Rational Design of Sustainable Liquid Microcapsules for Spontaneous Fragrance Encapsulation

Abstract The high volatility, water‐immiscibility, and light/oxygen‐sensitivity of most aroma compounds represent a challenge to their incorporation in liquid consumer products. Current encapsulation methods entail the use of petroleum‐based materials, initiators, and crosslinkers as well as mixing, heating, and purification steps. Hence, more efficient and eco‐friendly approaches to encapsulation must be sought. Herein, we propose a simple method by making use of a pre‐formed amphiphilic polymer and employing the Hansen Solubility Parameters approach to determine which fragrances could be encapsulated by spontaneous coacervation in water. The coacervates do not precipitate as solids but they remain suspended as colloidally stable liquid microcapsules, as demonstrated by fluorescence correlation spectroscopy. The effective encapsulation of fragrance is proven through confocal Raman spectroscopy, while the structure of the capsules is investigated by means of cryo FIB/SEM, confocal laser scanning microscopy, and small‐angle X‐ray scattering.


Hansen Solubility Parameters (HSP)
Approximately 500 mg of PEG-g-PVAc was weighed in glass vials, and solvent was poured to obtain 10 %wt solutions. The vials were left in an orbital shaker at 25 °C for 24 h. A solvent was deemed "good" (score = 1) if capable of dissolving the polymer completely and "bad" (score = 0) if the polymer remained visibly undissolved. An intermediate score (= 2) was assigned to solvents that yielded cloudy solutions, such as butanol and ethanol. The list of solvents used, along with their HSP, is given in Table S1. All data S2 are part of the database provided with the software used in this work, HSPiP (Hansen Solubility Parameters in Practice, © 2008-20 Steven Abbott and Hiroshi Yamamoto, www.hansen-solubility.com). The solubility parameter "distance" Ra between two materials, based on their respective partial solubility parameter components, is given by: 3 This equation is used to calculate the best sphere encompassing the good solvents and leaving out the bad ones (see main text). For each solvent, the RED (= Ra / R0, where R0 is the interaction radius) number is also given in Table S1 as a measure of the fit quality: good solvents have RED < 1 (they are inside the sphere), while bad solvents have RED > 1 (they are outside the sphere).

Phase diagrams
PEG-g-PVAc/perfume/water ternary phase diagrams were constructed by weighing the appropriate amounts of water, polymer, and perfume in a glass vial using an analytical balance (Radwag AS R2, accuracy ± 0.1 mg); the polymer was first molten at 50 °C for ease of manipulation. Samples were vortexed until homogenization and stabilized at 25 °C in an oven for 14 d.

Sample shown in movie for Supporting Information
A movie is provided as Supporting Information to demonstrate the spontaneous formation of the microcapsules. The clip shows a sample that was prepared as follows: a 4-mL clean glass vial was placed in a holding support to immobilize it; 90 mg of molten PEG-g-PVAc was added to the glass vial, and the where N is the average number of fluorescent molecules detected inside the confocal volume ( = , with = #/! % # and the concentration), & is the decay time, and = % / % is the ratio between the axial and the lateral dimensions of the confocal volume, determined through the calibration procedure with Alexa 568. The diffusion coefficient D of the fluorescent molecules can be obtained from the relationship: Confocal Raman microscopy S5 Raman analysis and mapping were performed on a Renishaw InvIa Qontor confocal MicroRaman system equipped with 785 nm (solid state type, IPS R-type NIR785, 100 mW, 1200 l/mm grating) and 532 nm (Nd:YAG solid state type, 50 mW, 1800 l/mm grating) lasers, front-illuminated CCD camera (256 × 1024 px, working temperature −70 °C) and a research-grade Leica DM 2700 microscope.
References for pure compounds were collected using the 785 nm excitation wavelength for PEG-g-PVAc and 532 nm excitation wavelength for the fragrances; Raman spectra were recorded in the wavenumber range from 100 to 3500 cm -1 using the extended range mode. Bidimensional maps were acquired using a long working distance 50× objective in high-confocality and static spectral range modes. Spectra were acquired with steps of 0.5 to 1 µm (depending on the sample) along the x-y plane. Typical acquisition times per point were 1 or 2 s, acquiring a single scan. Raw data were processed using Renishaw software WiRE v.5.2 for maps generation.

Cryogenic Focused Ion Beam (FIB) Scanning Electron Microscopy (SEM) and image analysis
Cryo FIB/SEM analysis was carried out using a Zeiss Crossbeam 550. Samples were placed in 3-mm carriers and frozen using a Leica HPM100 high-pressure freezer. These samples were then mounted under liquid nitrogen and transferred into a Quorum Technologies PP3010 cryo preparation chamber under vacuum. A platinum coating ~25 nm thick was deposited using the sputter coater built into the cryo preparation chamber. Cross-sections were prepared by milling a rough trench with the beam at 30 kV and 7 nA before polishing with a reduced beam current of 700 pA. SEM images were acquired using an accelerating voltage of 2 kV and a beam current of 75 pA with the in-lens detector. Cross-sectional images were acquired using the SEM at an angle of 54°. Images were made clearer through Fast Fourier Transform filtering to remove some curtaining and charging, followed by normalizing local contrast across 40 pixels with a standard deviation of 3. This image processing was done using FIJI (ImageJ). 5 The same software was used to measure interlamellar spacing in the FIB/SEM image shown in the main text (Fig. 6F). Namely,  VS is the total volume of the particle, and RS = Rc + t is the total radius of the particle, with t the shell thickness.
A Shultz distribution of radii, to account for particles polydispersity, was used: The interaction between polymer particles in the multicompartment capsules was interpreted according to a hard-sphere structure factor 7 : being ( ) a function of the following type: H' the hard-sphere radius and H' the hard spheres volume fraction. The potential can be described as follows: In the fitting procedure, the SLD of the linalool and water were fixed.
where I and I0 are the signal intensities respectively in the presence and absence of the applied field gradient, q = γgδ is the so-called scattering vector (γ being the gyromagnetic ratio of the observed nucleus), t = (Δ -δ/3) is the diffusion time, and Δ is the delay time between the encoding and decoding gradients.
Errors were estimated lower than 2% on the basis of repeated measurements. Figure S1. Raman spectrum of pure PEG-g-PVAc. Figure S2. Raman spectrum of pure L-carvone. S10 Figure S3. Raman spectrum of pure linalool. Figure S4. Raman spectrum of the mixture 83% anisaldehyde / 17% α-pinene. Figure S5. Representative Raman spectrum taken inside a capsule in the PEG-g-PVAc/L-carvone/water system. Red squares indicate peaks used in 2D mapping. Figure S6. Representative Raman spectrum taken inside a capsule in the PEG-g-PVAc/linalool/water system. Red squares indicate peaks used in 2D mapping. Figure S7. Representative Raman spectrum taken inside a capsule in the PEG-g-PVAc/anisaldehyde/α-pinene/water system. Red squares indicate peaks used in 2D mapping. Figure S8. CLSM images of a representative sample taken in the droplet region of the PEG-g-PVAc/isoeugenol/water system. The red signal originates from the RBITC-labelled polymer, excited with a 561 nm laser line.   The sample containing capsules (70% H2O and 18% linalool, indicated by a red circle) exhibits a main peak that can be related to the first broad peak observed in the two more structured samples. The samples containing linalool (55% H2O 18% linalool, blue square) and isoeugenol (55% H2O and 13.5% isoeugenol, green triangle) show a similar broad peak at low Q. This suggests that structured samples contain at least another phase, probably consisting of polymer capsules.    Fig. S12). Water increases from 55 %wt to 60 %wt and 65 %wt. The broad peak in the low-Q region is not significantly altered by dilution but shifts to lower Q values, suggesting the swelling of the capsules present in the samples. The structural peaks become less intense with increasing dilution, suggesting the partial disruption of ordered areas.  Using the data of Table S3, the following results were obtained: