Ethan B. Secor*,
Daniel Yeboah
and
Livio Gamba
Iowa State University, Ames, IA, USA 50011. E-mail: esecor@iastate.edu
First published on 4th August 2025
Droplet jetting technologies offer a versatile, digital platform to fabricate functional devices from nanomaterial building blocks. Inkjet, aerosol jet, and electrohydrodynamic jet printing constitute three distinct technologies for precise patterning of functional materials in an additive, digital, and noncontact manner. While the unique physical mechanism of each technology endows it with specific advantages and disadvantages, commonalities in materials compatibility, patterning capabilities, and application domains motivate a holistic assessment of nanomaterial integration with these methods. This report will highlight progress across ink formulation, process design, and application development from recent years, with an emphasis on emerging materials and practical applications in this evolving field of research. This includes an overview of the three printing technologies, a survey of ink formulation and printing efforts across conductive, insulating, and semiconducting materials, an examination of compelling application demonstrations in electronics, sensing, and energy, as well as discussion of key emerging themes related to artificial intelligence, multimaterial printing, and nonplanar patterning.
![]() | ||
Fig. 1 Overview of droplet-based printing technologies for functional devices. Inkjet, aerosol jet, and EHD jet printing are the focus of this report, given their high-level similarities in materials requirements and process integration characteristics. Graphics adapted with permission from ref. 14 and 15 (CC BY), ref. 16 and 179 (Copyright 2009 American Chemical Society). |
While arguably not the most glamorous application, a broad and impactful need for nanomaterial inks is for patterning of conductive wires for basic circuitry. Comparable to wiring of printed circuit boards, power and signal connections that can be adapted to flexible, conformal, or large-area surfaces have widespread application. For multilayer circuitry, this requires insulating layers to form signal crossovers, and functionality of the conductor encompasses electrical conductivity, print resolution, and print aspect ratio. This drives a large, established market for conductive inks.12 While this spans high throughput patterning methods (i.e., screen, gravure, flexographic printing), digital droplet-based methods offer excellent precision, improved research and prototyping characteristics, and unmatched agility. For single-layer printing on a flat surface, the case for droplet-based methods is more challenging to make outside of prototyping; for multi-layer printing that requires registration to underlying features, the ability to adapt on the fly is a key strength. Moreover, extension to conformal patterning on 3D surfaces enables an application space with limited competition,13–16 particularly for broad materials.17,18
A second broad application area is encompassed by hybrid electronics – the combination of conventional microelectronics technologies with printed components.19,20 At one end of this spectrum, this can describe electrical connections to discrete passive or active devices, such as resistors and capacitors, paralleling the functionality of printed circuit boards but with some capability to generalize. In higher precision embodiments, this is exemplified by interconnects to bare die, such as replacing wire bonds,21 along with emerging interest for heterogeneous integration and advanced packaging.22,23
For the previously mentioned application areas, the basic requirement to print metal and dielectric structures with high geometric precision and material quality is the core foundational technology. As more diverse functionality is added, the application potential for these methods broadens significantly. Two areas in particular include active electronics for logic, with the core unit being the thin film transistor,24 and sensing, which can describe a wide range of transduction mechanisms and target analytes.25–27 While printed transistors will not be competitive with conventional microelectronics in terms of performance, reliability, integration density, and power, they can provide a compelling fit for applications requiring low integration density – i.e., a relatively small number of transistors distributed over a large area. Backplanes for display technologies are a prime example,28,29 and more generally such functionality could support some level of signal conditioning or edge computing.30 Sensing, encompassing mechanical, chemical, electromagnetic, and other stimuli, is a diverse application space with significant opportunities for printing technologies. More than any other application discussed here, this benefits from the functional diversity of printable materials and often fits the low volume, high mix production aligned with digital manufacturing.
These benchmark applications shape the development of nanomaterial inks. Conductive inks are a workhorse for the broad scope of electronics applications, with silver currently the most prevalent material, but needs for alternative metal and non-metal inks for specific features.30–32 Dielectric inks compose a second broad class defined by electronic functionality, and while crucial for practical applications, these have not benefited from the same level of research and development in recent years.33 The last major class comprises semiconducting inks,34 which augment the diversity of materials and applications by enabling active logic, optoelectronics, expanded sensing functionality, and other benefits.
![]() | ||
Fig. 2 Key droplet-based printing technologies and their distinguishing benefits. Graphics adapted with permission from ref. 39 (CC BY), ref. 70 and 176 (CC BY), and ref. 280 (CC BY-NC-ND). |
Alongside the bottom-up requirements based on process science, these three jetting technologies occupy a similar space in terms of attributes and applications. In short, these methods are digital, non-contact, and high resolution. Digital: all three are digital patterning methods, meaning the target pattern to fabricate is embedded in a digital format and executed by a computer-controlled printing system. This is of course advantageous for research and prototyping environments, for which continuous modification of printed patterns supports rapid refinement of designs and process parameters. This also facilitates operation in less conventional modes, such as additive repair,35 that must be responsive to the actual geometry of the substrate or pre-existing patterns. In the same way, digital methods can serve an important role for multi-layer patterning.36 Rather than propagating and amplifying errors in lower-level patterns, these digital techniques could detect and accommodate such errors to provide a more robust and resilient manufacturing process. Non-contact: by depositing small droplets onto a surface from some meaningful standoff distance, these three jetting methods allow patterning within new constraints. Printing on sensitive surfaces becomes feasible, such as biological or deformable substrates. In addition, these methods allow more straightforward printing on complex 3D surfaces compared to most contact-based methods, with the standoff distance providing some tolerance for accurately contouring the 3D surface.37 High resolution: print resolution is a primary metric that defines the suitable application space of a given technology. While there are differences across these three methods, they all span ∼100 nm–100 μm size range for the ‘unit’ process, i.e., a single printed droplet or line. This is a promising size range for relevance in electronics packaging and board-level circuitry.
These similar attributes lead to broad similarities in materials development, allowing a cohesive overview. While there are methods that serve similar applications – i.e., transfer-based printing methods, extrusion printing, and screen printing – these do not share enough similarities in fundamental physics or technological attributes to justify inclusion here. One common method that is omitted, in particular, is direct ink writing or extrusion printing.38 While many qualitative applications are similar, the wide range of nozzle size for DIW allows micron-scale particles, while the close proximity of the print nozzle and surface put it in an ambiguous realm regarding contact vs. non-contact deposition.
The three methods are distinguished by their droplet generation and deposition mechanisms. While this imposes differences on how inks are tuned for the individual methods, and what features, advantages, and disadvantages characterize each, broadly speaking the generation and deposition of fine droplets to pattern materials leads to commonalities in materials compatibility, downstream processing, and applications. For each of the three printing technologies discussed, we will introduce the method and provide broader context, and then discuss the mechanism of the process, typical patterning metrics, benefits and drawbacks, and several illustrative use cases.
Inkjet | Aerosol Jet | EHD Jet | |
---|---|---|---|
a Approx. 0.5–10 mPa s suitable for ultrasonic atomization, with higher range possible using pneumatic atomization for droplet generation. | |||
Resolution | 20–100 μm | 10–100 μm | 0.1–100 μm |
Viscosity | 1–20 mPa s | 0.5–1000a mPa s | 0.5–10000 mPa s |
Dominant fluid characteristics | Surface tension, viscosity | Viscosity, vapor pressure, surface tension | Surface tension, electrical conductivity, viscosity |
Advantages | Parallel operation (throughput), technological maturity | High standoff distance, limited nozzle clogging | High resolution, high viscosity range |
Challenges | Nozzle clogging, narrow viscosity range | Overspray | Standoff distance, print throughput |
Ink formulation for inkjet printing is driven by three key requirements: engineering for suitable jetting, wetting, and drying. Jetting criteria are well established based on fluid dynamics and are captured by nondimensional numbers related to droplet formation. In particular, the inverse Ohnesorge number is frequently used to screen materials for jettability, identifying a suitable viscosity range and, to a lesser extent, surface tension, while considering the nozzle size.40 Excessively viscous inks prevent droplet formation by dissipating the pressure wave within the fluid, while overly inviscid inks can lead to satellite droplet formation and an accompanying deterioration in print quality.41 Wetting describes how droplets spread and coalesce with neighboring droplets on a substrate. As with jetting, there is a balance to achieve here. Excessive wetting reduces precision and can lead to nonuniform deposition of material, while insufficient wetting can prevent formation of continuous lines and films or lead to unstable liquid migration on the surface. As opposed to jetting, which is primarily contingent on the interplay of the ink and the printer, wetting is dictated by the interaction of the ink and the substrate surface. Both surface energy and surface roughness have a role, along with the surface tension of the ink. It is common to qualitatively assess wetting characteristics by measuring the ink contact angle on the surface, although this is typically a static contact angle and not wholly representative of dynamic phenomena. Importantly, nanomaterial inks will typically exhibit some degree of contact line pinning arising from deposition of functional materials, leading to a quasi-stable contact line.42 This behavior is challenging to predict and quantitatively assess, but can often be initiated via ink binders.
Drying is the final stage of patterning, before any post-processing. Because liquid inks commonly contain a high volume percentage of solvent, this carrier fluid must be removed by evaporation to yield a solid deposit (many UV curable inks are an exception to this statement but are less representative of typical nanomaterial formulations). Evaporation of solvent from the fluid bead leads to internal temperature and composition gradients that drive transport. This can cause the functional material to deposit near the periphery of the droplet/feature, in what is known as the coffee ring effect.43 Manipulation of the ink, surface, and drying environment can mitigate or reverse this effect, with significant impacts on the uniformity of printed material.44,45 For some applications, controlling this stage of the process is critical, and less straightforward than the relatively explicit criteria required for droplet formation. Coffee ring formation has also been deliberately exacerbated and exploited for several laboratory demonstrations,46–48 although the preponderance of applications require reduction in this effect to improve uniformity.
One of the limitations for inkjet printing is in scaling to smaller feature size. While this is not a hard constraint, as the production of finer nozzles is certainly feasible, the physics of the process become more demanding at smaller length scales. On one hand, particles with finite size can affect the propagation of pressure waves through the ink in the droplet formation step and can also physically clog the nozzle when they are on a similar order of magnitude in size scale. In addition, the energy required to form droplets, effectively overcoming surface tension, exhibits a relative increase when moving to smaller dimensions. If the print resolution is constrained primarily by the droplet size, and this is dependent on the nozzle size, this establishes a constraint to scaling to finer and finer features.
The stages of aerosol jet printing differ from inkjet printing in that droplet generation, transport, and deposition are largely decoupled; wetting and drying phenomena, meanwhile, parallel those for inkjet printing with some notable differences. Droplet generation for aerosol jet printing is less controlled and precise than for other droplet-based printing methods, in that the process can tolerate a polydisperse distribution of aerosol droplets. With ultrasonic atomization, the classical understanding of droplet formation is based on pinch-off from a standing capillary wave on the liquid surface, in which the droplet size is largely based on the surface tension and ultrasonic frequency, and droplet generation is limited by viscous damping of the acoustic energy that imparts an upper limit on ink viscosity.51,52 More viscous inks, therefore, rely on pneumatic atomization, in which high velocity two-phase (ink and carrier gas) flow through a narrow orifice results in liquid sheets and jet break-up for the generation of a wide distribution of droplet sizes. Much of the complexity of this process is covered over, as large droplets are collected within the cartridge, moderate-sized droplets settle out gravitationally before they reach the printhead, and very small droplets can be removed via a virtual impactor, which has the important purpose of reducing the gas flow rate.
Within the printhead, the sheath gas prevents droplets from impinging on the interior surface of the nozzle, helping prevent clogging and allowing feature sizes much smaller than the nozzle dimensions (i.e., 10–40%). This sheath gas also has an important influence on the droplet composition, as a dry sheath gas induces evaporation of solvents in the inks.53,54 This has the upside of increasing fluid viscosity and decreasing solvent volume prior to droplet impact on the surface, which can reduce liquid-phase spreading and improve stability of printed features.55 It also has a downside, in that the reduction in droplet size, and thus inertia, that results from evaporation on the periphery of the aerosol stream contributes meaningfully to overspray, a characteristic of AJP that describes diffuse deposition of material along the edges of printed patterns.56
These two mechanisms – droplet formation and droplet evaporation within the printhead – constrain and guide ink formulation. The requirement for droplet formation limits ink viscosity, particularly for ultrasonic atomization, and constrains the particle size as well. The drying action of the sheath gas constrains the vapor pressure of ink solvents, with excessively dry (high vapor pressure) inks resulting in significant overspray and granular morphology, and excessively wet (low vapor pressure) inks resulting in liquid-phase spreading, which can become uncontrolled under the high velocity gas jet. As implied above, this in-line drying modulates the influence of wetting and drying mechanisms that have been previously discussed in the context of inkjet printing.
Continuing efforts to improve AJP capabilities have in many cases focused on fine line patterning. While resolution in the 10–50 μm range is reasonably achievable, this 10 μm threshold has recently been exceeded through rigorous process optimization,57,58 along with creative strategies to exploit non-aerodynamic focusing via acoustic fields.59,60 These demonstrate that 10 μm is not a fundamental limit, and generalizing these results to broad material sets and higher deposition rates will further expand the capabilities readily accessible to AJP technology.
During EHD jet printing, the functional ink is passed through a charged nozzle, which forms a closed circuit with the substrate or support. The electric field between the nozzle and substrate causes charge build-up on the fluid surface, which deforms to create a Taylor cone. At suitable conditions of liquid back-pressure, material properties, and electric field, the liquid can form a stable cone jet, resulting in break-off of very fine (∼fL) droplets which then deposit on the substrate surface. Different jetting modes are possible, and application of a pulsed electric field to induce droplet formation can offer better control of frequency and droplet size.64 It is also possible to use alternating current to address challenges of residual charge buildup for inks on insulating substrates.65 Given the mechanism of droplet formation – relying on an electrical field to pull off droplets rather than a pressure wave to eject material – the requirements for ink formulation are quite distinct from traditional inkjet printing. With droplet formation being driven primarily by competition between surface tension and electrostatic phenomena, electrical conductivity and dielectric permittivity of the ink are important properties, alongside surface tension. Viscosity remains important as a secondary variable, implying a wide range of viscosities can be effectively printed but that higher viscosity inks can influence dynamics of droplet formation and the relaxation timescale of the ink.66 Additives can be incorporated in inks to modulate surface tension, electrical conductivity, and viscosity to tailor characteristics for EHD jet printing. Alongside the material properties, printer system characteristics including nozzle size and surface characteristics, nozzle-substrate distance, and voltage are significant determinants of the print capability and quality.67
Following deposition on the substrate, the wetting and drying characteristics of EHD jet printing parallel those for inkjet and aerosol jet. While electrostatic repulsion in the charged droplet can result in somewhat lower contact angle and large wetting radius, this is a modest effect in comparison to the substantially reduced droplet volume compared to other droplet-based printing methods. As with AJP, EHD jet printing employs droplets much smaller than the nozzle size. While this can aid in reducing nozzle clogging, it does not entirely eliminate it. Solvents with moderately high boiling point, and small particle size with high dispersion quality, remain effective for mitigating nozzle clogging during printing.
Based on its unique physics, EHD jet printing has significant benefits for high resolution patterning.67 The flip side of this is a reduced print throughput, given the frequent use of single-nozzle printing, extremely fine droplet size, and physical limitations in the jetting frequency. Moreover, because EHD jet printing is driven by an electric field between the nozzle and substrate, it is poorly equipped to accommodate a high amount of roughness, nonuniformity, or 3D topography in surfaces.68 Novel approaches to EHD jet printing continue to be introduced, including methods potentially suited for patterning on curved surfaces by exploiting surface polarization,69 along with strategies to dramatically increase jetting frequency.70
Custom inks also provide an opportunity to tailor nanomaterial morphology for custom functionality. While silver nanoparticles are effective for reasonably dense, conductive lines, they form optically opaque patterns. Silver nanowires, on the other hand, can form sparse percolation networks to support both electrical conductivity and optical transparency for use as a transparent conductor. For example, a custom silver nanowire ink was synthesized with silver nitrate, polyvinylpyrrolidone, and ethylene glycol, and by controlling the nanowire length, the authors were able to formulate an ink compatible with inkjet deposition using a desktop printing system.76 More generally, silver nanowires are considered for applications in microelectronics, thin-film solar cells, and biosensors.77–80 For these 1D nanomaterials, there is a general tradeoff in the morphology and printing performance, with high aspect ratio desired for efficient percolation to support transparent conductors but making inks more prone to droplet formation difficulties, limited solids concentration, and complex rheology.76
Although silver is extensively used for fabricating electronics and prototypes, in general the materials research community has shifted to other printed conductors given the maturity of commercial silver inks (Fig. 3). With silver as the benchmark, these other materials must demonstrate some realistic advantage on at least one dimension to justify development. While not an exhaustive list, we will highlight here recent research efforts focused on other metals, along with select inorganic conductors beyond traditional metals.
![]() | ||
Fig. 3 Overview of common materials and classification for droplet-based printing technologies focused on electronic characteristics. |
Gold is among the more established materials for nanoparticle inks beyond silver given its historical importance in nanotechnology and tolerant synthesis and processing. For applications, the primary benefits that gold offers include its extreme resistance to oxidation and its biocompatibility. Of course, the cost of gold is substantially higher than that of silver, so significant benefit must be justified for this to be a logical replacement for silver. Because of its higher melting point, gold nanoparticle inks in many cases exhibit higher sintering temperatures than necessary for silver. A notable recent effort to develop low-temperature sintering gold inks compatible with both inkjet and aerosol jet printing resulted in resolutions less than 20 μm while achieving electrical resistivity as low as 9.6 × 10−8 Ω m for 400 nm thick films.81,82
Zinc is an alternative conductive material with unique characteristics for biocompatibility and biodegradability. However, the printing and sintering approach for zinc is generally quite distinct from traditional silver nanoparticle inks due to the ready oxidation of nanoscale zinc particles. Majee et al. inkjet printed a zinc nanoparticle ink made from glycols (i.e., dipropylene and ethylene) and polyvinyl butyral on a polyethylene terephthalate (PET) substrate. Thereafter, the printed patterns were chemically sintered using acetic acid. Such a method removes the need for thermal sintering and thus allows integration with thermally sensitive substrates and other materials, in this case yielding conductivities on the order of 105 S m−1.83 Furthermore, although inks made from zinc may act as alternative conductors to silver, they also hold promise for printed primary batteries, and zinc's biodegradability aligns with opportunities for transient electronics.84–86
Coppers inks have been a significant focus for fundamental and applied research in printed electronics for years. The most commonly cited driver for this is cost and abundance, and copper offers comparable electrical conductivity to silver. In addition, copper is a benchmark material for printed circuit boards and microelectronics, and so many peripheral materials and methods – for example, solders – have been developed to be compatible with copper. In some cases, printed silver is not an effective drop-in replacement within this larger processing environment. While some commercial distributors of copper inks exist, the technical challenges and broad potential for this material motivate continued research efforts. In particular, while bulk copper has high conductivity, nanoparticle-derived films typically exhibit significant porosity, and the high surface area and local curvature of such films make them prone to oxidation. While films derived from silver nanoparticles are tolerant to this – they have a higher reduction potential than copper and remain conductive in the oxidized state – copper has proven to be much more sensitive. This can lead to both immediate effects (significantly greater challenges in achieving conductivity comparable to bulk), and long term stability concerns.87 Such instability can reduce the electrical conductivity of copper prints and hamper reliability in applications. For this reason, some researchers have employed strategies to maintain good conductivity while curbing the problem of instability. One example is the core–shell approach, in which a layer of oxidation-resistant material covers a copper core within each nanoparticle.88 The use of capping agents in copper ink formulation also has a significant role for oxidation stability, forming a barrier between the copper particle and the outside environment to slow down the oxidation process prior to printing.89–92
Apart from pure copper conductive inks, inks made of copper alloys can find application in sensing systems, such as CuNi for thermocouples. For these applications, reduced electrical conductivity compared to bulk is not a strong performance driver, so this can be more tolerant to the challenges with oxidation and porosity. Several researchers have therefore demonstrated printing of such alloys. Gu, et al. made a printable, aqueous, oxidation-resistant constantan ink using polyvinyl pyrrolidone, ammonium chloride, and glycerol, from which a sensor was fabricated via inkjet printing, yielding a temperature coefficient of resistance of 4 × 10−5 K−1.93 Sheng, et al. featured a thermocouple made of Cu–CuNi nanoparticle ink and printed via inkjet deposition. It showed high sensitivity (20.6 μV °C−1) while maintaining stability under thermal cycling.94 Likewise, a temperature sensor of high sensitivity (∼40 μV °C−1) was fabricated via aerosol jet printing using Cu and CuNi inks.95
Among other conductors, platinum is an expensive material like gold, but one with exceptional properties for certain high value applications. Having a high melting point, broad catalytic activity, biocompatibility,96 and electrochemical stability, platinum has been explored for fuel cells,97 dye-sensitized solar cells,98 electrochemical sensors,99 and high temperature electronics.100 Platinum inks can include both particle-free and colloidal formulations, given the high redox potential of platinum.100,101 Nickel inks can be magnetized and withstand high temperature oxidizing environments.102 It has been shown that nickel ink made from nickel acetate, polyvinyl pyrrolidone and ethylene glycol is compatible with aerosol jet printing on various substrates for high temperature applications.102 Apart from the coinage metals, a titanium-based ink in the form of hydride nanoparticles has also been demonstrated for aerosol jet printing, representing an alternative approach.103
In a broad sense, metal inks are commonly used for their high electrical conductivity. However, issues such as cost (silver, gold, platinum), oxidation susceptibility (e.g., copper), high sintering temperatures, and concerns of toxicity have motivated the development of alternative conductive inks. In this regard, graphene- and MXene-based inks have been extensively developed for various printing technologies. Because the latter are described thoroughly in recent review articles,104–106 we limit the discussion here, but MXene inks have been demonstrated for inkjet, EHD, and aerosol jet patterning technologies, with applications including electrochemical energy storage and sensing.107–109 The ability to form stable, dispersant-free formulations of MXenes significantly aids in obtaining highly conductive patterns with low processing temperature, and tailoring ligand chemistry can further support solvent versatility.110
Both graphene and MXene printed materials are composed of high aspect ratio flakes or platelets, which allow fairly dense film packing. Macroscale functionality of such printed patterns are based on both the material properties of individual flakes and, crucially, the flake-to-flake interconnections.111 As another point of commonality, these materials do not sinter in the traditional sense, such that the flake-to-flake junction resistance can be meaningful. As a result, printed conductors often favor larger particle sizes for electrical functionality, but this can come at a tradeoff for dispersion quality, solids loading, and ease of printing due to rheological effects of high aspect ratio particles and size limitations where dispersed particles begin to influence droplet formation physics.112
Graphene is touted to have good mechanical strength, thermal stability, and electrical properties.113,114 However, there is a wide range of achieved conductivity for printed graphene in various works, ranging from <300 to >70000 S m−1.115–118 This variation may be in part attributed to the quality of graphene produced for inks, related to flake size, defect density, and heteroatom content, along with flake-to-flake interconnections. While there are a variety of uses for printed graphene, in many cases the target is either high conductivity or electrochemical activity, and the target graphene properties can thus vary depending on the application.
For highly conductive printed patterns, low defect density, low heteroatom content, and large flake size is usually a preference.119 In developing graphene inks, issues related to particle aggregation must be carefully addressed to ensure favorable and stable deposition.120 This often requires dispersants or stabilizing agents, which influence electrical conductivity and required post-processing steps.121 To achieve low processing temperature, it is preferred to limit the use of such stabilizing agents or eliminate them altogether, but this must ideally be achieved without compromising good dispersion of the ink.122
In addition to its use for conducting patterns,123 graphene offers a high surface area due to its 2D structure, and can contain chemically active sites in the case of oxidized materials.124 It is therefore capable of chemically interacting with molecules as well as facilitating carrier transport.125,126 These interactions can thus be harnessed in the area of electrochemical devices for sensing and energy storage.127,128 While the electrical conductivity remains relevant for these applications, in many cases it is outweighed by surface area and electrochemical considerations, with an inherent tradeoff between highly active, poorly stacked materials for electrochemical activity and densely stacked, highly pure materials for electrical conductivity and electrochemical inertness.
A common device that relies on dielectrics is the thin-film transistor (TFT), a basic building block for more complex electronic circuits. Here, the dielectric couples a gate bias into the active semiconducting channel to modulate charge transport while maintaining minimal leakage current through this gate dielectric. For this application, high capacitance is preferred to reduce the working voltage of the device. For printed TFTs, the gate dielectric must also withstand subsequent processing steps, which could include solvent and thermal exposure, while maintaining a pinhole-free barrier. High-k metal oxide dielectrics are a common target for this application, including compositions based on molecular precursors of alumina, zirconia, and hafnia to achieve thin, smooth, amorphous films.132–136 A common challenge for these materials is the traditionally high processing temperature required to achieve dense, pinhole-free films with suitable dielectric characteristics. Sol–gel processing for metal oxide thin films traditionally requires temperatures exceeding 400 °C, which are incompatible with flexible substrates. Strategies to alleviate these concerns include alternative annealing methods such as intense pulsed light and deep UV exposure,137,138 along with tailoring chemical precursors to generate heat during exothermic decomposition and overcome the energy barrier for conversion to oxides with less external heating.139 A recent example used a solution-processable aluminum oxide film annealed by intense pulsed light for the gate dielectric of a metal oxide transistor, which showed a large capacitance of 109 nF cm−2 and low leakage current density of <10−8 A cm−2 at a working voltage of 10 V.140 Achieving thin, pinhole-free films for transistors remains a challenge, particularly for high purity dielectric materials that can support high frequency operation with minimal loss.
While sol–gel methods can yield dense dielectric films, nanomaterial-based dielectrics can also be employed in certain applications. In particular, 2D nanomaterials can exhibit favorable film morphology to support electrical isolation. An example within this category is hexagonal boron nitride. With a bandgap of 6 eV, chemical and thermal stability, high mechanical strength, and a basal plane free of dangling bonds, boron nitride is a promising dielectric for electronics.141 However, as with graphene, the material quality achieved for printed devices significantly lags that for single-flake demonstrations. Liquid phase exfoliation of bulk hexagonal boron nitride flakes has been employed for ink preparation, with additional stabilization methods to reduce aggregation and support printing of thin, pinhole-free films.142 In contrast to the low dielectric constant of boron nitride, researchers have synthesized high dielectric constant (k ∼ 192) perovskite nanosheets, such as Ca2NaNb4O13. By printing single and multilayer devices using this family of nanosheets, Zhang, et al. demonstrated fully-printed capacitors with a capacitance density as high as 346 nF cm−2.143,144
Polymer dielectrics provide distinct processing and performance characteristics compared to most inorganic printed materials. In general, they offer better process compatibility at low temperature, at the tradeoff of limited thermal stability. While often the thermal stability is sufficient for applications, it restricts subsequent process steps, such as sintering metal nanoparticle inks. While this report overall is largely focused on inorganic materials, within the scope of dielectric materials the use of nanocomposites can provide an effective means to tailor properties while leveraging the inherent properties of inorganic materials. For example, Wu, et al. developed a nanocomposite dielectric ink based on nanosheets of Ca2Nb3O10 in a PMMA matrix.145 This ink was patterned by aerosol jet printing to form the gate dielectric for a thin film transistor. In the same way, Abdolmaleki, et al. demonstrated inkjet printing of a BaTiO3/PVDF ink.146 Here, the BaTiO3 affected the crystallinity of the PVDF to enhance ferroelectric and piezoelectric characteristics. In addition, a BaxSr1−xTiO3/PMMA ink for inkjet printed capacitors exhibited high dielectric constant of 20–42 at 1 kHz.147 Such composites harness the low temperature processing and flexibility of polymers while leveraging inorganic nanomaterials to tailor performance.
Although some studies have been done to utilize composite dielectric inks for inkjet148 and aerosol jet130 techniques, in general dielectric materials are lagging conductors in their development for non-contact printing methods, and clear standards and performance criteria would aid broad development in this space. While conductive materials have relatively explicit and easy to measure metrics, dielectric performance can often be context-dependent and relate to pinholes, reliability, stability to subsequent processing steps, and other factors that are more challenging to generalize.
Irrespective of application or dielectric type, a quality of critical importance in printing dielectric inks is the homogeneity of the print, as this directly influences device performance characteristics such as insulator performance, consistency, and reliability.149,150 Thus, careful attention must be given to both ink engineering and printing parameters to achieve uniform deposition. Here, there is significant work required to better understand how material development and process design intersect to influence tradeoffs in patterning resolution, uniformity, and microstructural quality.
Because of their flexibility, intrinsic carrier mobility, good chemical stability, and compatibility with wide-area printing, carbon nanotubes (CNTs) have been used for thin-film transistors using non-contact printing, primarily as a p-type semiconductor. The clear motivation for flexible CNT devices has driven research for >15 years,155–157 leading to a relatively mature technology currently that is pushing into new areas in research labs. There are several strategies for developing printed CNT devices, which due to their high aspect ratio, and the desire to have long nanotubes for improved properties, can pose a challenge for reliable deposition. Common strategies involve inks containing semiconducting CNTs based on water/surfactant mixtures, select polar organic solvents, or stabilizing conjugated polymers in hydrocarbon solvents.158–162 A key challenge for transitioning CNT-based devices is reliability and control over the interplay between CNT purity, size, concentration, printing characteristics, density and morphology on the surface, and device functionality. While several groups have effectively calibrated the process to balance device mobility (favored by higher density of CNTs) and on/off ratio (favored by lower density, high purity), generalizing this and controlling it well in a production setting requires clear focus.158
Metal oxide semiconductors gained popularity for display applications, offering a transparent semiconductor if oxygen vacancies can be well controlled in derivatives of zinc, tin, and indium oxides. Sol–gel approaches for deposition of these materials are a frequent strategy, with similar thermal limitations as described for metal oxide dielectrics. However, when the process temperature can be reduced or accommodated, these materials can offer competitive mobility to organic materials and sufficient performance for a range of device applications. Because their electronic properties are driven by oxygen vacancies, effectively controlling vacancy concentration via composition and processing is key to achieving stable, high-performance devices. Indium–gallium–zinc oxide (IGZO) is one example in this category which is suited for transistors. This n-type semiconductor gained prominence in display technologies due to high carrier mobility characteristics, low off current, and good optical transparency. Due to its ternary nature, the composition can be varied to tailor the overall characteristics of printed devices.163 The fabrication of IGZO channels for thin film transistors via inkjet printing has been realized in a wide range of literature reports.164–166 As with CNT devices, this is reasonably mature within the research community,167 and more recent efforts have extended these materials to broader application domains such as sensors, memory, and neuromorphic computing.168–170
While the optical transparency of metal oxide semiconductors motivates their use in transparent electronics, in the context of photovoltaics and optical sensing alternative materials are desired. Chalcogenides and perovskites are particularly suited for optoelectronic applications and have demonstrated compatibility with non-contact printing methods. Cu2ZnSnS4, for example, is an effective absorber for printed thin film solar cells. In the case of perovskites, devices fabricated via aerosol jet printing have been demonstrated for detection of X-rays.171,172 In both cases, it is generally more common to print precursor or molecular inks, which are converted to the semiconductor composition and crystallinity following printing and post-processing. Certain chalcogenide materials can also be leveraged for their thermoelectric properties. Ag2Se, for example, has good thermoelectric properties even at modest temperatures. It was adopted for inkjet printing via the dispersion of Ag2Se nanoparticles in ethanol followed by sonication to obtain stable inks.173 Others such as Sb2Te3 and Bi2Te2.7Se0.3 have also been patterned on various substrates via aerosol jet printing to demonstrate thermoelectric devices with good power factor.174,175
In addition to these bulk semiconductors, there have been many efforts to develop semiconducting 2D materials for printable inks.176 Because this is thoroughly discussed in more focused literature reviews elsewhere,177,178 we forego repetition here. In general, this class of materials offers diverse functionality with inherent characteristics for liquid-phase processing. Due to the polydisperse nature of flakes, extra effort is demanded to control flake geometry, as it impacts both printing characteristics and device functionality.
There is clear potential for printed devices of various functionalities to be realized with semiconducting inks, making them a continued focus area for research efforts. However, the diverse fragmentation of applications, and thus key metrics, makes it challenging to develop meaningful figures of merits that generalize between use contexts, and complicates the business case for transitioning successful laboratory demonstrations of materials beyond research labs.
In many laboratory efforts, crude formulation of inks centers on solvent selection and composition. Blends of solvents frequently provide greater flexibility to tailor chemical, fluid, and transport characteristics of inks compared to single-solvent compositions. While typical commercial inks contain complex combinations of solvents, cosolvents, and additives, many laboratory-developed formulations rely on only one or two solvents. This is often practical for demonstration and material evaluation, but more sophisticated formulations are expected to be useful to improve printability and tailor characteristics for manufacturing-centric metrics. Explicit standards driven by manufacturing process compatibility and maturity, rather than material functionality for a one-off laboratory demonstration, could more broadly accelerate advances in printability that consider stability and variability relevant to a production environment.
![]() | ||
Fig. 4 Post processing approaches for nanomaterial inks. (a) Illustration of drying and sintering mechanisms for nanoparticle-based inks; adapted with permission from ref. 194 (CC BY). (b) Classification of various methods to achieve sintering or annealing in printed materials. (c) Example of energy-time map for thermal and pseudo-thermal post-processing techniques. |
A second broad alternative to thermal curing is chemical sintering. This allows for room temperature sintering of printed conductive traces. With ink chemistry tailored to this, one demonstration of this leverages exposure to an acidic vapor environment,83 which in this case caused desorption of additives and subsequent particle fusion. This method is particularly beneficial for temperature sensitive substrates, but can complicate processing, particularly if multiple materials are printed. Another approach involves printing the acidic solution onto the conductive trace and subsequently drying any solvent at low temperature.83 Plasma sintering has also been studied, as it enables lower temperatures and reduced time to be used in the process of sintering due to the chemical activation of surfaces from the plasma.191 Low temperature nonthermal plasma sintering is an alternative to standard plasma sintering as it mitigates thermal effects on delicate and flexible substrates.193 One motivation for plasma sintering is the ability to leverage chemical activation while retaining reasonable suitability for scalable processing. Fig. 4 illustrates broad methods used to post-process printed electronic devices in literature.194 Several more recent approaches to enhance nanoparticle sintering include induction heating, electrical sintering, damp heating sintering, and near-IR sintering to reduce processing time, limit substrate heat exposure, and improve material properties.195–198
![]() | ||
Fig. 5 Demonstrator applications for droplet-based printed electronics. (a) Low-pass filter with capacitor and inductor elements. (b) Passive components for high frequency electronics, including transmission lines and radial stubs. (c) Humidity sensor printed directly onto a packaged integrated circuit. Graphics adapted with permission from ref. 201 and 208 (CC BY), and ref. 230 (CC BY), respectively. |
Of course, the ability to recreate functionality of basic circuit elements, which are widely available as standardized surface mount components at low cost and high reliability, is a necessary effort but not the ultimate objective. Improving on the functionality of these – for example, by supporting flexible or stretchable designs or accommodating operation in extreme environments – would provide a greater value proposition for these printing technologies.
In many cases, capacitance achieved in printed devices is limited by inherent challenges in coupling a thin dielectric layer with multilayer structures. Manufacturing process reliability and pinhole prevention make this a challenge even for basic metal–insulator–metal capacitors, with yield challenges if these are stacked to increase capacitance in a given footprint of the substrate. Inductors, on the other hand, often support limited current flow owing to the coil length and the generally thin nature of printed conductors in comparison to electrolytic copper. In contrast to planar inductors, Gu, et al. demonstrated relevant inductance values for solenoid-type devices using polymer, iron, and ferrite cores, leveraging the conformal patterning capability of aerosol jet printing.202
Antennas provide another case for which the large-area patterning, conformal printing, and rapid prototyping capabilities of non-contact printing methods carry significant advantages (Fig. 5b). A wide range of antennas have been demonstrated,206–208 including conformal devices and complex antenna arrays with completely printed building blocks.209 In addition, complex metasurface antenna designs exemplify the design flexibility of these printing methods, and highlight the complex interplay of device design and fabrication required for multi-axis printing.210 One challenge less apparent under low frequency conditions is the sensitivity to porosity, roughness, and defects. The issue of surface roughness and material quality associated with prints should therefore be more thoroughly addressed to reduce losses of these high frequency devices and allow them compete with conventional technologies.211,212
![]() | ||
Fig. 6 Rationale for hybrid circuits combining conventional and printed technologies; reprinted with permission from ref. 121. |
In this context, a key technical challenge is designing and controlling the interface between printed and microfabricated components.215,216 Printed conductors on polymer substrates often have poor compatibility with conventional solders,217,218 motivating strategies for printed interconnects.219 This is a compelling area for research, and provides a practical route to integrate printed materials and components into more complex systems. Extending this approach to photonic devices, as highlighted by printed waveguide structures, provides another promising area of research for practical applications.220–222
Effectively comparing the performance of sensors made by individual printing methods is difficult due to the widely varying characteristics of interest and metrics adopted for performance evaluation. Stability and reliability of printed sensors, which encompasses manufacturing process reproducibility, is of particular importance for maturing printed sensors to address the broad range of applications where they offer meaningful benefits. While many research efforts demonstrate individual sensors for specific stimuli, multiplexed sensors that can compensate for interference, allow spatial specificity, or increase the range of chemical analytes will continue to increase in sophistication. Machine learning methodologies to deconvolute, simplify, and interpret large amounts of data from printed sensors will be an important complement to the diversity of materials and device configurations.231
We summarize a subset of these efforts in Table 2. For AJP, this research commonly leverages image analysis to classify line quality, with several works including functional properties such as resistance. For the most part, these efforts vary the digitally controlled printing parameters, such as gas flow rates and print speed, and have not yet extended to ‘low frequency’ variables such as ink composition and nozzle size. For IJP, there have been several interesting studies applying image analysis and high throughput experimentation to the droplet generation process itself, tailoring the waveform for droplet ejection, which can include predicting droplet velocity and size based on the ink properties. In addition, imaging following deposition can be applied as a predictor of print quality. A similar approach can be leveraged for EHD printing,233 using imaging of the printer nozzle to classify the jet characteristics. In addition, as for AJP, print parameters such as flow rate, speed, and voltage can be connected to both geometric metrics of print quality and electrical characteristics.
Target output | Model input | ML methods | Ref. |
---|---|---|---|
Acronym legend – RF: random forest; GB: gradient boosting; DT: decision tree; MP: multilayer perceptron; K-NN: k-nearest neighbors; SV: support vector; GP: gaussian process; RL: reinforcement learning. | |||
Inkjet printing | |||
Drop size, velocity, formation | Voltage frequency | Back propagation neural network | 235 |
Drop velocity, formation | Waveform ink properties | MP | 236 |
Drop size, velocity, formation | Waveform ink properties nozzle size | Ensembles: DT, GB, RF | 237 |
Line quality electrical conductance | Environment drop metrics frequency | Convolutional neural net | 238 |
Aerosol jet printing | |||
Line quality (edge) | Gas flow rates print speed | Hybrid: SV, K-means | 239 |
Electrical resistance | Gas flow rates print speed | Sparse rep. classification | 240 |
Line thickness line width | Gas flow rates print speed | Hybrid: SV, GP | 241 |
Electrical resistance | Gas flow rates print speed | Hybrid: MP, RL | 242 |
EHD jet printing | |||
Electrical conductivity | Print speed ink flow rate voltage | RF K-NN | 243 |
Droplet diameter | Ink properties nozzle size, standoff, flow rate, voltage | Artificial neural network | 244 |
Droplet diameter ejection frequency | Ink properties pulse settings flow rate nozzle size, standoff | Various; GB regression best | 245 |
Line width | Settings (standoff, voltage, speed) In situ images | Ensemble | 246 |
These applications of machine learning serve a dual purpose. When narrowly applied, for example to a particular ink, they provide a methodology to accelerate process optimization. More broadly, when machine learning is used to tease out connections between ink properties, print parameters, and process outcomes, this can reflect insight on the physical mechanisms of the process. This second approach, while more ambitious and complex, has strong potential to complement traditional methods, but this is somewhat contingent on the transparency or interpretability of the machine learning models.
Optimization of machine parameters for printing is a fairly well-defined and compartmentalized problem that lends itself well to rapid iteration to build up datasets with full digital control. A noteworthy example from the space of direct ink writing, by Deneault, et al., developed an autonomous research system with Bayesian optimization (Fig. 7a).234 It is thus reasonable that this constitutes an initial major foray of machine learning into additive electronics, but it is not the complete picture. In particular, while iteration over process parameters can be performed digitally in a purpose-built fabrication system, iteration over materials or ink formulations largely remains manual and thus relatively slow. While this is thus a less realistic target for short-term research advances in efforts limited to an individual lab or small research team, the complexity of ink formulation and non-continuous nature of many variables makes it a compelling target for future research where the scale and capability of sophisticated data analytics tools offer meaningful benefits. In this context, more standardized approaches to ink evaluation – even without the context of machine learning – would be valuable to establish broad datasets that allow meaningful comparison. Moreover, framing the test methodology on foundational process physics could support more efficient development and generalization than purely empirical approaches.
![]() | ||
Fig. 7 Application of advanced computational methods for digital printing. (a) Example of autonomous research system for closed-loop optimization of direct ink writing, in which computer vision feedback is combined with Bayesian optimization and printer control to optimize print parameters; adapted with permission from ref. 234 (CC BY). (b) Hybrid machine learning approach for AJP, which allows direct optimization on the functional characteristics of the printed materials, in this case resistance measurements; adapted with permission from ref. 242. |
A more approachable near-term target may be multi-objective optimization with constraints on training data size. Additive electronics is not naturally a big data problem, at least if the focus is on process optimization for a single ink. The work of Du, et al. is a clarifying example to illustrate this,242 focusing on iterative optimization with minimal data, and including both geometric and functional properties (Fig. 7b).
Notably, this is a dynamic research area with broad interest, and will continue to evolve rapidly in coming years.247 Increasingly sophisticated methods incorporate multimodal data streams and multiobjective optimization across scales. Furthermore, machine learning toolsets are becoming more broadly accessible, allowing a wider range of practitioners with interests across materials, methods, and applications to leverage these capabilities. While many of the early demonstrations zero in on process parameter optimization, broader application in part design for additive manufacturing,248 along with digital twins for defect prediction or real-time adaptive control,249,250 provide just some examples of the compelling opportunities in this space to augment physics-based process understanding.
Going forward, it is instructive to observe the applications of machine learning in other research domains. In particular, there are broad and successful efforts to apply machine learning in the space of materials science, leveraging broad communities with common data standards. Merging materials and process optimization in a similar manner could have significant benefits for practitioners of printed electronics. Notably, several studies applying ML methods to printed electronics have sourced data from the literature,237,245 hinting at the logical transition to larger data sets. This is challenged by limited standardization in data collection and reporting across research labs, but larger efforts to address this could meaningfully change the scope with which artificial intelligence could be applied to augment material and process development.
Finally, there should also be fair consideration given to digital approaches to improve these processes that rely on more traditional and straightforward methods. Where robust theory and understanding exists, there is less need for opaque machine learning models. This is supported by various efforts to apply computational fluid dynamics for process modelling, along with deterministic and physics-based control approaches.58,251–255 While complex, printed electronics is not nearly as much so as the stochastic and often chaotic phenomena for which big data methodologies excel, and physics-based understanding of processes can better support generalization, adaptation, and extrapolation to new systems.
![]() | ||
Fig. 8 (a) Multimaterial aerosol jet printing, in which two ink streams are mixed in situ to create gradient patterns, in this case for combinatorial screening of materials. (b and c) Conformal printing of conductive traces onto a drone safety cage using AJP. (d) High aspect ratio microelectrode array printed with AJP. Graphics adapted with permission from ref. 262 (CC BY), ref. 268 and 189 (CC BY-NC), respectively. |
Early adopters of this peculiar manufacturing approach for AJP included Reitz, et al., for grading composition vertically in a solid oxide fuel cell stack,258 along with Wang, et al. for varying electrical properties in CNT/polyimide nanocomposites.259 More recently, Craton, et al. demonstrated RF components with the multimaterial aerosol jet printing technique.260,261 In one instance, the authors demonstrated spatial tailoring of film dielectric constant by mixing a polyimide dielectric aerosol stream with one containing BaTiO3, leading to a ∼1.4 μm thick film with a dielectric constant between 3.1 and 8.9.261 They leveraged the same technique to mix nickel-zinc-ferrite nanoparticles in a static ratio into a polyimide matrix, which resulted in a 40% increased inductance with respect to similar planar inductors.260 A boost in recognition for this capability arose due to thorough work by Zhang, et al., with the broad demonstration of combinatorial printing using multimaterial AJP,262 including demonstrations for both thermoelectrics and later solid state electrolytes for LIBs.263 Gamba, et al. investigated multimaterial aerosol jet printing for the manufacturing of functionally graded materials. In one instance, the researchers mixed metallic aluminum nanoparticles with copper(II) oxide with constant stoichiometric ratio to fabricate high resolution nanothermites, which were later deposited on a silver resonator for wireless RF ignition.264 This result was enabled by the unique setup of AJP, leading to high resolution single pass thermite lines while also preventing premature mixing and potential accidental ignition of the fuel, increasing operational safety. In the same study, functionally graded manufacturing was also demonstrated with the same inks. The effect of the printhead design was correlated to the microscopic morphology of the energetic composites and it was concluded that static mixing elements within the printhead significantly enhance material mixing in the final traces. In a separate work, Gamba, et al. tailored electrical properties of carbon nanomaterial traces by mixing graphene and carbon nano onions with different stoichiometries, printing composites with either fixed mixing ratio or a compositional gradient.265 This allowed patterning of functional materials with electrical resistivity spanning two orders of magnitude, which could allow resistance targeting over a broad range by tailoring both the physical geometry of a part and its composition and intrinsic properties.
A compelling application of gradient printing was recently demonstrated by Zhang, et al.266 In this work, the authors prepared a mechanically graded interface by combining polyurethane dispersions with distinct stiffness. This allowed the graded substrate to interface well with skin via a soft polyurethane, with a smooth gradation to a harder surface to support electronics. This configuration allowed high performance electronics integrated on skin with considerable strain insensitivity, and broadly highlights the utility of multimaterial printing for creating controlled, diffuse interfaces.
Multilayer electronic circuits are highly sought for functional devices, and in many cases parallel the architectures of more traditional built-up substrates used in electronics packaging. In these most common instantiations, multilayer electronics at their simplest require deposition of both conductive and dielectric materials, with the compatibility between inks a critical determinant of functionality. This encompasses both materials compatibility in the final part (e.g., adhesion, interfacial strength) and considerations of process compatibility (e.g., curing conditions, solvent orthogonality). Several commercial suppliers have moved into this space, with Nano Dimension an early entrant using inkjet technology. The value of these platforms remains highly dependent on material properties, with dielectric materials often imposing limitations in thermal, mechanical, or electrical performance. Among the three droplet-based printing methods discussed here, inkjet has seen the most maturation towards multilayer circuits, largely a result of its scaling advantages due to parallel jetting, and direct ink writing approaches offer a useful guide, albeit typically for circuits with coarser resolution.269–272 At a high level, droplet-based printing methods with digital control have a potential advantage for such multilayer circuits. In particular, the digital control offers a possibility to adjust the print pattern to compensate for variations in previous fabrication steps, a part-by-part versatility that is not matched by methods with hard tooling. However, the realization of this remains largely limited to very narrow, proof of concept demonstrations. One advantage that noncontact, direct write printing offers for multilayer circuitry is the ability to localize insulating material to crossovers and other key places, reducing the overall amount of material and weight of the part compared to a traditional printed circuit board. Extending this concept to optical materials with multimaterial inkjet printing, researchers have developed gradient index lenses and sophisticated optics that hint at the scope of applications beyond electronics.273
The second flavor of 3D patterning for droplet-based methods is conformal printing, which describes the fabrication of thin film electronics on curved, 3D surfaces (Fig. 8b and c). This is a compelling application space not well served by master-based printing methods or most traditional fabrication technologies, offering a clear value proposition for digital printing. Conformal integration of electronics directly on structural components offers form factor flexibility with associated space and weight savings. This is particularly relevant to support miniaturization in high-value constrained applications, such as wearable devices, aerospace and automotive systems, biomedical devices, and space electronics.274 Moreover, direct integration of components such as sensors without an intermediating substrate can provide closer contact with structures for more precise sensing, for example in the case of thermal sensing and structural health monitoring.275 Nonplanar geometries can furthermore offer unique functionality for optics applications,276 along with versatility for hybrid augmentation of populated circuit boards or in-place repair.277 Among the methods discussed here, AJP is a highly compelling technology for conformal printing, with a high velocity jet that can maintain fine resolution even up to standoff distances of 5 mm.278 This offers key tolerance for contouring complex surfaces. Moreover, the suitability of AJP for deposition in an oblique configuration, which is supported by partial drying and resulting rheology changes prior to droplets hitting the surface, gives additional versatility for motion planning.279 Some prominent examples of conformal electronics include sensing systems,17,268 biomedical devices,280 antennas,281 and microinductors.282
The last two, related flavors of microscale 3D printing discussed here are patterning for high aspect ratio (height:
width) features and spanning structures in more complex configurations. AJP is again a promising technology for the demonstration of these capabilities, owing in part to its high standoff capability and suitability for printing high viscosity materials. In the area of printing high aspect ratio structures, Panat, et al. have demonstrated compelling applications of microelectrode arrays for biological sensing (Fig. 8d).27,267 The same group pioneered AJP for spanning 3D structures that qualify as freeform 3D printing, showcasing complex lattice structures and their application as current collectors for batteries.27,283 In all these cases, self-supporting structures are enabled by printing materials that rapidly solidify following deposition, largely through the traditional means of solvent evaporation.284 In a different approach, Akyurtlu, et al. demonstrated patterning of a dielectric polymer by rapid UV curing to create sophisticated 3D structures,285 with the 100% solids loading ink supporting relatively rapid fabrication. EHD jet printing also offers interesting alignment for certain high aspect ratio structures, such as micropillars.286 In this context, the shape of the electric field around previously-deposited material can enable precise structures approaching micron-scale diameter.287
This journal is © The Royal Society of Chemistry 2025 |