Rapid Screening of Liquid Metal Wetting for a Materials Compatibility Library

Abstract

Wetting behavior of molten metals on solid substrates is a critical phenomenon influencing numerous industrial applications, including welding, anti-corrosion coatings, and metal additive manufacturing (AM). In particular, molten metal jetting (MMJ), an emerging AM technology, requires that the molten metal remain pinned at the nozzle exit. Thus, each new metal requires a specific nozzle material to ensure consistent droplet ejection and deposition, making it important to rapidly identify the appropriate wetting combinations. However, traditional measurements of wetting angles require expensive equipment and only allow one combination of materials to be investigated at a time which can be time consuming. This work introduces a rapid screening method based on sessile droplet experiments to evaluate wetting profiles across multiple metal–substrate combinations simultaneously. This study investigates the wetting interactions of molten Al alloy (Al4008), Cu, and Sn on various ceramic and metal substrates to identify optimal material combinations for MMJ nozzle designs. Results demonstrate that Al4008 achieves wetting on ceramic substrates such as AlN, TiO₂, and SiC, with varying mechanisms including chemical reactions and weak surface interactions. Additionally, theoretical predictions regarding miscibility gaps and melting point differences were verified for Cu and Sn on refractory metals like Mo and W. Findings from this study contribute to the establishment of a materials compatibility library, enabling the selection of wetting/non-wetting combinations for stable MMJ operation. This resource not only advances MMJ technologies but also provides valuable insights for broader applications such as welding, coating, and printed electronics.

1. Introduction

Wetting of a liquid on a solid surface occurs when the contact angle between a droplet of the liquid and the solid surface is below 90°. Three types of wetting occur when a molten metal contacts a solid substrate [1]. First, non-reactive wetting, wherein the molten metal spreads over the solid surface through the surface interactions at the liquid-solid interface [2,3]. Second, reactive wetting, where the liquid reacts with the solid surface and forms a layer of stable compound on the solid substrate that provides a preferable wetting surface for the liquid, in this case, the liquid is actually wetting to the new compound formed at the interface rather than the solid substrate [4]. Third is dissolutive wetting, where the molten metal droplet dissolves the solid surface until it becomes saturated. This dissolution lowers the observed contact angle through two mechanisms. First, the surface energy of the liquid decreases through dissolution of the solid surface, which typically occurs if the solid has a lower surface energy than the liquid. The other mechanism comes from the crater formed in the solid from the dissolutive process, which changes the geometry of the contact surface such that the true wetting is not measured flat against the surface but following the curvature of the crater at the solid/liquid/vapor junction. This angle has been found to be the intrinsic contact angle of the liquid droplet saturated with the solid while sitting on the flat solid surface, resulting in a lower observed contact angle [5]. In practical applications, dissolutive wetting is usually avoided, as it involves destroying the interface between the two materials, contaminating the liquid, and degrading the solid. Conversely, if the interactions between the molten metal and solid surface are weak/negligible, then the droplet remains in a non-wetting state with a contact angle greater than 90° [6]. Since wetting is fundamentally governed by interfacial phenomena, atomic-scale and nanometric interactions at the liquid–solid interface can critically influence the measured contact angle. Even subtle nanoscale chemical reactions, adsorption events, or rearrangements of surface atoms can modify the local surface energy landscape, thereby altering the macroscopic wetting behavior. Consequently, accurate measurement of equilibrium contact angles requires strict control of surface cleanliness and chemical stability—particularly the prevention of oxide formation or other surface films that can mask or artificially enhance true interfacial interactions [7,8].

Understanding the wetting behavior of molten metals on solid surfaces is crucial for many industrial and research applications, including welding [9], anti-corrosion coating [10], gas atomization [11], spray forming [12], suppressing dendrite formation in battery cycling [13], and, of particular relevance to this work, molten metal jetting (MMJ) [14]. MMJ is a metal additive manufacturing (AM) technology that uses droplet-on-demand ejection of molten metal droplets onto a build plate to build a part layer-by-layer [15]. The droplets are ejected by pressurizing molten metal inside a precisely fabricated nozzle. In MMJ, precise control over droplet formation and deposition is paramount, making the understanding and control of wetting phenomena near the ejection orifice even more critical. In particular, to achieve highly repeatable ejections where each droplet has the same size, velocity, and directionality, it is important that the meniscus of the molten metal rests at the exit edge of the orifice [15]. Maintaining the position of the meniscus at the exit edge before each ejection event improves directionality, consistency in droplet velocity and size, and reduces the chance of satellite droplet formation. We refer to this effect as “pinning the meniscus”. One method of pinning the meniscus at the outer edge of an orifice requires a combination of wetting (metallophillic) material and non-wetting (metallophobic) coating to be used for the nozzle to ensure the molten metal wets to the inside of the orifice but does not freely spread across the outside upon subsequent droplet ejections. This work investigates wetting interactions of several combinations of molten metal and solid substrates for deployment into emerging MMJ applications, such as metal additive manufacturing [14,16], printed electronics [17], part repair [18], powder production [19], and extreme ultraviolet light generation [20].

By understanding trends behind the various liquid metal–solid substrate interactions, we can select specific material combinations to achieve the desired wetting conditions as needed for technologies such as MMJ. To investigate these complex interactions, we employ a sealed, inert vacuum furnace that allows for straightforward sessile droplet experiments across multiple molten metal and substrate combinations simultaneously. Rather than a precise determination of contact angle, the screening method proposed in this work is optimized to search for material combinations that offer non-reactive wetting of liquid metals and nozzle materials over the course of hours at temperatures well above the melting point of the metal. This requirement is driven by the specifics of MMJ droplet deposition, which utilizes superheated molten metal in a crucible to generate droplets with precise size. The wetting screening approach proposed here aims to rapidly identify promising substrate/melt combinations. This rapid screening capability allows for compatible material interfaces to be identified with standard equipment that requires low investment. Furthermore, this method can be used to identify multiple molten-metal/substrates either using one molten-metal and multiple substrate options or multiple molten-metal place on one substrate to identify favorable wetting conditions in both MMJ and other wetting applications as well. The method outlined here contrasts with experimental methodologies employed when the exact contact angle measurements are required, such as traditional sessile drop experiments [21], dispensed drop technique [21], advancing and receding contact angle measurement [22], Wilhelmy plate tensiometry [23], and capillary rise method [23].

In the case of traditional sessile drop experiments, a small amount of solid metal is carefully placed on a substrate that is heated to the melting point of the metal, and a high-resolution, high-magnification camera is used to monitor the static contact angle. In contrast, the dispensed drop technique leverages the dispensing of a molten metal droplet onto a substrate, where the contact angle can be monitored in both dynamic and static conditions. The high magnification of the cameras and highly localized nature of the dispensed drop technique require that only one metal/substrate combination be investigated at a time. Lastly, the Wilhelmy plate and capillary rise methods involve immersing a vertically oriented substrate plate into a molten bath of molten metal. These methods require that the surface of the liquid remains clean and undisturbed outside of the substrate, and thus can only accommodate one molten metal and substrate combination at a time to obtain high precision measurements [23]. Furthermore, the Wilhelmy plate and capillary rise methods require that the substrate plate be submerged in the molten bath, which requires a large volume of the molten material. This requirement makes these methods prohibitively expensive for measuring the wetting of rare or expensive metals and can raise safety concerns when dealing with toxic or radioactive metals.

In contrast, our approach involves a sessile drop experiment setup with a wider field of view camera, which can effectively evaluate the wetting and non-wetting profiles for multiple metal–substrate combinations at a time, while sacrificing precision in contact angle measurements. By observing multiple material combinations within a single measurement, we can reduce the time needed to find compatible materials for wetting/non-wetting application needs. Since material combinations are all tested together, we can also dynamically increase the temperature and identify the point at which each wetting event occurs, further reducing the number of measurements needed to explore the vast material and parameter space. This aims to provide a rapid screening technique to determine the wetting/non-wetting nature of these molten metal/substrate combinations and assess whether the wetting is reactive, non-reactive, or dissolutive. This technique also allows cheaper equipment to be utilized as a standard tube furnace, vacuum pump, and USB camera, with a 750 mm zoom lens, which is all that is needed to set up the experiment. The results of this work are intended to form the basis for a materials compatibility library that will be useful across multiple disciplines that rely heavily on the wetting nature of metals on various substrates, which can support numerous application spaces. In this paper, we specifically focus on the wetting characteristics of molten Al alloy (Al4008), Cu, and Sn on various ceramic and metal surfaces, given that these three metals are currently the focus of much MMJ work across applications [17,18,24,25]. Three additional requirements for MMJ nozzle materials are that

(1)

The wetting material is manufacturable into bulk components via milling, pressing and sintering, casting, or additive manufacturing;

(2)

The non-wetting material can be applied as a coating on the outside of the nozzle via spray coating, vapor deposition, or chemical growth methods;

(3)

The wettable material, the coating, and their interface are all stable—well above the melting point of the target liquid metal.

Thus, by utilizing our rapid screening system, we propose to find the optimal nozzle/coating combinations to ensure a pinned meniscus, which will advance the quality of MMJ systems to eject metals efficiently and provide a screening method for other use cases.

Despite the convenience of our proposed method to screen wetting materials, it is still necessary to narrow the field of candidate materials using a theoretical framework. The fundamental theory of molten metal wetting suggests that pure metals with large differences in melting temperatures and large miscibility gaps should exhibit perfect wetting behavior when the lower melting point metal is in contact with the solid high melting point metal [1,6,26]. We verify these theoretical predictions for the cases of molten Cu on Mo and combinations of molten Sn and solid Mo and solid W. Cu and Cu alloys exhibit exceptional electrical conductivity and thus represent a promising use case for MMJ in printed electronics, making it important to understand its wetting properties [17]. Additionally, Sn has been a staple material in the generation of extreme UV radiation (EUV) for nanolithography through the laser-liquid Sn interactions and thus serves as a good opportunity to illustrate that this method can support various applications outside of MMJ as an AM technology.

According to binary alloy phase diagrams of Al, there are no candidate metals that have both a higher melting point than Al and do not form intermetallic compounds with Al. Thus, the wetting of Al4008 is only achieved using non-metal/ceramic substrates. Our investigation identifies AlN, TiO₂, and SiC as possible wetting surfaces for Al4008, where all three show a low observable contact angle with molten Al4008. These findings indicate that the wetting of Al4008 on TiO₂ and SiC is induced by chemical reactions at the metal/ceramic interface, while wetting of Al4008 to AlN is primarily due to the weak interactions between surface nitrogen atoms and liquid aluminum atoms, allowing for wetting without any reactions [27,28,29].

In establishing a materials compatibility library based on our wetting studies, we aim to provide a resource for the MMJ community. This library will enable the selection of the required wetting/non-wetting combinations needed to pin the meniscus of a desired molten metal at the orifice exit, ensuring stable jetting. This will help greatly expand the viable material space for MMJ. Furthermore, engineers and researchers in other fields can use such findings to quickly assess the wettability of various metal–substrate combinations, thereby facilitating informed decisions in applications such as welding, coating, and additive manufacturing.

2. Materials and Methods

2.1. Sample Preparation

The wetting of molten metals to the substrates is sensitive to the quality of the interface between the two materials. To ensure that the wetting measurements provide an accurate indication of the wetting behavior, the molten metal and substrate surfaces must be free of any contaminants and oxides which require careful preparation of both materials before the experiment begins. The Mo, and W substrates are cut from sheets of each substance. The AlN, SiC, and TiO₂ ceramic substrates were purchased as 10 mm × 10 mm single-crystal samples from MTI Corporation (MTI corporation, Richmond, CA, USA). The Al₂O₃ and hBN (hexagonal boron nitride) substrates were purchased from McMaster-Carr (McMaster Carr, Los Angeles, CA, USA) and Kennametal (Kennametal, Pittsburgh, PA, USA), respectively. The process begins with the preparation of sample metals, which are polished with 600-grit SiC sandpaper and cleaned using isopropanol to remove any contaminants that could affect the wetting behavior. It should be noted that surface roughness has, in the past, been shown to affect wetting conditions. Previous literature has shown that a rougher surface exacerbates the natural wetting behavior between a liquid and solid surface, such that non-wetting interactions are more non-wetting and wetting interactions show stronger wetting [1]. This result comes from two effects: the increase in the actual surface area, and the second is the pinning of the solid/liquid/vapor junction by sharp defects. In a system with strong wetting interactions, the increase in surface results in more spreading and a lower observed contact angle. In systems with poor wetting interactions, the defects due to roughness lead to the formation of composite interfaces, partly solid–liquid and partly solid vapor. In this case, the droplet perimeter is pinned and even detaches during cooling due to adhesive rupture under solidification shrinkage. Thus, to make precise measurements of the equilibrium contact angle, a highly flat and smooth surface is required. However, here, the goal is to create a rapid screening technique to categorize the wetting interaction as either wetting or non-wetting, and then ascertain the type of wetting. In particular, we are motivated by practical wetting scenarios of molten metal jetting. Achieving a mirror-polished surface within the bore of a capillary channel is extremely difficult; therefore, performing highly polished surface experiments would not be informative for these applications. This consideration is important for our results to be used in practical design for MMJ and other liquid metal wetting systems. Using a rough surface is therefore more representative of the application space and helps to exacerbate the natural tendencies of the surface interactions. After polishing the metal surfaces, both the sample metal and substrate surfaces are wiped with isopropanol to remove dust or other contaminants from the surfaces. Once prepared, the samples are placed in an alumina crucible and transferred into a tube furnace with a platform welded to the inside of the tube. The welded platform ensures the samples remain on a flat plane while loading, such that the metals do not roll off the substrates during loading or heating

2.2. Wetting Experiment Setup

The experimental setup utilized for this study is illustrated in Figure 1a, which shows the tube furnace with the welded platform. The camera uses a zoom lens such that the field of view allows for 3–4 combinations to be viewed at the same time. The high temperatures achieved in the tube furnace result in large amounts of IR radiation, which is blocked by the IR filter window at the end of the furnace. The backlight illuminates the samples to make them visible. The furnace uses Kanthal APM FeCrAl tubes (Kanthal Corporation, Amherst, NY, USA), which can withstand high temperatures, and have welded KF fittings to prevent contamination of the atmosphere during heating. Prior to heating, the furnace is pumped down to a pressure of 10⁻⁴ Torr and backfilled with Ar gas three times [30] to ensure a clean environment. Following this, a high vacuum pump is employed to reduce the pressure within the furnace to ~10⁻⁶ Torr. This low-oxygen environment is essential for maintaining a clean atmosphere during the heating process, thereby preventing excessive oxidation during heating that would hinder wetting of the molten metals [22]. It also allows any residual isopropanol to evaporate before the heating process begins. The tube furnace is then heated to 200 °C at a rate of 5 °C/min and held for 1 h to remove any moisture or impurities from the tube furnace environment. Then, the tube is heated to the target temperature again at 5 °C/min. The furnace is then held at the target temperature for 5 h and then allowed to cool to room temperature at 5 °C/min. It should be noted that when the temperature is below 200 °C, the furnace naturally cools at a slower rate than 5 °C/min, so the heater turns off at that point. During the heating phase and while the target temperature is held, images of the molten droplets are captured every minute using an Allied Vision U-240m camera (Allied Vision, Stadroda, Germany) with a fixed 100 mm focal length zoom lens. A LabVIEW control program is used to capture and save one image every minute. The images taken by the camera allow observation of the contact angle between the molten metal and solid substrate, as shown schematically in Figure 1b–e. When the wetting angle is 0–90°, the molten metal fully wets the substrate (Figure 1b–d), and if the contact angle is much greater than 90°, the molten metal is non-wetting to the substrate (Figure 1d). This time-lapse imaging allows for detailed observation of the wetting behavior as the molten metal interacts with the substrate. The time-lapse shown in the supplementary information helps pinpoint when wetting begins. For example, some wetting interactions rely on reactive wetting where the liquid may not initially wet the substrate at the target, but over time, the reaction between the solid and liquid may progress and help lower the contact angle. By continuously monitoring the samples, we can gain an understanding of the kinetics that dictate the rate of this reaction, which would not be apparent from ex situ observations. Images correlating to the stabilized wetting angle were extracted as snapshots to illustrate the static wetting angle in this work. After the heating phase, the furnace is flooded with Ar, and samples are removed for post-mortem analysis.

2.3. Microscopy Preparation and Elemental Mapping

For wetting combinations, we employ Scanning Electron Microscopy (SEM) (Thermo Fisher Scientific, Waltham, MA, USA) coupled with Energy Dispersive X-ray Spectroscopy (EDX) (Thermo Fisher Scientific) analysis techniques to observe the post-heated interface between each metal and substrate after removing the samples from the furnace. We mount the wetted metal/substrate samples in epoxy and grind the surface with 320, 600, 800, 1000, and 1200 grit SiC sandpaper to expose the interface. After grinding, the surfaces are polished with 8 µm, 3 µm, and 1 µm diamond suspensions, and finally, samples are polished with vibratory polishing for 20–30 min using an 80 nm oxide polishing suspension (OPS). Using EDX, we examine the interface between the molten metal and the wetting substrate in detail and confirm that there is no erosion or dissolution of the substrate material during the experiment. By measuring the distribution of constituent elements of the molten droplets and substrate materials at the wetting interface, we can determine which type of wetting occurred. The various types of wetting are shown in Figure 1b–d. If the wetting interface is stable and there is no elemental migration across the interface, then the nature of the wetting was non-reactive (Figure 1b). If elemental migration does occur across the interface but a stable phase forms to prevent erosion of the interface, then the relationship is identified as reactive wetting (Figure 1c). Finally, if the interface is eroded and the solid substrate material is incorporated into the melt, then the interaction is labeled as dissolutive wetting (Figure 1d). It is also important to recognize that nano-scale reactions or atomic interdiffusion processes may occur even in systems that exhibit slow reaction kinetics. Such localized interactions can promote wetting by lowering interfacial energy, even when no discernible reaction product is detected at the microscale level. Conventional techniques such as SEM–EDX lack the spatial resolution to detect elemental migration or bonding changes at these nanometric length scales. Therefore, when a system exhibits wetting behavior without an apparent interfacial compound, it may still be driven by sub-surface atomic interactions that play a decisive yet experimentally subtle role.

3. Results and Discussion

3.1. Wetting of Metals with Negligible Mutual Solubility

The wetting behavior of a liquid drop on a solid substrate can be effectively described using Young’s equation [6], which relates the contact angle ($\theta$) of a liquid on a solid surface to the surface energies at the solid–vapor, liquid–vapor, and solid–liquid interfaces. Specifically, Young’s equation is expressed as follows:

$$\begin{array}{c}\mathit{cos} \theta ={\displaystyle \frac{{\sigma }_{sv}-{\sigma }_{sl}}{{\sigma }_{lv}}}\end{array}$$

(1)

where ${\sigma }_{sv}$, ${\sigma }_{sl}$, and ${\sigma }_{lv}$ represent the surface energies associated with the solid–vapor, solid–liquid, and liquid–vapor interfaces, respectively. When considering pure metals with negligible mutual solubility for both the liquid and solid substrate, this equation can be reformulated to [6]

$$\begin{array}{c}\mathit{cos} \theta ={\displaystyle \frac{{\sigma }_{LV}^A}{{\sigma }_{LV}^B}}-{\displaystyle \frac{\lambda }{L_e^B}}\end{array}$$

(2)

In this formulation, ${\sigma }_{LV}^A$, and ${\sigma }_{LV}^B$ denote the liquid–vapor surface energies of the solid metal substrate A and the molten metal B, respectively. The term $\lambda$ serves as an approximation of the regular solution parameter, calculated as the average of the enthalpy of solution at infinite dilution for each metal when considered as solvent and solute. Finally, $L_e^B$ represents the molar heat of evaporation of the molten metal B. Notably, the first term in this equation is typically much larger than the second. For example, in the Pb/Fe system, the first and second terms are 4 and 0.7, respectively, leading to the prediction that pure metals with negligible solubility should exhibit perfect wetting [6]. It should be noted that this approximation diverges from experimental observations due to three factors. Firstly, the presence of oxide impurities, on either the molten metal surface or the substrate surface, creates a barrier between contacting surfaces, preventing pure wetting between material A and material B. Second, the adsorption of gaseous oxygen on the substrate surface does not change the metallic bonding nature of the solid/liquid interface, but it does alter the surface energies such that perfect wetting is prevented [6]. Finally, surface roughness can increase wetting by increasing the contact surface for wetting combinations, and it can pin the droplet perimeter and arrest spreading at sharp defects in non-wetting scenarios [1]. As will be shown in this work, metal systems with low mutual solubility still show good but not perfect wetting under experimental conditions.

To test this prediction, we selected 3 metallic pairings with low to negligible mutual solubility: Cu/Mo, Sn/Mo, and Sn/W. To explore these wetting interactions, we first examine the wetting behavior of the molten metals in a controlled environment, as depicted in Figure 1, and then we utilize post-heating metallographic characterizations, such as SEM imaging and EDX mapping, to understand the molten metal–solid interactions. The experimental observations made in this study are in agreement with the previously mentioned theoretical framework. For instance, Cu is placed in contact with Mo and hexagonal boron nitride (hBN) and then melted at 1100 °C in our vacuum furnace setup. Cu and Mo have low mutual solubility, and thus it would be expected that they would wet well. This prediction is confirmed in Figure 2a, which illustrates the wetting behavior of molten Cu on Mo and hBN at 1100 °C, revealing that molten Cu effectively wets Mo while demonstrating non-wetting behavior on hBN (a video of this wetting process is available in Supplementary Information, Supplementary Video S1). To verify the non-reactive nature of this interaction, we analyze the interface of the Cu and Mo using SEM and EDX mapping, depicted in Figure 2b–d. In this elemental mapping, we see a clear interface between Cu and Mo and no mutual diffusion across the interface. This observation confirms that there are no reactions or mixing at the Cu-Mo interface. The equilibrium phase diagram of the Cu-Mo system shows negligible mutual solubility between these two metals, which aligns well with the experimental observations shown in Figure 2.

Sn has been a staple material in the generation of extreme UV radiation (EUV) for nanolithography through the laser-liquid Sn interactions. In contrast to Cu, Sn has a much lower melting point at 232 °C. It also has negligible mutual solubility with W and Mo, which both have significantly higher melting temperatures. Thus, based on the conclusions derived from Equation (2), it is expected that Sn would wet W and Mo once it is fully melted at relatively low temperatures. However, this is not the case, as seen in Figure 3a, where the wetting behavior of fully molten Sn on Mo, W, and BN at 600 °C suggests that Sn does not wet any of these substrates. The snapshot of the wetting setup at 600 °C reflects the temperature at which Sn will be jetted from an MMJ system and is thus representative of the wetting expected at operating conditions. In fact, the full video (supplementary Video S2) of the wetting procedure shows that Sn does not wet these substrates at low temperatures at all. However, when the temperature increased to 1100 °C, as shown in Figure 3b, Sn begins to wet both Mo and W, while still not wetting BN. Interestingly, despite Sn melting at around 230 °C, it does not exhibit wetting behavior on Mo and W until the temperature exceeds 800 °C (a video of this wetting process is available in Supplementary Information). This delayed wetting is likely attributed to the presence of interfacial impurities such as oxides from air exposure or residual carbides from the polishing procedure. These impurities can create a barrier between the molten metal and substrate, which can inhibit the wetting interaction. In fact, previous studies have shown that minor impurities as well as oxide skins on both the substrate and droplet surfaces can pin the droplet and prevent typical wetting interactions [31]. By leveraging the combined conditions of high vacuum and elevated temperature, the impurities can volatilize [32], leading to strong wetting predicted by the theoretical framework. It should be noted that a similar effect could be reached through using a reducing atmosphere, which could remove surface oxides and decrease the partial pressure of oxygen in the system to ensure pure contact between metal surfaces without impurities. Under such conditions, it would be expected that the Sn would wet the W substrate at much lower temperatures due to the lower impurity presence. Further investigation into the Sn-Mo and Sn-W interfaces, as shown in Figure 3c–h, via EDX elemental mapping reveals stable interfaces with no observable reactions or mixing between Sn and either the Mo or W solid substrates. The binary phase diagrams for both Sn-Mo and Sn-W systems indicate low mutual solubility of Sn with Mo and W. Thus, these findings underscore the ability to predict excellent, non-reactive wetting between elemental metals with low mutual solubility.

3.2. Wetting of Metals with Ceramics

While there is a clear benefit to selecting a nozzle material that has low reactivity and negligible mutual solubility with the molten jetting material, such as Mo for Cu or W for Sn, such combinations are not always practical to search for. For example, metals such as Al form intermetallic compounds readily. These reactions will tend to erode any wetting material during MMJ; thus, we cannot rely solely on metals as solid substrates when searching for wetting candidates. The wetting behavior of such reactive metals on ceramic substrates is a complex phenomenon influenced by the interactions of molten metal atoms with ionically or covalently bonded atoms on ceramic surfaces. Al-based alloys are especially important for many applications where strong lightweight materials are required, such as aerospace, automotive, naval, and the construction industry. In order to achieve widespread usage, it is imperative that MMJ systems be able to consistently eject Al-based alloys, which require strong wetting inside the orifice without deleterious reactions, and maintain a non-wetting coating around the outside of the orifice exit.

Reactive molten metals such as Al also tend to exhibit a wide range of interactions with solid ceramic substrates, leading to various outcomes in terms of wetting. For instance, as illustrated in Figure 4a, molten Al4008 demonstrates effective wetting on aluminum nitride (AlN), titanium dioxide (TiO₂), and silicon carbide (SiC) substrates at 1100 °C. Based on supplementary Video S3, Al4008 does not begin to wet the AlN, TiO₂, or SiC until after the temperature exceeds 1000 °C. In a previous example, it was seen that Sn did not wet Mo or W at lower temperatures due to the presence of impurities. However, while Sn wets metal substrates due to metallic bonding interactions, the mechanism by which Al4008 wets AlN, TiO₂, and SiC can vary depending on the interaction between the components of Al 4008 and the constituent elements of the ceramic substrates. The snapshot at 1100 °C is shown to illustrate the point at which maximum wetting is achieved for the substrates. In contrast, Figure 4b shows that Al4008 does not wet Al₂O₃ [33] or BN at 1000 °C.

Further analysis of the wetting behavior of AlN, SiC, and TiO₂ reveals significant differences in the interactions between Al4008 and the various ceramic substrates. Note that Al4008 contains about 7 wt.% Si, so regions of elemental Si are expected on the Al side of the interface in all conditions. The SEM EDX maps presented in Figure 4c–f illustrate that the interface between Al4008 and AlN remains stable, suggesting non-reactive wetting. Despite both Al4008 and AlN containing a significant amount of Al, a distinct contrast is observed at the interface. Notably, Si is localized on one side, while N is confined to the other, indicating that N does not diffuse into the Al4008. Although EDX cannot quantitatively measure N levels, this qualitative evidence supports the notion that the wetting of Al4008 on AlN is primarily due to non-reactive interactions. Conversely, the EDX-SEM maps in Figure 4g–j indicate that the interface between Al4008 and TiO₂ is reactive, with Al and Ti diffusing across the interface. This suggests that the formation of intermetallic compounds, which include Al and Ti, is responsible for the wetting of Al4008 rather than the direct interaction of Al4008 with the substrate. Additionally, Figure 4k–m shows a map of Al4008 melted onto SiC. Large (>100 µm) elemental Si crystals are seen in the Al alloy, as well as other large Al-Si-X crystals. Because Al4008 is a hypoeutectic Al-Si alloy, the typical phase distribution after solidification and cooling consists of eutectic colonies of FCC Al and elemental Si. Fine Mg₂Si can form at low temperature due to the ~0.3 wt.% Mg content, but that phase is not pertinent to this discussion. The presence of large elemental Si and Si-rich crystals in the droplets indicates that liquid Al absorbed significant Si from the SiC substrate. These phases could have only formed through isothermal flux growth and/or hypereutectic growth on cooling in a liquid with Si enriched far above the normal value for Al 4008. It appears that Si enrichment in the melt also enabled impurities in the melt to react and form Al-Si-X ternary phases. A lower magnification, SEM-EDX analysis is shown in Supplementary Information (Figure S1), which helps illustrate the different phases present in the structure. The substrate is unusually smooth for a surface that underwent significant dissolution; however, this surface is single-crystal SiC in the [0 0 0 1] orientation. As a single crystal, every point on the surface can be considered energetically equivalent, enabling uniform dissolution of SiC as a flat stable interface. The combination of Si crystal growth and of ternary precipitation suggests that the wetting of Al4008 onto SiC is driven by a combination of reactive and dissolutive, while the flat interface suggests that the single-crystal nature of the substrate encourages uniform migration of Si into the molten droplet.

The underlying mechanisms governing non-reactive wetting of molten metals on ceramic substrates are not as well understood as those for metal-to-metal wetting [29]. While metal-ceramic wetting sometimes requires a reaction product to facilitate wetting, this is not always the case. For example, in the report from Eustathopoulos [1], liquid metals can wet ceramics such as carbides, nitrides, or borides of transition metals without reacting because a significant part of the cohesion of these materials is provided by metallic-like bonds. Thus, the selection of the correct ceramic material can result in wetting, even if the primary cohesion of the ceramic is similar to that of the metal wetting to it. Previous studies, such as those by Cao et al., have explored the Al/AlN system using first-principles calculations [29]. Their density functional theory (DFT) simulations revealed that the lowest wetting angle occurs when the AlN surface is terminated with N atoms, and the Al(111) atoms are aligned directly above these N atoms. This specific configuration results in a distance of 1.92 Å between the Al atoms on the Al(111) surface and the N atoms on the AlN(0001) surface, closely matching the 1.90 Å distance between Al and N atoms found in the bulk AlN [29]. The electron density at the interface resulting from this configuration suggests strong interactions between the Al atoms and the N atoms, leading to stable bonding and a low wetting angle, even though no chemical reaction occurs. This example illustrates how atomic-level structure and electronic interactions can strongly influence macroscopic wetting behavior, even in systems that appear non-reactive. In addition to the macroscopic mechanisms of wetting, atomic- and nanoscale interactions at the liquid–solid interface play a significant role in determining the measured contact angles. Even when no visible reaction layer or bulk interfacial product forms, local charge redistribution, atomic rearrangement, and short-range bonding effects can modify the interfacial energy and promote wetting. These findings highlight that wetting cannot be understood solely in terms of macroscopic thermodynamic parameters; instead, it must also consider the atomic-scale structure and bonding that govern the true nature of the metal–ceramic interface.

While this explanation works well for the Al/AlN systems, there is some evidence in the literature that this concept can be generalized to other metal-ceramic systems [28]. One metric that can help describe the wettability between a liquid and a solid is known as the work of adhesion ($W_a$), which describes the change in the interfacial energy of a system when a solid surface is covered by a liquid. This parameter is given by:

$$-W_a={\sigma }_{SL}-\left({\sigma }_{SV}+{\sigma }_{LV}\right)$$

(3)

where ${\sigma }_{SL}$, ${\sigma }_{SV}$ and ${\sigma }_{LV}$ represent the surface energies of the solid–liquid, solid–vapor, and liquid–vapor interfaces. Larger values for the work of adhesion imply stronger wetting interactions between the solid and liquid. For example, Taranets and Naidich measure the wetting of many metals on a solid AlN substrate across a wide range of temperatures [28]. Through these experiments, they find that molten Al and molten Si wet strongly to AlN, while many other molten metals do not wet to AlN. They suggest that the wetting of molten metals to ceramic compounds depends on the free energy of formation of the substrate and the free energy of formation of compounds formed by the molten metals with the components of the substrate. The balance of these quantities defines the strength of the interaction between the molten metals and solid substrates, which they show is proportional to the work of adhesion and thus governs the degree of wetting. When calculating these quantities, Taranets and Naidich show that Al, which has a strong affinity to N in the AlN substrates, has a large work of adhesion, and that this affinity explains why molten Al would strongly wet to AlN. This conclusion is in agreement with the observations made in this work and the explanation given by Cao et al. [29].

Given that Al4008 is composed of ~92 wt.% Al, the non-reactive wetting properties observed in pure Al can be extrapolated to Al4008 as the elements in the alloy do not contribute to any reactions. Thus, the wetting of Al4008 on AlN is likely driven by strong interactions between the surface Al atoms in the molten alloy and the surface N atoms in AlN. This understanding of wetting interactions implies that ceramics with atoms that may have a strong surface interaction with the molten metal atoms encourage wetting without undergoing a chemical reaction. Furthermore, this helps explain why stable ceramics that tightly bind their constituent atoms so that their interactions with surface atoms of the molten metal are weakened and thus, lead to non-wetting interactions. This concept elucidates why hBN remains non-wetting to all the metals studied, as it is significantly more stable than any compounds that could form from reactions between boron or nitrogen and the metals in question, as supported by formation energy data [34,35]. This data is indicated in

Table 1. Comparison of the free energy of formation of nitrides at 1000 °C to illustrate the non-wetting nature of molten metals on hBN.

Compound System	Estimated Free Energy of Formation at or Around 1000 °C (kJ/mol)
Cu3N	Decomposes above 500 °C [37]
SnNx	Decomposes above 500 °C [38]
AlN	−343.1 [35]
hBN	−272.0 [35]
CeN	−376.6 [35]

, where Cu and Sn only form metastable nitrides that easily decompose; thus, they would have very weak interactions with hBN. It should be noted that while Table 1 suggests AlN should be more stable than hBN; thus, one would expect a reaction between Al and hBN to occur. However, at temperatures up to 1000 °C, previous literature has shown that the reaction kinetics are so slow as to not be measurable over long periods of time [36]. Thus, even though hBN may experience reactive wetting under very controlled conditions, the practical setting of MMJ allows us to use it as a non-wetting surface thanks to very slow reaction kinetics. This is a particularly useful finding as hBN is also highly machinable, with a high melting point, making it a great candidate for the crucible material when melting many metals. While hBN is a strong non-wetting candidate for the metals and alloys in this work, it may not be suitable for use with more reactive metals. For example, CeN has a much lower free energy of formation than hBN over a large temperature range, and Ce is a metal that is known for high reactivity. This would suggest that Ce would likely react with a hBN substrate under wetting conditions to form CeN, leading to the dissolution of the substrate. It is thus important to select a metal/ceramic combination such that the ceramic substrate is more stable than the possible compounds formed by any elements in the molten metal with any constituents in the ceramic substrate.

3.3. Selection of Wetting/Non-Wetting Combinations

The results of our wetting studies provide a foundation for selecting appropriate materials for wetting and non-wetting combinations of materials for various applications. Specifically, we can identify a suitable wetting material for a nozzle material used in Molten Metal Jetting (MMJ), which will be positioned at the nozzle exit, and a non-wetting coating for the outer surface. A requirement for this application is that the nozzle material must exhibit non-reactive wetting properties with the molten metal. This ensures the metal will remain wetted for long operation times.

Based on our results, we have developed a heuristic method for identifying compatible material candidates. When searching for a non-wetting candidate, it is essential to select a highly stable ceramic that exhibits weak or negligible interactions with the molten metal. This means the substrate compound must have a large and negative free energy of formation that would indicate high stability. If this free energy of formation is much greater than the free energy of formation of the molten metal with any of the substrate constituents, then the molten metal is likely to have a low work of adhesion to the substrate, which will prevent wetting. hBN serves this purpose effectively for the metals studied. Conversely, when seeking a material that the target molten metal will wet to, our first approach is to identify a pure metal with a significant miscibility gap relative to any components within the molten alloy, such as molybdenum (Mo) for molten copper (Cu) and tin (Sn). In this case, the metallic bonding at the interface is enough to guarantee strong wetting without any chemical reactions taking place. Thus, a stable wetting interface can be established that will not corrode the solid substrate and will not introduce impurities into the liquid. If a suitable pure metal is not available—due to the tendency of the molten metal to form intermetallic compounds readily (i.e., Al)—we then look for a ceramic material that can interact strongly with the molten metal atoms while remaining stable and unlikely to react, such as AlN for aluminum-based alloys. This can first be selected by finding ceramic substrates that contain components that will interact strongly with the molten. As long as the free energy of formation of the ceramic substrate is slightly larger than the free energy of formation of the molten metal with any of the substrate components, we can ensure that the work of adhesion between the molten and the substrate will remain strong enough to encourage wetting. Additionally, the substrate will remain stable enough that deleterious reactions may be suppressed. It should be noted that if the energy of formation between one or more of the substrate components and the molten metal is larger than the energy of formation of the solid substrate, reactive wetting may occur. In cases where no ideal non-reactive materials can be found, materials with very slow reaction kinetics can be selected, assuming the operation time for the application is within reasonable limits. We may also consider coating the substrates with a material that could react with the molten metal to form a passivating layer, thereby promoting wetting without compromising the integrity of the underlying material.

This systematic approach not only enhances our understanding of material compatibility in MMJ applications but also provides a framework for future material selection in similar contexts.

4. Conclusions

In this work, we have established a systematic method for rapidly screening materials to identify suitable wetting and non-wetting combinations for various molten-metal/solid–substrate interactions. This technique also allows cheaper equipment to be utilized to rapidly screen material combinations to form the basis for a materials compatibility library that will be useful across many application spaces that rely heavily on the wetting nature of metals on various substrates. In particular, this method is faster and cheaper than traditional wetting measurement systems that require long measurement times, extensive care, and expensive equipment to control surface roughness, atmospheric environment, and material purity for precise measurements of equilibrium wetting angles. Our findings reveal that Mo serves as an effective wetting substrate for molten Cu and that both Mo and W are compatible as wetting materials for molten Sn. The use of EDX elemental mapping has confirmed that these wetting interactions are non-reactive, aligning with established theories of metal-metal wetting characterized by low mutual solubility.

However, the situation is different for reactive metals such as Al, which has a propensity to react and form intermetallic compounds with other metals. In our study, we identified AlN, TiO₂, and SiC as effective wetting materials for molten Al4008. EDX mapping further elucidated that TiO₂ exhibits a reactive wetting interaction with Al4008, while AlN demonstrates a non-wetting interaction. This non-wetting behavior between AlN and Al4008 can be attributed to the strong interactions between the surface Al atoms in the molten alloy and the surface N atoms in AlN.

Across all metals tested, hexagonal boron nitride (hBN) consistently displayed non-wetting characteristics. This behavior is attributed to the exceptional stability of hBN, which effectively prevents significant surface interactions between hBN and the molten metals. The insights gained from this study contribute to the development of heuristic guidelines for identifying compatible wetting and non-wetting candidates in future applications.