- Email: [email protected]
Surface properties of recycled titanium oxide recovered from paint waste
Progress in Organic Coatings 125 (2018) 279–286
Contents lists available at ScienceDirect
Progress in Organic Coatings journal homepage: www.elsevier.com/locate/porgcoat
Surface properties of recycled titanium oxide recovered from paint waste a,⁎
b
a
c
T
d
Mikael C.F. Karlsson , Zareen Abbas , Romain Bordes , Yu Cao , Anders Larsson , Antonin Rollanda, Phil Taylore, Britt-Marie Steenaria a
Department of Chemistry and Chemical Engineering, Chalmers University of Technology, Kemivägen 4, 412 96 Gothenburg, Sweden Department of Chemistry and Molecular Biology, University of Gothenburg, Kemigården 4, 412 96 Gothenburg, Sweden Department of Industrial and Materials Science, Chalmers University of Technology, Hörsalsvägen 7B, 412 58 Gothenburg, Sweden d RISE Research Institutes of Sweden, The Unit of Surface, Process and Formulation, Drottning Kristinas Väg 45, 114 28 Stockholm, Sweden e AkzoNobel Decorative Paints Global Research & Development Open Innovation, AkzoNobel Decorative Paints, Wexham Road, Slough Berkshire, SL2 5DS, UK b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Titanium oxide Pigment Recycling Paint Zeta potential Surface charge
Aluminium oxide coated rutile pigment was extracted from a paint matrix by means of a thermal recycling process. The objective was to investigate the effect of the recycling process on the surface properties of the pigment. The pigment was analysed using powder x-ray diffraction (XRD), surface area measurements (BET), laser diffraction for particle size analysis and X-ray photoelectron spectroscopy (XPS) before and after the recycling process. Investigations on the zeta potential and the surface charge were performed as well. It was concluded that the rutile crystalline core and the aluminium oxide coating of the pigment were still intact after the recycling process. The particle size distribution of the recycled pigment was slightly broader compared to the virgin pigment. The measured magnitude in zeta potential of the recycled pigment was lower than for the virgin pigment. This difference is thought to be caused by alteration in the surface hydroxyl concentration. Surface charge titrations showed differences between the virgin and the recycled pigment at alkaline pH and at low salt concentrations.
1. Introduction Titanium minerals are of great technical importance in today´s society. In 2011, the world production of titanium containing minerals was 6.7 million metric tonnes [1]. Of that, the vast majority was used to produce TiO2 for use as a white pigment in paints, coatings, plastics and papers [2,3]. TiO2 is the major white pigment used by the coatings industry [4] due to its abundance and its ability to scatter visible light while being chemically inert. Typically, TiO2 is produced using a sulphate or chloride route, both of which have a high carbon footprint per kilogram of TiO2 produced. It has been reported that even if new innovative production routes are developed the carbon footprint per kg TiO2 produced would still be high [5]. The European Union (EU) has recognized the environmental impact of TiO2 production and has consequently set restrictions on the amount of TiO2 allowed to be used in paint formulations if the paints are to qualify for the voluntary Ecolabel labelling [6]. In the future, waste management of old paint residues may also be included in the Ecolabel criteria [7]. Even before regulatory guidelines were in place, the coatings industry has strived to find a replacement for TiO2 due to its relatively ⁎
high cost compared to other coating components [8]. One appealing approach is the recovery of TiO2 from paint waste, as it could benefit the coatings industry in two ways. First, recovered TiO2 may be a cheaper and more environmentally friendly replacement for virgin TiO2 produced by conventional high carbon footprint routes. Second, a successful TiO2 recovery process could be a cornerstone in the waste management of old paint residues and production waste from manufacturing plants. TiO2 is often mentioned as the major white pigment but commercial grade pigment is frequently not pure TiO2. The crystalline TiO2 core is the active ingredient that provides the pigment's optical functions. However, the surface of the TiO2 particles does not consist solely of titanium and oxygen. As the TiO2 crystals grow during manufacturing, insoluble components accumulate on the surface. The components can be impurities from the ore or additives introduced deliberately to control the crystal structure, crystal growth, and particle agglomeration [9,10]. Besides these contaminants, the surface of a pigment is deliberately altered to suit the final application of the pigment. For instance, to reduce photoactivity and to improve the compatibility between the pigment and the other paint components, the surface of the TiO2 pigment is commonly treated with silicon, aluminium and zirconium
Corresponding author. E-mail address: [email protected] (M.C.F. Karlsson).
https://doi.org/10.1016/j.porgcoat.2018.09.012 Received 28 March 2018; Received in revised form 31 August 2018; Accepted 1 September 2018 0300-9440/ © 2018 Elsevier B.V. All rights reserved.
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
properties of the recycled TiO2 and the effects of the recycling process on the TiO2 pigment specifically. The formulation, mixing and paint preparation was performed at Akzo Decorative Paints in Slough, UK.
oxides. The surface of uncoated TiO2 pigment shows a complex behaviour with different active groups, but a coated pigment can have an even more complex surface that is chemically very different to the bulk TiO2 phase [11]. For the optimum opacity and durability of a paint film, a well dispersed pigment is a must. So, although it is no uncomplicated task, the surface characterization of a coated pigment is of importance in understanding pigment-medium interactions and its performance the final application [12]. In a previous TiO2 recycling study [13] a white paint was pyrolysed at 500 °C in order to recover the inorganic components. In that study, TiO2 was extracted together with the mineral extenders present in the paint, generating a pigment-extender mix consisting of roughly 45% TiO2 and 55% mineral extenders, mainly dolomite. The recycled material was then used in a new paint formulation. It was concluded that using pyrolysis as a means to recycle TiO2 resulted in a pigment-extender mix that could be used in a low-quality, flat paint, formulation if it was not for the large, millimetre size, agglomerates present in the paint. Since the recovered fraction was a mix of inorganic components, it was not possible to study in detail the agglomeration mechanism or the effect of the recycling process on the properties of the TiO2 pigment. Therefore, it was decided to perform the next step of the work using a model paint based on TiO2 as the sole pigment. In the present work, a paint formulation based on alumina coated TiO2 pigment (rutile) was prepared. The pigment was extracted from the paint matrix through a pyrolysis based recycling process. The pigment was analysed prior to and after the recycling process in terms of surface chemistry, both with electrokinetic measurements and with direct titration of surface groups. Additionally, the pigment were characterized using powder x-ray diffraction (XRD), surface area measurements (BET), laser diffraction for particle size analysis and X-ray photoelectron spectroscopy (XPS). The main purpose of this study was to understand the effects of the thermal recycling process on the surface characteristics of the pigment and identify the changes caused by the process. The knowledge of the surface characteristics will be useful in future studies when evaluating the dispersion stability of the pigments and their interactions with other paint components, such as dispersants and binder.
2.2. Recovery process The TiO2 was recovered from the model paint presented in Table 1 by means of a pyrolysis process. The temperatures chosen in the thermal recovery process were based on previously published results [13]. By heating the paint in an inert atmosphere, the organic fraction was volatilized, separating it from the inorganic fraction. The volatilized organics were cooled and collected as an oil with the potential to be used as feedstock for a chemical process or as an energy source. After pyrolysis, the inorganic fraction was oxidized in air to remove residual carbon and non-pyrolyzed material, improving the colour of the recovered pigment. To further purify the recovered material it was dispersed in water with a mixture of ion exchangers. To finalize the recycled pigment the ion exchangers were separated from the pigment dispersion, which was dried and homogenized into a fine powder. A schematic of the recycling process can be found in the supplementary information. 2.2.1. Pyrolysis The pyrolysis experiments were conducted in a Rohde, ME 45-13 furnace, fitted with a pyrolysis retort in corrosion resistant steel. The inner dimensions of the pyrolysis retort were 300 × 300 × 150 cm. During the experiments, the temperature was controlled using a TC 504 temperature controller. In addition, an external thermocouple was used to monitor the furnace temperature in relation to the set temperature. Before each experiment, the retort was filled with nitrogen gas (purity 99.9%) to an overpressure of 0.5–1 bar, emptied, and refilled again with nitrogen to overpressure. This was repeated three times to create an oxygen free environment. The paint was dried at 150 °C, followed by pyrolysis done under atmospheric pressure at 500 °C. During the pyrolysis experiments, the retort was flushed with nitrogen, 0.85–0.95 L/ min. The volatilized organics and liquid produced during the pyrolysis were led out through the retort outlet. In the present work, the focus was on the inorganic components of the sample materials and therefore the oil and gas fractions were not collected for analysis. To reduce the exposure of hazardous volatile pyrolysis products, the sample was cooled to 50 °C, under continuous nitrogen flow, in the pyrolysis retort before being removed. Once cooled, the solid residue was collected and weighed. The solid product was finally homogenized with a mortar and pestle.
2. Materials and methods 2.1. Model paint A white model paint was produced according to the formulation shown in Table 1. A commercially available pigment quality TiO2, rutile coated with aluminium oxide, and all other paint components were acquired from Akzo Nobel Decorative Paints, UK. Typically, waste paint streams do not only contain TiO2 as the inorganic component. Inorganic extenders are used in paints to give the final coating the desired properties, such as gloss or scrub resistance. However, those components were excluded from the model formulation. A simplified version of a paint waste stream will allow for a more in-depth analysis of the
2.2.2. Oxidation The pyrolysis product mainly contained the inorganic components in the paint formulation (in this case, TiO2) and carbon residues from pyrolysis of the organics. To remove the residual carbon and non-volatilized organic material, the pyrolysis residue was oxidized by spreading the powder in a thin layer (< 4 mm thick) in an alumina crucible and heat treating it in 470–500 °C for 1.5 h in air. After cooling, the material was collected, weighed, and homogenized using a M20 universal mill from IKA.
Table 1 Model paint formulation produced for the TiO2 recovery process. Raw Materials
wt.%
Water Binder (vinyl acrylic) Antifoam Non-ionic surfactant Anionic surfactant Hydroxyethylcellulose (HEC) TiO2 Incan preservative pH modifier
22.92 44.25 0.51 0.51 0.46 0.51 30.47 0.12 0.25
2.2.3. Purification of pigment with ion exchanger resin During the pyrolysis process all organic components of the paint are decomposed into smaller entities and volatized, and were therefore separated from the inorganic fraction of the paint. This inorganic fraction would contain the inorganic pigments and unwanted salt residues from the original paint. In a later stage, when the recycled pigments were re-dispersed in an aqueous system, these ionic species would dissolve and increased the ionic strength of the solution. It is known that an increased ionic strength have negative effects on colloidal stability [14] and the performance of certain dispersing aids commonly used in paints [15]. Therefore a washing of the pyrolysis product was designed. 280
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
thickness of 100 nm.
The oxidized residue (see Section 2.2.2) was dispersed in distilled water. Ion exchanger resin (Amberlite IR120, hydrogen form) was added to the suspension under continuous stirring and pH measurement. The amount of ion exchanger added was calculated to present an excess of exchanging sites. pH rapidly decreased from the initial pH (> 9) to stabilize around pH 2.5–3.0 after 5 min. After 30 min the suspension was neutralized using an ionic exchanger resin (Amberlite IRN78, hydroxide form) under continuous stirring. pH slowly increased to stabilize around 7 after 1.5 h. The pigment suspension with the 2 ionic exchangers was left under continuous stirring for 4.5 h. The pigment was separated from the ionic exchangers through sieving and the recovered pigment was dried in 60 °C, before being homogenized into a fine powder with a mortar and pestle. This product is referred to as recycled pigment, unless stated otherwise. Conductivity was measured to get an indication of the effect of the washing step. To this end the virgin pigment, the oxidized pyrolysis residue and the recycled pigment was dispersed (10 wt. %) in Milli-Q water (> 18.2 MΩ/cm). After 96 h mixing, samples of the water phases were collected, diluted with Milli-Q water and analysed with inductively coupled plasma mass spectrometry (ICP-MS) for cations (atomic mass ranges 7–238) and ion chromatography (IC) for anions (Cl−, PO43−, NO3−). The measurement of conductivity and ion concentration was done in triplicate. All the conductivity measurements were done with a 5-ring conductivity measuring cell calibrated with a 100 μS/cm conductivity standard from Metrohm. All pH measurements were done with a LLAquatrode plus from Metrohm calibrated using a 5-point calibration (pH 2, 4.01, 7, 10 and 12, Certipur, Merck). The Amberlite IR120 (in the hydrogen form from Sigma-Aldrich) was washed with Milli-Q water and regenerated with 0.1 M HCl before use. The Amberlite IRN78 (in the hydroxide form from Sigma Aldrich) was used as received from supplier. The ICP-MS instrument was a Thermo iCap Q and the IC a Dionex DX-100 with anion column IonPac AS-4A SC.
2.3.2. Particle size and BET The particle size distribution was measured by laser diffraction, using a Mastersizer MicroPlus from Malvern Instruments. The particles were measured in a dilute suspension of deionized water. Before measurement, the suspension was treated with ultrasound (200 W for 40 s applied to 30 mL of 1 wt.% suspension, using a Sonics VibraCell sonicator) to break up agglomerates. The specific surface area of the virgin pigment and the recycled pigment was determined by BET measurements as described by Brunauer et al. [16]. The exact procedure for the experiments is described elsewhere [13]. 2.4. Dynamic mobility and zeta potential The zeta potential analysis was performed using a ZetaProbe from Colloidal Dynamics. The ZetaProbe operates at a field frequency range of 0.3–3 MHz. When the alternating field is applied, the particles (and ions) start to oscillate. As the particles move, they displace an equal volume of liquid. If the density of the particles is different from the density of the liquid, this will give rise to a sound wave. This so called Electrokinetic Sonic Amplitude (ESA) signal depends on the particle velocity and the dynamic mobility, μd, can be determined from this. The dynamic mobility of the particles depends on their size, their zeta potential and the frequency of the applied field. For dilute colloidal dispersions (up to about 5 vol percent), the relationship between the ESA effect and the dynamic mobility, μd, of the charged particles is given by [17]
ESA = A (ω) ϕ
Δp μ po d
(1)
where ESA is measured in (Pa/(V/cm)), A(ω) is an instrument factor, φ is the volume fraction of the particles and Δρ is the density difference between the particles and that of the solvent (ρ0). The ZetaProbe measures the maximum in the pressure of the wave (Pa) per unit applied electric field strength (V/cm). The ZetaProbe measures the ESA at 8 different frequencies and obtains a mobility spectrum of the sample. The zeta potential is then calculated from the dynamic mobility spectrum using the O´Brien formula, which is valid for spheroidal particles with thin electrical double layers [18]. Before the measurement, as the presence of excess ions can affect the measurement, the virgin pigment was cleaned with the use of ion exchangers with the same procedure as used for the recycled pigment, see Section 2.2.3. The measuring cell contained 250 mL of sample with a stirring speed of 150 rpm. The charge of the particles was determined in a pH range from 4 to 10 by titration using a particle concentration between 1–5 wt. % in 2, 10 mM NaCl and 10 mM NaNO3 (≥99.0% ACS regents from Merck dissolved in distilled water) using a computer controlled titrator. A particle concentration below 2.5 wt.% had a significant influence on the results, for that reason 5 wt.% was chosen for this study. The pH-electrode was calibrated daily using a 3-point calibration (pHstandards 4, 7, 10) with standards from Hamilton. The ESA signal was calibrated daily with a 0.25 S/m Potassium 12-tungstosilicate hydrate standard from Colloidal Dynamics. TiO2 with a density of 4.26 g/mL and a dielectric constant of 40 and the properties for water at 20 °C was chosen calculating the zeta potential. 1 M HCl and 1 M HNO3 was used for titration of samples dispersed in NaCl and NaNO3, respectively. For both electrolytes 1 M NaOH was used as base. All bases and acids were prepared from Titrasol concentrates (Merck). All data presented was corrected for the signal from the background electrolyte, although not corrected for change in pH. The influence of pH was checked, and the signal contribution was less than 2% except when approaching a pH of 2, where the influence could be up to 10% of the total ESA signal. As titration was done in pH 4–10 this contribution was neglected.
2.3. Characterization of recycled pigment Characterization of the recycled pigment was done with the virgin pigment used in the model paint, see Table 1, as a reference. The purpose of the characterization was to compare the properties of the virgin and recycled pigment and to see the effect of the recycling process on the pigment. 2.3.1. X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) To confirm that the main crystal structure of the rutile was intact, the virgin pigment and the recycled pigment were analysed with X-ray diffraction (XRD). This procedure has been described previously [13]. The virgin pigment and the recycled pigment were both analysed with X-ray photoelectron spectroscopy (XPS) using a Versaprobe III from Physical Electronics equipped with monochromatic Al Kα X-ray source (1486.6 eV). The virgin pigment was cleaned with the use of ion exchangers, as described in Section 2.2.3, prior to XPS analysis. The Xray beam diameter was 100 μm and the power was 25 W. Acquisition conditions for the survey spectra (0–1100 eV) were 112 eV pass energy, 45° take off angle and 0.5 eV/step. Selected region spectra were recorded covering the Ti 2p, Al 2p, O 1s, Si2p and C 1s photoelectron peaks. The acquisition conditions were, then, 55 eV pass energy, 45° take off angle and 0.1 eV/step. Samples were mounted on a steel sample holder using double sided tape. Dual beam charge compensation was carried out by flooding the sample with low energy electrons and low energy argon ions. The C 1s peak from adventitious C located at 284.8 eV was used as energy scale reference. The element distribution in depth was estimated by depth profiling performed using Ar+ ion sputtering with 2 kV accelerating voltage and 7 mA current. The etch rate was 5.4 nm/min calibrated on a Ta2O5 standard sample with a
281
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
2.5. Surface charge titrations
Table 2 Conductivity and selected ion concentration of 10 wt.% pigment suspensions. Uncertainties given are one standard deviation from the mean.
The point of zero charge for the virgin and the recycled pigment was determined experimentally by potentiometric titration. A Metrohm 905 Titrando equipped with stand, stirrer, acid and base dispensers and a pH-electrode (LL-Aquatrode plus). The pH-electrode was calibrated using a 5-point calibration (pH 2, 4.01, 7, 10 and 12, Certipur, Merck). The Tiamo software from Metrohm was used to monitor and control all calibrations and titrations. Prior to the titration was the virgin pigment was cleaned with the use of ion exchangers with the same procedure used for the recycled pigment, described in Section 2.2.3. Acids, bases, and background electrolytes for the titration were made from Titrasol concentrates and NaCl respectively KCl salts dried before use. All dilutions were made using Ultra pure Milli-Q water. The following reagents were used: 2 mM, 10 mM and 100 mM NaCl respectively KCl solutions were used as the background electrolyte. For each electrolyte concentration, 0.1 M NaOH and 0.1 M HCl, with the same salt concentration as the electrolyte, were prepared as well. A dispersion of 10.00 ± 0.01 g pigment and 50 mL background electrolyte was made. The dispersion was acidified with 2 mL acid for 30 min and then titrated with 4 ml of base added in 50 μL increments. Between each step, the potential was measured. The time between steps was 60–600 s, depending on equilibrium time. Equilibrium was considered to be reached if the change in the potential was less than 0.5 mV/min, corresponding to about 0.01 pH-unit/min. After the titration, the potential of the pH-electrode was converted into pH-units using the latest pH-calibration curve. For the background titrations, pure electrolytes were titrated similarly but the maximum time between each titration step was increased to 1800 s. The increased waiting time was necessary in order for the system to reach equilibrium in each step, especially for points around 0 mV. Experiments were also performed where the pigment dispersion and the background electrolyte respectively were continuously bubbled with nitrogen gas (purity 99.9%) during the titration. This was done to remove CO2, which can dissolve into solution and potentially affect the measurement. It was concluded that the nitrogen bubbling only had a minor effect on the final result, for which reason all results presented in this article are based on titrations made without nitrogen bubbling. The surface charge at each ionic strength was calculated, analogous to Kosmulski [19], by
σ0 =
FcΔV mA
Sample
Conductivity [mS/ cm]
Na [mM]
Al [mM]
Cl [mM]
Virgin pigment Recycled pigment prior to washing Recycled pigment after washing
0.2 ± 0.1 2.4 ± 0.1
11 ± 1 130 ± 10
< 0.1 13 ± 2
0.2 ± 0.1 0.1 ± 0.1
< 0.001
< 1.0
< 0.1
< 0.1
3.1.1. Purification of pigment with ion exchanger As can be seen in Table 2 the suspension of the recycled pigment had a conductivity compared to the suspension of virgin pigment. Of all the ions measured (see Section 2.2.3) the difference between the suspensions in the Na and the Al ion concentrations (see Table 2) were the most important. The increased Na and Al concentrations for the recycled sample are likely to come from the ionic dispersing aids commonly used in paint formulations [20], from the counter ion of the stabilizer of the binder and from the inorganic coating of the pigment particles. The increase in ion concentration could have a negative effect on the stability of the dispersion [14] and the functionality of used dispersants in a paint system [15]. After purification with ionic exchangers all the measured ion concentrations for the recycled pigment were below that of the virgin pigment. 3.2. Characterization of pigment 3.2.1. X-ray powder diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) The XRD spectrum for the residue confirmed that the rutile structure of the recycled pigment was intact. This is as expected, because the temperatures in the recovery process are relatively low and the results are analogous with previously published results [13]. The main purpose of the XPS analyses was to identify possible differences between the virgin and the recycled pigment. In the XPS spectrum, five elements were distinctly detected: C, Ti, Al, Si, and O. The C 1s peaks for both samples (see Fig. 1a) are identical and will therefore be treated as adventitious carbon, thus used as the energy reference by setting the main line of the C 1s spectrum to 284.8 eV. The peaks for Ti p3/2 and Ti p1/2 (shown Fig. 1b) were located at 458.2 and 463.8 eV respectively, as expected for TiO2 [21]. In Fig. 1c it can be seen that the Al 2p peak is located at 74.1 eV for both samples. This matches the binding energy of an aluminium oxide, Al2O3, Al (OOH) or Al(OH)3. No further investigation were carried to disambiguate the exact nature of the signal, as this task is complex [22] and outside the scope of this study. The center of the Si 2p peak for both the virgin and the recycled was located on 102.1 eV (Fig. 1d). It is common to coat pigments with silica (crystalline or amorphous SiO2) but the Si 2p peak at 102.1 eV did not correspond the binding energy of SiO2, which is located at higher energies (around 103.5 eV) [23]. This suggest that the silicon is in the valence state +3 [23], thus indicating that Si is a contaminant that has accumulated at the surface. It should be mentioned that siloxane compounds provided a good fit of the peaks, with Si 2p, O 1 s and C 1 s at 102.1, 532.4 and 284.4 eV respectively [24]. These types of compounds are sometimes used to improve the dry flow of pigment particles [9]. On the other hand, it has been reported that these kind of compounds degrade in the temperature range [25] used in the recovery process presented in this paper. As the Si 2p is clearly visible in both the virgin and the recycled sample is it more likely that the Si is incorporated in the matrix as contaminate rather than in the form of siloxane. A XPS survey spectra (see Supplementary information) of the virgin pigment collected prior to the purification step with ion exchanger (see Section 2.2.3) is lacking the Si 2p peak at 102.1 eV. This
(2)
where F is Faradays constant, c the concentration of acid or base, m the mass of solid particles, A the specific surface area measured by BET and ΔV the difference in volume of acid and base to reach the same pH in the dispersion as in the electrolyte solution.
3. Results and discussion 3.1. Recovery process The total TiO2 content of the processed model paint (see Table 1) was 30.47 wt. %. The average weight of the residues after pyrolysis was 32.8 ± 0.2 wt. % of the starting material, while the average weight of the residue after oxidation was 30.5 ± 0.2 wt. % of the starting material. This suggests that, after pyrolysis, the remaining organic material and residual carbon are effectively removed during the oxidation step. There was no visible colour difference between the virgin and the recycled pigment. These results show that it is possible to extract TiO2 (or other inorganic pigment) from the paint matrix by means of pyrolysis and oxidation. 282
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
Fig. 1. XPS photoelectron peaks a) C 1s, b) Ti 2p, c) Al 2p and d) Si 2p for virgin pigment (solid line) and recycled pigment (dashed line). Energy is normalized and the peak for recycled pigment is shifted to higher intensity for clarity. Peaks are relative to C 1 s = 284.8 eV.
belonging to the oxide bulk structure. The second peak, around 532 eV, is due to surface hydroxyl groups. The third peak, with the highest energy at roughly 534 eV, is from chemisorbed water. As the other components have their chemical state intact, the most likely explanation is a difference in concentration of water that has been chemisorb from the ambient surroundings on to the surfaces. This chemisorb water would increase the intensity at higher energies. It is also possible that the water could cover the surface and reduce the measured intensity at lower energies (bulk O2- oxygen). Thus, the alteration in the O 1s spectrum is probably not coming from an actual chemical change but rather from increased amount of chemisorbed water on the recycled pigment compared to the virgin pigment. Another aspect of importance is the concentration of the different species at the surface. One of the most important characteristics for a coated pigment is the ratio of bulk Ti to coating elements [28–30], in this case Al. In Fig. 3 a depth profile for the ratio between Ti and Al is shown for the virgin pigment and the recycled pigment. Si was excluded as its chemical state does not suggest a traditional coating characteristic. As seen in Fig. 3 the ratio of Ti and Al is intact at the surface after the recycling. However, going further into the particles the Al seems to drop in concentration more rapidly than for the virgin pigment. This indicates that some of the coating has been etched away during the recycling process. This corresponds well with the increased Al concentration measured in a suspension of non-purified recycled pigments (see Table 2). Overall, it seems that the recycling process does not alter the surface by the creation of new species but it reduces the thickness of the alumina coating.
Fig. 2. O 1s XPS peak for virgin pigment (solid line) and recycled pigment (dashed line). Energy is normalized and the peak for recycled pigment is shifted to higher intensity for clarity. Peaks are relative to C 1s = 284.8 eV.
leads to the conclusion that the Si peak is a contamination from the purification step. In comparison with the other element peaks, the O 1s signal showed more pronounced differences between the virgin and the recycled pigment (Fig. 2). Whereas the main centre of the peaks was at the same binding energy, virgin sample exhibited a shoulder towards lower energies while the recycled sample has a shoulder towards higher energies. This difference cannot reasonably be attributed to an alteration of the surface species, as all the other major elements exhibited minor changes. It has previously been documented that three different peaks O 1s can be observed for metal oxides [26,27]. The lowest energy peak (around 529 eV) is from the contribution of the O2− oxygen atoms
3.2.2. Particle size and specific surface area The particle size distribution of the virgin and the recycled pigment is show Fig. 4. TiO2 virgin pigment particles has a narrow particle size distribution, centred around 300 nm in diameter, in order to scatter light most efficiently [31,32]. Compared to the virgin material the recycled sample had a slightly broader particle size distribution. Published specific surface areas for rutile pigments with various 283
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
Fig. 5. Measured zeta potential for 5 wt. % pigment suspensions in 2 mM NaCl (■), 10 mM NaCl (●), and 10 mM NaNO3 (▴) as a function of pH for virgin (red) and recycled (blue) pigments. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Table 3 IEPs for 5 wt. % virgin and recycled pigment suspensions.
Fig. 3. Cation depth profile of virgin (red) and recycled (blue) pigments. Only Ti and Al are included. The atomic concentration of Ti (dashed lines) and Al (dotted line) as a function of sputter time. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
coatings lie in the range of 5–35 m2/g with the majority around 20 m2/g [28,29,33–35]. The BET surface area for the virgin pigment used in this work was measured to 17 m2/g and for the recycled pigment 16 m2/g. The measured surface areas are within the margin of error so it cannot be stated that there is any significant difference between the samples.
Sample
2 mM NaCl
10 mM NaCl
10 mM NaNO3
Washed virgin pigment Washed recycled pigment
7.4 7.6
7.4 7.6
7.4 7.8
Table 4 IEP values for commonly used coated TiO2 pigments.
3.3. Dynamic mobility and zeta potential The results of the zeta potential measurements are shown in Fig. 5 with the corresponding IEPs summarized in Table 3. Reference IEP values for coated rutile pigments are given in Table 4, with most pigment qualities having an IEP in the pH range of 7–8. The IEP tends to be related to the average surface-area of its components, where increased silica gives a more acidic IEP and alumina gives a more basic IEP [34]. Different IEPs for pure aluminium (hydr)oxide [36] and pure rutile [19] have been reported, use of the average IEP is suggested by the authors
Pigment
IEP
Source
Pigments coated with alumina and zirconia Pigment coated with alumina and silica Pigment coated with alumina Pure rutile Pure alumina
7.8 ± 0.3 2.8–8.1 8.2–9.2 4.9 9.2
[34] [28,29] [30] [35] [35]
of this work which is 9.0 and 5.6 respectively. Using these averages and the ratio 0.46 Ti to 0.54 Al in the outmost layer of the pigments given by the XPS data (see Fig. 3) an IEP of 7.4 was calculated, which is in good agreement with the IEP measured, for the virgin pigment. This
Fig. 4. Particle size distribution of virgin (●) and recycled (○) pigments expressed as volume percent (a) and accumulated volume percent (b). Data presented is the average of three measurements and the uncertainty is given in one standard deviation. 284
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
Fig. 6. Surface charge determined by potentiometric titrations in 1 mM (black), 10 mM (red), and 100 mM (green) NaCl solution for virgin (a) and recycled (b) pigments respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
charge at acidic pH but the recycled pigment has more surface charge at higher pH and this is especially noticeable at 1 mM and 10 mM NaCl. This is opposite to zeta potential measurements (see Section 3.3) in which the recycled pigment always showed a lower magnitude of charge compared to the virgin pigment. Such differences between the surface charge and zeta potentials have recently been analysed for SiO2 nanoparticles in NaCl solution [38]. The difference stems from the fact that decrease in zeta potential at increasing electrolyte is due to the compression of stern layer whereas increase in surface charge density at increasing electrolyte concentration is due to protonation/deprotonation reactions. The decreased zeta potential at pH > 7 for recycled TiO2 and increased surface charge density at such pH values can be rationalized due to the above mentioned reasons. Differences in surface charge between the samples could as well be explained by assuming that the exposure to water of the heat treated recycled pigment led to rehydroxylation of the surface. It is well known the rehydroxylation of a heat treated oxide surface is not instantaneous [39] and the kinetics is influenced by pH [40]. The surface charge was calculated using Eq. (2). It was assumed that no reactions takes place and that the difference in volume of acid and base to reach the same pH in the dispersion as in the electrolyte solution was due to protonation or deprotonation of surface hydroxyl groups. Formation of hydroxyl groups on the particle surface could lead to a change in pH of the suspensions, which would result in an artificially more negatively charged surface calculated with Eq. (2). This effect would be less visible in the case of zeta potential measurements, as the total titration time was much shorter, i.e. 1 h. During that time the hydroxyl formation was limited. The potentiometric titration took approximately 8 h during which the hydroxyl formation occurred and modified the surface chemistry. Although there are differences in surface charge which cannot be fully
demonstrate that the rutile core is not completely encapsulated by the alumina coating. The IEP of the recycled pigment is 0.2–0.4 pH-units higher than that of the virgin pigment (Table 3), and, as shown in Fig. 5, the magnitude of the zeta potential variation under different salt conditions was smaller for the recycled pigment compared to the virgin pigment over the measured pH scale. Moreover, the IEPs of both virgin and recycled TiO2 do not show concentration dependence which indicates that there is no specific ion interactions. It has previously been shown [37] that temperatures used in the recycling process can reduce the concentration of hydroxyl groups on surfaces. A lower concentration of hydroxyl groups would result in a lower magnitude in zeta potentials, as fewer groups are available for protonation/deprotonation. For electrostatic stabilized systems it is generally admitted that a larger magnitude of the zeta potential usually results in more stable particle suspension. In practice, the lower magnitude observed may come from partial dehydroxylation, which implies that the dispersibility by surface active agent would be less effective. 3.4. Surface charge titrations Surface charge, σ0, determined by potentiometric titrations, are shown in Fig. 6. Only results for the experiments with NaCl as electrolyte are shown as the KCl based electrolyte showed very similar results. It can be suggested that the virgin pigment has a common intercept point (CIP) for σ0 = 0 between pH 7.1–7.6. This correspond well with the measured IEP at 7.4 (see Table 3). However, for the recycled pigment no clear CIP at σ0 = 0 could be determined. When comparing the titration curves for the two pigments in respective electrolyte (Fig. 7) it can be seen that the pigments display a very similar surface
Fig. 7. Surface charge determined by potentiometric titrations in (a) 1 mM, (b) 10 mM, and (c) 100 mM NaCl solution for virgin (■) and recycled ( ) pigments respectively.
285
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
explained; these differences are most prominent at lower salt concentrations. At higher salt concentrations, which is more likely to be found in a paint formulation [41], the difference in surface charge is negligible.
[7] S. Jiannis, R.K. Kougoulis, Ben Walsh, Katherine Bojczuk, Trevor Crichton Revision of EU European Ecolabel and Development of EU Green Public Procurement Criteria for Indoor and Outdoor Paints and Varnishes, (2012). [8] M.J.A. Ruszala, N.A. Rowson, L.M. Grover, R.A. Choudhery, Low carbon footprint TiO2 substitutes in paint: a review, Int. J. Chem. Eng. Appl. 6 (5) (2015) 331–340. [9] J.H. Braun, Titanium dioxide—a review, J. Coat. Technol. 69 (868) (1997) 59–72. [10] T.C. Patton, Characterization and physical relationships, Pigment Handbook vol. 3, Wiley, New York, 1973. [11] D.H. Solomon, D.G. Hawthorne, Chemistry of Pigments and Fillers, Wiley, New York, 1983. [12] G.D. Parfitt, The role of the surface in the behavior of titanium dioxide pigments, Croat. Chem. Acta 53 (2) (1980) 333–339. [13] M.C.F. Karlsson, D. Corr, C. Forsgren, B.-M. Steenari, Recovery of titanium dioxide and other pigments from waste paint by pyrolysis, J. Coat. Technol. Res. 12 (6) (2015) 1111–1122. [14] G.D. Parfitt, Dispersion of Powders in Liquids: With Special Reference to Pigments, Applied Science Publishers, London, 1981. [15] S. Farrokhpay, A review of polymeric dispersant stabilisation of titania pigment, Adv. Colloid Interface Sci. 151 (1–2) (2009) 24–32. [16] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (2) (1938) 309–319. [17] R.W. O’Brien, The electroacoustic equations for a colloidal suspension, J. Fluid Mech. 212 (1990) 81–93. [18] R.W. O’Brien, Electro-acoustic effects in a dilute suspension of spherical particles, J. Fluid Mech. 190 (1988) 71–86. [19] M. Kosmulski, Surface Charging and Points of Zero Charge. Surfactant Science Series vol. 145, CRC Press, Boca Raton, 2009, pp. 0081–9603. [20] J. Bieleman, Additives for Coatings, Wiley-VCH, Weinheim, 2000. [21] A.P. Dementjev, O.P. Ivanova, L.A. Vasilyev, A.V. Naumkin, D.M. Nemirovsky, D.Y. Shalaev, Altered layer as sensitive initial chemical state indicator, J. Vac. Sci. Technol. A 12 (2) (1994) 423–427. [22] J.T. Kloprogge, L.V. Duong, B.J. Wood, R.L. Frost, XPS study of the major minerals in bauxite: gibbsite, bayerite and (pseudo-)boehmite, J. Colloid Interface Sci. 296 (2) (2006) 572–576. [23] K.T. Queeney, M.K. Weldon, J.P. Chang, Y.J. Chabal, A.B. Gurevich, J. Sapjeta, R.L. Opila, Infrared spectroscopic analysis of the Si/SiO2 interface structure of thermally oxidized silicon, J. Appl. Phys. 87 (3) (2000) 1322–1330. [24] M.R. Alexander, R.D. Short, F.R. Jones, M. Stollenwerk, J. Zabold, W. Michaeli, An x-ray photoelectron spectroscopic investigation into the chemical structure of deposits formed from hexamethyldisiloxane/oxygen plasmas, J. Mater. Sci. 31 (7) (1996) 1879–1885. [25] G. Camino, S.M. Lomakin, M. Lazzari, Polydimethylsiloxane thermal degradation Part 1. Kinetic aspects, Polymer 42 (6) (2000) 2395–2402. [26] H. Perron, J. Vandenborre, C. Domain, R. Drot, J. Roques, E. Simoni, J.J. Ehrhardt, H. Catalette, Combined investigation of water sorption on TiO2 rutile (110) single crystal face: XPS vs. periodic DFT, Surf. Sci. 601 (2) (2007) 518–527. [27] E. McCafferty, J.P. Wightman, Determination of the concentration of surface hydroxyl groups on metal oxide films by a quantitative XPS method, Surf. Interface Anal. 26 (8) (1998) 549–564. [28] T. Losoi, Surface studies of titanium dioxide pigments, J. Coat. Technol. Res. 61 (776) (1989) 57–63. [29] K. Siwinska-Stefanska, A. Krysztafkiewicz, T. Jesionowski, Effect of inorganic oxides treatment on the titanium dioxide surface properties, Physicochem. Probl. Miner. Process. 42 (2008) 141–151. [30] M.L. Taylor, G.E. Morris, R.S.C. Smart, Influence of aluminum doping on titania pigment structural and dispersion properties, J. Colloid Interface Sci. 262 (1) (2003) 81–88. [31] J. Fisher, T.A. Egerton, Titanium compounds, inorganic, Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Inc., 2000. [32] M.P. Diebold, Application of Light Scattering to Coatings: A User’s Guide, Springer International Publishing, 2014. [33] J.-P. Boisvert, J. Persello, J.-C. Castaing, B. Cabane, Dispersion of alumina-coated TiO2 particles by adsorption of sodium polyacrylate, Colloids Surf. A: Physicochem. Eng. Asp. 178 (1–3) (2001) 187–198. [34] S. Farrokhpay, G.E. Morris, D. Fornasiero, P. Self, Influence of polymer functional group architecture on titania pigment dispersion, Colloids Surf. A: Physicochem. Eng. Asp. 253 (1–3) (2005) 183–191. [35] G.E. Morris, W.A. Skinner, P.G. Self, R.S.C. Smart, Surface chemistry and rheological behaviour of titania pigment suspensions, Colloids Surf. A: Physicochem. Eng. Asp. 155 (1) (1999) 27–41. [36] G.V. Franks, Y. Gan, Charging behavior at the alumina–water interface and implications for ceramic processing, J. Am. Ceram. Soc. 90 (11) (2007) 3373–3388. [37] M. Primet, P. Pichat, M.V. Mathieu, Infrared study of the surface of titanium dioxides. I. Hydroxyl groups, J. Phys. Chem. 75 (9) (1971) 1216–1220. [38] M.A. Brown, A. Goel, Z. Abbas, Effect of electrolyte concentration on the stern layer thickness at a charged interface, Angew. Chem. Int. Ed. 55 (11) (2016) 3790–3794. [39] S. Shioji, M. Hanada, Y. Hayashi, K. Tokami, H. Yamamoto, Kinetic study of alkoxylation and rehydroxylation reactions on silica surfaces, Adv. Powder Technol. 18 (4) (2007) 467–483. [40] R.K. Iler, The surface chemistry of silica, Chemistry of Silica–Solubility, Polymerization, Colloid and Surface Properties and Biochemistry, (1979), pp. 644–645. [41] T. Kobayashi, Pigment dispersion in water-reducible paints, Prog. Org. Coat. 28 (2) (1996) 79–87.
4. Conclusions The objective of this work was to study the effect of a thermal recycling process on the surface properties of alumina coated TiO2 pigment recovered from a paint matrix. The recycled pigment was analysed using powder x-ray diffraction (XRD), surface area measurements (BET), laser diffraction for particle size analysis and X-ray photoelectron spectroscopy (XPS). Investigations on the zeta potential and the surface charge were also performed. The recycling process was not shown to affect the rutile crystalline core nor the pigment chemical composition. However, the presence of the other paint components led to a significantly higher salt content in a suspension of recycled pigments when compared to its virgin counterpart. The surface elements of the recycled pigment were similar to the virgin pigment and the ratio of alumina to titanium was kept unchanged, suggesting that the coating was unharmed by the recycling process. However, the XPS depth profile indicated some etching of the alumina coating which is important when considering the life cycle of the pigment, i.e. if it is recycled multiply times. The particle size distribution was slightly broader after the recycling process. Even though this is a crucial parameter for the performance of the pigment, it was deemed that the increase in particle size is acceptable. The main striking difference was seen for the surface charge of the pigments. They carry the same sign of charge over the measured pH range but the magnitude differs. The zeta potential showed a decrease in overall magnitude suggesting a decrease in the hydroxyl group concentration on the surface of the recycled pigment. At pH > 7 and at low salt concentrations was the surface charge of the recycled pigment higher than for the virgin pigment. At higher salt concentrations (100 mM NaCl) the difference in surface charge of the two samples was negligible. Overall the recycling process does have minor effects on the recycled pigment. The surface properties between the virgin and the recycled pigment appeared to be very similar. This shows that calcination is a sound route for pigment recycling. However, the presence of salts, as well as the change in hydroxylation state of the surface suggest that the pigment would need further purification and pre-treatment before being introduced as a replacement to virgin pigment. Acknowledgments The authors acknowledge thefinancial support from Swedish Governmental Agency for Innovation Systems (grant number: P375471), Akzo Decorative Paints UK, and Stena Metal AB Sweden. John Steele from Akzo Decorative Paints UK is acknowledged for his assistance. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.porgcoat.2018.09.012. References [1] Survey, U.S.G, Mineral Commodity Summaries 2012: U.S. Geological Survey, (2012). [2] M.A. Imam, F.H. Froes, K.L. Housley, Titanium and titanium alloys, Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Inc., 2000. [3] TZMI, Introduction to the Titanium and Zirconium Value Chains, (2013). [4] J.H. Braun, A. Baidins, R.E. Marganski, TiO2 pigment technology: a review, Prog. Org. Coat. 20 (2) (1992) 105–138. [5] S. Middlemas, Z.Z. Fang, P. Fan, LCA comparison of emerging and traditional TiO2 manufacturing processes, J. Clean. Prod. 89 (2015) 137–147. [6] COMMISSION DECISION of 28 May 2014 establishing the ecological criteria for the award of the EU Ecolabel for indoor and outdoor paints and varnishes, T.E. COMMISSION, Editor. 2014.
286
Contents lists available at ScienceDirect
Progress in Organic Coatings journal homepage: www.elsevier.com/locate/porgcoat
Surface properties of recycled titanium oxide recovered from paint waste a,⁎
b
a
c
T
d
Mikael C.F. Karlsson , Zareen Abbas , Romain Bordes , Yu Cao , Anders Larsson , Antonin Rollanda, Phil Taylore, Britt-Marie Steenaria a
Department of Chemistry and Chemical Engineering, Chalmers University of Technology, Kemivägen 4, 412 96 Gothenburg, Sweden Department of Chemistry and Molecular Biology, University of Gothenburg, Kemigården 4, 412 96 Gothenburg, Sweden Department of Industrial and Materials Science, Chalmers University of Technology, Hörsalsvägen 7B, 412 58 Gothenburg, Sweden d RISE Research Institutes of Sweden, The Unit of Surface, Process and Formulation, Drottning Kristinas Väg 45, 114 28 Stockholm, Sweden e AkzoNobel Decorative Paints Global Research & Development Open Innovation, AkzoNobel Decorative Paints, Wexham Road, Slough Berkshire, SL2 5DS, UK b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Titanium oxide Pigment Recycling Paint Zeta potential Surface charge
Aluminium oxide coated rutile pigment was extracted from a paint matrix by means of a thermal recycling process. The objective was to investigate the effect of the recycling process on the surface properties of the pigment. The pigment was analysed using powder x-ray diffraction (XRD), surface area measurements (BET), laser diffraction for particle size analysis and X-ray photoelectron spectroscopy (XPS) before and after the recycling process. Investigations on the zeta potential and the surface charge were performed as well. It was concluded that the rutile crystalline core and the aluminium oxide coating of the pigment were still intact after the recycling process. The particle size distribution of the recycled pigment was slightly broader compared to the virgin pigment. The measured magnitude in zeta potential of the recycled pigment was lower than for the virgin pigment. This difference is thought to be caused by alteration in the surface hydroxyl concentration. Surface charge titrations showed differences between the virgin and the recycled pigment at alkaline pH and at low salt concentrations.
1. Introduction Titanium minerals are of great technical importance in today´s society. In 2011, the world production of titanium containing minerals was 6.7 million metric tonnes [1]. Of that, the vast majority was used to produce TiO2 for use as a white pigment in paints, coatings, plastics and papers [2,3]. TiO2 is the major white pigment used by the coatings industry [4] due to its abundance and its ability to scatter visible light while being chemically inert. Typically, TiO2 is produced using a sulphate or chloride route, both of which have a high carbon footprint per kilogram of TiO2 produced. It has been reported that even if new innovative production routes are developed the carbon footprint per kg TiO2 produced would still be high [5]. The European Union (EU) has recognized the environmental impact of TiO2 production and has consequently set restrictions on the amount of TiO2 allowed to be used in paint formulations if the paints are to qualify for the voluntary Ecolabel labelling [6]. In the future, waste management of old paint residues may also be included in the Ecolabel criteria [7]. Even before regulatory guidelines were in place, the coatings industry has strived to find a replacement for TiO2 due to its relatively ⁎
high cost compared to other coating components [8]. One appealing approach is the recovery of TiO2 from paint waste, as it could benefit the coatings industry in two ways. First, recovered TiO2 may be a cheaper and more environmentally friendly replacement for virgin TiO2 produced by conventional high carbon footprint routes. Second, a successful TiO2 recovery process could be a cornerstone in the waste management of old paint residues and production waste from manufacturing plants. TiO2 is often mentioned as the major white pigment but commercial grade pigment is frequently not pure TiO2. The crystalline TiO2 core is the active ingredient that provides the pigment's optical functions. However, the surface of the TiO2 particles does not consist solely of titanium and oxygen. As the TiO2 crystals grow during manufacturing, insoluble components accumulate on the surface. The components can be impurities from the ore or additives introduced deliberately to control the crystal structure, crystal growth, and particle agglomeration [9,10]. Besides these contaminants, the surface of a pigment is deliberately altered to suit the final application of the pigment. For instance, to reduce photoactivity and to improve the compatibility between the pigment and the other paint components, the surface of the TiO2 pigment is commonly treated with silicon, aluminium and zirconium
Corresponding author. E-mail address: [email protected] (M.C.F. Karlsson).
https://doi.org/10.1016/j.porgcoat.2018.09.012 Received 28 March 2018; Received in revised form 31 August 2018; Accepted 1 September 2018 0300-9440/ © 2018 Elsevier B.V. All rights reserved.
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
properties of the recycled TiO2 and the effects of the recycling process on the TiO2 pigment specifically. The formulation, mixing and paint preparation was performed at Akzo Decorative Paints in Slough, UK.
oxides. The surface of uncoated TiO2 pigment shows a complex behaviour with different active groups, but a coated pigment can have an even more complex surface that is chemically very different to the bulk TiO2 phase [11]. For the optimum opacity and durability of a paint film, a well dispersed pigment is a must. So, although it is no uncomplicated task, the surface characterization of a coated pigment is of importance in understanding pigment-medium interactions and its performance the final application [12]. In a previous TiO2 recycling study [13] a white paint was pyrolysed at 500 °C in order to recover the inorganic components. In that study, TiO2 was extracted together with the mineral extenders present in the paint, generating a pigment-extender mix consisting of roughly 45% TiO2 and 55% mineral extenders, mainly dolomite. The recycled material was then used in a new paint formulation. It was concluded that using pyrolysis as a means to recycle TiO2 resulted in a pigment-extender mix that could be used in a low-quality, flat paint, formulation if it was not for the large, millimetre size, agglomerates present in the paint. Since the recovered fraction was a mix of inorganic components, it was not possible to study in detail the agglomeration mechanism or the effect of the recycling process on the properties of the TiO2 pigment. Therefore, it was decided to perform the next step of the work using a model paint based on TiO2 as the sole pigment. In the present work, a paint formulation based on alumina coated TiO2 pigment (rutile) was prepared. The pigment was extracted from the paint matrix through a pyrolysis based recycling process. The pigment was analysed prior to and after the recycling process in terms of surface chemistry, both with electrokinetic measurements and with direct titration of surface groups. Additionally, the pigment were characterized using powder x-ray diffraction (XRD), surface area measurements (BET), laser diffraction for particle size analysis and X-ray photoelectron spectroscopy (XPS). The main purpose of this study was to understand the effects of the thermal recycling process on the surface characteristics of the pigment and identify the changes caused by the process. The knowledge of the surface characteristics will be useful in future studies when evaluating the dispersion stability of the pigments and their interactions with other paint components, such as dispersants and binder.
2.2. Recovery process The TiO2 was recovered from the model paint presented in Table 1 by means of a pyrolysis process. The temperatures chosen in the thermal recovery process were based on previously published results [13]. By heating the paint in an inert atmosphere, the organic fraction was volatilized, separating it from the inorganic fraction. The volatilized organics were cooled and collected as an oil with the potential to be used as feedstock for a chemical process or as an energy source. After pyrolysis, the inorganic fraction was oxidized in air to remove residual carbon and non-pyrolyzed material, improving the colour of the recovered pigment. To further purify the recovered material it was dispersed in water with a mixture of ion exchangers. To finalize the recycled pigment the ion exchangers were separated from the pigment dispersion, which was dried and homogenized into a fine powder. A schematic of the recycling process can be found in the supplementary information. 2.2.1. Pyrolysis The pyrolysis experiments were conducted in a Rohde, ME 45-13 furnace, fitted with a pyrolysis retort in corrosion resistant steel. The inner dimensions of the pyrolysis retort were 300 × 300 × 150 cm. During the experiments, the temperature was controlled using a TC 504 temperature controller. In addition, an external thermocouple was used to monitor the furnace temperature in relation to the set temperature. Before each experiment, the retort was filled with nitrogen gas (purity 99.9%) to an overpressure of 0.5–1 bar, emptied, and refilled again with nitrogen to overpressure. This was repeated three times to create an oxygen free environment. The paint was dried at 150 °C, followed by pyrolysis done under atmospheric pressure at 500 °C. During the pyrolysis experiments, the retort was flushed with nitrogen, 0.85–0.95 L/ min. The volatilized organics and liquid produced during the pyrolysis were led out through the retort outlet. In the present work, the focus was on the inorganic components of the sample materials and therefore the oil and gas fractions were not collected for analysis. To reduce the exposure of hazardous volatile pyrolysis products, the sample was cooled to 50 °C, under continuous nitrogen flow, in the pyrolysis retort before being removed. Once cooled, the solid residue was collected and weighed. The solid product was finally homogenized with a mortar and pestle.
2. Materials and methods 2.1. Model paint A white model paint was produced according to the formulation shown in Table 1. A commercially available pigment quality TiO2, rutile coated with aluminium oxide, and all other paint components were acquired from Akzo Nobel Decorative Paints, UK. Typically, waste paint streams do not only contain TiO2 as the inorganic component. Inorganic extenders are used in paints to give the final coating the desired properties, such as gloss or scrub resistance. However, those components were excluded from the model formulation. A simplified version of a paint waste stream will allow for a more in-depth analysis of the
2.2.2. Oxidation The pyrolysis product mainly contained the inorganic components in the paint formulation (in this case, TiO2) and carbon residues from pyrolysis of the organics. To remove the residual carbon and non-volatilized organic material, the pyrolysis residue was oxidized by spreading the powder in a thin layer (< 4 mm thick) in an alumina crucible and heat treating it in 470–500 °C for 1.5 h in air. After cooling, the material was collected, weighed, and homogenized using a M20 universal mill from IKA.
Table 1 Model paint formulation produced for the TiO2 recovery process. Raw Materials
wt.%
Water Binder (vinyl acrylic) Antifoam Non-ionic surfactant Anionic surfactant Hydroxyethylcellulose (HEC) TiO2 Incan preservative pH modifier
22.92 44.25 0.51 0.51 0.46 0.51 30.47 0.12 0.25
2.2.3. Purification of pigment with ion exchanger resin During the pyrolysis process all organic components of the paint are decomposed into smaller entities and volatized, and were therefore separated from the inorganic fraction of the paint. This inorganic fraction would contain the inorganic pigments and unwanted salt residues from the original paint. In a later stage, when the recycled pigments were re-dispersed in an aqueous system, these ionic species would dissolve and increased the ionic strength of the solution. It is known that an increased ionic strength have negative effects on colloidal stability [14] and the performance of certain dispersing aids commonly used in paints [15]. Therefore a washing of the pyrolysis product was designed. 280
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
thickness of 100 nm.
The oxidized residue (see Section 2.2.2) was dispersed in distilled water. Ion exchanger resin (Amberlite IR120, hydrogen form) was added to the suspension under continuous stirring and pH measurement. The amount of ion exchanger added was calculated to present an excess of exchanging sites. pH rapidly decreased from the initial pH (> 9) to stabilize around pH 2.5–3.0 after 5 min. After 30 min the suspension was neutralized using an ionic exchanger resin (Amberlite IRN78, hydroxide form) under continuous stirring. pH slowly increased to stabilize around 7 after 1.5 h. The pigment suspension with the 2 ionic exchangers was left under continuous stirring for 4.5 h. The pigment was separated from the ionic exchangers through sieving and the recovered pigment was dried in 60 °C, before being homogenized into a fine powder with a mortar and pestle. This product is referred to as recycled pigment, unless stated otherwise. Conductivity was measured to get an indication of the effect of the washing step. To this end the virgin pigment, the oxidized pyrolysis residue and the recycled pigment was dispersed (10 wt. %) in Milli-Q water (> 18.2 MΩ/cm). After 96 h mixing, samples of the water phases were collected, diluted with Milli-Q water and analysed with inductively coupled plasma mass spectrometry (ICP-MS) for cations (atomic mass ranges 7–238) and ion chromatography (IC) for anions (Cl−, PO43−, NO3−). The measurement of conductivity and ion concentration was done in triplicate. All the conductivity measurements were done with a 5-ring conductivity measuring cell calibrated with a 100 μS/cm conductivity standard from Metrohm. All pH measurements were done with a LLAquatrode plus from Metrohm calibrated using a 5-point calibration (pH 2, 4.01, 7, 10 and 12, Certipur, Merck). The Amberlite IR120 (in the hydrogen form from Sigma-Aldrich) was washed with Milli-Q water and regenerated with 0.1 M HCl before use. The Amberlite IRN78 (in the hydroxide form from Sigma Aldrich) was used as received from supplier. The ICP-MS instrument was a Thermo iCap Q and the IC a Dionex DX-100 with anion column IonPac AS-4A SC.
2.3.2. Particle size and BET The particle size distribution was measured by laser diffraction, using a Mastersizer MicroPlus from Malvern Instruments. The particles were measured in a dilute suspension of deionized water. Before measurement, the suspension was treated with ultrasound (200 W for 40 s applied to 30 mL of 1 wt.% suspension, using a Sonics VibraCell sonicator) to break up agglomerates. The specific surface area of the virgin pigment and the recycled pigment was determined by BET measurements as described by Brunauer et al. [16]. The exact procedure for the experiments is described elsewhere [13]. 2.4. Dynamic mobility and zeta potential The zeta potential analysis was performed using a ZetaProbe from Colloidal Dynamics. The ZetaProbe operates at a field frequency range of 0.3–3 MHz. When the alternating field is applied, the particles (and ions) start to oscillate. As the particles move, they displace an equal volume of liquid. If the density of the particles is different from the density of the liquid, this will give rise to a sound wave. This so called Electrokinetic Sonic Amplitude (ESA) signal depends on the particle velocity and the dynamic mobility, μd, can be determined from this. The dynamic mobility of the particles depends on their size, their zeta potential and the frequency of the applied field. For dilute colloidal dispersions (up to about 5 vol percent), the relationship between the ESA effect and the dynamic mobility, μd, of the charged particles is given by [17]
ESA = A (ω) ϕ
Δp μ po d
(1)
where ESA is measured in (Pa/(V/cm)), A(ω) is an instrument factor, φ is the volume fraction of the particles and Δρ is the density difference between the particles and that of the solvent (ρ0). The ZetaProbe measures the maximum in the pressure of the wave (Pa) per unit applied electric field strength (V/cm). The ZetaProbe measures the ESA at 8 different frequencies and obtains a mobility spectrum of the sample. The zeta potential is then calculated from the dynamic mobility spectrum using the O´Brien formula, which is valid for spheroidal particles with thin electrical double layers [18]. Before the measurement, as the presence of excess ions can affect the measurement, the virgin pigment was cleaned with the use of ion exchangers with the same procedure as used for the recycled pigment, see Section 2.2.3. The measuring cell contained 250 mL of sample with a stirring speed of 150 rpm. The charge of the particles was determined in a pH range from 4 to 10 by titration using a particle concentration between 1–5 wt. % in 2, 10 mM NaCl and 10 mM NaNO3 (≥99.0% ACS regents from Merck dissolved in distilled water) using a computer controlled titrator. A particle concentration below 2.5 wt.% had a significant influence on the results, for that reason 5 wt.% was chosen for this study. The pH-electrode was calibrated daily using a 3-point calibration (pHstandards 4, 7, 10) with standards from Hamilton. The ESA signal was calibrated daily with a 0.25 S/m Potassium 12-tungstosilicate hydrate standard from Colloidal Dynamics. TiO2 with a density of 4.26 g/mL and a dielectric constant of 40 and the properties for water at 20 °C was chosen calculating the zeta potential. 1 M HCl and 1 M HNO3 was used for titration of samples dispersed in NaCl and NaNO3, respectively. For both electrolytes 1 M NaOH was used as base. All bases and acids were prepared from Titrasol concentrates (Merck). All data presented was corrected for the signal from the background electrolyte, although not corrected for change in pH. The influence of pH was checked, and the signal contribution was less than 2% except when approaching a pH of 2, where the influence could be up to 10% of the total ESA signal. As titration was done in pH 4–10 this contribution was neglected.
2.3. Characterization of recycled pigment Characterization of the recycled pigment was done with the virgin pigment used in the model paint, see Table 1, as a reference. The purpose of the characterization was to compare the properties of the virgin and recycled pigment and to see the effect of the recycling process on the pigment. 2.3.1. X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) To confirm that the main crystal structure of the rutile was intact, the virgin pigment and the recycled pigment were analysed with X-ray diffraction (XRD). This procedure has been described previously [13]. The virgin pigment and the recycled pigment were both analysed with X-ray photoelectron spectroscopy (XPS) using a Versaprobe III from Physical Electronics equipped with monochromatic Al Kα X-ray source (1486.6 eV). The virgin pigment was cleaned with the use of ion exchangers, as described in Section 2.2.3, prior to XPS analysis. The Xray beam diameter was 100 μm and the power was 25 W. Acquisition conditions for the survey spectra (0–1100 eV) were 112 eV pass energy, 45° take off angle and 0.5 eV/step. Selected region spectra were recorded covering the Ti 2p, Al 2p, O 1s, Si2p and C 1s photoelectron peaks. The acquisition conditions were, then, 55 eV pass energy, 45° take off angle and 0.1 eV/step. Samples were mounted on a steel sample holder using double sided tape. Dual beam charge compensation was carried out by flooding the sample with low energy electrons and low energy argon ions. The C 1s peak from adventitious C located at 284.8 eV was used as energy scale reference. The element distribution in depth was estimated by depth profiling performed using Ar+ ion sputtering with 2 kV accelerating voltage and 7 mA current. The etch rate was 5.4 nm/min calibrated on a Ta2O5 standard sample with a
281
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
2.5. Surface charge titrations
Table 2 Conductivity and selected ion concentration of 10 wt.% pigment suspensions. Uncertainties given are one standard deviation from the mean.
The point of zero charge for the virgin and the recycled pigment was determined experimentally by potentiometric titration. A Metrohm 905 Titrando equipped with stand, stirrer, acid and base dispensers and a pH-electrode (LL-Aquatrode plus). The pH-electrode was calibrated using a 5-point calibration (pH 2, 4.01, 7, 10 and 12, Certipur, Merck). The Tiamo software from Metrohm was used to monitor and control all calibrations and titrations. Prior to the titration was the virgin pigment was cleaned with the use of ion exchangers with the same procedure used for the recycled pigment, described in Section 2.2.3. Acids, bases, and background electrolytes for the titration were made from Titrasol concentrates and NaCl respectively KCl salts dried before use. All dilutions were made using Ultra pure Milli-Q water. The following reagents were used: 2 mM, 10 mM and 100 mM NaCl respectively KCl solutions were used as the background electrolyte. For each electrolyte concentration, 0.1 M NaOH and 0.1 M HCl, with the same salt concentration as the electrolyte, were prepared as well. A dispersion of 10.00 ± 0.01 g pigment and 50 mL background electrolyte was made. The dispersion was acidified with 2 mL acid for 30 min and then titrated with 4 ml of base added in 50 μL increments. Between each step, the potential was measured. The time between steps was 60–600 s, depending on equilibrium time. Equilibrium was considered to be reached if the change in the potential was less than 0.5 mV/min, corresponding to about 0.01 pH-unit/min. After the titration, the potential of the pH-electrode was converted into pH-units using the latest pH-calibration curve. For the background titrations, pure electrolytes were titrated similarly but the maximum time between each titration step was increased to 1800 s. The increased waiting time was necessary in order for the system to reach equilibrium in each step, especially for points around 0 mV. Experiments were also performed where the pigment dispersion and the background electrolyte respectively were continuously bubbled with nitrogen gas (purity 99.9%) during the titration. This was done to remove CO2, which can dissolve into solution and potentially affect the measurement. It was concluded that the nitrogen bubbling only had a minor effect on the final result, for which reason all results presented in this article are based on titrations made without nitrogen bubbling. The surface charge at each ionic strength was calculated, analogous to Kosmulski [19], by
σ0 =
FcΔV mA
Sample
Conductivity [mS/ cm]
Na [mM]
Al [mM]
Cl [mM]
Virgin pigment Recycled pigment prior to washing Recycled pigment after washing
0.2 ± 0.1 2.4 ± 0.1
11 ± 1 130 ± 10
< 0.1 13 ± 2
0.2 ± 0.1 0.1 ± 0.1
< 0.001
< 1.0
< 0.1
< 0.1
3.1.1. Purification of pigment with ion exchanger As can be seen in Table 2 the suspension of the recycled pigment had a conductivity compared to the suspension of virgin pigment. Of all the ions measured (see Section 2.2.3) the difference between the suspensions in the Na and the Al ion concentrations (see Table 2) were the most important. The increased Na and Al concentrations for the recycled sample are likely to come from the ionic dispersing aids commonly used in paint formulations [20], from the counter ion of the stabilizer of the binder and from the inorganic coating of the pigment particles. The increase in ion concentration could have a negative effect on the stability of the dispersion [14] and the functionality of used dispersants in a paint system [15]. After purification with ionic exchangers all the measured ion concentrations for the recycled pigment were below that of the virgin pigment. 3.2. Characterization of pigment 3.2.1. X-ray powder diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) The XRD spectrum for the residue confirmed that the rutile structure of the recycled pigment was intact. This is as expected, because the temperatures in the recovery process are relatively low and the results are analogous with previously published results [13]. The main purpose of the XPS analyses was to identify possible differences between the virgin and the recycled pigment. In the XPS spectrum, five elements were distinctly detected: C, Ti, Al, Si, and O. The C 1s peaks for both samples (see Fig. 1a) are identical and will therefore be treated as adventitious carbon, thus used as the energy reference by setting the main line of the C 1s spectrum to 284.8 eV. The peaks for Ti p3/2 and Ti p1/2 (shown Fig. 1b) were located at 458.2 and 463.8 eV respectively, as expected for TiO2 [21]. In Fig. 1c it can be seen that the Al 2p peak is located at 74.1 eV for both samples. This matches the binding energy of an aluminium oxide, Al2O3, Al (OOH) or Al(OH)3. No further investigation were carried to disambiguate the exact nature of the signal, as this task is complex [22] and outside the scope of this study. The center of the Si 2p peak for both the virgin and the recycled was located on 102.1 eV (Fig. 1d). It is common to coat pigments with silica (crystalline or amorphous SiO2) but the Si 2p peak at 102.1 eV did not correspond the binding energy of SiO2, which is located at higher energies (around 103.5 eV) [23]. This suggest that the silicon is in the valence state +3 [23], thus indicating that Si is a contaminant that has accumulated at the surface. It should be mentioned that siloxane compounds provided a good fit of the peaks, with Si 2p, O 1 s and C 1 s at 102.1, 532.4 and 284.4 eV respectively [24]. These types of compounds are sometimes used to improve the dry flow of pigment particles [9]. On the other hand, it has been reported that these kind of compounds degrade in the temperature range [25] used in the recovery process presented in this paper. As the Si 2p is clearly visible in both the virgin and the recycled sample is it more likely that the Si is incorporated in the matrix as contaminate rather than in the form of siloxane. A XPS survey spectra (see Supplementary information) of the virgin pigment collected prior to the purification step with ion exchanger (see Section 2.2.3) is lacking the Si 2p peak at 102.1 eV. This
(2)
where F is Faradays constant, c the concentration of acid or base, m the mass of solid particles, A the specific surface area measured by BET and ΔV the difference in volume of acid and base to reach the same pH in the dispersion as in the electrolyte solution.
3. Results and discussion 3.1. Recovery process The total TiO2 content of the processed model paint (see Table 1) was 30.47 wt. %. The average weight of the residues after pyrolysis was 32.8 ± 0.2 wt. % of the starting material, while the average weight of the residue after oxidation was 30.5 ± 0.2 wt. % of the starting material. This suggests that, after pyrolysis, the remaining organic material and residual carbon are effectively removed during the oxidation step. There was no visible colour difference between the virgin and the recycled pigment. These results show that it is possible to extract TiO2 (or other inorganic pigment) from the paint matrix by means of pyrolysis and oxidation. 282
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
Fig. 1. XPS photoelectron peaks a) C 1s, b) Ti 2p, c) Al 2p and d) Si 2p for virgin pigment (solid line) and recycled pigment (dashed line). Energy is normalized and the peak for recycled pigment is shifted to higher intensity for clarity. Peaks are relative to C 1 s = 284.8 eV.
belonging to the oxide bulk structure. The second peak, around 532 eV, is due to surface hydroxyl groups. The third peak, with the highest energy at roughly 534 eV, is from chemisorbed water. As the other components have their chemical state intact, the most likely explanation is a difference in concentration of water that has been chemisorb from the ambient surroundings on to the surfaces. This chemisorb water would increase the intensity at higher energies. It is also possible that the water could cover the surface and reduce the measured intensity at lower energies (bulk O2- oxygen). Thus, the alteration in the O 1s spectrum is probably not coming from an actual chemical change but rather from increased amount of chemisorbed water on the recycled pigment compared to the virgin pigment. Another aspect of importance is the concentration of the different species at the surface. One of the most important characteristics for a coated pigment is the ratio of bulk Ti to coating elements [28–30], in this case Al. In Fig. 3 a depth profile for the ratio between Ti and Al is shown for the virgin pigment and the recycled pigment. Si was excluded as its chemical state does not suggest a traditional coating characteristic. As seen in Fig. 3 the ratio of Ti and Al is intact at the surface after the recycling. However, going further into the particles the Al seems to drop in concentration more rapidly than for the virgin pigment. This indicates that some of the coating has been etched away during the recycling process. This corresponds well with the increased Al concentration measured in a suspension of non-purified recycled pigments (see Table 2). Overall, it seems that the recycling process does not alter the surface by the creation of new species but it reduces the thickness of the alumina coating.
Fig. 2. O 1s XPS peak for virgin pigment (solid line) and recycled pigment (dashed line). Energy is normalized and the peak for recycled pigment is shifted to higher intensity for clarity. Peaks are relative to C 1s = 284.8 eV.
leads to the conclusion that the Si peak is a contamination from the purification step. In comparison with the other element peaks, the O 1s signal showed more pronounced differences between the virgin and the recycled pigment (Fig. 2). Whereas the main centre of the peaks was at the same binding energy, virgin sample exhibited a shoulder towards lower energies while the recycled sample has a shoulder towards higher energies. This difference cannot reasonably be attributed to an alteration of the surface species, as all the other major elements exhibited minor changes. It has previously been documented that three different peaks O 1s can be observed for metal oxides [26,27]. The lowest energy peak (around 529 eV) is from the contribution of the O2− oxygen atoms
3.2.2. Particle size and specific surface area The particle size distribution of the virgin and the recycled pigment is show Fig. 4. TiO2 virgin pigment particles has a narrow particle size distribution, centred around 300 nm in diameter, in order to scatter light most efficiently [31,32]. Compared to the virgin material the recycled sample had a slightly broader particle size distribution. Published specific surface areas for rutile pigments with various 283
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
Fig. 5. Measured zeta potential for 5 wt. % pigment suspensions in 2 mM NaCl (■), 10 mM NaCl (●), and 10 mM NaNO3 (▴) as a function of pH for virgin (red) and recycled (blue) pigments. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Table 3 IEPs for 5 wt. % virgin and recycled pigment suspensions.
Fig. 3. Cation depth profile of virgin (red) and recycled (blue) pigments. Only Ti and Al are included. The atomic concentration of Ti (dashed lines) and Al (dotted line) as a function of sputter time. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
coatings lie in the range of 5–35 m2/g with the majority around 20 m2/g [28,29,33–35]. The BET surface area for the virgin pigment used in this work was measured to 17 m2/g and for the recycled pigment 16 m2/g. The measured surface areas are within the margin of error so it cannot be stated that there is any significant difference between the samples.
Sample
2 mM NaCl
10 mM NaCl
10 mM NaNO3
Washed virgin pigment Washed recycled pigment
7.4 7.6
7.4 7.6
7.4 7.8
Table 4 IEP values for commonly used coated TiO2 pigments.
3.3. Dynamic mobility and zeta potential The results of the zeta potential measurements are shown in Fig. 5 with the corresponding IEPs summarized in Table 3. Reference IEP values for coated rutile pigments are given in Table 4, with most pigment qualities having an IEP in the pH range of 7–8. The IEP tends to be related to the average surface-area of its components, where increased silica gives a more acidic IEP and alumina gives a more basic IEP [34]. Different IEPs for pure aluminium (hydr)oxide [36] and pure rutile [19] have been reported, use of the average IEP is suggested by the authors
Pigment
IEP
Source
Pigments coated with alumina and zirconia Pigment coated with alumina and silica Pigment coated with alumina Pure rutile Pure alumina
7.8 ± 0.3 2.8–8.1 8.2–9.2 4.9 9.2
[34] [28,29] [30] [35] [35]
of this work which is 9.0 and 5.6 respectively. Using these averages and the ratio 0.46 Ti to 0.54 Al in the outmost layer of the pigments given by the XPS data (see Fig. 3) an IEP of 7.4 was calculated, which is in good agreement with the IEP measured, for the virgin pigment. This
Fig. 4. Particle size distribution of virgin (●) and recycled (○) pigments expressed as volume percent (a) and accumulated volume percent (b). Data presented is the average of three measurements and the uncertainty is given in one standard deviation. 284
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
Fig. 6. Surface charge determined by potentiometric titrations in 1 mM (black), 10 mM (red), and 100 mM (green) NaCl solution for virgin (a) and recycled (b) pigments respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
charge at acidic pH but the recycled pigment has more surface charge at higher pH and this is especially noticeable at 1 mM and 10 mM NaCl. This is opposite to zeta potential measurements (see Section 3.3) in which the recycled pigment always showed a lower magnitude of charge compared to the virgin pigment. Such differences between the surface charge and zeta potentials have recently been analysed for SiO2 nanoparticles in NaCl solution [38]. The difference stems from the fact that decrease in zeta potential at increasing electrolyte is due to the compression of stern layer whereas increase in surface charge density at increasing electrolyte concentration is due to protonation/deprotonation reactions. The decreased zeta potential at pH > 7 for recycled TiO2 and increased surface charge density at such pH values can be rationalized due to the above mentioned reasons. Differences in surface charge between the samples could as well be explained by assuming that the exposure to water of the heat treated recycled pigment led to rehydroxylation of the surface. It is well known the rehydroxylation of a heat treated oxide surface is not instantaneous [39] and the kinetics is influenced by pH [40]. The surface charge was calculated using Eq. (2). It was assumed that no reactions takes place and that the difference in volume of acid and base to reach the same pH in the dispersion as in the electrolyte solution was due to protonation or deprotonation of surface hydroxyl groups. Formation of hydroxyl groups on the particle surface could lead to a change in pH of the suspensions, which would result in an artificially more negatively charged surface calculated with Eq. (2). This effect would be less visible in the case of zeta potential measurements, as the total titration time was much shorter, i.e. 1 h. During that time the hydroxyl formation was limited. The potentiometric titration took approximately 8 h during which the hydroxyl formation occurred and modified the surface chemistry. Although there are differences in surface charge which cannot be fully
demonstrate that the rutile core is not completely encapsulated by the alumina coating. The IEP of the recycled pigment is 0.2–0.4 pH-units higher than that of the virgin pigment (Table 3), and, as shown in Fig. 5, the magnitude of the zeta potential variation under different salt conditions was smaller for the recycled pigment compared to the virgin pigment over the measured pH scale. Moreover, the IEPs of both virgin and recycled TiO2 do not show concentration dependence which indicates that there is no specific ion interactions. It has previously been shown [37] that temperatures used in the recycling process can reduce the concentration of hydroxyl groups on surfaces. A lower concentration of hydroxyl groups would result in a lower magnitude in zeta potentials, as fewer groups are available for protonation/deprotonation. For electrostatic stabilized systems it is generally admitted that a larger magnitude of the zeta potential usually results in more stable particle suspension. In practice, the lower magnitude observed may come from partial dehydroxylation, which implies that the dispersibility by surface active agent would be less effective. 3.4. Surface charge titrations Surface charge, σ0, determined by potentiometric titrations, are shown in Fig. 6. Only results for the experiments with NaCl as electrolyte are shown as the KCl based electrolyte showed very similar results. It can be suggested that the virgin pigment has a common intercept point (CIP) for σ0 = 0 between pH 7.1–7.6. This correspond well with the measured IEP at 7.4 (see Table 3). However, for the recycled pigment no clear CIP at σ0 = 0 could be determined. When comparing the titration curves for the two pigments in respective electrolyte (Fig. 7) it can be seen that the pigments display a very similar surface
Fig. 7. Surface charge determined by potentiometric titrations in (a) 1 mM, (b) 10 mM, and (c) 100 mM NaCl solution for virgin (■) and recycled ( ) pigments respectively.
285
Progress in Organic Coatings 125 (2018) 279–286
M.C.F. Karlsson et al.
explained; these differences are most prominent at lower salt concentrations. At higher salt concentrations, which is more likely to be found in a paint formulation [41], the difference in surface charge is negligible.
[7] S. Jiannis, R.K. Kougoulis, Ben Walsh, Katherine Bojczuk, Trevor Crichton Revision of EU European Ecolabel and Development of EU Green Public Procurement Criteria for Indoor and Outdoor Paints and Varnishes, (2012). [8] M.J.A. Ruszala, N.A. Rowson, L.M. Grover, R.A. Choudhery, Low carbon footprint TiO2 substitutes in paint: a review, Int. J. Chem. Eng. Appl. 6 (5) (2015) 331–340. [9] J.H. Braun, Titanium dioxide—a review, J. Coat. Technol. 69 (868) (1997) 59–72. [10] T.C. Patton, Characterization and physical relationships, Pigment Handbook vol. 3, Wiley, New York, 1973. [11] D.H. Solomon, D.G. Hawthorne, Chemistry of Pigments and Fillers, Wiley, New York, 1983. [12] G.D. Parfitt, The role of the surface in the behavior of titanium dioxide pigments, Croat. Chem. Acta 53 (2) (1980) 333–339. [13] M.C.F. Karlsson, D. Corr, C. Forsgren, B.-M. Steenari, Recovery of titanium dioxide and other pigments from waste paint by pyrolysis, J. Coat. Technol. Res. 12 (6) (2015) 1111–1122. [14] G.D. Parfitt, Dispersion of Powders in Liquids: With Special Reference to Pigments, Applied Science Publishers, London, 1981. [15] S. Farrokhpay, A review of polymeric dispersant stabilisation of titania pigment, Adv. Colloid Interface Sci. 151 (1–2) (2009) 24–32. [16] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (2) (1938) 309–319. [17] R.W. O’Brien, The electroacoustic equations for a colloidal suspension, J. Fluid Mech. 212 (1990) 81–93. [18] R.W. O’Brien, Electro-acoustic effects in a dilute suspension of spherical particles, J. Fluid Mech. 190 (1988) 71–86. [19] M. Kosmulski, Surface Charging and Points of Zero Charge. Surfactant Science Series vol. 145, CRC Press, Boca Raton, 2009, pp. 0081–9603. [20] J. Bieleman, Additives for Coatings, Wiley-VCH, Weinheim, 2000. [21] A.P. Dementjev, O.P. Ivanova, L.A. Vasilyev, A.V. Naumkin, D.M. Nemirovsky, D.Y. Shalaev, Altered layer as sensitive initial chemical state indicator, J. Vac. Sci. Technol. A 12 (2) (1994) 423–427. [22] J.T. Kloprogge, L.V. Duong, B.J. Wood, R.L. Frost, XPS study of the major minerals in bauxite: gibbsite, bayerite and (pseudo-)boehmite, J. Colloid Interface Sci. 296 (2) (2006) 572–576. [23] K.T. Queeney, M.K. Weldon, J.P. Chang, Y.J. Chabal, A.B. Gurevich, J. Sapjeta, R.L. Opila, Infrared spectroscopic analysis of the Si/SiO2 interface structure of thermally oxidized silicon, J. Appl. Phys. 87 (3) (2000) 1322–1330. [24] M.R. Alexander, R.D. Short, F.R. Jones, M. Stollenwerk, J. Zabold, W. Michaeli, An x-ray photoelectron spectroscopic investigation into the chemical structure of deposits formed from hexamethyldisiloxane/oxygen plasmas, J. Mater. Sci. 31 (7) (1996) 1879–1885. [25] G. Camino, S.M. Lomakin, M. Lazzari, Polydimethylsiloxane thermal degradation Part 1. Kinetic aspects, Polymer 42 (6) (2000) 2395–2402. [26] H. Perron, J. Vandenborre, C. Domain, R. Drot, J. Roques, E. Simoni, J.J. Ehrhardt, H. Catalette, Combined investigation of water sorption on TiO2 rutile (110) single crystal face: XPS vs. periodic DFT, Surf. Sci. 601 (2) (2007) 518–527. [27] E. McCafferty, J.P. Wightman, Determination of the concentration of surface hydroxyl groups on metal oxide films by a quantitative XPS method, Surf. Interface Anal. 26 (8) (1998) 549–564. [28] T. Losoi, Surface studies of titanium dioxide pigments, J. Coat. Technol. Res. 61 (776) (1989) 57–63. [29] K. Siwinska-Stefanska, A. Krysztafkiewicz, T. Jesionowski, Effect of inorganic oxides treatment on the titanium dioxide surface properties, Physicochem. Probl. Miner. Process. 42 (2008) 141–151. [30] M.L. Taylor, G.E. Morris, R.S.C. Smart, Influence of aluminum doping on titania pigment structural and dispersion properties, J. Colloid Interface Sci. 262 (1) (2003) 81–88. [31] J. Fisher, T.A. Egerton, Titanium compounds, inorganic, Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Inc., 2000. [32] M.P. Diebold, Application of Light Scattering to Coatings: A User’s Guide, Springer International Publishing, 2014. [33] J.-P. Boisvert, J. Persello, J.-C. Castaing, B. Cabane, Dispersion of alumina-coated TiO2 particles by adsorption of sodium polyacrylate, Colloids Surf. A: Physicochem. Eng. Asp. 178 (1–3) (2001) 187–198. [34] S. Farrokhpay, G.E. Morris, D. Fornasiero, P. Self, Influence of polymer functional group architecture on titania pigment dispersion, Colloids Surf. A: Physicochem. Eng. Asp. 253 (1–3) (2005) 183–191. [35] G.E. Morris, W.A. Skinner, P.G. Self, R.S.C. Smart, Surface chemistry and rheological behaviour of titania pigment suspensions, Colloids Surf. A: Physicochem. Eng. Asp. 155 (1) (1999) 27–41. [36] G.V. Franks, Y. Gan, Charging behavior at the alumina–water interface and implications for ceramic processing, J. Am. Ceram. Soc. 90 (11) (2007) 3373–3388. [37] M. Primet, P. Pichat, M.V. Mathieu, Infrared study of the surface of titanium dioxides. I. Hydroxyl groups, J. Phys. Chem. 75 (9) (1971) 1216–1220. [38] M.A. Brown, A. Goel, Z. Abbas, Effect of electrolyte concentration on the stern layer thickness at a charged interface, Angew. Chem. Int. Ed. 55 (11) (2016) 3790–3794. [39] S. Shioji, M. Hanada, Y. Hayashi, K. Tokami, H. Yamamoto, Kinetic study of alkoxylation and rehydroxylation reactions on silica surfaces, Adv. Powder Technol. 18 (4) (2007) 467–483. [40] R.K. Iler, The surface chemistry of silica, Chemistry of Silica–Solubility, Polymerization, Colloid and Surface Properties and Biochemistry, (1979), pp. 644–645. [41] T. Kobayashi, Pigment dispersion in water-reducible paints, Prog. Org. Coat. 28 (2) (1996) 79–87.
4. Conclusions The objective of this work was to study the effect of a thermal recycling process on the surface properties of alumina coated TiO2 pigment recovered from a paint matrix. The recycled pigment was analysed using powder x-ray diffraction (XRD), surface area measurements (BET), laser diffraction for particle size analysis and X-ray photoelectron spectroscopy (XPS). Investigations on the zeta potential and the surface charge were also performed. The recycling process was not shown to affect the rutile crystalline core nor the pigment chemical composition. However, the presence of the other paint components led to a significantly higher salt content in a suspension of recycled pigments when compared to its virgin counterpart. The surface elements of the recycled pigment were similar to the virgin pigment and the ratio of alumina to titanium was kept unchanged, suggesting that the coating was unharmed by the recycling process. However, the XPS depth profile indicated some etching of the alumina coating which is important when considering the life cycle of the pigment, i.e. if it is recycled multiply times. The particle size distribution was slightly broader after the recycling process. Even though this is a crucial parameter for the performance of the pigment, it was deemed that the increase in particle size is acceptable. The main striking difference was seen for the surface charge of the pigments. They carry the same sign of charge over the measured pH range but the magnitude differs. The zeta potential showed a decrease in overall magnitude suggesting a decrease in the hydroxyl group concentration on the surface of the recycled pigment. At pH > 7 and at low salt concentrations was the surface charge of the recycled pigment higher than for the virgin pigment. At higher salt concentrations (100 mM NaCl) the difference in surface charge of the two samples was negligible. Overall the recycling process does have minor effects on the recycled pigment. The surface properties between the virgin and the recycled pigment appeared to be very similar. This shows that calcination is a sound route for pigment recycling. However, the presence of salts, as well as the change in hydroxylation state of the surface suggest that the pigment would need further purification and pre-treatment before being introduced as a replacement to virgin pigment. Acknowledgments The authors acknowledge thefinancial support from Swedish Governmental Agency for Innovation Systems (grant number: P375471), Akzo Decorative Paints UK, and Stena Metal AB Sweden. John Steele from Akzo Decorative Paints UK is acknowledged for his assistance. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.porgcoat.2018.09.012. References [1] Survey, U.S.G, Mineral Commodity Summaries 2012: U.S. Geological Survey, (2012). [2] M.A. Imam, F.H. Froes, K.L. Housley, Titanium and titanium alloys, Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Inc., 2000. [3] TZMI, Introduction to the Titanium and Zirconium Value Chains, (2013). [4] J.H. Braun, A. Baidins, R.E. Marganski, TiO2 pigment technology: a review, Prog. Org. Coat. 20 (2) (1992) 105–138. [5] S. Middlemas, Z.Z. Fang, P. Fan, LCA comparison of emerging and traditional TiO2 manufacturing processes, J. Clean. Prod. 89 (2015) 137–147. [6] COMMISSION DECISION of 28 May 2014 establishing the ecological criteria for the award of the EU Ecolabel for indoor and outdoor paints and varnishes, T.E. COMMISSION, Editor. 2014.
286