Binder Jetting for Functional Testing of Ceramic Sanitaryware

Abstract

Additive manufacturing (AM) of ceramics presents a promising approach for the production of complex sanitaryware prototypes, offering advantages in terms of cost and time to market. This study explores binder jetting (BJ) as an optimal AM technique due to its ability to process ceramic materials without thermal stress, accommodate various compositions, and produce large components without support structures. A combination of refractory cement, feldspathic sands, quartz, and calcined alumina was used to formulate 19 different compositions, ensuring adequate green strength and minimizing shrinkage during sintering. A hydration-activated binding method with a water-based binder was employed to enhance part formation and mechanical properties. The results indicate that compositions containing calcined alumina exhibited lower pyroplastic deformation, while optimized gelling agent concentrations improved green strength and dimensional accuracy. The final selected material (SA18) demonstrated high compressive strength, low shrinkage, and a surface roughness comparable to traditional sanitaryware. The application of an engobe layer improved glaze adherence, ensuring a homogeneous surface. This study highlights binder jetting as a viable alternative to traditional ceramic processing, paving the way for its adoption in industrial sanitaryware manufacturing.

1. Introduction

Ceramic materials are broadly used in industrial applications, ranging from the production of sanitaryware to advanced mechanics. Despite their widespread use, their production process still involves inefficient product development procedures, often based on a trial-and-error approach [1].

The product development process of ceramic sanitaryware encompasses a specific sequence of steps that a company follows to bring new offerings to the market to meet a particular demand. For many years, traditional prototyping has been utilized to develop, design, and create accurate models of diverse products. This prototyping method often requires extensive training and substantial time investment. Skilled craftsmen are essential to construct temporary prototypes of complex components based on blueprints. This approach enables the craftsmen to fabricate either a complete or partial model. Nevertheless, the prototype may not always work, leading to the creation of additional prototypes using different materials and crafting techniques. Typically, a near-final product must be produced to test its ultimate properties. This iterative cycle, spanning from conceptualization to production, entails considerable investments of time, materials, energy, and financial resources. This is precisely the stage where additive manufacturing (AM) can provide value.

AM allows for the fabrication of prototypes with greater ease and flexibility, eliminating the need for highly skilled workers. By directly translating intricate digital 3D models into tangible products, AM minimizes the time, labor, and materials traditionally expended in manual construction processes [2]. In addition to the time and cost savings of introducing AM at the pre-production stage, AM, and in particular binder jetting (BJ), have proven to be more sustainable processes than traditional manufacturing methods like casting or injection molding [3,4].

While AM processes can produce larger build parts, in the current framework, limitation on the availability of initial raw materials for obtaining feedstock is still a challenge in the AM of ceramics [5]. A significant constraint is related to the availability of printable materials suitable for glazing and that can be processed within an industrial cycle. The fabrication of large ceramic components with no cracks or deformation is still challenging [6]. In addition, during binder removal and sintering, large ceramic parts show problems that can be attributed to significant variations in sectional thickness and differences in thermal features and shrinkage.

The integration of ceramic AM continues to grow, driven by improvements in processing techniques and material properties, making it a viable solution for complex and high-performance applications [7]. Traditional ceramics have been largely overlooked in AM material development, as research and industrial applications have predominantly focused on advanced ceramics such as zirconia (ZrO₂), alumina (Al₂O₃), silicon carbide (SiC), and hydroxyapatite (HA) for their application in biomedical, aerospace, and energy applications [8]. As a result, the evolution of ceramic AM has been driven by materials that offer enhanced processability, structural integrity, and application-specific performance, resulting in the limited integration of conventional ceramics within this technological paradigm.

This study aims to develop prototypes using traditional ceramics, evaluating their feasibility in meeting the technical requirements and certification standards of conventional sanitaryware. The objective is to assess the suitability and reproducibility of the fabricated components, ensuring that additive manufacturing can effectively integrate traditional ceramic materials while maintaining industry-specific performance criteria.

In order to achieve the main objective of this project, some of the most influential factors in the development of the materials were identified as follows:

• The material must be able to be processed by BJ (more on technology selection later).
• Sufficient green strength. Printed parts do not need to have high mechanical strength at this stage, but they must be able to withstand further processing.
• Adequate pyroplastic deformation after sintering. The material must guarantee a dimensional stability after firing.
• The material must be able to withstand glazing to achieve the characteristics of the real end product.
• Adequate final roughness (after glazing). To be able to test some of the properties of the sanitaryware, the roughness of the final pieces must be equal to that obtained with pieces manufactured by the usual industrial process.
• To ensure that the glaze has the correct properties for further testing, the material needs to be fired using an industrial cycle.

In short, the required green strength was set at 2 MPa, as this was considered sufficient for the part to be handled and post-processed safely. The final mechanical strength was established at 4 MPa, as no mechanical strength tests were to be carried out. The main tests for the design would be related to fluid mechanics, and this would require the highest precision in the prototype and achieving the same surface roughness after glazing. As the part would be fired with an industrial cycle, the prototypes must withstand firing temperatures up to 1230 °C with a pyroplastic deformation lower than 15·10⁵ cm⁻¹.

2. Materials and Methods

2.1. Additive Manufacturing Technology

As illustrated in Table 1, a comparison is provided of the current AM technologies that are capable of processing ceramic materials. Of all the additive manufacturing technologies currently available, binder jetting was selected because it has certain advantages over other AM processes:

• It does not use heat during the build process. Other technologies use a heat source that can create residual stress that must be removed in secondary post-processing.
• It has a high compatibility with materials such as ceramics and refractory metals and a large group of material types that are difficult to process.
• There is no need for support structures, as the loose powder supports protrusions and stacked or suspended objects.
• It allows for the printing of large components and is often more cost-effective than other additive processes.

Table 1. Comparison of various 3D printing technologies for manufacturing ceramics [9].

Process	Layer Formation Technique	Accuracy/Resolution (µm)	Temperature Process/Speed	Support Structure	Build Size	Feedstock Cost/Process Cost
SLA	Photopolymerization	High/10	Very low	Required	XS/S/M	Low/Medium
FDM	Extrusion	Good/50–200	Low/Low	Required	M/L/XL	Low/Low
SLS	Powder fusion	Low/80–250	High/Low	Not required	M/L	Low/High
SLM	Sheet lamination	High/80–250	High/High	Required	M	Low/Low
BJ	Binder bonding	High/<50	Low/High	Not required	M/L/XL	Low/Medium
DED	Material deposition	Very low/-	High/	-	Versatile	Low/High

Table 1. Comparison of various 3D printing technologies for manufacturing ceramics [9].

Process	Layer Formation Technique	Accuracy/Resolution (µm)	Temperature Process/Speed	Support Structure	Build Size	Feedstock Cost/Process Cost
SLA	Photopolymerization	High/10	Very low	Required	XS/S/M	Low/Medium
FDM	Extrusion	Good/50–200	Low/Low	Required	M/L/XL	Low/Low
SLS	Powder fusion	Low/80–250	High/Low	Not required	M/L	Low/High
SLM	Sheet lamination	High/80–250	High/High	Required	M	Low/Low
BJ	Binder bonding	High/<50	Low/High	Not required	M/L/XL	Low/Medium
DED	Material deposition	Very low/-	High/	-	Versatile	Low/High

Binder jetting is an additive manufacturing process in which a liquid bonding agent is selectively deposited to join powder materials [10]. The machine comprises two platforms: the feed platform, where the raw material is placed, and the build platform, where the part is built up layer by layer.

First, a blade spreads a layer of the powder from the feed platform onto the build platform. Next, the print head, which consists of one or more inkjet nozzles, injects the binder liquid into the powder following a pattern defined by the CAD model. The binder helps to join the powder particles in the chosen areas. Once the layer is complete, the build platform lowers to a set height, the feed platform raises, and the blade spreads a new layer of fresh powder onto the build platform. The process is repeated until the part is completely built.

Once the printing process is complete, the parts are left in the machine for 60 min to facilitate drying and ensure proper binding. Subsequently, excess powder is gently brushed away, leaving a bed of powder surrounding the parts.

Figure 1 shows all the steps involved in binder jetting technology.

2.2. Specimen Preparation

2.2.1. Base Material

In order to improve the material’s compatibility with the printing process, ceramic atomized materials were mixed with additives. The primary function of these additives is to maintain the structural integrity of the green body, given that compaction by pressure is not performed during the printing process. The additives provide the necessary binding power to shape the ceramic material. After removing the excess support powder, the green piece remains fragile, making the role of additives crucial at this stage.

The raw base material used was a refractory cement (d50 = 15.6 µm). To improve strength and hardness, a mixture of felspathic sands and ceramic frit with two different mean particle sizes was used: FS&CF1 (d50 = 184.8 µm) and FS&CF2 (d50 = 232.6 µm). To reduce shrinkage during the drying and firing processes and to impart rigidity to the body, quartz sand (d50 = 264.4 µm) and micronized quartz (d50 = 17.0 µm) were incorporated. To provide strength against warping during firing and to add fired strength, calcinated alumina (d50 = 3.0 µm) was selected. The calcinated alumina was obtained from Alteo (Gardanne, France), while all other materials were supplied by ROCA Sanitario S.A. (Gavà, Spain).

Variations in the percentage of the cement and the different additives were used to create nineteen different compositions to test the printability and the properties of the printed parts (Figure 2).

2.2.2. Binder

In the present project, the hydration-activated binding method was employed. This method involves the utilization of a simple liquid (e.g., water) to selectively bind the self-binding powder bed. The absence of an adhesive is a consequence of the sufficiency of hydration for activating the setting.

A water-based liquid binder (85–95% water) was selected. As part of the liquid binder formulation, glycerol (Merck KGaA, Darmstadt, Germany, ≥99.5%) was dissolved in deionized water. Glycerol was utilized to increase the viscosity and surface tension of the liquid binder, with the objective of improving droplet definition and infiltration kinetics. The aqueous solution was magnetically stirred for 10–15 min to obtain a clear solution, a viscosity within the range of 1.1 and 1.3 Pa·s, and a surface tension of 72 ± 0.3 mN/m.

An organic gelling agent was utilized as a solid binder, exhibiting a solubility rate greater than 99.9% and a viscosity of 0.075 Pa·s (after 24 h in 25% of H₂O).

2.2.3. Preparation of Powders

In order to achieve a homogeneous mixture, the ceramic material, the additives, and the solid binder were dry-mixed with an Eirich planetary mixer for 10 min at 150 rpm.

To reduce porosity and improve printing resolution, small particle size distributions are favorable, as they allow for smaller layer thicknesses [11]. However, that same packing property, which is desirable for sintering, becomes a problem for the flowability of the material [12].

The flowability of solids is complex and linked to many parameters. It is an important property for dried particles, which is measured using the Carr Index (CI) and the Hausner Ratio (HR). The powder flowability (%), as determined by the CI value, is classified as follows: <15 very good, 15–20 good, 20–35 fair, 35–45 bad, and >45 very bad. The powder cohesiveness, as determined by the HR value is classified follows: <1.2 low, 1.2–1.4 intermediate, >1.4 high [13].

To ensure the printability, all the compositions made had a similar flowability to that shown in

Table 2. Powder characteristics.

Humidity H (%)	Aerated Bulk Density ${\mathbf{\rho }}_0$ (kg/m3)	Packed Bed Density ${\mathbf{\rho }}_{\mathbf{f}}$ (kg/m3)	Carr Index CI (%)	Hausner Ratio HR (${\mathbf{\rho }}_{\mathbf{f}}$/${\mathbf{\rho }}_0$)
0.92	1287 ± 28	1730 ± 21	25.61	1.35 ± 0.04

. In accordance with the findings of previous studies [14], a HR of 1.35 (intermediate) was considered to be correct. The Carr Index value obtained was deemed to be satisfactory according to the available bibliography. All the tested powders demonstrated comparable spreadability.

2.2.4. Printing Parameters

In addition to the powder-based parameters, there are other factors involved in BJ that impact the mechanical properties of sintered parts, such as the printing parameters [15]. In this study, the printer parameters were kept constant at the following values (

Table 3. M10 printing parameters.

Parameter	Value
Layer height	0.16 mm
Feed box ratio	1.3
Printing	2
Nozzle	2
Slow axis	12,000 pps

).

The layer height was established based on the mean particle size of the compositions.

The feed box ratio parameter allows for the provision of a greater quantity of powder. The value of 1.3 means that the build platform is lowered by 1 mm, and then the feed platform is raised by 1.3 mm, creating a greater compaction of the material.

The printing parameter refers to the number of printing repetitions in the same position, which is to be adjusted according to the amount of binder used.

The nozzle parameter sets the number of nozzle slots to be used (1 to 4, with the understanding that an increase in the number of nozzle slots results in higher binder saturation).

The slow axis parameter establishes the speed of the axis where the nozzle is mounted, and this can range from 16,000 pps (3840 mm/s) to 4000 pps (960 mm/s).

2.3. Drying and Sintering

After printing, the green body and construction bed containing the remaining powder were dried in an oven at 80 °C for four hours.

The parts were then removed from the oven, and any remaining unbound powder was carefully removed with a metal bristle brush.

Finally, the parts were sintered in an electric furnace according to an industrial firing cycle (Figure 3) in a laboratory electric furnace (Carbolite RHF, Carbolite Gero Ltd., Derbyshire, UK).

2.4. Glazing

A standard sanitary glaze was used for the glazing. The glaze was sprayed onto the green parts using different grammages. After firing, it was observed that the glaze had been partially absorbed, resulting in slight deformation of the shape and an irregular surface (Figure 4).

It was therefore decided to use a highly refractory engobe that would allow the glaze to be applied and stretched correctly. The engobe was applied to the green part with a grammage of ~908 g/m² and, after drying, the glaze was applied with a grammage of ~1024 g/m². The parts were then fired according to the above firing cycle. The fired parts showed a well stretched and homogeneous glaze with no signs of distortion on the piece (Figure 5).

2.5. Characterization of the Printed Parts

2.5.1. Pyroplastic Deformation

The magnitude of pyroplastic deformation is defined by the pyroplastic index (PI), which indicates the tendency of a specimen of given dimensions to be deformed by gravity during firing under specified conditions. The procedure used to determine the pyroplasticity index consists of measuring the curvature of a specimen resting on two refractory supports during firing using the following expression (1):

$$PI={\displaystyle \frac{4·e^2·s}{3·L^4}}$$

(1)

where e is the thickness of the part (cm), s is the maximum deformation (cm), and L is the distance between the supports (cm). The support spacing used in this case was 5.25 cm.

The specimens were fired at the maximum temperature of 1230 °C in an electric laboratory oven. The time spent at maximum temperature was 60 min.

2.5.2. Bulk Density

The density of the parts was measured using the Archimedes principle by submerging the specimens in a mercury bath and observing the increase in the weight of the bath following the method described by Amorós [16]. Mercury was used because it has a high degree of non-wetting properties and the parts do not need to be sealed for the measurement. The parts were measured before sintering. The bulk density is given by the following expression (2):

$$\rho ={\displaystyle \frac{m·\rho \mathrm{H}\mathrm{g}}{m\mathrm{f}\mathrm{l}}}$$

(2)

where ρ is the density measured in g/cm³, m is the weight of the part measured in g, ρ Hg is the density of mercury (13.53 g/cm³), and mfl is the weight of the fluid displaced by the specimen (g).

2.5.3. Dimensional Accuracy

The dimensional accuracy obtained by printing each material was analyzed by determining the linear shrinkage. The length of each part was measured, before and after sintering, using a digital caliper with an accuracy of 0.01 mm (Mitutoyo Europe GmbH, Neuss, Germany).

For the final selected material, the accuracy of the printing was also studied by scanning the green parts with a white-light-emitting hand scanner with a resolution of 50 µm (Academia 50, Creaform Inc., Quebec, Canada) and comparing them with the original digital files.

2.5.4. Compressive Strength

The compressive strength test was carried out on 8 cubic specimens measuring approximately 20 × 20 × 20 mm. The test consisted of applying a progressive compressive force perpendicular to the base of the specimens at a rate of 0.5 mm/min until the specimens broke.

The compressive strength was calculated using the following expression (3):

$$\mathsf{\sigma }={\displaystyle \frac{F_{max}}{A}}$$

(3)

where F_max is the maximum strength before the breaking point (N), and A is the compression area (mm²).

2.5.5. Surface Roughness

The surface roughness of the final (sintered) parts was established using a roughness meter (Hommelwerke TM 8000, Microtécnic Ibérica, Barcelona, Spain). Surface roughness measurements were performed in accordance with international standards DIN 4768 [17], BS 1134 [18], and ISO 4287 [19] to ensure consistency and comparability.

There are numerous parameters used to define roughness [20], and for this research, the roughness parameters studied were the roughness average (R_a) and the ten-point height (RzISO). Ra is the most commonly used roughness parameter for general quality control. Rz is more sensitive to occasional high peaks or deep valleys than Ra [20]. Since using Ra alone may cause some points, such as single protrusions, to be overlooked, it is important to use both Ra and Rz together.

Ra is the arithmetic mean of the absolute values of the distance of the points that make up the profile to a mean line (4):

$$R_a={\displaystyle \frac{1}{l_m}}{\int }_0^{lm}\left|y\left(x\right)dx\right|$$

(4)

where l_m is the evaluated length (5/6 of the total length, eliminating the first and last points).

RzISO is the arithmetic mean of the vertical distance between the highest peak and the deepest valley of each of the five fractions into which the evaluated profile has been divided, without the profile having been divided into equal sections (5).

$$R_{ZISO}={\displaystyle \frac{1}{5}}\left(\sum _{i-1}^5\left|Y_{pi}\right|+\sum _{i-1}^5\left|Y_{vi}\right|\right)$$

(5)

3. Results and Discussion

In the raw state, the primary factor influencing densification is the compaction of the powders, which is contingent, in part, on the pressure applied by the machine; in this study, the pressure applied remained constant for all compositions. The other main factor in the compaction of the compositions is the particle size distribution, which exhibited a direct correlation between the width of the particle size distribution and the bulk density obtained on the parts, as shown in Figure 6. The correlation between the width of the particle size distribution and the resulting density was therefore direct. This can be explained by a rearrangement of the particles between large and small populations.

3.1. Pyroplastic Deformation

As the parts are required to withstand 1230 °C, materials exhibiting high pyroplastic deformation were excluded from further evaluation. It is important to note that for deformations greater than 12 mm, the pyroplastic index was deemed excessive and, consequently, was not quantified. The results are presented in

Table 4. Pyroplastic deformation of the parts at 1230 °C.

Specimen	Pyroplastic Deformation Index
SA1	>12 mm
SA2	>12 mm (disintegrated)
SA3	>12 mm
SA4	Did not withstand handling
SA5	>12 mm
SA6	>12 mm (melted)
SA7	>12 mm (melted)
SA8	>12 mm
SA9	>12 mm
SA10	>12 mm
SA11	>12 mm
SA12	40.9·105 cm−1
SA13	>12 mm
SA14	11·105 cm−1
SA15	8.9·105 cm−1
SA16	9.3·105 cm−1
SA17	13.2·105 cm−1
SA18	12.2·105 cm−1
SA19	7.1·105 cm−1

.

As shown in the table above, compositions SA1 to SA13 exhibited very high pyroplastic deformation. Consequently, only compositions SA14 and onwards were further analyzed in this study.

These results indicate that calcined alumina is crucial for adequate performance during the firing phase, as only the compositions containing alumina had a pyroplastic deformation index lower than 14·10⁵ cm⁻¹.

3.2. Compressive Strength

A critical additive in the green state is the gelling agent, which acts as a binder for the composition. However, due to its organic nature, an increased amount of gelling agent will result in higher shrinkage and/or deformation during sintering. Compositions SA1, SA4, and SA10 were identical except for the amount of gelling agent (Figure 2). As expected, the composition without the gelling agent (SA4) crumbled on contact. From the gathered data, it was determined that the optimum range for the organic gelling agent was between 5–10% to achieve adequate green strength.

Figure 7 illustrates the compressive strength of the studied compositions.

3.3. Dimensional Accuracy

In addition to the need to manipulate the parts, the green strength is also needed to maintain the shape of the parts. As the prototypes must be true to scale, the amount of powder that the printed part must support during the printing process results in a significant weight that can distort the shape of the printed part (Figure 8a).

The higher compressive strength exhibited by SA18 (Figure 7) resulted in printed parts that could withstand the weight of the powder, thereby maintaining dimensional accuracy, as shown in Figure 8b.

Another critical parameter in achieving precise dimensional accuracy is contraction. Greater contraction in a printed part makes it more challenging to achieve high dimensional accuracy.

Figure 9 shows the linear shrinkage of the parts after sintering. All compositions exhibited a low shrinkage.

3.4. Surface Roughness

The engobed and glazed parts produced with the SA18 composition were subjected to a surface roughness analysis. A sanitaryware sample was used as the reference standard for the desired roughness. Figure 10 shows the topographic maps of each part.

Table 5. Surface roughness values.

Specimen	Ra (µm)	RzISO (µm)
Sample	0.11 ± 0.03	0.77 ± 0.18
SA18	0.11 ± 0.03	0.87 ± 0.13

shows the corresponding roughness values. The R_a value obtained in both parts was identical. However, there was a slight variation in R_zISO, which remained within acceptable tolerance limits. Figure 11 illustrates both parts and shows that the amount of glaze and engobe applied to achieve the observed roughness was still less than the amount used on the sample, leaving room for improvement in roughness if necessary.

4. Conclusions

The results presented here demonstrate the feasibility of producing full-size ceramic parts using binder jetting for the rapid prototyping and testing of sanitaryware components. The composition that best meets the various requirements is SA18. With the highest green strength, it can support the weight of the powder during printing, maintain dimensional stability, and allow for safe handling during post-processing. The measured pyroplastic deformation of 12.2·10⁵ cm⁻¹ fell within acceptable limits, confirming its suitability for a conventional firing cycle. Additionally, the application of an engobe facilitated proper glaze adhesion, resulting in a surface roughness comparable to that of industrially manufactured components. This ensured that subsequent testing would meet the technical requirements and certification standards of conventional sanitaryware.

As a final test, a full-size washbasin was printed, sintered, and glazed to prove the feasibility of binder jetting prototyping with ceramics. The positive results can be seen in Figure 12.