Transforming stereolithographic prototypes into metal or ceramic models by polymer substitution with titanium powder

Introduction

Additive manufacturing using various materials finds widespread application across diverse fields. The list of such technologies and materials continually expands with the addition of new items [1–8]. Laser stereolithography, employed in photocurable polymers, stands out as one of the most extensively used additive technologies due to its diverse material options and a wide price range for equipment, ranging from costly industrial machines to affordable household devices. This equipment demonstrates the capability to produce highly accurate objects.

Converting polymer prototypes into metal or ceramic components requires specific equipment and skilled personnel, such as through methods like molding or cavityless casting.

In the realm of additive manufacturing for ceramic models, several noteworthy technological implementations exist. These include layer-by-layer sintering of powder ceramics, curing of photopolymerizing compositions with ceramic fillers, and local extrusion of ceramic-filled material [2–6]. Notably, these technologies involve high-temperature processing stages to eliminate the polymer binder, achieve surface gloss, conduct additional pore-filling impregnation, among other steps. Regarding additive manufacturing of metal models, methods encompass layer-by-layer sintering of metal powders, layer-by-layer model formation via polymerization of filled compositions, and local extrusion of filled material (followed by sintering) [4–12].

This paper explores the potential expansion of the application scope for stereolithography equipment that work with photopolymers. It specifically examines new possibilities for converting prototypes into ceramic or metal parts.

Experimental part

In our view, one solution to broaden the applicability of stereolithography involves producing polymer stereolithographic parts with available free volume. Such models could undergo subsequent cold filling with ceramic or metal powder, followed by heat treatment to eliminate the polymer component of the model (referred to as annealing) and the sintering of metal or ceramics (henceforth referred to as “sintering” without separately delineating the polymer burnout process).

During sintering, there is potential to synthesize a ceramic compound from the substance used to fill the model body [13; 14]. The sintering stage in this technology shares similarities with the method utilizing filled resins.

We have explored two straightforward methods to create models suitable for subsequent filling with alternate materials. These methods involve producing a thin-walled inverse fillable model, essentially a matrix, and a contourless model exhibiting significant free volume uniformly distributed throughout its body.

The process of obtaining the inverse model, or matrix, is depicted in Fig. 1. In this developed model, a matrix is created by indenting the working surface of the model by a specified thickness. The design incorporates a port for powder filling and includes a covering lid as integral parts.

This alternative approach shares some resemblance to sintering within a non-burnable mold [15]. However, in our scenario, the mold is combustible, and its intricacy is solely constrained by the designer’s capabilities. In place of matrix models, we suggest employing original contourless prototypes. When creating a typical stereolithographic part, the process involves outlining both external and internal contours while filling the space between them. Additionally, denser outer hatching is necessary for horizontal outer surfaces.

To implement the contourless technology in stereolithography, several modifications to the conventional process are required, such as:

– eliminating contours;

– removing denser outer hatching;

– using only one coordinate (either Х or Y) on each layer; the direction of hatching alternates through the layer;

– increasing the hatching step.

The sparse hatching serves as an internal frame structure that gets filled with the initial liquid resin during manufacturing. By the end of the process, the liquid resin naturally flows out from the model or is expelled by compressed air.

Fig. 2 illustrates both the conventional method and the validated contourless methods for creating a stereolithographic part of a cylinder.

The model produced using this technology adopts a cellular polymer structure, openly accessible on all sides, containing a uniformly distributed free volume internally. Experimental findings indicate that manipulating technological parameters, such as layer thickness (h) and hatching step (H_s), within the ranges of h = 100 to 200 μm and H_s = 0.5 to 0.9 mm, results in varying proportions of free volume within the model, ranging from 30 to 80 %. A designer can replicate a similar internal hollow structure that fills the model. However, in such cases, the model’s complexity significantly impacts the file size, potentially increasing it by several orders of magnitude. This could pose challenges for subsequent software processing to prepare the model for printing, potentially rendering it difficult or even impossible. In contrast, contourless models achieve the same effect by simply adjusting the model growing technology without burdening the file size with excessive complexity.

We opted for titanium powder with particle sizes ranging from 15 to 45 μm to explore the replacement of prototype material. Titanium is highly sought-after in engineering and medical applications. Additionally, both titanium and previously used silicon can serve as intermediate materials capable of forming ceramic compounds (such as nitride, oxide, and carbide) under suitable conditions, offering a wide array of structural properties [17–19].

The matrix models were dry-filled through the open facet, followed by sealing the closing lid with initial resin cured under a UV lamp.

For the filling of contourless models, a suspension of titanium powder and water was prepared. Filling the free volume in the models was conducted within an MK-mini vacuum chamber (“MK-Technology,” Germany). Experimental findings revealed optimal results when the filler constituted approximately 50 ± 5 vol. % of the suspension.

Subsequently, all samples underwent heat treatment. Both matrix and contourless models were sintered under similar conditions using an SUOL-0.25.1/12-I1 furnace (manufactured in Russia). Sintering temperatures ranged from 900 to 1200 °C. The process was executed with gas purging (argon, nitrogen) and with exposure to atmospheric air. Sintering durations did not exceed 10 min.

Experimental results

Fig. 3 displays photographs of the test contourless models captured at various stages of the technological process.

The model obtained after sintering remarkably retained its original shape and dimensions, showcasing the initial color of the titanium powder and exhibiting adequate mechanical strength and electrical conductivity. The furnace utilized in our experimentation is well-suited for gas purging (specifically, Ar and N₂ in our case) and effectively prevents the temperature from exceeding 1200 °C. When using Ar purging, a minor quantity of weakly sintered metal powder (forming a loose layer up to 100 microns thick) was observed on the surface of the models. Sintering conducted at lower temperatures (1000 and 1100 °C) also yielded integral samples, albeit with inferior surface quality. Considering that the melting temperature of titanium is 1670 °C, it can be inferred that elevated temperatures might enhance the final model’s quality.

As previously mentioned, the resulting samples exhibited the same colour as the original metal powder. Sintering the Ti-polymer model in an argon environment may instigate a reaction between titanium and the polymer of the initial prototype. Under such conditions, stoichiometric titanium carbide TiC, which conducts electrical current similar to Ti, can be formed, along with several non-stoichiometric carbides, such as Ti_xC_y [19; 20]. According to literature sources, titanium actively reacts with atmospheric gases like nitrogen and oxygen at high temperatures. This reaction results in the formation of titanium nitride and titanium oxide, respectively. It’s essential to note that not only metallic titanium but also its compounds, such as carbides and nitrides, have the potential to interact with oxygen [18–22].

Sintering the contourless models in a nitrogen environment resulted in samples that maintained the original shape and dimensions, akin to the experiments conducted in argon. These samples exhibited electrical conductivity, although at times, surface cleaning was necessary to verify this property. Both titanium nitride and titanium carbide act as metal-type conductors, particularly in the case of stoichiometric TiN or TiC compounds, while TiO₂ oxide functions as a dielectric material [23–25].

Sintering in a nitrogen environment imparted a reddish-brown color to the samples, characteristic of titanium nitride TiN (Fig. 4, a). However, upon discontinuing nitrogen purging and before the samples cooled completely, a layer of white or yellow color formed on their surfaces, typical of titanium oxide TiO₂. Fig. 4, b illustrates a sample obtained through heat treatment in atmospheric air. Both samples (Fig. 4, a, b) displayed a slightly molten surface, which is noticeable when compared with the sample in Fig. 3, c. According to reference data [26], the enthalpies of titanium dioxide and nitride formation are ΔH_f° = –944 ÷ –938 kJ/mol (for different modifications) and ΔH_f° = –323 kJ/mol, respectively. During sintering, after the removal of the polymer, the structure of the samples was notably porous, displaying a branched surface. Due to chemical reactions, the temperature on their surface increased, resulting in localized areas of melted material.

The matrix models underwent a sintering process akin to the previously described sintering of contourless prototypes. Fig. 5 illustrates the stages of sample preparation and the outcome of sintering with argon purging.

A layer of orange-colored substance developed on the surface of the sintered model, particularly noticeable on the protruding sections. A similar occurrence was observed on the contourless models with nitrogen purging, albeit in the recessed areas. The models generated in this manner, depicted in Fig. 5, c, more faithfully replicate the shape of the original model. Structurally, they also exhibit strength, and the orange layer can be removed from the surface using abrasive materials or tools. Post-removal, the sample adopts a grey hue, akin to titanium powder.

Spectroscopy (FTIR) of the substance extracted from the surface of matrix models indicated (Fig. 6) the closest resemblance to the spectrum of non-stoichiometric Ti₃C₂ carbide. Notably, the spectrum obtained lacked characteristic lines indicative of Ti–N bonds [19; 27–34].

Figs. 7 and 8 depict matrix samples sintered under nitrogen purging and in an air environment, respectively. In the case of nitrogen purging, the samples sintered in this medium exhibited a more oxidized surface compared to those sintered in an air environment. The matrix samples prepared for sintering were densely packed with metal powder, as previously described. The metal was encapsulated from the surrounding medium by the matrix itself, formed by a layer of cured polymer. Heating initiated the thermal decomposition or combustion of the polymer, prompting the metal to react with the resulting decomposition products. Previous findings indicated that in an inert argon environment, this led to the formation of a layer composed of titanium-carbon compounds on the surface. As the polymer layer burned out, it gradually thinned and became permeable to atmospheric air or other gases, initiating competing processes between the formation of titanium carbide and titanium nitride [35; 36]. Titanium nitride reacted with heavier products of polymer thermal decomposition, containing a significant oxygen content, and transformed into oxide. It’s noteworthy that oxidation of titanium nitride commences at lower temperatures compared to titanium carbide. The simultaneous occurrence of these exothermic reactions on the sample surface induced additional heating and surface melting, irrespective of the temperature used for heat treatment (refer to Fig. 7).

Samples sintered in atmospheric air (Fig. 8) exhibited distinct differences from those treated with nitrogen purge. One notable characteristic was their sensitivity to the sintering temperature: lower temperatures (within specific limits) resulted in better retention of the sample’s shape and a cleaner surface. Moreover, these samples, despite exposure to atmospheric oxygen, showed minimal amounts of generated titanium oxide, observed only at temperatures surpassing 1100 °С.

We hypothesize that such sintering behavior can be elucidated by the fact that in the presence of atmospheric oxygen, the polymer matrix underwent combustion rather than decomposition. The primary combustion products (H₂O and CO₂) were not retained on the surface of the sintered sample. Additionally, the outflow of hot gaseous combustion products from the reaction area hindered the entry of nitrogen and oxygen. Combustion of the polymer also resulted in a decreased presence of carbide on the titanium surface. Notably, at a sintering temperature of 900 °C, the samples with the cleanest surfaces, in comparison to others, were produced.

Results and discussion

Fig. 9 presents cross-sectional photos of the sintered samples purged with argon. The contourless sample retained a mesh structure at the edges, inverse to that of the original prototype. In the center, there’s an area of sintered powder characterized by discontinuous pinpoint metallic luster. The matrix sample exhibited a darker outer layer, while the rest of the sample displayed a shiny metal appearance.

These investigations validate the feasibility of producing 3D metal models using a stereolithography apparatus, without necessitating foundry technologies. Each of the alternative technologies presented here holds distinct characteristics to be considered during the development of a final technological solution. Selecting a viable solution should follow a phase of experiments and tests evaluating the strength of the model.

All models created using the described methods required post-processing to eliminate the outer layer of the material. This post-processing step should be factored in during the computer modeling phase.

We are intrigued by the prospect of merging contourless and matrix technologies into one by creating a contourless matrix (as shown in Fig. 10). The wall of such a matrix is anticipated to be filled with a metal-water suspension, while the inner cavity may be filled either with dry powder or the same suspension. By selecting an appropriate sintering mode that facilitates easy removal of the weakly sintered outer porous part, the final model will resemble a sintered powder devoid of channels from polymer elements, minimizing its contact with external media during sintering.

Conclusions

Therefore, our experiments have validated the feasibility of transforming stereolithographic prototypes into metal or metal-ceramic forms. Through the use of titanium as an exemplary material, we have effectively showcased the genuine potential for such a transformation. By employing the described prototypes, we expand the utility of stereolithography, enhancing its versatility. Notably, the proposed solution does not necessitate any alterations to the photopolymer.

In contrast to additive technologies applied to metals or ceramics, which typically involve the use of significantly more expensive equipment, stereolithography equipment has consistently stood out as one of the most precise and user-friendly types of additive technology.

Additionally, it’s noteworthy that the methods we propose facilitate the manufacturing of models with varying degrees of porosity. This aspect holds significance in both technical and medical applications of such models.