Side-opened out-of-plane microneedles for microfluidic ... - IEEE Xplore

296

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 12, NO. 3, JUNE 2003

Side-Opened Out-of-Plane Microneedles for Microfluidic Transdermal Liquid Transfer Patrick Griss, Member, IEEE, and Göran Stemme, Member, IEEE

Abstract—We present the first hollow out-of-wafer-plane silicon microneedles having openings in the shaft rather than having an orifice at the tip. These structures are well suited for transdermal microfluidic applications, e.g., drug or vaccine delivery. The developed deep-reactive ion etching (DRIE) process allows fabrication of two dimensional, mechanically highly resistant needle arrays offering low resistance to liquid flows and a large exposure area between the fluid and the tissue. The presented process does not require much wafer handling and only two photolithography steps are required. Using a 3 3 mm2 chip in a typical application, e.g., vaccine delivery, a 100 l volume of aqueous fluid injected in 2 s would cause a pressure drop of less than 2 kPa. The presented nee[786] dles are approximately 210 m long. Index Terms—Deep-reactive ion etching (DRIE), delivery, drug, microneedle, transdermal, vaccine.

I. INTRODUCTION

T

HE OUTER-MOST skin layer, i.e., the Stratum Corneum (SC), is perhaps the most versatile biological barrier in the human body. It is an excellent electrical insulator and it prevents the uptake of infectious agents while restricting water loss [1]. The delivery of small amounts of liquids through the SC of humans into the underlying tissue or the sampling of fluids from the underlying tissue is becoming increasingly important in biomedical applications. Microsystem Technology provides means for the fabrication of microscaled liquid transfer needles, i.e., microneedles N . In the last few years, activity in the N field has been steadily growing. Due to their small dimensions, they can be inserted into the skin painlessly and cause less tissue damage than conventional hypodermic needles. Microneedles have the potential to become the preferred drug delivery device in applications where the transdermal aspect is essential. For example, biotechnology has produced a generation of novel compounds with great therapeutic promise that generally consist of active macromolecules, e.g., proteins. Their oral administration is complicated and the passive diffusion of those compounds across the skin is not a realistic option. Different hollow out-of-plane N for transdermal applications have been presented before. They are arranged in two-dimensional arrays to decrease flow resistance through the device [2]–[6]. The array can be achieved with wafer level processing. The openings are at the top of the needle, which increases the risk for clogging. Manuscript received December 6, 2001; revised September 28, 2002. Subject Editor R. T. Howe. The authors are with the Department of Signals, Sensors and Systems, Microsystem Technology, Royal Institute of Technology, 100 44 Stockholm, Sweden (e-mail: [email protected]). Digital Object Identifier 10.1109/JMEMS.2003.809959

In-plane N have been developed earlier and are characterized by the opening at the shaft of the needle and are less prone to clogging [7]–[13]. These needles are generally longer than out-of-plane needles. The fabrication of two-dimensional arrays is more difficult to achieve since it cannot be done on wafer level. Our own group, in collaboration with Datex-Ohmeda (a division of Instrumentarium Corp.), have reported on solid silicon N arrays successfully used for biopotential measurements [14]. The mechanical strength of those N arrays was observed to be surprisingly high, in particular during measurements of the activity of the brain where the arrays were applied on the forehead of test subjects. The mechanical strength of barbed N was also observed when measuring the attachment force of their arrays pressed into different types of materials [15]. Very low failure rate is a requirement for a micromachined N device to be used in commercial applications. In the case of hollow N designed for transdermal liquid transfer, they must be robust enough to penetrate biological tissue and withstand harsh treatment. Coating in-plane single crystalline silicon N with Parylene provides a way to prevent catastrophic failure [12]. This allowed the retraction of N from pierced gelatin membranes, even if the silicon core is fractured. Two-dimensional needle arrays are less prone to fracturing when exposed to shear forces during penetration than single needles of the same material and dimensions since the shear stress created by the tissue is distributed over a large amount of N. The goal of the presented work was to develop a micromachined structure that has the potential to be used in transdermal fluidic applications due to its low flow resistance, high structural strength, large area of drug exposure to the tissue and low risk of clogging. These considerations led to a novel type of out-of-plane N array where the N have openings at the side of the needle rather than at the top, as depicted in the conceptual drawing of Fig. 1. Therefore, when pressed into the tissue, the sharp N tip sections the tissue rather than stamping out a piece of it. The size of the side openings can be controlled through process parameters. The area of drug exposure is increased for a side-opened needle when compared to a tip-opened one, given the same diameter of the liquid channel in the needle. In the following we present and discuss the wafer level fabrication process for two different types of side-opened N arrays as well as assembly into a package allowing fluidic flow measurements and application on human skin. Subsequently the mechanical stability is studied and the flow-pressure characteristics is measured and discussed.

1057-7157/03$17.00 © 2003 IEEE

GRISS AND STEMME: SIDE-OPENED OUT-OF-PLANE MICRONEEDLES FOR LIQUID TRANSFER

297

Fig. 1. Concept of the side-opened out-of-wafer-plane microneedle compared to tip opened ones.

(a)

(b)

(c)

Fig. 3. Process flow showing a top view and a cut through the needle along the A-A line. Refer to the text for details.

Fig. 2. (a) Vertical walls of DRIE etched high aspect ratio silicon structures stay vertical during an isotropic plasma etch. (b) If a silicon dioxide mask is underetched and subsequently anisotropically etched, the resulting section of the structure corresponds to the mask. (c) Basic processing principle yielding side openings in high Aspect ratio structures (using DRIE technology).

II. EXPERIMENTAL A. Needle Design and Fabrication The fabrication of side-opened N is based on the triple DRIE process shown earlier by our group where it was observed that in this process vertical walls of DRIE etched high aspect ratio silicon structures stay vertical during an isotropic plasma etch [14], [15], as shown in Fig. 2(a). Further, it was observed that if a silicon dioxide mask is underetched and subsequently anisotropically etched, the resulting cross section of the structure corresponds to the mask, see Fig. 2(b). Combining these observations with a vertical high aspect ratio hole from the back side, a DRIE-based process for the fabrication of side-opened out-of-plane N was established. A simplified artistic drawing of the process principle is shown in Fig. 2(c).

Fig. 4. Photograph of the assembly used for flow-pressure measurements.

The detailed process flow yielding side-opened N is depicted in Fig. 3. A circular high aspect ratio hole is etched into the back side of the wafer using a silicon dioxide SiO mask in an inductively coupled plasma etcher (ICP) [see Fig. 3(a) and (b)]. The anisotropic etching is based on the Bosch process [16]. The hole serves as a liquid channel connecting the back

298


(a)

(a)

(b)

(b)

Fig. 5. SEM images of side-opened microneedles, the hole beginning at the base of the needle. The length of the structure is 210 m.

side of the chip to the front side. After wet oxidation, a SiO cross-shaped mask is aligned to the hole on the front side of the wafer. The diameter of the hole is smaller than the diagonal dimension of the cross [see Fig. 3(b)]. A first isotropic ICP step underetches the SiO front mask [see Fig. 3(c)] and is followed by an anisotropic ICP step, which creates a cross-shaped out-of-plane structure without side openings [see Fig. 3(d)]. The subsequent isotropic etch decreases the cross sectional area of the structure without altering the angle of the sidewalls, thus creating side openings in the walls that are still closed by a thin SiO membrane [see Fig. 3(e1)]. This etch also sharpens the four pillars of the cross-shaped structure, each pillar having a knife like edge at the top. This step is stopped before the mask is completely underetched at the center. A complete underetch of the mask would destroy the structure since the mask would fall off and probably stick to the sidewall. If side-openings that start at the base of the needle are desired, no additional plasma etch is required [see Fig. 3(e1)]. If it is desired that the N has a part where there are no side slits, another anisotropic plasma etch can be performed, which will result in a side hole placed above the needle base. This anisotropic etch step creates

Fig. 6. SEM images of side-opened microneedles, the hole beginning approximately 50 m above the base of the needle. The length of the structure is 210 m.

a self-aligned pillar structure on top of which the side-opened structure is placed [see Fig. 3(f1)]. The top mask can be removed by a final wet oxidation followed by a SiO HF-strip [see Fig. 3(e2) or (f2)]. The oxidation growth and removal also sharpens the tip apex of the needle. Process steps c to e1 or f1, respectively, are executed in one load of the ICP machine, thus the total process is uncomplicated and does not require much wafer handling. Only two photolithography steps are required to yield a relatively complex three-dimensional microstructure. B. Packaging/Assembly The measurement of the flow resistance requires assembly of the chip containing a side opened N array onto a carrier which allows a connection to fluid tubing, as shown in Fig. 4. The carrier is made of brass and was manufactured using conventional mm chip is fixed to the carrier machining methods. The by means of ultraviolet light curing epoxy (Epotek OG 198). The square geometry of the chip and the circular geometry of the through hole of the carrier results in the blockage of some N at the chip corner. Twenty-one N are not blocked and contribute to the flow through the device.


Fig. 7. Artistic drawing of the microneedles presented in this work. The position of the side opening is defined by process parameters. a) For a given hole and needle mask, the width of the side opening t as well as the position of the side opening (i.e., distance s from the needle tip and distance r above the base) are defined by process parameters. b) If it is desired to start the side opening at the base of the needle, r can be chosen to be zero.

299

Fig. 9. Measurement and calculation of the pressure drop over the chip caused by water flow.

Fig. 10. Penetration of a 10 m thick aluminum foil by side-opened microneedles. Note that no damage can be observed on at the needle.

Fig. 8. SEM photograph of a side-opened microneedle before the removal of the silicon dioxide front side mask. Where the side opening is located, silicon dioxide is visible and not removed yet.

III. RESULTS AND DISCUSSION Figs. 5 and 6 depict scanning electron microscope (SEM) images of two types of side-opened silicon N. In Fig. 5, the side opening starts at the base of the needle, whereas in Fig. 6 the N has a part at the base where there is no side opening, i.e., the opening starts above the needle. This feature can be important to prevent leakage of the liquid when the N is inserted into the skin. Both needle types can be achieved with the same mask. As shown in Fig. 7, for a given mask, the width of the side opening as well as the position of the side opening (i.e., distance from the needle tip and distance above the base) are defined by process parameters. This allows for great freedom of design and enables the fabrication of N optimized for a specific application. Fig. 8 depicts a side-opened N before mask removal and sharpening of the apex. A membrane consisting of SiO still covers the side opening. Gravimetric flow measurements resulted in the pressure-flow characteristics as shown in Fig. 9. The depicted characteristics are those of a N array (21 needles) and not of a single needle. Theoretical calculations of the flow characteristics are shown in

the same figure and take into account the viscous shear force of the Poiseuille flow inside a circular tube and the inertia efacross the channel is fects [3], [17]. The total pressure drop due to laminar friction (i.e., the sum of the pressure drop required to accelerate the Poiseuille) and the pressure drop liquid. For a tubular liquid channel this is calculated according to [17]:

where is the viscosity of the liquid, the flow generating the , the density of the liquid, is the radius of pressure drop is a numerical the channel and the length of the channel. factor and in this case 1.2. In a typical transdermal application, e.g., vaccine delivery, a 100 l volume of aqueous fluid injected through a chip containing 21 side-opened N in 2s would cause a pressure drop of less than 2 kPa. The flow resistance can further be decreased by increasing the number of N (i.e., the number of N per area unit). Since an anisotropic etch mainly defines the length of the N, very high density can be achieved since the mask is only slightly larger than the resulting N. Therefore, the maximum needle density which is inherently limited by the capability of the skin to be permeated can be achieved without being restricted by technology.

300


Fig. 11. Artistic drawing of two side-opened microneedles that can be manufactured using the presented process. a) has one single side opening and b) has four side openings. The primary difference between these two microneedles is the hole mask which defines the channel in the N.

If aqueous liquid is presented to the back side of the chip without applying a pressure difference between the front side and the back side, the liquid is sucked into the channels in the chip base by capillary forces. The liquid meniscus is stopped at the side openings without wetting the front side. A pressure of approximately 1 kPa was measured to break through this barrier. Fig. 10 exemplifies the mechanical stability of the obtained microneedles. A 10 m thick aluminum foil is penetrated by a N having side openings without breaking. In this figure, the side openings start at the base of the N. This type is theoretically more fragile than those where the side openings start above the needle base. Note that the shown structure was not oxidized long enough to yield a sharp apex, in contrast to the one shown in Fig. 5. It was observed that assembled side-opened N penetrate into human skin without causing pain. The application site was the inner forearm. The mechanical stability is excellent, the needle arrays can be inserted and removed from human skin repeatedly without breaking. Several other designs of out-of-plane side-opened N can be imagined. Using a circular needle mask and a cross shaped hole mask, a circular needle with four side holes result as depicted in Fig. 11(b). Two uncentered circular masks will result in a circular needle featuring one single side opening, as exemplified in Fig. 11(a). Further work is required to explore the full potential of the presented manufacturing method.

IV. CONCLUSION We demonstrate a new technology to fabricate arrays of hollow out-of-plane microneedles that have openings at the shaft rather than at the tip apex. The size and position of the side openings are defined by process parameters and not by the specific mask design.

Such needles allow new opportunities for transdermal liquid transfer. The measured flow resistance of a packaged side-opened needle array was measured to be low (and can further be decreased if needed by increasing the needle density). The mechanical strength of the needle arrays was found to be high. Subsequent penetration and removal to and from the skin did not result in the destruction of the needles. The mechanical strength is also demonstrated by the ability to pierce aluminum. Potentially, the shown structures are less prone to clogging than tip-opened counterparts and the large size of the side openings allow a large area of liquid exposure to the skin. ACKNOWLEDGMENT The authors would like to thank MicroJoining AB (Tyresö, Sweden) for kindly providing UV-curing epoxy, Kjell Norén for his invaluable help during packaging and Wouter van der Wijngaart for the support during flow measurements. REFERENCES [1] Y. N. Kalia, V. Merino, and R. H. Guy, “Transdermal drug delivery,” in Dermatol. Clin., Apr. 1998, vol. 16. [2] D. V. McAllister, F. Cros, S. P. Davis, L. M. Matta, M. R. Prausnitz, and M. G. Allen, “Three-dimensional hollow microneedle and microtube arrays,” in Proc. 10th International Conference on Solid-State Sensors and Actuators, Sendai, Japan, 1999. [3] B. Stoeber and D. Liepman, “Fluid injection through out-of-plane microneedles,” in Proc. 1st Annual International IEEE-EMBS Special Topic Conference on Microtechnologies in Medicine & Biology, Lyon, France, 2000. [4] A. E. Guber, H. D. Dittrich, M. Heckele, D. Herrmann, A. Musilija, W. Pfleging, and Th. Schaller, “Polymer micro needles with through-going capillaries,” in Micro Total Analysis Systems 2001, J. M. Ramsey and A. van den Berg, Eds., pp. 155–156. [5] E. Mukerjee, S. D. Collins, R. L. Smith, and R. Isseroff, “Microneedle array for transermal bio-fluid sampling and drug delivery,” in Micro Total Analysis Systems 2001, J. M. Ramsey and A. van den Berg, Eds., pp. 379–380.


[6] A. van den Berg, “Integrated micro- and nanofluidics: silicon revisited,” in Micro Total Analysis Systems 2001, J. M. Ramsey and A. van den Berg, Eds., pp. 207–209. [7] J. Chen, K. D. Wise, J. F. Hetke, and S. C. Bledsoe Jr., “A multichannel neural probe for selective chemical delivery at the cellular level,” IEEE Trans. Biomed. Eng., vol. 44, Aug. 1997. [8] N. H. Talbot and A. P. Pisano, “Polymolding: two wafer polysilicon micromolding of closed-flow passages for microneedles and microfluidic devices,” in Soli-State Sensor and Actuator Workshop, Hilton Head Island, SC, 1998. [9] K. S. Lebouitz and A. P. Pisano, “Microneedles and micro-lancets fabricated using SOI wafers and isotropic etching,” in Meet. Electrochem. Soc., Boston, MA, 1998. [10] L. Lin and A. P. Pisano, “Silicon-processed microneedles,” J. Microelectromech. Syst., vol. 8, Mar. 1999. [11] J. D. Brazzle, I. Papautsky, and A. B. Frazier, “Micromachined needle array for drug delivery or fluid extraction,” IEEE Eng. Med. Biol., Nov./Dec. 1999. [12] P. A. Stupar and A. P. Pisano, “Silicon, parylene, and silicon/parylene microneedles for strength amd toughness,” in Proc. 11th Int. Conf. on Solid-State Sensors and Actuators, Munich, Germany, 2001. [13] K. Oka, S. Aoyagi, Y. Isono, G. Hashiguchi, and H. Fujita, “Fabrication of a micro needle for a trace blood test,” in Proc. 11th Int. Conf. on Solid-State Sensors and Actuators, Munich, Germany, 2001. [14] P. Griss, P. Enoksson, H. K. Tolvanen-Laakso, P. Meriläinen, S. Ollmar, and G. Stemme, “Micromachined electrodes for biopotential measurements,” J. Microelectromech. Syst., vol. 10, pp. 10–16, Mar. 2001. [15] P. Griss, P. Enoksson, and G. Stemme, “Micromachined barbed spikes for mechanical chip attachment,” Sens. Actuators: A Phys., to be published. [16] F. Larmer and A. Schilp, German patent DE4 241 045. [17] M. Richter and P. Woias, “Microchannels for applications in liquid dosing and flow-rate measurements,” Sens. Actuators: A Phys., vol. 62, 1997.

301

Patrick Griss (M’01) was born 1973 in Lucerne, Switzerland. He received the M.Sc. degree in microengineering from the Swiss Federal Institute of Technology in Lausanne (EPFL), Switzerland, in 1999 and the Ph.D. degree in electrical engineering from the Royal Institute of Technology in Stockholm, Sweden, in 2002. His research is focused on the application of microdevices to the biomedical field, including transdermal interfaces and biochemical lab-on-chip. He is currently a Postdoctoral Fellow at the Roche Diagnostics Microtechnology Center, Rotkreuz, Switzerland. Dr. Griss won, together with two colleagues, the final of the Swedish Innovation Cup in 2001.

Göran Stemme (M’98) received the M.Sc. degree in electrical engineering and the Ph.D. degree in solid state electronics from the Chalmers University of Technology, Gothenburg, Sweden, in 1981 and 1987, respectively. In 1981, he joined the Department of Solid State Electronics, Chalmers University of Technology, Gothenburg, Sweden. In 1990, he became an Associate Professor (Docent) heading the silicon sensor research group. In 1991, he was appointed a Professor at The Royal Institute of Technology, Stockholm, Sweden, where he heads the Microsystem Technology group at the department of Signals, Sensors and Systems. His research is devoted to microsystem technology based on micromachining of silicon. He has published more than 80 research journal and conference papers and has been awarded eight patents. Dr. Stemme was a member of the International Steering Committee of the Conference series IEEE Microelectromechanical Systems (MEMS) between 1995 and 2001 and he was General Co-Chair of that conference in 1998. He is a member of the Editorial Board of the IEEE/ASME JOURNAL OF MICROELECTROMECHNICAL SYSTEMS and of the Royal Society of Chemistry journal Lab On A Chip.In 2001, he won, together with two colleagues, the final of the Innovation Cup.