Bioinspired Vacuum Generation via Pressure-to-Vacuum Conversion for Manipulating all Phases of Matter

Abstract

Animal diaphragm-lung systems are soft organs that generate a controllable vacuum. Elephants, as rare land animals, can manipulate all three states of matter using their lung-generated vacuum. In soft robotics, however, current vacuum generation relies on rigid components, and no single soft device effectively handles all states of matter. Traditional soft pumps and grippers are limited in scope: soft pumps provide continuous liquid flow but cannot directly manipulate solids, while grippers manage solids but are ineffective with liquids and gases. Inspired by lung functionality, we present a soft pressure-to-vacuum converter that provides precise control over the suction, holding, and release of solids, liquids, and gases through a single entry and exit point based on negative pressure. Through the selection of appropriate material properties and design variations, our soft device achieves vacuum levels up to −18 kPa, enabling intermittent control and sequential handling of various media without the need for additional components. We demonstrate diverse applications of our soft device, including artificial lungs, liquid blending, vacuum gripping, coffee preparation, and liquid-gas vaporization. This bioinspired device not only provides a safe and adaptable solution for vacuum generation but also addresses a critical gap in soft robotics, offering a multifunctional system capable of manipulating all states of matter.

Keywords

soft robotics soft vacuum generators artificial lung multi-state manipulator

Introduction

Vacuum has numerous applications in engineering,¹ medical,^2–4 and soft robotics^5–10 fields. However, in soft robotics, we currently rely on traditional rigid components to generate vacuum,^5–10 which underscores the necessity for softer alternatives to enhance safety. Animal lungs^11,12 and muscular suction cups of octopuses^13–15 are examples of biological vacuum generators. In mammals, the collaboration between soft organs—diaphragms and lungs—acts as a soft vacuum generator. Lungs create negative pressure (a vacuum) for inhaling air and then generate positive pressure for exhaling air.^11,12,16,17 Elephants are rare land animals that use the controllable vacuum in their lungs not only for breathing (gas) but also for drinking water (liquid) and delicately manipulating objects (solid),^11,18 effectively manipulating all three states of matter (see our observation in Supplementary Movies S1 and S2).

Existing soft robotic devices are limited in handling multiple states of matter.^19,20 Soft grippers, like suction cups, use rigid vacuum components to manipulate solids but are ineffective with liquids and gases.^{7,10,19,21–28} Soft pumps, on the other hand, generate continuous liquid flow using positive pressure with separate inlets and outlets.^29–40 However, robotic applications often require more than just continuous flow; they demand intermittent control, such as suction, holding, and release of liquids. Soft pumps are generally not designed for this and, most importantly, are incapable of handling solids directly and are also ineffective with gases alone. For example, the elephant’s heart acts as a steady soft pump for continuous blood flow, whereas its lungs function as versatile manipulators, capable of sucking, holding, and releasing materials across different states. Just as the heart cannot directly manipulate solids or gases, traditional soft pumps are limited in scope.

In the future, to assist with all daily tasks, robots need a single soft device capable of skillfully manipulating solids, liquids, and gases simultaneously.²⁰ Inspired by lung functionality, we present a bioinspired soft device that generates controllable negative pressure (vacuum) to directly handle solids, liquids, and gases through a single port. Unlike conventional soft grippers or pumps, which are optimized for handling only a single state of matter, our device functions as a versatile manipulator. This adaptability supports applications requiring periodic suction and release, especially in tasks that involve sequential handling of different states of matter.

In this study, first, we investigated the fundamentals and detailed the thermodynamics of vacuum creation. Using these foundational principles, we designed soft pressure-to-vacuum converters—hereafter referred to as soft vacuum generators—which consist of soft diaphragms and a soft vacuum chamber, although they rely on rigid components for pneumatic actuation. We then investigated the effects of various material properties and geometrical factors on vacuum generation through simulations. Subsequently, we developed a lung-inspired vacuum generator and assessed its capability to manipulate all three states of matter. Furthermore, we introduced a double-acting soft vacuum generator that can produce both vacuum and air jets in single-diaphragm actuation. This dual function allows for the simultaneous blowing of solid particles and vacuum suction.

Finally, we showcase the diverse applications of our soft vacuum generator, serving as artificial lungs, a soft liquid manipulator, gripper, and gas manipulator. Its safety and versatility make it invaluable across engineering, medicine, and the rapidly expanding domain of soft robotics.

Materials and Methods

Expansion of gas in a soft body as a means of generating vacuum

During the process of inhalation, animals experience a slight vacuum due to the expansion of air in their lungs. Conversely, during exhalation, the opposite occurs, as illustrated in Figure 1A and 1B. In a closed system, vacuum is generated when the gas pressure inside a chamber decreases from $p_{1}$ ( $p_{atm}$ ) to $p_{2}$ as the volume expands from $v_{1}$ to $v_{2}$ . This vacuum is given by:

p_{vacuum} = p_{2} - p_{atm} = p_{atm} ({(\frac{v_{1}}{v_{2}})}^{\frac{c_{p}}{c_{v}}} - 1)

(1)

FIG. 1.

Vacuum creation inside a soft body. (A) Lungs in animals expand and relax, (B) producing both negative pressure (vacuum) and positive pressure (where $p_{alv}$ is alveolar pressure, $p_{atm}$ is atmospheric pressure, and $v_{L}$ is the volume of breath). (C) Lung volumes in animals vary from milliliters to liters. (D) A plot of the vacuum pressure ( $p_{vacuum}$ ) against the final volume ( $v_{2}$ ) of a soft body, considering various initial volumes ( $v_{1}$ ) ranging from milliliters to 5 liters and an initial pressure of 1 atmosphere, illustrates how initial and final volumes influence the resulting vacuum. (E) Schematic of soft vacuum generator, (F) fabricated soft vacuum generator before actuation, and (G) after actuation.

Here, p, v, and $\frac{c_{p}}{c_{v}}$ represent gas pressure, volume, and the heat capacity ratio (for air, $\frac{c_{p}}{c_{v}}$ = $\frac{7}{5}$ at normal room temperature), respectively (for detailed information, refer to Supplementary Section S1 in the Supplementary Data).

Equation 1 highlights how the final volume ( $v_{2}$ ) and the initial volume ( $v_{1}$ ) influence the vacuum output. For better understanding, we plot a graph of final pressure, $p_{2}$ , (indicating vacuum) against final volume for various initial volumes that correspond to the residual lung volume in certain animals (Fig. 1C). Assuming a maximum volume change of 5 liters ( $v_{2} - v_{1} = 5$ L), Figure 1D depicts the resulting graph, demonstrating the change in output pressure $(p_{vacuum})$ concerning volume change. Vacuum generation increases with increasing final volume and volume change, particularly as the initial volume approaches zero. Hence, maximum vacuum output requires a substantial volume change ( $v_{2} ≫ v_{1}$ ) from the smallest initial volume ( $v_{1} \to 0$ ). However, the decrease in vacuum performance with increasing initial volume—when the volume change is fixed—is not a fundamental limitation. Systems with larger initial volumes can maintain or even enhance vacuum generation by proportionally increasing the volume change. Therefore, optimizing the expansion ratio is an essential consideration in the design of soft vacuum systems across different size scales.

A multilayered soft vacuum generator inspired by mammalian lungs and diaphragms

Figure 1E, shows our soft pressure-to-vacuum converter, inspired by mammalian lungs and operated pneumatically. Several actuation methods were considered, including tethered mechanical pulling, shape memory alloys, and dielectric elastomer actuators. However, these alternatives were less suited for replicating the smooth, isotropic, and cyclic volume changes required to actuate the soft membrane. Pneumatic actuation, despite requiring an external pump, offers uniform, tunable deformation and better mimics the natural expansion–contraction behavior. Constructed with silicone rubbers, our soft device comprises three layers, each serving a specific function. The top two layers are bonded closely, forming an airtight diaphragm with interconnected microchannels known as diaphragm cavities, facilitating pneumatic actuation. The bottom layer, connected circumferentially to the second layer, forms the vacuum chamber (refer to Supplementary Fig. S6 and Supplementary Data S4).

Introducing compressed air into the diaphragm cavities causes the diaphragm to move away from the third layer (Fig. 1F and 1G), increasing the air volume inside the vacuum chamber from initial volume ( $v_{1}$ ) to final volume ( $v_{2}$ ). This creates a decrease in pressure, mimicking vacuum generation in animal lungs. Deflating the diaphragm restores atmospheric pressure within the vacuum chamber.

Our design intentionally keeps the diaphragm close to the third layer, minimizing the initial volume ( $v_{1}$ ) of the vacuum chamber, crucial for achieving maximum vacuum output (Fig. 1D).

Results and Discussion

Maximizing the final volume of vacuum chamber to improve vacuum

We first investigated the influence of the shape changes of the diaphragm on the final volume of the vacuum chamber by analyzing two disc-shaped diaphragms: SHS and HSH. Both diaphragms have two disc-shaped layers, with Layer 1 at the top and Layer 2 at the bottom. They are made from functionally graded materials (refer to Supplementary Fig. S6). SHS diaphragm features a stiff inner annulus (Dragon skin 30, $E_{1} \approx$ 600 kPa) and a softer outer annulus (Ecoflex 10, $E_{2} \approx$ 55 kPa) with a stiffness ratio of $E_{1} / E_{2}$ = 11 (Fig. 2A) along its radius. In contrast, the layers in diaphragm HSH have the opposite stiffness configuration, with a Dragon skin 30 (D30) outer annulus ( $E_{2} \approx$ 600 kPa) and an Ecoflex 10 (E10) inner annulus ( $E_{1} \approx$ 55 kPa), yielding a stiffness ratio of $E_{1} / E_{2}$ = 0.09 (see Fig. 2A). However, the middle spongy part, skin, and soft fibers of both diaphragms are composed of Ecoflex 50.

FIG. 2.

Diaphragm characteristics and deformation. (A) A schematic diagram showing Layer 1 (the same for Layer 2) of two diaphragms, SHS (violet) and HSH (pink), with two distinct annuli composed of Ecoflex 10 (E10) and Dragon Skin 30 (D30). (B) Simulation results obtained after inflating both diaphragms, SHS and HSH, at 150 kPa. (C) Deformed cross-sectional profiles for SHS and HSH obtained from both experiments (violet for SHS and pink for HSH) and finite element (FE) analysis (black). (D) Snapshots of the SHS diaphragm during inflation. (E) The curvature ( $k$ ), and (F) radius of curvature ( $k^{- 1}$ ) for the first half of the deformed cross-sectional profiles of SHS and HSH.

Based on finite element analysis (FEA) at a pressure of 150 kPa,^41–43 we observed that the outer annulus (softer part) of SHS deforms more than its inner annulus, while the inner annulus of HSH deforms more than the outer annulus (Fig. 2B). Consequently, despite having equivalent global strains along the diameter ( $ϵ_{SHS} = 16 %$ and $ϵ_{HSH} = 15 %$ , where $ϵ$ represents linear strain), both diaphragms produced significantly different shapes under the same pressure due to differences in local strains along the diameter (refer to Supplementary Section S2 and S3 for more details).

We further experimentally verified the shape change of both diaphragms at $\sim$ 150 kPa, depicted in Figure 2C, with the violet curve representing SHS and the pink curve representing HSH. The experimental results closely match the simulation. We then estimated revolution volume for both diaphragm profiles by parameterizing the cubic Bézier function as a curve (changing the parameter r from 0 to 1 generated half of these curves, each symmetrical about the axis passing through the center of the disc) and found that the final volume of the SHS profile is approximately 1.25 times greater than that of the HSH profile. Figure 2D illustrates an example of snapshots taken before and after the actuation of SHS, clearly showing that the outer annulus is deformed more than the inner annulus (see Supplementary Movie S3), and the shape produced is similar to that generated by an FEA model.

In Figure 2E, we show the curvature rates (k) for both deformed diaphragms, plotted over half of each profile (r from 0 to 1). SHS exhibits a positive curvature, indicating a convex-shaped profile, while HSH shows a curvature that transitions from negative to positive, indicating a change in curvature shape from concave to convex. Further analysis of the radius of curvature ( $k^{- 1}$ in Fig. 2F) implies that the SHS profile has a significantly larger radius around its outer (r = 0) and inner (r = 1) areas but remains almost consistently positive in between. In contrast, the HSH profile has a negative radius of curvature between r = 0 and r = 0.5, then turns positive between r = 0.5 and r = 1. The larger area under the curve and revolution volume of the SHS profile suggests that positive curvature and positive radius with significantly larger values at the circumference and at the center are preferred over negative curvature or negative radius to maximize final volume and thus the vacuum output.

Motivated by the aforementioned results, we conducted further experimental investigations into the initial and final volumes of the vacuum chambers for two soft vacuum generators, constructed using two different diaphragms: SHS and HSH. These vacuum generators will be referred to as VSHS and VHSH, respectively. The experiment involved inflating the diaphragms using compressed air while allowing stored water from a syringe to enter the vacuum chamber simultaneously via gravity (refer to Supplementary Fig. S8).

As depicted in Figure 3A and 3B, the final volume of VSHS (22 mL) is approximately 1.29 times larger than that of VHSH (16.5 mL). This ratio aligns with the output predicted by FEA (1.25 times), despite both diaphragms producing an approximate height of 21 mm. There is a slight difference in the initial volume ( $\sim$ 1 mL) of the two vacuum generators, which can be attributed to imperfections introduced during vacuum generator fabrication. Lastly, based on the volume ratio of VSHS and VHSH and using equation 1, we predict that the ratio of the final pressure inside VHSH to that of VSHS should be around 1.14. Therefore, VSHS is expected to produce more vacuum than VHSH.

FIG. 3.

Vacuum produced by the vacuum generators. The initial and final volumes of the soft vacuum generators (A) VSHS and (B) VHSH, determined by a water experiment. (C) The measurement of vacuum within the VSHS (light pink) and VHSH (dark pink) vacuum generators using pressure sensors after the diaphragms were inflated (violet). (D) Measurement of vacuum within VSHS under quasi-static conditions. The final pressure within the VSHS vacuum generator under two conditions: (E) with varying outlet tube lengths and (F) with different outlet tube opening diameters (0.5, 1.6, 2.4, 3.8 mm) during varying heights of diaphragm inflation (DE is diaphragm expansion, and DR is diaphragm relaxation).

Next, we investigated the vacuum generation capacity of both soft vacuum generators by inflating the diaphragms of VSHS and VHSH together to an average input pressure of 130 kPa (the violet-shaded areas in Fig. 3C), while simultaneously monitoring pressure changes in both vacuum chambers using pressure sensors (Supplementary Fig. S9). In Figure 3C, we show that as the input pressures in both diaphragms exceed atmospheric pressure 0 kPa (violet region), the pressure inside the vacuum chambers of VSHS (light pink) and VHSH (dark pink) drops below atmospheric pressure (0 kPa), creating a vacuum and reaching peak values, approximately −18 kPa and −8 kPa, respectively. Deflating the diaphragm restored the chamber pressure to atmospheric levels. Even though both diaphragms received the same input pressure, VSHS generated 125% more vacuum (approximately −18 kPa) due to its larger final volume achieved through the stiffness configuration, compared to VHSH (approximately −8 kPa). Note that the vacuum in our device can be controlled by the deformation of the diaphragm, similar to lungs, which is very useful for manipulation tasks.

The vacuum output in SHS, shown in Figure 3C, illustrates a dynamic scenario in which the diaphragm is pressurized almost instantaneously. It takes approximately 1.5 s to reach a stable vacuum level, followed by a vacuum recovery over $\sim$ 3 s after the diaphragm is abruptly depressurized. This behavior is likely influenced by viscoelastic effects in the diaphragm material, as well as airflow resistance introduced by the narrow input tubing and the porous structure of the diaphragm. To further investigate the pressure–vacuum relationship under quasi-static conditions, we conducted a test using a triangular pressure waveform. This was applied using a pressure regulator (ITV0030-3BS, SMC) driven by a triangular signal generated by a waveform generator (33622A, Keysight Technologies). The results of this test are presented in Figure 3D. Together, these experiments provide a comprehensive view of the performance of the soft vacuum generators under both dynamic and quasi-static operating conditions.

The impact of length of outlet tube on the vacuum

In our design, the volume of the outlet tube also contributes to the initial volume (refer to Supplementary Fig. S10B), and therefore, we investigated the impact of extending the length of the vacuum outlet tube on vacuum output. For this, we connected the vacuum chamber of VSHS to an outlet tube with a constant diameter of 1.8 mm but with varying lengths (L = 55, 110, 165, and 220 mm). We then inflated the diaphragm of the VSHS vacuum generator to a height of 15 mm while simultaneously monitoring the pressure variations within the chamber. The results, depicted in Figure 3E, revealed that tube with the shortest length (L = 55 mm) produced the highest vacuum (around −7 kPa), whereas tube with the longest length (L = 220 mm) generated the least vacuum (around −2 kPa). These observations suggest that the length of the outlet tube could impact the output vacuum.

The impact of opening diameter of the outlet tube on the vacuum

So far, vacuum measurements have been performed by operating our soft vacuum generator as a closed system. Next, we investigated the vacuum levels when it operated as an open system, allowing atmospheric air to enter the vacuum chamber. For this, first, we attached an outlet tube with a diameter of 1.8 mm and a length of 55 mm to the vacuum chamber of VSHS. Subsequently, we affixed reducing direct connectors with various diameters to the tip of the tube, ranging from 0.5 mm to 3.8 mm (refer to Supplementary Fig. S10C), to adjust the opening diameter as needed.

Figure 3F illustrates the change in pressure in the vacuum chamber during diaphragm expansion (with a height of 13.5 mm) and relaxation for various opening diameters. The results show that the tube with the smallest tip diameter, i.e., 0.5 mm, produced the highest vacuum, reaching approximately −2 kPa. However, as the diameter increased from 0.5 mm to 3.8 mm, the level of vacuum produced gradually decreased and approached zero. This phenomenon occurs due to the difference in airflow rate ( $\dot{m} = ρ A V$ , where $\dot{m}$ is the mass flow rate, $ρ$ is the density, A is the cross-sectional area, and V is the velocity of the air), which is a function of the cross-sectional area of the tubes (Supplementary Fig. S10C). A narrower outlet diameter introduces greater fluidic resistance, analogous to electrical resistance in a circuit where pressure corresponds to voltage and flow to current. This increased resistance limits the inflow of atmospheric air, thereby sustaining a pressure difference across the vacuum chamber for a longer duration and resulting in a higher measurable vacuum.

Moreover, Figure 3F reveals a fascinating discovery: with a tip diameter of 0.5 mm, the pressure inside the vacuum chamber exceeds atmospheric pressure (0.4 kPa) at the end of diaphragm relaxation (black rectangular blocks in Fig. 3F), due to the compression of air that entered the chamber. This finding suggests the possibility of compressing air within the vacuum chamber and releasing it as an air jet—a concept similar to how elephants expel air from their lungs to manipulate objects that are out of reach.¹⁷

Double-acting soft vacuum system for active air jet and vacuum simultaneously

The preceding results inspired us to create a double-acting soft vacuum generator, as depicted in Figure 4A–D, capable of concurrently producing both vacuum and air jets using only one diaphragm. Despite having only one diaphragm, our design incorporates two vacuum chambers. Chamber 1, positioned above the diaphragm in Figure 4A, possesses a small initial volume. In contrast, the chamber 2, located below the diaphragm, is larger and has a substantial initial volume. Figure 4E shows the pressure changes in both chambers as they operate as a closed system. This plot highlights three key features: First, when the diaphragm is actuated (p $\approx$ 40 kPa), the pressure in chamber 1 decreases, resulting in an approximate vacuum of −2.5 kPa due to air expansion. Simultaneously, the pressure within chamber 2 increased by approximately 6 kPa from atmospheric levels, as a result of air compression (Fig. 4B). This is indicated by the pressure change between lines ⓐ and ⓑ in Figure 4E. Second, when we opened the outlet of chamber 2, compressed air from chamber 2 was expelled as an air jet (Fig. 4B), resulting in a pressure decrease (Fig. 4E). Furthermore, upon simultaneous diaphragm relaxation, the pressure in chamber 1 returned to atmospheric pressure. At the same time, the pressure in chamber 2 decreased further, resulting in a vacuum effect (achieving −5 kPa) due to air expansion (Fig. 4C). This is demonstrated by the transition from lines ⓑ to © in Figure 4E. Third, as the diaphragm returned to its original position (Fig. 4D), the pressure in both chambers tended to revert to its initial level, as shown between lines ©, and ⓓ in Figure 4E.

FIG. 4.

Double-acting soft vacuum generator. (A) A schematic diagram of a double-acting soft vacuum system with one diaphragm and two chambers. (B) Actuation of the diaphragm causes air in chamber 1 to expand (vacuum), and air in chamber 2 to compress, resulting in an air jet. (C) Diaphragm relaxation causes pressure recovery in chamber 1, and creation of a vacuum in chamber 2, (D) followed by re-admission of air into chamber 2. (E) A plot of pressure against time for a double-acting vacuum generator shows a simultaneous increase in pressure in chamber 2 and a decrease in chamber 1 during diaphragm actuation, and vice versa during diaphragm relaxation.

These findings demonstrate how a single diaphragm can simultaneously produce vacuum and an air jet by coupling two chambers in a mechanically antagonistic arrangement. While this is not a direct imitation of biological lungs—which operate these functions sequentially—it achieves a bioinspired functionality (refer to Fig. 1B) through engineered agonist—antagonist behavior. This approach could enable versatile applications in various fields of engineering and robotics.

Next, we showcase the versatility of our soft vacuum generator by manipulating all three states of matter: liquid (varying viscosity), solid (various size and shape), and gas.

Soft vacuum generator as liquid manipulator

In Figure 5A and 5B, we present a soft liquid manipulator designed to suction-feed liquid from a glass container (Supplementary Movie S6). Figure 5C illustrates that liquids, such as soft drinks, can flow through a 1.8 mm diameter tube at a speed of approximately 40 cm/s, with the diaphragm operating at a speed of around 10 cm/s. It should be noted that the reported flow speed reflects the combined performance of the vacuum generator and the upstream pneumatic actuation system, including the pressure source, tubing, and valve dynamics. Compared with traditional soft liquid pumps, which are designed for continuous liquid flow,^29–34 our controllable vacuum-driven approach offers precise liquid manipulation, capable of suction, retention, and delivery of liquids at variable speeds (refer to Supplementary Movie S7 for liquid blending), regardless of their viscosity (see Supplementary Movies S8 and S9 for the handling of high-viscosity liquids, yogurt, and honey), making it highly suitable for robotic applications.

FIG. 5.

A soft liquid manipulator. (A, B) An example of soft vacuum generator drawing liquid (soft drink) from a glass container by generating a vacuum. The inset shows a high-viscosity liquids, honey and yogurt, both of which can be transported using the vacuum generator. (C) A plot of the flow speed of soft drink versus the upward speed of the diaphragm.

A soft vacuum generator for manipulating solids and gas

In Figure 6A and Figure 6B, we further illustrate the ability of a double-acting vacuum generator to handle solids by attracting, holding, and releasing a puffed grain (Supplementary Movies S10 and S11). Furthermore, Figure 6C demonstrates that increased pressure in chamber 2 creates air jets, propelling small solid particles (Supplementary Movie S12), showcasing interactions beyond normal reach. This process mirrors the behavior of elephants using their lungs: they use air jets to move objects beyond their grasp and vacuums to capture small objects.^11,17 In Figure 6D–G (Supplementary Movie S13), we further showcase coffee preparation, employing vacuum to handle various states of matter such as sugar (small particles) and coffee (liquid), while air jets (gas) assist in delivering them into the glass, all using a single soft device. Finally, in Figure 6H, we demonstrate the conversion of liquid from the reservoir into white fumes using the vacuum effect, and then subsequently expel it as jet fumes into the atmosphere (Fig. 6I and 6J) (Supplementary Movie S14). These practical applications underscore the versatility of our soft device spanning all three states of matter compared to existing material handling soft devices.

FIG. 6.

Soft vacuum generator that grips and blows. (A, B) Double-acting vacuum generator gripping puffed grain using vacuum. (C) An illustration of the pressure change in chamber 2 shows how an air jet created during diaphragm actuation can blow solid particles (beyond reach), and the vacuum created during diaphragm relaxation can lift solids. (D) The vacuum generator picks up sugar, (E) delivers it to the coffee glass, (F) then picks liquid coffee and (G) delivers it to the coffee glass. (H) The vacuum generator activates a heating device, converting liquid from the reservoir into white fumes (gas), which it then inhales. (I, J) The fumes are subsequently exhaled as jet from the soft chamber into the atmosphere.

Conclusions

This study draws inspiration from the remarkable ability of elephants to manipulate all three states of matter through lung-generated vacuum. Inspired by this, we developed soft pressure-to-vacuum converters capable of handling solids, liquids, and gases, offering secure and versatile solutions for soft robotics. Our 95 mm diameter vacuum generator produced a vacuum of approximately −18 kPa at a diaphragm pressure of 130 kPa when used as a closed system. This level of vacuum is suitable for handling gases, liquids, and tiny solids.

Our soft vacuum generator offers several advantages over existing material handling systems in soft robotics: (i) unlike soft grippers or soft pumps, which can handle only a single state of matter, our soft device can manage all states, either individually or in combinations; (ii) while soft grippers struggle with handling tiny solids, our device excels at managing small particles; (iii) unlike most soft pumps, which are primarily engineered for liquid pumping, our device can manipulate liquids by drawing, holding, and releasing them as needed; (iv) our double-acting vacuum generator, capable of creating both vacuum and an air jet, facilitates interactions with materials that are out of reach, akin to the unique behavior found in elephants.¹⁷

In the last section of our study, we explored the diverse applications of these soft vacuum generators. In summary, the proposed bio-inspired soft vacuum generator offers a soft robotics solution for manipulating solids, liquids, and gases. Its performance across diverse tasks demonstrates its potential utility in various soft robotics applications involving all three states of matter. Scalability is an important consideration, and upcoming work will investigate how vacuum performance can be preserved across different size scales. We will also pursue formal optimization of material properties and diaphragm geometry to further improve vacuum performance.

The current system relies on multiple rigid solenoid valves to control different physical states and actions. While effective, this setup increases bulk and limits integration. To address this, upcoming designs will explore replacing rigid components with soft valves or fluidic circuits to reduce hardware complexity and improve compliance. Nonetheless, achieving fully untethered operation remains a significant challenge due to the need for onboard pressure and power sources. Promising directions include replacing the pneumatic diaphragm with electrically driven actuators, such as dielectric elastomer membranes, and integrating compact battery-powered pneumatic pumps. These approaches will be investigated to progressively enhance autonomy.

Authors’ Contributions

R.C. and B.M. conceptualized the lung-inspired soft vacuum generators and B.M. supervised the study. R.C. designed, fabricated, and conducted experiments on soft vacuum generators. A.M. provided various technical assistance and insights required for building experiment setups and conducting analyses. R.C., A.M., and B.M. collectively contributed to writing the article.

Footnotes

Acknowledgments

The authors thank Kottur Elephant Sanctuary in Thiruvananthapuram, Kerala, for assisting in observing native Asian elephants in the Western Ghats of India. We express our gratitude to our colleague C. Filippeschi (Istituto Italiano di Tecnologia) for the technical assistance provided during the fabrications. LLM tools are utilized for grammar and spelling checks in the article content. This work has received support from the European Union’s Horizon 2020 research and innovation program under grant agreement no. 863212 (PROBOSCIS project).

Author Disclosure Statement

The authors declare no financial or nonfinancial competing interests.

Funding Information

No funding was received for this article.

Data Availability

Data required to support the conclusions of this article are included in the main text or the . The datasets generated during and/or analyzed in this study are available from the corresponding author upon reasonable request.

Supplemental Material

References

Colligon

. A history of vacuum technology from 5th century bc and a personal view of progress in the field, including medical applications, from 1959 to 2021. Vacuum, 2022; 198:110821.

Putnis

, Khan

, Wong

JM-L

. Negative pressure wound therapy—a review of its uses in orthopaedic trauma. Open Orthop J, 2014; 8(1):142–147.

Marangi

, Segreto

, Morelli Coppola

, et al. Management of chronic seromas: A novel surgical approach with the use of vacuum assisted closure therapy. Int Wound J, 2020; 17(5):1153–1158.

Novelli

, Daleffe

, Birra

, et al. Negative pressure wound therapy in complex cranio-maxillofacial and cervical wounds. Int Wound J, 2018; 15(1):16–23.

Jiao

, Ji

, Zou

, et al. Vacuum-powered soft pneumatic twisting actuators to empower new capabilities for soft robots. Adv Mater Technol, 2019; 4(1):1800429.

Robertson

, Paik

. New soft robots really suck: Vacuum-powered systems empower diverse capabilities. Sci Robot, 2017; 2(9):eaan6357.

Song

, Drotlef

D-M

, Son

, et al. Adaptive self-sealing suction-based soft robotic gripper. Adv Sci, 2021; 8(17):2100641.

Coutinho

, Rodrigue

. Fluidic hardware strategies for powering combined negative- and positive-pressure artificial muscles. Adv Eng Mater, 2023; 25(14):2300071.

Choi

, Kim

, Lee

, et al. Soft, multi-dof, variable stiffness mechanism using layer jamming for wearable robots. IEEE Robot Autom Lett, 2019; 4(3):2539–2546.

10.

Sui

, Zhu

, Zhao

, et al. A bioinspired soft swallowing gripper for universal adaptable grasping. Soft Robot, 2022; 9(1):36–56.

11.

Schulz

, Ning Wu

, Ha

SYS

, et al. Suction feeding by elephants. J R Soc Interface, 2021; 18(179):20210215.

12.

Carrier

. Ventilatory action of the hypaxial muscles of the lizard iguana iguana: A function of slow muscle. J Exp Biol, 1989; 143(1):435–457.

13.

Kier

, Smith

. The structure and adhesive mechanism of octopus suckers. Integr Comp Biol, 2002; 42(6):1146–1153.

14.

Tramacere

, Beccai

, Kuba

, et al. The morphology and adhesion mechanism of octopus vulgaris suckers. PLoS ONE, 2013; 8(6):e65074–e7.

15.

Tramacere

, Kovalev

, Kleinteich

, et al. Structure and mechanical properties of octopus vulgaris suckers. J R Soc Interface, 2014; 11(91):20130816.

16.

De Troyer

, Boriek

. Mechanics of the respiratory muscles. Compr Physiol, 2011; 1(3):1273–1300.

17.

Mizuno

, Irie

, Hiraiwa-Hasegawa

, et al. Asian elephants acquire inaccessible food by blowing. Anim Cogn, 2016; 19(1):215–222.

18.

Dagenais

, Hensman

, Haechler

, et al. Elephants evolved strategies reducing the biomechanical complexity of their trunk. Curr Biol, 2021; 31(21):4727–4737.e4.

19.

Shintake

, Cacucciolo

, Floreano

, et al. Soft robotic grippers. Adv Mater, 2018; 30(29):1707035.

20.

Xian

, Zhu

, Xu

, et al. Fluidlab: A differentiable environment for benchmarking complex fluid manipulation. In: Proceedings of the Eleventh International Conference on Learning Representations. ICLR: 2023; doi: 10.48550/arXiv.2303.02346

21.

Kuwajima

, Shigemune

, Cacucciolo

, et al. Active suction cup actuated by electrohydrodynamics phenomenon. In: Proceedings of the 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems. IROS; 2017, pp. 470–475.

22.

Baik

, Kim

, Park

, et al. A wet-tolerant adhesive patch inspired by protuberances in suction cups of octopi. Nature, 2017; 546(7658):396–400.

23.

Xie

, Domel

, An

, et al. Octopus arm-inspired tapered soft actuators with suckers for improved grasping. Soft Robot, 2020; 7(5):639–648.

24.

Kim

, Coutinho

, Park

, et al. Octopus-inspired suction cup array for versatile grasping operations. IEEE Robot Autom Lett, 2023; 8(5):2962–2969.

25.

Takahashi

, Suzuki

, Aoyagi

. Octopus bioinspired vacuum gripper with micro bumps. In: Proceedings of the 11th IEEE Annual International Conference on Nano/Micro Engineered and Molecular Systems. NEMS; 2016, pp. 508–511.

26.

Zhou

, Ren

, Chen

, et al. Bio-inspired soft grippers based on impactive gripping. Adv Sci (Weinh), 2021; 8(9):2002017.

27.

Kim

, Kim

, Cho

, et al. Electrothermal soft manipulator enabling safe transport and handling of thin cell/tissue sheets and bioelectronic devices. Sci Adv, 2020; 6(42):eabc5630.

28.

Koivikko

, Drotlef

D-M

, Balda Dayan

, et al. 3D-printed pneumatically controlled soft suction cups for gripping fragile, small, and rough objects. Adv Intell Syst, 2021; 3(9):2100034.

29.

Zhou

, Qi

, Xia

, et al. Miniaturized soft centrifugal pumps with magnetic levitation for fluid handling. Sci Adv, 2021; 7(44):eabi7203.

30.

Diteesawat

, Helps

, Taghavi

, et al. Electro-pneumatic pumps for soft robotics. Sci Robot, 2021; 6(51):eabc3721.

31.

Lin

, Yang

, Gong

, et al. Bio-inspired magnetic-driven folded diaphragm for biomimetic robot. Nat Commun, 2023; 14(1):163.

32.

Cacucciolo

, Shintake

, Kuwajima

, et al. Stretchable pumps for soft machines. Nature, 2019; 572(7770):516–519.

33.

Cao

, Gao

, Conn

. A magnetically coupled dielectric elastomer pump for soft robotics. Adv Mater Technol, 2019; 4(8):1900128.

34.

Tang

, Zhang

, Zhong

, et al. Customizing a self-healing soft pump for robot. Nat Commun, 2021; 12(1):2247.

35.

, Nunez

, Souri

, et al. A compact dea-based soft peristaltic pump for power and control of fluidic robots. Sci Robot, 2023; 8(79):eadd4649.

36.

Jiang

, Tang

, Dong

, et al. Soft pocket pump for multi-medium transportation via an active tubular diaphragm. Adv Funct Materials, 2023; 33(50):2305289.

37.

Fuhrer

, Schumacher

, Zeltner

, et al. Soft iron/silicon composite tubes for magnetic peristaltic pumping: Frequency-dependent pressure and volume flow. Adv Funct Materials, 2013; 23(31):3845–3849.

38.

Wang

, Liao

, Sun

, et al. Bioinspired construction of artificial cardiac muscles based on liquid crystal elastomer fibers. Adv Mater Technol, 2022; 7(1):2100934.

39.

Cao

, Xuan

, Hu

, et al. 3D printing-assistant method for magneto-active pulse pump: Experiment, simulation, and deformation theory. Appl Phys Lett, 2020; 117(24):241901.

40.

, Zhu

, Foo

, et al. A robust dual-membrane dielectric elastomer actuator for large volume fluid pumping via snap-through. Appl Phys Lett, 2017; 111(21):212901–212911.

41.

Polygerinos

, Wang

, Overvelde

JTB

, et al. Modeling of soft fiber-reinforced bending actuators. IEEE Trans Robot, 2015; 31(3):778–789.

42.

Xavier

, Fleming

, Yong

. Finite element modeling of soft fluidic actuators: Overview and recent developments. Adv Intell Syst, 2021; 3(2):2000187.

43.

Marechal

, Balland

, Lindenroth

, et al. Toward a common framework and database of materials for soft robotics. Soft Robot, 2021; 8(3):284–297.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

3.74 MB

3.31 MB

2.88 MB

10.02 MB

5.49 MB

10.02 MB

5.49 MB

11.96 MB

8.22 MB

3.25 MB

6.54 MB

6.89 MB

6.25 MB

6.70 MB

5.48 MB

12.98 MB