Towards outperforming conventional sensor arrays with fabricated individual photonic vapour sensors inspired by Morpho butterflies

Combining vapour sensors into arrays is an accepted compromise to mitigate poor selectivity of conventional sensors. Here we show individual nanofabricated sensors that not only selectively detect separate vapours in pristine conditions but also quantify these vapours in mixtures, and when blended with a variable moisture background. Our sensor design is inspired by the iridescent nanostructure and gradient surface chemistry of Morpho butterflies and involves physical and chemical design criteria. The physical design involves optical interference and diffraction on the fabricated periodic nanostructures and uses optical loss in the nanostructure to enhance the spectral diversity of reflectance. The chemical design uses spatially controlled nanostructure functionalization. Thus, while quantitation of analytes in the presence of variable backgrounds is challenging for most sensor arrays, we achieve this goal using individual multivariable sensors. These colorimetric sensors can be tuned for numerous vapour sensing scenarios in confined areas or as individual nodes for distributed monitoring.


Supplementary Note 1. Effects of fabrication tolerances
To evaluate performance variability of developed nanostructured multivariable sensors, their fabrication tolerances should be determined. Tolerances related to the physical structure will control the reproducibility of the reflectance spectra, while tolerances related to structure functionalization will control vapour-selectivity. Common methodologies of determination of fabrication tolerances of photonic nanostructures will be implemented. 1-3 Calibration will be performed as a batch-or individual-sensor calibration, dependent on the required sensor accuracy. Calibration will include development of transfer functions of sensor responses for vapours at identified environmental conditions. The determined fabrication tolerances will aid the development of the robust transfer functions.

Supplementary Note 2. Differential reflectance spectral response R()
Existing photonic nanostructured sensors exhibit a significant color change only at relatively high concentrations of detected vapours. 4,5 When the spectral shifts are relatively small, the differential reflectance spectral response R() of a sensor is measured before and after analyte exposure, as 5 where R() and R 0 () are sensor spectra upon exposure to an analyte and a blank carrier gas, respectively. Thus, the common features in the two spectra before and after vapour exposure cancel and the R() spectrum accentuates the subtle differences due to vapour response. 6

Supplementary Note 3. Sensor ability for quantitation of individual vapours in mixtures
Recently, the limit of recognition (LOR) has been introduced 33-35 as a criterion for evaluating sensor arrays and defined as the concentration below which a vapour can no longer be recognized from its response pattern. To take into the account not only vapour recognition but also vapour quantitation in mixtures, two more conservative estimations may be appropriate. Such estimations are the root mean square error of calibration (RMSEC) and the root mean square error of cross-validation (RMSECV) obtained from a multivariate regression model that provides the ability to recognize and quantify analytes in the presence of interferences. 36 For example, multivariate PLS models are a standard tool in quantitation of analytes in mixtures in numerous industrial applications that utilize Process Analytical Technology methodologies. 37 The errors of concentration values of analytes obtained using a multivariate PLS model account for all statistical noise and are very conservative in order to provide robust quantitation.
We have built two PLS models to describe responses of the developed sensors to (1) methanol/propanol and (2) Table 1. In this work, we have not focused to minimize RMSEC or RMSECV values but focused instead on selectivity enhancement of individual sensors. "Discussion" section of the main text outlines our approach toward increasing of sensor signal.

Supplementary Note 4. Vapour selectivity of natural Morpho structures, bio-inspired nanofabricated sensors, and conventional sensor arrays
The first two PCs were indicative of the main dispersion directions 7,8 upon exposure to two individual vapours (see Supplementary Table 2). For methanol and propanol exposures, the contribution of PC 2 for Morpho and nanofabricated structures were 18% and 30%, respectively. For methanol and ethanol exposures, the contribution of PC 2 for bare nanofabricated structure and QCM sensor array were < 1%, while for the MOX sensor array it was 19% and for the FSfunctionalized nanofabricated structure it was 7%. Although the MOX sensor array had the response dispersion larger than that of the FS-functionalized nanofabricated structure, the latter had a better ability to operate at variable humidity without significant non-linearity effects (see Fig. 6 and Supplementary Figs. 16, 19). The prediction ability of the developed sensors to quantify vapours in their mixtures is illustrated in Supplementary Figs. 17