An Optical Actuation System and Curvature Sensor for a MR-compatible Active Needle

Abstract

A side optical actuation method is presented for a slender MR-compatible active needle. The needle includes an active region with a shape memory alloy (SMA) wire actuator, where the wire generates a contraction force when optically heated by a laser delivered though optical fibers, producing needle tip bending. A prototype, with multiple side heating spots, demonstrates twice as fast an initial response compared to fiber tip heating when 0.8 W of optical power is applied. A single-ended optical sensor with a gold reflector is also presented to measure the curvature as a function of optical transmission loss. Preliminary tests with the sensor prototype demonstrate approximately linear response and a repeatable signal, independent of the bending history.

I. INTRODUCTION

MR imaging is a very interesting technique to identify early stage cancers. Radiologists now consider its use for visualization purposes, but also to perform biopsies for diagnostics. The needles used for such procedures exhibit significant deflection during their insertion [1]. In [2], we therefore proposed to integrate an active degree of freedom in the needle to compensate for the deflections, and we introduced a new optically actuated MR-compatible active needle.

An optical actuator absorbs light and converts it into mechanical motion. It has unique features such as immunity to electric and magnetic noise sources, contact-less power transmission and a potential for integration into small and very narrow devices when light is transmitted by optical fibers. Due to these advantages, several authors have considered light-driven actuation systems for special applications. S. Lu et al. developed optically driven carbon nanotube actuators using visible light for micro manipulation [3]. S. Sarkisov et al. used films of polyvinylidene fluoride (PVDF) that bends in one direction due to a thermal expansion effect, independent from the direction of light [4]. M. Yamada et al. developed a light-driven plastic motor using UV light to induce a phase transition of liquid crystals [5]. This plastic motor can switch its rotational direction by changing the illumination direction.

Among the possible actuation methods for pairing with optical power transmission, employing a shape memory alloy (SMA) is particularly promising where a high specific force is required. In the form of a wire, it combines a large mechanical strain with a large force to volume ratio, and it presents an adequate energy conversion efficiency in tension mode [6]. K. Yamaguchi et al. developed a light-driven rotary motor with SMA wire, and showed that 0.97 % of the maximum energy conversion efficiency is possible instantly [7].

The energy reflected from the SMA surface, was more than half of the initially radiated energy. For the proposed active needle, most of the lost energy is transferred to tissues surrounding the needle, leading potentially to excessive tissue heating. Increasing the actuator efficiency is therefore of great importance both from a medical point of view, and to optimize the dynamics of the deflection compensation. In this paper, we investigate the design of a new optical heating system with higher transmission efficiency.

Compensation of the needle deflection requires estimation of the needle curvature. In a solid needle stylet, embedded strain sensors such as optical fiber Bragg gratings (FBGs) can provide an accurate measurement of the bending strain associated with curvature [8]. In the present case, due to the construction of the needle, it is more useful to measure curvature directly. M. Kovacevis et al. built an analytical model of the power loss in an optical curvature gauge for the case of a set of tooth-like indentations removed from the exterior surface of an optical fiber [9]. According to their model, the optimal length of the set of “teeth” depends on the number and the shape of the teeth. Using this approach, the magnified transmission loss for bending a 250 μm diameter plastic optical fiber (POF) is insensitive to changes in stress or temperature. However, some temperature dependence of bending loss has been reported with unmodified single mode fibers [10]. The discrepancy may be due to the fact that losses arising from controlled surface alterations dominate over other effects. Y. Fu et al. proposed a light intensity modulation curvature sensor with a multi-mode POF [11]. The sensor response presents good dynamics and linearity up to a curvature of 16.7 (1/m), as well as a polarity distinguishing positive and the negative bending.

In the present application, a drawback to the previously described optical curvature sensors is that they require an emitter at one end and a detector at the opposite end. In the case of the needle, there is not room to loop the optical fiber. Therefore, we have developed a single-ended solution, using reflected light at the distal end of the fiber.

In the following section, the design, manufacturing, and experimental evaluation of the optical heating system are introduced in section. Section III presents the optical curvature sensor and the results of testing it for linearity. Section IV presents conclusions and future work.

II. High Efficiency Optical actuation system

In the proposed active needle (Fig. 2), the main design feature is the active element located near the tip. It is composed of a SMA wire used as an actuator, a superelastic NiTi tube with a set of slits, a multi-lumen polytetrafluoroethylene (PTFE) tube placed inside the tube, and four optical fibers inserted into the dedicated lumens of the PTFE tube. Two optical fibers are used for the SMA wire heating, while the other two fibers constitute the curvature and temperature sensors. Analysis and tests of force/deflection capabilities of this design are reported in [2]. The bending capability was demonstrated to reach up to 10° when using Joule heating to activate the SMA wire. Evaluation of the optical heating requirements showed also that distributed heating would be necessary to achieve desired deflections without locally damaging the SMA wire. The new optical heating solution is presented here.

Fig. 2.

Design of active needle and the inner components

The new design conveys optical fibers through a multi-lumen tube inside the needle (Fig. 2). The tube material, PTFE, is relatively IR-transparent [12] and has adequate mechanical stiffness to maintain the alignment of the fibers and adequate thermal conductivity to promote heat transfer. The low friction makes it easy to slide fibers into the lumens and the high working temperature (above 260°C) prevents damage. In addition, the PTFE tube supports the SMA wire so that its contraction force is always parallel to the long axis of the needle.

The IR transparency of PTFE allows the light diffused by the fibers either to reach the superelastic NiTi tube or to escape through the set of slits machined on the tube. In order to solve this leaking power issue, we coat the outer surface of the multi-lumen tube with gold, except for the channel that encapsulates the SMA wire. This gold coating acts an efficient IR reflector. The light that does not immediately impinge on the SMA wire is therefore reflected and trapped, and ultimately is converted to heat within the needle (Fig. 3). With this coating, the energy conversion efficiency is improved, which improves the actuation dynamics and reduces heat transferred to the surrounding tissue.

Fig. 3.

The effect of the introduction of the proposed components for efficient optical actuation: (a) Diffused side power, (b) Gold coat reflecting optical power that would otherwise be lost

A. Optical fiber with side heating

The initial design introduced in [2] was based on the idea of side heating using tilted fiber Bragg gratings (tFBG). However, this solution suffers from low efficiency and relatively high cost. A simpler and more effective solution is to achieve controlled optical side loss over a length of the fiber.

A conventional optical fiber is composed of three layers: the protective plastic jacket, the cladding and the core. Light is confined within the core by total internal reflection (TIR) at the interface between cladding and core. TIR results from the higher refractive index of the cladding. Light loss can therefore be voluntarily introduced by removing the jacket and cladding. Fiber machining to introduce power loss consists mostly of fiber polishing [13] or chemical etching with hydrofluoric (HF) acid [14], [15]. Such techniques require extremely careful handling as the fiber becomes weak when the cladding is removed, particularly if any small cracks or defects result from the material removal process. Direct writing on the fiber using a commercial laser was also investigated as a processing method [16], but requires a relatively high power laser with a precisely focused spot.

In the present case, satisfactory results have been obtained using a glass etching cream (Armour Etch, Armour Products, Hawthorne, NJ), designed for the decoration of silica glass. The cream does not provide a consistent etching on large areas. However, it is effective for multi-mode glass fibers (MM-S105-125-22A, Nufern, East Granby, CT) that have a 125 μm cladding diameter, after removal of the the jacket with a stripper. The etched spot can be clearly visualized (Fig. 4), which is an advantage during the needle integration: the direction of the emitted light can easily be rotated toward the SMA wire.

Fig. 4.

Example of chemically etched four side heating spots on two fibers

The side-etched region inherently diffuses the laser light into the SMA wire (Fig. 3), allowing a higher optical power without damaging the wire. The total amount of contraction is proportional to the number of heating segments. Similarly to the segmented SMA actuator proposed by B. Selden et al. [17], a first way to control very simply the wire contraction could be to use a binary control of light from each of the two fibers. Two positions of the needle tip would then be obtained.

A 30-minute etching session allows the removal of about 6 μm on a fiber of 125 μm cladding diameter. The fiber core has a diameter of 105 μm. At least two etching steps are therefore necessary. A triple-etched spot or a spot longer than 2 mm allows most of the optical power to exit. This is undesirable because it limits the number of achievable emission regions along the fiber length. Finally, the manufacturing consists in a double etching of each of the two fibers, each for two locations, where each spot is less than 2 mm.

The multi-lumen tube is manufactured by the extrusion of PTFE resin (Zeus Inc, Orangeburg, SC). The 500 nm gold coating is obtained by sputtering. This multi-lumen tube is also used to stabilize and protect the fibers during the etching process. The tube is machined to partially remove its side sections, through which the etching cream is applied and later washed out, while the tube supports the optical fibers (Fig. 4). After etching, one of the fibers is pulled back slightly along the tube to stagger the four optical emission spots along the actuator length.

B. Experimental evaluation

Four different prototypes were built to investigate the effectiveness of the distributed heating and the gold coating, and to analyze the influence of the heating spot locations. Each prototype was equipped with two optical fibers, with different optical heating conditions, as summarized in Table I. The length L is the length between the SMA wire ends, approximately 22 mm.

TABLE I.

Optical heating conditions of prototypes

Prototype		1	2	3	4
PTFE tube	Gold coating	No	No	Yes	Yes
Heating spot design	Type	Tip	Side	Side	Side
	Thermal cement	Yes	No	No	No
	Number	2	4	4	8
	Length [mm]		1.5	1.5	1
Position of heating spots	First spot	1/3L	1/7L	1/7L	0
	Last spot	2/3L	6/7L	6/7L	6/7L

For a baseline comparison, prototype 1, heating is performed with the two fiber tips, instead of side heating. High temperature thermal cement was applied at the tips as a diffuser for a better heat transfer. As presented in Fig. 3, the laser from the tip is highly directional so that it is hard to illuminate the SMA wire evenly. The thermal cement, mixed with carbon powder in a ratio of 3:1, absorbs about 30 % of light while diffusing the remainder. The heat absorbed by the cement is delivered to the SMA wire through conduction and the diffused light is radiated on the wire. The thermal cement helps to prevent possible damage to the SMA wire; however it also increases the thermal mass of the system, resulting in the need for an increased power during heating and a longer cooling time.

The prototypes 2 and 3 have the same heating configurations, using side heating with two spots per fiber and with (3) or without (2) a gold coating on the PTFE tube. The influence of this coating can therefore be evaluated.

For the prototype 4, each fiber has 4 spots for a wide distribution. The first spot is located under the clamped part of the wire to take advantage of the better anchoring of this side of the wire.

Optical fibers are connected to 976 nm lasers (Alfalight Inc, Madison, WI) and 0.38 W of optical power is applied to each fiber. This amount of power was proved to be safe for 250 μm diameter SMA wire in previous experiments, even in the case of direct contact between the wire and the fiber tips. The prototypes have different maximum bending angles, but they all bend more than 5°. After reaching the maximum angle, the tip deflection often drops by about 1~2 mm, due to a release of the wire clamp on the actuator at the proximal end. To overcome this slippage at maximum force, it would be desirable to clamp or laser weld both ends of the actuator in future prototypes, while taking care to avoid stress concentrations that could reduce actuator cycle life [18].

The activations of prototypes up to 5° of tip angle are plotted in Fig. 5. Although all of the actuation times are relatively long, as a consequence of using a low optical power for the tests, the results demonstrate that distributed side heating significantly improves the response, shortening the time to reach 5° from 19 s to 8.5 s. The gold coated PTFE tube also has the desired effect on the response, reaching 5° for 7.5 s with the fastest initial movement. The efficiency of power conversion with prototype 3 is approximately 250 % that obtained with prototype 1. Note also that the tip angle, which is proportional to the actuator force, increases approximately linearly with time over the duration of the test. This indicates that we are not approaching a steady-state condition. Therefore, considerably faster response can be obtained with higher laser power before thermal conduction into the surrounding tissue becomes the dominating constraint. A simple model of the thermal mass and rate of heat loss by conduction into the surrounding tissue suggests that doubling the laser power to 0.8 W per fiber can lead to less than 4 s of activation up to 5° for the same conditions. However, this needs to be confirmed with additional tests.

Fig. 5.

Initial response of needle prototypes with different heating conditions at 0.38 W optical power per fiber

Note also that prototype 4 with the largest number of heating spots shows a slightly lower speed compared to prototype 3. The likely reason is that some of the optical power from the first spot is heating the outer NiTi tube rather than contracting the SMA wire. In addition, the slightly increased total spot length reduces optical power density, which may be not enough to rapidly initiate SMA contraction at the power level used. Further investigation is necessary.

III. Single-ended optical curvature sensor

A. Sensitive zone and Reflector

During the activation of the needle, the whole active part of the needle experiences heating. Since FBG sensors are sensitive to both temperature and strain, an FBG is integrated to evaluate the temperature, and a separate curvature measurement is used to de-correlate the influence of the strain on the measured temperature, as well as to estimate the tip bending angle.

The fundamental principle of a bending loss based optical curvature sensor is simply to have optical power loss through a pattern on the cladding and the core when the fiber is bent. The pattern is machined at the location of the required curvature measurement, called the sensitive zone (Fig. 6). In contrast to most previously reported curvature sensors based on the loss of transmitted light, for the active needle, this curvature sensor should have only one access point for sending and receiving light at the needle base.

Fig. 6.

Schematics of the proposed single-ended curvature sensor based on enhanced bending loss

Gold is a virtually perfect reflector for visible red light as well as the mid-IR region, and it can be coated on the fiber tip by sputtering. An unjacketed plastic fiber (POF) is used for an easy machining, with 250 μm of cladding diameter and 240 μm of core diameter. Preliminary tests with gold coated POF show that less than 15 % of light is lost in the coating, and the reflected light intensity is high enough to be measured at the base. A 2-to-1 optical coupler is used to send and receive light simultaneously from the single embedded POF.

B. Sensor prototyping

A customized fiber slitting tool was built using a fixed blade and micro-positioning stages (Fig. 7). The POF (SK-10, Mitsubishi Yayon Co.LTD., Japan) is embedded on a grooved wax block during the slitting. The resulting slit depth is controlled with an accuracy of 60 μm. The optimal number of slits and distance between slits were determined through experimentation, resulting in eight slits with a spacing of 300 μm. The set of slits begins 10 mm from the needle tip, centered on the active element. If the fiber is used immediately after the machining of the slits, light is reflected in an unpredictable way, causing noisy signal. Thus, the small particles, by-products of the slitting process, are rinsed. The fiber tip is cut and polished initially with 10 μm grit sandpaper and then with 1 μm abrasive paper. After the polishing, 500 nm of gold is deposited by sputtering. A thicker coating of 1 μm might be preferable because of the maximum roughness of the polished tip. However, a thicker coating is more easily peeled from the tip.

Fig. 7.

Preparation of the POF for the curvature sensor

C. Experimental evaluation

The prototype of optical curvature sensor is integrated with a 150 mm long 18G biopsy needle for the evaluation. A red LED and a phototransistor (IF-E96-R and IF-D93 respectively, Industrial Fiberoptics, Tempe, AZ) are connected to the machined POF through a 50:50 power ratio, 2-to-1 optical coupler (IF 562, Industrial Fiberoptics, Tempe, AZ). The sensitive zone of the prototype is placed at the needle base. The voltage signal obtained from the reflected light for the various tip deflections is compared with the local curvature for the same tip deflection, estimated using Finite Element Analysis with Ansys software. The results are plotted in Fig. 8. A linear relationship is obtained for curvatures up to about 5 (1/m), that corresponds to the 5° of active needle tip bending. Linearity should be obtained for larger curvature ranges with an optimized pattern design.

Fig. 8.

Curvature sensor test: (a) voltage response to curvature change (b) signal for a series of step loads

Another experiment is performed by manually applying successive deflection steps equal to 1 cm (Fig. 8b). One can see the characteristic pattern for step loading, with some relaxation after each step. The reflected voltage signal is independent from the time history of the loading, though there is a slight drift of the signal. This may be due to the clearance between the plastic fiber and the lumen in the PTFE tubing.

IV. Conclusion

In this paper, a new distributed optical heating method using side-etched optical fibers is presented. This optical heating system, including a gold coated multi-lumen tube, is integrated with an MR-compatible active needle prototype. The results demonstrate a significant improvement in the efficiency of optical power conversion to SMA actuation over previous prototypes. The ability to distribute the heat along the wire opens the possibility to apply considerably higher power for better dynamic performance. Considering that these achievements are obtained from a prototype using a relatively high temperature SMA wire with 70°C of austenite transition, the proposed design of the optical heating system should be quite acceptable for the future active needle prototypes using a lower temperature SMA wire and higher optical power.

A new single-ended optical curvature sensor based on bending loss is also presented. The sensor is only sensitive to the local curvature. Its combination with FBGs allows us therefore to measure continuously the temperature and the curvature of the device. With the current prototype, the sensor response presents a good linearity up to 5° and a good repeatability. Further investigation will be focused on widening the linear range and stabilizing the voltage signal. Other issues related to integration in needle prototype, such as the power interference between the sensing light and the heating laser, will be examined.

Further improvement should be obtained optimizing the design of the optical heating spots and the pattern of curvature sensors. The investigation of the needle dynamics in tissue phantoms will be conducted after integrating the optimized version of the system with the temperature sensor. Modeling of the active needle for its control with low temperature SMA wire will then be studied.

Fig. 1.

A functional prototype of active element with optical actuation capability