1. Introduction

In the field of soft robotics, the use of soft materials enables the reduction of deformation and damage to objects during contact or grasping [1]. Soft robots fabricated using conventional materials such as poly(dimethyl siloxane) (PDMS) and synthetic resins have been reported [2–5]. In recent years, the fabrication of soft robots using hydrogels as flexible materials has been reported [6–8]. Hydrogels exhibit responsiveness to various external stimuli, including temperature, pH, and electric fields, making them suitable for actuation and other applications [9–11]. However, the mechanical strength of hydrogels is generally lower compared to silicon-based elastomers or metals. To ensure optimal and long-term use for various applications, there is a need to modify the mechanical properties of hydrogels or enhance their strength. Various approaches have been explored, including molecular design of hydrogel polymers [12,13], incorporation of fillers [14,15] or nanoclays [16], and the fabrication of double-network hydrogels (DN gels) using multiple hydrogel polymers crosslinked independently [17–20]. DN gels, in particular, are known to exhibit high mechanical strength due to the multi-stage crosslinking of dissimilar hydrogel monomers. These hydrogels, where two distinct three-dimensional polymer networks physically (non-covalently) interpenetrate and entangle, are also referred to as interpenetrating polymeric network gels (IPN gels). For example, DN gels using poly(2-acrylamido-2-methyl-1-propanesulfonic acid) (PAMPS) as the first network and polyelectrolyte and polyacrylamide (PAAm) as the second network demonstrated a high compressive strength of 17.2 MPa [17]. On the other hand, besides imparting suitable mechanical strength for specific applications, maintaining the flexibility of hydrogels is also necessary to ensure the functional capabilities of gel-based applications. For instance, soft robotic hands require the ability to grasp objects with diverse shapes and hardness, necessitating the use of materials with flexibility on the contact surface with the objects [21]. Spatially selective arrangement of both flexible and rigid structures is desired for such applications.

Multi-photon polymerization (MPP), a known technique for photo-crosslinking hydrogels or polymer materials, has been widely used to fabricate intricate three-dimensional structures [22–24]. Recently, research efforts have reported the fabrication of volume diffractive gratings by utilizing femtosecond lasers for the densification of hydrogels [25]. It holds promise as a fabrication method for soft robots consisting of microscale hydrogels. It has also been reported that the mechanical properties of the fabricated structures can be varied by adjusting the irradiation conditions of the laser pulses used for MPP [26]. Therefore, it is expected that by locally altering the laser irradiation conditions, the mechanical properties of specific regions within the microstructures can be selectively modified. However, the range of achievable changes is believed to be solely dependent on variations in the crosslinking density of a single polymer network. If spatially selective crosslinking of DN gels becomes possible, it is anticipated to enable a wider range of mechanical property modifications attributed to the presence of multiple polymer networks, contributing to the realization of highly functional micro-soft robots.

In this study, we attempted spatially selective photo-crosslinking of hydrogel monomers embedded within pre-crosslinked regions using femtosecond laser pulses, employing MPP as a selective crosslinking method for DN gels. Evaluation of the mechanical properties of the fabricated structures demonstrated the formation of a double network only in the regions subjected to two-stage crosslinking, indicating a change in mechanical strength.

2. Materials and methods

2.1 Double network (DN) gel preparation

Polyethylene glycol diacrylate with an average molecular weight of 4000 (PEGDA4000, Polysciences) and a photoinitiator (Irgacure2959, Sigma-Aldrich Co. LLC) were dissolved in pure water and stirred with a magnetic stirrer for approximately 20 minutes. Additionally, as a different hydrogel monomer, polyethylene glycol diacrylate (PEGDA) with an average molecular weight of 700 (PEGDA700, Sigma-Aldrich, Co., LLC), along with a photoinitiator and pure water, were stirred using a magnetic stirrer for approximately 20 minutes. The concentration of the photoinitiator was 1 wt% for all PEGDA solutions with average molecular weight.

Figure 1(a) shows conceptual illustration of the network structure inside the hydrogel. Double network gels were prepared using three different methods in this study. The first method involved cross-linking the 1st and 2nd gels through a linear optical absorption process using UV irradiation (Fig. 1(b)). For this experiment, PEGDA with average molecular weights of 700 and 4000 was used as the hydrogel monomer, and the DN gels were prepared in block form. The second method entailed cross-linking the 1st gel with UV irradiation, followed by cross-linking of the 2nd gel with femtosecond laser pulse irradiation (Fig. 1(c)). By cross-linking the 2nd gel with MPP, a portion of the 1st gel was selectively transformed into a DN gel in a spatially-specific manner. Samples were used to experimentally demonstrate the feasibility of cross-linking the 2nd gel through MPP and to investigate local changes in mechanical properties associated with the spatially-selective cross-linking of DN gels. The third method involved fabricating DN gels using MPP induced by femtosecond laser pulse irradiation for cross-linking both the 1st and 2nd gels (Fig. 1(d)). This method was employed to examine spatially-selective changes in mechanical properties depending on the fabrication conditions. The hydrogel microstructures fabricated using this method were also utilized to experimentally demonstrate micro-object grasping. The details of the three methods are described in the following subsections.

 

Fig. 1. (a) Conceptual illustration of the network structure inside the hydrogel and (b-d) cross-linking methods of hydrogel. (b) Preparing DN gel blocks only by UV irradiation. (c) Fabricating DN gel microstructure by MPP inside the UV-cured SN gel block. (d) DN gel microstructure cross-linked only by MPP.

Download Full Size | PDF

2.1.1 Preparation of DN gel blocks by UV irradiation

The first method of cross-linking DN gels was prepared by a linear optical absorption process by UV irradiation (Fig. 1 (b)). The prepared PEGDA4000 solution was dropped into a silicone mold (3 × 3 × 3 mm³), and the hydrogels were cross-linked in a block shape (1st gel) by irradiation with light from UV lamp (LUV-16, ASONE) at a center wavelength of 365 nm for 40 minutes. The hydrogels were then immersed in 50 ml of pure water for 24 hours to remove residual unreacted monomers inside. Next, the cross-linked PEGDA4000 SN gels were immersed in a PEGDA700 solution for 24 hours to allow the monomer to permeate inside. Finally, the immersed hydrogels were placed in the same silicone mold and irradiated with a UV lamp for 40 minutes to cross-link the 2nd gel and form DN gel blocks. After preparation, the hydrogels were immersed in 50 ml of pure water for 24 hours to remove residual monomers.

2.1.2 Preparation of DN gel by MPP inside a SN gel block prepared by UV irradiation

The second method of cross-linking DN gels was to cross-link the 1st gel through a linear optical absorption process using UV irradiation and then cross-link the 2nd gel using MPP with femtosecond laser pulse irradiation (Fig. 1 (c)). After removing the residual monomers, SN gel blocks containing PEGDA700 monomers were prepared by immersing them in a PEGDA700 solution for 24 hours. The SN gel block, consisting of PEGDA4000, after immersion in the PEGDA700 solution, was placed on a cover glass, and microstructures were fabricated inside it using MPP. A femtosecond laser (HighQ-2, Spectra-Physics) with a center wavelength of 522 nm, a pulse width of 192 fs, and a repetition rate of 63 MHz was used as the light source. The femtosecond laser pulses were focused and scanned using a water immersion objective (NA = 1.0, Olympus) and an XYZ stage (OptoSigma) to induce a polymerization reaction inside the SN gel block and fabricate a three-dimensional structure. The planar structure fabricated by laser irradiation had dimensions of 360 × 360 µm². The three-dimensional structure was created by forming five adjacent layers of planar structures, with each layer spaced 30 µm apart along the Z-axis direction. The scanning speed of the XYZ stage was 200 µm/s, and the laser power was 15 mW, which corresponds to ∼2.4 nJ per pulse.

2.1.3 Preparation of DN gels by MPP

The third method of cross-linking DN gels involved creating microstructures by inducing cross-linking with MPP in both the 1st and 2nd gels (Fig. 1 (d)). An acrylic plate with a thickness of 0.2 mm was used as a spacer for the microstructures, and a hole measuring 2 × 2 mm² was drilled in the center of a 20 × 20 mm² acrylic plate using a laser cutting machine (Speedy 300, trotec). PEGDA4000 solution was dropped onto the spacer placed on a cover glass, and another cover glass was placed on top to seal the solution. After cross-linking the PEGDA4000 as the 1st gel, the fabricated structure was immersed in pure water for 1 hour and then immersed in PEGDA700 solution for 1 hour. The immersed 1st gel was sealed in a spacer with PEGDA700 solution, and the 2nd gel was cross-linked by focusing and scanning the laser pulse again. The laser power was 10 mW (∼0.16 nJ per pulse), and the scanning speed was 200 µm/s. The scanning trajectory with the stage was set at a spacing of 3 µm between adjacent lines in XY plane. In the process of fabricating multilayered DN gels, a layer-to-layer spacing of 20 µm was employed. The laser pulses were focused and scanned in the direction of the depth of the solution irradiation, starting and ending inside the cover glass between the spacers. This procedure ensured that the 2nd gel (PEGDA700) was cross-linked even at the interface between the cover glass and the solution, i.e., the surface of the fabricated structure.

2.2 Evaluation of mechanical properties of DN gel blocks

The sample was positioned on a transparent acrylic stage, and a digital force gauge (DST-50N, IMADA) was securely attached to the uniaxial displacement stage. Compression testing was conducted by applying pressure to the sample through a compression platen connected to the force gauge. The initiation point of the measurement was defined as the position at which the force gauge first detected vertical load. The sample was compressed at a constant rate of 25 µm/s until fracture occurred. During the compression process, the stress was calculated based on the observed sample area, and stress-strain curves were generated. The following equations were employed to determine stress (σ) and strain (ε):

2.3 Evaluation of mechanical properties of DN gel microstructures prepared by MPP inside the hydrogel blocks

The block-shaped sample with DN microstructure was positioned on a cover glass and examined using a bright-field optical microscope (Eclipse Ti-E, Nikon). To induce compression, the sample was compressed from both sides using a digital micrometer (DM025, ASONE) with glass plates attached. The alterations in the internal structure during compression and the resulting local strain were observed.

For the assessment of local strain, metal dot arrays comprising silver nanoparticles were created within the hydrogel through multi-photon photo reduction to visualize positional coordinates. Multi-photon photo reduction is a technique that employs femtosecond laser pulses to selectively induce the reduction of metal ions in a spatially controlled manner. It has been reported that this method enables the fabrication of structures within hydrogels [27,28]. In this study, metal dot arrays consisting of silver nanoparticles were formed within a 0.9 × 0.9 mm² region on the same plane, with a spacing of 30 µm. The laser power was set to 15 mW. The fabrication of metal dot structures was conducted subsequent to the cross-linking of the hydrogel. To eliminate the residual prepolymer, the fabricated gels were immersed in deionized water for more than 24 hours, followed by immersion in a silver nitrate aqueous solution.

2.4 Evaluation of mechanical properties of DN gel microstructures fabricated by MPP

The mechanical properties of the fabricated DN gel microstructures were assessed using a Nanoindenter (iNano, Nanoindenter, Nanomechanics Inc.). A flat platen with a contact surface diameter of 20 µm was utilized for the measurements. Load-displacement curves were recorded, with a maximum load of 2 mN.

2.5 Micro object grasping

U-shaped microstructures were created by cross-linking the hydrogels using MPP. Two types of structures were prepared: U-shaped SN (single-network) gel microstructures and SN gel microstructures with a portion selectively transformed into DN gel. To fabricate the U-shaped SN gel microstructures, a PEGDA4000 solution with a concentration of 0.2 g/ml was utilized. For the 2nd gel, a PEGDA700 solution with a concentration of 0.4 g/ml was employed. During the fabrication process, the laser power was set to 10 mW (∼0.16 nJ per pulse), and the scanning speed was set at 200 µm/s. Glass beads (BZ-01, ASONE) with an approximate diameter of 100 µm were used as the objects to be grasped. The fabricated structure and glass bead were placed on top of a cover glass and surrounded by water droplets. External mechanical tweezers (Sampling Station, Micro Support) equipped with a digital microscope were employed to open and close the fabricated structure, allowing for the grasping of the glass bead.

3. Results and discussions

3.1 Mechanical properties of DN gel blocks

Monomers of PEGDA4000 and PEGDA700 were used as the first and second gels, respectively, to prepare DN gels. The mechanical properties of the DN gel blocks, cross-linked by UV irradiation, were measured using compression tests, as shown in Fig. 2(a). Figure 2(b) and 2(c) illustrate the stress-strain curves obtained from the compression tests for the DN gel blocks. The elastic modulus of the DN gel blocks increased with an increasing concentration of the PEGDA700 solution, which was cross-linked to form the second gel (Fig. 2(b)). During the cross-linking process of the hydrogel monomers, a higher concentration of monomers led to a higher cross-linking density of the resulting hydrogel, resulting in the formation of a network structure with a higher elastic modulus [29]. The observed increase in the elastic modulus suggests that PEGDA700 permeated the first gel and cross-linked to form a network for the second gel. Additionally, the fracture stress increased from approximately 150 kPa to 1830 kPa when the concentration of the second gel was increased from 0.2 g/ml to 0.6 g/ml. Furthermore, when the concentration of the PEGDA4000 (first gel) solution was increased from 0.1 g/ml to 0.2 g/ml, both the elastic modulus and fracture stress increased (Fig. 2(c)). This increase in the cross-linking density of the PEGDA4000 contributed to an overall increase in the collective cross-linking density of the DN gels.

 

Fig. 2. (a) A schematic diagram of the compression test for obtaining elastic modulus and fracture stress of UV-cured hydrogel blocks. (b,c) Stress-strain curves of DN gel blocks prepared with PEGDA4000 (1st gel) at concentrations of 0.1 g/ml (b) and 0.2 g/ml (c) with various PEGDA700 (2nd gel) concentration.

Download Full Size | PDF

3.2 Mechanical properties of DN gel microstructures fabricated by MPP inside SN gels

We attempted to form DN gel microstructures using MPP inside the SN gel blocks. The change in local mechanical properties induced by spatially-selective cross-linking of DN gels was investigated using compression tests, as shown in Fig. 3(a). Bright-field optical microscope images of the samples before and after compression are shown in Figs. 3(b) and 3(c), respectively. The formation of a new PEGDA700-derived network structure resulted in a change in the refractive index inside the hydrogel, allowing for the observation of the fabricated second gel structure under the optical microscope. When the entire hydrogel block was compressed, the internal second gel was also compressed. The dependence of dot distances on the compression distance of the entire hydrogel block is presented in Fig. 3(d). The distances between dots were measured in the SN gel (where MPP was not performed) and in the DN gel (where the second gel was cross-linked by MPP), and the averages were plotted. Before compression, the distances between dots were comparable for SN and DN gels. However, when the entire hydrogel block was compressed by 1400 µm, the distances between dots inside the SN and DN gels were 15.3 µm and 27.1 µm, respectively. This result indicates a change in stiffness in the DN gel, where the second gel was spatially-selectively cross-linked.

 

Fig. 3. (a) Schematic diagrams of the compression test for DN gel microstructure fabricated inside SN hydrogel block with metal dot array. (b, c) Bright-field optical microscope images of the fabricated structure (b) before and (c) after compression (1.2 mm). The scale bars indicate 100 µm. (d) Dependence of averaged inter-dot distance in SN and DN gels with compression.

Download Full Size | PDF

To measure the local strain, we measured the distance between metal dots, which were sequentially numbered in the horizontal direction for each compression distance (Fig. 4). Bright-field optical microscope images of the DN gel microstructure before and after compression are shown in Figs. 4(a) and 4(b), respectively. The metal dot arrays were numbered and identified within the red dotted square. Measurement points 1 to 5 and 20 to 26 represent dots created inside the SN gel, while measurement points 6 to 19 represent dots within the region where the DN gel was cross-linked. Figure 4(c) illustrates the local strain evaluated by the change in distance between the numbered dots during compression. The elastic modulus of the network structure derived from PEGDA700 was higher than that of the SN gel (PEGDA4000), resulting in the formation of a DN gel with a rigid, high elastic modulus network structure.

 

Fig. 4. Spatial distribution of the compression in hydrogel. (a, b) Bright-field optical microscope images of the fabricated structure (a) before and (b) after compression (1.2 mm). The scale bars indicate 100 µm. The bottom figures are magnified views of the red dotted squares. The distance between the dots was set to 30um. The dots were numbered to evaluate the local strain of the fabricated structure. (c) Local strain distribution of fabricated structure at each compression distance (0.4, 0.8. 1.2 mm).

Download Full Size | PDF

3.3 Mechanical properties of DN gel microstructures fabricated by MPP

DN gel microstructures were fabricated using MPP for both the first and second gels. Initially, single-layer planar structures were fabricated using PEGDA4000, followed by the fabrication of planar structures using PEGDA700 that intersected perpendicularly. The DN gel microstructure was formed at the intersection of these two structures. Figure 5(a) illustrates the bright-field optical microscopic image of the fabricated microstructure. The structures within the red, blue, and green squares indicate the DN gel microstructure, PEGDA700 SN gel, and PEGDA4000 SN gel, respectively. Figure 5(b) displays the measured mechanical properties of each structure. The shear modulus of the DN gel microstructure was 33 kPa, which was higher than that of the PEGDA4000 SN gel, suggesting that the relatively rigid PEGDA700 network structure, the second gel network, was locally cross-linked inside the first gel network. Conversely, the shear modulus of the PEGDA700 SN gel was approximately 62 kPa, which was higher than that of the DN gel microstructure. This discrepancy may be due to the relatively low density of PEGDA700 in the network structure of the DN gel. The concentration of monomer molecules in the second gel, which permeated the cross-linked first gel, could be lower than the monomer concentration.

 

Fig. 5. (a) Bright-field optical microscope image of the single-layered hydrogel microstructures fabricated by MPP on a cover glass. The structures consisting of PEGDA700 in the longitudinal direction (blue) and PEGDA4000 in the transverse direction (green) were fabricated. DN gels were formed at the intersections (red). The scale bar indicates 100 µm. (b) Shear modulus of fabricated microstructure measured by nanoindentations.

Download Full Size | PDF

To evaluate the mechanical properties, DN gel microstructures were fabricated in multiple layers. Figure 6 presents a schematic of the fabrication method (a), the composition of the fabricated structure (b), and a digital microscope image of the fabricated structure (c). The square structure was observed in a 300 × 300 µm² area where an additional laser was scanned, indicating that the second gel was cross-linked in the block hydrogel. Figure 7(a) and Fig. 7(b) depict the load-displacement curves and the shear modulus obtained from the DN gel, respectively. With an increase in the concentration of the second gel, the stiffness of the DN gel increased compared to the SN gel. The shear modulus of the DN gel was 87 kPa at a concentration of 0.2 g/ml and 5.8 MPa at a concentration of 0.6 g/ml, demonstrating that the shear modulus could be varied up to 66 times within the employed concentration range. This result indicates that the stiffness of the DN gel was successfully improved through spatially selective double cross-linking. The shear modulus of the SN gels prepared with PEGDA700 or PEGDA4000 is shown in Fig. 7(c). The shear modulus of the SN gels increased with an increasing concentration for both PEGDA4000 and PEGDA700 gels. The PEGDA4000 SN gel exhibited a shear modulus of ∼41 kPa when prepared at a concentration of 0.2 g/ml, and a maximum value of ∼180 kPa at the highest concentration of 0.6 g/ml, indicating that the change in shear modulus was limited to 4.39 times with the employed concentrations. These results suggest that, within these concentrations, the DN gel prepared using our method can achieve a wider range of change in shear modulus compared to the SN gel.

 

Fig. 6. (a) Schematic diagram for fabricating multi-layered DN gel microstructures by MPP. (b) Illustration of the fabricated structure. (c) Top-view of the fabricated structure obtained with digital microscope. The scale bar indicates 200 µm.

Download Full Size | PDF

 

Fig. 7. (a) Load-displacement curves of multi-layered DN gel microstructures, where the concentration of 2nd gel was varied from 0.2-0.5 g/ml. (b, c) Obtained shear modulus of DN gel (b) and SN gels of PEGDA700 and PEGDA4000 (c) fabricated in multiple layers

Download Full Size | PDF

When the DN gel microstructure was fabricated in multiple layers, the shear modulus of the DN gel fabricated at a concentration of 0.6 g/ml for the second gel exceeded that of the PEGDA700 SN gel fabricated at the same concentration. This result differed from the findings for the DN gel with a single-layered planar structure (Fig. 5(b)). The increased cross-linking density resulting from the fabrication of the structure in multiple layers is believed to have contributed to this difference. The polymerization reaction induced by focused laser pulse scanning could be initiated again within the cross-linked hydrogel by stacking the layers.

3.4 Micro object grasping by DN gel microstructure

The grasping of micro-objects was demonstrated using a structure where the stiffness was locally enhanced through spatially-selective formation of DN gel. Initially, U-shaped SN gel tweezers composed of PEGDA4000 SN gel were used to sandwich a glass bead with a diameter of 100 µm, as illustrated in Fig. 8(a). However, when force was applied from the top and bottom, the significant deformation of the SN gel caused the glass bead to slip through the gel structure (Figs. 8(b), (c), and [Visualization 1]). The structure was too flexible to effectively transmit external force from the mechanically driven tweezers to the contact area with the bead. Conversely, when the U-shaped structure incorporated DN gel microstructures near the contact area with the glass bead through spatially-selective cross-linking of PEGDA700, the glass bead was successfully grasped (Figs. 8(e), (f), and [Visualization 2]). This was possible due to the combination of flexible PEGDA4000 SN gel at the contact area with the bead and rigid DN gel microstructures, which facilitated the transmission of external force within the gel tweezers.

 

Fig. 8. Grasping demonstration of a glass bead by U-shaped hydrogel tweezers consisting (a-c) only with PEGDA4000 and (d-f) with reinforced DN structure. (a,d) Schematics of the fabricated structures. (b,c, e, f) Object grasping before (b, e) and after (c, f) applying the external force. The scale bars indicate 200 µm.

Download Full Size | PDF

4. Conclusion

In this study, we presented an experimental study evaluating the mechanical properties of DN gels crosslinked through UV irradiation and/or femtosecond laser pulse irradiation. When DN gels were crosslinked using a polymerization reaction mediated by one-photon absorption, we confirmed that the elastic modulus and fracture stress improved with an increase in the concentration of the 2nd gel. Next, we demonstrated the ability to create structurally modified systems where the mechanical properties were selectively altered within SN gel by locally crosslinking DN gel through MPP. Lastly, we showed that by solely employing MPP as the crosslinking method within micro-scale hydrogel microstructures, it was possible to achieve localized changes in mechanical properties through laser scanning during the fabrication of the 2nd gel. By enabling localized changes in mechanical properties, we highlighted the potential application of this structure to soft manipulators, which would exhibit minimal deformation and facilitate easy transmission of external forces during object manipulation. We believe that this research outcome contributes to the fundamental technologies required for the development of micro-soft robots.