Introduction

In modern engineering and medical fields, the demand for highly adaptable, lightweight, and efficient pump systems has become increasingly urgent1,2,3,4,5. Traditional rigid pumps, while effective in certain applications, suffer from significant drawbacks such as bulkiness, noise, and inflexibility. These limitations hinder the development of advanced systems in soft robotics, wearable devices, and precise temperature control, where seamless integration and adaptability are paramount.

The growing demand for flexible pumps has led to the exploration of two primary categories: mechanical structure pumps and non-mechanical structure pumps. Mechanical structure pumps, often employing smart materials with self-deformation capabilities and compliance6, have seen significant developments. These pumps utilize various activation methods, such as thermally responsive7, electrically responsive8,9,10, magnetically responsive11, and optically responsive materials12, to achieve high flow outputs relative to their size, either through increased response frequencies or significant chamber deformations. However, the fixed nature of their deformation chambers and external structures often makes these pumps rigid and limits their performance adaptability, failing to meet the dynamic needs of advanced systems.

Conversely, non-mechanical structure pumps, which generate responses directly within the fluid medium, offer innovative alternatives. These pumps utilize techniques such as chemical reactions13,14, temperature-induced phase changes15, and electro-hydrodynamics (EHD)16,17,18,19 to achieve seamless integration and enhanced functionality within soft systems. Each of these methods, however, presents its own set of challenges: chemical reactions can be uncontrollable and irreversible, temperature changes are slow and difficult to regulate, and EHD requires intricate designs of electrodes and channels to optimize performance.

In the realm of EHD, early enhancements involved the use of rigid materials and complex 3D electrode structures like triangular prism-to-slit electrodes20, needle-to-ring electrodes21, and disk electrodes22. While effective, these methods were inherently inflexible. Some studies also utilized flexible 3D electrode structures. For example, Tang et al.17,18 used silicone molds and 3D-printed conductive TPU to create flexible 3D needle-ring electrode structures, offering rapid responsiveness and expandable configurations. However, the inherent design of needle-ring electrodes affects the flexibility of the flow channel. Excessive bending within the channel can disrupt the alignment between the 3D electrodes, thereby diminishing the pump’s effectiveness.

Recent advances have shifted towards 2D flexible planar electrode configurations, enhancing deformability. Vito Cacucciolo et al.16 proposed using interdigitated electrodes with conductive silicone and printed silver ink to fabricate dual-layer integrated pumps, achieving flexible bending and stretchability in EHD pumps. Mao et al.23 explored the impact of different structural parameters of interdigitated electrodes on the output performance of EHDs. Zhu et al.24 used 2D-printed metal circuits to create a range of planar electrodes, facilitating multidirectional driving capabilities in EHD pumps. Despite these advancements, the typical rectangular cross-sectional flow channels of these designs are prone to bulging under high pressure, which can detract from their performance.

To address these limitations, Michael Smith et al.19 created a 3D spiral structure by winding copper wires and TPU strings around circular rods. This design’s high integration, lightweight, compact size, and superior output performance, along with the fiber structure’s ability to maintain functionality when bent or coiled, result from its fabrication method, which synchronously processes electrodes and flow channels. However, relying on copper wires as electrodes limits exploration into varied electrode forms. Continuous innovation in electrode designs remains necessary to enhance the versatility and applicability of EHD pumps in various technological and industrial applications.

Here, we propose a simple and straightforward fabrication method for customizable soft fiber pumps (CSFPs) that affords tunable electrode configurations, adjustable cross-sectional geometries, and reconfigurable overall morphologies. This strategy simplifies manufacturing and broadens EHD pump functionality by delivering higher output, greater voltage tolerance, and broader design flexibility. It also allows CSFPs to be tailored to meet specific performance goals. In detail, this method offers several significant advantages: 1) Flexible electrode design. By transforming planar electrodes into 3D helices, our method expands the design space, allowing us to explore how variations in line width and electrode geometry affect performance and to direct flow both along and around the pump axis. Using this capability, we built a cyclone CSFP and operated it as a liquid separator, demonstrating improved fluid dynamics and purification efficiency. 2) Customizable cross-section. Thermoplastic molding allows us to vary both diameter and shape: Reducing the diameter lifts the pressure gradient to 1.39 kPa/cm, about 15.8% higher than the previous record, whereas changing the cross-section raises the specific flowrate to 160 mL/min/g, an improvement of 29.8%. This flexibility also allows the pump profile to be matched to curved or planar heat-exchange surfaces, broadening its practical utility. 3) Re-shapeable overall pump structure: The pump can be reheated and pre-bent into tight radii without kinking, preventing flow blockage or electrical shorting. We have applied the CSFP to drive soft actuators and to cool hot surfaces, confirming its value in applications that benefit from post-fabrication shape customization.

Results

Working principle and fabrication method

Figure 1a shows the EHD driving principle, applying high voltage across electrodes creates an electric field that charges fluid molecules, propelling the fluid molecules in a directed flow between the electrodes25. The Eq. (1)25 provides a general description of the pressure generated between two plane-parallel electrodes:

$$P(x)-{P}_{0}=\frac{\varepsilon }{2}\cdot (E{(x)}^{2}-{E}_{0}^{2})$$
(1)

where\(P\) is the pressure within the fluid, \(E\) is the electric field inthensity, \(\varepsilon\) is the permittivity. The intensity of the electric field within the flow channel affects the pump’s output pressure. Thus, variations in the electrodes’3D structure can notably alter this intensity, thereby influencing pump performance.

Fig. 1: Working principle and fabricating process of the CSFP.
figure 1

a A schematic illustration depicts the operating principle of the CSFP. Liquid molecules ionize at the negative electrode by accepting electrons. These charged particles not only move toward the positive electrode but also drag other liquid molecules along, creating a directional fluid flow toward the positive electrode. b Foundational procedures of CSFP, including electrode manufacturing, electrode winding, and thermoplastic molding. In step 3, the electrode film melts onto the tube, fusing the electrode and the tube into a unified structure. c Modifications enhancing the functionality of CSFP. (i) Electrodes can be crafted into various shapes. (ii) The diameter of the CSFP can be precisely changed by additional heating in any section. (iii) The cross-sectional geometry of the CSFP is adjustable. (iv) Overall structural deformation of the CSFP can be achieved and repeated through controlled heating. d The Actual Structure of CSFP (i) A CSFP heating-deformed into a herial structure. (ii) The copper electrodes printed on a piece of EVA film and embedded within the walls of a PE tube.

Full size image

We propose a simplified and customizable soft fiber pump (CSFP) fabrication method that utilizes straightforward steps to achieve tunable electrode configurations, modifiable cross-sectional geometries, and customizable overall morphologies (depicted in Fig. 1b, detailed in Methods and Supplementary Movie S1). The first step, electrode printing, involves screen printing conductive copper ink onto an Ethylene Vinyl Acetate Copolymer (EVA) film, which is then cut into strips to form the winded electrodes. Compared to etching26, flexible circuit printing24, and laser cutting27, screen printing offers advantages such as rapid processing, low cost, and high precision of electrodes designed to meet the specific requirements of different applications. In the second step, electrode winding, the cut electrodes are winded around a metal central stick, facing inward, and encased in a PE tube. The third step, thermoplastic molding, heats this assembly, causing the tube to contract and the electrode film to adhere to the inner tube wall, simultaneously establishing the electrode placement and constructing the flow channel. This process culminates in the initial CSFP structure, which can be further refined through repeated applications of heat-shrinking to enhance output performance.

This method not only simplifies the production process but also enhances the pump’s adjustability at multiple levels (in Fig. 1c). Firstly, it allows for customizable electrode configurations through precision screen printing, enhancing CSFP performance by enabling innovative electrode designs. Secondly, it allows for precise adjustments to the pump’s diameter and cross-sectional shape by replacing the central stick and applying heat, improving the pump’s adaptability for practical applications such as cooling systems. Lastly, the use of a thermoplastic PE tube enables the final structure to be thermally set into diverse forms, such as curls and bends, allowing use in complex environments. This streamlined process leverages readily accessible materials and tools, offering significant flexibility and customization compared to traditional fabrication methods.

The completed CSFP assembly is shown in Fig. 1d(i) with a helical structure deformed post-heating. A microscopic examination (Fig. 1d(ii) and Supplementary Fig. S1) reveals that the copper-printed EVA film has completely melted onto the PE tube wall, ensuring a secure bond that facilitates synchronous deformation and further manipulations like bending and twisting.

Exploration of pump electrode configurations

Utilizing the electrode design ability, we investigated how different electrode shapes influence the internal electric-field intensity of the pump and its corresponding output performance. Three electrode configurations were selected based on established designs from prior EHD pump research16,19,23. The simple parallel line electrodes (named as string electrodes), electrodes with unilateral protrusions, and interdigitated electrodes. These configurations are illustrated in Fig. 2a, with actual CSFP samples shown in Supplementary Fig. S2. We chose these electrode structures specifically to examine how incorporating microstructures, such as unilateral protrusions, or altering electrode patterns into an interdigitated arrangement, affects pump performance relative to a baseline parallel-electrode geometry. To ensure reliability, each data point in Fig. 2 represents averaged measurements from three independently fabricated pumps. Additionally, Novec 7100 fluid from 3M, known for its low conductivity and viscosity and widely employed in high-voltage fluidic applications25,28, was used as the working liquid to maintain consistent and robust pump operation throughout the experiments. Our experiments, shown in Fig. 2b, indicated that string electrodes perform best at lower voltages, providing the highest output pressure at 7 kV. By contrast, unilateral protrusion electrodes outperform at higher voltages, sustaining operation up to 13 kV and achieving a maximum pressure of 10.5 kPa, thereby offering broader voltage adaptability. To understand the impact of electrode shapes on pump performance, we utilized COMSOL simulations to model the coupling of electric and flow fields within the pump. The simulations are based on the theoretical model presented in Supplementary Text 1, and details are discussed in Supplementary Text 2 and Supplementary Fig. S7. We obtained steady-state results at different voltage levels applied to the CSFP and plotted them as a dot-line graph in Fig. 2b. The simulation results in Fig. 2b align closely with our experimental data. In Fig. 2c, the axial electric field intensity generated by different electrode shapes correlates with the observed trends in pressure output. This demonstrates that variations in axial electric field intensity, dictated by electrode shape, are directly correlated with the observed differences in pressure output.

Fig. 2: The impact of electrode variations on pump performance.
figure 2

a Three types of designed electrodes. b Pressure outputs from different electrodes under varying voltages, based on simulations and experiments. c Simulated axial mean electric field intensity for the three electrode types, demonstrating a correlation with the trends observed in output pressure. d The diagram of line width and line spacing parameters in the string electrode. e Pressure changes for different string electrode line spacings in CSFPs with a 0.5 mm line width in the electrode, a diameter of 1.5 mm, and a length of 80 mm in the pump. f Pressure changes for different string electrode line widths in CSFPs with a 1.0 mm line spacing in the electrode, a diameter of 1.5 mm, and a length of 80 mm in the pump. g Correlation between output pressure changes and electric field strength variations at 7 kV DC voltage with different line widths. h Flowrate changes at different string electrode line spacings in CSFPs with a 0.5 mm line width in the electrode, a diameter of 1.5 mm, and a length of 80 mm in the pump. i Flowrate changes for different string electrode line widths in CSFPs with a 1.0 mm line spacing in the electrode, a diameter of 1.5 mm, and a length of 80 mm in the pump.

Full size image

Focusing on string electrodes, we explored structural parameters such as electrode line spacing and line width (in Fig. 2d) to understand their effects on the CSFP’s pressure and flowrate performance. We maintained an electrode winding angle of ~30°, achieved by fixing the total electrode strip width at 5 mm and winding it around a central stick with a diameter of 3 mm. And the resulting pump had an inner diameter of 1.5 mm and a length of 80 mm. Figure 2e shows the effect of line spacing when the line width is fixed at 0.5 mm. At an applied voltage of 7 kV, the pressure peaks at 7.6 kPa when the spacing is 0.8 mm. Reducing the spacing intensifies the electric field and can enhance the pressure, but it also increases the risk of dielectric breakdown and electrode damage. At voltages appreciably higher than 7 kV, these failure risks dominate, and the pressure no longer improves. Given this trade‑off between performance and reliability, we adopted a line spacing of 1.0 mm for all subsequent experiments to ensure stable operation at elevated voltages.

Figure 2f shows that, with the total electrode width fixed at 5 mm and a constant line spacing of 1.0 mm, a line width of 0.5 mm yielded the highest performance: an 80-mm-long CSFP (inner diameter 1.5 mm) reached a maximum pressure of 7.5 kPa at 7 kV. According to Eq. (1), this pressure rise correlates directly with the local electric-field intensity. COMSOL simulations performed at various line widths closely matched the experimental results, confirming a clear relationship between pressure output and axial electric-field intensity (Fig. 2g). With these geometric constraints, varying the electrode line width simultaneously influenced the electric field within each electrode pair and the spacing between adjacent pairs. For widths below 0.5 mm, incremental widening enhanced the forward-directed electric field, raising the pressure. Conversely, beyond 0.5 mm, further widening reduced the spacing between pairs, amplifying an opposing field component, lowering the net electric-field strength, and thus decreasing pressure output. These results indicate that carefully tuning electrode geometry effectively controls the electric field and consequently the pump’s performance.

According to Poiseuille’s law, in Eq. (2):

$$Q=\frac{\pi {r}^{4}\Delta P}{8\eta L}$$
(2)

The output flowrate Q is directly related to the pressure \(\varDelta P\) differential within the pump when the radii of the pump \(r\) viscosity \(\eta\) and the length of the \(L\) are settled. In other words, the higher the output pressure of the CSFP, the greater the flowrate. The experimentally tested flow output results are consistent with the anticipated trend in flow changes. As shown in Fig. 2h, at varying line spacings, smaller electrode spacings result in higher flow outputs, with a 0.8 mm spacing producing a flowrate of 35 mL/min under 7 kV; In Fig. 2i, when changing line widths, the 0.5 mm line width electrode produced the maximum flowrate of 38 mL/min under 7 kV.

Exploration of pump cross-sectional geometries and overall morphologies

In the electric field generated by the 3D electrode, changes in the 2D electrode shape affect the electric field, as do changes in the axial cross-section shape and size. Focusing on CSFP, we explored the impact of inner diameter size on output performance. Maintaining the electrode winding lean angle at 30°, electrode line spacing at 1.0 mm, line width at 0.5 mm, and pump length at 80 mm, we tested the cylindrical fiber pumps under various inner diameters for output pressure and flowrate. In Fig. 3, data points represent the mean value obtained from three independent pumps, with error bars indicating standard deviation. As shown in Fig. 3a, an inner diameter of 0.8 mm produced the highest output pressure, reaching 11.1 kPa under a 5 kV input voltage. This is due to the reduction in inner diameter, effectively decreasing the spacing between the electrodes radially, thereby increasing the electric field intensity and consequently the output pressure. However, reducing the inner diameter also decreases the flowrate, thus, under a 0.8 mm diameter, the pump produced the minimum flowrate of 7.8 mL/min (Fig. 3b).

Fig. 3: Impact of different cross-sectional geometries and overall morphologies on pump functionality.
figure 3

a Pressure-voltage curves for CSFPs with varying diameters, constructed with string electrodes featuring 1.0 mm line spacing and 0.5 mm line width, and a pump length of 80 mm. b Generated flowrate-voltage curves of different pump diameters. c Pressure-flowrate characteristics of a baseline-CSFP structure under varying applied voltages, with linear fitting results included. This CSFP model utilizes a string electrode design with 1.0 mm line spacing and 0.5 mm line width. The pump measures 80 mm in length and 1.5 mm in diameter. d, e Demonstrate how variations in cross-sectional shapes influence pressure and flowrate under a 7 kV voltage. The CSFPs, all constructed with string electrodes and fundamental parameters, exhibit experimental trends that align with simulation results. Cross-sectional geometries tested include a circular shape with a diameter of 1.5 mm, and rectangular shapes with dimensions of 1.3 mm × 1.8 mm and 1.0 mm × 2.0 mm. f Demonstrates how thermoforming prevents CSFP’s clasp under small-radius folding. g Two different small-radius bends with radii of 3.0 mm and 2.0 mm. h Pressure-flowrate characteristics of CSFP under small radius bending, showing minimal impact on pump performance. i Pressure-flowrate characteristics of CSFPs subjected to repeated 3.0 mm radius bending, illustrating performance degradation with increased deformation cycles.

Full size image

Based on previous research into pump fabrication parameters, we used string electrodes with a spacing of 1.0 mm and a line width of 0.5 mm to construct a baseline-CSFP with an inner diameter of 1.5 mm and a length of 80 mm. Under a 7 kV input, it produced a maximum output pressure of 7.5 kPa and a maximum flow rate of 38 mL/min. The pressure-flowrate linear fitting curves at different input voltages are shown in Fig. 3c with a volume of only 0.424 cm³ and a weight of 0.3 g.

In addition to changeable cross-section sizes, the cross-section can also be deformed in two ways: one is by replacing the central stick and thermoplastic shrinking into the designed dimensions, and the other is by directly pressing the softened condition onto a hard object’s surface during heating, causing external deformation of the shape. Taking rectangular cross-sections as examples, we fabricated pumps with dimensions of 1.3 mm × 1.8 mm and 1.0 mm × 2.0 mm, respectively. Figure 3d, e shows the maximum output pressure and flowrate measured experimentally under the same structural and operating conditions with a supplied voltage of 7 kV, alongside simulation results. The data indicates that changing the cross-section shape similarly alters the pump’s output performance.

When a tube is bent to a small radius, its wall tends to collapse and fold. This cross-sectional flattening can restrict the flow and, in the CSFP, may also let the embedded electrodes touch, causing a short-circuit. We therefore treat tight bending as a coupled deformation that involves both curvature and cross-section shape. By locally heating the CSFP, we impose the required bend while at the same time reforming the cross-section so that it stays open. After cooling, the PE tube locks in this pre-shaped geometry, preventing collapse and preserving electrical isolation during operation (Fig. 3f). To evaluate how bending affects the CSFP, we formed right-angle bends using molds with radii of 3 mm and 2 mm, respectively (Fig. 3g). In Fig. 3h, the performance of CSFP after bending does not decrease, still maintaining normal pumping operation, demonstrating the CSFP’s capability to function under deformed conditions. Furthermore, the CSFP can be reheated after fabrication and bent or coiled. After cooling, the pump maintains the bent shape, and upon reheating, the pump automatically returns to a straight configuration. This process can be repeated more than 4 times, and the breakdown possibility will also develop as time increases (Fig. 3i).

Demonstration of pump driving performance

To demonstrate the CSFP’s driving performance, we used the pump to independently drive a single-direction bending soft finger, and measured the internal pressure of the finger cavity with a pressure sensor, as shown in Fig. 4a and Methods. The pump utilized in our experiment was a baseline-CSFP, featuring string electrodes with a spacing of 1.0 mm and a line width of 0.5 mm, constructed with an inner diameter of 1.5 mm and a length of 80 mm. Figure 4b and Supplementary Movie S2, Supporting Information, show the deformation process of the finger under a 7 kV voltage, the finger reached its maximum bending angle, with the cavity’s internal pressure at 4.48 kPa. The inner pressure’s variation at changing voltage can be seen in Fig. 4c. Further expanding on the pump’s capabilities, we engineered a bi-directionally bendable finger using silicone, integrating the pump and actuator seamlessly. The pump we utilized for actuation was similarly configured as a baseline-CSFP, as described earlier. This configuration, illustrated in Fig. 4d and demonstrated in Supplementary Movie S3, Supporting Information, allowed the pump to transfer fluid between cavities on either side of the actuator under deformed conditions. By reversing the voltage polarity, we could change the pump’s driving direction, enabling bi-directional bending. This experiment showcased the pump’s capability to manage both positive and negative pressures and to adapt fluid flow directions, enhancing its utility with soft actuators. The results demonstrate the CSFP’s compact structure, flexibility, and bi-directional driving capability.

Fig. 4: Actuation application using CSFP.
figure 4

a Integration of the CSFP with a silicone finger, monitored using a pressure sensor. b Process of one-directional deformation of the silicone finger driven by CSFP. c Variation in pressure within the silicone finger in response to changing voltage. d Bi-directional bending of a silicone finger, which was linked to a single CSFP looped into a circle. The pump, controlled by a sinusoidal signal, reversed the flow direction when the voltage was reversed, resulting in bidirectional bending of the finger.

Full size image

Demonstration of pump cooling cycle function

Previous studies have frequently used EHD pumps for active thermal regulation16,19,29, and in such applications, the device’s conformity to the target surface plays a key role in cooling performance. Therefore, we conducted a comparative experiment to demonstrate the impact of thermoplastic deformation on the CSFP’s performance during the cooling process (Supplementary Movie S4, Supporting Information). To increase the flowrate, we designed connectors to parallel four pumps, connecting the pumps to a 6 kV DC voltage and putting the pumps beneath the heating plate to achieve a better observation of the measurement point’s temperature changing on the heating plate (in Fig. 5a and Methods). This setup enabled the pumps to extract solution from a pool and recirculate it back through the pump body, achieving a cooling effect. In the experiment, we placed the system on a 70 °C flat plate for cooling, stopping when the temperature dropped to 55 °C. Figure 5b demonstrates the CSFP’s ability to deform under pressure and revert to its original shape with heating, this unique characteristic lays the foundation for significant functional improvements in the cooling system.

Fig. 5: Cooling application with CSFPs.
figure 5

a The cooling system uses CSFPs. Four-parallel-connected CSFPs have been used to cool down a heating plate, and a measurement point has been set on the plate. b A circle and non-circle cross-section deformation and recovery loop. c Temperature-time curves at measurement points under different cooling conditions. d The cooling process of CSFPs with different cross-sections. (i) The baseline-CSFP with a round cross-section showing limited cooling area attachment, in a time of 70 s, cooling a plate from 54.5 °C to 35.4 °C. (ii) After thermoforming, the cross-section was altered to a square, expanding the attached cooling area, and enhancing the cooling efficiency from 57.6 °C to 30.1 °C in the same period, marking a 44% increase in efficiency.

Full size image

We used a temperature camera for a real-time detection of the plate’s temperature change. The temperature variation of the measurement point has been shown in Fig. 5c. Initially, with the pump’s cross-section being circular and placed directly on the flat plate (Fig. 5d(i)), the contact area was minimal and likely to have gaps, resulting in an insignificant cooling effect. The temperature only decreased to 35.4 °C, and the cooling was uneven. To enhance the pump’s cooling performance, we heated and softened it, then pressed its outer wall to change its cross-section to one with higher adherence to the flat plate (Fig. 5d(ii)). After reshaping, the pump achieved a cooling temperature down to 30.1 °C with a more uniform cooling area. The outcome demonstrates that the thermoplastic deformation of CSFPs improves their cooling efficiency by 44%, underscoring the practical benefits of our design in real-world applications.

Demonstration of cyclone separation system using CSFP

In the past few decades, hydrocyclones have been adopted in both heavy and light industries30, because their simple, motion-free geometry efficiently separates fine solid or immiscible liquid impurities. This ease of operation has recently prompted their use in biotechnology and food processing31. The CSFP can replicate cyclone behavior through the modification ability of electrode shape, enabling orientation changes beyond the axial direction. Specifically, we positioned the interdigitated electrodes parallel to the axis to induce a spiral flow, effectively transforming the CSFP into a cyclonic separator. Figure 6a and Supplementary Movie S5 in the Supporting Information illustrate the basic model of cyclone CSFP design. The CSFP operates with Novec 7100 fluid, commonly used for cleaning and cooling electronic devices. It is crucial that the fluid remains free of impurities to prevent damage to sensitive electronic components. As such, we enhanced the cyclone structure to demonstrate its capability as an effective separator (as shown in Fig. 6b): When activated, the CSFP generates a spiral flow causing floating impurities to descend while the cleaner liquid centrifuges upward, overflowing from the pump walls. We also integrated this cyclonic CSFP separator with a baseline-CSFP, enabling active enhancement of the cyclone effect and allowing for the control of the overflow rate through voltage adjustments. Figure 6c depicts our proposed design for an advanced separation system utilizing CSFP technology. In Fig. 6d, we established a simplified system to validate our design, setting the activation voltage for both the pump and cyclone CSFP separator at 2.5 kV. Supplementary Movie S6 in the Supporting Information demonstrates the system’s effective separation capabilities. With both floating and precipitated impurities introduced into each part of the separation system, the CSFP separator successfully purified the liquid and directed it into a designated reservoir on the right, confirming its functionality as a posterior separator in a contaminated system.

Fig. 6: Cyclone separation using CSFP.
figure 6

a Basic model of cyclone CSFP. Illustrates the basic model with (i) interdigitated electrodes wound parallel to the axis, enhancing cyclonic action, and (ii) displays the actual construction of the Cyclone CSFP, measuring 6 mm in diameter and 20 mm in length. b Cyclone CSFP separator model. The cyclone CSFP separator generates a spiral flow when powered, causing the liquid to spin and overflow from the pump’s walls while driving floating impurities downward. c Conceptual integrated separation system. This proposed system combines a baseline-CSFP with a cyclone CSFP, allowing for precise control over liquid pumping and overflow speeds. It includes a storage tank and supports the cleaning or cooling of electronics. d Simplified separation system for validation. The system operates with both floating and precipitated impurities in the fluid. This system functions with the baseline-pumping CSFP and the Cyclone CSFP separator under an applied pressure of 2.5 kPa, ensuring effective separation.

Full size image

Discussion

In this study, we developed a simple and straightforward fabrication method for Customizable Soft Fiber Pumps (CSFP), featuring tunable electrode configurations, adjustable cross-sectional geometries, and customizable morphologies. The method streamlines production and extends the functional range of EHD pumps, delivering higher output, greater voltage tolerance, and wider geometric adaptability. A survey of every fiber-format pump19,32, reported to date (Supplementary Fig. S3, data in Supplementary Table S1) shows that the CSFP achieves high pressure and flowrate while offering new levers for further optimization. Reducing the fiber diameter lifts the pressure gradient to 1.39 kPa/cm, a 15.8% increase over the previous record; changing the cross-section raises the specific flowrate to 160 mL/min/g, a 29.8% improvement. Adding unilateral-protrusion micro-structures pushes the breakdown limit to 13 kV and provides an additional boost in peak pressure. These findings indicate that the CSFP can be customized to meet different performance goals: the baseline-CSFP with string electrodes offers an overall balance of pressure and flowrate; to maximize pressure, a smaller diameter (0.8 mm) design with unilateral protrusion electrodes is preferred; and to achieve higher flowrate, a rectangular cross-section proves most effective. Beyond these core metrics, the CSFP’s bidirectional drive operates reliably under large deformation, and its reconfigurable geometry improves surface conformity in heat-exchange applications. Re-orienting the printed electrodes enables swirling flow, demonstrated by a cyclone CSFP separator that removes particulates from contaminated liquid. All devices are built from off-the-shelf materials at a cost below $0.10 per pump (Supplementary Table S2), substantially lowering the barrier to future EHD pump research and deployment.

In previous experiments, the CSFPs were only tested at a length of 80 mm. Due to the manual fabrication process, individual variations are unavoidable, and longer pump lengths may lead to greater discrepancies. Currently, multiple pumps can be connected in series to achieve higher output pressure. In the future, this limitation can be addressed through standardized and mechanized fabrication, effectively facilitating the scaling up of CSFPs. Although we have comprehensively analyzed the performance improvements attributed to variations in the string electrode parameters, the complexity of the impacts will increase with more complicated electrode structures like the unilateral protrusion pattern. A thorough investigation of electrode structures through extensive, systematic experiments is still required and will eventually culminate in a guideline for designing electrodes based on different functional needs. However, beyond the design of electrode structures, the operational efficiency and longevity of these pumps are currently limited by electrode degradation, which impacts both rigid33, and flexible electrodes18, thereby diminishing the pumps’ performance and lifespan. The CSFP can work over 24 h, demonstrating its ability for long-term operation (as shown in Supplementary Fig. S4). Inevitably, the performance of the CSFP gradually decreases during the pumping process. Under high voltage, copper electrodes also experience electromigration, leading to electrode aging. Additionally, due to manual fabrication, inconsistencies in electrode quality can result in weak points and defects that are more susceptible to breakdown under high voltage, causing electrode damage. Longer lifetimes will therefore depend either on automated, defect-free production or on more durable conductors. Protective coatings of corrosion-resistant polymer34, or gold35, and the self-healing electrode networks recently demonstrated in resilient EHD pumps36, both represent promising strategies for extending pump longevity. By integrating such measures, future CSFPs should combine greater durability with true roll-to-roll scalability, opening the way to broader use in soft robotics, thermal management, and biomedical pumping.

Methods

Electrode printing section

Fabrication of various electrode types begins with processing the screen-printing plate. Detailed parameters of the designed electrodes are displayed in Supplementary Fig. S5. EVA films with 0.1 mm thickness (Dongguan Langtaosha Electronics Co., Ltd.) are cut to appropriate sizes and screen printed with conductive copper ink (Shenzhen DELO Technology Development Co., Ltd.). These films are heated at 85 °C for 30 min to strengthen the ink-film adhesion. The heated films are then cut into 5 mm wide strips and fitted with copper tape at both ends to facilitate circuit connections.

Electrode winding section

Electrode strips (printed with string or unilateral protrusion patterns) are wound around a 3 mm diameter metal stick in a helical shape, with the strip’s head secured using tape to ensure edge-to-edge attachment, preventing overlap. This process forms a spiral structure with an approximate lead angle of 30 degrees and a pitch of about 5.7 mm, resulting in roughly 175 loops per meter. The interdigitated electrode is designed to wrap entirely around the pump’s circumference, fitting snugly around the central stick. An 80 mm piece of PE tube is then placed over the electrode to create a pre-structure for the subsequent molding step. To achieve the final inner diameter of 1.5 mm, a tube that shrinks to half its original size (3 mm) is selected. Once the electrode strip width and central stick diameter are chosen, the winding lead angle is determined accordingly.

Thermoplastic molding section

This step involves assembling the two main components of the CSFP: the electrode and the PE tube (Shanghai Niushi Electric Co., Ltd.). Considering the pump’s small size, a heat gun set between 90 °C and 100 °C is used for heating. Moving the heat gun slowly ensures even shrinkage of the tubing. The film melts onto the tube wall, forming the initial CSFP structure.

Experimental procedures of tests

Our testing system, detailed in Supplementary Fig. S8, Supporting Information, includes the pump, a flowrate sensor (AFD4100 flow sensor from ASAIR), a pressure sensor (MIK-Y290 produced by Asmik), a fluid reservoir, a throttle valve (G55100-ec1from Guangzhou Mifanxi Trading Co., Ltd.), and a high-voltage DC power supply (PA1020 from Pintech) capable of delivering up to 20 kV. The working media we used is a solution of Novec 7100 (3 M). The system operates in three modes. With the throttle closed, no fluid passes, allowing pressure measurement by the sensor. With the throttle fully open, the system is at atmospheric pressure, and the flowrate sensor records the pump’s maximum flowrate. Adjusting the throttle provides an external load, enabling simultaneous recording of pressure and flowrate behaviors. To present a more reliable performance of the CSFP, we averaged the flowrate over 20 s for each data point and the pressure over 5 s.

Finger motivation

The soft fingers that were presented in the demonstration were both made of Ecoflex 00-10 (Smooth on) and fabricated by mould-casting. In the one-direction bending test, a pressure sensor integrated into the finger monitored bending pressure. The bending experiments utilized a high-voltage amplifier (Trek 610C) controlled by a signal generator (UTG2060b from UNI-T). For one-direction bending, a sloped signal with a 30-cycle was used. The bi-directional bending experiment employed a sinusoidal signal with a maximum voltage of 6.5 kV and the same 30 s cycle.

Cooling process

A specially designed parallel connection, fabricated using a Stratasys J750 3D printing machine, was employed to link four pumps together, effectively enhancing the flowrate. This connection included a strip of copper tape to ensure all pumps were simultaneously connected to the electrical circuit. An infrared temperature camera (Fotric 616C-L30-M50) was used for real-time temperature detection. The heating plate of 60 mm width and 80 mm long from Jiangyin Huachenqi Electric Heating Appliances Co., Ltd., was initially set to 90 °C, and the cooling pumps were activated once the temperature decreased to 70 °C. The system was powered at an applied voltage of 6.5 kV.

Separation system with cyclone CSFP

The fabrication process of the Cyclone CSFP is detailed as follows: Initially, the designated electrodes are screen printed (the interdigitated electrode structure is depicted in Supplementary Fig. S6). The electrodes are then wrapped around a 6 mm central stick with the copper layer facing inward. Next, an 8 mm diameter, 20 mm long thermoplastic tube is placed over the electrodes, and the assembly is heated to cause the tube to shrink, ensuring the electrodes adhere tightly to the tube’s interior. After the heating process, the central stick is removed, completing the Cyclone CSFP assembly. For the separator configuration, the basic Cyclone CSFP is coupled with a silicone tube measuring 9 mm in diameter and 30 mm in length. Both components are mounted on a specially designed base that features liquid inlet and outlet ports at the bottom for the flow of substances. For the separation trials, we utilized polystyrene particles with a diameter of 0.3 mm (from Shanghai Tuomiao Biotechnology Co., Ltd.), and silica particles ranging from 0.1 mm to 0.3 mm in diameter (provided by Henan Hengnuo Co., Ltd.). This setup was employed to evaluate the efficacy of the cyclone CSFP in particle separation under controlled conditions.