Conformal printed electronics on flexible substrates and inflatable catheters using lathe-based aerosol jet printing

Lathe printer attachment conceptualization and design

To achieve cylindrical-coordinate-based control of printing, a custom attachment for the x–y platen of a commercial AJ printer (Optomec AJ300) was developed, as shown in Fig. 1. In the computer aided design (CAD) diagram of the lathe mechanism, the various features that provide the lathe functionality can be readily identified. The primary feature worth noting is the mechanism for translating linear motion in the cartesian x-axis to rotational motion in the cylindrical θ-axis while maintaining linear motion in the axial direction along the y/z-axis. This translation is achieved by the rack, pinion, and axle clamp, which serve as the junction between the printer-controlled platen and the suspended rotating axle. The pinion is affixed on an axis above the rack by a linear motion shaft with screw-tapped ends that allow for on-axis attachments like the pinion and axle clamp. The pinion interlocked with the platen-fixed rack provides the rotation to the suspended axle, and the axle clamp, which locks onto a ball bearing in a guide rail, providing the linear motion to the shaft regardless of position on the x-axis (see Fig. 1a). This shaft is suspended above the rack by a position-adjustable support with two ball bearings to allow smooth rotation. This support is attached to the printer body separately from the platen and has adjustable guides to set the rotating axle directly under the nozzle tip. Depending on the direction of the platen movement, the suspended axle can freely rotate within the bearings, slide through the bearings to allow linear axial motion, or both simultaneously. The length of the shaft determines the range of axial motion, and full circumferential rotation enables deposition onto substrates of any diameter <40 mm, which is the largest substrate-holding mandrel that can be fit under the nozzle (see Supplementary Fig. 2). The significant advantage of utilizing a platen-mounted rack to translate linear motion to axial rotation is that this is a readily scalable solution, such that multiple suspended axles can be supported by the system without the addition of any motors or components beyond the additional substrate-supporting mandrels.

a CAD diagram of LAJ printer attachment and closeup image of the rack and pinion with axle clamp. b Picture of LAJ printer attachment secured to the printing system with deposition nozzle located above cylindrical rotating mandrel. c Illustration of the angle of print incidence for planar (stage linear motion in x) and lathe (substrate rotational motion in θ) AJ printing. d Images of planar and LAJ printed graphene traces at positions 1, 2, and 3 separated by ~5.9 mm for each method (dashed line indicates cross-section location with scale bar of 250 µm). e Thickness profiles across the width of planar AJ printed graphene traces at positions 1, 2, and 3—note the significant drop in thickness at position 3 where the incident angle was ~75°. f Thickness profiles across the width of LAJ printed graphene traces at positions 1, 2, and 3 showing consistency of thickness distribution at each site.

Using this mechanism, cylindrical-coordinate LAJ printing is achieved, making it possible to conformally print around the entire circumference of rotationally symmetric bodies without manual interaction. Examples of substrates that this configuration can conformally print onto include flexible films wrapped around cylinders, 3D rotational solids (cylinders, cones, helices, etc.), and hollow rotational solids such as tubing or balloons. Cylindrical bodies are the ideal case for this system, as the print incidence will be perfectly normal for the entire curved surface. However, with the mechanism it is possible to conformally print around 360° of rotation for any solid with a definable axis, which supports a high degree of applicability to any number of substrate geometries.

Significance of the angle of print incidence in conformal printing

The angle of print incidence for AJ printing is defined here as the angle between the target substrate and the aerosol stream that is responsible for material deposition onto the substrate. In the ideal printing scenario, the angle of incidence is always kept at 90° such that the deposition is normal to the substrate and factors affecting thin-film formation (e.g., surface hydrophobicity, inclination) are limited^63,64,65. Maintaining normal incidence during deposition is important because at increasing degrees of hydrophobicity, a droplet has less surface adhesion and will roll across a surface more easily at oblique angles of incidence. Through the practice of treating surfaces to lower hydrophobicity and keeping the print nozzle orthogonal over the substrate, print adhesion improves greatly for a variety of substrates. In planar AJ printing onto a flat substrate, it is trivial to maintain normal incidence; however, for conformal printing applications where the substrate topography presents varying degrees of nonplanarity, the ideal normal incidence is almost entirely lost with planar AJ printing. While the loss of normal incidence does not strictly mean an inability to print conformally, it does establish a discrepancy from the assumed ideal case that will create issues in print uniformity and design at increasingly steep angles of incidence. An example of this issue is presented in Fig. 1c–f, wherein planar and LAJ conformal printing of graphene ink onto a polyimide sheet wrapped about a cylindrical body is compared. The illustration in Fig. 1c shows that for planar AJ printing, the angle of print incidence is only normal at the acme of the cylinder (position 1) and continues to increase as the cylinder slopes away from the x-axis (positions 2 and 3). In contrast, LAJ printing maintains normal print incidence around the entire circumference of the cylinder because of the θ-axis rotation of the substrate, which holds the three positions as the acme of the cylinder directly under the nozzle throughout printing. It should be noted that in this experiment, the standoff distance for the planar case is also increasing with angle of incidence because the nozzle is not actively moving in the z-axis to correct for the change in distance above the substrate. However, even with active z-axis control the angle of incidence will be non-orthogonal and achieving adequate conformity around a circumference remains a major challenge.

The significance of the difference between planar and LAJ printing on a curved surface becomes visually clear in Fig. 1d when looking at the quality of printed traces between the two print methods at positions 1, 2, and 3 (each ~5.9 mm apart). At positions 1 and 2, the planar and lathe approaches produce visually consistent traces because of the near-normal incidence in the planar case. At position 3, however, when the angle of incidence is steepest (>160°) for the planar approach, the printed trace becomes thin and patchy in comparison to the lathe printed trace, which maintains a uniform appearance consistent with positions 1 and 2. Cross-sectional thickness profiles of these printed traces at the corresponding positions are shown in Fig. 1e, f for the planar and lathe approaches, respectively. The planar traces exhibit nonuniformity in thickness between the three positions, whereas the conformal traces produced by LAJ printing are effectively the same thickness regardless of position thanks to rotation of the substrate keeping the incident angle fixed. From this demonstration, it is shown that for successful conformal printing onto 3D curved substrates it is critical to use a system with rotational movement because a planar cartesian motion system cannot provide normal print incidence for curved features of any kind.

Translation of motion and the importance of the diametric ratio

The ratio between the substrate diameter (D_S) and the pinion diameter (D_P), referred to as the diametric ratio (D_S/D_P) is of critical design importance because it can amplify or diminish the actual motion and print speed for the substrate under the nozzle. If the two diameters are equal, the x– and y-direction movement of the printer platen is translated at a 1:1 ratio with the θ– and z-direction movement of the rotating substrate. However, if the diameters are unequal then the translation of the x-direction platen movement to the θ-direction mandrel rotation will be multiplied by the diametric ratio, as illustrated in Supplementary Fig. 3. As a result, the true displacement (d) and print speed (v) of the rotating substrate in the θ– and z-directions will differ from the movement and print speed of the planar platen according to the four equations provided in the Supplementary Information.

Although it is ideal to match the diameters of the pinion and a given substrate, it is much easier to acknowledge the difference in the two diameters, as mapped out in Supplementary Fig. 4. For the work presented here, four pinions were tested with diameters of 7.5, 15, 25, and 50 mm, and a 15 mm cylindrical mandrel (Supplementary Fig. 5) to vary the diametric ratio as illustrated in Fig. 2a. Due to the effect of the diametric ratio, it is important to adjust the design file by either elongating or shortening dimensions in the θ-direction to achieve the desired pattern. In Fig. 2b, the as-printed traces for each of the diametric ratios were imaged, and each trace is shown to be 10 mm in length as intended, despite the diametric ratio affecting the design dimensions. Since the diametric ratio will also have an effect on the rotational speed, it is important to consider how the true rotational speed of the substrate will impact printing. For the graphene traces, a linear print speed of 2 mm/s was set for the motorized platen, and the actual print speed of the rotating mandrel was determined according to the diametric ratio, as plotted in Fig. 2c for the four example cases. In this plot, the thickness of the printed traces is also provided relative to the true print speed. As would be expected, for faster print speeds at higher diametric ratios the printed traces become thinner than when they are printed at lower diametric ratios with slower print speeds. Consequently, for the thicker traces printed with the lowest diametric ratios there is substantially less sheet resistance than for the relatively thinner traces produced at higher ratio values, as shown in Fig. 2d. This trend in sheet resistance is also expected, as the thicker traces contain a higher density of conductive material and thus exhibit lower sheet resistance than the thinner traces with less material. These results (Fig. 2) highlight the importance of understanding and characterizing the effect of the diametric ratio on LAJ printing so that appropriate design choices can be made.

a Illustration of the ratio-adjusted design of lathe aerosol jet (LAJ) printed graphene traces for the four diametric ratios studied. b Microscope images of LAJ printed graphene traces using different diametric ratios, scale bar is 1 mm. c Thickness of graphene traces and rotational print speed for different diametric ratios with error bars indicating standard deviation using n = 3 samples for each ratio. d Plot of sheet resistance for graphene traces printed using different diametric ratios with error bars indicating standard deviation using n = 3 samples for each ratio.

Characterizing performance of lathe mechanism

Considering the custom nature of this lathe mechanism prototype, it was critical to determine the accuracy of movement that is provided by the translation to cylindrical-coordinate motion. To measure the accuracy achieved by the lathe system, graphene traces were printed with center-to-center pitch from 500 µm down to 20 µm in both the circumferential \((\hat\theta )\) and axial \((\hatz)\) directions with a 1:1 diametric ratio (Supplementary Fig. 6). From the observed precision between the designed and measured pitch between lines, as shown in Fig. 3, the lathe mechanism was demonstrated to have controllable and accurate movement down to 20 µm in the circumferential and axial directions. Both the achieved resolution and uniformity of the lathe printing process on a cylindrical substrate are equal to the capabilities of planar printing onto flat surfaces because parity is provided by the rack and pinion motion translation mechanism. Additionally, to show the circumferential and axial directions can also be accurately moved simultaneously, traces were printed with center-to-center pitches of 500 µm down to 50 µm in ±45° angles relative to the axial direction with a 1:1 ratio (see Supplementary Fig. 7). Lastly, to verify that the lathe system accuracy is highly repeatable and maintained throughout printing, an experiment was devised in which the word “Duke” was printed three times in rapid succession with a 1:1 ratio and overlaid in the same location each time for print speeds of 2, 4, 6, 8, and 10 mm/s. The resulting prints are shown in Supplementary Fig. 8 with videos of the lathe mechanism operating at print speeds 4–10 mm/s provided in Supplementary Movie 1. The distinct lack of shifting and the clearly legible “Duke” for each of the print speeds is indicative of the accuracy and repeatability of the lathe mechanism across multilayer and variable-speed printing.

Plotted comparison of targeted pitch to the actual measured pitch for the a circumferential and b axial directions with error bars indicating standard deviation using n = 4 samples for each data point and the dashed line representing a 1:1 relation between the targeted and measured pitch (insets are magnified plots of the smallest pitches).

Conformal and flexible electronic devices using LAJ printing

To demonstrate conformal and flexible electronics fabricated with LAJ printing, electronic devices of resistive traces, capacitors, and thin-film transistors (TFTs) were printed onto flexible substrates wrapped around cylindrical mandrels with a 1:1 diametric ratio. First, it was discovered that the resistance of infilled traces of printed graphene is dependent on the arc length (defined by the central angle and radius of the arc) and the infill direction, with axial infill yielding higher resistance at longer arc lengths, as seen in Fig. 4a. Resistance was measured in a two-terminal fashion by positioning micromanipulator probes at the ends of the long side of a trace. The two infill directions were tested by printing graphene traces along the circumference of a substrate-wrapped 15 mm diameter mandrel with different central angles of 5°, 15°, 45°, and 90° defining the arc lengths. These traces exhibited mostly similar values of resistance up to an arc length of 11.78 mm (central angle of 90°), where the resistance of the arc trace began to deviate from previous values. This deviation is due to the axial-infilled trace being printed perpendicular to the arc length direction, which uses more individual microstrip lines to form a given trace. In contrast, the circumferential-infilled trace used longer printed lines with increasing arc length, but the number of lines used to form the trace remained the same. Thus, the circumferential-infilled traces maintained a linear resistance across increasing arc lengths while the axial-infilled traces showed a jump to higher resistance for long arc lengths because it was composed of more individual lines than the circumferential infill. Therefore, the infill patterns for the materials throughout this work were chosen in accordance with having the fewest individual lines being printed for each geometry on a device layer to ensure consistent resistance.

a Resistance comparison between multiple arc lengths of graphene traces using different infill directions with error bars indicating standard deviation using n = 3 samples for each data point. b Resistance comparison between graphene traces printed on substrate-wrapped mandrels of different diameters and arc lengths. c Normalized capacitance across frequency for graphene/CNC capacitors. d Charging and discharging curves for graphene/CNC capacitors. e Fabrication process flow for LAJ printed CNT-TFTs on paper substrates on a 25 mm mandrel. f Device subthreshold characteristics of CNT-TFTs (inset is picture of devices with channel length and width of 225 and 200 µm, respectively). Data for both curves represent the average (solid line) ± the standard deviation (shaded region) for seven devices.

In Fig. 4b, resistive traces of varying arc length were printed along the circumferences of 15- and 25-mm diameter mandrels using a diametric ratio of 1:1 and circumferential infill. Regardless of substrate diameter between the two mandrels, the resistance of the traces showed an expected trend of increasing resistance with arc length. These graphene traces were then measured by the four-point probe method and a relatively consistent range of sheet resistance values was observed across increasing arc length between the two mandrels (Supplementary Fig. 9). These results indicate the reliable performance of LAJ printing along increasing circumferential dimensions while also showing scalable application to substrates with different diameters.

While successful single-layer LAJ printing of electronic materials was critical to establish, the ability to fabricate multilayer electronics is more difficult and relies on accurate patterning and positioning between layers to achieve correct functionality. Given that AJ printing has shown great potential for producing eco-friendly and sustainable devices such as all-carbon recyclable electronics⁶⁶, cellulose nanocrystal (CNC), a carbonaceous ionic dielectric material, was chosen for use in the presented multilayer electronics as an insulating layer. To demonstrate a three-layered carbon-based device, a two-plate capacitor of graphene contacts and CNC was fabricated on a paper substrate wrapped around a 15 mm diameter mandrel using LAJ printing at a 1:1 ratio (Fig. 4c, d). The normalized capacitance across frequency for the capacitors of 0.81 mm² overlapping plate area decreases in a nonlinear fashion in accordance with the ionic nature of the CNC dielectric^67,68. Furthermore, under applied voltages of 0.5, 1, and 2 V, exponential behavior of the capacitor charging and discharging currents is observed with measured leakage current for the 2 V case that diminishes for the lower voltages. These results demonstrate that simple multilayer electronic devices are readily fabricable with LAJ printing.

By expanding to four layers and using a top-gate transistor design that includes printed carbon nanotubes (CNTs) as the semiconducting layer, TFTs were fully LAJ printed on paper in accordance with the process flow illustrated in Fig. 4e. Briefly, source and drain contacts of graphene along with fiducial marks were printed onto a paper substrate mounted on a 25 mm diameter cylindrical mandrel. Then, the CNT channel was printed to bridge the two contacts followed by the CNC gate dielectric and a graphene top gate, all of which utilize the fiducial marks for correct alignment between the four TFT layers. It is worth noting that the toluene-based CNT ink used for these devices could pose some compatibility issues for substrates that display susceptibility to the toluene content, in which case, an alternative water-only CNT ink should be used to avoid any negative interactions between ink and substrate⁶⁹. The switching characteristics of these TFTs are shown in Fig. 4f, wherein the drain current (I_D) and gate current (I_g) were measured for applied gate voltages (V_GS) from −1 to 1 V with a constant drain to source voltage (V_DS) of −0.5 V. The measured CNT-TFT devices possess repeatable and consistent switching behavior with an on/off-current ratio of ~10² and a full sweep range of only 2 V due to the ionic nature of the CNC gate dielectric drastically lowering the threshold voltage requirements. These devices have similarly low threshold voltage requirements (V_T < 1 V) and comparable device current ratios (I_ON/I_OFF: 10²–10³) to previous examples of all-carbon printed TFT devices on paper despite the presented devices having no added ions or treatment to the wrapping polymer in the CNT channel⁶⁶. However, these devices also suffer from a relatively high level of gate leakage current (I_g > 1 nA) that is a notable problem of devices with ionic dielectrics. The successful printing of these three-terminal, multilayer devices further corroborate the versatility and accuracy of LAJ printing across multiple layers and for different materials and substrates. LAJ printing is demonstrably capable of fabricating electronic devices onto nonplanar surfaces for conformal and flexible electronic applications alike.

Conformal printing onto complex shapes with LAJ printing

From the demonstrations already presented, it is clear that LAJ printing achieves ideal orthogonal print incidence for conforming onto cylindrical bodies and flexible substrates bent around a convex radius. Coincidentally, without any modification to the actual lathe mechanism, the LAJ printing system also allows the print nozzle to be positioned at the focal point of concave curvature, as shown in Fig. 5a for a 15 mm diameter hollow half cylinder. In Fig. 5b, c, example patterns of graphene were deposited onto the concave mandrel with and without a fixed substrate to show that this setup can also be used to print onto flexible substrates bent into a compressive state at a concave angle.

a Picture of AJ print nozzle fixed at the focal point above a concave mandrel. Pictures of printed graphene in b a meander line onto concave mandrel and c parallel lines on concavely bent paper substrate. d Picture of conical mandrel with printed AgNW traces. e Pictures of LAJ printing along the length of conical mandrel showing movement up the length of the cone after 5 s of printing. Pictures of f printing a graphene helix on 1000 µL pipette tip and g a magnified view of graphene helix on the pipette tip.

Despite offering a simple and efficient method for conformal printing onto cylindrical curvature, the lathe-based approach still has limitations when working with multidirectional curves like spherical surfaces and nonuniform 2D curvatures. For complex geometries like these, the current maneuvering capabilities of the lathe prototype cannot achieve the ideal normal print incidence of 90°, and it may be better to consider systems with advanced functionality like the 5-axis printer. However, because of the standoff distance and jet-stream deposition of AJ printing, getting approximate uniformity in a printed trace is still possible for moderate angles of non-normal incidence. This tolerance of moderately non-normal incidence in conformal applications is demonstrated for LAJ printing in Fig. 5d, e by printing silver nanowires (AgNWs) onto a 3D-printed conical mandrel with a taper angle of 15°, resulting in a non-normal print incidence of 105° across the cone. For different substrates (e.g., cylindrical, conical, flat), the effects of printing with non-normal (oblique) incidence are identical and determined by the angle of print incidence from the nozzle position relative to the substrate⁴³. In situations where the angle of incidence becomes increasingly oblique, like the example in Fig. 1 of printing onto a cylinder using planar movement, it was observed that thickness and uniformity of the printed films will quantifiably reduce. Oblique incidence will similarly impact cylindrical printing onto conical substrates, resulting in negligible effect for small taper angles but forming thinner and less defined printed traces for large tapers with extreme obliquity. As a further example of the adaptability of conformal AJ printing for small angles of obliquity, a 1000 µL pipette tip with a gradual taper of 7.5° was used as the substrate for a conformal graphene helix, which was printed spiraling down the length of the pipette tip from its narrow end as imaged in Fig. 5f, g. These examples serve to represent the tolerance of LAJ printing to gradual multidirectional curvature with its simple motion, however, there will still be limitations to the conformal complexity that can be realized without adding further axes of control.

Catheter balloon functionalized with on-surface LAJ printed graphene sensor

Successfully fabricating conformal electronics onto 3D curvilinear substrates like cylinders is a non-trivial challenge that LAJ printing is well-suited to solve in a simple and readily scalable fashion. However, AM of conformal electronics becomes increasingly difficult for substrates like catheters and catheter balloons, which possess soft polymeric surfaces and non-rigid inflatable bodies that add complication to the print process. Although difficult to achieve, conformal fabrication of electronics directly onto catheter substrates is an important step in advancing catheter device technology for use in critical medical applications ranging from cancerous tissue ablation^56,58 to implanting long-term devices within the body^70,71. During these crucial medical procedures catheters are integral, yet knowledge of conditions around the catheter such as temperature, contact/degree of inflation, and external pressure are practically unknown without built-in electronic functionality.

To enable the fabrication of conformal electronics onto catheters and catheter balloons, LAJ printing was modified to include a system for on-axis inflation that permits hollow and inflatable rotational bodies to be mounted on the lathe mechanism, as portrayed in Fig. 6a. Since the on-axis inflation required no other modification to the lathe mechanism beyond the inclusion of a rotating air inlet, the system capabilities for printing onto any substrate <40 mm in diameter hold true, even for inflated catheters. In Fig. 6b, examples of six commercially available catheter balloons are plotted according to the possible balloon diameters for the product and their medical application space^{72,73,74,75,76,77}. Also indicated in this plot is the 14 mm diameter thermoplastic polyurethane (TPU) catheter balloon used in this work, as shown in Fig. 6c, which is sized to be used for endoscopic applications⁷⁸. These catheter dimensions compared to the compatible range of the lathe system make evident that the lathe mechanism presented in this work can be utilized for essentially any sized catheter regardless of medical application.

a CAD diagram and picture of lathe mechanism showing the pathway for pumped air to achieve on-axis catheter balloon inflation. b Comparison plot of commercial catheter balloon diameters relative to lathe system range, which can support up to 40 mm diameter. c Picture of 14 mm diameter inflatable catheter balloon used for demonstration and relevant for endoscopy procedures. d Fabrication process for graphene sensor on catheter balloon. e Picture of LAJ printing graphene meander line sensor around the circumference of inflated catheter balloon. Plots of normalized resistive change of printed graphene sensor f versus temperature showing sensitive response and polynomial modeling with high R² fit and g over time as the catheter balloon is inflated and deflated illustrating the ability of the sensor to monitor inflation.

The six-step fabrication process flow for LAJ printing onto the inflated catheter balloon is detailed in Fig. 6d. The first pretreatment step is critical for cleaning the balloon substrate and improving the surface adhesion properties of the polymeric material to ensure good print quality. Then, after mounting the catheter and printing graphene traces onto it, an encapsulating layer of polydimethylsiloxane (PDMS/silicone) is dripped over the printed material for a few reasons: to ensure printed films do not delaminate from the surface during inflation/deflation, to provide a biocompatible coating over the chosen printed material, and also to reduce noise for printed sensors.

Using this fabrication process, a graphene meander line sensor was printed directly onto the surface of the inflated 14 mm TPU catheter balloon, as shown in Fig. 6e and Supplementary Movie 2. Graphene was selected for this sensor demonstration because of the environmentally sensitive material properties it offers, which cause a resistive change in the printed graphene based on temperature, stress/strain, and humidity⁷⁹. Given that the graphene sensor is encapsulated by PDMS, humidity sensitivity is removed from the sensor leaving behind sensitivities to temperature and stress/strain. To demonstrate the value of having a functionalized catheter, the normalized resistive change of the graphene sensor was measured relative to temperature on the catheter balloon from 25 to 80 °C, which encompasses physiological body temperatures as well as higher temperatures that are important to consider when heating tissue for procedures like catheter ablation. From the results in Fig. 6f, the sensor provides a resistive change that is well modeled by the quadratic equation given in the figure with an R² value of 0.987.

Using the best fit quadratic equation, the average normalized percent change in resistance can be calculated for any temperature in the modeled range, which can then be used to estimate the sensitivity of the sensor. For example, the sensitivity of the sensor is calculated as \(-0.42\frac \% ^\circ \rmC\) in the body temperature range of 35–42 °C whereas the sensitivity for 42–80 °C is \(-1.48\frac \% ^\circ \rmC\). This difference in sensitivities is due to the quadratic nature of the resistive response to temperature, however, future designs will modify the graphene sensor to be most sensitive in the body temperature range in addition to reducing measured noise in the data caused by contact inconsistencies. These improvements will be achieved through optimized design tuning of the print parameters controlling graphene density in the film as well as the overall thickness while also revising the contact method for measurement to eliminate erroneous noise signals presenting in the data from the simple contact approach.

Furthermore, because the graphene sensor also exhibits a response to mechanical movement independent of temperature response, the sensor could be used for the secondary function of detecting inflation and deflation of the catheter balloon in real time as shown in Fig. 6g. When inflated, the graphene sensor on the catheter balloon is held taut in its originally printed position and shows little change from the initial resistance value. However, when the catheter balloon is deflated, the graphene sensor crumples up with the balloon surface and the sensor drops into a low-resistance state with a distinct 18% change from the initial resistive value. Drift of the resistive value for the inflated state is attributed to mechanical forces acting on the printed graphene film during the initial inflation/deflation cycles. Through collapse and expansion, the thin-film graphene eventually evolves into a more stable morphology and an approximate steady state is reached. For the graphene sensor on the catheter surface, drift of the inflated state resistive value R₀ was characterized in Supplementary Fig. 10, which shows that the sensor had an initial R₀ of ~8.9 kΩ that reached a relatively steady state value of ~10.6 kΩ after 18–20 cycles. In Fig. 5g, inflated and deflated states were detected during cycles 13–20 with a drift of ~1.5% in the inflated resistance value, which is smaller than the resistive change due to deflation and less than the drift during the wearing-in cycles of a new sensor. With LAJ printing this sensor with two optional functionalities was trivial to fabricate directly onto the catheter balloon such that no secondary substrate or additional processing was required. The capabilities of lathe-based AJ printing coupled with the modularity of the lathe mechanism ensures that this method of DW printing can be used for virtually any catheter without requiring interaction or modification of the mechanism.

We have developed a mechanism for actively translating cartesian-coordinate motion to cylindrical-coordinate motion for use in the AM of conformal electronic devices, specifically by using AJ printing. This lathe-like mechanism in combination with AJ printing enables fully controlled and high-resolution fabrication of thin-film materials and electronic devices onto 3D rotational shapes that are difficult to accommodate without using more complex 5-axis motion systems, which are significantly expensive and difficult to scale. The most unique application of this mechanism is demonstrated by printing flexible electronic sensors directly onto an endoscopic catheter balloon. Functionalized catheters are very difficult to produce, and although there are existing AM methods that have demonstrated fabrication of catheter electronics, the lathe-based approach presented here is low-cost, efficient, and simply designed for easy scaling to commercial production demands. Additionally, given the modularity of the mechanism design, LAJ printing can be employed to produce conformal electronic devices in a number of versatile applications including for any commercially available catheter or catheter balloon.