In recent years, soft robotics has emerged as a transformative field, with pneumatic soft grippers serving as key components in applications ranging from medical rehabilitation to rescue operations. These grippers, often constructed from highly elastic materials like silicone, offer superior compliance and safety compared to traditional rigid robots. However, manufacturing complexities, particularly in forming hollow structures, pose significant challenges. Conventional methods, such as layered fabrication with adhesive bonding, frequently result in defects like poor sealing and reduced durability. To address these issues, we explore an advanced lost wax casting technique combined with multi-step pouring, leveraging diffusion theory to enable monolithic fabrication of pneumatic soft grippers. This approach not only simplifies模具 design but also enhances structural integrity and airtightness. In this article, we detail the design, manufacturing process, simulation, and experimental validation of a multi-cavity pneumatic mesh soft gripper, emphasizing the efficacy of lost wax casting in soft robotics.
The structural design of our pneumatic soft gripper incorporates a multi-cavity pneumatic network, which facilitates controlled bending deformation upon inflation. As illustrated in the conceptual diagram, the actuator comprises a驱动层 with seven semi-cylindrical air chambers connected by a central channel and a quarter-spherical chamber, all bonded to a restrictive base layer. When pressurized air is introduced, the chambers expand uniformly, but the non-deformable base layer constrains elongation, resulting in a downward bending motion. This principle, known as pneumatic network actuation, allows for significant deformation with minimal input pressure. The design ensures high force output and rapid response, making it ideal for delicate tasks like object manipulation in unstructured environments. To mathematically model the bending behavior, we consider the pressure-deformation relationship, where the bending angle $\theta$ relates to the applied pressure $P$ and material properties. For a single chamber, the expansion can be described by:
$$ \Delta V = \frac{P \cdot A}{E} $$
where $\Delta V$ is the volume change, $A$ is the effective area, and $E$ is the elastic modulus of the silicone. For the entire gripper, the cumulative effect of multiple chambers leads to the bending angle, approximated as:
$$ \theta = k \cdot P \cdot L $$
Here, $k$ is a constant dependent on material stiffness and geometry, and $L$ is the length of the gripper. This linear relationship underscores the predictability of the actuation, which we further validate through simulations and experiments.
The manufacturing process centers on lost wax casting, a method that enables the creation of intricate internal cavities without adhesives. Traditional approaches often involve bonding multiple layers, which can compromise密封性 and longevity. In contrast, lost wax casting involves creating a wax core that defines the internal channels, embedding it in silicone, and then melting the wax to leave a hollow structure. However, challenges such as poor silicone fluidity and core displacement persist. To overcome these, we integrate multi-step pouring based on diffusion theory, where liquid silicone interdiffuses with partially cured layers, forming a seamless interface. This process involves two main stages: wax core fabrication and silicone casting. The wax core, made from recyclable paraffin, is molded using 3D-printed molds split along the symmetry axis for easy demolding. The silicone gripper is then cast in steps—first, the驱动层 is poured and cured, followed by embedding the wax core and pouring the base layer. The diffusion at the interface ensures molecular bonding, eliminating the need for adhesives. Key parameters influencing diffusion include solubility parameters and interfacial tension, which we optimize for Dragon Skin 30 silicone, as detailed in Table 1.
Table 1: Material Properties of Dragon Skin 30 Silicone
| Parameter | Value |
|---|---|
| Color | Translucent |
| Mix Ratio (A:B) | 1:1 by weight |
| Tear Strength (pli) | 108 |
| Shrinkage (in./in.) | < 0.001 |
| Elongation at Break (%) | 364 |
| 100% Modulus (psi) | 86 |
| Tensile Strength (psi) | 500 |
| Specific Gravity (g/cm³) | 1.08 |
| Hardness (Shore A) | 30 |
| Cure Time (hours) | 16 |
| Pot Life (minutes) | 45 |
| Viscosity (cps) | 30,000 |
Mold design is critical to the success of lost wax casting. We employ 3D-printed molds for both the wax core and silicone gripper, incorporating features like alignment pins and vent holes to mitigate air entrapment and core misalignment. The wax core mold consists of two halves with 5° draft angles on locating pins for easy assembly and disassembly. For the silicone gripper, a four-part mold system includes a base, spacer, and two cover plates. The cover plates feature 2 mm diameter vents that allow excess silicone and air to escape during pressing, ensuring bubble-free casting. This design reduces the likelihood of wax core deflection, a common issue in traditional lost wax casting. The step-by-step casting process begins with heating paraffin to 75°C for liquefaction, pouring it into the core mold, and demolding after 15–20 minutes. Next, the驱动层 silicone is mixed, degassed, and poured into the base mold with Cover A, then cured. The wax core is embedded, and the base layer is cast using the spacer and Cover B. After full curing, the wax is melted out in a 75°C water bath or oven, leaving a monolithic gripper. This method enhances repeatability and reduces material waste, aligning with sustainable practices.
To validate the manufacturing process, we conducted finite element analysis (FEA) using ABAQUS and experimental tests on the bending performance. The FEA model incorporated hyperelastic material properties for Dragon Skin 30, derived from uniaxial test data. Simulations applied pressures from 0 to 30 kPa in 5 kPa increments, measuring the resulting bending angles. The governing equation for large deformation in soft materials is based on the Neo-Hookean model:
$$ W = \frac{\mu}{2} (I_1 – 3) + \frac{\kappa}{2} (J – 1)^2 $$
where $W$ is the strain energy density, $\mu$ is the shear modulus, $\kappa$ is the bulk modulus, $I_1$ is the first invariant of the Cauchy-Green tensor, and $J$ is the volume ratio. The simulation results showed a near-linear increase in bending angle with pressure, consistent with theoretical predictions. Experimentally, we fabricated grippers using the described lost wax casting technique and measured bending angles under the same pressure range. A comparison revealed that experimental angles slightly exceeded simulations due to material nonlinearities and manufacturing tolerances. Table 2 summarizes the data, while the relationship is expressed as:
$$ \theta_{\text{exp}} = \alpha \cdot P + \beta $$
where $\alpha$ and $\beta$ are empirical constants. The close agreement confirms the工艺’s reliability for functional soft grippers.
Table 2: Bending Angle vs. Pressure: Simulation vs. Experiment
| Pressure (kPa) | Simulation Angle (°) | Experimental Angle (°) |
|---|---|---|
| 5 | 12.5 | 14.2 |
| 10 | 25.1 | 27.8 |
| 15 | 37.8 | 40.5 |
| 20 | 50.3 | 53.1 |
| 25 | 62.9 | 66.0 |
| 30 | 75.4 | 78.9 |
In conclusion, our innovative approach to lost wax casting with multi-step pouring effectively addresses the limitations of conventional soft gripper manufacturing. By harnessing diffusion theory, we achieve monolithic structures with improved sealing and mechanical performance. The optimized mold design and process parameters facilitate high-quality production, while simulations and experiments validate the design’s functionality. This lost wax casting method not only simplifies fabrication but also paves the way for broader applications in soft robotics, such as wearable devices and adaptive grippers. Future work will focus on refining material formulations and exploring complex geometries to further enhance performance. Ultimately, this工艺 represents a significant advancement in the pursuit of robust, efficient, and scalable soft robotic systems.