Calcite Surfaces Modified with Carboxylic Acids (C2 to C18): Layer Organization, Wettability, Stability, and Molecular Structural Properties

A fundamental understanding of the interactions between mineral surfaces and amphiphilic surface modification agents is needed for better control over the production and uses of mineral fillers. Here, we controlled the carboxylic acid layer formation conditions on calcite surfaces with high precision via vapor deposition. The properties of the resulting carboxylic acid layers were analyzed using surface-sensitive techniques, such as atomic force microscopy (AFM), contact angle measurements, angle resolved X-ray photoelectron spectroscopy (XPS), and vibrational sum-frequency spectroscopy. A low wettability was achieved with long hydrocarbon chain carboxylic acids such as stearic acid. The stearic acid layer formed by vapor deposition is initially patchy, but with increasing vapor exposure time, the patches grow and condense into a homogeneous layer with a thickness close to that expected for a monolayer as evaluated by AFM and XPS. The build-up process of the layer occurs more rapidly at higher temperatures due to the higher vapor pressure. The stability of the deposited fatty acid layer in the presence of a water droplet increases with the chain length and packing density in the adsorbed layer. Vibrational sum frequency spectroscopy data demonstrate that the stearic acid monolayers on calcite have their alkyl chains in an all-trans conformation and are anisotropically distributed on the plane of the surface, forming epitaxial monolayers. Vibrational spectra also show that the stearic acid molecules interact with the calcite surface through the carboxylic acid headgroup in both its protonated and deprotonated forms. The results presented provide new molecular insights into the properties of adsorbed carboxylic acid layers on calcite.


Vapour pressure and aqueous solubility of carboxylic acids
Surface modification via carboxylic acid vapour exposure is efficient only at high enough vapour pressure.2] For the latter case, a third-order polynomial fit of the form shown in Eq.S1 was used: where p is the vapour pressure,  the temperature (K), and A, B, C and D fitted constants summarized in Table S1. 1 The vapour pressures of the different fatty acids used under our experimental conditions are presented in Table S2.
Table S1.Coefficients of Eq.S1 for stearic acid (valid up to the boiling temperature: 648.1 K) Stearic acid 5.17E8 -5.69E6 5.47E3 3.27 Table S2.Vapour pressure for water and carboxylic acids (C 2 -C 8 3 , C 12 4 , C 18 1 ).Liquid state of the media is marked as ( liq ).The melting point and aqueous solubility of the carboxylic acids used in the present work are provided in Table S3.

Table S3
Normal melting point of carboxylic acids and their aqueous solubility in grams per kilogram of water at 21-23 °C.The mark ∞ means that the carboxylic acid is completely miscible with water 3 .As compared to immediately after preparation (Figure 3 in the main text), the surface is less homogeneous.

Figure S3
. AFM topography and nanomechanical mapping images of modified calcite surface exposed for 4 h in C 8 vapour at room temperature (⁓25 °C) and stored overnight in ambient air.
Deformation images recorded at the same time as the topography and adhesion images reported in Figure 5 of the main text are reported in Figure S4: S6

XPS calculations
The relative amounts of carbon species with different bonds to oxygen were determined from the highresolution carbon C 1s spectra by deconvolution into five different carbon peaks using the Gaussian -Lorentzian peak shape model.Clearly, as also shown in Figure 6a, the C4-and C5-carbon peaks overlap and are not readily resolved into two separate peaks.For this reason, the theoretical atomic ratio Ca / C5-carbon = 1 in calcite was used to calculate the size of the C5-carbon peak.The remaining part in the broad C4 -C5 carbon peak is then the smaller C4 carbon peak.
The adsorbed layer, overlayer, thickness  , was calculated under the assumption that a uniform stearic acid layer is present on a homogeneous calcite substrate, as for the substate-overlayer model presented in Figure S5.In this case, the photoelectron peak intensity from element a in the substrate,  , (area of the peak after background subtraction, see Figure 6 in the main text) is given by 6 : where  is the empirically derived atomic sensitivity factor (also known as relative sensitivity factor RSF) for element a,  the volume atomic density of element a,  the photoelectron inelastic mean free path (IMFP) in the stearic acid overlayer for photoelectrons emitted from element a in the substrate, and  the photoelectron take-off angle.The reduced thickness,  / , was determined from Eq. S2 by plotting the natural logarithm of the substrate signal intensity  (where  is the substrate signal Ca 2p) as a function of 1/sin.The data was fitted using linear regression and  / was obtained from the slopes.Thus, when  is known, the overlayer thickness can be determined.Here,  was calculated using the procedure developed for organic materials by Cumpson. 7][10] The method was later extended to other substrates, such as cellulose, 8 gold, 11 and silica. 6n our analysis, we instead utilize the known bulk chemical composition of calcite as reference, following the procedure described earlier and used for studying adsorption to silica surfaces. 6Using this approach, the number of C1-carbon atoms from the stearic acid  (atoms/nm 3 ) on the calcite surface was calculated from Eq. S3: where I d is the photoelectron peak intensity from element d in the overlayer (the C1-carbon contribution to the C 1s peak),  the relative sensitivity factor for element d,  the IMFP in the stearic acid overlayer for photoelectrons emitted from element d in the overlayer (calculated according to Cumpson).The volume atomic density of the substrate element,  , was calculated from the structural formula of calcite, CaCO 3 , with molecular weight 100.09g/mol, Avogadro's number and the density of calcite (2.71 g/cm 3 ). 3The calculated theoretical atomic densities for calcite were 16.3 Ca/nm 3 , 16.3 C/nm 3 and 48.9 O/nm 3 .In calcite, the theoretical atomic ratio O/Ca = 3, and after adjusting for oxygen present in the organic C2-C4-carbon peaks, the experimentally determined O/Ca atomic ratios were found to be close to the theoretical ratio.Therefore, we used the theoretical atomic density of 16.3 Ca/nm 3 in the calculations of adsorbed amount according to Eq. S3.
For quantification of the adsorbed amount of stearic acid expressed as stearic acid molecules/nm 2 the number of carbon atoms  (atoms/nm 3 ) was divided by 17, the number of C1-carbon atoms in one stearic acid molecule, and then multiplied with the layer thickness determined using Eq.S2.In the calculations, we considered two assumptions: i) none of the C1-carbon (aliphatic carbon) found on freshly cleaved calcite was displaced by stearic acid, and ii) all C1-carbon found on the freshly cleaved calcite surface was displaced by stearic acid.

XPS: Adsorption to calcite in ultra-high vacuum XPS conditions
We investigated the chemical stability of the freshly cleaved calcite surface by keeping it for almost 24 h in ultra-high vacuum conditions (pressure below 1.33 ꞏ 10 -5 Pa).The organic carbon on the surface increased by about 3.0 atomic % after 3 h under X-ray irradiation and a further increase by 1.5 atomic% after 24 h in high vacuum with X-ray irradiation.We can thus conclude that even under high vacuum, organic molecules physisorbed to the calcite surface, but the adsorption increase is small and occurs at a low rate.

XPS: Effects of X-ray irradiation
The possible damaging effect of X-ray irradiation on the stearic acid layer during long exposure times was evaluated (Figure S6) using calcite surfaces exposed to stearic acid vapour for 4 h.Here, we considered changes in the atomic ratio of the C 1s (total carbon) to Ca 2p elemental signals.The C 1s / Ca 2p atomic ratio decreased with X-ray exposure time, see Figure S6.This suggests fragmentation and/or desorption of the stearic acid layer either directly by X-ray radiation, by interactions with photoelectrons or due to heating of the sample surface and sample holder, which both could induce desorption.Heating of the sample holder has been noted in other studies after long X-ray irradiation times. 12 a result of continuous X-ray irradiation, the signal from the stearic acid layer decreased, whereas that from the calcite substrate increased.We note, however, that all XPS measurements for quantification of layer thickness and adsorbed amount were done after short X-ray irradiation exposure times (below 10 minutes, where the carbon to calcium ratio decreased just to about 95 % of the starting value, that is, the atomic ratio of C (total carbon) / Ca decreased from 4.21 at 5 min to 4.01 at 10 min).Any underestimation of thickness and adsorbed amount due to degradation is, therefore small.

Vibrational sum frequency spectroscopy.
Sum frequency spectra of stearic acid adsorbed on the calcite surface after a 24 h exposure for all azimuthal angles measured under the SSP polarization are presented in Figure S7.The spectra were subsequently fitted using a convolution of Lorentzian and Gaussian line shapes of the form presented in Eq.S4.The Lorentzian and Gaussian components account for the homogeneous and inhomogeneous line broadening, respectively. 13 where  represents the non-resonant contribution,  is the amplitude of v th resonant mode,  is the IR frequency and  ,  and  are the peak position, Lorentzian width and Gaussian line width, respectively.The obtained fitting parameters for the most relevant bands at the different azimuthal angles are listed in Table S4.Note that the Lorentzian line widths were constrained to 2 cm -1 , while  were allowed to vary between 1.8 -3 cm -   S8b.In an isotropic planar surface, the intensity in the latter polarization is expected to be zero. 14However, the fact that peaks are detected provides additional proof that the stearic acid monolayer on calcite is anisotropic in the plane.

Figure S8.
Vibrational sum frequency spectra of a stearic acid monolayer adsorbed (24 h) on calcite, measured at different polarization combinations for two selected azimuthal angles ().To help visualize the azimuthal angle dependence for the different vibrational modes detected in the CH stretching region, the patterns presented in Figure 8 of the main paper were fitted according to Eq. S5.
The fitted values are shown in Table S5.6][17] These will be presented in a separate study.
S9 shows the sum frequency spectrum of a freshly cleaved calcite surface in the absence of stearic acid.The sharp feature observed in the spectra at 1432 cm -1 unambiguously stems from the calcite surface and is assigned to the antisymmetric carbonate stretch ( 3 calcite ).

*
Estimated for temperatures outside the valid range of the Antoine equation

Figures
Figures S1 and S2show deformation images of bare calcite and calcite modified by carboxylic acids, respectively.These images are recorded at the same time as the topography and adhesion images reported in Figure2and Figure3, respectively, in the main text.

Figure S1 .
Figure S1.AFM deformation images of an uncoated freshly cleaved calcite surface (similar behaviour to that of a freshly cleaved calcite heated for 20 min at 85 °C.

Figure S2 .
Figure S2.AFM deformation images of modified calcite surface exposed for 4 h in (a) octanoic acid vapour at room temperature (25 °C), and (b) stearic acid vapour at 105 °C

Figure S4 .
Figure S4.AFM deformation images of stearic acid organization on calcite surfaces at different exposure temperatures and times, as well as their stability over longer exposure to ambient air (measured a few days after deposition).Stearic acid modified calcite surface deposited for (a) 10 min at 85 °C, (b) 10 min at 105 °C, (c) 20 min at 85 °C, and (d) 20 min at 105 °C.

Figure
Figure S5.(a) Schematic diagram of the substate-overlayer model with photoelectrons e -at take-off angle  from the sample surface while assuming a homogeneous and fully covering stearic acid overlayer d with thickness t o on a homogeneous calcite substrate surface a.(b) The dependence of the natural logarithm of the peak intensity for the substrate signal (I) as a function of the photoelectron take-off angle (θ), where θ was 45°, 30 ° and 20°.For a calcite sample exposed to stearic acid vapour at 105 °C for 4 h, the data are shown for two substrate signals: Ca 2p (in violet), and for comparison O 1s (in olive).The linear fitting was used for calculating  / values).

Figure S6 .
Figure S6.Degradation and / or desorption of the stearic acid layer in the XPS instrument under the influence of X-ray radiation.(a) The C 1s spectra evaluated during the first scan (straight line), and during the 18 th scan 93 min later (dashed line).The sample was exposed to stearic acid vapor for 4 h at 105 °C.(b) The C 1s/ Ca 2p atomic ratio relative to the value at 5 min (4.21) as a function of X-ray irradiation time fitted with a polynomial function.

Figure S7 .
Figure S7.Vibrational sum frequency spectra of a stearic acid monolayer adsorbed (24 h) on calcite, measured as a function of the azimuthal angle () as defined in the main article.Polarization combination: SSP.

b
Vibrational Sum Frequency.Fitting the anisotropy patterns in the polar plots

18
FigureS9shows the sum frequency spectrum of a freshly cleaved calcite surface in the absence of stearic acid.The sharp feature observed in the spectra at 1432 cm -1 unambiguously stems from the calcite surface and is assigned to the antisymmetric carbonate stretch ( 3 calcite ).18

Figure S9 .
Figure S9.Vibrational sum frequency spectra of freshly cleaved and unmodified calcite sample at three different polarizations (without the stearic acid monolayer)