Blanc Fixe Micro (BaSO₄, Sachtleben), Ethanesulfonic acid 95% (186260 Aldrich), Butyric acid ≥ 99% (B103500 Aldrich), Trimethoxy(propyl) silane 97% (662275 Aldrich), Phosphoric acid 2-hydroxyethyl methacrylate > 90% (SR 9054, Sartomer), Toluene anhydrous 99.8% (244511 Sigma-Aldrich), and N-Butyl Acetate (Univar) were used as received. All model molecules are presented in Figure 1.

The average particles diameter of BaSO₄ is 860 ± 22 nm (determined by dynamic light scattering—Malvern Mastersizer, dry process) and the specific area is equal to 2.5 m²/g (determined by BET analysis).

To perform the chemical surface treatment of barium sulfate particles, 10 g of BaSO₄, 4 g of model molecules diluted in 100 ml of appropriate solvent (toluene for butyric acid and trimethoxy(propyl) silane, butyl acetate for phosphoric acid 2-hydroxyethyl methacrylate and ethanesulfonic acid) are mixed and heated at 80˚C under magnetic stirring. After 1 h, the mixture is cooled down to room temperature and then centrifuged to remove solvent. 4 cycles of washing with appropriate solvent are performed by centrifugation to remove the species in excess. The BaSO₄ powder is then dried 15 h at 80˚C in an oven and then pelletized by compression molding. In case of untreated BaSO₄, 10 g of BaSO₄ are mixed with toluene (or butyl acetate) and heat at 80˚C during 1 hour under magnetic stirring. Then, the mixture is cooled down to room temperature, centrifuged to remove solvent and finally dried 15 h at 80˚C in an oven.

FT-IR analyses were performed in order to detect organic molecules onto the surface of inorganic particles (Figure 2). Strong absorptions were observed from 600 to 1200 cm⁻¹ corresponding to stretching bands of barium sulfate. The presence of a large band at 3400 cm⁻¹ is attributed to the stretching vibration of absorbed water on the hydrophilic surface of particles [5] [10] [21]. Spectrum highlights the difficulty to analyze organic residue due to the strong absorption of barium sulfate. In fact, the characteristic groups of organic molecules (trialkoxysilane, phosphoric acid or sulfonic acid groups) have bands located in the same zone as those of barium sulfate. So no significant difference is observed between the spectra, except for particles treated with phosphoric acid 2-hydroxyethyl methacrylate for which barium sulfate absorption bands are largest, which is probably due to the phosphorous compounds (hydrogen bonds).

In order to identify the characteristic groups of the various organic molecules outside the barium sulphate absorption zone, zooms were carried out between 1400 and 4000 cm⁻¹ (Figure 3 and Figure 4).

IR spectra of particles treated in butyl acetate with ethanesulfonic acid or phosphoric acid 2-hydroxyethyl methacrylate are presented in Figure 3. Analyses were focused on two areas: 2800 - 3000 cm⁻¹ and 1450 - 1750 cm⁻¹. The presence of absorption bands at 2900 cm⁻¹ only for phosphoric acid compound treated particles is attributed to the vibration mode of alkyl groups (ν C-H), corresponding to aliphatic chains of phosphoric acid 2-hydroxyethyl methacrylate [21]. Thus, other band at 1650 cm⁻¹, which is assigned to the C=O stretching vibration mode of the ester groups of phosphoric acid 2-hydroxyethyl methacrylate, confirms the existence of phosphoric acid groups onto the surface of barium sulfate. In counterpart, no specific bands of ethanesulfonic acid are observed. To conclude, FT-IR analyses of modified particles show that only phosphoric acid group is able to interact with the BaSO₄ surface.

Figure 4 presents the inorganic particles treated in toluene with butyric acid or trimethoxy(propyl) silane. As previously, the large peak at 3000 cm⁻¹ can be attributed to the stretching vibration mode of alkyl groups for both trialkoxysilane and carboxylic acid derivates, and underlines the presence of organic groups. Thus, from 1500 to 1700 cm⁻¹, a new peak appears for particles modified with butyric acid. It was assigned to the C=O stretching vibration mode of carboxylic acid. The shift of this band compared to the C=O of free butyric acid (1700 cm⁻¹ observed by FT-IR of raw material) indicates that the complex between COO⁻ and Ba²⁺ ions has been formed [10].

The presence of silane at the surface of the BaSO₄ particles, however, cannot be observed by IR.

Thermogravimetric analyses were performed on all samples to quantify the presence of residual grafted molecules for treated BaSO₄ compared to an untreated one.

Figure 5 shows the TGA curves obtained for untreated BaSO₄ and treated BaSO₄. The curves (a) and (b) correspond to thermogravimetric analyses of untreated BaSO₄ mixed with butyl acetate and toluene respectively. The curves are very close (less than 10% of standard deviation), suggesting that powders contain no solvent residues. Until 200˚C, the weight loss corresponds to physical desorption of water at the barium sulfate surface. Beyond 200˚C (from 200˚C to 550˚C), the weight loss is due to the elimination of chemisorbed water and to decomposition of barium hydroxides [5] [22].

The curves (c), (d), (e) and (f) refer to thermogravimetric analyses of BaSO₄ treated with ethanesulfonic acid, phosphoric acid 2-hydroxyethyl methacrylate, butyric acid and trimethoxy(propyl) silane respectively. Except for ethanesulfonic acid, for which the curve is similar to the reference, all curves present a significant weight loss compared to the raw BaSO₄. In fact, a more important weight loss than for untreated BaSO₄ is observed until nearly 300˚C. These weight losses relate to the adsorption of organic molecules at the surface of particles.

However, the weight losses are very low (less than 2%), because of the weak specific area of BaSO₄ (2.5 m²/g). That means that the grafted density of molecules onto BaSO₄ surface is very low.

In spite of the low weight losses, the grafted densities of organic molecules were calculated.

Firstly, if the weight losses of untreated particles are only due to water desorption because of decomposition of barium hydroxides, their weight loss contribution on treated particles can be determined (%ref = 0.809% and 0.879% for untreated BaSO₄ in butyl acetate and toluene respectively).

So, the weight losses due to the degradation of organic molecules grafted onto the surface (ΔmGM) can be deduced from Equation (1).

Δ m GM = Δ m − % ref ∗ m (1)

where, Δm_GM is the weight loss of the treated filler after the degradation of organic molecules used as models, Δm the total weight loss of treated particles determined by TGA, %_ref the percentage weight loss due to adsorbed water and hydroxide groups condensation, and m the mass of treated particles used as sample for thermogravimetric analysis.

From Δm_GM, the grafting density (molecules/nm²) of organic molecules can be calculated using Equation (2).

Grafting density = Δ m GM ∗ N A ∗ 10 18 S p ∗ m ∗ M TGA (2)

where, Δm GM is the weight loss of the treated filler due to the degradation of organic molecule, N_A the Avogadro constant, Sp the specific surface area of particles (2.5 m²/g), m the mass of particles used as sample for thermogravimetric analysis and MTGA the molar mass of degraded organic molecule.

Some hypotheses were taken into account for the molar mass values MTGA. For ethanesulfonic acid and butyric acid, all molecules were degraded as gases: carbon dioxide, carbon monoxide, alkenes for butyric acid and sulfur dioxide and alkenes for ethanesulfonic acid [23]. So the molar masses of raw materials were used.

However, in case of trialkoxysilane and phosphoric acid compounds, only the organic parts of initial molecules are thermally degraded. The silicon and the phosphorus remained at the BaSO₄ surface in form of silica [SiOx] and phosphorus oxide [POx] residues respectively. Water adsorbed on particles and thermal activation lead to the hydrolysis and condensation of these species (Figure 6) [23] [24] [25].

In order to evaluate the stability of trimethoxy(propyl) silane during the surface modification process, chemical treatment was renewed in benzene-d6 (instead of anhydrous toluene) at 65˚C during 1h30. The hydrolysis of trialkoxysilane was followed by ¹H NMR, using the evolution of the proton peaks corresponding to both the methyl groups attached to the silane (-Si-(OCH₃)₃) and those released during the hydrolysis (CH₃-OH), appearing at 3.44 and 3.07 ppm, respectively (Figure 7) [26] [27]. Analysis of supernatant revealed the presence of about 2% molar of methanol already observed before the treatment. Thus, no evolution of methyl groups attached to the silane was highlighted. Therefore, this trialkoxysilane is stable in anhydrous solution. However, the weight loss observed by TGA for BaSO₄ treated with alkoxysilane suggests that the condensation of trialkoxysilane onto BaSO₄ surface in anhydrous solvents can take place. In fact, it was reported that the traces of water physisorbed onto oxide surface can hydrolyze alkoxysilanes. Then, the hydrolyzed species can react with the hydroxyl groups present on the surface of the oxide via a hydrogen bond or a covalent bond. Finally, the growth of the organosilane layer occurs by condensation of hydrolyzed silanes [28] [29].

It is likely that the alkoxysilane was only partially hydrolyzed during the treatment. It was assumed that only one of methyl groups attached to the silane was hydrolyzed. Methanol produced by the hydrolysis of trialkoxysilane was removed by the cycles of washing before TGA analysis. The weight losses observed were probably due to the alkane gases issued from the degradation of the carbon chain (considered as propane) and the methanol issued from the degradation of methyl groups which were not hydrolyzed during chemical treatment (considered as methanol). The molar mass MTGA of the organic part degraded from trimethoxy(propyl) silane is given by Equation (3).

Msilane TGA = MPropane + 2 ∗ MMethanol = 1 0 8 . 18g/mol (3)

For phosphoric acid 2-hydroxyethyl methacrylate, only organic part of molecules, i.e. 2-hydroxyethyl methacrylate (HEMA) is degraded.

There are three possibilities according to the supplier data. R1 = R2 = RA named AA, R1 = RA/R2 = RB named AB and R1 = R2 = RB named BB (Figure 8).

For this reason, grafting densities taking into account the three cases AA, AB and BB described previously have to be calculated. The molar masses of degraded organic molecules due to these three species are summarized in Table 1.

³¹P NMR spectroscopy was performed in order to determine the relative percentage of different species AA, AB and BB (Figure 9). The ³¹P NMR spectrum reveals that phosphoric acid 2-hydroxyethyl methacrylate contains two major compounds as orthophosphate, monoester (AB) and diester (BB) of phosphoric acid 2-hydroxyethyl methacrylate [30] [31]. The triester (AA), orthophosphate and polyphosphates are in minority. These minor species were neglected for the determination of grafting density. The relative molar ratio of AB and BB were calculated using peak integration (Equation (4)). The results show a molar content of 56.7% of BB and 43.3% of AB.

% mol BB = ( BB ) peak integration ( BB ) + ( AB ) peak integration ∗ 100 (4)

The molar mass Mphosphoric acid TGA of degraded phosphoric acid 2-hydroxyethyl methacrylate calculated is 186.44 g/mol, using Equation (5).

M phosphoric acid TGA = ( 2 − % mol BB ) ∗ M HEMA (5)

where % mol BB is the relative molar percentage of phosphoric acid monoester compared to total of monoester and diester and MHEMA the molar mass of 2-hydroxyethyl methacrylate (130.14 g/mol).

The molar mass values MTGA used and the calculated grafting densities for all organic compounds are summarized in Table 2. The grafting density for ethanesulfonic acid is very low, in accordance with IR analysis. The conclusion is that sulfonic acid is not a good anchoring group to establish strong interactions with barium sulfate. The grafting density of trialkoxysilane, carboxylic acid and phosphoric acid derivates are more important. Hypotheses used for the calculated grafted density of alkoxysilane exclude the formation of multilayer. However, it is possible that the hydrolysis of methoxy groups is upper than 25%, and leads to the formation of multilayers. Despite this, the calculated value is considered to be representative of the actual grafting density.

^(a)Calculated using Equation (1) with untreated BaSO₄ in butyl acetate as reference: %ref = 0.809% (obtained by TGA); ^{(b)Calculated} using Equation (1) with untreated BaSO₄ in toluene as reference: %ref = 0.879% (obtained by TGA); ^(c)According to Equation (2).

The grafted densities for butyric acid and phosphoric acid 2-hydroxyethyl methacrylate are the highest. They are upper than 10 molecules/nm². Carboxylic and phosphoric acids are therefore good candidates as anchoring groups on the surface of barium sulfate. Thus, phosphoric acid groups seem to be more appropriate than carboxylic acid groups according to calculated grafting density.

To conclude, TGA made it possible to distinguish the effectiveness of several functions in anchoring to the surface of barium sulfate. It underlines that phosphoric acid and carboxylic acid groups are good anchoring groups for barium sulfate.