Optimization of Printing Parameters for Extrusion 3D Printing of Ceramic Clay ^†

Abstract

With substantial advancements in the additive manufacturing (AM) of polymers and metals in the last few decades, the AM of ceramic material has recently gained research traction. Manufacturing complex ceramic parts is an extremely challenging process that can be revolutionized by AM. While many industrial-grade expensive AM systems are available for printing ceramic parts, the use of inexpensive desktop-type extrusion AM systems for ceramic parts is particularly challenging, which mandates printing parameter optimization. In this paper, the printing parameters of an inexpensive extrusion AM system, such as layer height, line width, extrusion speed, infill, and flow percentage, were optimized, which resulted in ceramic parts with desired qualities in terms of part density. For this, Clay Body 370 was chosen to produce clay with different percentages of water content ranging from 5% to 15%. Printed samples were subjected to sintering in a furnace to sinter at 1100 °C, followed by shrinkage and weight loss measurements. Finally, the printing parameters that resulted in good surface quality and less dimensional shrinkage with higher compression strength were reported.

1. Introduction

Ceramics were discovered centuries before metal and polymers, but now this field lags behind metal and polymer industries, and there seems to be a void in this industry, which could open up good research opportunities in this field. Ceramics are one of the oldest materials made by mankind. Ceramics are used in different fields ranging from art to aerospace engineering for their properties like high-temperature tolerance, chemical stability, high strength, corrosion and abrasion resistance, etc. [1,2,3]. Ceramics can be formed from clays. Clay is referred to a type of powder with a small particle size. As the chemical composition varies, it can named kaolinite, vermiculite, smectites, mica, talc, etc. The most abundant clay type is found among alumino-silicate clays named kaolinite, a mixture of SiO₂ and Al₂O₃ with the formula 2SiO₂·Al₂O₃·2H₂O. It is widely used in cosmetics, pottery, paper, ceramics, and paintings [4].

One of the methods used in ceramics production is pottery. Pottery was among the first ancient tools used by mankind about 80 centuries ago, but it has not lost its importance even in today’s daily life and industry [5,6,7]. Traditional methods in pottery are of great value, but we are not yet able to manufacture advanced samples with the details that are needed for today’s industry [4]. Other methods of ceramic production include slip casting [8] and press molding [9]. The above-mentioned methods of ceramics production come with some drawbacks, including being time-consuming, constraints on dimensions, being labor-intensive, expensive molds, and being not good for small, detailed samples [10,11]. The technology that can help in this regard is additive manufacturing (3D printing), which enables us to produce more accurate and detailed ceramic parts. This being the case, additive manufacturing is of great interest to the ceramic industry [12,13,14]. Initially, 3D printing began with polymer-based fused deposition modeling (FDM), evolved into 3D metal printing, and is now transitioning towards 3D printing of ceramics [7,8,9]. Additive manufacturing has progressed, and it is increasing the printability of more complex designs. In contrast with thermoplastic printing, the printed sample does not turn into a solid after being extruded from the nozzle; for clays, it takes time to dry and maintain its shape, which makes clay printing more challenging [5].

Different 3D printing methods have been developed to produce ceramics like Powder Bed Fusion [15], Binder Jetting [16], Selective Laser Sintering [17], etc. They have pitfalls like not being able to produce high-density samples in short times, struggling to produce high-quality products of intricate geometries with smooth surfaces [18], and being susceptible to cracking due to rapid heating and cooling cycles [19]. Among these newly developed methods, ceramic extrusion 3D printers seem to have their benefits, namely being able to produce micro-sized samples [20] to help in construction [21]. This technique is not limited to a specific material and can print any flowable material. Extrusion 3D printing uses incremental forming processes where the paste is spread layer by layer through the nozzle onto the substrate until the 3D object is completed [22,23].

To perform ceramic 3D printing, both material properties and devices can impact the final quality of the ceramic [24]. There are many factors impacting clay properties including particle size, the type of impurities, shape, etc. [25]. On the other hand, the final quality of a printed sample can be significantly changed by changing printing parameters like flow, speed, printing path, etc. [26]. Quality is not accidental, and to obtain a high-quality sample, the printing parameters need to be optimized. Also, ceramics shrinkage after sintering greatly affects the final dimensions and quality of the samples, and the effect of printing parameters on the shrinkage needs to be studied [27,28,29,30,31]. We aim to increase the accuracy of ceramic printing by optimizing four of the printing parameters of an affordable 3D printer that works based on material extrusion.

2. Experimental Setup

An extrusion-based clay 3D printer (Model: Moore 2 Pro; Make: Shenzhen Tronxy Technology Co. Ltd.; Shenzhen, China) was utilized that uses a separate plunger-based feeding system to supply the clay to the deposition head via a flexible hose. The ceramic clay used in this work was M370 clay (Plainsman Clays Ltd., Medicine Hat, AB, Canada). The supplier-specified physical properties of this clay, as obtained, are as follows: drying shrinkage of 5.5–6.5%, water content of 21.7–22.7%, and loss on ignition (LOI) of 5.0–6.0%; the chemical composition is given in

Table 1. The chemical composition of the M370 clay (%) [28].

CaO	K2O	KNaO	MgO	Na2O	TiO2	Al2O3	P2O5	SiO2	Fe2O3	FeO	MnO	LOI
0.2	1.2	0.1	0.2	2.6	0.7	22.4	0.0	65.6	0.5	0.0	0.0	6.4

.

The obtained ceramic clay was diluted with 7% of water. To mix with water, a mechanical mixer (Vevor; JJ-1, Shanghai, China) with blades was used to form a smooth paste. Then, the clay was added to the printer’s container in small portions to avoid creating air gaps in the clay. A 3D model of a solid box with the dimensions of 40 × 40 × 20 mm³ was designed and used to print all samples. As a substrate, a paper layer was placed on the printer bed.

To optimize the printing parameters, as the main goal of this research, four main parameters were changed, namely layer height (LH), line width (LW), flow, and infill density. As a result, 32 samples were reported, with each changing one of these parameters. Before printing the 32 final samples, many trial-and-error tests revealed the approximate optimized parameters and some parameters were fixed and seemed to have a negligible effect on printing or were out of the scope of this research like nozzle size, speed, and infill pattern. We used layer heights of 1.5, 2, 2.5, and 3 mm; line widths of 0.8, 1, 1.2, and 1.4 mm, flows of 0.7, 0.9 mm/s; and infill densities of 40% and 55%.

Once the samples were printed, they were left to air-dry for at least two weeks and then sintered in a furnace (Model: 3-550, Make: Vulcan, Luoyang City, China), following the heating curve. Regarding the ramping rate, the furnace temperature was increased by a rate of 0.8 °C per minute to reach 150 °C and held for 60 min; then, the rate was increased by 0.16 °C per minute up to 600 °C and held for 10 min and, finally, by 3.3 °C per minute, reaching 1100 °C. The samples were left to cool down without opening the furnace door to prevent any thermal shock resulting in possible cracks. No signs of cracks were observed on the sintered samples. The dimensions of each of the 32 samples were measured using a caliper before and after sintering to calculate the shrinkage percentage. The density of the samples was also calculated by dividing the dried samples’ weights by their volumes before and after sintering.

3. Results and Discussion

Figure 1 presents 4 of the 32 final samples, which reveal the effect of the selected parameters. By comparing the side walls of the first two samples, the effect of layer height (LH) can be observed, which increases from 1.5 to 2 mm. Upon increasing this parameter from 1.5 to 2 mm, the density of samples is almost constant, but upon increasing it from 2.5 to 3 mm, the density increases. By increasing the LH, we need a smaller number of layers to complete a wall for a sample with the same height. Lower LHs can be used when using lower flows and infill densities, and higher LHs are good for samples with higher flows and infill densities to give the extruded material more space to deposit in samples 115 and 116. Excessive material is observed on sample 115, which cannot be properly sited on each layer, and by printing the next layer so close to the previous one, these materials find ways to escape from the walls and make a wall full of defects. However, by increasing LH in sample 116, the same amount of material had a proper place to be deposited and had a much better effect on the walls.

To observe the effect of line width (LW), samples 102 and 104 can be compared. By comparing the top surfaces of these two samples, it is easy to observe that by increasing LW from 0.8 mm for sample 102 to 1 mm for sample 104, the distance between the lines increases and the number of lines required to complete a square surface decreases. While increasing the LH can give more vertical space to extruded material, LW can make more room in the horizontal direction.

While LH and LW have negligible effects on the density, the flow and infill density can change a sample’s density noticeably. The effect of flow is compared by changing this parameter from 0.7 mm/s for samples 101–104 to 0.9 mm/s for samples 109–112. Comparing these shows that by increasing this parameter, more material can be extruded from the nozzle, resulting in a more compact and denser sample with lower free space, as the average density of the first row of samples is 66.6 g/cm³, which is increased to an average of 88.8 g/cm³. By comparing any line of samples with the next line, the effect of increasing the infill density from 40% to 55% can be noticed. More material is then extruded, resulting in a more condensed structure and less space. To see the effect of infill density on the samples’ densities, the average of the densities of the first row of samples is calculated to be 66.5 g/cm³, while the second row is 88.6 g/cm³.

Table 2. Final sample’s parameters, density, and shrinkage.

Sample	Layer Height (mm)	Line Width (mm)	Flow (mm/s)	Infill (%)	Density (g/cm3)	Density (Reduction%)	Shrinkage (%)
101	1.5	0.8	0.7	40	65.1	7.91	4.9
102	2	0.8	0.7	40	63.8	7.00	4.7
103	1.5	1	0.7	40	68.2	6.90	5.4
104	2	1	0.7	40	68.8	6.78	5.9
105	1.5	0.8	0.7	55	88.6	6.70	5.6
106	2	0.8	0.7	55	87.9	6.86	6.4
107	1.5	1	0.7	55	89.5	6.67	5.5
108	2	1	0.7	55	88.4	6.73	6.1
109	1.5	0.8	0.9	40	86.8	6.71	6.2
110	2	0.8	0.9	40	86.6	9.21	4.6
111	1.5	1	0.9	40	91.0	6.83	4.4
112	2	1	0.9	40	90.8	6.66	4.6
113	1.5	0.8	0.9	55	111.9	8.46	5.5
114	2	0.8	0.9	55	112.1	8.56	6.1
115	1.5	1	0.9	55	116.0	8.60	8.3
116	2	1	0.9	55	115.8	10.25	6.7
117	2.5	1.2	0.7	40	74.4	4.90	5.2
118	3	1.2	0.7	40	77.1	6.73	6.2
119	2.5	1.4	0.7	40	74.2	6.52	6.3
120	3	1.4	0.7	40	74.4	5.95	6.0
121	2	1.2	0.7	55	91.1	6.53	7.0
122	2.5	1.2	0.7	55	93.0	6.00	5.7
123	3	1.4	0.7	55	94.2	6.77	4.9
124	2.5	1.4	0.7	55	96.7	6.87	5.8
125	3	1.2	0.9	40	91.1	6.12	6.3
126	2	1.2	0.9	40	100.4	7.20	5.6
127	2.5	1.4	0.9	40	96.7	7.88	6.5
128	3	1.4	0.9	40	100.4	7.63	5.5
129	2.5	1.2	0.9	55	115.3	9.21	7.0
130	3	1.2	0.9	55	124.4	10.26	8.5
131	2.5	1.4	0.9	55	120.9	9.82	7.7
132	3	1.4	0.9	55	127.4	10.04	7.1

contains all four parameters for each sample, in addition to the calculated density of each sample after two weeks of drying and before sintering. It also shows the change in density and average shrinkage for each sample. Density reduction varies from 4.9 to 10.26 with an average of 7.47%. The average shrinkage in the X dimension is 6.2%, in Y is 6.0%, and in Z is 5.7%. The total average of all samples in all directions is 6.01%.

4. Conclusions

Additive manufacturing has helped the ceramics industry to make more accurate and detailed samples that were not possible to make using handmade methods. A total of 32 ceramic samples were made from kaolin clay containing 65.6% SiO₂ and 22.4% Al₂O₃ using extrusion 3D printing. Mixing the clay with 7% water gave the desired viscosity and printability. Samples were printed based on changing four printing parameters, layer height, line width, infill, and flow percentage, which resulted in ceramic parts with desired qualities in terms of part density. The samples showed that a higher layer height can be used when we need higher flow and infill density, as it gives more space for material to deposit and causes minimum defects by requiring a smaller number of layers for a sample. From the top view of the samples, we learned that the samples could have smaller numbers of lines, with more distance between each line, by increasing the line width, which could also be useful for achieving higher flow and infill densities.

The result of the dimensional measurement of samples before and after sintering revealed an average shrinkage of 6%, with 8.5% being the maximum and 4.4% being the minimum. While changing the layer height and line width affected the density of the samples negligibly, changing the flow and infill density had a great effect on the density. By increasing the flow from 0.7 to 0.9 mm/s and the infill density from 40% to 55%, more material was extruded, resulting in a more compact and denser sample with less free space.