Ceramics International 47 (2021) 18443–18454

Available online 19 March 20210272-8842/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Metal cation complexes as dispersing agents for non-aqueous powder suspensions

Paulina Zubrzycka a,b,*, Marta Radecka b, Thomas Graule a, Michael Stuer a,**

a Laboratory for High Performance Ceramics, Empa, Überlandstrasse 129, 8600, Dübendorf, Switzerland b Faculty of Materials Science and Ceramics, AGH University of Science and Technology, Al. Adama Mickiewicza 30, 30-059, Krakow, Poland

A R T I C L E I N F O

Keywords: Organic dispersions MgAl2O4

Dispersing agent Particle stabilization

A B S T R A C T

Formulation of high solid load slurries is an important aspect in achieving dense, high quality ceramic green body shaping by additive manufacturing techniques. To this end, the understanding and control of interparticle interactions in non-aqueous organic slurry systems is imperative. With nanosized powders gaining increasing interest, particle stabilization by long-chain polymeric dispersants is undesirable due to the high exclusion volume associated and the subsequent decrease in the maximum solid load of the suspension. In this study, we focus on the stabilization mechanisms provided by metal cation complexes in methyl ethyl ketone and isopropanol-based spinel slurries. Special focus is given to chromium (III) 2-ethylhexanoate. Hemi-micelle for-mation at the powder surface in methyl ethyl ketone is suggested as a highly efficient stabilization mechanism. Complexes of metal dopant cations may open new pathways towards future advances in ceramic slurry formulation, especially in the field of additive manufacturing. Furthermore, the importance of solvation forces is discussed on the basis of the protic or aprotic character of the solvent used.

1. Introduction

Production of fully dense and defect-free ceramics is an important milestone to boost the technological relevance of additive manufacturing (AM) in the ceramic industry. Achievement of highly dense and homogeneous green bodies is hereby key and remains a topic of research for many material systems [1–3]. It requires the formulation of stable powder suspensions with high solid loads that respect the rheological property requirements of a given AM technology [4,5]. Such progress will eventually allow AM of transparent ceramics with complex shapes, which have only scarcely been reported to date [6].

Within the field of AM, considerable progress has been made in recent years on the formulation of organic-based powder slurries through the evaluation of varying dispersing agents [7–16]. Using commercial dispersants such as MelPers 4343 [17] or Solsperse 41000 [18], solid loadings of ca. 48 vol% with viscosities reaching down to 1 Pa s at 200 s− 1 were achieved using powders or mixtures of powders with a dv50 particle size of 150–500 nm. Higher solid loadings (60 vol%) have only been reported by Zhang et al. [19] using much larger particle sizes (dv50 = 10.34 μm), albeit with higher viscosities (15.4 Pa s at 200

s− 1) and relying on yet another commercial dispersant (KOS110). These results illustrate the challenges associated with the increasing exclusion volume with decreasing particle size following steric or electro-steric stabilization mechanisms. The majority of the published work relies on proprietary commercial dispersants. Consequently, the lack of detailed information regarding their chemical composition leads to a rather widespread empirical slurry optimization within the AM ceramic community. The associated lack of in-depth particle stabilization knowledge subsequently leads to difficulties in reproducing the reported results or relying on pre-acquired know-how once the boundary condi-tions have slightly changed. Furthermore, it can be observed that use of sub-micron or nanoparticles is reported to a much lesser extent [20,21]. There is, however, a considerable amount of literature on non-aqueous colloidal processing. The influence of the chemical structure of disper-sants on their stabilizing effects [22–28] as well the effect of micelle formation [29,30] on the colloidal stability in organic media has been reported. Despite this body of literature, its application to ceramic powder stabilization in non-aqueous systems remains challenging, as demonstrated by numerous empirical studies.

To match the performance of conventional ceramic sintering paths,

* Corresponding author. Laboratory for High Performance Ceramics, Empa, Überlandstrasse 129, 8600, Dübendorf, Switzerland. ** Corresponding author.

E-mail addresses: paulina.zubrzycka@empa.ch (P. Zubrzycka), michael.stuer@empa.ch (M. Stuer).

Contents lists available at ScienceDirect

Ceramics International

journal homepage: www.elsevier.com/locate/ceramint

https://doi.org/10.1016/j.ceramint.2021.03.168 Received 4 January 2021; Received in revised form 10 March 2021; Accepted 17 March 2021

Ceramics International 47 (2021) 18443–18454

18444

the use of sub-micron powders is crucial. With decreasing particle size, the relative volume occupied by long-chain polymeric dispersants be-comes increasingly significant. This proportionately increasing exclu-sion volume directly prevents formulation of high solid load particle slurries [17,31,32]. Dispersants for organic slurry formulations there-fore have to be optimized with respect to their molecular weight, conformation and charge. Hereby, a sharp steric repulsion increase at the lowest interparticle distance will allow minimization of the differ-ence between absolute and effective solid loading. This requires a fine-tuned affinity of the dispersant for the powder and the solvent. Furthermore, the solvent must be selected with respect to its effects on powder particle solvation, particle charging, configuration of adsorbed species (e.g. Flory-Huggins parameter) and micelle formation.

For aqueous systems, interparticle forces and certain slurry proper-ties can be predicted and studied using the DLVO approach [33–38]. For AM, weakly flocculated suspensions with close secondary minima are generally preferred to maximize the solid load while keeping problems linked to powder settling at a minimum. Attempts to do so using the well-known DLVO approach have been reported [32,39–41]. The role of non-DLVO forces (i.e. solvation forces due to solvent polarity), however, limit the success in accurately predicting the final slurry properties [42].

In this work, we explore new pathways for the formulation organic Mg-Spinel powder slurries. The insights gained in this study may be applied in future to the formulation of UV-curable organic slurries for AM (e.g. Digital Light Processing technology). We examine the use of metal cation complexes with varying cation valence as dispersing agents to control the interparticle forces. To investigate the interactions be-tween solvents, metal-complexes and particles from a more fundamental standpoint, single solvent systems were chosen. The particle stabiliza-tion mechanisms brought forward in the present work may lay out the pathway towards a more knowledge-based selection of appropriate (dopant compatible) dispersing agents for future slurry formulation in the AM ceramic field.

2. Experimental

2.1. Materials and reagents

The powder used in this study was MgAl2O4 (S25CR, Baikowski, France) with a specific surface SBET = 23.1 m2/g, dv,50 = 0.2–0.3 μm and ≥99% spinel crystal phase according to the supplier data sheet and confirmed by in-house measurements after milling. The powder was dispersed in Methyl Ethyl Ketone (MEK) (≥99.5%, Merck, Germany) and Isopropanol (IPA) (Reag. Ph Eur ≥99.8%, analysis grade EMSURE® ACS, Merck, Germany) respectively (see Table 1). Chromium (III) 2-eth-ylhexanoate 70 wt% in mineral spirits (abcr GmbH, Germany) and Zinc 2-ethylhexanoate 80 wt% in mineral spirits (abcr GmbH, Germany) were evaluated as dispersants as well as 2-ethylhexanoic acid (≥99%, Sigma Aldrich).

2.2. Suspension preparation

All suspensions were prepared according to the following procedure (see Fig. 1). First, the desired amount of dispersant was mixed with the selected organic dispersing liquid (MEK or IPA) and stirred with a magnetic stirrer for 10 min in a covered polypropylene (PP) container Prior to powder addition, a first conductivity measurement was per-formed. Powder was then added to reach slurries with 10 vol. % solid loading. The slurries were then first stirred with a spatula before vigorous shaking for 30 s and an ultrasonic bath treatment for 10 min. The slurries were finally left ball rolling on a rolling bench for 24 h (with 15 g of 3 mm ZrO2 balls for 60 g of powder slurry) for final homoge-nization and stabilization, prior to slurry characterization. This pro-cedure was chosen well aware that it would not lead to full powder deagglomeration, but to provide equal particle distributions to compare the effects of the various dispersing conditions. A final planetary ball- milling step was only performed in the case of Cr3+-complexes in MEK and IPA respectively to confirm their effectiveness as dispersants in the de-agglomeration process. Note that the acid concentrations were cho-sen to compare the effect of 2-ethylhexanoate group additions over an equal concentration range compared to the metal complex additions and ensure that the full powder surface coverage was reached and exceeded.

2.3. Characterization methods

A conductometer (712 Conductometer, Metrohm, Switzerland) equipped with a conductivity measuring cell for low conductivity media (c = 0.1 cm− 1) and a temperature sensor Pt1000 was used to evaluate the conductivity of the electrolyte.

Turbidimetry (Turbidimeter Eutech TN-100, Fisher Scientific, USA) was used to detect the potential presence of micelles in the supernatant following slurry centrifugation.

The zeta-potential was evaluated using Laser Doppler Electropho-resis (Malvern Zetasizer Nano ZS, UK) using the Huckle model. A glass dip cell (Universal dip cell kit, Malvern, UK) was used to avoid solvent compatibility issues.

The particle size distribution was measured using static light scat-tering (LS 13320, Beckman Coulter, USA) in MEK or IPA to match the dispersing liquid of the sample to be analyzed. Mie scattering theory was used for the calculation of the particle size from the raw scattering data.

The absorbance spectra of the received and surface-modified dry powder samples were measured by a Fourier transform infra-red spec-trometer (FTIR Bruker Tensor 27, USA) to gain some qualitative infor-mation regarding the surface modification step of the powder.

The powder sedimentation velocity was determined following the sediment height at 24 h intervals over one week.

The rheological behavior of the slurries was measured using a Rheometer (MCR302, Anton Paar, Austria) with a concentric cylinder setup (CC27/T200/SS) with the temperature regulated at 25 ◦C.

Table 1 Properties of used solvents.

Solvent Dynamic viscosity (mPa∙s) at 20 ◦C Dielectric constant at 20 ◦C (− ) Density at 20 ◦C (g/cm3) Hydrogen bond index (1-10) Purity

IPA - isopropanol 2.4 18.6 0.785 8.9 ≥99.8% MEK - methyl ethyl ketone 0.42 18.5 0.805 5.0 ≥99.5%

Fig. 1. Flowchart of suspension preparation procedure.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18445

3. Results and discussion

3.1. Conductivity measurements and turbidimetry

The conductivity measurements were performed to evaluate the conductivity increase of the suspension following dispersant addition (i. e. complex or acid). The rate of conductivity increase depends on the dispersant dissociation and adsorption ratio respectively as a function of the dispersant concentration prior to surface saturation and on disper-sant dissociation only afterwards. Two measurements were performed: one in the solution prior to powder addition, and one in the supernatant after slurry centrifugation.

From Fig. 2a) and b) and c) it can be seen that the conductivity in-creases with acid/complex addition, and particularly so in the presence of spinel powder. This conductivity increase with increasing dispersant concentration suggests at least a partial dissociation – and thus ionic nature – of the added complexes and acids (Fig. 2). The strongest con-ductivity increase is observed for chromium complexes, where 232 μS/ cm is reached at 0.145 mol/dm3 in the supernatant of a MEK-based suspension. This compares with 12 μS/cm for Zn2+-complexes and 2 μS/cm for acid additions in MEK. The observed conductivity increase in the surfactant upon powder addition furthermore indicates that surface adsorption of selected species occurs, increasing the overall ionic strength of the suspension.

In Fig. 2 d) molar conductivity (Λm = κ/c) of Cr3+/MEK is plotted as a function of the square root of the complex concentration (

c

√). The

strong change in molar conductivity as a function of c

√depicts a weak

electrolyte nature of the chromium complex in MEK. This further illus-trates that the dissociation constant is small and that the conductivity increase in the supernatant relates to surfactant adsorption on the powder surface, limiting the ion release in the liquid as a result of additional complex dissociation.

From Fig. 2, one can also observe that there appears to be a certain

“critical” complex/acid concentration – decreasing with increasing cation charge – above which the conductivity of the supernatant in-creases noticeably with respect to the initial electrolyte solution. Although the origins of this critical concentration have not been further elucidated, its existence may suggest that surfactant adsorption only kicks in beyond this critical concentration. Furthermore, it can be observed that the gap in the conductivities (beyond the critical complex/ acid concentrations) is more pronounced in MEK compared to IPA. This suggests that surfactant adsorption strongly differs in both media, either changing the amount of complex/acid dissociated upon complex/acid addition and/or proton exchange reaction on the powder surface. The former would be explained through a dominating adsorption of one ion charge polarity, shifting the dissociation equilibrium towards increased free ions. The latter would follow earlier studies suggesting that a con-ductivity increase in the supernatant was a result H+ release following a surface reaction of oxide particles in contact with alcohol [43,44]. Alternatively, proton exchange of acidic particles and basic dispersant resulting in protonated (positively charged) dispersant molecules and deprotonated (negatively charged) acid sites on powder surface could also occur [45].

Furthermore, it can be observed from Fig. 2 that in the case of Cr3+- complexes, unlike in other cases, the conductivity in MEK and IPA are seemingly identical in the primary electrolyte solution. After powder exposure, however, higher conductivities can be observed in MEK. Both media having a similar dielectric constant, the conductivity disparity underscores the importance of the differing surface chemistries in MEK and IPA suspensions, as we will further corroborate with zeta-potential measurements (section 3.2.). In the case of Cr3+-complex additions (Fig. 2b), the data may suggest a flattening of the slope of the conduc-tivity increase following a first steep increase with increasing complex concentration (especially for IPA). Although additional work would be required to confirm, a change in slope may occur at the dispersant concentration required for full powder surface coverage, beyond which

Fig. 2. Conductivity vs. complex concentration of primary solutions before adding powder (solid lines) and of supernatants (dash lines): a) Cr3+-complex; b) Zn2+- complex; c) 2-ethylhexanoic acid and d) Molar conductivity vs. the square root of the complex concentration.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18446

the conductivity increase returns to a dissociation constant-dominated regime.

Following the bulk conductivity data, it is also possible to determine the critical micelle concentration (CMC) [30,46–48]. Interestingly, turbidimetry data of the supernatants in MEK following Cr3+-complex addition (Fig. 3) reveal two distinct maximum peaks that coincide with slope changes in the conductivity increase (Fig. 2a). Since an increase in turbidity goes along with an increased micelle concentration, this result suggests configuration changes in ad-layer on the particle surface affecting the surfactant concentration. Alternatively, the sudden de-creases around 0.02 mol/dm3 and 0.08 mol/dm3 might also be the consequence of a restructuration of surfactant aggregate shape, such as a change in the micelle size or fusion of globular micelles to elongated worm-like micelles as reported elsewhere [52,53]. Note that the clear difference observed in turbidimetry between IPA and MEK excludes that turbidity is a result of residual powder content (e.g. same particle size distribution). Finally, it should also be noted that the presence of water can furthermore promote reverse micelle formation in non-aqueous media [49–51].

From these conductivity and turbidimetry measurements, a distinct difference between MEK and IPA can be observed in terms of adsorption behaviors and complex stability.

3.2. Zeta-potential

Following a first assessment using conductivity and turbidimetry measurements that suggested distinct adsorption changes between MEK and IPA, zeta-potential measurements were performed to further eluci-date the differences observed between the surfactants and the nature of dispersing liquid. Following reports by Lyklema [54], a distinction be-tween non-polar and semi-polar/polar media arises if the dielectric constant lies below or above 10, respectively. The dielectric constant of both IPA and MEK being above 10, they thus can be considered polar. Consequently, the concept of a zeta-potential applies even to such non-aqueous particle suspensions. A charged slipping plane is formed, the details of which are described elsewhere [55,56].

Zeta-potential measurements (Fig. 4) revealed a significant differ-ence in the charging behavior of spinel in MEK and IPA. While zeta- potentials are negative in MEK, they are positive in IPA over the full range of evaluated acid/complex concentrations. This difference can be explained considering the protic/aprotic character of IPA and MEK, respectively, and their resulting acid-base interactions with the surface [45,50,57,58]. An in-depth discussion of the charging of colloidal par-ticles in non-polar solvents in the presence of surfactants can be found elsewhere [30]. Following Kitabara et al. [59], the relation between the

zeta-potential and the surfactant concentration can be classified in three types, all of which have been observed in this work. For type A, observed for Cr3+/MEK (Fig. 4a), the zeta-potential rapidly increases before reaching a distinct maximum value followed by a steep drop. For type B, observed for Cr3+/IPA and Zn2+/IPA (Fig. 4b and d), the zeta-potential slowly decreases upon a surfactant concentration increase. For type C, all remaining cases in Fig. 4, the zeta-potential remains mostly unchanged following surfactant concentration changes.

The charge reversal effect observed for type A behavior may follow various mechanisms. Labib and Williams reported a moisture-induced charge reversal effect as a result of a donor/acceptor character change of the solid oxide surface [60]. Charge reversal can alternatively also be induced through the formation of tail-tail associated hemi-micelles at the powder surface, which has been equally reported for aqueous [61] and organic-based [44] suspensions. This latter mechanism is suggested to apply for the observed charge reversal observed in the case of spinel in Cr3+/MEK (Fig. 4a). Starting from − 48 mV, the zeta-potential first in-creases sharply with complex addition before reaching a maximum (− 23.3 mV) around 0.0012 mol/dm3. Beyond this, further complex addition leads a rapid zeta-potential decrease until it reaches a floor around − 64 mV.

For Zn2+-complex and acid addition in MEK (Fig. 4c and e), no sig-nificant zeta-potential change can be observed with varying concen-trations. The observed fluctuation is not fully understood but might be linked to slight configurational changes on the surface and/or configu-rational changes of the complexes and micelles in MEK, all of which may lead to fluctuation of the zeta-potential plane position. Similar behavior is observed in the case of acid addition in IPA (Fig. 4f). For Cr3+-complex (Fig. 4b) and Zn2+-complex (Fig. 4d) addition in IPA, however, a net charge neutralization effect is observed, reducing the zeta-potential.

In the following, we will suggest a mechanism able to explain the above behavior through the adsorption/desorption of positively and negatively charged species. In the Cr3+/MEK case (Fig. 5 top), Cr3+

cations released from complex dissociation are initially strongly attracted to the negatively charged surface following the Schulze-Hardy rule [62] (Fig. 5b). Following the adsorption, a rapid increase in zeta-potential is observed. At the maximum, the 2-ethyl-hexanoate groups that adsorb alongside the Cr3+ cations, rather than a desorp-tion mechanism of Cr3+, is expected to induce the charge reversal trend. Based on literature, a tail-tail association of the surfactant to form hemi-micelles is suggested (Fig. 5c) [44,61]. According to the definition of hemi-micelles provided by Tadros [63], they are typically formed at the solid/liquid interface at concentrations below the critical micelle concentration (CMC) and may be observed as a secondary increase in surfactant adsorption beyond a first saturation point.

In the case of zinc complex (Zn2+/MEK) and acid (H+/MEK) addi-tions in MEK, the absence of clearly identifiable peaks to not allow for any equivalent interpretation of the adsorption mechanism. The results suggest, though, that there is no hemi-micelle formation occurring in those cases, a hypothesis supported by that fact that both show a distinctly different rheological behavior (section 3.6) compared to the Cr3+/MEK case. Indeed, anchoring of the charged species to the surface may not be as stable in the absence of Cr3+ and some adsorption- desorption mechanisms may occur. It may be assumed that following the hemi-micelle formation (i.e. hard shell) in the case of Cr3+-complex addition in MEK, particle stabilization is dominated by electrostatic repulsion. In the other cases (Zn2+/MEK and H+/MEK), however, steric repulsion may be the predominant particle stabilization mechanism. Here the valence of the cation appears to play a significant role, since the same Cr3+/MEK behavior could be reproduced with Y3+/MEK (Yttrium (III) 2-ethylhexanoate, 10% w/v in n-hexane) in a follow-up experiment. Despite these suggestions, further investigation would be required to explicitly confirm the suggested stabilization mechanism, exceeding the scope of the present study.

In IPA (Fig. 5 bottom), positively charged particles predominantly attract 2-ethylhexanoate anions, leading to a net charge reduction of the

Fig. 3. Turbidimetry of Cr3+-complex supernatants in MEK and IPA.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18447

surface. Co-adsorption of some metal cations may not be excluded, though, and explain the larger zeta-potential decrease for Zn2+/IPA (Fig. 4d) compared to Cr3+/IPA (Fig. 4b). In this configuration, the COO− heads from the 2-ethyl-hexanoate adsorb to the surface, leaving the chain extended into the solution to provide a steric hindrance against agglomeration. This is supported through the calculation of the Flory-Huggins interaction parameter which gives 0.39 for MEK and 0.35 for IPA with respect to their interaction with the 2-ethylhexanoate groups. Indeed, a value below 0.5 is required to provide effective ste-ric hindrance. This once more clearly indicates that a pure steric stabi-lization effect cannot explain the differences observed between the various powder suspensions, which will be further corroborated analyzing the viscosity and agglomeration degree between both solvents.

Following these results, we will concentrate on the effects of Cr3+- complex addition in MEK and IPA respectively, except for the rheology measurements in section 3.6.

3.3. Particle size distribution in pure solvents

Although zeta-potential measurements showed an opposite charge for spinel powder in MEK and IPA, their absolute values were close to each other, providing a similar repulsive potential component. The comparison of particle size distributions in pure MEK and IPA (Fig. 6) reveals that both solvents are insufficient to stabilize the suspension, which remains strongly agglomerated in both cases. Without surfac-tants, repulsion is governed mainly by electrostatic repulsion forces, with attractive van der Waals forces prevailing. Also note that with MEK and IPA having similar dielectric constants, attractive van der Waals forces are believed comparable in both systems. Explaining the strong gel formation in MEK (Fig. 4a) thus requires additional non-DLVO forces.

As reported by Hellebrand [61], powder dispersion requires wetting by the dispersing liquid, with less dispersion work required for lower contact angles. There is no straightforward rule and hence there are different approaches to predict the wetting behavior of a dispersing liquid [64]. One approach considers comparison of hydrogen bond

Fig. 4. Zeta-potential in MEK for a) Cr3+-complex c) Zn2+-complex and e) acid. Zeta-potential in isopropanol for b) Cr3+-complex d) Zn2+-complex and f) acid.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18448

indexes. Following this approach, the higher hydrogen bonding index found in isopropanol might contribute to the formation of a solvation layer, and subsequently provide a more efficient powder wetting and dispersing capacity. Weak wetting could furthermore explain the for-mation of a percolating powder network in MEK, leading to a gel formation.

3.4. FTIR

In order to corroborate the assumptions made from zeta-potential measurements, FTIR spectroscopy was performed to qualitatively identify surface adsorbed groups and the presence of 2-ethylhexanoate.

The FTIR spectra of dried powder samples are provided in Fig. 7, after being suspended in the different solvents/surfactants. Despite the qualitative nature of the analysis, potential surface modifications may be followed through the identification of characteristic functional groups.

Broad and weak characteristic peaks for O–H vibrations can be located between 3600 and 3200 cm− 1. FTIR, however, does not allow distinction between their origin from water molecules and/or hydroxide groups. There are, however, no significant changes observed with increasing Cr3+-complex addition in MEK or between MEK and IPA. Furthermore, there is no reason to believe that the water/moisture content should change significantly among the investigated samples that have been prepared and measured at the same time at equal conditions. In the following, O–H groups have therefore been mostly omitted in the interpretation of the FTIR results.

Focusing on the characteristic C–H groups, it can be observed that for spinel powder modified in 0.0012 mol/dm3 chromium (III) 2-ethylhex-anoate, there are no significant changes in the spectra compared to MEK only exposure, explaining the remaining gel-like structure of these slurries. For the complex concentrations ≥0.0036 mol/dm3, however, characteristic C–H absorption peaks appear around 2840-3000 cm− 1. This threshold aligns with the critical complex concentration observed for a fast viscosity drop within the Cr3+/MEK powder slurries (Fig. 7 inlay). From 0.036 mol/dm3 onwards, a characteristic peak for C––O bonds at 1680 cm− 1 develops in parallel with the increasing C–H peaks. These results clearly indicate that adsorption of 2-ethylhexanoate onto the powder surface occurs. Although these results do not allow conclu-sions to be made regarding the tail-tail association of 2-ethyl-hexanoate in MEK – furthermore noticing solvent removal may significantly disrupt the ad-layer structure formed in solution – they also do not refute the hypothesis brought up previously from zeta-potential measurements.

3.5. Sedimentation height and settling velocity

The sediment tests are simple tests that allow assessment of the long- term stability of a suspension. Small sediment heights are hereby

Fig. 5. Schematic representation of proposed stabilizing mechanism (top) in MEK: a) negatively charged slipping plane, b) adsorption of Cr3+ ions and partial adsorption of 2-ethylhexanoate molecules, c) hemi-micelle formation on spinel surface and (bottom) in IPA, d) positively charged slipping plane and e) adsorption of 2-ethylhexanoate molecules.

Fig. 6. Particle size distribution measured by Static Light Scattering method for MgAl2O4 powder in pure solvents (isopropanol and MEK) after ultra-sonic treatment.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18449

indicative of a well-dispersed powder suspension, preventing particle agglomeration and allowing particle rearrangement for dense sediment formation. Increasing sediment height might be representative of long- range repulsive forces, weakly flocculated suspensions and finally par-ticle agglomeration. Although flocculation and agglomeration may impact the sedimentation speed, it is also a function of the viscosity of the medium, temperature (e.g. Brownian motion) and particle size. Because of the gelated suspension characteristics, especially in IPA, two distinct sedimentation tests were performed in the present study to

confirm the general trend, as will be discussed in the following. Considerable changes between the subsequent experiments are high-lighted by circles.

From the sedimentation results in Fig. 8, it can be observed that in pure MEK and IPA, nearly no sedimentation occurs because of the highly gelated character of the suspension. Following acid/complex addition, the sediment heights drop in both solvents. There are, however, distinct differences between IPA and MEK. Acid and Zn2+-complex addition in MEK lead only to limited changes in sedimentation heights with a

Fig. 7. FTIR data comparison in the functional group region (4000-1500 cm− 1) with viscosity as a function of Cr3+-complex concentration in MEK in the figure inlay.

Fig. 8. Natural sedimentation after 7 days in a) MEK; b) control test in MEK; c) IPA (encircled point is an artefact, which is supported by d) and zeta potential values (inlay)); d) control test in IPA.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18450

plateau reached at a sediment height around 90% of the initial sus-pension height. The situation, however, significantly changes upon Cr3+-complex addition. The sediment height rapidly drops and reaches a plateau at about 50% of the original suspension height. This drop in sediment height suggests a packing of particles close to the theoretical 56% expected from a random close packing of particles. This is only achievable in well-dispersed and non-cohesive particle slurries. This observation is consistent with the viscosity measurements that will be discussed in the following section (section 3.6.). Note also the substan-tial changes in the sediment height between both tests (Fig. 8a and b) at Cr3+-complex concentrations around 0.0036 and 0.018 mol/dm3 where charge inversion occurs, which may lead to reproducibility issues.

In IPA (Fig. 8c–d), the added species (with different cation valences) do not seem to play a significant role. The sediment height drops quickly, albeit only slightly upon acid/complex addition, reaching a plateau around 90% of the initial suspension height. Whereas for Cr3+/ IPA and Zn2+/IPA the changes in sediment heights occur relatively fast, more acid addition is required to reach an equivalent change, which is explained through the higher 2-ethyl-hexanoate mole content in the metal complexes. These results suggest that the 2-ethyl-hexanoate con-centration required for minimal steric repulsion under these conditions is relatively low. Indeed, from the current results it may be suggested that this minimum concentration leads to a weak minimum in sediment height (i.e. maximum following the notation in Fig. 8c and d), beyond which increased 2-ethyl-hexanoate surface density (increased exclusion volume) and/or surface charge neutralization leads again to a slight increase in sediment height. Overall, the sediment height shows low variability upon complex/acid addition beyond the initial sediment height drop. These results support the previous hypothesis that steric repulsion from adsorbed ethylhexanoate is the main stabilization mechanism in IPA, with the cation valence playing no significant role. Finally, these results also suggest that IPA powder suspensions are never in a well-dispersed state for any of the studied dispersants.

Fig. 9 illustrates the sedimentation tests over a 7 days period to

illustrate the sedimentation velocity. It can be observed that unless in the case of MEK, sedimentation speed is equivalent in IPA regardless of the chromium complex concentration. In MEK, however, the sedimen-tation speed increases with increasing Cr3+-complex concentration. The results suggest that for Cr3+/MEK there are no or much weaker long- range inter-particle forces than for Cr3+/IPA, limiting the flocculation phenomenon as will be discussed in the next section. Furthermore, the results may suggest that long-range interparticle forces are reduced with increasing chromium complex concentration in MEK.

3.6. Rheology measurement

Viscosity measurements revealed (Fig. 10a) a low-viscosity near- Newtonian behavior beyond a Cr3+-complex concentration of 0.018 mol/dm3. Net particle interaction forces thus appear to rapidly decrease upon Cr3+-complex addition. In the case of acid and Zn2+-complex ad-ditions in MEK (Fig. 10e and c), a shear thinning behavior is observed, representative of flocculated particle slurries and regardless of their respective concentrations. This fits well with the observations made from the sedimentation tests. Rheology was not evaluated for 0.0036 mol/dm3 Zn2+/MEK and 0.0072 mol/dm3 acid in MEK because of their fully gelated nature, like spinel in pure MEK.

Unlike in MEK, a shear-thinning behavior appeared for all surfactant types in IPA, once more supporting the steric repulsion hypothesis following 2-ethy-hexanoate adsorption (Fig. 10b, d, 10f). However, the shear thinning behavior is most pronounced in the case of acid addition, leading to the suggestion that the metal cations interfere in the ad-layer or its vicinity, as suggested from the zeta-potential measurements.

Comparing viscosities at constant shear rate (Fig. 11), it can be observed that for Cr3+ in IPA, the viscosity decreases systematically with an increase of the complex concentration. This is in contrast to the sit-uation observed for Zn2+-complex and acid additions, where the vis-cosity increases after reaching a minimum following a rapid initial drop. The minimum may be associated with an optimal dispersant

Fig. 9. The settlement velocity of a) Cr3+/MEK and b) Cr3+/IPA after 1 day, 4 days and 7 days.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18451

concentration for the respective system, supporting the results from the sedimentation trials (Fig. 8c and d). Above this optimum dispersant concentration, viscosity increases are commonly observed, whether linked to viscosity changes of the medium or stiffening effect of the adsorbed layer, increasing the effective solid loading. These optimum dispersant concentrations have been found at 0.0036 mol/dm3 and 0.0071 mol/dm3 for the Zn2+-complex and acid, respectively. For Cr3+/MEK, the viscosity drops significantly after reaching a complex concentration of 0.0036 mol/dm3 and remains equally low upon further increase of the complex concentration.

The differences in the effects of selected complexes/acids on the viscosity observed here highlight the influence of the dispersing media on the dispersion. In particular, the protic (e.g. IPA) or the aprotic (e.g. MEK) nature of the solvent can have a dominant role on the interaction between dispersants and the powder surface. In IPA, the viscosity is lowest following acid addition and highest upon Cr3+-complex addition. The reason why in the case of Cr3+-complex addition in IPA the viscosity keeps dropping is not fully understood, but is presumably linked to a co-

adsorption of Cr3+ in the adsorption layer or its location in the vicinity thereof. Corroboration thereof, however, would require additional ex-periments out of the scope of the present paper. The situation, however, is the opposite in MEK, where chromium complex addition provides the lowest viscosities by far. Addition of chromium (III) 2-ethylhexanoate in MEK reduces the interparticle forces, presumably changing the config-uration of 2-ethylhexanoate adsorption at the powder surface. The lowering of the interparticle forces and/or their distance range inhibits gel formation and long-range particle repulsion, consistent with the observations from the sedimentation experiments.

3.7. Ball milling

In order to confirm the good dispersion capability of chromium complex additions in MEK, Fig. 12 shows the particle size distribution of Cr3+/MEK and Cr3+/IPA spinel suspensions as a function of the ball milling time. Both suspensions present considerable amounts of ag-glomerates prior to milling. Sixty minutes of milling is sufficient to fully

Fig. 10. Viscosity curves for suspensions with a) Cr3+-complex in MEK; b) Cr3+-complex in IPA; c) Zn2+-complex in MEK; d) Zn2+-complex in IPA; e) 2-ethylhexanoic acid in MEK; f) 2-ethylhexanoic acid in IPA. The inflection points in a), c) and e) are measuring artefacts due to the low viscosity and some powder sedimentation.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18452

remove any agglomerates in Cr3+/MEK, leading to a monomodal par-ticle size distribution with a dv50 = 0.139 μm. In the case of Cr3+/IPA, agglomeration subsists after 60 min of milling, illustrating the poor particle stabilization character of chromium complexes in IPA. This result confirms the particle stabilization capability of Cr3+ 2-ethyl-hex-anoate complexes in aprotic solvents such as MEK.

4. Conclusions

We have evaluated the use of 2-ethyhexanoate metal complexes as novel dispersing agents for oxide powder ceramics in protic (iso-propanol, IPA) and aprotic (methy ethyl ketone, MEK) dispersing media. Chromium (III) 2-ethylhexanoate has been identified as a promising small-chained surfactant to effectively disperse spinel powders with a dv,50 < 0.2 μm in an aprotic solvent. While other powder systems have been more carefully studied, such as titania in pure alcohols [43] or alumina in various solvents [23], investigation of spinel stabilization mechanisms has mostly focused on aqueous-based suspensions [65,66]. Consequently, the mechanisms behind the spinel suspension gelation in aprotic organic solvents is less well documented. While the selected system was chosen to avoid the common use of commercial dispersants, it has simultaneously laid out a pathway towards the selective use of dispersants as powder dopant sources.

Magnesium aluminate spinel forms a strongly gelated suspension in

aprotic dispersing media (e.g. MEK) with negatively charged particle surfaces. In protic dispersing media (e.g. IPA), however, gelation is weaker and particles are charged positively. These opposite particle charges are attributed to the differing proton donor-acceptor character of MEK and IPA. The formation of the gel structure is presumably linked to the different energy required for particle wetting in protic and aprotic conditions. A similar behavior is expected during the use of acrylates which may express a more protic (e.g. hydroxyl functional acrylates like HEA or HEMA) or aprotic (e.g. HDDA) character.

Combining various measurement techniques, the influence of IPA and MEK on the adsorption and particle stabilization mechanisms of (Zn2+, Cr3+) 2-ethyl-hexanoate complexes and 2-ethylhexanoic acid could be established. The surface polarity and the cation valence could hereby be identified as prime factors governing the particle stabilization performance of such dispersants. In particular, Cr3+ 2-ethyl-hexanoate has been identified as a promising dispersant for particle stabilization in aprotic dispersing media, such as MEK. Hemi-micelle formation was suggested as the governing stabilizing mechanism behind this promising performance. The “hard” shell configuration following the hemi-micelle formation apparently leads to electrostatic-governed particle repulsion. Further investigation would be required to provide a final confirmation for hemi-micelle formation on the powder surface.

Combined with the strong electrostatic particle repulsion, the higher conductivity values obtained for Cr3+ 2-ethyl-hexanoate in MEK effi-ciently suppress long-range particle interaction, shifting the rheological behavior from a shear thinning to a Newtonian liquid behavior. This could be confirmed through sedimentation tests and ball milling trials, both of which confirmed the good particle stabilization and packing density performance.

Because of the limited use of Cr3+ as a dopant in ceramics, a similar behavior could be confirmed with Y3+ and is expected with complexes of other trivalent metal cations. A metal-complex based dispersion system like the one investigated in the present work may thus be promising for use in other aprotic media such as aprotic acrylates commonly employed in stereolithographic-based additive manufacturing. Application of the knowledge brought forward in the present paper to non-aqueous UV- curable systems may thus prove relevant in the rapidly expanding field of ceramic additive manufacturing. Therein, a more knowledge-based approach may pave the way towards efficient dispersion and stabiliza-tion of high solid load slurries required for reliable high-performance ceramics.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence

Fig. 11. Viscosity at a shear rate of 100 s− 1 for dispersants in MEK (red line) and isopropanol (black line). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 12. Particle size distribution before and after planetary ball milling of 10 vol% suspension with chromium (III) 2-ethylhexanoate in a) isopropanol – ag-glomerates present; b) methyl ethyl ketone - fully de-agglomerated after 60 min of milling.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18453

the work reported in this paper.

Acknowledgements

The authors would like to thank Prof. Paul Bowen (EPFL Lausanne, Switzerland) for the fruitful discussion.

References

[1] H. Wu, W. Liu, R. He, Z. Wu, Q. Jiang, X. Song, Y. Chen, L. Cheng, S. Wu, Fabrication of dense zirconia-toughened alumina ceramics through a stereolithography-based additive manufacturing, Ceram. Int. 43 (2017) 968–972.

[2] A. Ghazanfari, W. Li, M.C. Leu, J.L. Watts, G.E. Hilmas, Additive manufacturing and mechanical characterization of high density fully stabilized zirconia, Ceram. Int. 43 (2017) 6082–6088.

[3] A. Vyatskikh, R.C. Ng, B. Edwards, R.M. Briggs, J.R. Greer, Additive manufacturing of high-refractive-index, nanoarchitected titanium dioxide for 3D dielectric photonic crystals, Nano Lett. 20 (2020) 3513–3520.

[4] C. Hinczewski, S. Corbel, T. Chartier, Ceramic suspensions suitable for stereolithography, J. Eur. Ceram. Soc. 18 (1998) 583–590.

[5] D.A. Komissarenko, P.S. Sokolov, A.D. Evstigneeva, I.A. Shmeleva, A.E. Dosovitsky, Rheological and curing behavior of acrylate-based suspensions for the DLP 3D printing of complex zirconia parts, Materials (2018) 11.

[6] D. Carloni, G. Zhang, Y. Wu, Transparent alumina ceramics fabricated by 3D printing and vacuum sintering, J. Eur. Ceram. Soc. 41 (2021) 781–791.

[7] H. Xing, B. Zou, Q. Lai, C. Huang, Q. Chen, X. Fu, Z. Shi, Preparation and characterization of UV curable Al2O3 suspensions applying for stereolithography 3D printing ceramic microcomponent, Powder Technol. 338 (2018) 153–161.

[8] K. Li, Z. Zhao, The effect of the surfactants on the formulation of UV-curable SLA alumina suspension, Ceram. Int. 43 (2017) 4761–4767.

[9] J. Sun, J. Binner, J. Bai, Effect of surface treatment on the dispersion of nano zirconia particles in non-aqueous suspensions for stereolithography, J. Eur. Ceram. Soc. 39 (2019) 1660–1667.

[10] Z. Chen, J. Li, C. Liu, Y. Liu, J. Zhu, C. Lao, Preparation of high solid loading and low viscosity ceramic slurries for photopolymerization-based 3D printing, Ceram. Int. 45 (2019) 11549–11557.

[11] A. Goswami, K. Ankit, N. Balashanmugam, A.M. Umarji, G. Madras, Optimization of rheological properties of photopolymerizable alumina suspensions for ceramic microstereolithography, Ceram. Int. 40 (2014) 3655–3665.

[12] M. Borlaf, A. Serra-Capdevila, C. Colominas, T. Graule, Development of UV-curable ZrO2 slurries for additive manufacturing (LCM-DLP) technology, J. Eur. Ceram. Soc. 39 (2019) 3797–3803.

[13] X. Li, H. Zhong, J. Zhang, Y. Duan, H. Bai, J. Li, D. Jiang, Dispersion and properties of zirconia suspensions for stereolithography, Int. J. Appl. Ceram. Technol. 17 (2019) 239–247.

[14] Y.-H. Lee, J.-B. Lee, W.-Y. Maeng, Y.-H. Koh, H.-E. Kim, Photocurable ceramic slurry using solid camphor as novel diluent for conventional digital light processing (DLP) process, J. Eur. Ceram. Soc. 39 (2019) 4358–4365.

[15] S. Zhang, N. Sha, Z. Zhao, Surface modification of alpha-Al2O3 with dicarboxylic acids for the preparation of UV-curable ceramic suspensions, J. Eur. Ceram. Soc. 37 (2017) 1607–1616.

[16] G. Ding, R. He, K. Zhang, M. Xia, C. Feng, D. Fang, Dispersion and stability of SiC ceramic slurry for stereolithography, Ceram. Int. 46 (2020) 4720–4729.

[17] Y. De Hazan, J. Heinecke, A. Weber, T. Graule, High solids loading ceramic colloidal dispersions in UV curable media via comb-polyelectrolyte surfactants, J. Colloid Interface Sci. 337 (2009) 66–74.

[18] M. Borlaf, N. Szubra, A. Serra-Capdevila, W.W. Kubiak, T. Graule, Fabrication of ZrO2 and ATZ materials via UV-LCM-DLP additive manufacturing technology, J. Eur. Ceram. Soc. 40 (2020) 1574–1581.

[19] K. Zhang, C. Xie, G. Wang, R. He, G. Ding, M. Wang, D. Dai, D. Fang, High solid loading, low viscosity photosensitive Al2O3 slurry for stereolithography based additive manufacturing, Ceram. Int. 45 (2019) 203–208.

[20] K. Zhang, R. He, C. Xie, G. Wang, G. Ding, M. Wang, W. Song, D. Fang, Photosensitive ZrO2 suspensions for stereolithography, Ceram. Int. 45 (2019) 12189–12195.

[21] J. Sun, J. Binner, J. Bai, 3D printing of zirconia via digital light processing: optimization of slurry and debinding process, J. Eur. Ceram. Soc. 40 (2020) 5837–5844.

[22] B.I. Lee, J.P. Rives, Dispersion of alumina powders in nonaqueous media, Colloid. Surface. 56 (1991) 25–43.

[23] J.C. Liu, J.H. Jean, C.C. Li, Dispersion of nano-sized 7-alumnina powder in non- polar solvents, J. Am. Ceram. Soc. 89 (2006) 882–887.

[24] S. Zürcher, T. Graule, Influence of dispersant structure on the rheological properties of highly-concentrated zirconia dispersions, J. Eur. Ceram. Soc. 25 (2005) 863–873.

[25] L. Jay Deiner, K.A. Piotrowski, T.L. Reitz, J. Ferreira, Mechanisms of fatty acid and triglyceride dispersant bonding in non-aqueous dispersions of NiO, J. Am. Ceram. Soc. 96 (2013) 750–758.

[26] S. Farrokhpay, A review of polymeric dispersant stabilisation of titania pigment, Adv. Colloid Interface Sci. 151 (2009) 24–32.

[27] T. Kakui, T. Miyauchi, H. Kamiya, Analysis of the action mechanism of polymer dispersant on dense ethanol alumina suspension using colloidal probe AFM, J. Eur. Ceram. Soc. 25 (2005) 655–661.

[28] I. Sushumna, R.K. Gupta, E. Ruckenstein, Effective dispersants for concentrated, nonaqueous suspensions, J. Mater. Res. 7 (1992) 2884–2893.

[29] M.F. Hsu, E.R. Dufresne, D.A. Weitz, Charge stabilization in nonpolar solvents, Langmuir 21 (2005) 4881–4887.

[30] G.N. Smith, J. Eastoe, Controlling colloid charge in nonpolar liquids with surfactants, Phys. Chem. Chem. Phys. 15 (2013) 424–439.

[31] Wenbing Hu, in: Polymer Physics, A Molecular Approach, 1, Springer, Vienna, 2013. https://doi.org/10.1007/978-3-7091-0670-9e.

[32] A.R. Studart, E. Amstad, M. Antoni, L.J. Gauckler, Rheology of concentrated suspensions containing weakly attractive alumina nanoparticles, J. Am. Ceram. Soc. 89 (2006) 2418–2425.

[33] E.J.W Verwey, J.Th.G. Overbeek, Theory of the Stability of Lyophobic Colloids, Elsevier Publishing Company Inc., Amsterdam, 1948.

[34] B. Derjaguin, L. Landau, Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solutions of electrolytes, Acta Physicochimica U.R.S.S. 14 (1941) 633–662.

[35] Q. Ran, P. Somasundaran, C. Miao, J. Liu, S. Wu, J. Shen, Effect of the length of the side chains of comb-like copolymer dispersants on dispersion and rheological properties of concentrated cement suspensions, J. Colloid Interface Sci. 336 (2009) 624–633.

[36] M. Stuer, P. Bowen, Yield stress modelling of doped alumina suspensions for applications in freeze granulation: towards dry pressed transparent ceramics, Adv. Appl. Ceram. 111 (2012) 254–261.

[37] R.J. Flatt, P. Bowen, Yield stress of multimodal powder suspensions: an extension of the YODEL (Yield stress mODEL), J. Am. Ceram. Soc. 90 (2007) 1038–1044.

[38] U. Aschauer, O. Burgos-Montes, R. Moreno, P. Bowen, Hamaker 2: a toolkit for the calculation of particle interactions and suspension stability and its application to mullite synthesis by colloidal methods, J. Dispersion Sci. Technol. 32 (2011) 470–479.

[39] L. Bergstrm, K. Shinozaki, H. Tomiyama, N. Mizutani, Colloidal processing of a very fine BaTiO3 powder-effect of particle interactions on the suspension properties, consolidation, and sintering behavior, J. Am. Ceram. Soc. 80 (2005) 291–300.

[40] F. Rubio-Hernandez, Is DLVO theory valid for non-aqueous suspensions? J. Non- Equilibrium Thermodyn. 24 (1999) 75–79.

[41] M.J.D. Bower, T.L. Bank, R.F. Giese, C.J. van Oss, Nanoscale forces of interaction between glass in aqueous and non-aqueous media: a theoretical and empirical study, Colloid. Surface. Physicochem. Eng. Aspect. 362 (2010) 90–96.

[42] J. Israelachvili, Solvation, Structural, and Hydration Forces, 2011, pp. 341–380. [43] M. Farrokhi-rad, M. Ghorbani, Stability of titania nano-particles in different

alcohols, Ceram. Int. 38 (2012) 3893–3900. [44] T. Chartier, E. Jorge, P. Boch, Dispersion properties of BaTiO3 tape-casting slurries,

J. Eur. Ceram. Soc. 11 (1993) 387–393. [45] F.M. Fowkes, H. Jinnai, M.A. Mostafa, F.W. Anderson, R.J. Moore, Mechanism of

electric charging of particles in nonaqueous liquids, Coll. Surf. Reprogr. Technol. (1982) 307–324.

[46] Z. Khan, Micellization and thermodynamic properties of cationic surfactant cetyltrimethylammonium bromide in non-aqueous mixture of lauric acid, Int. J. Electrochem. Sci. (2017) 4528–4542.

[47] N. Kumaraguru, K. Santhakumar, S. Kalyanasundharam, Studies on thermodynamics of micellization and solvophobic interactions of novel surfactant–Cr(III) complexes in non-aqueous solvents, J. Solut. Chem. 40 (2011) 1673–1686.

[48] T. Warnheim, Aggregation of surfactants in nonaqueous, polar solvents, Curr. Opin. Colloid Interface Sci. 2 (1997) 472–477.

[49] W.H. Briscoe, R.G. Horn, Direct measurement of surface forces due to charging of solids immersed in a nonpolar liquid, Langmuir 18 (2002) 3945–3956.

[50] I.D. Morrison, Electrical charges in nonaqueous media, Colloid. Surface. Physicochem. Eng. Aspect. 71 (1993) 1–37.

[51] H.-F. Eicke, H. Christen, Is water critical to the formation of micelles in apolar media, Helv. Chim. Acta 61 (1978) 2258–2263.

[52] A.S. Rafique, S. Khodaparast, A.S. Poulos, W.N. Sharratt, E.S.J. Robles, J.T. Cabral, Micellar structure and transformations in sodium alkylbenzenesulfonate (NaLAS) aqueous solutions: effects of concentration, temperature, and salt, Soft Matter 16 (2020) 7835–7844.

[53] G.V. Jensen, R. Lund, J. Gummel, T. Narayanan, J.S. Pedersen, Monitoring the transition from spherical to polymer-like surfactant micelles using small-angle X- ray scattering, Angew Chem. Int. Ed. Engl. 53 (2014) 11524–11528.

[54] J. Lyklema, Principles of interactions in non-aqueous electrolyte solutions, Curr. Opin. Colloid Interface Sci. 18 (2013) 116–128.

[55] R. Parsons, The electrical double layer in non-aqueous solvents, Electrochim. Acta 21 (1976) 681–686.

[56] E.J.W. Verwey, The electrical double layer of oxidic substances especially in non- aqueous media, Recl. Trav. Chim. Pays-Bas 60 (2010) 625–633.

[57] T. Hao, Exploring the charging mechanisms in non-aqueous multiphase surfactant solutions, emulsions and colloidal systems via conductivity behaviors predicted with eyring’s rate process theory, Phys. Chem. Chem. Phys. 18 (2016) 476–491.

[58] G.H. Wang, P. Sarkar, P.S. Nicholson, Influence of acidity on the electrostatic stability of alumina suspensions in ethanol, J. Am. Ceram. Soc. 80 (1997) 965–972.

[59] A. Kitabara, M. Amano, S. Kawasaki, K. Kon-no, The concentration effect of surfactants on zeta-potential in non-aqueous dispersions, Colloid Polym. Sci. 255 (1977) 1118–1121.

[60] M.E. Labib, R. Williams, The effect of moisture on the charge at the interface between solids and organic liquids, J. Colloid Interface Sci. 115 (1987) 330–338.

[61] H. Hellebrand, Tape Casting, Wiley-VCH Verlag GmbH1995.

P. Zubrzycka et al.

Ceramics International 47 (2021) 18443–18454

18454

[62] G. Trefalt, I. Szilagyi, M. Borkovec, Schulze-Hardy rule revisited, Colloid Polym. Sci. 298 (2020) 961–967.

[63] T. Tadros, Hemi-micelle, in: T. Tadros (Ed.), Encyclopedia of Colloid and Interface Science, Springer Berlin Heidelberg, Berlin, Heidelberg, 2013, p. 631, 631.

[64] T.A. Ring, Wetting, Deagglomeration, and Adsorption, Fundamentals of Ceramic Powder Processing and Synthesis 1996, pp. 359-420.

[65] A. Talimian, D. Galusek, Aqueous slip casting of translucent magnesium aluminate spinel: effects of dispersant concentration and solid loading, Ceram. Int. 45 (2019) 10646–10653.

[66] P. Ramavath, P. Biswas, K. Rajeswari, M.B. Suresh, R. Johnson, G. Padmanabham, C.S. Kumbhar, T.K. Chongdar, N.M. Gokhale, Optical and mechanical properties of compaction and slip cast processed transparent polycrystalline spinel ceramics, Ceram. Int. 40 (2014) 5575–5581.

P. Zubrzycka et al.