A promising azeotrope-like mosquito repellent blend

Topical repellents play a key role in reducing the outdoor transmission of mosquito-borne diseases by reducing human-vector contact. Excellent repellents are available, but there is always room for improvement. This article reports on a particularly effective binary repellent blend of ethyl butylacetylaminopropionate and nonanoic acid. A composition containing 25 mol% of the acid exhibits negative pseudo-azeotrope behaviour at 50 °C, meaning that the liquid vapour pressure is lower than that of the parent compounds and evaporation occurs without a change in the liquid composition. In tests performed using the South African Medical Research Council’s cup-on-arm procedure, this mixture provided better protection for a longer time than the “gold standard of mosquito repellents”, namely N,N-diethyl-m-toluamide, commonly known as DEET.


SI. 1. The physical concept behind formation of pseudo-azeotropes
Raoult's law states that the equilibrium vapour pressure of an ideal liquid vaporising as an ideal gas is simply the mole fraction weighted mean of the vapour pressures of the neat components. i.e. When the activity coefficients assume values below unity, this equation predicts that the vapour pressure of the mixture will be less than the value predicted by Raoult's law. That is generally the case when the strength of the attractive interactions between the unlike molecules exceed those between the like molecules in the liquid mixture. When these attractive interactions between the unlike molecules are sufficiently strong, the predicted liquid vapour pressure curve may even dip below the vapour pressures of the two parent compounds. The location of the resulting minimum in the vapour pressure curve represents a negative pseudo-azeotrope composition since the observed vapour pressure will be below the value predicted by Raoult's law.

SI. 2. Description of the spectral features of nonanoic acid and IR3535 FTIR spectra
The spectrum for the nonanoic acid has, among others, the following distinguishing features.
The strongest peak, located at 1706 cm −1 , arises from the carbonyl stretch vibration. This peak is actually a composite of several overlapping absorptions contributed by the carboxylic acid moiety. In order from high to lower wavenumbers, it includes contributions from the terminal groups of linear dimers and polymers, monomers, the cyclic dimer and the inner groups of linear dimers/oligomers 1 . A very broad peak for the nonanoic acid is centred approximately at 3000 cm −1 . It is superimposed on the C−H stretch bands, but it is much broader, having a

SI. 3. The spectral residuals for IR3535-nonanoic acid binary mixtures
Spectral residual amplifies differences arising from intermolecular interactions. ∆Amix is defined by: where Amix is the absorbance measured for the actual mixture with compositions corresponding to mole fractions xA and xC; the subscripts A and C denoting IR3535 and nonanoic acid respectively. If the interactions the molecules experience in the mixture were identically the same as in the neat liquids, then ∆Amix would be exactly zero. However, this is not the case here. Some bands are characteristic of functional groups present on only one of the constituents. Others arise from the interactions between the unlike molecules. In the former case, negative values for ∆Amix associated with a given band indicates that, compared to the situation in the neat liquid, the effect of the molecular vibration is less than expected.

SI. 4. The relative absorptions band intensities for nonanoic acid-IR3535 binary mixtures
To calculate the intensity of absorption relative to the fraction of molecules present in the mixture, it is more helpful to consider the relative changes in the peak intensities in the following way: These variables should assume values of unity if the absorption band characteristic for the given compound under investigation is not affected by the presence of the other compound.

SI. 5. Deconvolution and curve-fitting of the peak related to H(OH)-O(OH) in partial radial distribution function of nonanoic acid
The peak is deconvoluted to Curve 1 (3.3 Å), Curve 2 (3.9 Å), and Curve 3 (4.6 Å) and curvefitted on Fit Sum. The first peak corresponds to the cyclic dimer while the second peak is attributed to the presence of higher order aggregates in the liquid 6 . Partial radial distribution of H(OH)-O(OH) reveals an interesting aspect of the hydrogen bonding in nonanoic acid.
Although there are two oxygen atoms capable of forming hydrogen bonds in the nonanoic acid molecule 6 , only the carbonyl oxygen is involved. This is reflected in absence of any major peak at short distances, around 1.8 Å, i.e. where hydrogen bonds usually peak.   SI. 9. Added parameters to AUA forcefield

SI. 10. Regression of the oven evaporation test and repellence data
Non-linear regression, using the "nls" function in R 19 was used for fitting the models in this section. For the oven evaporation test (Table SI.10.1), the Avrami model was fitted to the data obtained for starting IR3535 concentrations of 40 mol% and higher: where x() is the concentration (mol%) of IR3535 in the remaining liquid at the point where a portion  of it has evaporated; xazeo is the pseudo azeotrope composition; xo is the initial concentration IR3535 in the liquid; λ and n are model parameters. This model was explicitly coded with initial values corresponding to manual guesses. The logistic function was fitted to the repellence data (Table SI. where the proportion refers to the protection rendered by the repellent in the test at time t. The parameter S can be associated with an estimate for the highest protection possible c is a shift parameter and τ provides a rough measure of the maximum time duration of protection.
In both cases, the "nlsBoot" function from the "nlstools" package 20 was used in order to obtain non-parametric bootstrap confidence intervals for all model parameters 21 . The bootstrap sample was set to 5,000 in all cases, but may effectively be less because of incomputable sample combinations.