Concentration gradients in evaporating binary droplets probed by spatially resolved Raman and NMR spectroscopy

Abstract

Understanding the evaporation process of binary sessile droplets is essential for optimizing various technical processes, such as inkjet printing or heat transfer. Liquid mixtures whose evaporation and wetting properties may differ significantly from those of pure liquids are particularly interesting. Concentration gradients may occur in these binary droplets. The challenge is to measure concentration gradients without affecting the evaporation process. Here, spectroscopic methods with spatial resolution can discriminate between the components of a liquid mixture. We show that confocal Raman microscopy and spatially resolved NMR spectroscopy can be used as complementary methods to measure concentration gradients in evaporating 1-butanol/1-hexanol droplets on a hydrophobic surface. Deuterating one of the liquids allows analysis of the local composition through the comparison of the intensities of the C–H and C–D stretching bands in Raman spectra. Thus, a concentration gradient in the evaporating droplet was established. Spatially resolved NMR spectroscopy revealed the composition at different positions of a droplet evaporating in the NMR tube, an environment in which air exchange is less pronounced. While not being perfectly comparable, both methods—confocal Raman and spatially resolved NMR experiments—show the presence of a vertical concentration gradient as 1-butanol/1-hexanol droplets evaporate.

Keywords: NMR; Raman; concentration gradient; evaporation; sessile droplet.

Fig. 1.

(A) Raman spectra of 1-butanol-d₉ (red spectrum), 1-hexanol (blue spectrum), and a mixture of both substances in a ratio of 50:50 mol % (violet spectrum). The spectrum of the mixture shows a superposition of both pure components. The peaks used to calculate the concentration are 2,000 to 2,400 cm⁻¹ (C–D, 1-butanol-d₉) and 2,800 to 3,000 cm⁻¹ (C–H, 1-hexanol). (B) Sketch of the Raman experimental setup with microscope objective and table, substrate, laser, binary liquid mixture droplet and a Raman 3D scan. (C) Raman 2D depth scan with axes. In every pixel a spectrum (as in A) was taken and filtered. Many of these 2D scans taken after each other were combined in temporal sequences in D for the 1-butanol-d₉ and in E for the 1-hexanol signals. (F) Calculated concentration profiles for each scan. Error bars are shown for one measurement are representative and similar for the other data points. (G) The 1-butanol-d₉ concentration gradient over experimental data at a position slightly off-center of the droplet as sketched in B.

Fig. 2.

(A) Chemical structures of 1-butanol and 1-hexanol. CH₂ resonances are highlighted. (B) ¹H NMR spectra (400 MHz, 300 K) of pure 1-butanol, a 50 mol % mixture of 1-butanol and 1-hexanol, and pure 1-hexanol measured as isotropic bulk samples in 5-mm NMR tubes. The NMR spectra were acquired with a single scan and 64,000 data points over a sweep width of 20 ppm within an acquisition time of 4.089 s. For all three spectra, the integrals of the hydroxy groups (-OH), methylene groups near the oxygen (O-CH₂), and methyl groups (-CH₃) show the same value (gray boxes). For pure 1-butanol, for the remaining methylene groups (-CH₂, green box), a signal integral of four is expected and obtained. For pure 1-hexanol, the remaining CH₂ groups have a signal integral of eight (red box). Hence, for a 50:50 mol % mixture of these two alcohols, a signal integral of six (blue box) is observed. This indicates the expected linear behavior with concentration.

Fig. 3.

(A) Sagittal RARE (40) VTR (Rapid Acquisition with Relaxation Enhancement with variable repetition time TR) image and (B) axial RARE (Rapid Acquisition with Relaxation Enhancement) image (40) of a fresh 1-butanol/1-hexanol droplet on a PFDTS surface. Both images are acquired at 400-MHz proton resonance frequency (9.4 T). The positions of the voxels used for PRESS are highlighted with colored boxes. Both images were acquired with a 128 × 128 matrix with a field of view of 4.5 × 4.5 mm and a slice thickness of 1 mm. For the sagittal image, a single scan with an echo time (TE) of 20 ms, a repetition time (TR) of 1,000 ms, and a RARE factor of 8 was used. For the axial image, two scans with TE = 20.83 ms and TR = 2,000 ms and a RARE factor of 8 were accumulated.

Fig. 4.

Fraction of 1-butanol in four different voxels (0.2 × 0.2 × 0.2 mm) during evaporation over time t at 298.5 K. The voxel positions are indicated within the drawing. The PRESS NMR spectra were acquired with 64 scans, a repetition time of 4,000 ms and an echo time TE of 20 ms. For each free induction decay, 8,000 data points were acquired within an acquisition time of 998.4 ms. Bandwidth of the pulses was set to 5,400 Hz. The orange voxel is shifted downward compared to the green voxel to avoid a moving liquid/gas interphase within the voxel. The 1-butanol fraction is calculated by comparing the signal integrals of CH₂ and CH₃ resonances.