Nanometric 3D Printing of Functional Materials by Atomic Layer Deposition

Abstract

Atomic layer deposition (ALD) is a chemical vapour deposition (CVD) method that allows the layer-by-layer growth of functional materials by exposing a surface to different precursors in an alternative fashion. Thus, thanks to gas-solid reactions that are substrate-limited and self-terminating, precise control over thickness below the nanometer level can be achieved. While ALD was originally developed to deposit uniform coatings over large areas and on high-aspect-ratio features, in recent years the possibility to perform ALD in a selective fashion has gained much attention, in what is known as area-selective deposition (ASD). ASD is indeed a novel 3D printing approach allowing the deposition of functional materials (for example metals to oxides, nitrides or sulfides) with nanometric resolution in Z. The chapter will present an introduction to ALD, which will be followed by the description of the different approaches currently being developed for the ASD of functional materials (including initial approaches such as surface pre-patterning or activation, and newer concepts based on spatial CVD/ALD). The chapter will also include a brief overview of recent works involving the use of ALD to tune the properties of 3D printed parts.

1. Introduction

Vapor-phase techniques are powerful approaches for the deposition of functional thin films of different materials, including metals and compounds such as oxides, nitrides, and even organic materials and composites onto a substrate [1, 2, 3]. There are two types of vapor deposition methods, namely physical and chemical. Physical vapor deposition (PVD) methods involve a change of state (i.e., evaporation and recondensation) of a source, and include, among other, sputtering, pulsed laser deposition or different evaporation approaches [4]. While PVD methods yield materials of high quality with tunable properties, they are performed in high vacuum and often high temperatures, using sophisticated equipment. Finally, the low vacuum process results in a line-of-sight type of coating (i.e., only taking place on the directly exposed surfaces.

The possibility to pattern and 3D print materials at different scales has a tremendous impact on many technologies and applications. Over the years. Different 3D printing approaches have been developed allowing such patterning. This include, to name a few, aerosol jet printing (also known as Maskless Mesoscale Materials Deposition or M₃D) [5], ink jet and screen printing [6], laser chemical vapor deposition (LCVD) [7], laser-induced forward transfer (LIFT) [8, 9] or micro stereo lithography and multiphoton lithography [10]. The interested readers are encouraged to the cited references for more details on these methods. In this chapter, we focus on 3D printing approaches based on ALD. A brief introduction to CVD and ALD is thus presented next.

Chemical vapor deposition (CVD) approaches on the other hand rely on chemical reactions between different precursors on and over a surface. In conventional CVD, the precursors are injected in the reactor at the same time and the reaction is activated by heat (hot substrate) or by other energy sources, such as plasma. A scheme representing the reaction chamber is shown in Figure 1a [11]. This technique allows the deposition of high-quality films [12], and is largely used by the industry. Nonetheless, CVD is governed by the diffusion of the different gas precursors, and therefore, the deposition of extremely thin films with a thickness control at the sub-nanometer level [13], and the uniform coating of large areas or high-aspect-ratio/porous features is extremely difficult [14].

Figure 1.

(a) CVD mechanism where the precursor is adsorbed on the surface at relatively high-temperature followed by the film growth and a release of volatile byproducts, (b) ALD process: Schematic of one ALD cycle of monolayer growth. The first step consists in exposing the substrate to the precursor followed by a purge step to remove all the byproducts an excess precursor, then another step with a co-reactant agent and the final step in which the byproducts an excess precursors are purged again, (c) illustration of edge coverage for ALD, CVD, and PVD.

Such limitations prompted the development of an alternative method, namely, atomic layer deposition (ALD). ALD is indeed a CVD method but it is characterized by having the substrate exposed to the different precursors one at a time, and not simultaneously as in CVD. Thus, in typical ALD processes, a precursor is first injected in a deposition chamber where the substrate is located. The precursor can then react with active sites on the surface (i.e., undergoing a chemisorption) until the latter is saturated. A purge step is then applied to eliminate excess precursor and reaction byproducts. Then a second reactant is injected that will react with the preciously adsorbed layer. After the reaction is completed, again a purge step is necessary to eliminate excess reactant and reaction byproducts. Such an ALD cycle is shown in Figure 1b. As a result of this sequential exposure to the different reactants, the ALD process is surface-selective and self-terminating, which in turn offers unique control over film thickness at the angstrom level (i.e., a given growth per cycle, GPC, being obtained for each process as a function of the reactor geometry and precursors used) and allows the conformal coating of porous, complex and high-aspect-ratio substrates. The films are also compact and free of pinholes and can be obtained at low temperatures (even room temperature) due to the high reactivity of ALD precursors. The reader is referred to reviews and books dedicated to ALD for more information [1, 2, 3]. Figure 1c shows a sketch of the different types of coating obtained over the high-aspect-ratio features when using the different techniques discussed.

Over the years, the number of materials that can be deposited by ALD has grown enormously, including pure elements (e.g., metals), nitrides, sulfides, oxides, fluorides, etc. (see the atomic limits site, with includes an ALD materials database that is permanently being updated [15]). While at the origin the main motors of the ALD development were the deposition of homogeneous coatings over large areas or high aspect-ratio features, in the last years, there have been innovative developments in the ALD field that allow the localized and topological deposition of functional materials. This opens the door to its application as a new nano-to-macro 3D printing technology based on gas precursors. These recent developments, namely, area-selective deposition (ASD) and different spatial approaches, are presented in Sections 2 and 3, respectively. Finally, the unique assets of the ALD technique are ideal to tune the properties of pieces fabricated by conventional 3D printing approaches. Section 4 presents a brief overview of recent results on this line. The chapter finishes with some conclusive remarks.

2. Area-selective deposition (ASD)

For more than 50 years, the shrinking of microelectronic devices has involved successive steps of deposition, lithography and etching. Indeed, unlike building a house, it is not possible to directly draw the walls or pillars of a chip on a 300 mm substrate. It is therefore necessary to first cover the whole substrate with a thin layer, before removing part of it by the steps of lithography (to draw the object) and etching (to remove what should not remain on the surface). This is called a top-down approach. The reduction of the dimensions of microelectronic devices in the last 10 years to nanometric scales has greatly complicated these steps and increased their cost. Indeed, for many years, the wavelengths used to draw were greater than the desired line thickness. It was therefore necessary to make lithography more complex by integrating etching/deposition steps to achieve the desired dimensions, such as multi-patterning (Self Aligned Double or Quadrupole Patterning—SADP and SAQP). Thus, these steps allowing to obtain locally nanometric materials on the substrate are now complex, time-consuming and expensive. They must also be done with nanometric placement precision, which is already a real challenge.

The alternative solution to this increasingly complex approach is to deposit the material directly and selectively on the desired surface without having to resort to lithography steps. This so-called selective growth on a surface is a bottom-up approach and is known as area-selective deposition (ASD) [16, 17]. In an ideal ASD process, a thin film should be uniformly deposited in the desired growth region while no deposition should be observed in the desired no-growth region. This requires the use of a surface selective deposition process, with controlled growth at the atomic scale, and thus ALD is the one that seems to be the most adapted. Indeed, a growing number of researchers working on the ALD process are now trying to establish strategies from this process to have a material deposited selectively on a surface. The three main strategies are: (i) to use an inherent selectivity of the precursor/substrate couple [18, 19]; (ii) to block the growth on the no-growth area by a pre-deposition treatment [20, 21]; (iii) to promote the growth on the growth area by a pre-deposition treatment [22]. Whatever the strategy, we observe growth on all surfaces after a certain number of cycles, or at best, a little defectivity with nuclei on the no-growth area, i.e., the selectivity fades out during the successive ALD cycles. It was then proposed to regularly add the surface treatment step (passivation step) in the ALD cycles changing a cycle from a (treatment + AB) process to an (ABC) cycle with the treatment reinjected regularly [23]. Another proposed solution is to use super-cycles with the injection of etching steps every xALD cycles, this is called ASD by super-cycles of deposition-etching [24, 25, 26, 27, 28]. In the end, with these different ASD strategies, passivation or super-cycles, [29] the possibility to deposit thin films of more than 10 nm on a previously selected surface has been successfully demonstrated. However, the additional steps (passivation, etching) increase the processing time. It takes more than 24 h for passivation by a self-assembled monolayer (SAM) using chemical baths. The insertion of etching steps in an ALD process also increases the processing time by a factor of 2–3 compared to a conventional ALD process. Finally, the insertion of passivation or etching chemistries can induce contamination of the deposited layer as well as process drifts (with for example a modification of the chemistry of the ALD reactor walls during the etching step). Nevertheless, ASD has the potential to deposit functional materials with complex 3D shapes and nanometric resolution, well beyond the possibilities offered by standard 3D printing methods. This can be done by different approaches, as shown in Figure 2.

Figure 2.

(a) Illustration of four different strategies for an area selective deposition where a is the growth area and B the no-growth area; (b–d) examples of ASD using deposition and etch: (b) TiO₂ on TiN vs. Si/SiO₂ (reprinted with permission) [25], (c) patterning at the microscale of TiO₂ and W films (reprinted with permission) [30], and (d) topographical selective deposition of Ta₂O₅ on Si (reprinted with permission) [31].

Although, ASD offers a huge potential for the 3D printing of functional materials at resolutions orders of magnitude below what can be achieved with conventional 3D printing approaches, the different steps it implies (i.e., surface pre-pattering, regeneration of the selectivity) make them harder to work with. It would thus be desirable to develop an ALD approach that could allow the direct deposition of patterned materials. This can indeed be achieved, as detailed in the next section, by using different spatialapproaches. The question that arises now is whether this approach can be used to develop ASD processes on larger surfaces and at dimensions that are no longer nanometric but micrometric or millimetric. The natural answer is yes. Indeed, an optimized ALD process is uniform across m-scale objects. Using a spatial ALD technology, we can then imagine selectively printing on a large surface with ALD: 3D printing. This is described in the next section.

3. 3D printing of functional materials based on spatial CVD/ALD approaches

As explained in the introduction, the unique assets of ALD are the result of having a surface-limited, self-terminated reaction between gas reactants and the surface of a substrate. To limit the reaction to the surface, the ALD is based in alternate exposures of the precursors to the substrate. Traditionally, this has been done by sequential injection of the precursors in a deposition chamber followed by purging steps, thus in a temporal approach, as detailed in Figure 1b above and in the scheme below (Figure 3a). An alternative approach consists in having a continuous injection of the different reactants but in different locations of the reactor, keeping them separated by a region of inert gas. Then, by alternatively exposing the substrate to the different regions, the ALD cycle is reproduced (Figure 3b). This approach is known as Spatial ALD (SALD) [33, 34, 35, 36]. The first advantage of processing in the spatial mode is that the process can become much faster (up to two orders of magnitude) since no purging step is required.

Figure 3.

(a) Schematic representation of the classical temporal ALD approach with the different characteristic steps of an ALD cycle: (1) injection of the first precursor, (2) purging step, (3) injection of the second precursor, (4) purging step, separated in time. (b) Schematic representation of the spatial ALD approach, where the precursors are injected continuously in the reaction chamber in different locations separated by an inert gas and the sample is exposed to the different regions to reproduce the ALD cycle. (c) Scheme of the close-proximity AP-SALD approach based on a manifold injection head: the precursors are carried out from the containers of the head where they are distributed in parallel alternative channels. (d) COMSOL simulation of the mass fraction of each precursor present in different areas of the substrate (left). In these cases the evacuation of the precursors is not efficient and thus cross-talk is observed, yielding a CVD reaction on the zones where the precursors meet (see COMSOL simulation in the center). If a deposition is made in static mode (i.e., without moving the substrate), 4 lines of oxide can be obtained, as shown in the optical image (right) where 4 lines of ZnO have been deposited on a Si wafer in this way (adapted from Ref. [32]).

The SALD concept is very versatile and can indeed be applied in different ways [33, 37]. SALD can even be performed at atmospheric pressure (i.e., no vacuum processing) and even in the open air (i.e., no deposition chamber), and this is sometimes referred to as Atmospheric-Pressure SALD (AP-SALD). This is the case of the close-proximity approach based on a manifold injection head, originally presented by Kodak [38]. In this particular approach, the different reactants are carried to the injection head where they are distributed along alternate parallel channels (Figure 3c) [39]. By proper design of the head, the different flows can be kept separated provided the substrate is at close proximity of the head (i.e., 50–200 μm). Then by scanning the substrate back and forth under the head the ALD cycles are achieved. It is worth noting that since the size and area of the deposition depend on the head size and substrate scan distance, this SALD approach can already be seen as an ASD approach at the cm scale.

Close-proximity SALD approaches based on injection heads have several extra appealing advantages. The first one is that deposition can be also performed in spatial CVD (SCVD) mode. Then, crosstalk between the different reactants above the surface of the substrate is allowed. In this case, the deposition rate can be faster, but care must be taken since the properties of the materials deposited could change [32]. The impact of the change in the film properties when passing from the SALD to SCVD mode has to be evaluated depending on the intended application, but several works have demonstrated that the SCVD can be used to deposit components for functional devices [40]. In addition, the possibility of having SCVD opens the door to a new ASD approach. Indeed, the CVD reaction can be located in different areas above the substrate. Figure 3d presents a computational fluid dynamics (CFD) simulation that shows the areas over the substrate where the different reactants meet and thus react when the deposition is performed in certain SCVD conditions. Then, by performing a static deposition (i.e., without the substrate scan that is needed to perform the spatial ALD cycles) growth of the films can be localized to the regions where the reactants meet (see the four ZnO lines obtained by this approach in Figure 3d). This constitutes a new alternative approach of ASD at a higher scale and much faster deposition rate than the traditional ASD approaches based on ALD that have been described in the previous section [32].

The second advantage of using a close-proximity SALD approach based on an injection head is that the system can be customized by simply modifying the injection head. While this is so, the modification and fabrication of the head can result very difficult, if not impossible, thus limiting the potential of the approach (see Figure 4a where the scheme of a standard SALD head is shown. It comprises several parts that need to be fabricated separately and then soldered, and the distribution of the different gas flows to the head is quite complex involving many pipes). To overcome this limitation, D. Muñoz-Rojas’ group at the Laboratoire des Matériaux et du Génie Physique (LMGP, Grenoble, France) has introduced the utilization of 3D printing for the fabrication of customized SALD injection heads [41]. This allows having more freedom to design the head and, for example, the gas distribution can be incorporated in the body of the head (Figure 4b and c) [41, 42, 43]. Plastic heads can be printed for depositions taking place at low temperatures while metal 3D printing is also possible for higher temperatures [44]. Thanks to 3D printing, the design of the heads can be easily customized. This is very convenient to easily modify the area of deposition, and also to have free-form patterns when performing SCVD with custom heads (Figure 4d) [41].

Figure 4.

(a) Scheme of a close-proximity SALD head made of several parts and fabricated by conventional approaches. (b) 3D scheme of a head design integrating the gas distribution for the different gases inside its body: metallic precursor in green, co-reactant in red, inert gas in blue and exhaust in black. (c) Head printed with clear resin where the distribution channels can be observed. (d) 3D scheme of a head designed for circular shape deposition in static SCVD mode. ZnO circles with different thicknesses are shown. (e) Picture of a printed SALD pen (left), bottom view of the concentric gas outlets in the SALD pen approach allowing deposition in any direction (right). (f) Scheme of the SALD pen installed in a 3D table. (g) Scheme of a SALD pen implemented in the XYZ table and drawing ZnO in a circular pattern. (h) LMGP initials on a Si wafer drawn with the 3D printed SALD pen (adapted with permission from Ref. [41]).

The possibility to deposit free-form patterns without having to modify the head for each design would also be appealing. This can indeed be done if instead of using parallel channels, the head is designed so that concentric channels are used. In this way, no matter which direction the head moves, the substrate will be exposed to the different reactants, thus leading to ALD film growth (Figure 4e). Such a head can again be readily implemented by 3D printing. D. Muñoz-Rojas’ group demonstrated that such a SALD pen can be printed and used to deposit free-form patterns when installed in an XYZ table, in this case with a resolution going down to several mm (Figure 4f–h) [41]. This represents a new 3D printing approach that is based on gas precursors and that offers nanometric resolution in Z. Here again, the resolution of the obtained patterns in X-Y depends on the head design and the possibility to scale it down. Indeed, the latter work by Midani et al. presented a similar concept in which sub-millimiter resolution was achieved by inserting a capillary in the central metal precursor channel of the SALD pen [45].

Certainly, the advances in the different 3D printing technologies will allow de fabrication of SALD heads with smaller channels, which will extend the possibilities of SALD for depositing patterns of functional materials down to the micrometer scale in X-Y.

4. ALD for conventional additive manufacturing: recent works involving the use of ALD to tune the surface and properties of 3D printed parts

Additive manufacturing (AM), also known as 3D printing, is recognized as a revolutionary technology, which has primarily been used in the field of engineering to create customized prototypes [46, 47, 48]. 3D printing has now become a subject of great interest and is extensively applied in many areas, such as prototyping, medicine [49] or aerospace [50], since it allows new products with complex geometries and microarchitecture (multiple pore shape and size) to be imagined, designed and fabricated. However, the material from which the designed products are made is still limited by the 3D-printing material itself. Even if the number of available materials that can be printed is expanding [48], most of the manufactured objects are made of polymer or stainless steel. Thus, a post-treatment may be required to control the nature and chemistry of the product surface and offer it its desired functionality. As illustrated in Figure 5 and discussed below, ALD is a highly appealing technique to expand the potential of 3D printing through coating or infiltration of the printed parts.

Figure 5.

Illustration of the combination of 3D printing and atomic layer processing. Depending on the 3D printed material, either a coating is obtained, allowing for the tuning of the surface properties (typical ALD); or inorganic components are introduced to the subsurface of the 3D objects (atomic layer infiltration, ALI [51,52]), permitting the tuning of the matrix.

As seen previously, ALD can be used as an innovative and novel 3D printing route, to prepare customized and complex 3D structures at the nano-to-cm scale. In addition, this technology can also be used to precisely tune the surfaces of 3D printed objects that were manufactured using more “conventional” additive manufacturing approaches such as fused deposition modeling, inkjet printing, stereolithography, selective laser sintering (SLS), powder bed fusion or even bioprinting [46, 47, 48]. ALD allows the preparation of thin films with a sub-nanometer thickness control, high uniformity and excellent conformality even on high aspect ratios substrates, a unique capability, as discussed in the first section of this chapter [3, 53, 54, 55, 56, 57, 58]. As ALD allows the conformal coating of complex substrates with nanolayers made of an expanding number of materials [2, 15], such as oxides [59, 60], metals [61], nitrides [62] and sulfides [63], the combination of this route with 3D printing can be beneficial to a myriad of applications.

A large number of 3D printed objects are made of polymers, the current mainstream materials being ABS (acrylonitrile butadiene styrene) and PLA (polylactic acid). However, when performing ALD on 3D printed objects based on such polymeric materials, some considerations must be taken. The first obvious consideration is related to the ALD process temperature, which has to be lower than the polymer melting point. 3D printing materials such as ABS and PLA will already be deformed when the temperature is higher than 200°C. The ALD processes must therefore be compatible with rather low temperatures. Furthermore, as depicted by the review of Parsons and co-workers [64], the ALD precursors often infiltrate and react with polymeric substrates, which can alter the eventual 3D printed products. As shown by numerous studies, the risk of reaction between the polymer and the precursors increases with temperature and long exposures times. In addition, the presence of functional groups in the polymer chains also increases the potential infiltration of the ALD precursors [51, 52, 59, 64, 65, 66, 67, 68]. Thus, even if most of the ALD processes are compatible with the coating of 3D printed objects, these considerations must be taken into account and the processes have to be tuned accordingly to coat certain 3D printed materials. When the processes developed are compatible, the unique capability of ALD to coat complex objects with such control over the layer deposited, makes this route particularly relevant and attractive. This innovative combinatorial approach has been used for different and various applications, such as aerospace, photoelectrocatalysis, filtration, biomedicine, or solid-state batteries.

Kestila et al. combined polymeric additive manufacturing and an ALD-coating to produce satellite propulsion components with improved structural integrity and thermal resistance [69]. The components were made of two different polymers, namely acrylonitrile butadiene styrene (ABS) and polyamide, and were coated with alumina by ALD. The Al₂O₃ layer allowed to enhance the structural integrity for the polymeric restrictors and progressively smoothed out the PA surface improving the argon flow through the restrictor, which might be due to increased surface smoothness [69]. Heikinen et al. have recently shown that ALD of alumina on porous 3D printed ABS plastics permits to considerably lower their vacuum degassing. Nyman et al. have also confirmed the low outgassing of ABS, but also polyether ether ketone (PEEK), polycarbonate (PC), and nanodiamond-doped polylactide (ND-PLA) 3D printed materials with an Al₂O₃ ALD coating [70]. Thus, the combination of plastic 3D printing with ALD opens prospects for the fabrication of laboratory vacuum tools, and is also suited for spacecraft tools and in-space manufacturing applications [70, 71]. Moll et al. also coupled powder bed additive manufacturing with CVD and ALD of nitrides, to prepare 3D Ti-6Al-4 V structures highly resistant to high-temperature oxidizing environments. Coupling CVD and ALD on the 3D printed objects permitted to obtain thick coating and roughness reduction by CVD, and filling of narrow defects and reactivity mitigation by ALD [72].

Browne et al. employed additive manufacturing and ALD for photoelectrocatalysis, by depositing TiO₂ onto 3D-printed electrodes. These electrodes were initially printed in inert stainless steel, and gained their catalytic functionality thanks to the ALD coating. The conformality allowed by ALD successfully permitted these 3D-printed electrodes to be used as photoanodes for water oxidation. The results presented have shown that the 3D-printed stainless steel electrode coated with ALD of TiO₂ were considerably more active towards the water oxidation, and that the catalytic activity was enhanced by increasing the number of ALD cycles applied [73]. The team of Pumera et al. recently applied ALD to 3D-printed nanocarbon/polylactic acid electrodes to coat them with metal dichalcogenide MoS₂ nanolayers [74]. The MoS₂ coated electrodes were then successfully applied for photoelectrocatalytic hydrogen evolution reaction (HER). Recently the group of M. Bechelany from the Institut Européen des Membranes (IEM, Montpellier, France) has developed in collaboration with the University of Zaragoza the functionalization of 3D printed ABS filters with MOF (Metal-Organic Framework) for toxic gas removal [75]. The fabrication approach at low temperature includes ALD of Zinc oxide on the ABS 3D printed filter followed by the hydrothermal conversion of ZnO to ZIF-8, Zeolitic Imidazolate Framework. The obtained filters show a good adsorption performance for dimethyl methylphosphonate, thus demonstrating their potential for toxic gas capture applications. Such types of 3D printed filters with an active MOF layer could have a wide range of applications in environmental fields such as adsorption systems for removing toxic gases or water pollutants.

In the biomedical field, the combinatorial approach has been applied to prepare silver-coated titanium orthopedic implants. [76] Using the selective laser melting (SLM) 3D-printing technique, titanium orthopedic implants have been fabricated with intricate geometries. The surface chemistry of the prepared implants has then been modified by coating them with a silver nanolayer by ALD. The inhibition of bacterial colonization obtained thanks to the silver coating resulted in the drastic reduction of the pathogenic biofilm. This result, combined with the increase of the vascularization and the osseointegration observed, opens a new path to this combinatorial approach for clinical orthopedic applications [76]. The “pure marriage” between 3D printing and ALD has also been exploited by Xue et al., who tailored the surface of 3D printed plastic earplugs using plasma-assisted ALD [75]. By combining 3D printing, plasma-assisted ALD and hydrothermal process, they loaded a layer of ZnO nanoarrays on the surface of the earplugs and thus improved the antibacterial properties of the earplugs, which enhanced the safety of the ear devices. In addition, they have shown that the sound insulation performances were higher than those of traditional earplugs. Finally, the field of solid-state batteries benefited as well from the combination of 3D printing and ALD. For example, thanks to an innovative 3D-printing ink formulation, a cell-based on a 3D-printed stacked array of LLZ (Li₇La₃Zr₂O₁₂, a solid lithium conductor) and lithium electrodes was fabricated, and ALD of alumina has been performed at the surface of the LLZ to allow the wetting of lithium [77, 78]. The ability to 3D-print solid electrolytes enables the manufacturing of unique ordered structures, and ALD permits their efficient functionalization, improving the overall efficiency of the battery device.

These few selected studies demonstrate the great potential of combining additive manufacturing and ALD. The combinatorial approach allows the fast prototyping of functional products with the additional precise control over their surface chemistry. As depicted in the presented examples, the benefits of combining 3D printing and ALD nanocoatings can be applied to many complex surfaces, and the lack of materials that can be 3D printed is at least partially solved by the use of ALD coatings. Thus, this novel approach allows synthesizing precisely integrated and customized architectures with tailored surface performance, and/or eventually the bulk properties of the materials thanks to ALI, paving the way towards innovative and functional products, and opening prospects for many potential applications.

5. Conclusion

Although ALD was initially developed to exploit the possibility it offered to obtain continuous, pin-hole-free thin films even over large areas, in the last years there have been different approaches to perform ALD in a localized fashion, giving rise to the ASD field. As it has been discussed, these methods are based on different approaches allowing either a high spatial resolution in XY (at the nanometer) or simpler more direct approaches that provide direct patterning at the millimeter and micrometer level in XY. In any case, and given these approaches are based on the ALD method, the control in Z is nanometric. The possibility to have spatial control over the ALD process can be exploited as a new gas-based technique for the 3D printing of functional materials at different scales, providing a unique approach to the fabrication of functional materials with complex shapes. Beyond using ALD as a 3D printing technique in itself, the possibility it offers to coat (even infiltrate) complex shapes in a highly controlled way and with a large amount of different materials is ideal to nanoengineer the properties of pieces obtained by standard 3D printing approaches, thus expanding the range of applications that can be achieved. ALD should thus experience an important penetration in the 3D printing field in the coming years.

Acknowledgments

D.M.-R. acknowledges support from the European Union’s Horizon 2020 FETOPEN-1-2016-2017 research and innovation program under Grant Agreement 801464, and through the Marie Curie Actions (FP7/2007-2013, Grant Agreement No. 63111). The Agence Nationale de la Recherche (ANR, France) is also acknowledged for funding via the programs ANR-16-CE05-0021 (DESPATCH) and ANR-20-CE09-0008 (ALD4MEM). The French National Research Agency (in the framework of the “Investissements d’avenir” program (No. ANR-15-IDEX-02) through the project Eco-SESA) is acknowledged for a PhD Grant.