**2.1. Fabrication of artificial pectoral fin**

226 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

generation of autonomous underwater vehicles.

to control the volume of the bladder underwater.

discussed in Section 4.

**Figure 3.** (a) Sperm whale (*Physeter macrocephalus*); (b) Golden fish (*Carassius auratus*).

The rest of this Chapter is organized as the follows. Section 2 is focused on the development of a bio-inspired robotic manta ray powered by IPMC artificial pectoral fin. Fabrication and characterization of the artificial pectoral fin and design of the robotic manta ray are presented. Section 3 is focused on bio-inspired depth control device enabled by IPMC enhanced water electrolysis, where the buoyancy control mechanism, device design, and open loop testing are demonstrated. Conclusions and future work on both studies are

(a) (b)

solution is not feasible for implementation in small devices. In order to build more efficient buoyancy control devices, researchers have turned to biology for inspiration for the next

Biology has many novel and effective depth control mechanisms suitable for a variety of environments. For example, sperm whales (Fig. 3(a)) achieve buoyancy control by using their spermaceti oil. An adult sperm whale contains about 4 tons of spermaceti oils in their spermaceti organ, which represents approximately 8% of its total mass (Shibuya et al, 2006). The spermaceti oil has a low melting point and its density depends largely on the temperature of the oil. By manipulating the arterial blood flow through the spermaceti organ, the sperm whales can regulate the temperature of the oil and are thus able to control their buoyancy. There have been recent demonstrations of buoyancy control concepts manipulating temperature to change the density of oil (Shibuya et al, 2006) or wax (McFarland et al, 2003). However, the response times are slow (on the order of 10 minutes), and it is inefficient for small devices because a constant power must be supplied to maintain the temperature of the oil while cruising at a certain depth. Ray-finned fishes, such as one depicted in Fig. 3(b), change the buoyancy of their body using a swim bladder (Bond, 1996). Expansion of the bladder results in increased volume, thus making the body more positively buoyant and vice versa. Inspiration for the artificial bladder presented in this chapter comes from these ray-finned fishes. The challenges arise from how to generate and release the gas

The proposed artificial pectoral fin must be able to generate oscillatory motion with a twist angle as observed in swimming of the manta ray, under hydrodynamic loads. An artificial pectoral fin was fabricated by combining one IPMC actuator with a PDMS elastomer in a mold to create a predefined planform shape. The design of pectoral fin is shown in Fig. 4(a). The outline shape of the fin mimics that of the manta ray. The fin is divided into two areas: IPMC beam on the leading edge and PDMS passive membrane on the trailing edge. Note that the size and shape of the IPMC is chosen to generate enough bending moment with limited power consumption. Optimal design of the fin will be focused in the future work.

**Figure 4.** IPMC powered artificial pectoral fin.

The first step in creating the artificial pectoral fin is to fabricate the IPMC actuator. Many groups have developed different IPMC fabrication processes for various purposes (Kim & Shahinpoor, 2003; Chung et al, 2006; Lee et al, 2006; Kim & Shahinpoor, 2002). In our fabrication method, we followed most of the procedure outlined by K. Kim and M. Shahinpoor (Kim & Shahinpoor, 2003) but added additional platinum/gold plating process to reduce the surface resistance of the electrodes (Lee et al, 2006). The following supplies are used to fabricate the IPMC beams: (1) Nafion ion exchange membrane Nafion 1110 (258 µm thick, DuPont); (2) tetraammineplatinum chloride 98% (Aldrich); (3) sodium borohydride (NaBH4, Aldrich); (4) dilute ammonium hydroxide solution (NH4OH 29% solution); 5) dilute hydrochloric acid (HCl aq, 1.0 N solution); and 6) de-ionized water. Fig. 5 shows the IPMC fabrication process. The major fabrication steps are as follows:

**Figure 5.** IPMC fabrication process.

1. Treatment of Nafion with HCl (cleaning). This step is to remove metal particles and other impurities from the film. The Nafion film (10 cm by 5 cm) is boiled in 1.0 N hydrochloride acid (HCl) at 80 °C for 30 min. The film is then rinsed with DI water to remove acid residue.


After fabrication of the IPMC actuator, the next step is to bond the IPMC with a PDMS elastomer membrane. The PDMS bonding process (shown in Fig. 6) is described in the following 5 steps. (1) A Delrin polymer (McMaster) mold was made using a Computer Numerical Control (CNC) rapid milling machine (MDX-650, Roland). The mold has two concaved areas to house the PDMS passive membrane and IPMC actuator. The thicknesses of PDMS membrane and IPMC are 600 µm and 280 µm, respectively. (2) The IPMC was cut into the shape shown in Fig. 3(a). (3) About 3% glass bubbles (Glass bubble K37, 3M Inc) were added into PDMS gel (Ecoflex 0030, Smooth-on Inc.) to gain a neutrally buoyant pectoral fin. (4) The IPMC and the PDMS gel were then clamped with the mold and the PDMS was cured at room temperature for 3 hours. (5) The IPMC/PDMS artificial pectoral fin (shown in Fig. 4(b)) was removed from the mold. Note that there is about 150 µm thick PDMS covered on the top of IPMC to guarantee good bonding between passive and active areas. The additional layer of PDMS stiffens the IPMC active area. However, since the PDMS (Ecoflex) is an extremely compliant material (Young's modulus, 0.06 MPa), compared to the Nafion (Young's modulus 114 MPa), the stiffening effect can be neglected.

### **2.2. Characterization of pectoral fin**

228 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

steps are as follows:

**Figure 5.** IPMC fabrication process.

remove acid residue.

The first step in creating the artificial pectoral fin is to fabricate the IPMC actuator. Many groups have developed different IPMC fabrication processes for various purposes (Kim & Shahinpoor, 2003; Chung et al, 2006; Lee et al, 2006; Kim & Shahinpoor, 2002). In our fabrication method, we followed most of the procedure outlined by K. Kim and M. Shahinpoor (Kim & Shahinpoor, 2003) but added additional platinum/gold plating process to reduce the surface resistance of the electrodes (Lee et al, 2006). The following supplies are used to fabricate the IPMC beams: (1) Nafion ion exchange membrane Nafion 1110 (258 µm thick, DuPont); (2) tetraammineplatinum chloride 98% (Aldrich); (3) sodium borohydride (NaBH4, Aldrich); (4) dilute ammonium hydroxide solution (NH4OH 29% solution); 5) dilute hydrochloric acid (HCl aq, 1.0 N solution); and 6) de-ionized water. Fig. 5 shows the IPMC fabrication process. The major fabrication

1. Treatment of Nafion with HCl (cleaning). This step is to remove metal particles and other impurities from the film. The Nafion film (10 cm by 5 cm) is boiled in 1.0 N hydrochloride acid (HCl) at 80 °C for 30 min. The film is then rinsed with DI water to The pectoral fin was characterized in terms of tip deflection, twist angle, and power consumption. These characteristics are useful in providing comparison data in the design of the bio-inspired robot. To characterize the actuating response of the pectoral fin, three testing points (A, B, C) are defined on the membrane (shown in Fig. 4(a)).

**Figure 6.** Fabrication process of pectoral fin.

#### *2.2.1. Tip displacement*

To measure the tip deflection, the fin was actuated in a transparent tank containing water. A laser sensor (OADM 20I6441/S14F, Baumer Electric) was fixed outside of the tank to measure the bending displacement at point A. Note that due to large deflection of the fin, the laser sensor was unable to capture the bending displacement of the tip. So we moved the measurement point from the tip to the point A. The tip deflection was normalized by dividing the bending displacement by the length at the point A. Fig. 7 shows the tip deflection when a square wave actuation voltage (0.09 Hz, 4 V) was applied to the IPMC. It shows a peak-to-peak deflection of 100% in the span-wise direction. One can achieve a larger deflection by applying a higher actuation voltage. But there is a limit to the size of voltage applied across the IPMC—anything greater than 6 V risks dielectric breakdown through the IPMC (Lee et al, 2006).

**Figure 7.** Tip deflection of the pectoral fin.

Ionic Polymer-Metal Composite Artificial Muscles in Bio-Inspired Engineering Research: Underwater Propulsion 231

#### *2.2.2. Bode plot*

230 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

To measure the tip deflection, the fin was actuated in a transparent tank containing water. A laser sensor (OADM 20I6441/S14F, Baumer Electric) was fixed outside of the tank to measure the bending displacement at point A. Note that due to large deflection of the fin, the laser sensor was unable to capture the bending displacement of the tip. So we moved the measurement point from the tip to the point A. The tip deflection was normalized by dividing the bending displacement by the length at the point A. Fig. 7 shows the tip deflection when a square wave actuation voltage (0.09 Hz, 4 V) was applied to the IPMC. It shows a peak-to-peak deflection of 100% in the span-wise direction. One can achieve a larger deflection by applying a higher actuation voltage. But there is a limit to the size of voltage applied across the IPMC—anything greater than 6 V risks dielectric breakdown

0 5 10 15 20

0 5 10 15 20

Time (S)

**Figure 6.** Fabrication process of pectoral fin.

through the IPMC (Lee et al, 2006).



Voltage (V)

0

Tip deflection (%)

50

**Figure 7.** Tip deflection of the pectoral fin.

*2.2.1. Tip displacement* 

To capture the Bode plot of the pectoral fin, a series of sinusoidal actuation signals with amplitude of 4 V and frequencies ranging from 0.05 Hz to 0.9 Hz were applied to the IPMC. The tip deflection at the point B and the actuation voltage were measured. The magnitude and phase shift of the tip deflection over the actuation voltage were calculated. The Bode plot (Fig. 8) demonstrates that the actuation dynamics of the pectoral fin behaves as a lowpass filter with a 0.4 Hz cut-off frequency. This is to be expected as the ions in the IPMC cannot move very rapidly (Nemat-Nasser, 2002) and hydrodynamic force dumps high frequency vibration (Chen & Tan, 2010).

**Figure 8.** Bode plot of pectoral fin.

#### *2.2.3. Characterization of twisting motion*

In order to characterize the 3-dimensional (3D) kinematics of the fin, two laser sensors (OADM 20I6441/S14F, Baumer Electric) were used to measure the bending displacements at the points B and C, dB and dC respectively. The twist angle was calculated by

$$\alpha = \tan^{-1} \left( \frac{d\_B - d\_C}{BC} \right) \tag{1}$$

where BC=20 mm.

A series of square wave voltage signals were generated via LabVIEW (National Instruments), amplified using power amplifiers, and then applied to the IPMC actuator. All signals have the same amplitude of 4 V but varying frequencies from 0.06 Hz to 1 Hz. As the IPMC actuator is being used to generate thrust in underwater vehicles, the 3D kinematics was quantified in water. The fin was placed in a water tank and the laser sensors were fixed outside of the tank. Fig. 9 shows the twisting and flapping motion on the fin at f=0.09 Hz. The upper figure shows the bending displacements dB and dC, which indicates a phase delay

and an amplified magnitude between the bending motions at the points B and C. The lower figure shows the calculated twist angle. The peak-to-peak twist angle is 40o. Fig. 10 shows the plot of twist angle versus operating frequency. The twist angle decreases as the frequency increases. The cut-off frequency of twist angle is around 0.1 Hz. Since the twist angle plays an important role in generating a thrust, the optimal operating frequency of the robot will be set around 0.1 Hz.

**Figure 9.** Twisted flapping motion on the fin.

**Figure 10.** wist angle versus operating frequency.

#### *2.2.4. Power consumption*

For un-tethered bio-inspired robot applications, key questions regarding power consumption and optimal power management must be addressed. In this section, we study the power consumption of the pectoral fin under a square wave actuation voltage, which is easy to generate on board. The power consumed by the IPMC was calculated using the following equation:

$$P = \frac{1}{T} \int\_0^T i\left(t\right) v\left(t\right) dt\tag{2}$$

where i(t), v(t), and T are measured actuation current, voltage, and period, respectively. A 4 V, 0.1 Hz square wave voltage was applied to the IPMC of the pectoral fin, which was fixed under water. Fig. 11 shows the actuation voltage and current. The power consumption at 0.1 Hz is 1.09 W.

**Figure 11.** Actuation current and voltage under a 0.1 Hz square wave signal.

Under an electrical voltage signal, the IPMC behaves as a pseudo capacitor. When the applied voltage changes its polarization, the current reaches its peak, and then drops down to a steady-state current. The steady-state current is due to the DC resistance of the polymer and the water electrolysis on the electrodes. To reduce the power consumption of the pectoral fin, it would be better to eliminate the water electrolysis by coating a waterproof protecting layer on the IPMC, which will be our future focus. To study the relationship between power consumption and operating frequency, a series of square wave actuation signals with amplitude of 4 V and frequencies ranging from 0.1 Hz to 1 Hz were applied to the fin. Fig. 12 shows the power consumption versus operating frequency. It demonstrates that they are positively related and the power consumption is below 1.5 W.

#### **2.3. Design of robotic manta ray**

232 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

robot will be set around 0.1 Hz.

**Figure 9.** Twisted flapping motion on the fin.

Twist angle (Degree)

Twisting angle (Degree)

Displacement (mm)

**Figure 10.** wist angle versus operating frequency.

*2.2.4. Power consumption* 

and an amplified magnitude between the bending motions at the points B and C. The lower figure shows the calculated twist angle. The peak-to-peak twist angle is 40o. Fig. 10 shows the plot of twist angle versus operating frequency. The twist angle decreases as the frequency increases. The cut-off frequency of twist angle is around 0.1 Hz. Since the twist angle plays an important role in generating a thrust, the optimal operating frequency of the

<sup>0</sup> <sup>5</sup> <sup>10</sup> <sup>15</sup> <sup>20</sup> <sup>25</sup> <sup>30</sup> -20

<sup>0</sup> <sup>5</sup> <sup>10</sup> <sup>15</sup> <sup>20</sup> <sup>25</sup> <sup>30</sup> -30

Time (S)

dB dC

For un-tethered bio-inspired robot applications, key questions regarding power consumption and optimal power management must be addressed. In this section, we study the power consumption of the pectoral fin under a square wave actuation voltage, which is

10-1 <sup>100</sup> <sup>0</sup>

Frequency (Hz)

In our previous work, a free swimming and IPMC enabled robotic manta ray was developed for the first time (Chen et al, 2011). We used four IPMC beams bonded with thin PDMS membrane to create artificial pectoral fin. By independently controlling the bending of each IPMC beam, the fin can generate undulatory motions. However, in the free-swimming test of the robot, only a flapping motion was generated due to the complexity of generating the four control signals on-board. In this paper, we present an IPMC-enabled robotic manta ray capable of free swimming. Since there is only one IPMC beam in the leading edge, a single signal is generated on-board, greatly simplifying the control strategy.

**Figure 12.** Power consumption versus operating frequency.
