosuspensions Studied by Small-Angle X-ray and Neutron Scattering and Computer Simulations. J. Phys. Chem. B 2014, 118, 8808−8818. (25) Schmiele, M.; Gehrer, S.; Westermann, M.; Steiniger, F.; Unruh, T. Formation of Liquid Crystalline Phases in Aqueous Suspensions of Platelet-like Tripalmitin Nanoparticles. J. Chem. Phys. 2014, 140, 214905. (26) Seetapan, N.; Bejrapha, P.; Srinuanchai, W.; Puttipipatkhachorn, S.; Ruktanonchai, U. Nondestructive Rheological Measurement of Aqueous Dispersions of Solid Lipid Nanoparticles: Effects of Lipid Types and Concentrations on Dispersion Consistency. Drug Dev. Ind. Pharm. 2010, 36, 1005−1015. (27) Schmiele, M.; Schindler, T.; Unruh, T.; Busch, S.; Morhenn, H.; Westermann, M.; Steiniger, F.; Radulescu, A.; Lindner, P.; Schweins, R.; et al. Structural Characterization of the Phospholipid Stabilizer Layer at the Solid−Liquid Interface of Dispersed Triglyceride Nanocrystals with Small-Angle Xray and Neutron Scattering. Phys. Rev. E 2013, 87, 062316. (28) Westesen, K.; Siekmann, B.; Koch, M. H. Investigations on the Physical State of Lipid Nanoparticles by Synchrotron Radiation X-Ray Diffraction. Int. J. Pharm. 1993, 93, 189−199. (29) Helgason, T.; Awad, T.; Kristbergsson, K.; McClements, D.; Weiss, J. Influence of Polymorphic Transformations on Gelation of Tripalmitin Solid Lipid Nanoparticle Suspensions. J. Am. Oil Chem. Soc. 2008, 85, 501−511. (30) van Langevelde, A.; Peschar, R.; Schenk, H. Structure of βTrimyristin and β-Tristearin from High-Resolution X-Ray Powder Diffraction Data. Acta Crystallogr., Sect. B: Struct. Sci. 2001, 57, 372− 377. (31) Nagle, J. F.; Tristram-Nagle, S. Structure of Lipid Bilayers. Biochim. Biophys. Acta, Rev. Biomembr. 2000, 1469, 159−195. (32) Zhang, F.; Ilavsky, J.; Long, G.; Quintana, J.; Allen, A.; Jemian, P. Glassy Carbon as an Absolute Intensity Calibration Standard for Small-Angle Scattering. Metall. Mater. Trans. A 2010, 41, 1151−1158. (33) Unruh, T.; Bunjes, H.; Westesen, K.; Koch, M. H. J. Observation of Size-Dependent Melting in Lipid Nanoparticles. J. Phys. Chem. B 1999, 103, 10373−10377. (34) Bunjes, H.; Koch, M. H. J. Saturated Phospholipids Promote Crystallization but Slow Down Polymorphic Transitions in Triglyceride Nanoparticles. J. Controlled Release 2005, 107, 229−243. (35) Bunjes, H.; Unruh, T. Characterization of Lipid Nanoparticles by Differential Scanning Calorimetry, X-Ray and Neutron Scattering. Adv. Drug Delivery Rev. 2007, 59, 379−402. (36) The program XNDiff is available free of charge from https:// sourceforge.net/projects/xndiff/. (37) Unruh, T.; Bunjes, H.; Westesen, K.; Koch, M. H. J. Investigations on the Melting Behaviour of Triglyceride Nanoparticles. Colloid Polym. Sci. 2001, 279, 398−403. (38) Bunjes, H.; Westesen, K.; Koch, M. H. Crystallization Tendency and Polymorphic Transitions in Triglyceride Nanoparticles. Int. J. Pharm. 1996, 129, 159−173. (39) Bunjes, H.; Steiniger, F.; Richter, W. Visualizing the Structure of Triglyceride Nanoparticles in Different Crystal Modifications. Langmuir 2007, 23, 4005−4011.
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded by the Deutsche Forschungsgemeinschaft (DFG) through the project UN267/2-1 and the Cluster of Excellence Engineering of Advanced Materials. S.G. expresses her thanks to C. Bär for designing the sample holder for the Ganesha camera. We acknowledge the RRZE for using the Lima HPC cluster.
■
REFERENCES
(1) Veerman, J. A. C.; Frenkel, D. Phase behavior of disklike hardcore mesogens. Phys. Rev. A 1992, 45, 5632−5648. (2) Wensink, H.; Lekkerkerker, H. Phase Diagram of Hard Colloidal Platelets: A Theoretical Account. Mol. Phys. 2009, 107, 2111−2118. (3) Marechal, M.; Cuetos, A.; Martínez-Haya, B.; Dijkstra, M. Phase Behavior of Hard Colloidal Platelets Using Free Energy Calculations. J. Chem. Phys. 2011, 134, 094501. (4) van der Kooij, F. M.; Lekkerkerker, H. N. W. Formation of Nematic Liquid Crystals in Suspensions of Hard Colloidal Platelets. J. Phys. Chem. B 1998, 102, 7829−7832. (5) van der Kooij, F.; Kassapidou, K.; Lekkerkerker, H. Liquid Crystal Phase Transitions in Suspensions of Polydisperse Plate-like Particles. Nature 2000, 406, 868−871. (6) van der Beek, D.; Lekkerkerker, H. N. W. Nematic Ordering vs Gelation in Suspensions of Charged Platelets. Europhys. Lett. 2003, 61, 702−707. (7) van der Beek, D.; Lekkerkerker, H. N. W. Liquid Crystal Phases of Charged Colloidal Platelets. Langmuir 2004, 20, 8582−8586. (8) van der Beek, D.; Petukhov, A.; Oversteegen, S.; Vroege, G.; Lekkerkerker, H. Evidence of the Hexagonal Columnar Liquid-Crystal Phase of Hard Colloidal Platelets by High-Resolution SAXS. Eur. Phys. J. E 2005, 16, 253−258. (9) Mourad, M.; Wijnhoven, J.; van’t Zand, D.; van der Beek, D.; Lekkerkerker, H. Gelation versus Liquid Crystal Phase Transitions in Suspensions of Plate-like Particles. Philos. Trans. R. Soc., A 2006, 364, 2807−2816. (10) Mourad, M. C. D.; Petukhov, A. V.; Vroege, G. J.; Lekkerkerker, H. N. W. Lyotropic Hexagonal Columnar Liquid Crystals of Large Colloidal Gibbsite Platelets. Langmuir 2010, 26, 14182−14187. (11) Brown, A. B. D.; Clarke, S. M.; Rennie, A. R. Ordered Phase of Platelike Particles in Concentrated Dispersions. Langmuir 1998, 14, 3129−3132. (12) Verhoeff, A. A.; Bakelaar, I. A.; Otten, R. H. J.; van der Schoot, P.; Lekkerkerker, H. N. W. Tactoids of Plate-Like Particles: Size, Shape, and Director Field. Langmuir 2011, 27, 116−125. (13) Verhoeff, A. A.; Lekkerkerker, H. N. W. Direct Observation of Columnar Liquid Crystal Droplets. Soft Matter 2012, 8, 4865−4868. (14) zur Mühlen, A.; Schwarz, C.; Mehnert, W. Solid Lipid Nanoparticles (SLN) for Controlled Drug DeliveryDrug Release and Release Mechanism. Eur. J. Pharm. Biopharm. 1998, 45, 149−155. (15) Müller, R. H.; Mäder, K.; Gohla, S. Solid Lipid Nanoparticles (SLN) for Controlled Drug DeliveryA Review of the State of the Art. Eur. J. Pharm. Biopharm. 2000, 50, 161−177. (16) Westesen, K. Novel Lipid-Based Colloidal Dispersions as Potential Drug Administration SystemsExpectations and Reality. Colloid Polym. Sci. 2000, 278, 608−618. (17) Wissing, S.; Kayser, O.; Müller, R. Solid Lipid Nanoparticles for Parenteral Drug Delivery. Adv. Drug Delivery Rev. 2004, 56, 1257− 1272. (18) Joshi, M. D.; Müller, R. H. Lipid Nanoparticles for Parenteral Delivery of Actives. Eur. J. Pharm. Biopharm. 2009, 71, 161−172. (19) Bunjes, H. Structural Properties of Solid Lipid Based Colloidal Drug Delivery Systems. Curr. Opin. Colloid Interface Sci. 2011, 16, 405−411. 11395
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
(40) Kuntsche, J.; Horst, J. C.; Bunjes, H. Cryogenic Transmission Electron Microscopy (Cryo-TEM) for Studying the Morphology of Colloidal Drug Delivery Systems. Int. J. Pharm. 2011, 417, 120−137. (41) Illing, A.; Unruh, T. Investigation on the Flow Behavior of Dispersions of Solid Triglyceride Nanoparticles. Int. J. Pharm. 2004, 284, 123−131. (42) Siekmann, B.; Westesen, K. Submicron-Sized Parenteral Carrier Systems Based on Solid Lipids. Pharm. Pharmacol. Lett. 1992, 1, 123− 126. (43) Freitas, C. Stability determination of solid lipid nanoparticles (SLN TM) in aqueous dispersion after addition of electrolyte. J. Microencapsulation 1999, 16, 59−71.
11396
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
Liquid Crystalline Phase Formation in Suspensions of Solid Trimyristin Nanoparticles Simone Gehrer,† Martin Schmiele,† Martin Westermann,‡ Frank Steiniger,‡ and Tobias Unruh*,† †
Physik Department, Friedrich-Alexander-Universität Erlangen−Nürnberg, Staudtstrasse 3, 91058 Erlangen, Germany Center for Electron Microscopy, Jena University Hospital, Ziegelmühlenweg 1, 07743 Jena, Germany
‡
S Supporting Information *
ABSTRACT: The presence of liquid crystalline phases in suspensions of solid lipid nanoparticles can increase the risk of their gelling upon administration through fine needles. Here we study the formation of liquid crystalline phases in aqueous suspensions of platelet-like shaped solid lipid nanoparticles. A native lecithin-stabilized trimyristin (20 wt %) suspension was investigated at different dilution levels by small-angle X-ray scattering (SAXS) and visual inspection of their birefringence between two crossed polarizers. For trimyristin concentrations φMMM < 6 wt %, the dispersed platelets are well separated from each other whereas they start to self-assemble into stacked lamellae for 6 wt % ≤ φMMM < 12 wt %. For φMMM ≥ 12 wt %, the SAXS patterns become increasingly anisotropic, which is a signature of an evolving formation of a preferred orientation of the platelets on a microscopic scale. Simultaneously, the suspensions become birefringent, which proves the existence of an anisotropic liquid crystalline phase formed in the still low viscous liquid suspensions. Spatially resolved SAXS scans and polarization microscopy indicate rather small domains in the (sub)micrometer size range in the nematic liquid crystalline phase and the presence of birefringent droplets (tactoids). The observed critical concentrations for the formation of stacks and the liquid crystalline phase are significantly higher as for equivalent suspensions prepared from triglycerides with longer chains. This can be explained with the lower aspect ratio of trimyristin platelets. Special emphasis is put on the isotropic−liquid crystalline phase transition as a function of the ionic strength of the dispersion medium and φMMM. Higher salt concentrations allow shifting of the phase transition to higher trimyristin concentrations. This can be attributed to a partial screening of the repulsive forces between the platelets, which allows higher packing densities within the platelet stacks and of remaining isolated platelets.
■
INTRODUCTION Platelet-like shaped colloidal particles can form different types of liquid crystalline phases in suspension. Theoretical calculations and simulations predicted the existence of nematic and columnar phases for suspensions of hard colloidal platelets.1−3 Experimentally, nematic and columnar phases were observed for inorganic Gibbsite4−10 and nickel hydroxide11 suspensions. In phase-separated Gibbsite suspensions with coexisting isotropic and nematic phases, droplets with a nematic order of the platelets (tactoids) have been observed.4−12 On a microscopic scale, these droplets can be regarded as the precursor of a macroscopic nematic phase. At sufficiently high concentrations also columnar tactoids were found, where the platelets are ordered in hexagonally arranged columns.13 Aqueous suspensions of lecithin-stabilized triglyceride nanocrystals have been studied in the context of drug carrier systems for parenteral administration of poorly water-soluble drugs.14−19 The nanoparticles are produced by high pressure melt homogenization. Upon cooling, the emulsion droplets crystallize typically into platelet-like nanocrystals in the stable β-modification of the triglycerides.19,20 © 2014 American Chemical Society
In previous SAXS and TEM studies, it was demonstrated that the platelet-like triglyceride nanoparticles can self-assemble at sufficiently high concentrations into stacked lamellae21−24 whereby the stack formation is completely reversible upon dilution. Recently, it has been demonstrated for lecithinstabilized tripalmitin (PPP) nanoparticles that at even higher concentrations the stacks can transform into a nematic phase.25 The formation of the liquid crystalline phases could adversely affect the rheological properties and the stability of such suspensions upon administration through fine needles and storage. For native PPP suspensions (10 wt % PPP) stabilized with the lecithins 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC) and soybean lecithin (S100), the isotropic−nematic phase transitions started at PPP concentrations of only 7 and 9 wt %, respectively,25 and thus in the range relevant for pharmaceutical applications. The difference of the critical concentrations of the phase transition could be attributed to Received: July 8, 2014 Revised: September 7, 2014 Published: September 8, 2014 11387
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
until a count rate in the range between 50 and 100 kcps was reached. Because the crystallization temperature of the MMM emulsion is significantly lower than 20 °C (cf. Results), the measurements of the sample were performed at room temperature (21 °C) in the state of a suspension and emulsion (after heating to 60 °C and subsequent cooling to 21 °C). Micro Differential Scanning Calorimetry (μDSC). The μDSC measurements were performed with a Micro DSC III microcalorimeter (Setaram, Caluire-et-Cuire, France). The sample cell was filled with 150 mg of the native MMM suspension, the reference cell with the same amount of purified water. The sample was scanned at a rate of 0.1 K/min from 6 to 60 °C and afterward cooled at the same rate back to 6 °C. The enthalpies of fusion and crystallization were determined by an integration of the heating and cooling curves with respect to their baselines. Freeze-Fracture Transmission Electron Microscopy (TEM). A small droplet of the native MMM suspension was embedded between two copper sandwich profiles. The sandwich was rapidly frozen in a liquid propane/ethane mixture (1:1) cooled by liquid nitrogen, placed in a freeze-fracture unit (BAF 400D, BAL-TEC, Liechtenstein), and fractured in a vacuum chamber at −140 °C and 10−6 mbar. The fractured samples were shadowed under an angle of 35° with platinum/carbon covering the fracture with a 2 nm layer of Pt/C. Subsequently 15−20 nm of carbon was deposited from the top in order to stabilize the replica. The replicas were detached from the copper profiles in deionized water, cleaned for 1 h in a mixture of methanol and chloroform (1:1), fished up on copper grids, and transferred to a transmission electron microscope (Zeiss CEM 902A, Carl Zeiss AG, Oberkochen, Germany) operated at 80 kV. Polarization Microscopy. Polarization micrographs were collected at 21 °C for the native MMM suspension, placed between a microscope slide and a coverslip with an Eclipse LV100D microscope (Nikon, Tokyo, Japan) equipped with a 2 mega pixel CCD camera (Nikon DS-2Mv). Small-Angle X-ray Scattering (SAXS). The SAXS patterns were collected with a highly customized Ganesha 300 XL + SAXS instrument (SAXSLAB ApS, Skovlunde, Denmark) at our institute. The system was equipped with a Cu Kα (λ = 1.5418 Å) Microfocus X-ray source (GeniX, 30 W, Xenocs, Sassenage, France) and a FOX 3D multilayer optics. The beam was collimated by two automated double slit systems (aperture sizes g1 and g2) with a distance of about 1.2 m. The second slit system consists of four “scatterless” silicon single crystal blades. The sample position was located directly between the second slit system and the evacuated detector tube. A 2D Pilatus 300 K detector (Dectris Ltd., Baden, Switzerland) was used to collect the scattered radiation. The 3D motion of the detector inside the detector tube including the adjustment of the sample−detector distance was motorized. All measurements were performed at 21 °C. The sample− detector distance (sdd) and the beam center were determined using a silver behenate standard. With these parameters, λ and the size of the quadratic detector pixels of 172 μm, the scattering angle 2θ of each pixel, can be converted to an s-scale (s = Q/(2π) = (2/λ) sin (2θ/2) in units of (1/nm)). The collected scattering patterns of the samples and water were corrected for transmission, acquisition time, and sample thickness and were put on an absolute scale (in units of (1/cm/ sr)) using a glassy carbon standard,32 kindly provided by the 15ID-D USAXS beamline at the Advanced Photon Source,
different aspect ratios of the platelets induced by the different lecithin stabilizers. Here, the critical concentrations for the formation of stacks and for the isotropic−liquid crystalline phase transition are studied for a lecithin-stabilized trimyristin (MMM) suspension in a broad concentration range of MMM between 4 and 20 wt %. In contrast to the previous study, MMM is used here to study the influence of the matrix triglyceride on the critical concentrations. The critical concentration for the isotropic− liquid crystalline phase transition is determined in a dilution series by visual inspection of the samples between crossed polarizers and by the manifestation of an orientation in the collected 2D SAXS patterns. Using a two-dimensional SAXS scan and polarization microscopy, we estimate the domain size of the liquid crystalline phase. The influence of different ionic strengths on the isotropic−liquid crystalline phase transition is investigated.
■
EXPERIMENTAL METHODS Sample Preparation. A native trimyristin suspension (20 wt % MMM, 4.8 wt % stabilizer, and 1.2 wt % costabilizer) was prepared by high pressure melt homogenization as follows: MMM (Dynasan 114, 98% purity,26 Sasol GmbH, Witten, Germany) and the stabilizer S100, a purified soybean lecithin (≥94% phosphatidylcholine with mostly unsaturated C18 fatty acids,27 Lipoid GmbH, Ludwigshafen, Germany) were heated until a clear yellowish melt was obtained. Sodium glycocholate (NaGC, >98% purity, TCI Europe N.V., Zwijndrecht, Belgium) was dissolved in purified water and heated to the same temperature. The costabilizer NaGC is required to prevent the gelation of the dispersion upon crystallization.28,29 The hot mixture of both liquids was predispersed by using an Ultra-Turrax T25 Basic disperser (IKA-Werke GmbH & Co, Staufen, Germany) with a speed of 25000 rpm for 3 min. The thus obtained pre-emulsion was passed through a heated (65 °C) high pressure melt homogenizer APV-2000 (APV Deutschland GmbH, Unna, Germany) at 1 kbar for 4 min. The hot emulsion was allowed to cool down to 6 °C and stored at this temperature. Upon cooling, the emulsion droplets crystallize to platelet-like nanocrystals.19 For a dilution series, the native suspension was diluted with purified water to reach final MMM concentrations in the range between 4 and 20 wt %. Similarly, for different NaCl concentrations (10−4, 10−3, 10−2, and 10−1 mol/L), appropriate aliquots of the native suspension were mixed with adequate volumes of NaCl solutions. In the following, except for the salt concentrations, all sample compositions are given in wt %. Because the aqueous phase and the lipid ingredients possess weight densities close to 1 g/cm3 (1.036 g/cm3 for β-MMM,30 about 1−1.05 g/cm3 for phosphatidylcholines31), the concentrations given in wt % are nearly the same as in vol %. Thus, the native suspension has a total lipid volume fraction of about 26 vol %. Photon Correlation Spectroscopy (PCS). The particle size and the polydispersity index were determined for the native MMM dispersion in the state of an emulsion and suspension by PCS, using the cumulant method to analyze the measured correlation function. The correlation function was measured with a photon correlation spectrometer (Brookhaven Instruments Corporation, Holtsville, NY, USA), consisting of a MiniL 30 compact diode laser (30 mW, 637.6 nm) and a BI-200 SM goniometer with a photomultiplier, at a scattering angle of 90°. A few droplets of the dispersion were diluted in purified water 11388
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
method. For this analysis, the scattering patterns of the dilute dispersion are simulated by N = 10 ensembles, each consisting of 500 single platelets with a parallelepipedal shape. All platelets in the ith ensemble (i = 1,...N) possess a thickness of i MMM unit cells along the G001 direction. Thus, the thickness hi of the nanocrystals in ensemble i reads as hi = i·d001 = i·3.57 nm (excluding the lecithin stabilizer layer). The two lateral diameters of the parallelepipedal platelets in the ensembles are subject to a Gaussian distribution with a mean and variance of 200 and 50 nm, respectively. The lecithin stabilizer layer covering the MMM nanocrystals is modeled by two concentric shells representing the hydrophobic acyl chains of the lecithin molecules (inner layer) and their hydrophilic headgroups (outer layer), respectively. Both the thicknesses (disl, dosl) and the electron densities (ρisl, ρosl) of the inner and outer stabilizer layers (isl, osl), can be determined later by a fit of the simulation data to the SAXS pattern. The scattering of the particle core is calculated on the basis of the triclinic structure of β-MMM with Z = 2 molecules per unit cell, which has been determined by van Langevelde et al.30 Here we use the convention a < b < c for the three lattice parameters a, b, and c. The rearranged lattice parameters from ref 30 for the triclinic unit cell read as a = 5.4588 Å, b = 12.0626 Å, c = 41.7140 Å, α = 118.274°, β = 73.388°, and γ = 100.408°. To ensure two full MMM molecules within the unit cell, the transformed fractional coordinates x′, y′, and z′ must be used: x′ = z + 1/2, y ′ = x + 1/2, and z′ = y (first molecule) and x′ = 1/2 − z, y′ = 1/2−x, and z′ = 1 − y (second molecule), where x, y, and z are the fractional coordinates reported by van Langevelde et al.30 In the next step, the macroscopic scattering cross section (dΣ/dΩ) of the measured SAXS pattern can be fitted by a linear combination of the simulated scattering patterns (dΣ/ dΩ)i including a constant background term A > 0,
Argonne, IL, USA. Subsequently, from the sample patterns, the pattern of the water reference was subtracted. One-dimensional SAXS patterns were obtained by an angular average of the 2D patterns. A dilution series of the native MMM suspension was measured in the concentration range between 4 and 20% MMM with the configuration g1 = 0.7 mm, g2 = 0.4 mm and sdd = 761 mm, providing a beam size of about 0.5 mm × 0.5 mm at the sample position. Acquisition times varied between 1800 and 9000 s depending on concentration. The samples were contained between two 5−7 μm thick mica windows in circular cells with a diameter of 5 mm and a sample thickness of 1 mm. To investigate the domain sizes and the spatial distribution of the favored orientations of the platelet stacks in the native suspension, the native suspension was measured between two 20−25 μm thick mica windows in a circular sample cell with a diameter of 20 mm and a thickness of 1.25 mm. SAXS patterns were recorded for 480 s at 81 different positions of a quadratic grid centered in the middle of the sample cell (nine positions each in steps of 0.7 mm along the horizontal and vertical direction, corresponding to a total scanned area of 5.6 × 5.6 mm2). The same configuration for the collimation system was used as above (beam diameter of about 0.5 mm), and the sdd was set to 810 mm. A single SAXS pattern was recorded with higher angular resolution using a configuration with g1 = 0.35 mm, g2 = 0.27 mm, and sdd = 2078 mm for a measuring time of 1200 s. To study the formation of the liquid crystalline phases in their native form, it was attempted to avoid shear-stress as far as possible. All sample holders were filled using pipettes or syringes without needles. However, when sealing the sample cells with a mica window, shear effects cannot be fully avoided. Additional shear stress, caused by, e.g., pushing the suspension through fine needles, could promote the formation of a liquid crystalline phase.
⎛ dΣ ⎞ ⎜ ⎟ = ⎝ dΩ ⎠
■
THEORETICAL METHODS Determination of the Platelet Thicknesses. With the Xray powder pattern simulation analysis (XPPSA) method, it is possible to reproduce the SAXS patterns of suspensions of triglyceride nanocrystals on an absolute scale.23,27 MMM emulsions stabilized with S100 and NaGC crystallize predominantly into the stable β-modification where the nanocrystals adopt a platelet-like shape (cf. Results).21,33,34 The β-modification of triglycerides possess a triclinic crystal structure.20,30 Regarding their thickness, the platelets consist of only a few molecular layers of MMM molecules.35 Assuming the c-axis to be the longest axis of the unit cell, the height of one molecular layer is given by the crystallographic d001 value, which represents the height of one unit cell along the direction of the reciprocal lattice vector G001. Because of the small thickness of the platelets (i.e., a low number of MMM unit cells along the G001 direction) and a fairly large d001 spacing of 3.57 nm for β-MMM,30 a broad 001 Bragg peak can be observed in the SAXS patterns. This broadened Bragg peak represents a key feature in the SAXS patterns and allows determination of the distribution function of the platelet thicknesses and thus in combination with the platelet diameters to estimate the aspect ratios of the platelets. For that purpose, the SAXS pattern of a dilute MMM suspension which possesses essentially no self-assembled but only well separated particles is analyzed with the XPPSA
N
⎛ dΣ ⎞i ⎟ + A dΩ ⎠
∑ ci⎝⎜ i=1
N
∑ ci = 1 i=1
(1)
The fitted linear coefficients ci can be interpreted as the volume fractions of all platelets in the suspension with a thickness of i MMM unit cells.27 In the fitting procedure, they must obey the completeness relation given on the right-hand side of eq 1.27 As mentioned above also the stabilizer shell thicknesses disl and dosl as well as the electron densities ρisl and ρosl, which are implicitly included in the (dΣ/dΩ)i, are fitted, too. A scaling factor 0.8 < χ < 1.2 for the simulated SAXS pattern is permitted to account for possible errors in the determination of the absolute scale. A volume averaged thickness of the nanocrystals, ⟨i⟩V, in units of d001 = 3.57 nm, can be calculated as27 N
⟨i⟩V =
∑i ci·i N
∑i ci
N
=
∑ ci·i i
(2)
Thus, the mean platelet thickness ⟨h⟩V is then given by ⟨h⟩V = ⟨i⟩V ·d001 + 2·(d isl + dosl)
(3)
The computer simulations were performed with the C++ program XNDiff36 on the high performance computing cluster Lima at the Regionales Rechenzentrum Erlangen (RRZE). Details on the computer simulations and the fitting routine can be found elsewhere.27 11389
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
■
Article
RESULTS AND DISCUSSION Sample Characterization. The native sample was characterized by PCS to get an estimate for the platelet diameters and by mDSC and freeze-fracture TEM to characterize the polymorphic state of the nanoparticles. The PCS measured correlation functions (Figures S1 and S2 in the Supporting Information) can be fitted for both the emulsion and suspension state with a monomodal distribution with moderate polydispersities, respectively. The z-averaged mean particle diameter in the case of the suspension is 111.5 nm and the polydispersity index 0.165. However, for platelets, the z-average provides only a crude approximation of the platelet diameters. For the spherical emulsion droplets, the determination of the mean diameters is more reliable. The determined z-average of dPCS = 100.3 nm and polydispersity index of 0.185 for the emulsion droplets indicate that the platelet diameters for their crystallized counterparts are significantly higher than the measured 111.5 nm. A better estimation for the platelet diameters can be obtained under the assumption of equal volumes for the emulsion droplets and the nanocrystals with the mean platelet thickness ⟨h⟩V = 18.9 nm, which is determined with the XPPSA method below. A mean platelet diameter ⟨D⟩ can be estimated as ⟨D⟩ =
3 2 dPCS = 188 nm 3 ⟨h⟩V
Figure 2. Freeze-fracture TEM micrograph taken from the native MMM suspension.
platelet-like shape which is characteristic for the β-modification of the triglycerides.39,40 The platelet diameters typically range between 50 and 300 nm. Spherical nanoparticles which are characteristic for the thermodynamically instable α-modification39 could not be observed. Dilution Series. A dilution series of the native MMM suspension was studied in the concentration range between 4 and 20% MMM on a microscopic level by SAXS and on a macroscopic scale by the observation of the samples placed between crossed polarizers. The aim was to examine the formation of stacks and in particular liquid crystalline phases and to determine their respective critical concentrations. Observation under Crossed Polarizers. For the visual inspection, the suspensions were filled at a temperature of 21 °C in quartz cuvettes with a path length of 0.5 mm, placed between two crossed polarizers, illuminated through one of the polarizers by a standard tungsten halogen lamp with an optical diffuser (cf. Figure S3 in the Supporting Information), and photographed using a standard digital camera (cf. Figure 3). In the range between 4 and 12% MMM, no birefringence can be observed. At 13% MMM, a weak lighted strip can be seen in the middle of the cuvette (might be visible in the electronic version
(4)
The μDSC heating and cooling runs are shown in Figure 1. The melting curve shows the characteristic fractionated melting
Figure 1. Microcalorimetric heating (bottom) and cooling (top) scans of the native MMM suspension. The gray shaded areas determine the enthalpy of fusion and crystallization in the heating and cooling run, respectively.
pattern of β-MMM platelets in the temperature range between about 35 and 55 °C. In this range, the melting curve consists of many partially overlapping melting peaks, where each peak belongs to the melting of platelets with a particular thickness (i.e., a certain number of molecular layers of β-MMM).33,37 The MMM emulsion crystallizes in the subsequent cooling run only at 10.5 °C due to a strong supercooling effect.38 The determined enthalpies of fusion and crystallization for the native suspension (20% MMM) are 33.4 and −30.3 J/g, respectively. This corresponds to 167 and −151.5 J/g for 100% MMM, respectively. The comparison with reported enthalpy of fusion for bulk β-MMM of 181 J/g38 demonstrates that the platelets mostly crystallized in the stable β-modification. Freeze-fracture TEM micrographs (cf. Figure 2) taken from a native suspension support this finding. The particles exhibit a
Figure 3. Photographs of cuvettes filled with MMM suspensions of different dilutions (4−20% MMM), demonstrating the birefringence of the dispersions. For reference, a corresponding photograph of a water sample is displayed. The white cross indicates the orientations of the polarizers. Starting at a concentration of about 7% MMM, the images become brighter. Above 13% MMM, the samples exhibit an inhomogeneous pattern of birefringent areas. 11390
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
only), this effect gets stronger at 14%, and at 15% MMM large bright spots appear which further grow with rising concentration. In contrast to the Gibbsite suspensions, no phase separation could be observed, probably because the weight density difference between MMM and the aqueous phase is too small and the viscosity of the dispersions are significant at high concentrations. It is worth noting that at concentrations ≥7% a diffuse brightening over the complete cuvette can be observed. For concentrations lower than 7%, this happens only in the middle of the cuvette (maximum intensity of the light source). It is assumed that this brightening is caused by the stacked lamellae formed in the dispersions at the corresponding concentrations. SAXS Measurements. The circularly averaged SAXS patterns of the dilution series are visualized in Figure 4. The
Figure 5. Azimuthal distribution of the intensity along the 001 Debye−Scherrer ring for MMM nanosuspensions at different MMM concentrations. The curves were shifted vertically for better visualization.
ranges 0−180° and 180−360°, respectively, were averaged for getting better statistics. Because the reciprocal lattice vector G001 is parallel to the director of the platelets (i.e., perpendicular to the large (001) faces of the platelets) and, furthermore, also parallel to the stacking direction of the platelets, an elevated scattering contribution at a particular azimuthal angle of the 001 Bragg powder ring and the stackrelated maxima corresponds to a preferential orientation of platelets and platelet stacks in a direction under this angle. Up to a concentration of 11%, the intensity profiles are completely flat, i.e., the 2D scattering patterns are fully isotropic. At approximately 12%, the intensity shows a tiny anisotropy around 75°. At 13% and above, the anisotropy becomes clearly visible and the magnitudes of the amplitudes grow with rising concentration. From the SAXS data, it is found that the formation of stacks starts at the critical concentration of 6% MMM. This agrees well with the diffuse brightening found for samples with ≥7% MMM in the optical examination. The brightening is most probably caused by the platelet stacks which cause small birefringent contributions. Also the deformation of the 001 Bragg reflection with the scattering contributions from stacked lamellae and the pronounced rise in the amplitudes of the stack-related peaks for MMM concentrations ≥13% coincides with the optically observed phase transition. This is caused by a higher degree of order of the platelets in the liquid crystalline phase. An increasing proportion of the platelets in the liquid crystalline phase orientates over length scales of at least several hundred nanometers in a preferred direction. Below the critical concentration of the phase transition, the stacks are rather randomly orientated isolated objects with a poorer order of the platelets in the stacks.
Figure 4. Representation of the azimuthally averaged 2D-SAXS patterns of a MMM dispersion at different concentrations. The curves were shifted vertically by a factor of 2.5 for a better visualization. For the suspension with 4% MMM, a fit obtained with the XPPSA method is drawn in red. The inset in the lower left corner shows the fitted platelet thickness distribution, where the bars represent the volume fractions ci of platelets with a thickness of i = 1···10 unit cells MMM.
SAXS pattern for the suspension with 4% MMM features the 001 Bragg reflection besides the diffuse small-angle scattering at small s-values. Starting at about 6% MMM, additional interference maxima in the low s-range become visible (labeled with numbers in Figure 4 in the case of the native suspension). They can be assigned to the lamellar order of self-assembled platelets. With rising MMM concentration, the intensity of the maxima grows (in particular for ≥13% MMM) and their positions shift to larger s, i.e., the amount of self-assembled stacks rises, while the repeat distance of the platelets within the stacks decreases. At higher concentrations, higher orders of the interferences become apparent, indicating a higher degree of order of the platelets within the stacks. At 13% MMM, the maxima also start to interfere with the 001 Bragg reflection, resulting in its strong deformation. From the 2D SAXS patterns, intensity profiles of the 001 Bragg peak as a function of the azimuthal angle (along the 001 Bragg powder ring) could be extracted and are displayed in Figure 5. The two intensity profiles achieved from the angular 11391
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
Figure 6. (a) Assembly of the 2D detector images from a SAXS scan over different postions of the native MMM suspension (20% MMM). (b) Extracted preferential platelet orientations at the different sample positions. At a particular position, the direction of the dashes represents the preferential platelet orientations. The lengths of the dashes describe the magnitude of the anisotropy. Areas with similar dash patterns are outlined with (red) solid lines.
PPP platelets from a previous study (dPCS = 84.2 nm, disl = 7.0 Å and dosl = 3.5 Å, ⟨i⟩V = 2.79·d001, d001 = 4.02 nm).25 The lower critical concentrations of 4% PPP for the stack formation in PPP suspensions22,25 is in good agreement with the observation that the propensity of triglyceride platelets to self-assemble in stacks rises with the length of the fatty acids of the triglyceride in the sequence trilaurin (C12) < trimyristin (C14) < tripalmitin (C16) < tristearin (C18).21 Triglycerides with longer fatty acids produce platelets with higher aspect ratios that promote the formation of stacks and increase the viscosity of the suspensions.26,41 Accordingly, the significantly higher critical concentration of 12% MMM for the isotropic−liquid crystalline phase transition found here for the MMM suspension with respect to the reported 9% PPP for PPP suspensions can be attributed to the lower aspect ratio for MMM platelets with respect to PPP platelets. Domain Size. 2D SAXS Scan for the Native Suspension. We performed a two-dimensional SAXS scan for the native suspension to investigate how the spatial distribution of the preferred stack orientations varies in order to estimate the domain sizes of the liquid crystalline phase. The 81 positions of a quadratic grid (5.6 × 5.6 mm2), centered in the middle of the circular sample holder with a diameter of 20 mm, were scanned. The step size of 0.7 mm in horizontal and vertical directions was larger than the beam size of 0.5 mm. The assembled 2D SAXS patterns of all positions are visualized in Figure 6a. For all measured positions, the 2D SAXS patterns are anisotropic, and thus the total scanned range is liquid crystalline. At many positions, two preferential orientations are found. Exemplary Figure 7 shows the 2D SAXS pattern for such a position. The intensity profile along the 001 powder ring of such a 2D pattern has been extracted as described above and is shown in Figure S4 in the Supporting Information. The intensity profiles I(ϕ) were fitted with up to two Gaussian functions with amplitudes Ai, mean values ϕi (i = 1,2), variances σi, and a constant A0:
Both the SAXS measurements and visual inspection with crossed polarizers provide similar critical concentrations for the stack formation (6%) and the phase transition (12%). The concentration-dependent transition from randomly oriented isolated platelets via randomly oriented platelet stacks to preferentially oriented platelets and stacks in the liquid crystalline phase is schematically visualized for the probed concentration range in the graphical abstract. The slightly lower values for the critical concentrations determined by SAXS can be attributed to the fact that SAXS probes changes in the sample on a microscopic level. Eventually, these microscopic changes become apparent at slightly higher concentrations on a macroscopic scale (photographs). Determination of the Platelet Thicknesses. From the SAXS pattern of the dilute suspension containing 4% of MMM, the thickness distribution of the platelets in the MMM suspensions has been determined by using the XPPSA method. The fitted SAXS curve and its fitted distribution of the volume fractions ci of platelets with a thickness of i molecular layers of MMM are visualized in Figure 4. Two-thirds of the nanocrystals possess a thickness of 3−5 unit cells. On average, the thickness of the MMM core of the platelets is ⟨i⟩V = 4.74 in units of d001. The fitted inner and outer shell thicknesses are disl = 6.0 Å and dosl = 3.5 Å, respectively, are in good agreement with disl = 7.0 Å and dosl = 3.0 Å found in a previous study on S100-stabilized tripalmitin (PPP) suspensions.24 The fitted scaling factor χ = 0.88 for the SAXS pattern deviates by 12% from 1, which is suggested to be a reasonable value considering possible errors in the determination of the absolute scale (sample thickness, transmission, and calibration of the absolute intensity scale with the glassy carbon standard). The mean platelet thickness including the stabilizer layer becomes with eq 3 ⟨h⟩V = 18.9 nm. Together with the estimated mean platelet diameter ⟨D⟩ = 188 nm (cf. eq 4), the mean platelet aspect ratio ⟨D⟩/⟨h⟩V for the S100-stabilized MMM platelets is 9.9. This is significantly lower than the aspect ratio ⟨D⟩/⟨h⟩V = 13.0 that can be calculated in a similar way for S100-stabilized 11392
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
orientations over length scales of a few millimeters, explaining the large fairly homogeneous birefringent areas optically observed. Polarization Microscopy. To verify the model proposed above, the native sample has also been studied by polarization microscopy (cf. Figure 8). The micrographs in Figure 8a,b exhibit some birefringent droplets of spherical shape and diameters typically ranging between 5 and 30 μm beside some diffuse birefringent background. Some droplets appear as bright spheres with a dark cross (cf. Figure 8c). This birefringence pattern is comparable to the patterns of the nematic tactoids found in suspensions of Gibbsite platelets. The micrographs support the finding from the SAXS data that the domain sizes are rather small and do not exceed several micrometers. Most probably the observed birefringent droplets represent the largest domain sizes in the suspensions. Liquid Crystalline Order. To study the liquid crystalline phase in more detail, an additional high resolved SAXS pattern has been recorded at the center position of the 2D SAXS scan. The corresponding SAXS pattern is displayed in Figure 9.
Figure 7. 2D SAXS pattern of a native MMM nanosuspension exhibiting two main preferential orientations of the liquid crystalline phase (azimuth angles ϕ1 and ϕ2). The 001 Bragg peak and the stackrelated interference maxima (white arrows) are indicated. 2
I(ϕ) = A 0 +
⎛ (ϕ − ϕ)2 ⎞ i ⎟ ⎟ 2σi2 ⎠ ⎝
∑ Ai exp⎜⎜− i=1
(5)
In Figure 6, the favored orientations ϕi of the liquid crystalline phase and their magnitudes Ai are visualized for all measured positions in the sample. The dashes represent the preferred orientations of the normals of the platelets which are oriented along the G001 direction. The lengths of the dashes represent the magnitudes Ai of the anisotropy. Generally, ϕi and Ai vary from position to position. However, areas with similar patterns for the favored orientations can be identified and are framed with (red) solid lines. These areas vary in size and shape and can stretch over a few millimeters in space. A0 is a measure for a remaining fraction of randomly orientated stacks and single platelets which cause an isotropic scattering contribution to the 001 Bragg ring. A0 is within a few percent of a constant independent of the investigated position in the sample. In most cases the amplitudes A1 and A2 do not exceed A0 by a factor of 3 or more. The lack of very strong orientations in the 2D SAXS patterns and the presence of a non-negligible amount of randomly oriented stacks indicate that the liquid crystalline phase consists of many small domains with sizes in the (sub)micrometer size range rather than of one (or two consecutive) macroscopic domains. Within the domains, the platelets share the same directors. The directors of the small domains can probably vary significantly but obviously share one or two similar preferential
Figure 9. SAXS pattern of a native MMM suspension (20% MMM). The curve is assembled of two measurements at two different sample− detector distances. The stack-related interference maxima are labeled with numbers and the position of the 001 Bragg reflection is indicated. P1 marks a point that might be related to side−side correlations between the platelets.
Besides the peaks that can be assigned to the lamellar order of the platelets, the SAXS pattern exhibits a small broadened bump at s = 0.0057 nm−1. The corresponding length scale of about 175 nm agrees well with the estimated platelet diameter ⟨D⟩ = 188 nm. However, because the bump is very weak, there exist if at all only weak side−side correlations between the
Figure 8. Polarization microscopy images taken from the native MMM suspension at different magnifications (scale bars correspond to lengths of 100 μm (a,b) and 10 μm (c)). The white crosses indicate the polarizer orientations. 11393
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
ionic strength up to 10−2 mol/L NaCl, the viscosity of triglyceride suspensions dropped significantly without altering the stability of the suspensions significantly.41 Flocculation was also found at 10−1 mol/L NaCl. Thus, the formation of the liquid crystalline phase at elevated concentrations of the dispersed nanoparticles adversely affects the flow properties of the suspensions. Formulations exhibiting liquid crystalline phases already in their native state should be considered critical with regard to a potential application because they expose the risk of gelling upon administration through narrow needles. The addition of salts can increase the critical concentration for the isotropic−liquid crystalline phase transformation and by this improve the flow properties of the suspensions.
platelets, and thus, the liquid crystalline phase has only a nematic order rather than a nematic columnar one. This is also reasonable with regard to the significant polydispersity in size and shape of the MMM platelets (cf. TEM micrograph in Figure 2), which hinders the formation of liquid crystalline phases with a higher degree of order. Salt Concentration. For different ionic strengths (0, 10−4, −3 10 , 10−2, 10−1 mol/L NaCl) and MMM concentrations φMMM in the range between 2% and 18%, we established a phase diagram. For that purpose, a freshly prepared native MMM suspension was used. As it was found in the dilution series, it is sufficient to examine the samples between crossed polarizers because it provides nearly the same critical concentration for the isotropic−nematic phase transition as the SAXS measurements. We can distinguish three cases: isotropic phase, liquid crystalline phase, and flocculation. Samples are considered as flocculated when several larger macroscopic particles are observed. The phase diagram is shown in Figure 10.
■
CONCLUSION The physical state of suspensions of lecithin-stabilized trimyristin (MMM) platelets was studied over a large concentration range of the dispersed phase. In good agreement, the critical concentrations for the stack formation (6−7 wt % MMM) and a isotropic−nematic phase transition (12−13 wt % MMM) can be identified by SAXS and optical birefringence experiments. The critical concentrations for the stack formation and the phase transition are significantly higher than the reported values for similar tripalmitin (PPP) suspensions. This can be attributed to the lower aspect ratio of MMM platelets with respect to PPP platelets in equivalent S100-stabilized PPP suspensions. Thus, using monoacid triglycerides with shorter acyl chains (e.g., trimyristin and trilaurin) as matrix lipids allows shifting of the isotropic−nematic phase transition to higher triglyceride concentrations while those with longer chains (e.g., tripalmitin and tristearin) promote the phase transition. Raising the ionic strength can also shift the isotropic−nematic phase transition to higher MMM concentrations and thus improve the flow properties of the suspensions. This can be explained by a denser packing of the platelets as a result of a decrease of the repulsive forces between them. Spatial SAXS scans and polarization micrographs indicate that the liquid crystalline phase consists of rather small domains with sizes of a few micrometers. In the micrographs, birefringent droplets of several micrometers in diameter are found that resemble the nematic droplets (tactoids) previously found in suspensions of Gibbsite platelets. Optical birefringence measurements of such triglyceride nanosuspensions (in their native state or under shear stress) could complement rheological measurements and serve as an easy-to-use method to evaluate their applicability for an intravenous administration (injectability, syringeability).
Figure 10. Phase diagram for MMM suspensions at different MMM concentrations φMMM (or total lipid concentration φ = 1.3 φMMM) and ionic strengths.
Without salt, the observations are in good agreement with the previous dilution series and the suspensions become increasingly birefringent starting at 14% MMM. For low salt concentrations of 10−4 mol/L, the same results are obtained. As the salt concentration further rises, the isotropic−liquid crystalline phase transition is shifted to higher MMM concentrations. At 10−3 mol/L NaCl, the phase transition occurs at 15% MMM. For 10−2 mol/L NaCl, all studied samples are isotropic in the studied concentration range. Thus, adding moderate amounts of salt allows inhibition of a phase transition in the formulations. At the highest studied salt concentration, 10−1 mol/L NaCl, the suspensions with 14% MMM and above begin to flocculate. The triglyceride nanocrystals stabilized by lecithins and NaGC possess a negative ζ potential42 which causes repulsive forces between them. The shift of the phase transition to higher MMM concentrations with rising ionic strength can be explained by a reduction of the repulsive forces between the platelets. At higher ionic strengths, the platelets and stacks can be more densely packed. At sufficiently high ionic strengths and platelet concentrations, the stabilization of the platelets collapses and the suspension starts to flocculate. The effect of salts on the viscosity of suspensions of solid lipid nanoparticles shows a similar impact.41,43 When rising the
■
ASSOCIATED CONTENT
S Supporting Information *
Figures showing the PCS measured correlations functions of the native suspension and emulsion, the experimental setup of the optical birefringence measurements, and the evaluation of the scattered intensity along the 001 Debye−Scherrer rings. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Phone: +49-(0)9131-8525189. Fax: +49-(0)9131-8525182. 11394
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
Notes
(20) van Langevelde, A.; van Malssen, K.; Hollander, F.; Peschar, R.; Schenk, H. Structure of Mono-Acid Even-Numbered β-Triacylglycerols. Acta Crystallogr., Sect. B: Struct. Sci. 1999, 55, 114−122. (21) Unruh, T.; Westesen, K.; Bösecke, P.; Lindner, P.; Koch, M. H. J. Self-Assembly of Triglyceride Nanocrystals in Suspension. Langmuir 2002, 18, 1796−1800. (22) Illing, A.; Unruh, T.; Koch, M. Investigation on Particle SelfAssembly in Solid Lipid-Based Colloidal Drug Carrier Systems. Pharm. Res. 2004, 21, 592−597. (23) Unruh, T. Interpretation of Small-Angle X-Ray Scattering Patterns of Crystalline Triglyceride Nanoparticles in Dispersion. J. Appl. Crystallogr. 2007, 40, 1008−1018. (24) Schmiele, M.; Schindler, T.; Westermann, M.; Steiniger, F.; Radulescu, A.; Kriele, A.; Gilles, R.; Unruh, T. Mesoscopic Structures of Triglyceride Nanosuspensions Studied by Small-Angle X-ray and Neutron Scattering and Computer Simulations. J. Phys. Chem. B 2014, 118, 8808−8818. (25) Schmiele, M.; Gehrer, S.; Westermann, M.; Steiniger, F.; Unruh, T. Formation of Liquid Crystalline Phases in Aqueous Suspensions of Platelet-like Tripalmitin Nanoparticles. J. Chem. Phys. 2014, 140, 214905. (26) Seetapan, N.; Bejrapha, P.; Srinuanchai, W.; Puttipipatkhachorn, S.; Ruktanonchai, U. Nondestructive Rheological Measurement of Aqueous Dispersions of Solid Lipid Nanoparticles: Effects of Lipid Types and Concentrations on Dispersion Consistency. Drug Dev. Ind. Pharm. 2010, 36, 1005−1015. (27) Schmiele, M.; Schindler, T.; Unruh, T.; Busch, S.; Morhenn, H.; Westermann, M.; Steiniger, F.; Radulescu, A.; Lindner, P.; Schweins, R.; et al. Structural Characterization of the Phospholipid Stabilizer Layer at the Solid−Liquid Interface of Dispersed Triglyceride Nanocrystals with Small-Angle Xray and Neutron Scattering. Phys. Rev. E 2013, 87, 062316. (28) Westesen, K.; Siekmann, B.; Koch, M. H. Investigations on the Physical State of Lipid Nanoparticles by Synchrotron Radiation X-Ray Diffraction. Int. J. Pharm. 1993, 93, 189−199. (29) Helgason, T.; Awad, T.; Kristbergsson, K.; McClements, D.; Weiss, J. Influence of Polymorphic Transformations on Gelation of Tripalmitin Solid Lipid Nanoparticle Suspensions. J. Am. Oil Chem. Soc. 2008, 85, 501−511. (30) van Langevelde, A.; Peschar, R.; Schenk, H. Structure of βTrimyristin and β-Tristearin from High-Resolution X-Ray Powder Diffraction Data. Acta Crystallogr., Sect. B: Struct. Sci. 2001, 57, 372− 377. (31) Nagle, J. F.; Tristram-Nagle, S. Structure of Lipid Bilayers. Biochim. Biophys. Acta, Rev. Biomembr. 2000, 1469, 159−195. (32) Zhang, F.; Ilavsky, J.; Long, G.; Quintana, J.; Allen, A.; Jemian, P. Glassy Carbon as an Absolute Intensity Calibration Standard for Small-Angle Scattering. Metall. Mater. Trans. A 2010, 41, 1151−1158. (33) Unruh, T.; Bunjes, H.; Westesen, K.; Koch, M. H. J. Observation of Size-Dependent Melting in Lipid Nanoparticles. J. Phys. Chem. B 1999, 103, 10373−10377. (34) Bunjes, H.; Koch, M. H. J. Saturated Phospholipids Promote Crystallization but Slow Down Polymorphic Transitions in Triglyceride Nanoparticles. J. Controlled Release 2005, 107, 229−243. (35) Bunjes, H.; Unruh, T. Characterization of Lipid Nanoparticles by Differential Scanning Calorimetry, X-Ray and Neutron Scattering. Adv. Drug Delivery Rev. 2007, 59, 379−402. (36) The program XNDiff is available free of charge from https:// sourceforge.net/projects/xndiff/. (37) Unruh, T.; Bunjes, H.; Westesen, K.; Koch, M. H. J. Investigations on the Melting Behaviour of Triglyceride Nanoparticles. Colloid Polym. Sci. 2001, 279, 398−403. (38) Bunjes, H.; Westesen, K.; Koch, M. H. Crystallization Tendency and Polymorphic Transitions in Triglyceride Nanoparticles. Int. J. Pharm. 1996, 129, 159−173. (39) Bunjes, H.; Steiniger, F.; Richter, W. Visualizing the Structure of Triglyceride Nanoparticles in Different Crystal Modifications. Langmuir 2007, 23, 4005−4011.
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded by the Deutsche Forschungsgemeinschaft (DFG) through the project UN267/2-1 and the Cluster of Excellence Engineering of Advanced Materials. S.G. expresses her thanks to C. Bär for designing the sample holder for the Ganesha camera. We acknowledge the RRZE for using the Lima HPC cluster.
■
REFERENCES
(1) Veerman, J. A. C.; Frenkel, D. Phase behavior of disklike hardcore mesogens. Phys. Rev. A 1992, 45, 5632−5648. (2) Wensink, H.; Lekkerkerker, H. Phase Diagram of Hard Colloidal Platelets: A Theoretical Account. Mol. Phys. 2009, 107, 2111−2118. (3) Marechal, M.; Cuetos, A.; Martínez-Haya, B.; Dijkstra, M. Phase Behavior of Hard Colloidal Platelets Using Free Energy Calculations. J. Chem. Phys. 2011, 134, 094501. (4) van der Kooij, F. M.; Lekkerkerker, H. N. W. Formation of Nematic Liquid Crystals in Suspensions of Hard Colloidal Platelets. J. Phys. Chem. B 1998, 102, 7829−7832. (5) van der Kooij, F.; Kassapidou, K.; Lekkerkerker, H. Liquid Crystal Phase Transitions in Suspensions of Polydisperse Plate-like Particles. Nature 2000, 406, 868−871. (6) van der Beek, D.; Lekkerkerker, H. N. W. Nematic Ordering vs Gelation in Suspensions of Charged Platelets. Europhys. Lett. 2003, 61, 702−707. (7) van der Beek, D.; Lekkerkerker, H. N. W. Liquid Crystal Phases of Charged Colloidal Platelets. Langmuir 2004, 20, 8582−8586. (8) van der Beek, D.; Petukhov, A.; Oversteegen, S.; Vroege, G.; Lekkerkerker, H. Evidence of the Hexagonal Columnar Liquid-Crystal Phase of Hard Colloidal Platelets by High-Resolution SAXS. Eur. Phys. J. E 2005, 16, 253−258. (9) Mourad, M.; Wijnhoven, J.; van’t Zand, D.; van der Beek, D.; Lekkerkerker, H. Gelation versus Liquid Crystal Phase Transitions in Suspensions of Plate-like Particles. Philos. Trans. R. Soc., A 2006, 364, 2807−2816. (10) Mourad, M. C. D.; Petukhov, A. V.; Vroege, G. J.; Lekkerkerker, H. N. W. Lyotropic Hexagonal Columnar Liquid Crystals of Large Colloidal Gibbsite Platelets. Langmuir 2010, 26, 14182−14187. (11) Brown, A. B. D.; Clarke, S. M.; Rennie, A. R. Ordered Phase of Platelike Particles in Concentrated Dispersions. Langmuir 1998, 14, 3129−3132. (12) Verhoeff, A. A.; Bakelaar, I. A.; Otten, R. H. J.; van der Schoot, P.; Lekkerkerker, H. N. W. Tactoids of Plate-Like Particles: Size, Shape, and Director Field. Langmuir 2011, 27, 116−125. (13) Verhoeff, A. A.; Lekkerkerker, H. N. W. Direct Observation of Columnar Liquid Crystal Droplets. Soft Matter 2012, 8, 4865−4868. (14) zur Mühlen, A.; Schwarz, C.; Mehnert, W. Solid Lipid Nanoparticles (SLN) for Controlled Drug DeliveryDrug Release and Release Mechanism. Eur. J. Pharm. Biopharm. 1998, 45, 149−155. (15) Müller, R. H.; Mäder, K.; Gohla, S. Solid Lipid Nanoparticles (SLN) for Controlled Drug DeliveryA Review of the State of the Art. Eur. J. Pharm. Biopharm. 2000, 50, 161−177. (16) Westesen, K. Novel Lipid-Based Colloidal Dispersions as Potential Drug Administration SystemsExpectations and Reality. Colloid Polym. Sci. 2000, 278, 608−618. (17) Wissing, S.; Kayser, O.; Müller, R. Solid Lipid Nanoparticles for Parenteral Drug Delivery. Adv. Drug Delivery Rev. 2004, 56, 1257− 1272. (18) Joshi, M. D.; Müller, R. H. Lipid Nanoparticles for Parenteral Delivery of Actives. Eur. J. Pharm. Biopharm. 2009, 71, 161−172. (19) Bunjes, H. Structural Properties of Solid Lipid Based Colloidal Drug Delivery Systems. Curr. Opin. Colloid Interface Sci. 2011, 16, 405−411. 11395
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
The Journal of Physical Chemistry B
Article
(40) Kuntsche, J.; Horst, J. C.; Bunjes, H. Cryogenic Transmission Electron Microscopy (Cryo-TEM) for Studying the Morphology of Colloidal Drug Delivery Systems. Int. J. Pharm. 2011, 417, 120−137. (41) Illing, A.; Unruh, T. Investigation on the Flow Behavior of Dispersions of Solid Triglyceride Nanoparticles. Int. J. Pharm. 2004, 284, 123−131. (42) Siekmann, B.; Westesen, K. Submicron-Sized Parenteral Carrier Systems Based on Solid Lipids. Pharm. Pharmacol. Lett. 1992, 1, 123− 126. (43) Freitas, C. Stability determination of solid lipid nanoparticles (SLN TM) in aqueous dispersion after addition of electrolyte. J. Microencapsulation 1999, 16, 59−71.
11396
dx.doi.org/10.1021/jp506787v | J. Phys. Chem. B 2014, 118, 11387−11396
