Plasmonic Temperature-Programmed Desorption

Abstract

Temperature-programmed desorption (TPD) allows for the determination of the bonding strength and coverage of molecular mono- or multilayers on a surface and is widely used in surface science. In its traditional form using a mass spectrometric readout, this information is derived indirectly by analysis of resulting desorption peaks. This is problematic because the mass spectrometer signal not only originates from the sample surface but also potentially from other surfaces in the measurement chamber. As a complementary alternative, we introduce plasmonic TPD, which directly measures the surface coverage of molecular species adsorbed on metal nanoparticles at ultrahigh vacuum conditions. Using the examples of methanol and benzene on Au nanoparticle surfaces, the method can resolve all relevant features in the submonolayer and multilayer regimes. Furthermore, it enables the study of two types of nanoparticles simultaneously, which is challenging in a traditional TPD experiment, as we demonstrate specifically for Au and Ag.

Keywords: adsorption; metals; molecules; nanoparticles; plasmonic sensing; temperature-programmed desorption.

Figure 1

(a) SEM image of the nanofabricated Au nanoparticle array (particle ⌀ = 120 nm, h = 20 nm). (b) Extinction spectra obtained from the array at different temperatures (i.e., 100–300 K, 50 K steps). Note that an increase in temperature shifts the LSPR peak to a longer wavelength. (c) Continuous change in the LSPR peak spectral position, Δλ_peak, as a function of temperature. The dashed line is a linear fit to the experimental data.

Figure 2

(a) The difference between λ_peak of a clean and methanol-covered Au nanodisk array, Δλ_peak, for varying initial exposures and thus molecular coverages along a temperature ramp with a linear heating rate of β = 0.33 K/s. The initial increase in the Δλ_peak of 2.0 × 10⁵ L is likely due to small drift in the measurement. A positive correlation between Δλ_peak at 100 K and initial methanol exposure is observed, indicating the formation of molecular layers with different thickness, as depicted in the inset. Dashed line is a guide to the eye. (b) Negative of the first derivative of the data in (a) showing multiple peaks corresponding to desorption of methanol from amorphous and crystalline parts of the methanol multilayer structure, as discussed in the main text. Inset is a zoomed-in image of (b).

Figure 3

(a) The difference between λ_peak of clean and benzene-covered Au nanoparticles, Δλ_peak, with varying initial exposures measured for a temperature ramp with a heating rate of β = 1 K/s. A positive correlation between Δλ_peak at 100 K and initial benzene exposure is observed, indicating the formation of molecular multilayers with different initial thickness. The initial decrease in the Δλ_peak of 6.0 × 10³ L is likely due to small drift in the measurement. (b) Negative of the first derivative of the data in (a). Insets are zoomed-in versions of a certain area in the respective panels.

Figure 4

The Δλ_peak of the Au nanodisk arrays as a function of methanol and benzene molecular coverage in the submonolayer regime. The proportionality observed demonstrates the possibility to deduce the molecular layer thickness with submonolayer sensitivity. The data are inferred from the results shown in Figures 2 and 3.

Figure 5

(a) Top: Schematic of the sample with two chemically different nanoparticle arrays (Au and Ag with different diameters) on the either side of a sapphire substrate. Bottom: SEM images of the corresponding Au and Ag nanoparticles in each array obtained at the same magnification to highlight the size difference. Scale bars are 200 nm. (b) Extinction spectrum of the sample, exhibiting two distinct peaks that correspond to the Au (small particles–short wavelengths) and Ag (large particles–long wavelengths) nanoparticle arrays, respectively. (c) The difference between λ_peak of the clean and methanol-covered Au and Ag surfaces during a temperature ramp at a heating rate of β = 0.17 K/s. (d) Negative of the first derivative of the data in (c). It is clear that the two metals exhibit distinct desorption profiles. Note that the desorption profile of Au is shifted upward for clarity. Inset is a zoomed-in view of (d).