Supplementary Material for
Polymer-Brush Lubrication in Oil: Sliding Beyond the Stribeck Curve
Robert M. Bielecki, Maura Crobu, Nicholas D. Spencer*
ETH Zurich, Department of Materials, Laboratory for Surface Science and Technology, Wolfgang-Pauli-Strasse 10, HCI G517, 8093 Zurich, Switzerland
*Corresponding author.
Phone: +41-44-632-5850 Fax: +41-44-633-1027
E-mail address:
This supplementary material contains X-Ray Photoelectron Spectroscopy (XPS) characterization of the initiator-functionalised surfaces, complementary tribological data illustrating data reproducibility and some additional information concerning the coefficient-of-friction maps (triboscopy).
- XPS characterisation of initiator-functionalised surfaces
The aim of the supplementary XPS measurements was to analyse and compare the chemical composition of the initiator-functionalized surfaces, prior to synthesis of the polymers. By means of XPS it was possible to estimate the packing densities of the initiator layers formed on iron oxide substrates.
1.1.Experimental
The XPS analysis was performed using a VG Theta Probe (Thermo Fisher Scientific, East Grinstead, UK) spectrometer (a detailed description can be found in [1]). The spectra were acquired using a monochromatic Al Kα source with a beam size of 300 μm in the constant-analyzer-energy (CAE) mode. A pass energy (PE) of 100 eV and a step size of 0.1 eV (FWHM for silver Ag 3d5/2 = 0.88 eV) was used to acquire the high-resolution spectra. A PE of 200 eV with a step size of 1 eV was used for the survey spectra.
The angle-resolved mode spectra (16 emission angles) were acquired using a PE of 150 eV (FWHM for silver Ag 3d5/2 = 1.15 eV). The overlayer thickness of the iron oxide samples has been calculated considering the attenuation of the iron Fe 2p3/2 signal at two different emission angles (25˚ and 66˚). The inelastic mean free path, λ, was calculated using the Seah and Dench formula [2].
Two sample types for Si-based substrates were analysed (bare and BPCS-initiator functionalized) and three samples for Fe-bases substrates (bare, with nitrodopamine and with nitrodopamine coupled to an initiator group) were analyzed in standard (3 points per sample) and angle-resolved mode (1 point per sample). Two independent series of samples (number 1 and 2) were analyzed for a total of 10 samples
1.2.Results and discussion
The typical survey spectra of a bare silicon wafer and a silicon wafer functionalised with BPCS initiator are reported in Figure 1. C, O and Si signals were detected on the bare silicon wafers while C, O, Si and Br were present on the silicon wafers functionalised with the BPCS initiator. Traces of Pt were also found on the initiator-functionalised silicon wafers and can be assigned to the remaining Pt catalyst used for the synthesis the BPCS molecules. The Br 3d5/2 signal (see Figure 1) was found at the binding energy (BE) of 70.7±0.2 eV.
Figure 1 XPS survey spectra of a bare silicon wafer and a silicon wafer functionalised with BPCS initiator. The high-resolution spectrum of the Br 3d signal is reported as well, at the top-right of the graph.
Knowing the initiator layer thickness from the ellipsometric measurements (d=1.8 nm), and the molecular weight of this molecule (Mw=413.9 g/mol), a simple equation (Eq. 1) was used to calculate the packing density of the BPCS initiator molecules on the silicon oxide surface σ=2.9 molecules/nm2 (assuming density ρ=1.1 g/cm3).
(Eq. 1)
The bare iron oxide samples showed the presence of Fe, O and C (Figure 2), while Fe, O, C and N were detected in the spectra of the samples functionalised with nitrodopamine. Traces of sulphur on those surfaces can be explained by the usage of nitrodopamine sulphate in surface-functionalisation protocols.
The “iron oxide + catecholamine + initiator” samples showed the presence of Fe, O, C, N, and Br. In both samples, the Br is present in two different oxidation states (the two components of the Br 3d5/2 signal were found at 70.4±0.2 eV assigned to Br bound to a tertiary carbon in the form of an initiator, and 68.8±0.2 eV that could be assigned to bromide salts)
Figure 2: XPS survey spectra of a bare Fe2O3, a nitrodopamine functionalised Fe2O3 and a Fe2O3 functionalised with nitrodopamine and initiator.
From the stoichiometry of the molecule, the ratio of Br to N should be 1:2. However, the measured, corrected intensity ratio calculated from the XPS data appears to be roughly 1:3, indicating a yield of ca 60 % for the coupling reaction of the initiator molecules with the amines present on the substrate.
Further, from the angle-resolved data it was possible to calculate the overlayer thickness of 1.4 nm for the “catecholamine+initator” substrates on the iron oxide samples. Using the same calculation principle as for the BPCS molecules on silicon, and taking the molecular weight of a molecule where each amine is coupled with one molecule of α-bromoisobutyryl bromide – 346 g/mol (and assuming again density of 1.1 g/cm3) the packing density of the nitrodopamine based molecules on iron oxide was estimated to be 2.7 molecules/nm2. This calculation was performed, without taking into account a possible layer of contamination, and assuming a perfect yield of the amide-coupling reaction, both of which may lead to overestimation of the values.
Even if the actual initiator packing density is as low as 2.0 molecules/nm2, the measured differences between the packing density of Br-initiating groups on silicon oxide and iron oxide ensure a massive excess of the initiating groups, consideringtypical polymer grafting densities (approx. 0.2-0.3 chains/nm2, or using only around 10% of initiating groups).
- Complementary data displaying the tribological data reproducibility
Figure 4 As supplementary data to Figure 3 in the main manuscript obtained in PF350 oil,a total of 4 series of experiments displaying data reproducibility in the sliding speed - COF plot.
Figure 5 As supplementary data to Figure 3 in the main manuscript obtained in EO500 oil, a total of 4 series of experiments displaying data reproducibility in the sliding speed - COF plot.
Figure 6 Typical COF evolution with sliding distance, obtained for brush-brush contact at 0.1 cm/s at 20 mN and studied in hexadecane (140 nm dry P12MA thickness, data supplementary to Figure 5 in the manuscript).
Figure 7 Supplementary experiments displaying the coating-stability dependence on the sliding speed. Coating thickness 120 nm, brush-brush configuration. Data obtained in a linear module with reciprocating sliding (distance of 4 mm per cycle), as supplementary material to Figure 5 in the manuscript, in which the curves obtained at higher speeds (0.5, 1 and 5 cm/s, in hexadecane) are more susceptible to differences among samples. In this additional figure a logical trend can be observed, in which the coating stability is higher at higher sliding speeds, supporting the information contained in Figure 5 in the main text (In this case the effect on stability is even more profound since coating failure was achieved for all samples).
- Triboscopy of poly(alkyl methacrylates)
Coefficient-of-friction maps plotted as a function of position along the wear-track arc and sliding distance can be useful to identify the degradation mechanism of the polymer coating. This type of “triboscopy” diagram has been demonstrated by Belin as a useful tool to monitor degradation of thin coatings [3]. The images below display COF values that are an average of the values measured during the clockwise and counterclockwise movement at the given point on the sample.
Figure 8 Acquisition of triboscopic data: a) with the speed varying as a sinusoidal function as f(t) along a sliding arc of radius R and angular range a. At the turning points the speed is 0 and in the center (angle 0°) the speed reaches its maximum. b) The two possible coefficient-of-friction representations: COF plotting against an angular position or against sliding speed.
Figure 9 Triboscopy of a 70 nm thick P12MA coating studied in PF350 oil under 20 mN load and with maximal speed of 0.1 cm/s. The COF during the initial cycles was measured asapprox. 0.08 for the majority of the sliding track length (ca 75 %). Afterwards it decreased to ca 0.03, later increasing slightly to ca 0.04. The arrow indicates time evolution.
Figure 10 Triboscopy of P12MA 140 nm thick coating studied in hexadecane under 20 mN load and with a maximal speed of 0.5 cm/s. Initially coefficient-of-friction values start at a relatively uniform value along the entire length of the sliding track at ca 0.08, followed by steady sliding, with COF values below 0.05 at positions on the wear track where the speed varied between ca 0.5 cm/s and 0.25cm/s, and at slightly higher values over the remaining part of the sliding track. Ultimately the coating fails, reaching COF values above 0.15, preserving only a narrow area in the center of the sliding track (highest sliding speed), where COF remained below 0.1. The arrow indicates time evolution.
References
[1] Mangolini F., Rossi A., and Spencer N. D.: Chemical Reactivity of Triphenyl Phosphorothionate (TPPT) with Iron: An ATR/FT-IR and XPS Investigation. J. Phys. Chem. C 115, 1339-1354 (2011).
[2] M. P. Seah W. A. D.: Quantitative electron spectroscopy of surfaces: A standard data base for electron inelastic mean free paths in solids. Surf. Interface Anal. 1, 2-11 (1979).
[3] Belin, M., Lopez, J., Martin, J.M.: Triboscopy, a quantitative tool for the study of the wear of a coated material. Surface and Coatings Technology 70(1), 27-31 (1994).