New Approaches to Stretched Film Sample Alignment and Data Collection for Vibrational Linear Dichroism

Rapid measurements of vibrational linear dichroism (VLD) infrared spectra are shown to be possible by using stretched polymer films and an extension of existing instrumentation designed for vibrational circular dichroism spectroscopy. Earlier techniques can be extended using additional inexpensive polymer substrates to record good-quality VLD spectra of a significantly wider range of compounds with comparatively short sample-preparation times. The polymer substrates used, polyethylene and polytetrafluoroethylene, are commonly available and inexpensive, and samples are more easily prepared than that for many earlier stretched-film and crystal studies. Data are presented for neutral hydrophobic organic molecules on hydrophobic films including acridine, anthracene, fluorene, and recently synthesized S-(4-((4-cyanophenyl)ethynyl)phenyl)ethanethioate. We extend the approach to polar or ionic species, including 2,2′-bipyridine, 1,10-phenanthroline, and sodium dodecyl sulfate, by oxidizing polyethylene films to change their wetting properties. The combination of new instrumentation and modified sample preparation methods is useful in basic spectroscopy for untangling and assigning complicated infrared spectra. Nevertheless, it is not a panacea as surface-adsorbed molecules are often not monodispersed, and higher analyte concentrations can lead to aggregation and resonance phenomena that have previously been observed for infrared spectra on surfaces. These effects can be assessed by varying the sample concentration. The focus of this paper is experimental, and detailed analysis of most of the spectra lies outside its scope, including some well-studied compounds such as acridine and anthracene that allow comparisons with earlier research.


Procedure for baseline correction of infrared absorption and VLD spectra
Subtraction of polymer baseline spectra often did not produce analyte spectra with perfectly flat baselines, especially at the sensitivities that were possible using the Jasco FVS-6000 spectrometer.
Deviations were attributable to slight mismatches in sample placement and sinusoidal features, discussed below, arising from interference fringes caused by the interaction between the IR beam and the thin polymer films.The Spectra Manager software version 2.07.02 allows spectra to be added, subtracted, and scaled, and for baselines to be corrected by fitting straight lines or splines to different wavenumber segments across the spectrum.Absorbance spectra were routinely flattened in this way when the sharpness of the bands made the location of the baseline clear.The situation was more complicated for broader spectra, including large biomolecules, and correction relied on the comparison of replicate spectra (to account for residual interference fringes) and surface abrasion of the films to minimize the size of the fringes.Other software packages could be used for these purposes, as well as other protocols for baseline flattening.
Due to the baseline correction options only being available for the top channel and the VLD spectra being in the second channel for the Spectra Manager software, the VLD spectra were S3 flattened by exchanging the absorbance and VLD channels and using the baseline correction function on the VLD spectra, as summarised below.The flattened absorbance and VLD spectra were recombined into a single data file by exporting them to separate Microsoft Excel files, combining the data, and importing the resulting single file back into the Spectra Manager software.
Residual PE, PE OX and PTFE bands were removed from the spectra by deleting the relevant absorbance and VLD data points in the Excel files.These regions have been annotated in the Figures to show where the polymer bands have been removed.
The baseline correction software used in this study requires the operator to manually "pin" the correction line (or spline, if this option is chosen) to a series of points on the spectrum, corresponding to points where the location of the baseline is unambiguous.This works best for spectra with little or no undulation due to interference fringes, and sharp bands with good baseline separation.The greater the amount of undulation, the larger the number of "pinned" points that are needed to give a flat baseline between the absorption or VLD bands.See Figure 4 in the main text for an example of baseline correction for a spectrum with sharp spectral features superimposed on an undulating baseline.
Despite careful purging with dry nitrogen, some baseline-corrected spectra contained weak water-vapour rovibrational bands, especially for data collected on humid days, readily identified in the 1 cm -1 spectral bandwidth spectra.These were minimized by adding or subtracting small multiples of a standard water-vapour spectrum.

PE and PTFE spectra
When initially mounted in the film stretcher unstretched, and subsequently stretched by an additional factor of two, PE, PE OX , PE PnS and PE PnS,OX films (the longer dimension of the PE PnS roll is the manufacturing stretch direction) gave negative LD bands at 1463 and 1471 cm −1 .
(Figure SI1) These corresponded to C−H scissoring, 1 consistent with earlier studies. 2The signs of the bands are consistent with the oriented parts of the PE films having the C−C bonds aligned with the (vertical) stretch direction and C−H bonds perpendicular to it.Smaller positive and negative bands could be observed for other transitions.The LD r values of the strong negative bands for the Glad Press'n Seal film are typically about −0.7 indicating that the average orientation parameter of the polymers is S ~ 0.5 (assuming  = 90 o ).As Figure SI1 indicates, the Glad Snap Lock PE is less oriented, with S ~ 0.25.Regrettably, the PE PnS scatters and/or depolarizes UV-visible light so we could not record corresponding pairs of ELD and VLD spectra using this film.
The PTFE film has absorbance greater than 1.1 for the 1200 cm -1 polymer band and hence provides unreliable maximum PTFE LD magnitudes.The PTFE films had typical LD r values of −0.26 (1230 cm -1 ) and −0.32 (1158 cm -1 ), corresponding to the C−F bond stretches being perpendicular to the stretch direction, consistent with the spectra of shear-deposited PTFE reported by Ji et al.. 3 This gives an average orientation parameter for the polymers of S ~ 0.1 (using  = 90 o ).sh denotes shoulder; vw denotes very weak; (+) and (-) denote weak positive and negative LD; and (?) denotes that the observed band is too weak for the sign of the VLD to be discerned.

Fig. SI1 .
Fig. SI1.Absorption (a) and VLD (b) spectra of 2 stretched Glad Snap Lock PE and PE PnS,OX , and 1.55 stretched PTFE films.Baselines have not been flattened, so that interference fringes (PE) and vertical shifts due to scattering are visible.

Fig. SI2 .
Fig. SI2.Absorption (a) and VLD (b) spectra of acridine absorbed into 2 stretched Glad Snap-Lock PE; adsorbed onto PE PnS,OX ; and adsorbed onto 1.9 stretched PTFE.Polymer spectra have been subtracted; baselines flattened; and residual water, PE and PTFE bands removed.Spectra have been displaced vertically for clarity.

Note: 3 .
Absorbances showed a significant amount of inter-sample variation, depending on sample preparation.Stronger bands are highlighted in bold text.L = long in-plane axis (Long); M = short in-plane axis (Median); N = out-of-plane axis (Normal) For L-axis polarized transitions (positive VLD)  =   For M-and N-polarized transitions (negative VLD)  = −2  3 .

Table S1 .
Band wavenumbers, reduced LD (LD r ) and orientation parameters (S) of VLD spectra of acridine absorbed in 2 stretched polyethylene films.